MANUAL OF

ENVIRONMENTAL
MICROBIOLOGY
F O U R T H E D I T I O N
Editor in Chief Marylynn V.Yates
Editors
Cindy H. Nakatsu | Robert V. Miller | Suresh D. Pillai
 Copyright © 2016 American Society for Microbiology. All rights reserved. No part of this publication
may be reproduced or transmitted in whole or in part or reused in any form or by any means, electronic or
mechanical, including photocopying and recording, or by any information storage and retrieval system,
without permission in writing from the publisher.

Disclaimer: To the best of the publisher’s knowledge, this publication provides information concerning
the subject matter covered that is accurate as of the date of publication. The publisher is not providing legal,
medical, or other professional services. Any reference herein to any specific commercial products,
procedures, or services by trade name, trademark, manufacturer, or otherwise does not constitute or imply
endorsement, recommendation, or favored status by the American Society for Microbiology (ASM).
The views and opinions of the author(s) expressed in this publication do not necessarily state or reflect
those of ASM, and they shall not be used to advertise or endorse any product.

Library of Congress Cataloging-in-Publication Data

Names: Yates, M. V. (Marylynn V.), editor. | Nakatsu, Cindy H., editor. | Miller, Robert V. (Robert Verne),
1945- editor. | Pillai, Suresh D., 1962- editor.
Title: Manual of environmental microbiology / editor in chief, Marylynn V. Yates; editors,
Cindy H. Nakatsu, Robert V. Miller, Suresh D. Pillai.
Description: Fourth edition. | Washington, DC : ASM Press, [2016] |
Includes bibliographical references and index.
Identifiers: LCCN 2016014816 ( print) | LCCN 2016016986 (ebook) |
ISBN 9781555816025 (hardcover) | ISBN 9781555818821 ()
Subjects: LCSH: Microbial ecology–Laboratory manuals. | Sanitary microbiology–Laboratory manuals.
Classification: LCC QR100 .M36 2016 ( print) | LCC QR100 (ebook) | DDC 577.8–dc23
LC record available at https://lccn.loc.gov/2016014816

All Rights Reserved

Printed in Canada

10 9 8 7 6 5 4 3 2 1

Address editorial correspondence to

ASM Press, 1752 N St., N.W.,
Washington, DC 20036-2904, USA

Send orders to ASM Press, P.O. Box 605, Herndon, VA 20172, USA
Phone: 800-546-2416; 703-661-1593
Fax: 703-661-1501
E-mail: books@asmusa.org
Online: http://www.asmscience.org
 Contents

Editorial Board ix 2.3.2 PCR, Real-Time PCR, Digital PCR, and

Isothermal Amplification / 2.3.2-1
Contributors xi RACHEL A. BARTHOLOMEW, JANINE
R. HUTCHISON, TIMOTHY M. STRAUB,
AND DOUGLAS R. CALL
1.1.1 Introduction / 1.1.1-1 2.3.3 Microarray-Based Environmental
MARYLYNN V. YATES
Diagnostics / 2.3.3-1
DARRELL P. CHANDLER
2.3.4 Field Application of Pathogen Detection
GENERAL METHODOLOGY Technologies / 2.3.4-1
TIMOTHY M. STRAUB, DOUGLAS
VOLUME EDITOR: SURESH D. PILLAI R. CALL, CINDY BRUCKNER-LEA,
HEATHER COLBURN, CHERYL L. BAIRD,
SECTION EDITORS: YOICHI KAMAGATA, RACHEL A. BARTHOLOMEW,
CLEBER C. OUVERNEY, DOUGLAS R. CALL, RICHARD OZANICH, AND
STEFAN J. GREEN, YILDIZ T. CHAMBERS, AND KRISTIN JARMAN
JOHN SCOTT MESCHKE
2.1 CULTURE-BASED AND PHYSIOLOGICAL 2.4 MICROBIAL COMMUNITY ANALYSIS OF
DETECTION ENVIRONMENTAL SAMPLES WITH
2.1.1 Detection of Specific Taxa Using NEXT-GENERATION SEQUENCING
Chromogenic and Fluorogenic 2.4.1 Introduction to Microbial
Media / 2.1.1-1 Community Analysis of Environmental
MOHAMMAD MANAFI Samples with Next-Generation
2.1.2 Anaerobic Cultivation / 2.1.2-1 Sequencing / 2.4.1-1
TAKASHI NARIHIRO AND YOICHI STEFAN J. GREEN AND JOSH D. NEUFELD
KAMAGATA
2.4.2 Microbial Community Analysis Using
2.1.3 New Devices for Cultivation / 2.1.3-1 High-Throughput Amplicon
YOSHITERU AOI AND SLAVA EPSTEIN Sequencing / 2.4.2-1
2.2 MICROSCOPIC METHODS DANNY IONESCU, WILL A. OVERHOLT,
MICHAEL D. J. LYNCH, JOSH D. NEUFELD,
2.2.1 Gold-Based In Situ Hybridization for ANKUR NAQIB, AND STEFAN J. GREEN
Phylogenetic Single-Cell Detection of
Prokaryotes in Environmental 2.4.3 Functional Metagenomics: Procedures and
Samples / 2.2.1-1 Progress / 2.4.3-1
THILO EICKHORST AND HANNES SCHMIDT LAURA S. MORRIS AND JULIAN
R. MARCHESI
2.2.2 Assessment of Prokaryotic Biological Activity
at the Single-Cell Level by Combining 2.4.4 Metagenomics: Assigning Functional
Microautoradiography and Fluorescence Status to Community Gene
in situ Hybridization / 2.2.2-1 Content / 2.4.4-1
CLEBER C. OUVERNEY NASEER SANGWAN AND RUP LAL
2.4.5 Generation and Analysis of Microbial
2.3 TARGET-SPECIFIC DETECTION Metatranscriptomes / 2.4.5-1
2.3.1 Antibody-Based Technologies for NEHA SARODE, DARREN J. PARRIS, SANGITA
Environmental Biodetection / 2.3.1-1 GANESH, SHERRY L. SESTON, AND FRANK
CHERYL L. BAIRD AND SUSAN M. VARNUM J. STEWART

v
vi ▪ CONTENTS

2.5 QA/QC IN ENVIRONMENTAL 3.1.2 Best Practices for Cyanobacterial Harmful

MICROBIOLOGY Algal Bloom Monitoring / 3.1.2-1
TIMOTHY G. OTTEN AND HANS W. PAERL
2.5.1 Introduction to Principles of Quality
Assurance / 2.5.1-1 3.1.3 Assessing the Efficiency of Wastewater
KEVIN K. CONNELL Treatment / 3.1.3-1
GRACIELA RAMÍREZ TORO AND HARVEY
2.5.2 General Quality Control / 2.5.2-1 MINNIGH
ROBIN K. OSHIRO
3.1.4 Epidemiologic Aspects of Waterborne
2.5.3 Quality Control for Bacteriological Infectious Disease / 3.1.4-1
Analyses / 2.5.3-1 SAMUEL DOREVITCH
ELLEN BRAUN-HOWLAND
3.1.5 Waterborne Enteric Viruses: Diversity,
2.5.4 Quality Control for Virological Distribution, and Detection / 3.1.5-1
Analyses / 2.5.4-1 MORTEZA ABBASZADEGAN AND ABSAR
RICHARD E. DANIELSON ALUM
2.5.5 Quality Control for USEPA Method 1623 3.1.6 Detection of Protozoa in Surface and Finished
Protozoan Analysis and PCR Waters / 3.1.6-1
Analyses / 2.5.5-1 ABSAR ALUM, ERIC N. VILLEGAS, SCOTT
GEORGE D. DI GIOVANNI AND GREGORY P. KEELY, KELLY R. BRIGHT, LAURA
D. STURBAUM Y. SIFUENTES, AND MORTEZA
2.5.6 The Role of Statistical Thinking in ABBASZADEGAN
Environmental Microbiology / 2.5.6-1 3.1.7 Drinking Water Microbiology / 3.1.7-1
J. VAUN MCARTHUR AND R. CARY MARYLYNN V. YATES
TUCKFIELD
2.5.7 Study Design / 2.5.7-1
3.2 AIR
YILDIZ T. CHAMBERS AND ROBIN K. OSHIRO 3.2.1 Introduction to Aerobiology / 3.2.1-1
PAULA KRAUTER AND LINDA
2.6 SAMPLING METHODS D. STETZENBACH
2.6.1 Water Sampling and Processing 3.2.2 Sampling for Airborne
Techniques for Public Health-Related Microorganisms / 3.2.2-1
Microbes / 2.6.1-1 SERGEY A. GRINSHPUN, MARK P. BUTTNER,
VINCENT HILL GEDIMINAS MAINELIS, AND KLAUS
2.6.2 Surface Sampling / 2.6.2-1 WILLEKE
LAURA J. ROSE, JUDITH NOBLE-WANG, AND 3.2.3 Analysis of Bioaerosol Samples / 3.2.3-1
MATTHEW J. ARDUINO PATRICIA CRUZ AND MARK P. BUTTNER
2.6.3 Soil Sampling for Microbial Analyses / 2.6.3-1 3.2.4 Fate and Transport of Microorganisms in
JOHN BROOKS Air / 3.2.4-1
2.6.4 Microbiological Sampling of Wastewater and GARY S. BROWN AND ALAN JEFF MOHR
Biosolids / 2.6.4-1 3.2.5 Airborne Fungi and Mycotoxins / 3.2.5-1
NICOLETTE A. ZHOU, ERIC C. THOMPSON, DE-WEI LI, ECKARDT JOHANNING, AND CHIN
AND JOHN SCOTT MESCHKE S. YANG
3.2.6 Airborne Bacteria, Archaea, and
Endotoxin / 3.2.6-1
PETER S. THORNE, CAROLINE DUCHAINE,
ENVIRONMENTAL PUBLIC HEALTH AND PASCALE BLAIS LECOURS
3.2.7 Airborne Viruses / 3.2.7-1
MICROBIOLOGY SYED A. SATTAR, NITIN BHARDWAJ, AND
VOLUME EDITOR: MARYLYNN V. YATES M. KHALID IJAZ
SECTION EDITORS: GARY A. TORANZOS, MARK 3.2.8 Aerobiology of Agricultural
P. BUTTNER, ED TOPP, VALERIE J. HARWOOD, Pathogens / 3.2.8-1
AND MARYLYNN V. YATES ESTELLE LEVETIN
3.2.9 Legionellae and Legionnaires’
3.1 WATER Disease / 3.2.9-1
3.1.1 Current and Developing Methods for the CLARESSA E. LUCAS AND BARRY S. FIELDS
Detection of Microbial Indicators in
Environmental Freshwaters and Drinking 3.3 SOIL
Waters / 3.1.1-1 3.3.1 Pathogenic Viruses and Protozoa Transmitted
TASHA M. SANTIAGO-RODRIGUEZ, by Soil / 3.3.1-1
JULIE KINZELMAN, AND GARY A. PASCAL DELAQUIS, JULIE BRASSARD, AND
TORANZOS ALVIN GAJADHAR

3.3.2 Natural Soil Reservoirs for Human Pathogenic
and Fecal Indicator Bacteria / 3.3.2-1
MARIA LAURA BOSCHIROLI, JOSEPH
FALKINHAM, SABINE FAVRE-BONTÉ, SYLVIE
NAZARET, PASCAL PIVETEAU, MICHAEL
SADOWSKY, MURULEE BYAPPANAHALLI,
PASCAL DELAQUIS, AND ALAIN
HARTMANN
3.4 MICROBIAL SOURCE TRACKING
3.4.1 The Evolving Science of Microbial Source
Tracking / 3.4.1-1
VALERIE J. HARWOOD, CHARLES HAGEDORN,
AND MICHAEL SADOWSKY
3.4.2 Validation of Microbial Source Tracking
Markers and Detection Protocols:
Considerations for Effective
Interpretation / 3.4.2-1
ASJA KORAJKIC, DON STOECKEL, AND JOHN
F. GRIFFITH
3.4.3 Overview of Microbial Source Tracking
Methods Targeting Human Fecal Pollution
Sources / 3.4.3-1
ORIN C. SHANKS, HYATT GREEN, ASJA
KORAJKIC, AND KATHARINE G. FIELD
3.4.4 Methods of Targeting Animal Sources of Fecal
Pollution in Water / 3.4.4-1
ANICET R. BLANCH, ELISENDA BALLESTÉ,
JENNIFER WEIDHAAS, JORGE SANTO
DOMINGO, AND HODON RYU
3.4.5 Microbial Source Tracking: Field Study
Planning and Implementation / 3.4.5-1
JULIE KINZELMAN AND WARISH AHMED
3.4.6 Fecal Indicator Organism Modeling and
Microbial Source Tracking in Environmental
Waters / 3.4.6-1
MEREDITH B. NEVERS, MURULEEDHARA
N. BYAPPANAHALLI, MANTHA
S. PHANIKUMAR, AND RICHARD L. WHITMAN
3.5 MICROBIAL RISK ASSESSMENT
3.5.1 Risk Assessment Framework / 3.5.1-1
MARYLYNN V. YATES
3.5.2 Exposure Assessment / 3.5.2-1
SUSAN R. PETTERSON AND NICHOLAS
J. ASHBOLT
3.5.3 Dose-Response Modeling and Use: Challenges
and Uncertainties in Environmental
Exposure / 3.5.3-1
MARK H. WEIR
MICROBIAL ECOLOGY
VOLUME EDITOR: ROBERT V. MILLER
SECTION EDITORS: LARRY J. FORNEY, ROBERT
H. FINDLAY, BRIAN P. HEDLUND, AND JULIAN
R. MARCHESI
4.1 THEORY
CONTENTS ▪ vii
4.1.1 Phylogenomic Networks of Microbial
Genome Evolution / 4.1.1-1
TAL DAGAN, OVIDIU POPA, THORSTEN
KLÖSGES, AND GIDDY LANDAN
4.1.2 Evolutionary Ecology of Microorganisms:
From the Tamed to the Wild / 4.1.2-1
JAY T. LENNON AND VINCENT J. DENEF
4.2 AQUATIC ENVIRONMENTS
4.2.1 The Microbial Ecology of Benthic
Environments / 4.2.1-1
ROBERT H. FINDLAY AND TOM J. BATTIN
4.2.2 Heterotrophic Planktonic Microbes: Viruses,
Bacteria, Archaea, and Protozoa / 4.2.2-1
JED A. FUHRMAN AND DAVID A. CARON
4.2.3 Aquatic Biofilms: Development, Cultivation,
Analyses, and Applications / 4.2.3-1
JOHN R. LAWRENCE, THOMAS R. NEU,
ARMELLE PAULE, DARREN R. KORBER, AND
GIDEON M. WOLFAARDT
4.3 EXTREME ENVIRONMENTS
4.3.1 The Microbiology of Extremely Acidic
Environments / 4.3.1-1
D. BARRIE JOHNSON AND ANGELES
AGUILERA
4.3.2 Life in High Salinity Environments / 4.3.2-1
AHARON OREN
4.3.3 Microbial Life in Extreme Low-Biomass
Environments: A Molecular Approach / 4.3.3-1
KASTHURI VENKATESWARAN, MYRON T. LA
DUC, PARAG VAISHAMPAYAN, AND JAMES
A. SPRY
4.3.4 Life in High-Temperature
Environments / 4.3.4-1
BRIAN P. HEDLUND, SCOTT C. THOMAS,
JEREMY A. DODSWORTH, AND CHUANLUN
L. ZHANG
4.4 ANIMAL-GUT MICROBIOMES
4.4.1 Invertebrate Gut Associations / 4.4.1-1
DANIELE DAFFONCHIO, ALBERTO ALMA,
GUIDO FAVIA, LUCIANO SACCHI, AND
CLAUDIO BANDI
4.4.2 Studying the Mammalian Intestinal
Microbiome Using Animal Models / 4.4.2-1
FLOOR HUGENHOLTZ, JING ZHANG, PAUL
W. O’TOOLE, AND HAUKE SMIDT
4.4.3 Animal Gut Microbiomes / 4.4.3-1
RICHARD J. ELLIS AND CHRISTOPHER S.
MCSWEENEY
BIODEGRADATION AND
BIOTRANSFORMATION
VOLUME EDITOR: CINDY H. NAKATSU
SECTION EDITORS: CINDY H. NAKATSU AND
CHRISTOPHER RENSING
viii ▪ CONTENTS

5.1 BIODEGRADATION 5.1.5 A Basic Introduction to Aerobic

5.1.1 Genomic Features and Genome-Wide
Analysis of Dioxin-Like Compound
Degraders / 5.1.1-1
MASAKI SHINTANI AND KAZUHIDE
KIMBARA
5.1.2 Biodegradation of Organochlorine
Pesticides / 5.1.2-1
YUJI NAGATA, MICHIRO TABATA,
YOSHIYUKI OHTSUBO, AND
MASATAKA TSUDA
5.1.3 Anaerobic Degradation of Aromatic
Compounds / 5.1.3-1
WEIMIN SUN, VALDIS KRUMINS, DONNA
E. FENNELL, LEE J. KERKHOF, AND MAX
M. HÄGGBLOM
5.1.4 Microbial Electrochemical Technologies
Producing Electricity and Valuable Chemicals
from Biodegradation of Waste Organic
Matters / 5.1.4-1
TAEHO LEE, AKIHIRO OKAMOTO, SOKHEE
JUNG, RYUHEI NAKAMURA, JUNG RAE KIM,
KAZUYA WATANABE, AND KAZUHITO
HASHIMOTO
Biodegradation of Petroleum Aromatic
Compounds / 5.1.5-1
KENGO INOUE, ONRUTHAI PINYAKONG,
KANO KASUGA, AND HIDEAKI NOJIRI
5.1.6 Environmental Systems Microbiology of
Contaminated Environments / 5.1.6-1
TERRY C. HAZEN AND GARY S. SAYLER
5.2 BIOTRANSFORMATION
5.2.1 Breathing Iron: Molecular Mechanism of
Microbial Iron Reduction by Shewanella
oneidensis / 5.2.1-1
REBECCA E. COOPER, JENNIFER L. GOFF, BEN
C. REED, RAMANAN SEKAR, AND THOMAS
J. DICHRISTINA
5.2.2 Experimental Geomicrobiology: From Field to
Laboratory / 5.2.2-1
TIMOTHY S. MAGNUSON AND RHESA
N. LEDBETTER
5.2.3 Microbial Uses in the Remediation of
Metal-Impacted Soils / 5.2.3-1
TIMBERLEY ROANE AND MUNIRA LANTZ
Index I-1
 Editorial Board

Mark P. Buttner Section 3.2 Yoichi Kamagata Section 2.1

School of Community Health Sciences, University of Bioproduction Research Institute, National Institute of
Nevada, Las Vegas, Las Vegas, NV 89154 Advanced Industrial Science and Technology (AIST),
Tsukuba, Ibaraki 305-8566, Japan
Douglas R. Call Section 2.3
Paul G. Allen School for Global Animal Health, Julian R. Marchesi Section 4.4
Washington State University, Pullman, WA 99164 School of Biosciences, Cardiff University, Centre for
Digestive and Gut Health, Imperial College London,
Yildiz T. Chambers Section 2.5 Cardiff, Wales CF10 3AT, United Kingdom
CSC Science, Engineering, and Mission Support,
Alexandria, VA 22310 John Scott Meschke Section 2.6
Department of Environmental and Occupational Health
Robert H. Findlay Section 4.2 Sciences, University of Washington, Seattle, WA 98105
University of Alabama, Department of Biological
Sciences, Tuscaloosa, AL 35487 Cleber C. Ouverney Section 2.2
Department of Biological Sciences, San Jose State
Larry J. Forney Section 4.1 University, San Jose, CA 95192
Department of Biological Sciences, University of Idaho,
Moscow, ID 83844 Christopher Rensing Section 5.2
Department of Plant and Environmental Sciences,
Stefan J. Green Section 2.4 University of Copenhagen, Frederiksberg 1871, Denmark
Research Resources Center, University of Illinois at
Chicago, Chicago, IL 60612 Ed Topp Section 3.3
Agriculture and Agri-Food Canada, London, ON N5V
Valerie J. Harwood Section 3.4 4T3, Canada
Department of Integrative Biology, University of South
Florida, Tampa, FL 33620 Gary A. Toranzos Section 3.1
Department of Biology, University of Puerto Rico, San
Brian P. Hedlund Section 4.3 Juan, PR 00932, Puerto Rico
School of Life Sciences, Nevada Institute of Personalized
Medicine, University of Nevada, Las Vegas, Las Vegas,
NV 89154

ix
 Contributors

MORTEZA ABBASZADEGAN TOM J. BATTIN

Arizona State University, Tempe, AZ 85287
ANGELES AGUILERA
Centro de Astrobiología (INTA-CSIC), Madrid 28850, Spain
WARISH AHMED
CSIRO Land and Water Queensland Biosciences Precinct,
St. Lucia, Queensland 4067, Australia
ALBERTO ALMA
Department of Agriculture, Forestry and Food Sciences
DISAFA, University of Turin, Grugliasco I-10095, Italy
Stream Biofilm and Ecosystem Research Laboratory, Ecole
Polytechnique Fédérale Lausanne, CH-1015 Lausanne,
Switzerland
NITIN BHARDWAJ
Advanced Medical Research Institute of Canada, Sudbury, ON
P3E 5J1, Canada
PASCALE BLAIS LECOURS
Centre deRecherché, University Institute of Cardiology and
Pulmonology of Québec, Université of Laval, Québec, QC
G1K7P4, Canada

ABSAR ALUM ANICET R. BLANCH

Arizona State University, Tempe, AZ, 85287
YOSHITERU AOI
Institute of Sustainable Sciences and Development, Hiroshima
University, Hiroshima 739-8529 Japan, and Northeastern
University, Boston, MA 02115
MATTHEW J. ARDUINO
Centers for Disease Control and Prevention, Division of
Healthcare Quality Promotion, Atlanta, GA 30329
NICHOLAS J. ASHBOLT
School of Public Health, University of Alberta, Edmonton, AB
T6G 2G7, Canada
CHERYL L. BAIRD
Pacific Northwest National Laboratory, Biochemistry and
Structural Biology, Fundamental and Computational Sciences
Division, Richland, WA 99352
ELISENDA BALLESTÉ
Department of Microbiology, University of Barcelona, Barcelona
08028, Spain
CLAUDIO BANDI
Department of Veterinary Sciences and Public Health,
University of Milan, Milan I-20133, Italy
RACHEL A. BARTHOLOMEW
Pacific Northwest National Laboratory, Chemical and Biological
Signature Sciences Group, National Security Directorate,
Richland, WA 99354
Department of Microbiology, University of Barcelona, Barcelona
08028, Spain
MARIA LAURA BOSCHIROLI
ANSES French Agency for Food, Environmental &
Occupational Health Safety, Maisons-Alfort Animal Health
Laboratory, Bacterial Zoonoses Unit, Maisons-Alfort 94706,
France
JULIE BRASSARD
Agriculture and Agri-Food Canada, Food Research and
Development Centre, Saint-Hyacinthe, QC J2S 8E3, Canada
ELLEN BRAUN-HOWLAND
Laboratory of Environmental Biology, NYSDOH,
Wadsworth Center, Biggs Laboratory, Empire State Plaza ,
Albany, NY 12201
KELLY R. BRIGHT
University of Arizona, Tempe, AZ 85287
JOHN BROOKS
Genetics and Precision Agriculture Unit, USDA-ARS,
Mississippi State University, Mississippi State, MS 39762
GARY S. BROWN
Lockheed Martin Corporation, Scientific, Engineering,
Response, Analytical Services, Las Vegas, NV 89119
CINDY BRUCKNER-LEA
Pacific Northwest National Laboratory, Richland,
WA 99354

xi
xii ▪ CONTRIBUTORS

MARK P. BUTTNER THOMAS J. DICHRISTINA

School of Community Health Sciences, University of Nevada, School of Biology, Georgia Institute of Technology, Atlanta,
Las Vegas, Las Vegas, NV 89154 GA 30332

MURULEE BYAPPANAHALLI JEREMY A. DODSWORTH

USGS Great Lakes Science Center, Ann Arbor, MI 48105 Department of Biology, California State University, San
Bernardino, CA 92407
DOUGLAS R. CALL
Paul G. Allen School for Global Animal Health, Washington SAMUEL DOREVITCH
State University, Pullman, WA 99164 U of Illinois at Chicago, School of Public Health, Chicago,
IL 60612
DAVID A. CARON
Department of Biological Sciences, University of Southern CAROLINE DUCHAINE
California, Los Angeles, CA 90089 Department of Biochemistry and Microbiology, Université
Laval, Québec, QC G1K7P4, Canada
YILDIZ T. CHAMBERS
CSC Science, Engineering, and Mission Support, Alexandria, THILO EICKHORST
VA 22310 Soil Microbial Ecology, University of Bremen, Bremen 28359,
Germany
DARRELL P. CHANDLER
Akonni Biosystems, Inc., Frederick, MD 21701 RICHARD J. ELLIS
Animal and Plant Health Agency, Specialist Scientific Support
HEATHER COLBURN Department, New Haw, Surrey KT15 3NB, United Kingdom
Pacific Northwest National Laboratory, Richland,
WA 99354 SLAVA EPSTEIN
Department of Biology, Northeastern University, Boston,
KEVIN K. CONNELL MA 02115
Science & Engineering Line of Service, CSC, Alexandria, VA
22310 JOSEPH FALKINHAM
III, Department of Biological Sciences, Virginia Tech,
REBECCA E. COOPER Blacksburg, VA 24061
School of Biology, Georgia Institute of Technology, Atlanta,
GA 30332 GUIDO FAVIA
School of Biosciences and Biotechnology, University of
PATRICIA CRUZ Camerino, Camerino I-62032, Italy
School of Community Health Sciences, University of Nevada,
Las Vegas, Las Vegas, NV 89154 SABINE FAVRE-BONTÉ
Microbial Ecology Laboratory, UMR 5557, CNRS/University
DANIELE DAFFONCHIO Lyon I, Villeurbanne 69622, France
Department of Food, Environmental and Nutritional, Sciences,
DeFENS, University of Milan, Milan I-20133, Italy DONNA E. FENNELL
Department of Environmental Sciences, School of
TAL DAGAN Environmental and Biological Sciences, Rutgers, The State
Institute of Microbiology, Christian-Albrechts-University of University of New Jersey, New Brunswick, NJ 08901
Kiel, Kiel 24118, Germany
KATHARINE G. FIELD
RICHARD E. DANIELSON Oregon State University, Department of Microbiology,
BioVir Laboratories, Inc., Benicia, CA 94510 Corvallis, OR 97331

PASCAL DELAQUIS BARRY S. FIELDS

Agriculture and Agri-Food Canada, Pacific Agri-Food Research
Centre, Summerland, BC V0H 1Z0, Canada
VINCENT J. DENEF
Department of Ecology and Evolutionary Biology, University of
Michigan, Ann Arbor, MI 48109
GEORGE D. DI GIOVANNI
Environmental and Occupational Health Sciences, University of
Texas School of Public Health, El Paso Regional Campus, El
Paso, TX 79902
Division of Global Disease Detection & Emergency Response,
Center for Global Health, Centers for Disease Control and
Prevention, Atlanta, GA 30333
ROBERT H. FINDLAY
University of Alabama, Department of Biological Sciences,
Tuscaloosa, AL 35487
JED A. FUHRMAN
Department of Biological Sciences, University of Southern
California, Los Angeles, CA 90089
 CONTRIBUTORS ▪ xiii

ALVIN GAJADHAR FLOOR HUGENHOLTZ

Canadian Food Inspection Agency, Centre for Foodborne and Wageningen University, Laboratory of Microbiology, TI Food
Animal Parasitology, Saskatoon, SK S7N 2R3, Canada and Nutrition, Netherlands Consortium for Systems Biology,
University of Amsterdam, Wageningen 6703HB,
SANGITA GANESH The Netherlands
School of Biology, Georgia Institute of Technology, Atlanta, GA
30332 JANINE R. HUTCHISON
Pacific Northwest National Laboratory, Chemical and Biological
JENNIFER L. GOFF Signature Sciences Group, National Security Directorate,
School of Biology, Georgia Institute of Technology, Atlanta, Richland, WA 99354
GA 30332
M. KHALID IJAZ
HYATT GREEN R&D Surface Care and Germ Protection, Reckitt Benckiser
US EPA, Office of Research and Development, National Risk LLC., Montvale, NJ 07645
Management Research Laboratory, Cincinnati, OH 45268
KENGO INOUE
STEFAN J. GREEN University of Miyazaki, Biochemistry and Applied Biosciences,
Research Resources Center, University of Illinois at Chicago, Miyazaki 889-2192, Japan
Chicago, IL 60612
DANNY IONESCU
JOHN F. GRIFFITH Leibniz Institute for Freshwater Ecology and Inland Fisheries,
Southern California Coastal Water Research Program, Costa Neuglobsow, Stechlin 16775, Germany
Mesa, CA 92626
KRISTIN JARMAN
SERGEY A. GRINSHPUN Pacific Northwest National Laboratory, Applied Statisticals
University of Cincinnati, Center for Health-Related Aerosol and Computational Modeling Group, Fundamental and
Studies, Cincinnati, OH 45267 Computational Sciences Directorate, Richland,
WA 99354
CHARLES HAGEDORN
Department of Crop and Soil Environmental Sciences, Virginia ECKARDT JOHANNING
Tech, Blacksburg, VA 24061 Occupational & Environmental Life Science, Fungal Research
Group Foundation, Inc., Albany, NY 12203
MAX M. HÄGGBLOM
Department of Biochemistry and Microbiology, School of D. BARRIE JOHNSON
Environmental and Biological Sciences, Rutgers, The State College of Natural Sciences, Bangor University, Bangor LL57
University of New Jersey, New Brunswick, NJ 08901 2UW, United Kingdom

ALAIN HARTMANN SOKHEE JUNG

Agroecology Unit, UMR 1347, INRA/University of Burgundy/ School of Civil and Environmental Engineering, Yonsei
AgroSup Dijon, Dijon 21065, France University, Seoul 120-749, Korea

VALERIE J. HARWOOD YOICHI KAMAGATA

Department of Integrative Biology, University of South Florida, Bioproduction Research Institute, National Institute of
Tampa, FL 33620 Advanced Industrial Science and Technology (AIST), Tsukuba,
Ibaraki 305-8566, Japan
KAZUHITO HASHIMOTO
Department of Applied Chemistry, School of Engineering, KANO KASUGA
University of Tokyo, Tokyo 113-8656, Japan Akita Prefectural University, Department of Biotechnology,
Akita 010-0195, Japan
TERRY C. HAZEN
Department of Civil & Environmental Engineering, University SCOTT P. KEELY
of Tennessee/Oak Ridge National Laboratory, Knoxville, United States Environmental Protection Agency, Cincinnati,
TN 37996 OH 45268

BRIAN P. HEDLUND LEE J. KERKHOF

School of Life Sciences, Nevada Institute of Personalized Department of Marine and Coastal Sciences, School of
Medicine, University of Nevada, Las Vegas, Las Vegas, Environmental and Biological Sciences, Rutgers, The State
NV 89154 University of New Jersey, New Brunswick, NJ 08901

VINCENT HILL JUNG RAE KIM

Centers for Disease Control and Prevention, Atlanta, School of Chemical and Biomolecular Engineering, Pusan
GA 30333 National University, Pusan 609-735, Korea
xiv ▪ CONTRIBUTORS

KAZUHIDE KIMBARA DE-WEI LI

Department of Applied Chemistry and Biochemical Connecticut Agricultural Experiment Station Valley Laboratory,
Engineering, Graduate School of Engineering, Shizuoka Windsor, CT 06095
University, Hamamatsu, Shizuoka 432-8561, Japan
CLARESSA E. LUCAS
JULIE KINZELMAN Division of Bacterial Diseases, National Center for Infectious
City of Racine Health Department, Racine, WI 53403 Disease, Centers for Disease Control and Prevention, Atlanta,
GA 30333
THORSTEN KLÖSGES
Institute of Molecular Evolution, Heinrich-Heine University MICHAEL D.J. LYNCH
Düsseldorf, Düsseldorf 40225, Germany Department of Biology, University of Waterloo, Waterloo, ON
N2L 3G1, Canada
ASJA KORAJKIC
US Environmental Protection Agency, Cincinnati, OH 45268 TIMOTHY S. MAGNUSON
Department of Biological Sciences, Idaho State University,
DARREN R. KORBER Pocatello, ID 83209
Department of Food and Bioproduct Science, University of
Saskatchewan, Saskatoon, SK S7N 5A8, Canada GEDIMINAS MAINELIS
Rutgers University, New Brunswick, NJ 08901
PAULA KRAUTER
Sandia National Laboratories (retired), Livermore, CA 94550 MOHAMMAD MANAFI
Institute for Hygiene and Applied Immunology, Medical
VALDIS KRUMINS University of Vienna, Vienna 1090, Austria
Department of Environmental Sciences, School of
Environmental and Biological Sciences, Rutgers, The State JULIAN R. MARCHESI
University of New Jersey, New Brunswick, NJ 08901 School of Biosciences, Cardiff University, Centre for Digestive
and Gut Health, Imperial College London, Cardiff, Wales CF10
MYRON T. LA DUC 3AT, United Kingdom
Jet Propulsion Lab, California Institute of Technology, Pasadena,
CA 91109 J. VAUN MCARTHUR
Savannah River Ecology Laboratory, Aiken, SC 29803
RUP LAL
Department of Zoology, University of Delhi, Delhi 110007, CHRISTOPHER S. MCSWEENEY
India CSIRO Animal, Food and Health Services, Queensland
Biosciences Precinct, St. Lucia, Queensland 4067,
GIDDY LANDAN Australia
Institute of Microbiology, Christian-Albrechts-University of
Kiel, Kiel 24118, Germany JOHN SCOTT MESCHKE
Department of Environmental and Occupational Health
MUNIRA LANTZ Sciences, University of Washington, Seattle, WA 98105
Department of Integrative Biology, University of Colorado,
Denver, Denver, CO 80217 HARVEY MINNIGH
Gabriella and Paul Rosenbaum Foundation, Bryn Mawr, PA
JOHN R. LAWRENCE 19010
Environment Canada, Saskatoon, SK S7N3H5, Canada
ALAN JEFF MOHR
RHESA N. LEDBETTER Life Sciences Division, U.S. Army, Dugway Proving Ground,
Department of Chemistry and Biochemistry, Utah State Dugway, UT 84022
University, Logan, UT 84322
LAURA S. MORRIS
TAEHO LEE School of Biosciences, Cardiff University, Cardiff, Wales CF10
Department of Environmental Engineering, Pusan National 3AT, United Kingdom
University, Pusan 609-735, Korea
YUJI NAGATA
JAY T. LENNON Department of Environmental Life Sciences, Graduate School of
Department of Biology, Indiana University, Bloomington, Life Sciences Tohoku University, Hatahira, Sendai 980-8577,
IN 47405 Japan

ESTELLE LEVETIN RYUHEI NAKAMURA

Department of Biological Science, The University of Tulsa, Department of Applied Chemistry, School of Engineering,
Tulsa, OK 74104 University of Tokyo, Tokyo 113-8656, Japan
 CONTRIBUTORS ▪ xv

ANKUR NAQIB CLEBER C. OUVERNEY

DNA Services Facility, University of Illinois at Chicago, Department of Biological Sciences, San Jose State University,
Chicago, IL, 60613 San Jose, CA 95192

TAKASHI NARIHIRO WILL A. OVERHOLT

Bioproduction Research Institute, National Institute of School of Biology, Georgia Institute of Technology, Atlanta,
Advanced Industrial Science and Technology (AIST) Tsukuba GA 30332
605-8566, Japan, and Department of Civil and Environmental
Engineering, University of Illinois at Urbana-Champaign, RICHARD OZANICH
Urbana, IL 61801 Pacific Northwest National Laboratory, Richland,
WA 99355
SYLVIE NAZARET
Microbial Ecology Laboratory, UMR 5557, CNRS/University HANS W. PAERL
Lyon I, Villeurbanne 69622, France Institute of Marine Sciences, University of North Carolina at
Chapel Hill, Morehead City, NC 28557
THOMAS R. NEU
River Ecology, Helmholtz Centre for Environmental Research, DARREN J. PARRIS
Magdeburg 39114, Germany School of Biology, Georgia Institute of Technology, Atlanta,
GA 30332
JOSH D. NEUFELD
Department of Biology, University of Waterloo, Waterloo, ON ARMELLE PAULE
NSL 3G1, Canada Global Institute for Water Security, Saskatoon, SK S7N3H5,
Canada
MEREDITH B. NEVERS
U.S. Geological Survey, Great Lakes Science Center, Porter, SUSAN R. PETTERSON
IN 46304 Water & Health Pty Ltd, Salamander Bay, NSW 2317,
Australia
JUDITH NOBLE-WANG
Centers for Disease Control and Prevention, Division of MANTHA S. PHANIKUMAR
Healthcare Quality Promotion, Atlanta, GA 30329 Michigan State University, Department of Civil and
Environmental Engineering, East Lansing, MI 48824
HIDEAKI NOJIRI
The University of Tokyo, Biotechnology Research Center, ONRUTHAI PINYAKONG
Tokyo 13-8657, Japan1 Chulalongkorn University, Department of Microbiology,
Bangkok 10330, Thailand
YOSHIYUKI OHTSUBO
Department of Environmental Life Sciences, Graduate School of PASCAL PIVETEAU
Life Sciences Tohoku University, Hatahira, Sendai 980-8577, Agroecology Unit, UMR 1347 INRA/University of Burgundy/
Japan AgroSup Dijon, Dijon 21065, France

AKIHIRO OKAMOTO OVIDIU POPA

Department of Earth Sciences, University of Southern Institute of Microbiology, Christian-Albrechts-University of
California, Los Angeles, CA 90089 Kiel, Kiel 24118, Germany

AHARON OREN GRACIELA RAMÍREZ TORO

Department of Plant and Environmental Sciences, Institute of Centro de Educación, Conservación e Interpretación
Life Sciences, The Hebrew University of Jerusalem, Jerusalem Ambiental, Universidad Interamericana de Puerto Rico, San
91904, Israel Germán, PR 00683, Puerto Rico

ROBIN K. OSHIRO BEN C. REED

Engineering and Analysis Division, USEPA Headquarters, School of Biology, Georgia Institute of Technology, Atlanta,
Washington, DC 20460 GA 30332

PAUL W. O’TOOLE TIMBERLEY ROANE

School of Microbiology & Alimentary Pharmabiotic Centre, Department of Integrative Biology, University of Colorado,
University College Cork, Cork T12 YN60, Ireland Denver, Denver, CO 80217

TIMOTHY G. OTTEN LAURA J. ROSE

Department of Microbiology, Oregon State University, Centers for Disease Control and Prevention, Division of
Corvallis, OR 97331 Healthcare Quality Promotion, Atlanta, GA 30329
xvi ▪ CONTRIBUTORS

HODON RYU University of Amsterdam, Wageningen 6703HB, The

US EPA NRMRL/WSWRD/MCCB, Cincinnati, OH 45268 Netherlands

LUCIANO SACCHI JAMES A. SPRY

Department of Biology and Biotechnology “L. Spallanzani”, Jet Propulsion Lab, California Institute of Technology, Pasadena,
University of Pavia, Pavia I-27100, Italy CA 91109

MICHAEL SADOWSKY LINDA D. STETZENBACH

BioTechnology Institute, University of Minnesota, St. Paul,
MN 55108
NASEER SANGWAN
Department of Zoology, University of Delhi, Delhi 110007, India
University of Nevada, Las Vegas, Las Vegas, NV 89154
FRANK J. STEWART
School of Biology, Georgia Institute of Technology, Atlanta,
GA 30332

TASHA M. SANTIAGO-RODRIGUEZ DON STOECKEL

Department of Biology, Center for Applications in
Biotechnology, California Polytechnic State University, San
Luis Obispo, CA 93407
JORGE SANTO DOMINGO
US EPA NRMRL/WSWRD/MCCB, Cincinnati, OH 45268
Batelle Memorial Institute, Columbus, OH 43201
TIMOTHY M. STRAUB
Pacific Northwest National Laboratory, Chemical and Biological
Signature Sciences Group, National Security Directorate,
Richland, WA 99354

NEHA SARODE GREGORY D. STURBAUM

School of Biology, Georgia Institute of Technology, Atlanta, ALS Laboratory Group, Molecular Biology, Scoresby, VIC 3179,
GA 30332 Australia

SYED A. SATTAR WEIMIN SUN

Centre for Research on Environmental Microbiology, University Department of Biochemistry and Microbiology, Department
of Ottawa, Ottawa, ON K1H 8M5, Canada of Environmental Sciences, School of Environmental and
Biological Sciences, Rutgers, The State University of New Jersey,
GARY S. SAYLER New Brunswick, NJ 08901
Department of Microbiology, University of Tennessee,
Knoxville, TN 37996 MICHIRO TABATA
Department of Environmental Life Sciences, Graduate School of
HANNES SCHMIDT Life Sciences Tohoku University, Hatahira, Sendai 980-8577,
Soil Microbial Ecology, University of Bremen, Bremen 28359, Japan
Germany
SCOTT C. THOMAS
RAMANAN SEKAR School of Life Sciences, University of Nevada Las Vegas, Las
School of Biology, Georgia Institute of Technology, Atlanta, GA Vegas, NV 89154
30332
ERIC C. THOMPSON
SHERRY L. SESTON King County Environmental Laboratory, Seattle,
Department of Biology, Alverno College, Milwaukee, WA 98119
WI 53234
PETER S. THORNE
ORIN C. SHANKS Department of Occupational and Environmental Health,
US EPA, Office of Research and Development, National The University of Iowa, College of Public Health, Iowa City,
Risk Management Research Laboratory, Cincinnati, OH 45268 IA 52246

MASAKI SHINTANI GARY A. TORANZOS

Department of Applied Chemistry and Biochemical Department of Biology, University of Puerto Rico, San Juan,
Engineering, Graduate School of Engineering, Shizuoka PR 00932, Puerto Rico
University, Hamamatsu, Shizuoka 432-8561, Japan
MASATAKA TSUDA
LAURA Y. SIFUENTES Department of Environmental Life Sciences, Graduate School of
University of Arizona, Tempe, AZ 85287 Life Sciences, Tohoku University, Hatahira, Sendai 980-8577,
Japan
HAUKE SMIDT
Wageningen University, Laboratory of Microbiology, TI Food R. CARY TUCKFIELD
and Nutrition, Netherlands Consortium for Systems Biology, ECOSTATys LLC., Aiken, SC 29803
 CONTRIBUTORS ▪ xvii

PARAG VAISHAMPAYAN RICHARD L. WHITMAN

Jet Propulsion Lab, California Institute of Technology, Pasadena, U.S. Geological Survey, Great Lakes Science Center, Porter,
CA 91109 IN 46304

SUSAN M. VARNUM KLAUS WILLEKE

Pacific Northwest National Laboratory, Biochemistry and University of Cincinnati, Cincinnati, OH 45267
Structural Biology, Fundamental and Computational Sciences
Division, Richland, WA 99352 GIDEON M. WOLFAARDT
Department of Chemistry and Biology, Ryerson University,
KASTHURI VENKATESWARAN Toronto, ON M5B2K3, Canada
Jet Propulsion Lab, California Institute of Technology, Pasadena,
CA 91109 CHIN S. YANG
Prestige EnviroMicrobiology, Voorhees, NJ 08043
ERIC N. VILLEGAS
United States Environmental Protection Agency, Cincinnati, MARYLYNN V. YATES
OH 45268 Department of Environmental Sciences, University of
California, Riverside, Riverside, CA 92521
KAZUYA WATANABE
School of Life Science, Tokyo University of Pharmacy and Life CHUANLUN L. ZHANG
Sciences, Tokyo 192-0392, Japan State Key Laboratory of Marine Geology School of Ocean and
Earth Sciences, Tongji University, Shanghai 201804, China
JENNIFER WEIDHAAS
West Virginia University, Civil and Environmental Engineering, JING ZHANG
Morgantown, WV 26506 Wageningen University, Laboratory of Microbiology,
Wageningen 6703HB, The Netherlands
MARK H. WEIR
Department of Public Health, Department of Civil and NICOLETTE A. ZHOU
Environmental Engineering, Temple University, Philadelphia, Department of Environmental and Occupational Health
PA 19122 Sciences, University of Washington, Seattle, WA 98105
 INTRODUCTION
MARYLYNN V. YATES
1.1.1
There are bred certain minute creatures which cannot be seen by
the eyes, which float in the air and enter the body through the mouth
and nose and there cause serious diseases.
— Marcus Terentius Varro, De Re Rustica
[On Agriculture], 37 BC
Environmental microbiology might be considered by some to
be an ill-defined subject: Where does the environment begin,
and where does it end? From Marcus Terentius Varro's obser-
vations regarding unseen “minute creatures” more than two
millennia ago to Antonie van Leeuwenhoek's first glimpse
of the “animalcula” beneath his lens, there is no place on
Earth—from thermophilic, acidic springs to the air we
breathe to the deepest subsurface locations we have yet
been able to reach—where people have looked and not found
microorganisms of some type. The domain of what may be
considered environmental microbiology thus continues to
expand beyond the textbook definition of “the study of micro-
organisms existing in natural and artificial environments.” At
the same time, our knowledge of microorganisms is increasing
at an ever-more rapid rate as the result of incredible improve-
ments in analytical methodology, especially at the molecular
level. When compiling a manual of this nature, therefore,
how does one determine what to include and what to
exclude? In the end, the editors decided to showcase as
much information as possible on some of the most important
areas of environmental microbiology, to provide a clear sense
of the possibilities presented by the existence of microorgan-
isms in various environments. Further and more detailed
information can be found in the wealth of expertly chosen
references within each chapter.
This edition of the Manual of Environmental Microbiology
has been reorganized, with new and updated sections to pro-
vide recent information on topics of importance to the field.
The General Methodology volume, beginning with Section
2.1, is devoted to methodologies—the “how” of environmen-
tal microbial studies from analytical detection to sample col-
lection. The methods presented in this volume are the basis of
how we understand the microbial world around us and are
used across environmental microbial disciplines and applica-
tions. General Methodologies begins with tried-and-true cul-
ture-based and physiological detection methods, microscopy-
based methods, and molecular target–specific detection
doi:10.1128/9781555818821.ch1.1.1
1.1.1-1
methods, before moving on to address the increasingly impor-
tant study of microbes at the community level, thanks to the
advent of next-generation sequencing technologies. This vol-
ume concludes with information on critical topics that are
sometimes overlooked in environmental microbiology: qual-
ity assurance and quality control, which increase the useful-
ness and reliability of analytical data for downstream
decision making.
The next volume, Environmental Public Health Microbi-
ology, surveys the microorganisms in the water, air, or soil that
can cause illness when humans are exposed to them. These
potentially pathogenic microorganisms can take many ave-
nues en route to their target host, ultimately leading to sub-
stantial health impacts and considerable negative economic
and social consequences to the communities involved. The
importance of understanding microbial distribution and
detection for public health cannot be understated. This vol-
ume includes chapters describing the field of microbial source
tracking; methods and approaches for determining predomi-
nant sources of fecal pollution of water, which has become
a critical component of assessment; and determining ways
to better protect water from microbial contamination. It cul-
minates in a section on microbial risk assessment, integrating
the information in the earlier sections and providing perspec-
tive for informed decision making and rational project design.
The Microbial Ecology volume addresses the various ways
microorganisms can be classified and their interconnected
relationships, with each other and with macroorganisms,
and impacts on the environment. Some microorganisms are
native to their environment and perform essential services,
such as cycling nutrients or interacting with other organisms
in ways that enable them to perform functions and/or inhabit
environments where they would otherwise not exist. Other
microorganisms may be introduced to a new environment,
changing the overall ecosystem. This volume begins with
an exploration of the theories behind microbial genome evo-
lution and the evolutionary ecology of microorganisms. The
subject matter then delves into more specific environments,
including aquatic environments, environments with extreme
conditions (extreme acidity, high salinity, low biomass, and
high temperatures), and the unique environment inside the
gastrointestinal tract of animals, an area of increasing study
and data.
1.1.1-2 ▪ INTRODUCTION
The final volume, Biodegradation and Biotransformation,
focuses on the varied ways that microorganisms can transform
or degrade nearly any natural or anthropogenic chemical
compound. Since the earliest eras of the Industrial Revolu-
tion, study of microorganisms for these processes has been
an active area of research. Microbial transformations of natu-
ral substances can make otherwise unavailable nutrients
accessible by living organisms. Additionally, microorganisms
can be used to degrade harmful contaminants, either by using
the biochemical pathways naturally present or by employing
genetic modification. These abilities have been exploited in
numerous situations in which soil or water has been contami-
nated by organic or inorganic chemicals as well as for a variety
of industrial and commercial uses.
Microorganisms play a critical role in the health and well-
being of the planet and the human, animal, and plant life that
dwells here. Our understanding of them is crucial for main-
taining the environment. This fourth edition of the Manual
of Environmental Microbiology builds on the solid foundation
created by the previous three editions, and we are indebted
to those who came before us. This edition is truly a collabora-
tive triumph, and is the work of 19 editors and more than 180
contributing authors who generously gave their time, effort,
and expertise. It is my hope that this reference serves as an
informative and reliable source of information for current
and future endeavors in environmental microbiology and
inspires the next generation of researchers as they move along
their path in this growing and important area.
General Methodology
VOLUME EDITOR: SURESH D. PILLAI
SECTION EDITORS: YOICHI KAMAGATA, CLEBER C. OUVERNEY, DOUGLAS R. CALL, STEFAN J. GREEN,
YILDIZ T. CHAMBERS, AND JOHN SCOTT MESCHKE
2.1 CULTURE-BASED AND PHYSIOLOGICAL
DETECTION
2.1.1 Detection of Specific Taxa Using
Chromogenic and Fluorogenic
Media / 2.1.1-1
MOHAMMAD MANAFI
2.1.2 Anaerobic Cultivation / 2.1.2-1
TAKASHI NARIHIRO AND YOICHI
KAMAGATA
2.1.3 New Devices for Cultivation / 2.1.3-1
YOSHITERU AOI AND SLAVA EPSTEIN
2.2 MICROSCOPIC METHODS
2.2.1 Gold-Based In Situ Hybridization for
Phylogenetic Single-Cell Detection of
Prokaryotes in Environmental
Samples / 2.2.1-1
THILO EICKHORST AND HANNES SCHMIDT
2.2.2 Assessment of Prokaryotic Biological Activity
at the Single-Cell Level by Combining
Microautoradiography and Fluorescence
in situ Hybridization / 2.2.2-1
CLEBER C. OUVERNEY
2.3 TARGET-SPECIFIC DETECTION
2.3.1 Antibody-Based Technologies for
Environmental Biodetection / 2.3.1-1
CHERYL L. BAIRD AND SUSAN M. VARNUM
2.3.2 PCR, Real-Time PCR, Digital PCR, and
Isothermal Amplification / 2.3.2-1
RACHEL A. BARTHOLOMEW, JANINE
R. HUTCHISON, TIMOTHY M. STRAUB,
AND DOUGLAS R. CALL
2.3.3 Microarray-Based Environmental
Diagnostics / 2.3.3-1
DARRELL P. CHANDLER
2.3.4 Field Application of Pathogen Detection
Technologies / 2.3.4-1
TIMOTHY M. STRAUB, DOUGLAS R. CALL,
CINDY BRUCKNER-LEA,
HEATHER COLBURN, CHERYL L. BAIRD,
RACHEL A. BARTHOLOMEW,
RICHARD OZANICH, AND
KRISTIN JARMAN
2.4 MICROBIAL COMMUNITY ANALYSIS OF
ENVIRONMENTAL SAMPLES WITH
NEXT-GENERATION SEQUENCING
2.4.1 Introduction to Microbial
Community Analysis of Environmental
Samples with Next-Generation
Sequencing / 2.4.1-1
STEFAN J. GREEN AND JOSH D. NEUFELD
2.4.2 Microbial Community Analysis Using
High-Throughput Amplicon
Sequencing / 2.4.2-1
DANNY IONESCU, WILL A. OVERHOLT,
MICHAEL D. J. LYNCH,
JOSH D. NEUFELD, ANKUR NAQIB, AND
STEFAN J. GREEN
2.4.3 Functional Metagenomics: Procedures and
Progress / 2.4.3-1
LAURA S. MORRIS AND JULIAN
R. MARCHESI
2.4.4 Metagenomics: Assigning Functional
Status to Community Gene
Content / 2.4.4-1
NASEER SANGWAN AND RUP LAL
2.4.5 Generation and Analysis of Microbial
Metatranscriptomes / 2.4.5-1
NEHA SARODE, DARREN J. PARRIS, SANGITA
GANESH, SHERRY L. SESTON, AND FRANK
J. STEWART
2.5 QA/QC IN ENVIRONMENTAL 2.5.7 Study Design / 2.5.7-1
YILDIZ T. CHAMBERS AND
MICROBIOLOGY ROBIN K. OSHIRO
2.5.1 Introduction to Principles of Quality
Assurance / 2.5.1-1
KEVIN K. CONNELL 2.6 SAMPLING METHODS
2.5.2 General Quality Control / 2.5.2-1 2.6.1 Water Sampling and Processing
ROBIN K. OSHIRO Techniques for Public Health-Related
2.5.3 Quality Control for Bacteriological Microbes / 2.6.1-1
Analyses / 2.5.3-1 VINCENT HILL
ELLEN BRAUN-HOWLAND 2.6.2 Surface Sampling / 2.6.2-1
2.5.4 Quality Control for Virological LAURA J. ROSE, JUDITH NOBLE-WANG, AND
Analyses / 2.5.4-1 MATTHEW J. ARDUINO
RICHARD E. DANIELSON 2.6.3 Soil Sampling for Microbial
2.5.5 Quality Control for USEPA Analyses / 2.6.3-1
Method 1623 Protozoan Analysis and JOHN BROOKS
PCR Analyses / 2.5.5-1 2.6.4 Microbiological Sampling of
GEORGE D. DI GIOVANNI AND Wastewater and Biosolids / 2.6.4-1
GREGORY D. STURBAUM NICOLETTE A. ZHOU, ERIC C. THOMPSON,
2.5.6 The Role of Statistical Thinking in AND JOHN SCOTT MESCHKE
Environmental Microbiology / 2.5.6-1
J. VAUN MCARTHUR AND R. CARY TUCKFIELD
 Detection of Specific Taxa Using Chromogenic and
Fluorogenic Media
MOHAMMAD MANAFI
2.1.1
OVERVIEW OF FLUOROGENIC AND
CHROMOGENIC MEDIA IN ENVIRONMENTAL
MICROBIOLOGY
Over the past 20 years, a number of selective chromogenic
and fluorogenic media for detection and enumeration of
most important bacteria and yeast in particular in food and
water have been developed and marketed. Fluorogenic
enzyme substrates generally consist of a specific substrate for
the specific enzyme such as sugar or amino acid and a fluoro-
gen such as 4-methylumbelliferone, being able to convert UV
light (365 nm) to visible light. Chromogenic enzyme sub-
strates are compounds that act as the substrate for specific
enzymes and change color due to the action of the enzymes.
Most commercially available chromogenic media have
exploited indoxylic substrate. Indoxyl and its halogenated
derivatives can be derivatized to form a range of esters. Halo-
genation of the indoxyl molecule has an effect on the color
and intensity of the chromogen. 5-Bromo-4-chloro-indoxyl
forms a green/blue dye, whereas 5-bromo-6-chloro-indoxyl
forms a magenta dye. The incorporation of such substrates
into a selective medium can eliminate the need for subculture
and further biochemical tests to establish the identity of cer-
tain microorganisms. There have been some review papers
dealing with the use of these substrates in environmental
and clinical microbiology (1–3). A useful review describes
the chemistry of chromogens and fluorogens in culture
media (4). Some chromogenic or fluorogenic media (e.g.,
agar Listeria according to Ottaviani and Agosti [ALOA]
and tryptone-bile-glucuronic [TBX] have therefore been
taken up in the standardized ISO methods (ISO 11290-1/
A1:2004 and ISO 16649:2001, respectively). This review
describes some recent developments in chromogenic and flu-
orogenic culture media in microbiological diagnostic in par-
ticular in food and water microbiology.
GRAM-NEGATIVE BACTERIA
E. coli
E. coli is an important indicator of fecal contamination in
samples from the food processing and water purification
plants. The new enzymatic definition of E. coli is the posses-
sion of uidA gene coding for ß-D-glucuronidase (GUD) as an
indicator for E. coli. GUD is an enzyme that catalyzes
the hydrolysis of ß-D-glucopyranosiduronic acids into their
doi:10.1128/9781555818821.ch2.1.1
2.1.1-1
corresponding aglycons and D-glucuronic acid (5, 6). GUD
activity is measured by using different chromogenic and
fluorogenic substances such as 4-methylumbelliferyl-ß-D-
glucuronide (Fig. 1, MUG) or 5-bromo-4-chloro-3-indolyl-
ß-D-glucuronide. MUG has been incorporated into different
media, including lauryl sulfate broth, lactose broth, m-Endo
broth, EC broth, violet red bile agar, ECD agar, MacConkey
agar and m-FC agar, and are described earlier (7). The dis-
advantage of incorporating MUG into solid media is that
fluorescence diffuses rapidly from the colonies into the
surrounding agar. The chromogenic substrate 5-bromo-
4-chloro-3-indolyl-ß-D-glucuronide (BCIG) is added into
TBX agar. GUD cleaves BCIG, and the released chromo-
sphere causes distinct blue-green-colored colonies of E. coli
and complies with the ISO/DIS Standard 16649 for the enu-
meration of E. coli in food and animal feeding stuffs.
Coliform
The new definition of coliforms, which is not method related,
is the possession of lacZ gene coding for ß-D-galactosidase
which is responsible for the cleavage of lactose into glucose
and galactose. The determination of ß-D-galactosidase is
accomplished by using substrates such as 6-bromo-3-indolyl-
ß-galactopyranoside (Salmon-Gal, red), 5-bromo-4-chloro-
3-indolyl-ß-D-galactopyranoside (Fig. 2, blue) or fluorogenic
substrate 4-methylumbelliferyl-ß-D-galactopyranoside (blue
fluorescence under UV light). Attempts were made to
enhance the coliform assay response by adding 1-isopropyl-ß-
D-thiogalactopyronaside (IPTG) to the media (8) increas-
ing the ß-D-galactosidase activity by improving the transfer
of the substrate and/or enzyme across the outer membrane.
IPTG molecule induces the lac operon but unlike the natural
substrate it cannot be hydrolysed by ß-galactosidase.
E. coli and Coliforms
Commercially available media have been developed which
permit rapid simultaneous detection of E. coli and coliforms
(Table 1) such as the Colilert system (IDEXX, Branford,
CT), LMX broth, and Readycult coliforms (Merck, Ger-
many). Other representative examples are MI agar (Becton-
Dickinson, USA), CHROMagar CCA (CHROMagar,
France), or Chromocult coliforms (Merck). Comparative
studies have been done and are reviewed elsewhere (1).
Schets et al. (47) compared Colilert with Dutch standard
enumeration methods for E. coli and coliforms in water and
2.1.1-2 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
FIGURE 1 Structure of 4-Methylumbellieryl-beta-D-glucuronide (MUG) for detection of E. coli. doi:10.1128/9781555818821.ch2.1.1.f1
found that Colilert gave false-negative results in samples with
low numbers of E. coli or total coliforms. An evaluation of a
number of presence/absence (P/A) tests for coliforms and
E. coli, including LMX broth (Merck) and Colilert (IDEXX)
has been published under the Department of the Environ-
ment series in the United Kingdom (48). The study con-
cludes that there is no P/A test that is best at all locations
for both coliforms and E. coli, and as there can be marked eco-
logical differences between sources it is important that partic-
ular P/A tests are validated in each geographical area before
use. The MI agar method (10) containing indoxyl-ß-D-
glucuronide and 4-methylumbelliferyl-ß-D-galactopyranoside
for the simultaneous detection of E. coli and total coliforms,
was compared with the approved method by the use of
wastewater-spiked tap water samples. Overall, weighted anal-
ysis of variance (significance level 0.05) showed that the MI
FIGURE 2 Structure of X-GAL (5-bromo-4chloro-3-indolyl-beta-D-galactopyranoside) for detection of coliforms.
doi:10.1128/9781555818821.ch2.1.1.f2
agar recoveries of total coliforms and E. coli were significantly
higher than those of mEndo agar. MI agar method is approved
for detecting total coliforms and E. coli under the total coli-
forms rule and for enumerating total coliforms under the sur-
face water treatment rule in the United States. Byamukama
et al. (49) described the quantification of E. coli contamina-
tion with Chromocult coliform agar from different polluted
sites in a tropical environment. It proved to be efficient and
feasible for determining fecal pollutions in the investigated
area within 24 h. Blue coloration in the broth indicates the
presence of total coliforms and/or E. coli.
E. coli O157:H7
E. coli O157:H7 is an important foodborne pathogen and can
cause diarrhea, hemorrhagic colitis, and hemolytic uremic
syndrome. Several chromogenic media have been applied to
 2.1.1. Detection of Specific taxa Using Chromogenic and Fluorogenic Media ▪ 2.1.1-3

TABLE 1 Example of commercial chromogenic and fluorogenic culture media

Target Medium/colony color Company Reference
E. coli/coliforms Readycult (E. coli: blue fluorescence /coliforms blue green coloration) Merck (Germany) 9
MI-agar (E. coli: blue colonies/coliforms blue fluorescence) BD (USA) 10
LMX broth (E. coli: blue fluorescence /coliforms blue green coloration) Merck (Germany) 11
Chromocult (E. coli: blue colonies/coliforms red colonies) Merck (Germany) 12
Coli Complete (E. coli: blue fluorescence /coliforms blue Biocontrol (USA) 13
green coloration)
m-Coliblue (E. coli: blue colonies/coliforms red colonies) Hach (USA) 14
E. coli O157:H7 CHROMagar O157 (red) CHROMagar (France) 15
Rainbow agar O157 (black) Biolog (USA) 16, 17
BCM E. coli O157:H7 (blue-black) Biosynth (Switzerland) 18
Salmonella spp. SM ID medium and SM ID 2-medium (red) BioMérieux (France) 19
BBL CHROMagar Salmonella (mauve) BD (USA) 20
Harlequin Salmonella ABC (green) lab m (England) 21
Rambach agar (red) Merck (Germany) 22
Salmonella-Chromogen-agar Oxoid (England) 23
Cronobacter sakazakii DFI agar (blue) Oxoid (England) 24
ESPM (blue) R&F Laboratories 25
Yersinia enterocolitica CHROMagar Yersinia, pathogenic (mauve) and nonpathogenic (blue). CHROMagar (France) 26
ESBL ChromID ESBL bioMérieux (France) 27
Brilliance ESBL agar Oxoid (England) 27
Vibrio sp. CHROMagar Vibrio CHROMagar (France) 28
Bio-Chrome Vibrio medium BioMedix, USA 29
Staphylococcus aureus CHROMagar Staph.aureus CHROMagar (France) 30
S. aureus ID agar (SA ID) BioMérieux (France) 31
MRSA CHROMagar MRSA CHROMagar (France) 32
MRSASelect Bio-Rad (USA) 32
Listeria monocytogenes BCM Listeria monocytogenes plating medium Biosynth (Switzerland) 33
CHROMagar Listeria CHROMagar (France) 34
Oxoid Chromogenic Listeria agar (OCLA) Oxoid (England) 34
Rapid’L.mono Sanofi (France) 34
Agar Listeria (ALOA) AES (France) 34
Bacillus group BCM Bacillus cereus group Biosynth (Switzerland) 35
B. cereus/B. thuringiensis agar R&F Laboratories (USA) 36
Chromogenic Bacillus Cereus agar (CBC) Oxoid (England) 37
Bacillus anthracis agar R&F Laboratories (USA) 38
Enterococcus spp mEI agar (indoxyl-ß-D-glucoside) BD (USA) 39
Chromocult Enterococci broth and Readycult/blue Merck (Germany) 40
Chromocult Enterococci agar/red Merck (Germany) 41
Enterolert (4-methylumbelliferone -ß-D-glucoside)/ blue fluoresence Idexx (USA) 42
VRE ChromID VRE agar bioMérieux (France) 43
CHROMagar VanRE CHROMagar (France) 44
Clostridium perfringens Chromoselect CP/green Sigma (Switzerland) 45
Clostridium difficile ChromID C. difficile agar Biomérieux (France) 46

the detection of E. coli O157:H7 from food samples. Most
are based on similar principles, relying on nonfermentation
of sorbitol and/or rhamnose and lack of ß-glucuronidase
activity (50). A second chromogenic substrate (for α- or
ß-galactosidase) may be used to highlight the presence
of E. coli O157:H7 among nonreactive background flora.
Some selective media increase the effectiveness of E. coli
O157:H7 isolation, including Rainbow agar O157 (RB,
Biolog, Hayward, USA), BCM O157:H7 (BCM, Biosynth
AG, Staad, Switzerland), Fluorocult E. coli O157:H7 (H7,
Merck), and CHROMagar O157 (Chromagar).
The performance of three chromogenic agars—Rainbow
Agar O157, CHROMagar O157, and O157:H7 ID agar—
with a large collection of verotoxigenic and nontoxigenic
E. coli strains have been tested (15, 16). Another study
compared Rainbow Agar O157, BCM O157:H7, and Fluoro-
cult HC with the conventional Sorbitol MacConkey have
showed the clear advantage of the chromogenic media for
the isolation of E. coli O157:H7 from raw ground beef and
raw drinking milk (51). The behavior of E. coli O157:H7
was studied during the manufacture and ripening of raw
goat milk lactic cheeses using O157:H7 ID (bioMérieux) and
CT-SMAC agar (52). In conclusion, this new chromogenic
media O157:H7 ID proved to be well suited and simplified
many of the inherent problems associated with plate confir-
mation of E. coli O157:H7 using SMAC as subculture media.
Salmonella
The conventional media for the detection of Salmonella
have a very poor specificity, creating an abundance of false
2.1.1-4 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
positives (such as Citrobacter, Proteus) among the rare real
positive Salmonella. On Rambach agar, Salmonella strains
gave red colonies because of their ability to acidify neutral
red by fermentation of propylene glycol (22). This medium
employs X-gal for ß-D-galactosidase positive coliforms. How-
ever, the strains of S. typhi and S. paratyphi fail to produce
acid from propylene glycol, resulting in colorless colonies
on Rambach agar. Furthermore, the strains of S. enterica sub-
species S. arizona show blue-violet colonies on this medium
(53). On SM-ID agar (bioMérieux) Salmonella colonies
are detected by their distinctive red coloration, while coli-
forms appear blue, violet, or colorless. A rare study in food
microbiology using chromogenic media has been done
by Schönenbrücher et al. (23). The draft ISO 6579:2002
was compared to the European gold standard (DIN EN
12824:1998), including the three chromogenic plating
media AES Salmonella agar plate, Oxoid Salmonella chro-
mogen media, and Miller-Mallinson agar. MUCAP-test
(Biolife, Italy) is a confirmation test for Salmonella species
based on the rapid detection of caprylate esterase, using
fluorogenic 4-methylumbelliferyl-caprylate. In the presence
of C8 esterase the substrate is cleaved with the release of 4-
methylumbelliferone. Strong bluish fluorescence indicates
the presence of Salmonella spp. Since most of the false-
positive strains are oxidase positive, the combination of
MUCAP and the oxidase test is recommended (54). Other
chromogenic media for detection of Salmonella spp. are listed
in Table 1.
Shigella
The chromogenic Shigella spp. agar (R&F Laboratories, West
Chicago, IL), offers intermediate selectivity using bile salts
and antibiotic supplementation (55). Colony color enhance-
ments created by the chromogenic Shigella spp. agar results
from the lack of the Shigella to utilize the carbohydrates and
metabolize the chromogens. Presumptive positive colonies
of Shigella sp. appears as white/clear colonies 1.0–3.0 mm
diameter with or without a clear ring after 22–28 h incuba-
tion, whereas the colonies of other members of the Enterobac-
teriaceae are variously colored. The second medium, xylose
galactosidase medium contains bile salts, D-xylose and 5-
bromo-4-chloro-3-indolyl-ß-D-galactopyranoside and IPTG
as biochemical marker (56). S. sonnei give green colonies
(xylose negative and ß-gal positive) and colorless colonies
for other species of Shigella (xylose and ß-gal negative).
Cronobacter (Enterobacter) sakazakii
C. sakazakii is a recently described genus that is composed of
six genomospecies. C. sakazakii is an occasional contaminant
of powdered infant formula that can cause rare but severe
foodborne infections such as meningitis, necrotizing entero-
colitis, bacteremia, and sepsis in infants. C. sakazakii possesses
α-glucosidase activity and a number of selective, chromo-
genic agars for C. sakazakii are based on the detection of
this enzyme (57, 58, 59). 5-Bromo-4-chloro-3-indolyl-α-
D-glucopyranoside (X-α-Glc) is added to a basal medium to
differentiate C. sakazakii strains from other members of
the Enterobacteriaceae. The enzyme α-glucosidase hydrolyses
X-α-Glc, giving blue-colored colonies on DFI agar (Oxoid,
UK), ESIA agar (recommended by ISO/TS 22964, AES,
France), Compas (Biokar, France), or Chromocult E. sakazakii
(Merck,), which are commercially available. Hi Chrome
E. sakazakii agar (Sigma-Aldrich) also complies with the for-
mulation recommended in ISO/TS 22964.
A dual chromogenic medium (ESPM) is available
from R&F Laboratories (25), as well as from bioMérieux
(ChromID Sakazakii). A second substrate, 5-bromo-4-
chloro-3-indolyl-ß-cellobioside, is included in this medium.
A mix of carbohydrates such as D-sorbitol, D-arabitol, and
adonitol and antibiotics vancomycin and cefsulodin are
added. Cronobacter colonies appear blue-black, with most
other species appearing as colorless, yellow, or green colonies.
4-Methylumbelliferyl-α-D-glucoside, the fluorogenic sub-
strate of α-D-glucosidase, was used as a selective marker to
develop a differential medium for C. sakazakii. This bacterium
showed strong fluorogenic characteristics clearly distinguish-
able from other microorganisms. On the basis of reducing
background noise, an optimum basal medium and nitrogen
source were selected. Incubation conditions were optimized
(60, 61).
Extended-Spectrum β-lactamase (ESBL)
Rapid detection of extended-spectrum β-lactamase (ESBL)–
producing Gram-negative bacilli in surveillance samples of
high-risk patients allows early optimization of antimicrobial
therapy and timely introduction of infection control proce-
dures. BLSE (AES Chemunex), chromID ESBL (bioMér-
ieux), and Brilliance ESBL agar (Oxoid) for rapid detection
of ESBL-producing Enterobacteriaceae are the few available
chromogenic media (27). A comprehensive study from Ger-
many showing a high prevalence of TEM-, CTX-M-, and
SHV-type ESBLs in Enterobacteriaceae isolated from retail
chicken meat. The high rate of coresistance to different
classes of antibiotics in the ESBL producers might reflect
the common veterinary usage of these and related substances
(62).
Yersinia enterocolitica
Weagant (63) described an agar, so-called YeCM, for detec-
tion of potentially virulent Y. enterocolitica. This agar contains
cellobiose as the fermentable sugar, a chromogenic substrate,
and selective inhibitors for suppression of colony formation
by many competing bacteria. The strains of biotypes 1B,
and biotypes 2–5 formed convex, red bull’s-eye colonies on
YeCM. Y. enterocolitica biotype 1A and other related Yersinia
formed colonies that were purple/blue on YeCM while they
formed typical red bull’s-eye colonies on CIN agar. Commer-
cially available CHROMagar Yersinia is a new chromogenic
medium for the presumptive detection of virulent Y. enteroco-
litica in stools (26).
Campylobacter
R & F Campylobacter jejuni/C. coli is a new chromogenic plat-
ing medium. Presumptively positive colonies of C. jejuni/
C. coli appear as dark salmon flat to convex colonies, 1.0–
2.0 mm in diameter with and without a clear ring after 48 h
at 41–42°C under microaerophilic conditions. Brilliance
CampyCount Agar (Oxoid) is specifically designed for accu-
rate, specific, and easy enumeration of C. jejuni and C. coli
from poultry testing. It is a transparent medium on which
Campylobacter produce distinct dark red colonies, making
identification and counting of Campylobacter significantly
easier than on traditional charcoal or blood-containing agar.
Vibrio
Vibrio parahaemolyticus has been one of the most impor-
tant foodborne pathogens in Japan since the 1960s, and a
large epidemic was caused by the pandemic serotype O3:K6
from 1997 to 2001. The traditional most-probable number
(MPN) method using thiosulfate-citrate-bile-salts-sucrose
medium (TCBS) for detecting V. parahaemolyticus cannot dif-
ferentiate growth of V. parahaemolyticus from Vibrio vulnificus
 2.1.1. Detection of Specific taxa Using Chromogenic and Fluorogenic Media ▪ 2.1.1-5
or Vibrio mimicus. Suspicious colonies grown on TCBS need
to be confirmed with lengthy biochemical tests. Recently, a
chromogenic medium (Bio-Chrome Vibrio medium, Bio-
Medix, USA) was developed for differentiating V. parahaemo-
lyticus from other Vibrio species based on the formation of
purple colonies by V. parahaemolyticus (29, 64). To distin-
guish V. vulnificus from other pathogens that cause necro-
tizing fasciitis, Nakashima et al. (65) have developed a
selective isolation culture agar plate (Chromochecker Vibrio
Agar) for use in environmental monitoring and the clini-
cal setting. There are a few other chromogenic media for
detection of Vibrio spp., which shows different color between
V. parahaemolyticus, V. cholera, and V. vulnificus such as VVM
(66), Chromagar Vibrio (28, 67, 68), and Vibrio ID (bioMér-
ieux) and further evaluation studies are needed.
GRAM-POSITIVE BACTERIA
Enterococci
Enterococci are used as indicators of fecal contamination of
water and food. On the basis of 16S rRNA cataloging,
the genus Streptococcus was separated during the 1980s
into the three genera Enterococcus, Lactococcus, and Strepto-
coccus. Consequently, bacteria previously named S. faecalis,
S. faecium, S. avium, and S. gallinarum were transferred to the
revised genus Enterococcus as E. faecalis, E. faecium, E. avium,
and E. gallinarum, respectively (69). More than 100 modifica-
tions of selective media have been described in the past
for isolating fecal streptococci or enterococci from various
specimens. In most literature, Slanetz-Bartley (70) agar is
treated as a suitable MF medium to detect fecal enterococci
from a water sample.
The use of chromogenic and fluorogenic substrates
for detection of enterococci has received considerable
attention (39, 71, 72). Substrates such as 5-bromo-4-
chloro-3-indolyl-ß-D-glucopyranoside (XGLU) and indoxyl-
-ß-D-glucoside were already described for detection of
ß-D-glucosidase activity (40). Enterococci produce an insolu-
ble indigo blue complex, which diffuses into the surrounding
media, forming a blue halo around the colony. Chromocult
Enterococci broth (Merck) and Readycult Enterococci
(Merck) utilize the ß-D-glucosidase reaction as an indicator
for enterococci. XGLU is liberated and rapidly oxidized to
bromochloroindigo, which produces blue color in Chromo-
cult broth. Chromocult enterococci agar (EKA, Merck)
uses a chromogenic mix in a selective agar; enterococci cleave
chromogenic substrates in this medium and show red colonies
on the plate. Nonenterococci produce colorless, blue/violet,
or turquoise colonies (41).
4-Methylumbelliferyl-ß-D-glucoside (MUD) has been
proven useful to detect enterococci in water. This substrate,
when hydrolyzed by enterococcal ß-D-glucosidase, releases 4-
methylumbelliferone, which exhibits fluorescence under a
UV lamp (365 nm). Enterolert (IDEXX Laboratories), and
a Microtiter plate MUST (Sanofi, France) utilize the nutrient
indicator substrate MUD as a fluorogenic substrate detecting
ß-D-glucosidase. The MUD method tended to give slightly
lower recoveries than the agar cultivation methods with
some target species at 44°C, but recoveries were better than
at 41°C (73). It has to be mentioned that there may be
some problems with other glucosidase-positive bacteria
such as Aerococcus spp., in particular in natural contaminated
water samples.
Membrane-Enterococcus indoxyl-ß-D-glucoside agar (mEI
agar) is a selective culture medium used for the chromogenic
detection and enumeration of enterococci in water by the
single-step membrane filtration technique. It conforms to
the U.S. Environmental Protection Agency Approved
Method 1600: Enterococci in Water by Membrane Filtration.
Rhodes and Kator (74) evaluated mEI agar with respect to
specificity and recovery of enterococci from environmental
waters. Extending incubation from 24 to 48 h improved
enterococci recovery, but 77% of the colonies classified as
nontarget were confirmed as enterococci. Messer and Dufour
(75) modified the medium by reducing the triphenyltetrazo-
lium chloride from 0.15 to 0.02 g/liter and adding 0.75 g
indoxyl ß-D-glucoside per liter. The new MF medium, mEI
medium, detected levels of enterococci in 24 h comparable
to those detected by the original mE medium in 48 h, with
the same level of statistical confidence.
Using XGLU Chromocult Enterococci broth (CEB)
(Merck) and Readycult Enterococci (Merck) utilizes the
ß-D-glucosidase reaction as an indicator of the presence of
enterococci. XGLU is liberated and rapidly oxidized to bro-
mochloroindigo which produces blue color in Chromocult
broth as well as blue colored enterococci colonies in Chromo-
cult Enterococci agar. Other ß-D-glucosidase-producing
organisms were suppressed by the sodium azide content of
the broth (40). The results obtained with pure cultures
showed that 97% of the strains, which gave positive results,
were identified as enterococci (E. fecalis, E. faecium,
E. durans, E. casseliflavus, and E. avium). The false-positive
strains (3%) were Leucononostoc, Lactococcus lactis lactis,
and Aerococcus spp.
Clostridium perfringens
C. perfringens is an anaerobic, Gram-positive, spore-forming
rod-shaped bacterium. They are widespread in the environ-
ment and also found in the digestive systems of humans and
animals. C. perfringens is capable of surviving in soil and water
for extended periods of time. Detection and enumeration of
C. perfringens from food and environmental samples is usually
based on cultivation on agar plates, but an MPN technique
can also be applied. Several selective culture media for the
detection and enumeration of C. perfringens are based on
sulfite reduction as a differential characteristic and cycloser-
ine as a selective agent. Selectivity of cultivation can be
increased by using an elevated incubation temperature, since
C. perfringens is able to grow rapidly at 45°C. The shortcom-
ing of the selective culture and growth conditions is the poor
recovery of injured vegetative cells of C. perfringens.
TSC-4-methylumbelliferyl phosphate (MUP) agar incor-
porates sodium metabisulfite and ferric ammonium citrate as
an indicator for sulfite reduction and the disodium salt of
MUP for detection of acid phosphatase. D-Cycloserine inhib-
its the accompanying bacterial flora and causes the colonies
to remain smaller. All black colonies on this medium that
emit light blue fluorescence when exposed to a UV lamp
(365 nm) were counted as presumptive C. perfringens. The
disadvantage of incorporating MUP into agar is that fluores-
cence diffuses rapidly from the colonies into the surrounding
agar, and therefore the interpretation of the plates is difficult
and confusing. The use of (MUP and ortho-nitrophenyl-
β-d-galactopyranoside for the identification of C. perfringens
was investigated (76). A liquid assay containing both
compounds was a highly specific alternative method for
C. perfringens confirmation, reducing incubation time from
48 to only 4 h. The new Chromoselect CP agar is now avail-
able, which can be used in the water microbiology (45).
C. perfringens colonies give green-colored colonies, other
2.1.1-6 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
strains of clostridia give violet or blue colonies (Sigma,
Schwitzerland).
Clostridium dificille
C. difficile is the major cause of colitis and pseudomembranous
colitis associated with antibiotic treatment. ChromID
C. difficile agar (bioMérieux) is a new chromogenic medium
for isolation and identification of C. difficile in 24 h (46).
Listeria monocytogenes
L. monocytogenes is a human and animal pathogen that is
widespread in nature. The organism is a transient constituent
of the intestinal flora excreted by 1–10% of healthy humans.
This organism can survive for many years in the cold in
naturally infected sources. L. monocytogenes has been isolated
from a wide variety of foods, including dairy products, meats,
and fish. Although most of the foodborne listeriosis out-
breaks have been linked to the consumption of dairy prod-
ucts, recent sporadic cases have been associated with
meats as well as other foods. Several virulence genes have
been identified in pathogenic Listeria spp. The lecithinase
operon contains the gene plcB. PlcB encodes a lecithinase
involved in cell-to-cell spread. The gene plcA encodes
a phosphatidylinositol-specific phospholipase C (PI-PLC)
that may contribute to the lysis of the phagosomal membrane.
These virulence genes ( plcA, plcB) occur in pathogenic
L. monocytogenes and L. ivanovii (77, 78).
The detection of L. monocytogenes in food and environ-
mental samples by cultivation includes enrichment step(s)
for resuscitation of injured cells and concentration of the
cells, followed by plating on selective media and confirmation
of the tentative identifications of suspected colonies by bio-
chemical tests. Current conventional culture techniques
take approximately 1 week to complete. The recent method-
ology development has focused on the optimization of
enrichment steps and the development of new differential
culture media to obtain faster and more reliable detection of
L. monocytogenes. Selective chromogenic L. monocytogenes
plating media offer rapid economic detection and enumera-
tion of pathogenic Listeria spp. within 24 or 48 h of incuba-
tion at 36 ± 1°C.
Several studies have been done using chromogenic media
for the detection of L. monocytogenes (34). Their advan-
tages are direct detection and enumeration of pathogenic Lis-
teria spp. utilizing cleavage of substrates by the virulence
factor PI-PLC and by phosphatidylcholinphospholipase C
(PC-PLC). There are two groups of such media: the first uti-
lizes cleavage by PI-PLC of L-alpha-phosphatidyl-inositol,
forming a white precipitation zone around the colony,
combined with the chromogenic substrate 5-bromo-4-
chloro-3-indoxyl-beta-D-glucopyranoside for detection of
ß-D-glucosidase, which occurs in all Listeria spp. The typical
colony morphology of Listeria spp. is reported to be turquoise
blue (ALOA, CHROMagar Listeria). Pathogenic Listeriae are
additionally surrounded by a translucent halo.
The second group of media such as BCM, Rapid’ L.
mono, and LIMONO-Ident-agar utilizes enzyme substrate
5-bromo-4-chloro-3-indoxyl-myo-inositol-1-phosphate (X-
phos-Inositol, BIOSYNTH). Enzymatic cleavage of
X-phos-Inositol leads to turquoise colonies of pathogenic Lis-
teria spp. easy to enumerate (L. monocytogenes and L. ivanovii).
Nonpathogenic Listeria spp. appears clearly distinguishable as
white colonies. BCM L. monocytogenes Plating Medium II and
LIMONO-Ident-Agar additionally combine the cleavage of
X-phos-Inositol in forming turquoise colonies of pathogenic
Listeria spp. with production of a white precipitate surrounding
these colonies. This precipitate is formed by addition of a
selected lecithin–mixture (79, 80). For that reason, both vir-
ulence genes ( plcA and plcB) may be involved in forming this
precipitation zone (33). Rapid L’MONO agar (Sanofi,
France) is based on the detection of phosphatidylinositol
phospholipase C and xylose fermentation. In this way, the col-
onies of L. ivanovii appear blue sourrounded by a yellow halo
(xylose positive) whereas the colonies of L. monocytogenes
are blue without the halo (xylose negative).
Stessel et al. (81) evaluated six chromogenic media
similar to ALOA. Additionally, the ability of chromogenic
agars to facilitate growth of stressed L. monocytogenes strains
and mixed cultures with competitive non-Listeria strains was
estimated. They have found that the competitive flora of food
samples is able to overgrow low numbers of L. monocytogenes,
especially in half-Fraser enrichment. This might lead to the
underestimation of L. monocytogenes positive samples.
CHROMagar Listeria agar (CHROMagar), OCLA
(Oxoid), Compass L. mono (Biokar), OAA (bioMérieux),
and Chromoplate Listeria (Merck) are some of this kind of
chromogenic media.
B. cereus, B. thuringiensis, and B. anthracis Group
Bacillus species is a group of ubiquitous facultative anaerobic
spore-forming Gram-positive rods commonly found in soil.
The spores frequently contaminate a variety of foods, includ-
ing produce, meat, eggs, and dairy products. Several selec-
tive and nonselective culture media have been developed
for the detection of B. cereus from foods. The enumeration
of B. cereus in food and industrial samples is commonly based
on a plate-counting culture technique, except for samples
with low cell numbers (<10 CFU/g), for which the MPN
method is preferred. BCM Bacillus cereus group Plating
Medium (Biosynth) utilizes the production of the enzyme
PI-PLC cleaving substrate 5-bromo-4-chloro-3-indoxyl
myo-inositol-1-phosphate by members of the B. cereus group
(except B. anthracis). Typical dull blue-turquoise colonies
appear on this plating medium, easy to recognize and count
(35). B. anthracis grows on BCM Bacillus cereus group plating
medium in size and shape like B. cereus, but with white
colonies.
B. anthracis chromogenic agar (ACA) has been evaluated
by Juergensmeyer et al. (36). ACA contains the chromogenic
substrate 5-bromo-4-chloro-3-indoxyl-choline phosphate
that on hydrolysis yields blue green colonies indicating the
presence of PC-PLC activity. Examination of colony mor-
phology in 18 pure culture strains of B. anthracis required
48 h at 35–37°C for significant color production, whereas
only 24 h was required for B. cereus and B. thuringiensis.
This differential rate of PC-PLC synthesis in B. anthracis
(due to the truncated PlcR gene and PlcR regulator in
B. anthracis) allowed for the rapid differentiation on ACA
of presumptive colonies of B. anthracis from B. cereus and
B. thuringiensis in both pure and mixed cultures. The perform-
ance of two chromogenic plating media (CBC and BCM) was
compared with two standard selective plating media
(PEMBA and MYP) recommended by food authorities for
isolation, identification, and enumeration of B. cereus (37).
The Chromogenic Bacillus Cereus agar (CBC; Oxoid) con-
tains 5-bromo-4-chloro-3-indolyl-β-glucopyranoside that is
cleaved by β-D-glucosidase and results in white colonies
with a blue-green center. Tallent et al. (82) evaluated the
use of Bacara, a new chromogenic agar, as an efficient method
to identify and enumerate B. cereus group from food matrixes,
even in the presence of background flora.
 2.1.1. Detection of Specific taxa Using Chromogenic and Fluorogenic Media ▪ 2.1.1-7
Staphylococcus aureus
S. aureus is a major pathogenic bacterium found in clinical
field and in food industry. Nosocomial infections due to
S. aureus create an increasing number of problems, so it is
becoming more important to detect S. aureus in foods and
clinical microbiology. CHROMagar S. aureus is a selective
agar medium that employs a combination of chromogenic
enzyme substrates. S. aureus strains grow as typical pink-
colored colonies on this medium, whereas most other staph-
ylococci produce white or occasionally blue colonies (30).
S. aureus ID (bioMérieux) is similar, and the colonies of
S. aureus are distinctive green colonies because of production
of α-glucosidase. Other staphylococci form white or occa-
sionally pink colonies due to the hydrolysis of a substrate
for ß-glucosidase (31).
BIFIDOBACTERIA
The use of chromogenic X-α-D-galactoside for differential
enumeration of Bifidobacterium spp. and lactic acid bacteria
was described by Chevalier et al. (83). The X-α-D-
galactoside-based medium is useful to identify Bifidobacteria
among Lactobacillus or Streptococcus strains. Bifidobacteria
show blue-colored colonies on the agar plate.
Methicillin-Resistant S. aureus (MRSA)
MRSA is a nosocomial pathogen of worldwide importance
and an increasingly frequent cause of community-acquired
infection. Rapid laboratory diagnosis is critical for treating,
managing, and preventing MRSA infections. Different media
are available (32) to detect MRSA such as Brilliance MRSA
agar (Oxoid), ChromID (bioMérieux), MRSASelect (Bio-
Rad), CHROMagar, Spectra MRSA (Remel, Lenexa, KS),
and BBL-CHROMagar (BD Diagnostics).
Vancomycin-Resistant Enterococci (VRE)
Strains of enterococci are known to be nosocomial patho-
gens, which can be involved in infections of the urinary tract,
surgical wounds, and the bloodstream. An increasing num-
ber of these infections are caused by enterococci that are
resistant to vancomycin and to other antibiotics. Chromo-
genic media are available (43, 44) to detect VRE, such as
chromID VRE agar (bioMérieux) and CHROMagar VanRE
(BD Diagnostics).
CONCLUSIONS
Rapid detection and identification of microorganisms is of
high importance in a diverse array of applied and research
fields. Chromogenic and fluorogenic substrates have been
used in diagnosstic culture media for the past two decades.
The incorporation of such substrates into a selective medium
can eliminate the need for subculture and further biochemi-
cal tests to establish the identity of certain microorganisms.
REFERENCES
1. Manafi M. 2000. New developments in chromogenic
and fluorogenic culture media. Int J Food Microbiol 60:
205–218.
2. Manafi M, Kneifel W, Bascomb S. 1991. Fluorogenic and
chromogenic substrates used in bacterial diagnostics. Micro-
biol Rev 55:335–348.
3. Orenga S, James AL, Manafi M, Perry JD, Pincus DH.
2009. Enzymatic substrates in microbiology. J Microbiol
Meth 79:139–155.
4. Druggan P. 2012. Chromogens, fluorogens, trojan horses
and their use in culture media, p. 623–646. In Corry JEL,
Curtis GDW, Baird RM (eds), Handbook of Culture Media
for Food and Water Microbiology, 3rd ed. RSC Publishing,
UK.
5. Bredie WL, de Boer E. 1992. Evaluation of the MPN,
Anderson-Baird-Parker, Petrifilm E. coli and Fluorocult
ECD method for enumeration of E. coli in foods of animal
origin. Int J Food Microbiol 16:197–208.
6. Hartman PA. 1989. The MUG glucuronidase test for E. coli
in food and water, p. 290–308. In Turano A, (ed), Rapid
Methods and Automation in Microbiology and Immunology-
Brixia Academic Press, Brescia, Italy.
7. Frampton EW, Restaino L. 1993. Methods for E. coli iden-
tification in food, water and clinical samples based on
ß-glucuronidase detection. J Appl Bacteriol 74:223–233.
8. Manafi M. 1995. New medium for the simultaneous detec-
tion of total coliforms and E. coli in water. Poster, 95th Meet-
ing of the American Society for Microbiology. Washington,
DC. Abstr. P-43, p. 389.
9. Manafi M, Rosmann H. 1998. Evaluation of Readycult
presence-absence test for detection of total coliforms and
E. coli in water. Poster, 98th Meeting of the American
Society for Microbiology, Atlanta, GA. Abstr. Q-263,
p. 464.
10. Brenner KP, Rankin CC, Roybal YR, Stelma GN, Scar-
pino PV, Dufour AP. 1993. New medium for simultaneous
detection of total coliforms and E. coli in water. Appl Environ
Microbiol 59:3534–3544.
11. Hahn G, Wittrock E. 1991. Comparison of chromogenic
and fluorogenic substances for differentiation of coliforms
and E. coli in soft cheese. Acta Microbiol Hung 38:265– 271.
12. Alonso JL, Soriano K, Amoros I, Ferrus MA. 1998. Quan-
titative determination of E. coli and fecal coliforms in water
using a chromogenic medium. J Environ Sci Health 33:
1229–1248.
13. Feldsine PT, Falbo-Nelson MT, Hustead D. 1994. Coli-
complete® substrate-supporting disc method for confirmed
detection of total coliforms and E. coli in all foods, Compa-
rative study. J AOAC Int 77:58–63.
14. Grant MA. 1997. A new membrane filtration medium for
simultaneous detection and enumeration of E. coli and total
coliforms. Appl Environ Microbiol 63:3526–3530.
15. Bettelheim KA. 1998. Reliability of CHROMagar O157 for
the detection of enterohaemorrhagic E. coli EHEC O157 but
not EHEC belonging to other serogroups. Appl Environ
Microbiol 85:425–428.
16. Bettelheim KA. 1998. Studies of E. coli cultured on Rain-
bow agar O157 with particular reference to enterohaemor-
rhagic E. coli (EHEC). Microbiol Immunol 42:265–269.
17. Bochner BR. 1995. Rainbow Agar VTEC, a new chromo-
genic culture medium for detecting verotoxin-producing
E. coli. Poster, Australian Society for Microbiology, Poster 2.1.
18. Restaino L, Frampton EW, Turner KM, Allison DR. 1999.
A chromogenic plating medium for isolating E. coli O157:
H7 from beef. Lett Appl Microbiol 29:26–30.
19. Monnery I, Freydière AM, Baron C, Rousset AM, Tigaud
S, Boude-Chevalier M, de Montclosand H, Gille Y. 1994.
Evaluation of two new chromogenic media for detection of
Salmonella in stools. Eur J Clin Microbiol Infect Dis 13:
257–261.
20. Gaillot O, DiCamillo P, Berche P, Courcol R, Savage C.
1999. Comparison of CHROMagar Salmonella medium
and Hektoen enteric agar for isolation of Salmonellae from
stool samples. J Clin Microbiol 37:762–765.
21. Perry JD, Ford M, Taylor J, Jones AL, Freeman R, Gould
FK. 1999. ABC medium, a new chromogenic agar for the
selective isolation of Salmonella spp. J Clin Microbiol 37:
766–768.
2.1.1-8 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
22. Rambach A. 1990. New plate medium for facilitated differ-
entiation of Salmonella spp. from Proteus spp. and other
enteric bacteria. Appl Environ Microbiol 56:301–303.
23. Schönenbrücher V, Mallinson ET, Bülte M. 2008. A com-
parison of standard cultural methods for the detection of
foodborne Salmonella species including three new chromo-
genic plating media. Int J Food Microbiol 123:61–66.
24. Song KY, Hyeon JY, Shin HC, Park CK, Choi IS, Seo
KH. 2008. Evaluation of a chromogenic medium supple-
mented with glucose for detecting Enterobacter sakazakii.
J Microbiol Biotechnol 18:579–584.
25. Restaino L, Frampton EW, Lionberg WC, Becker RJA.
2006. Chromogenic plating medium for the isolation and
identification of Enterobacter sakazakii from foods, food
ingredients, and environmental sources. J Food Prot 69:
315–322.
26. Renaud N, Lecci L, Courcol RJ, Simonet M, Gaillot O.
2013. CHROMagar Yersinia, a new chromogenic agar for
screening of potentially pathogenic Yersinia enterocolitica iso-
lates in stools. J Clin Microbiol 51:1184–1187.
27. Willems E, Cartuyvels R, Magerman K, Verhaegen J.
2013. Evaluation of 3 different agar media for rapid detection
of extended-spectrum β-lactamase-producing Enterobacteria-
ceae from surveillance samples. Diagn Microbiol Infect Dis 76:
16–19.
28. Hara-Kudo Y, Nishina T, Nakagawa H, Konuma H, Hase-
gawa J, Kumugai S. 2001. Improved method for detection of
Vibrio parahaemolyticus in seafood. Appl Environ Microbiol
67:5819–5823.
29. Su YC, Duan J, Wu WH. 2005. Selectivity and specificity
of a chromogenic medium for detecting Vibrio parahaemoly-
ticus. J Food Prot 68:1454–1456.
30. Gaillot O, Wetsch M, Fortineau N, Berche P. 2000. Eval-
uation of CHROMagar S. aureus, a new chromogenic
medium, for isolation and presumptive identification of
Staphylococcus aureus from human clinical specimens.
J Clin Microbiol 38:1587–1591.
31. Perry JD, Rennison C, Butterworth LA, Hopley AL,
Gould FK. 2003. Evaluation of S. aureus ID, a new chromo-
genic agar medium for detection of Staphylococcus aureus.
J Clin Microbiol 41:5695–5698.
32. Denys GA, Renzi PB, Koch KM, Wissel CM. 2013. Three-
way comparison of BBL CHROMagar MRSA II, MRSA
Select, and spectra MRSA for detection of methicillin-
resistant Staphylococcus aureus isolates in nasal surveillance
cultures. J Clin Microbiol 51:202–205.
33. Restaino L, Frampton EW, Irbe RM, Schabert G, Spitz
H. 1999. Isolation and detection of L. monocytogenes
using fluorogenic and chromogenic substrates for phosphati-
dylinositol-specific phospholipase C. J Food Protect 62:
244–251.
34. Reissbrodt R. 2004. New chromogenic plating media for
detection and enumeration of pathogenic Listeria spp.-an
overview. Int J Food Microbiol 95:1–9.
35. Peng H, Ford V, Frampton EW, Restaino L, Shelef LA,
Spitz H. 2001. Isolation and enumeration of Bacillus cereus
from foods on a novel chromogenic plating medium. Food
Microbiol 18:231–238.
36. Juergensmeyer MA, Gingras BA, Restaino L, Frampton
EW. 2006. A selective chromogenic agar that distinguishes
Bacillus anthracis from Bacillus cereus and Bacillus thuringiensis.
J Food Prot 69:2002–2006.
37. Fricker M, Reissbrodt R, Ehling-Schulz M. 2008. Evalua-
tion of standard and new chromogenic selective plating
media for isolation and identification of Bacillus cereus.
Int J Food Microbiol 121:27–34.
38. Marston CK, Beesley C, Helsel L, Hoffmaster AR. 2008.
Evaluation of two selective media for the isolation of Bacillus
anthracis. Lett Appl Microbiol 47:25–30.
39. Dufour AP. 1980. A 24-hour membrane filter procedure for
enumerating enterococci. Abstract, 80th American Society
for Microbiology, Washington, DC. Q-69, p. 205.
40. Manafi M, Windhager K. 1997. Rapid identification of
enterococci in water with a new chromogenic assay. Poster,
97th Meeting of the American Society for Microbiology,
Miami, FL. Abstr. P-107, p. 453.
41. Miranda JM, Franco CM, Vázquez BI, Fente CA, Barros-
Velázquez J, Cepeda A. 2005. Evaluation of Chromocult
enterococci agar for the isolation and selective enumeration
of Enterococcus spp. in broilers. Lett Appl Microbiol 41:
153–156.
42. Budnick GE, Howard RT, Mayo DR. 1996. Evaluation of
Enterolert for enumeration of enterococci in recreational
waters. Appl Environ Microbiol 62:3881–3884.
43. Cuzon G, Naas T, Fortineau N, Nordmann P. 2008. Novel
chromogenic medium for detection of vancomycin-resistant
Enterococcus faecium and Enterococcus faecalis. J Clin Micro-
biol 46:2442–2444.
44. Stamper PD, Shulder S, Bekalo P, Manandhar D, Ross TL,
Speser S, Kingery J, Carroll KC. 2010. Evaluation of BBL
CHROMagar VanRE for detection of vancomycin-resistant
Enterococci in rectal swab specimens. J Clin Microbiol 48:
4294–4297.
45. Manafi M, Waldherr K, Kundi M. 2013. Evaluation of CP
Chromo Select Agar for the enumeration of Clostridium per-
fringens from water. Int J Food Microbiol 167:92–95.
46. Eckert C, Burghoffer B, Lalande V, Barbut F. 2013. Eval-
uation of the chromogenic agar chromID C. difficile. J Clin
Microbiol 51:1002–1004.
47. Schets FM, Medema GJ, Havelaar AH. 1993. Comparison
of Colilert with Dutch standard enumeration methods for
E. coli and total colifoms in water. Lett Appl Microbiol 17:
17–19.
48. Lee JV, Lightfoot NF, Tillett HE. 1995. An evaluation of
presence/absence tests for coliform organisms and E. coli.
Int. Conf. on Coliforms and E. coli, Problem or Solution?,
Leeds, UK.
49. Byamukama D, Kansiime F, Mach RL, Farnleitner AH.
2000. Determination of E. coli contamination with chromo-
cult coliform agar showed a high level of discrimination effi-
ciency for differing fecal pollution levels in tropical waters of
Kampala, Uganda. Appl Environ Microbiol 66:864–868.
50. Rice EW, Allen MJ, Brenner DJ, Edberg SC. 1991. Assay
for ß-glucuronidase in species of the genus Escherichia and
its application for drinking-water analysis. Appl Environ
Microbiol 57:592–593.
51. Manafi M, Kremsmair B. 2001. Comparative evalua-
tion of different chromogenic/fluorogenic media for detect-
ing E. coli O157H:7 in food. Int J Food Microbiol 71:
257–262.
52. Vernozy-Rozand C, Mazuy-Cruchaudet C, Bavai C, Mon-
tet MP, Bonin V, Dernburgand A, Richard Y. 2005.
Growth and survival of E coli O157:H7 during the manufac-
ture and ripening of raw goat milk lactic cheeses. Int J Food
Microbiol 105:83–88.
53. Kühn H, Wonde B, Rabsch W, Reissbrodt R. 1994. Eval-
uation of Rambach agar for detection of Salmonella subspe-
cies I to VI. Appl Environ Microbiol 60:749–751.
54. Humbert F, Salvat G, Colin P, Lahellec C, Bennejean G.
1989. Rapid identification of Salmonella from poultry
meat products by using “Mucap test”. Int J Food Microbiol
8:79–83.
55. Warren BR, Parish ME, Schneider KR. 2005. Comparison
of chromogenic Shigella spp. plating medium with standard
media for the recovery of Shigella boydii and Shigella sonnei
from tomato surfaces. J Food Prot 68:621–624.
56. García-Aguayo JM, Ubeda P, Gobernado M. 1999. Evalu-
ation of xylose-galactosidase medium, a new plate for the
 2.1.1. Detection of Specific taxa Using Chromogenic and Fluorogenic Media ▪ 2.1.1-9
isolation of Salmonella, Shigella, Yersinia and Aeromonas spe-
cies. Eur J Clin Microbiol Infect Dis 18:77–78.
57. Druggan P, Iversen C. 2009. Culture media for the isola-
tion of Cronobacter spp. Inter J Food Microbiol 136:169–178.
58. Iversen C, Forsythe SJ. 2007. Comparison of media for the
isolation of Enterobacter sakazakii. Appl Environ Microbiol 73:
48–52.
59. Lehner A, Nitzsche S, Breeuwer P, Diep B, Thelen K,
Stephan R. 2006. Comparison of two chromogenic
media and evaluation of two molecular based identification
systems for Enterobacter sakazakii detection. BMC Microbiol
6:15.
60. Leuschner RG, Bew J. 2004. A medium for the presumptive
detection of Enterobacter sakazakii in infant formula: interla-
boratory study. J AOAC Int 87:604–613.
61. Oh SW, Kang DH. 2004. Fluorogenic selective and differen-
tial medium for isolation of Enterobacter sakazakii. Appl Envi-
ron Microbiol 70:5692–5694.
62. Kola A, Kohler C, Pfeifer Y, Schwab F, Kühn K, Schulz K,
Balau V, Breitbach K, Bast A, Witte W, Gastmeier P,
Steinmetz I. 2012. High prevalence of extended-spectrum-
β-lactamase-producing Enterobacteriaceae in organic and
conventional retail chicken meat, Germany. J ntimicrob Che-
mother 67:2631–2634.
63. Weagant SD. 2008. A new chromogenic agar medium for
detection of potentially virulent Yersinia enterocolitica.
J Microbiol Meth 72:185–190.
64. Duan J, Su YC. 2005. Comparison of a chromogenic
medium with thiosulfate-citrate-bile salts-sucrose agar for
detecting Vibrio parahaemolyticus. J Food Sci 70:125–128.
65. Nakashima Y, Oho M, Kusaba K, Nagasawa Z, Komatsu
O, Manome I, Araki K, Oishi H, Nakashima M. 2007. A
chromogenic substrate culture plate for early identification
of Vibrio vulnificus and isolation of other marine Vibrios.
Ann Clin Lab Sci 37:330–334.
66. Cerdà-Céllular M, Jofr J, Blanch AR. 2000. A selective
media and a specific probe for detection of V. vulnificus.
Appl Environ Microbiol 66:855–859.
67. Blanco-Abad V, Ansede-Bermejo J, Rodriguez-Castro A,
Martinez-Urtaza J. 2009. Evaluation of different procedures
for the optimized detection of Vibrio parahaemolyticus in mus-
sels and environmental samples. Int J Food Microbiol 129:
229–236.
68. Rosec JP, Causse V, Cruz B, Rauzier J, Carnat L. 2012.
The international standard ISO/TS 21872–1 to study the
occurence of total and pathogenic Vibrio parahaemolyticus
and Vibrio cholerae in seafood: ITS improvement by use of
a chromogenic medium and PCR. Int J Food Microbiol
157:189–194.
69. Schleifer KH, Killper-Bälz R. 1984. Transfer of Streptococ-
cus faecalis and Streptococcus faecium to the genus Enterococ-
cus nom. rev. as Enterococcus faecalis comb. nov. and
Enterococcus faecium comb. nov. Int J Syst Bacteriol 34:
31–34.
70. Slanetz LW, Bartley CH. 1957. Numbers of enterococci in
water, sewage, and feces determined by the membrane filter
technique with an improved medium. J Bacteriol 74:
591–595.
71. Hernandez JF, Pourcher AM, Delattre JM, Oger C, Loeuil-
lard JL. 1993. MPN miniaturized procedure for the enumer-
ation of faecal enterococci in fresh and marine waters, the
MUST procedure. Wat Res 27:597–606.
72. Littel K, Hartman PA. 1983. Fluorogenic selective and dif-
ferential medium for isolation of faecal streptococci. Appl
Environ Microbiol 45:622–627.
73. Niemi RM, Ahtiainen J. 1995. Enumeration of intestinal
enterococci and interfering organisms with Slanetz-Bartley
agar, KF streptococcus agar and the MUST method. Lett
Appl Microbiol 20:92–97.
74. Rhodes MW, Kator H. 1997. Enumeration of enterococcus
sp. using a modified mE method. J Appl Microbiol 83:
120–126.
75. Messer W, Dufour A. 1998. A rapid, specific membrane fil-
tration procedure for enumeration of enterococci in recrea-
tional water. Appl Environ Microbiol 62:3881–3884.
76. Adcock PW, Saint CP. 2001. Rapid confirmation of C. per-
fringens by using chromogenic and fluorogenic substrates.
Appl Environ Microbiol 67:4382–4384.
77. Camilli A, Goldfine H, Portnoy DA. 1991. L. monocyto-
genes mutants lacking phosphatidylinositol-specific phos-
pholipase C are avirulent. J Exp Med 173:751–754.
78. Gouin E, Mengaud J, Cossart P. 1994. The virulence gene
cluster of L. monocytogenes is also present in Listeria ivano-
vii, an animal pathogen, and Listeria seeligeri, a non-
pathogenic species. Infect Immun 62:3550–3553.
79. Hechelmann H, Albert T, Reissbrodt R. 2002. Character-
ization of new chromogenic plating media for detection and
enumeration of L. monocytogenes. Poster , “Food Micro,” Lil-
lehammer, Norway.
80. Roczen D, Knödlseder M, Friedrich K, Schabert G, Spitz
H, Müller R, Reissbrodt R. 2003. Development of a new
chromogenic L. monocytogenes plating medium and compar-
ison with three other chromogenic plating media. Poster,
Meeting of the American Society for Microbiology, Wash-
ington DC, Poster O-48.
81. Stessl B, Luf W, Wagner M, Schoder D. 2009. Performance
testing of six chromogenic ALOA-type media for the detec-
tion of L. monocytogenes. J Appl Microbiol 106:651–659.
82. Tallent SM, Kotewicz KM, Strain EA, Bennett RW. 2012.
Efficient isolation and identification of Bacillus cereus group.
J AOAC Int 95:446–451.
83. Chevalier P, Royand D, Savoie L. 1991. X-α-Gal-based
medium for simultaneous enumeration of bifidobacteria and
lactic acid bacteria in milk. J Microbiol Meth 13:75–83.
 Anaerobic Cultivation
TAKASHI NARIHIRO AND YOICHI KAMAGATA
2.1.2
Anaerobes are key players in global cycles of elements and
nutrition in natural and anthropologic ecosystems and are
also causative agents of human and veterinary diseases.
Numerous studies have sought to develop culturing techni-
ques for anaerobes to enable the elucidation of their basic
physiology, pathogenic mechanisms, and ecological func-
tions. This chapter describes a brief history of the develop-
ment of anaerobic culturing techniques from the historical
Hungate technique to techniques and apparatuses commonly
used in modern laboratories. In addition, recent progress in
anaerobic culturing techniques (e.g., single-cell manipula-
tion and isolation, the six-well plating method, the coculture
method, bioreactor-based enrichment, and in situ/in vivo cul-
tivation) are described, with several examples of the applica-
tion of these techniques for the isolation of anaerobes from
natural and artificial ecosystems.
Anaerobic microorganisms (hereafter designated anae-
robes) are key players in global cycles of elements (e.g., carbon,
nitrogen, sulfur, and iron) in natural and anthropogenic ecosys-
tems, including aquifer/groundwater (1), animal/insect guts
(2, 3), rice paddy fields (4), rumen (5), terrestrial/oceanic sub-
surfaces (6, 7), wetlands (8), biodiesel production processes (9),
food production processes (10), and wastewater treatment
processes (11). In these anoxic ecosystems, various metabolic
groups of anaerobes play roles in specific biochemical reactions,
such as fermentation (12, 13), anaerobic ammonium oxidation
(anammox) (14), sulfate reduction (15), iron reduction (16),
syntrophic substrate oxidation (17), and methanogenesis
(18). In addition to these metabolic groups, anaerobes include
human and animal pathogens, which were the main focus in
the earliest era of modern microbiology. The challenges of cul-
tivating, characterizing, and controlling these anaerobic patho-
gens organisms spurred the establishment of modernized
microbiology (19–23).
A crucial difference between aerobic and anaerobic
culturing techniques is that creating and maintaining an
anoxic (O2-free) condition is a prerequisite for the growth
of anaerobes. This prerequisite long hindered the cultivation
of pure cultures. However, a number of researchers have suc-
cessfully isolated anaerobes using various means through a
trial-and-error process. This chapter describes the early his-
tory and current status of anaerobic culturing techniques.
The chapter comprises four sections: (a) a brief history of
anaerobic culturing techniques, tracing a path through the
development of the gold standard Hungate technique; (b)
doi:10.1128/9781555818821.ch2.1.2
2.1.2-1
apparatuses (e.g., gassing manifold, anaerobic chamber, and
jar-type container) and reducing agents currently used in
common laboratories; (c) successful examples using the Hun-
gate technique, specifically those involving the isolation of
anaerobes that had long been refractory to isolation, referred
to as-yet-uncultured organisms; and (d) novel techniques for
culturing anaerobic microorganisms.
EARLY HISTORY OF ANAEROBIC CULTURING
TECHNIQUES AND THE DEVELOPMENT OF
THE HUNGATE TECHNIQUE
Anaerobic cultivation was first described by Pasteur in 1861
(24), who observed the growth of microorganisms involved
in butyric fermentation under anoxic conditions produced
by a vacuum pump, while no growth was observed in the pres-
ence of air. Forty years later, more than 50 primitive but crea-
tive methods for culturing anaerobes had been developed and
were compiled by Hunziker in a review article in 1902 (25).
In this 100-year-old seminal review, anaerobic culturing
techniques were categorized on the basis of six concepts: (a)
formation of a vacuum; (n) replacement of air by inert anoxic
gases; (c) absorption of oxygen; (d) reduction of oxygen; (e)
exclusion of atmospheric oxygen by means of various physical
principles and mechanical devices; and (f ) the combined
application of any two or more of the foregoing principles.
Surprisingly, these concepts still underlie current techniques.
By the mid-20th century, anaerobic culturing techniques
had become varied and disorganized: for example, there
were reports of the use of pyrogallol, potassium hydroxide,
phosphorous, cysteine, and thioglycollate to remove oxy-
gen (26–28), whereas other studies used an inert gas-filled
jar without reducing agents (29–31) or petri dishes com-
bined with reducing agents (32, 33). In addition to these
approaches, Esmarch (34) developed the roll tube technique
in 1886; however, this technique did not become popular
because it was laborious and time-consuming compared
with the petri dish method (33, 35). Despite these difficulties,
Wilson (36) implemented the roll tube technique for viable
cell counting, and various types of mechanical devices for
producing a roll tube with a uniform distribution of solidi-
fied medium were developed (35, 37–39). These enormous
efforts were eventually integrated by Hungate to establish a
sophisticated culturing technique for fastidious anaerobes,
the so-called Hungate roll tube technique (40). The major
2.1.2-2 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
principles of the Hungate technique are (a) removal of oxy-
gen from the medium; (b) mimicking the conditions in which
the microbes were found (e.g., composition of the medium,
pH buffering, and redox potential); (c) minimization of oxy-
gen exposure during inoculation; and (d) rapid solidification
of agar with cold water either in a mechanical device or man-
ually (41). This technique was initially used for the isolation
of anaerobic cellulose-degrading bacteria from ruminal envi-
ronments (40, 41). By the end of the 1970s, the Hungate
technique had been further improved by using prereduced
anaerobically sterilized (PRAS) media (42), plugging the
tube with butyl rubber stoppers (43), sealing with crimp-
closure aluminum seals (44) or screw caps (45), inoculating
with syringes and needles (46), and using pressurized (e.g.,
2–3 atm) tubes or serum bottles (47) (Fig. 1). The improved
Hungate technique has now become a mainstream methodol-
ogy for the culturing of anaerobic microbes, including metha-
nogens (48, 49), syntrophic bacteria (50, 51), fermentative
bacteria (52), and anaerobic fungi (53).
BASIC APPARATUSES AND REAGENTS FOR
ANAEROBIC CULTIVATION
Gassing Manifold
In terms of routine preparation for anaerobic cultivation,
preparing many bottles/tubes of PRAS media is a time-
consuming process. To overcome this problem, Balch and
Wolfe (47) introduced a multichannel gassing manifold in
1976. This type of gassing manifold, including the Deoxygen-
ized Gas Pressure Injector types IP-8 and GR-8 (Sanshin
Industrial, Yokohama, Japan), is now commercially available
from several manufacturers (Fig. 1c). In these systems, high-
pressure gas cylinders are connected to a reactor containing
a copper-based catalyst. The anoxic gas passes over the heated
(approximately 220°C) copper reactor. The manifold is
equipped with a three-way valve, which is connected to the
gas injection, a vacuum pump, and eight valves connected
to injection/vacuum needles. The gas pressure in the mani-
fold is maintained using an ordinary two-stage regulator con-
nected to the anoxic gas cylinder. Such gassing manifolds
permit the preparation of large quantities of PRAS bottles/
tubes with an appropriate gas composition and pressure and
are compatible with the Hungate technique. More recently,
Wolfe and Metcalf (54) developed a vacuum-vortex techni-
que for the preparation of PRAS media with the gassing mani-
fold. This simple technique is applicable for small-volume
(approximately 50 ml) test tubes or serum bottles and com-
bines the strong mixing of the vortex machine and the nega-
tive pressure produced by the gassing manifold. The liquid
cultures prepared by this technique yielded a greater cell den-
sity of the methanogenic archaea Methanosarcina barkeri
strain Fusaro Δhpt-WWM85 compared to the Hungate tech-
nique modified by Balch et al. (49).
Anaerobic Chamber (Glove Box) and Anaerobic
Incubator
An anaerobic chamber (or glove box) provides sufficient
workspace for traditional surface cultivation on Petri dishes
under anaerobic conditions. The use of anaerobic chambers
for cultivation was first described in the 1960s (55–58). In
1969, Aranki et al. (59) developed a glove box with clear flex-
ible vinyl plastic to cover the chamber. They used this glove
box for the isolation of anaerobic bacteria from human gin-
giva and mouse cecum and found that the cultures prepared
in the glove box provided adequate recovery of the anaerobic
bacterial colonies from the samples compared with anaerobic
jar-based cultivation. Typical vinyl-type anaerobic chambers
are now available from COY Laboratory Products (Grass
Lake, MI, USA) (Fig. 1j), and the detailed configuration of
the anaerobic chambers was described in a review article by
Speers (60). Subsequently, Balish et al. (61) have introduced
a transparent plastic incubator for culturing anaerobes within
the anaerobic glove box, in effect making the anaerobic glove
box a tool for incubation. In addition, Metcalf et al. (62) have
demonstrated the suitability of an anaerobic intrachamber
incubator outfitted with a vacuum line, two gas-mixture sup-
ply lines, and a rectangular air lock for culturing Methanosar-
cina spp.
Jar-Type Apparatuses and Compact Gas Generators
The archetype of jar-type apparatuses for culturing anaerobes
was invented by Novy in 1893 (63), with the development of
derivatives until the early 20th century (25, 28, 31, 64). The
prototypic Novy jar was connected to a vacuum pump and
hydrogen gas tank via tubes, and a reducing agent (e.g., pyro-
gallol) solution was placed in the bottom of the jar. Although
the combination of vacuum pump, hydrogen gas, and reduc-
ing agent permitted the effective formation of anoxic condi-
tions, the entire system was relatively complicated, and it was
difficult to maintain the anaerobic condition during the incu-
bation period. As an alternative to this complex system, sev-
eral researchers have developed methods of generating anoxic
gas based on a simple chemical reaction and a metallic cata-
lyst. Brewer et al. (65) used sodium borohydride and cobalt
chloride to produce hydrogen gas and also tested the usability
of a mixture of magnesium metal, zinc chloride, and sodium
chloride for this purpose (66). They subsequently developed
an improved, safe jar-type anaerobic system that employs
sodium borohydride, a palladium catalyst, and a carbon diox-
ide tablet (containing sodium bicarbonate and citric acid) to
generate anoxic conditions, with an anaerobic indicator
based on the methylene blue–hydrogen reaction (67, 68).
These studies resulted in the development of commercial dis-
posable anoxic gas generators, such as GasPak (BD, Franklin
Lakes, NJ, USA) (69–71), GasKit (Don Whitley Scientific,
West Yorkshire, UK) (72), AnaeroGen (Oxoid, Hampshire,
UK) (73), Anaerocult (Merck, Darmstadt, Germany)
(74), and AnaeroPack (Mitsubishi Gas Chemical, Tokyo,
Japan) (75, 76) (some of which are no longer available).
The combination of a jar-type apparatus with gas-generating
envelopes is widely used to culture anaerobic microorgan-
isms on petri dishes (77). In addition, the Anoxomat, an
automatic evacuation-replacement system that provides
anaerobic conditions in a jar, was developed in 1989 (78).
This easy-to-operate apparatus enables the cultivation of
obligate anaerobic bacterial strains belonging to the genera
Bacteroides, Clostridium, Fusobacterium, Prevotella, Porphyro-
monas, and Peptostreptococcus (79, 80). The Anoxomat
Mark II is currently available from Mart Microbiology
(Drachten, the Netherlands).
Bicarbonate–Carbon Dioxide Buffering System
Bicarbonate–carbon dioxide is the most popular buffering sys-
tem for anaerobic culture media. In this system, carbonic acid
(H2CO3) is formed by the reaction of carbon dioxide (CO2)
with water (H2O), then produced carbonic acid is immedi-
ately dissociated into proton (H+) and bicarbonate (HCO3−):
CO2 þ H2 O ! H2 CO3 ! HCO3  þ Hþ
 2.1.2. Anaerobic Cultivation ▪ 2.1.2-3

(a) (f)

(b) (g)

(c) (h)

(i)
(d)

(e) (j)

FIGURE 1 (a) Gas cylinders for replacing head space (e.g., N2, N2/CO2, and H2/CO2). (b) Gassing manifold in safety cabinet. All operations
can be done aseptically. (c) Gassing manifold in draft chamber (Deoxygenized Gas Pressure Injector types IP-8, Sanshin Industrial, Yokohama,
Japan). (d) The medium is dispensed to serum vials and replaced the head space with anoxic gas. (e) Plugging the tube with butyl rubber stop-
pers. (f ), (g) sealing with crimp-closure aluminum seals. (h) Reducing agents (e.g., Na2S and L-cysteine hydrochloride) can be prepared (the
head space has to be replaced with N2) and autoclaved in advance. (i) Incubation in thermostatic chamber. ( j) Vinyl-type anaerobic chamber
(COY Laboratory Products, Grass Lake, MI, USA). doi:10.1128/9781555818821.ch2.1.2.f1
2.1.2-4 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION

Although chemical components other than sodium TABLE 1 Reducing agents used for anaerobic culturing
bicarbonate may affect the value of pH, the Henderson- 0

Hasselbalch’s formula can be used as a guideline for pH deter- Compound E0 (mV) Reference
mination of anaerobic media: Sodium ascorbate +58 (170, 171)
 
½HCO3   Sodium thioglycollate −140 (170, 171)
pH ¼ pKa þ log10  pCO2
kH Glutathione −230 (171)
where pKa is acid dissociation constant of carbonic acid in −340 (171)
water (assumed to be 6.4) (81); [HCO3−] is molar concentra- 2-Mercaptoethanol −260 (172)
tion of bicarbonate (mM); kH is Henry’s law constant of Sodium sulfide (S0/H2S) −243 (171)
carbon dioxide (assumed to be 0.034 mM/mmHg) (82); Sodium sulfide (S0/HS−) −250 (170)
pCO2 is partial pressure of carbon dioxide in the medium.
−270 (170, 173)
In 1969, Hungate (41) reported that 0.5% (∼60 mM) sodium
bicarbonate solution equilibrated with 100% carbon dioxide L-cysteine hydrochloride −325 (170)
atmosphere has a pH of about 6.7 consistent with the −340 (171)
Henderson-Hasselbalch’s formula: Dithiothreitol −330 (170, 172, 174)
 
60 Titanium(III) citrate −480 (104, 170)
pH ¼ 6:4 þ log10  100 ¼ 6:77
0:034 Sodium dithionite (SO2− 2−
3 /S2O4 ) −471 (171)
Another example is the medium developed by Widdel and −527 (171)
Pfennig (83) equilibrated with an atmosphere of N2/CO2 −574 (170)
(80:20, vol/vol), which is commonly used for the cultiva-
tion of anaerobic bacteria (84, 85) and for enrichment of
methanogenic communities (86, 87). In these cases, 0.25% reported that titanium(III) NTA, instead of titanium(III) cit-
(∼30 mM) sodium bicarbonate solution was added into the rate, was useful for the enrichment of acid-tolerant hydroge-
medium. The pH value is calculated to be 7.16 that is almost notrophic methanogens and the isolation of a methanogen,
consistent with the actual values we have measured in our lab- Methanoregula boonei strain 6A8, from the enrichment culture
oratory: (97). Furthermore, Carbonero et al. (108) found that tita-
  nium(III) NTA might have inhibitory effects on the growth
30 of the aceticlastic methanogen Methanosaeta spp. These find-
pH ¼ 6:4 þ log10  20 ¼ 7:16
0:034 ings strongly indicate the need for caution in selecting an
The other buffering systems such as bicarbonate-CO2 plus appropriate reducing agent for culturing fastidious, previously
HEPES (88, 89) and phosphate buffer plus MES, HEPES, or uncultured anaerobes.
TAPS (90) are also used for cultivation of anaerobic
microbes.
RECENT PROGRESS IN THE ISOLATION OF
Reducing Agents ANAEROBES: THE CHLOROFLEXI SUBPHYLUM
The removal of oxygen with a reducing agent is a major strat- I AS A CASE STUDY
egy for creating an anaerobic condition. In the early stages Great advances in the isolation of anaerobes are exemplified
of the development of the jar-based anaerobic culturing by the class Anaerolineae of the phylum Chloroflexi (formerly
technique, pyrogallol, phosphorus, and potassium hydrox- known as Chloroflexi subphylum I). Until 2003, no represen-
ide were used to create anoxic conditions (25, 27). Other tatives of subphylum I within the phylum Chloroflexi had been
reducing agents that are directly added to liquid and solid cultured, even though culture-independent molecular analy-
media, such as ascorbic acid (91, 92), L-cysteine hydrochlor- ses had indicated the ubiquitous nature of this subphylum.
ide (92, 93), dithiothreitol (92, 94), glutathione (92, 95), 2- Sekiguchi et al. (109) observed outgrowth of Chloroflexi sub-
mercaptoethanol (92, 96), 2-mercaptoethanesulfonic acid phylum I–related filamentous microorganisms on the surface
(coenzyme M) (92, 97), sodium dithionite (92, 98), sodium of sludge granules obtained from a thermophilic reactor for
thioglycollate (92, 99), sulfur compounds (e.g., sodium sul- the treatment of fried soybean curd–manufacturing waste-
fide [92, 100], sodium sulfite [101], sodium thiosulfate water. The tips of the projection-like parts on the surface of
[102], and amorphous ferrous sulfide [103]), titanium(III) cit- the granules were cut with a surgical knife under a binocular
rate (92, 104), and titanium(III) nitrilotriacetate (NTA) microscope and used as an inoculum for isolation with the
(105), have commonly been used since the early stages of Hungate technique. A filamentous bacterium, Anaerolinea
the development of anaerobic culturing techniques (Table thermophila strain UNI-1, was successfully isolated as the first
1). Because these reducing agents possess a variety of chemi- cultured representative of the Chloroflexi subphylum I (84),
cal and biological properties (e.g., the equilibrium redox and the novel class Anaerolineae was coined (85). Subse-
potential [E00 ], toxicity and nutritional properties), their quently, a number of newly isolated organisms within the
effects on the growth of anaerobic microorganisms have new genera Bellilinea (110), Leptolinea (85), Levilinea (85),
been investigated. Wachenheim and Hespell (106) reported Longilinea (110), Thermanaerothrix (111), Thermomarinilinea
that titanium(III) citrate might be inhibitory to ruminal bac- (112), and Ornatilinea (113), all of which are within the class
teria compared with the combination of L-cysteine hydro- Anaerolineae, were isolated in pure culture. These organisms
chloride and sodium sulfide. By contrast, Bräuer et al. (107) were characterized as anaerobic heterotrophic bacteria and
reported that a peat soil–inoculated enrichment culture sup- isolated from various environments, such as wastewater treat-
plemented with titanium(III) citrate yielded the highest rate ment processes, rice paddy soil, shallow submarine hydrother-
of methane production compared with other reducing agents, mal vents, and deep terrestrial hot aquifers (114), indicating
such as hydrogen sulfide (H2S) and cysteine. They also their ecological importance in anaerobic ecosystems.

NOVEL TECHNIQUES FOR CULTURING
ANAEROBIC MICROORGANISMS
Single-Cell Manipulation of Anaerobic Microbes
In 1904, Barber (115) developed a glass pipet-based single-
cell manipulation technique for the isolation of microorgan-
isms. This primitive single-cell manipulation technique was
used to cultivate several species of Clostridium (116). After
decades of use of this technique (117–119), Fröhlich and
König (120, 121) developed a single-cell isolation system
comprising a salinized pipette tip (diameter 1–10 µm), a
microscope, and a CCD camera. This system can be used
for single-cell isolation of anaerobes in an anaerobic glove
box (121). The authors successfully isolated an endosymbi-
otic methanogen and a Mycoplasma-related endosymbiont
from flagellates in termite gut (120). Although such pipette-
based single-cell manipulations are useful for the isolation
of microbes, highly skilled manual manipulation is required.
To simplify and automate single-cell isolation, flow cytometry
(FCM) technology has been implemented for single-cell
sorting. Hamilton-Brehm et al. (122) developed a high-
throughput cultivation method based on anaerobic FCM.
In this system, microbial cells are sorted as single cells into
a PRAS medium–containing 96-well plate through a N2
gas-pressurized sheath fluid and incubated in a jar-type
apparatus with a gas pack. This anaerobic FCM-based sin-
gle-cell sorting is applicable for the isolation of anaerobic
thermophilic bacteria (Caloramator- and Caldicellulosiruptor-
related strains) from hot spring water samples. To achieve
effective FCM sorting, Borner et al. (123) tested a single-cell
capsulation with alginate microbeads for Clostridium sulfatire-
ducens strain CCUG 50825 and microbes in a cellulolytic
compound-degrading community. Although micromanipu-
lator and single-cell sorting techniques are powerful tools
for the single-cell isolation of anaerobes, another simple and
efficient way to obtain a single microbial cell is the dilution-
based single-cell culturing method. First, total microbial
cell numbers of environmental samples are estimated by
microscope-based (or possibly FCM-based) cell counting.
Then, the samples are diluted and aliquoted into a tube (or
a well of a 96/384-well plate) such that a tube or well contains,
on average, <1 cell per tube or well on the basis of total cell
numbers (124). This approach has been successfully applied
for the isolation of anaerobes from rumens (125) and the
human oral cavity (126). Although a lengthy incubation
period (e.g., 1–18 months) is required, this method may be
effective for the cultivation of slow-growing microbes.
Six-Well Plating Method
Although the Hungate technique is well established for the
culture of anaerobes, improvements continue to be made.
Nakamura et al. (127) developed a simple technique for
the cultivation of fastidious anaerobic organisms using a six-
well plate and an anaerobic gas pack system. This technique
was based on previously reported anaerobic bag culture meth-
ods (128, 129) and facilitates handling of the culture. The
steps for this protocol are (a) prepare the PRAS medium
with gellan gum as a solidifying agent, with storage at 45–
55°C until use; (b) dispense the medium into the wells of six-
well plate on a clean bench; (c) inoculate the cell suspension
immediately into the plate using the dispenser, rotate the
plate flatly to mix, and close the lid with tape; (d) place the
plate and two gas packs into a nylon bag; (e) insert a needle
connected to a gas-exchange machine with a vacuum pump
and exchange the interior atmosphere with O2-free gas; (f )
2.1.2. Anaerobic Cultivation ▪ 2.1.2-5
zip the bag completely and place it in the incubator (Fig. 2).
This method is useful for the culturing of obligate anaer-
obes, including methanogens, syntrophic metabolizers, and
sulfate- or thiosulfate-reducers. A Peptococcaceae-associated
sulfate-reducing bacterial strain KNH (Fig. 3) (127) and a
thermophilic and hydrogenotrophic methanogen, Methano-
thermobacter tenebrarum strain RMAS (130), were successfully
isolated from a natural gas field using this technique.
Coculture Method
Under methanogenic conditions, the degradation of certain
types of substrates (syntrophic substrates, such as volatile fatty
acids, aromatic compounds, and primary alcohols) produced
by the fermentative degradation of high-molecular-weight
organic compounds (such as carbohydrates, proteins, and lip-
ids) is accomplished via a syntrophic association between
substrate-oxidizing syntrophic metabolizers (syntrophs) and
hydrogenotrophic methanogens (17). Because the oxidation
of these syntrophic substrates is thermodynamically unfavor-
able under methanogenic conditions, degradation proceeds
only if hydrogen concentrations are kept low by hydrogen-
consuming methanogens. Thus, hydrogen concentration is
a critical factor for the growth of hydrogenotrophic methano-
gens. Generally, a gas mixture of hydrogen and carbon
dioxide (H2:CO2, 80:20) is used for the cultivation of metha-
nogens by replacing the atmospheric gas of the culture bot-
tles/tubes with H2:CO2 (48). Although this standard
approach has been widely used for the isolation of methano-
genic archaea, such high concentrations of hydrogen may dif-
fer from those found in actual methanogenic environments.
Sakai et al. (131) developed a groundbreaking technique
for the isolation of fastidious methanogens associated with
the order Methanocellales (formerly known as Rice Cluster
I). These authors attempted to mimic the hydrogen concen-
tration in the habitat of a Rice Cluster I–related methanogen
by adding propionate as an indirect substrate and a
propionate-oxidizing H2-producing syntroph, Syntrophobacter
fumaroxidans strain MPOB, as the H2 provider in an enrich-
ment culture. Using the coculture method, they success-
fully enriched the Rice Cluster I–related methanogen with
an H2-producing partner and isolated it by consecutive serial
dilution for over a year using the deep-agar method with H2 as
the substrate (131). The obtained strain was proposed as the
first cultured representative of Rice Cluster I and named
Methanocella paludicola strain SANAE (132). Imachi et al.
(133) isolated a hydrogenotrophic methanogen, Methanoli-
nea tarda strain NOBI-1, from anaerobic digester sludge in a
similar manner. In addition to propionate-degrading cocul-
ture, Tian et al. (134) isolated a hydrogenotrophic methano-
gen, Methanoculleus hydrogenitrophicus strain HC, from a
butyrate-degrading coculture with Syntrophomonas erecta
subsp. sporosyntropha strain 5-3-Z. This coculture technique
was dependent on interspecies hydrogen/electron transfer
between syntroph and methanogen (135), suggesting that
such microbial interactions are an important factor for the
isolation of fastidious anaerobes.
Enrichment in Bioreactors
The bioreactor-based culturing strategy enables the establish-
ment of stable culture conditions with continuous feed-
ing of substrates and removal of by-products while closely
mimicking environmental conditions. The bioreactor-based
culturing approach has recently been employed to enrich
uncultured and fastidious anaerobes. For example, Kindaichi
et al. (136) reported that “Candidatus Scalindua”–related
anammox bacteria were highly enriched (85.5% ± 4.5% of
2.1.2-6 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
To the gas
injector
FIGURE 2 Schematic illustration of brief procedure of six-well plate method with aerobic inoculation. This figure is a reprint with slight
modifications from Nakamura et al. (127) with the permission of the Microbes and Environments. doi:10.1128/9781555818821.ch2.1.2.f2
total bacterial cells based on fluorescent in situ hybridization
analyses) from coastal sediment samples using an up-flow col-
umn reactor (136). Oshiki et al. (137) determined that anam-
mox bacteria associated with “Candidatus Scalindua sp.” and
“Candidatus Brocadia sinica” predominated (>90% of the
total bacterial cells based on fluorescent in situ hybridization
analyses) in membrane bioreactors with biomasses obtained
from activated sludge and marine sediment, respectively.
Imachi et al. (138) developed a down-flow hanging sponge
reactor for the cultivation of uncultivated anaerobes in sub-
seafloor sediment, including members of Chloroflexi and
methanogens. Etting et al. (139, 140) reported that denitrify-
ing methanotrophic bacteria were highly enriched from
Step 1 Prepare gellan gum medium
Step 2 Stored at 45 or 55°C until use
Step 3
Dispense the molten medium into wells of 6-well
plate (P) on a clean bench
Step 4
Inoculate culture dilutions into medium with a
dispenser, and rotate plate flatly to suspend
inoculated cultures homogeneously in medium
Step 5
Tape lid with the plate
Step 6
Put the plate and two catalyst sachets
(Cs) into nylon bag (Nb)
Step 7
Insert needle connected to Deoxidized Gas
Pressure Injector equipped with a degassing
pump and exchange interior atmosphere with
O2-free gas (N2/CO2 or H2/CO2)
Step 8
Zip the bag completely and transfer to incubator;
if filled with H2/CO2, secure top of the bag with a
clip (C) and a paper clip (Pc)
freshwater sediment samples using a bioreactor-based method
with a continuous supply of methane and nitrite. Subse-
quently, these authors identified a novel, nitrite-dependent
anaerobic methanotrophic system on the basis of physiologi-
cal tests and the complete genome sequence derived from the
enriched microbe, “Candidatus Methylomirabilis oxyfera”
(141). The bioreactor-based strategy has also been used to
enrich archaeal groups responsible for the anaerobic oxida-
tion of methane. Dominant anaerobic oxidation of methane
populations were observed in a submerged-membrane bio-
reactor system (142, 143) and in high-pressurized bioreactors
mimicking the pressure in the sub-seafloor environment
(144, 145). These bioreactor-based enrichment procedures

FIGURE 3 (a) Colonies of strain KNH grown in the gellum
gum medium in a well of six-well plate for 3 weeks of incubation at
45°C. (b) Phase contrast microphotograph of strain KNH. Bar indi-
cates 10 µm. This figure is a reprint from the Nakamura et al. (127)
with the permission of Microbes and Environments.
doi:10.1128/9781555818821.ch2.1.2.f3
are an effective means of isolating pure cultures as well as ini-
tiating culture-independent analyses such as metagenomics
and single-cell genomics.
In situ/in vivo Cultivation
Precisely replicating environmental conditions for anaerobic
cultivation is difficult, even when the bioreactor-based tech-
nique is used. A possible solution is in situ/in vivo cultivation,
which may provide almost identical environmental condi-
tions. Takai et al. (146) developed an “in situ colonization sys-
tem” with a self-recording temperature probe (STR-ISCS) for
the isolation of uncultured members of the class Epsironpro-
teobacteria from deep-sea hydrothermal vents. Briefly, the
ISCS comprises a stainless steel pipe filled with sterilized syn-
thetic pumice as the “carrier” for microbial growth. The
STR-ISCS is inserted into or placed on a vent orifice to cap-
ture and enrich the resident microbes in the environment.
Thermophilic hydrogen-oxidizing bacteria Hydrogenimonas
thermophila (147) and Lebetimonas acidiphila (148) of the
2.1.2. Anaerobic Cultivation ▪ 2.1.2-7
Epsironproteobacteria have been successfully isolated in this
way. More recently, Nunoura et al. (112) reported that
Anaerolineae-type filamentous microorganisms grew in an
ISCS placed on a hydrothermal vent (52°C) in the Taketomi
submarine hot spring shallow hydrothermal field located in a
coral lagoon in the southern part of the Yaeyama Archipelago
in Japan. After 5 days of in situ cultivation, a filamentous
white microbial mat was observed in the ISCS (Fig. 4). The
pumiceous materials were recovered from the STR-ISCS
and inoculated into a dilution series of enrichment cultures
at 55°C. Eventually, the authors successfully isolated a fila-
mentous bacterial strain SW7 and proposed the new genus
Thermomarinilinea with type species T. lacunofontalis. The in
situ enrichment approach has also been employed in subsur-
face environments (149). Sizova et al. (126) reported the
mini-trap method for the in vivo cultivation of both aerobic
and anaerobic microbes from the human oral cavity. The
mini-trap is comprised of three surgical steel metal plates
with a total of 72 holes (diameter 400 µm). The holes of
the center plate was filled with the appropriate medium, sol-
idified with 1% molten agar, and covered with polycarbonate
membranes with a 1.0 µm pore size. The center plate was cov-
ered by upper and bottom side plates, which were tightened
with 2 mm screws. After 48 h of in vivo cultivation, the
authors obtained phylogenetically novel bacteria from the
mini-trap system. These in situ/in vivo cultivation techniques
may provide an ideal method for the proliferation of resident
microorganisms from these ecosystems.
Other Endeavors: Mimicked Medium, Alternative
Gelling Agent, and Antibiotics
Kenters et al. (125) developed a synthetic medium containing
concentrations of ions (i.e., Ca2+, Mg2+, K+, Na+, PO3− 4 , and
NH4+) and total S and CO2 mimicking those in ruminal fluid
and isolated significant numbers of phylogenetically novel
anaerobic bacteria from the ruminal environment. However,
solidifying agents for anaerobic culturing techniques have
not been well investigated. In 1983, Shungu et al. (150)
reported the usability of gellan gum as an alternative solidifying
agent to traditional agar to culture some anaerobic bacteria.
Subsequently, various types of anaerobes, including methano-
gens (151, 152), sulfur- or sulfate-reducing bacteria (153–156),
(a) (b)
FIGURE 4 (a) Appearance of in situ colonization system (ISCS)
after 5 days of incubation in a hydrothermal vent emission. Inner
diameter and inner length of the ISCS were 45 and 250 mm, respec-
tively. Filamentous white microbial mat formations were observed.
(b) Pumiceous stuffing settled in the ISCS after the 5 days of incuba-
tions. Outer diameter and length of the white rings were 6 and 9 mm,
respectively. This figure is a reprint from the Nunoura et al. (112)
with the permission of Microbes and Environments.
doi:10.1128/9781555818821.ch2.1.2.f4
2.1.2-8 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
iron-reducing bacteria (130, 157, 158), and fermentative bac-
teria (159) have been isolated with gellan gum-solidified
medium. Imachi et al. (160) employed antibiotics (vancomy-
cin, penicillin G, and kanamycin) to eliminate undesired
bacterial growth and facilitate the growth of methanogenic
archaea, resulting in the successful isolation of an ethanol-
utilizing methanogen, Methanofollis ethanolicus strain HASU,
from a lotus field. Such individual approaches may be impor-
tant for the cultivation of anaerobes.
CONCLUSIONS
A wide variety of techniques for anaerobic cultivation have
been developed and successfully applied for the isolation of
anaerobes. Pure isolates of anaerobes may provide direct infor-
mation about physiological and biochemical traits in anaero-
bic ecosystems. However, the isolation and characterization
of as-yet-uncultured but functionally important microorgan-
isms are being replaced by omics-driven approaches without
cultivation to at least partially reveal their physiology and
functions (161–163). In response to this trend, a number of
microbiologists continue to promote the importance of the
cultivation and isolation of microorganisms (164–167).
Indeed, study proposals on novel species, genus, and even
higher taxonomic ranks have increased substantially, and in
recent decades, approximately 6,000 prokaryotic species
(including anaerobes and aerobes) have been isolated and
characterized (168). However, most of the described organ-
isms are readily cultivable ones, while as-yet-uncultured
microbes remain uncultivable. To fill the gap between culti-
vation techniques and state-of-the-art technologies that cir-
cumvent isolation, the development of approaches for the
manual cultivation of as-yet-uncultured organisms remains
one of the most exciting challenges in microbiology. Norbert
Pfennig, a prominent anaerobic microbiologist, stated in
his retrospective that “access to the unexplored realm was
gained only after certain microbes appeared and grew in our
lab and so became our guidestars to unusual companions”
(169). Anaerobic ecosystems are filled with millions of
unseen spheres. The challenge to find a stellar anaerobe still
continues.
The figures used in this chapter are kindly provided by Kohei Nakamura
(Figs. 2 and 3, Gifu University, Gifu, Japan) and Dr. Takuro Nunoura
(Fig. 4, Japan Agency for Marine-Earth Science & Technology, Yoko-
suka, Japan) via the permission of the Microbes and Environments.
This study was partly supported by JSPS KAKENHI Grant Number
26630250.
REFERENCES
1. Kleinsteuber S, Schleinitz KM, Vogt C. 2012. Key players
and team play: anaerobic microbial communities in
hydrocarbon-contaminated aquifers. Appl Microbiol Bio-
technol 94:851–873.
2. Ohkuma M. 2008. Symbioses of flagellates and pro-
karyotes in the gut of lower termites. Trends Microbiol 16:
345–352.
3. Sommer F, Backhed F. 2013. The gut microbiota—masters
of host development and physiology. Nat Rev Microbiol 11:
227–238.
4. Ishii S, Ikeda S, Minamisawa K, Senoo K. 2011. Nitrogen
cycling in rice paddy environments: past achievements and
future challenges. Microbes Environ 26:282–292.
5. Flint HJ. 1997. The rumen microbial ecosystem—some
recent developments. Trends Microbiol 5:483–488.
6. Gray ND, Sherry A, Hubert C, Dolfing J, Head IM.
2010. Methanogenic degradation of petroleum hydrocar-
bons in subsurface environments remediation, heavy oil
formation, and energy recovery. Adv Appl Microbiol 72:
137–161.
7. Schrenk MO, Huber JA, Edwards KJ. 2010. Microbial
provinces in the subseafloor. Ann Rev Mar Sci 2:279– 304.
8. Laanbroek HJ. 2010. Methane emission from natural wet-
lands: interplay between emergent macrophytes and soil
microbial processes. A mini-review. Ann Bot London 105:
141–153.
9. Almeida JR, Favaro LC, Quirino BF. 2012. Biodiesel bio-
refinery: opportunities and challenges for microbial pro-
duction of fuels and chemicals from glycerol waste.
Biotechnol Biofuels 5:48.
10. Bokkenheuser V. 1993. The friendly anaerobes. Clin
Infect Dis 16(Suppl 4):S427–S434.
11. Narihiro T, Sekiguchi Y. 2007. Microbial communities in
anaerobic digestion processes for waste and wastewater
treatment: a microbiological update. Curr Opin Biotechnol
18:273–278.
12. Jones DT, Woods DR. 1986. Acetone-butanol fermenta-
tion revisited. Microbiol Rev 50:484–524.
13. Abdel-Rahman MA, Tashiro Y, Sonomoto K. 2013.
Recent advances in lactic acid production by microbial fer-
mentation processes. Biotechnol Adv 31:877–902.
14. van Niftrik L, Jetten MS. 2012. Anaerobic ammonium-
oxidizing bacteria: unique microorganisms with excep-
tional properties. Microbiol Mol Biol Rev 76:585–596.
15. Pester M, Knorr KH, Friedrich MW, Wagner M, Loy A.
2012. Sulfate-reducing microorganisms in wetlands—
fameless actors in carbon cycling and climate change. Front
Microbiol 3:72.
16. Weber KA, Achenbach LA, Coates JD. 2006. Microor-
ganisms pumping iron: anaerobic microbial iron oxidation
and reduction. Nat Rev Microbiol 4:752–764.
17. Schink B, Stams AJM. 2013. Syntrophism among prokar-
yotes, p. 471–493. In Rosenberg E, DeLong E, Lory S,
Stackebrandt E, Thompson F (eds), The prokaryotes, 4th
ed, Springer, Berlin.
18. Liu YC, Whitman WB. 2008. Metabolic, phylogenetic,
and ecological diversity of the methanogenic archaea.
Ann N Y Acad Sci 1125:171–189.
19. Miyamoto K, Li J, McClane BA. 2012. Enterotoxigenic
Clostridium perfringens: detection and identification.
Microbes Environ 27:343–349.
20. Murphy EC, Frick IM. 2013. Gram-positive anaerobic
cocci—commensals and opportunistic pathogens. FEMS
Microbiol Rev 37:520–553.
21. Brook I. 2010. The role of anaerobic bacteria in bactere-
mia. Anaerobe 16:183–189.
22. Wexler HM. 2007. Bacteroides: the good, the bad, and the
nitty-gritty. Clin Microbiol Rev 20:593–621.
23. Hatheway CL. 1990. Toxigenic clostridia. Clin Microbiol
Rev 3:66–98.
24. Pasteur L. 1861. Animalcules infusoires vivant sans gaz
oxygène libre et déterminant des fermentations. Comp
Rend Acad Sci 52:344–347.
25. Hunziker OF. 1902. A Review of the Existing Methods for
Cultivating Anaerobic Bacteria. Cornell University Library,
Ithaca, NY.
26. Sonnenwirth AC. 1972. Evolution of anaerobic methodol-
ogy. Am J Clin Nutr 25:1295–1298.
27. Hall IC. 1929. A review of the development and appli-
cation of physical and chemical principles in the cultiva-
tion of obligately anaerobic bacteria. J Bacteriol 17:
255–301.
28. Ehrenberg H. 1927. Simple method of producing anaero-
bic conditions. J Bacteriol 14:391–393.

29. Noguchi H. 1911. Method for the pure cultivation of
pathogenic Treponema pallidum (Spirochaeta Pallida). J Exp
Med 14:99–108.
30. Bushnell LD. 1922. A method for the cultivation of anae-
robes. J Bacteriol 7:277–281.
31. Ayers SH. 1910. A new form of anaerobic jar. Am J Public
Hygiene 20:844–846.
32. Herrington BL, Knaysi G. 1930. A simple method of
growing anaerobes in petri dish cultures. J Bacteriol 19:
101–103.
33. Morse S, Kopeloff N. 1922. A simple method for anaero-
bic cultivation in petri dishes. Am J Public Health 12:
119–121.
34. Esmarch E. 1886. Ueber eine Modification des Koch’schen
Plattenverfahrens zur Isolirung und zum quantitativen
Nachweis von Mikroorganismen. Zeits Hyg 1:293–301.
35. Thompson R. 1934. A mechanical device which prepares
and inoculates rolled tubes. J Bacteriol 28:279–287.
36. Wilson GS. 1922. The proportion of viable bacteria in
young cultures with especial reference to the technique
employed in counting. J Bacteriol 7:405–446.
37. Gee H. 1932. Mechanical spinner for Esmarch cultures.
J Bacteriol 24:35–41.
38. Moir GM, Thomson GI. 1950. Device for preparing roll-
tube cultures. J Dairy Res 17:192–193.
39. Hirsch A. 1948. An apparatus made from Meccano parts
for the preparation of roll-tube cultures. J Gen Microbiol
2:123–125.
40. Hungate RE. 1950. The anaerobic mesophilic cellulolytic
bacteria. Bacteriol Rev 14:1–49.
41. Hungate RE. 1969. A roll tube method for culti-
vation of strict anaerobes, p. 117–132. In Norris JR, Rib-
bons DW (eds), Methods in Microbiology. Academic Press,
New York.
42. Moore WEC. 1966. Techniques for routine culture of fas-
tidious anaerobes. Int J Syst Bacteriol 16:173–190.
43. Hungate RE, Smith W, Clarke RT. 1966. Suitability of
butyl rubber stoppers for closing anaerobic roll culture
tubes. J Bacteriol 91:908–909.
44. Miller TL, Wolin MJ. 1974. A serum bottle modification
of the Hungate technique for cultivating obligate anae-
robes. Appl Microbiol 27:985–987.
45. Attebery HR, Finegold SM. 1969. Combined screw-cap
and rubber-stopper closure for Hungate tubes ( pre-reduced
anaerobically sterilized roll tubes and liquid media). Appl
Microbiol 18:558–561.
46. Macy JM, Snellen JE, Hungate RE. 1972. Use of
syringe methods for anaerobiosis. Am J Clin Nutr 25:
1318–1323.
47. Balch WE, Wolfe RS. 1976. New approach to the cultiva-
tion of methanogenic bacteria: 2-mercaptoethanesulfonic
acid (HS-CoM)-dependent growth of Methanobacterium
ruminantium in a pressureized atmosphere. Appl Environ
Microbiol 32:781–791.
48. Wolfe RS. 2011. Techniques for cultivating methanogens.
Meth Enzymol 494:1–22.
49. Balch WE, Fox GE, Magrum LJ, Woese CR, Wolfe RS.
1979. Methanogens—re-evaluation of a unique biological
group. Microbiol Rev 43:260–296.
50. Boone DR, Bryant MP. 1980. Propionate-degrading bac-
terium, Syntrophobacter wolinii sp. nov. gen. nov., from
methanogenic ecosystems. Appl Environ Microbiol 40:
626–632.
51. Beaty PS, Mcinerney MJ. 1987. Growth of Syntrophomo-
nas wolfei in pure culture on crotonate. Arch Microbiol
147:389–393.
52. Bryant MP. 1972. Commentary on the Hungate technique
for culture of anaerobic bacteria. Am J Clin Nutr 25:
1324–1328.
2.1.2. Anaerobic Cultivation ▪ 2.1.2-9
53. Joblin KN. 1981. Isolation, enumeration, and mainte-
nance of rumen anaerobic fungi in roll tubes. Appl Environ
Microbiol 42:1119–1122.
54. Wolfe RS, Metcalf WW. 2010. A vacuum-vortex techni-
que for preparation of anoxic solutions or liquid culture
media in small volumes for cultivating methanogens or
other strict anaerobes. Anaerobe 16:216–219.
55. Socransky S, Macdonald JB, Sawyer S. 1959. The cultiva-
tion of Treponema microdentium as surface colonies. Arch
Oral Biol 1:171–172.
56. Keefer CE, Buck TC. 1962. An oxygen free gas chamber
for isolating fastidious anaerobic bacteria. J Water Pollut
Con F 34:75–79.
57. Rosebury T, Reynolds JB. 1964. Continuous anaerobiosis
for cultivation of spirochetes. Proc Soc Exp Biol Med 117:
813–815.
58. Drasar BS. 1967. Cultivation of anaerobic intestinal bac-
teria. J Pathol Bacteriol 94:417–427.
59. Aranki A, Syed SA, Kenney EB, Freter R. 1969. Isolation
of anaerobic bacteria from human gingiva and mouse
cecum by means of a simplified glove box procedure. Appl
Microbiol 17:568–576.
60. Speers AM, Cologgi DL, Reguera G. 2009. Anaerobic
cell culture. Curr Protoc Microbiol Appendix 4:Appendix
4F.
61. Balish E, Brown JF, Wilkins TD. 1977. Transparent plas-
tic incubator for the anaerobic glove box. Appl Environ
Microbiol 33:525–527.
62. Metcalf WW, Zhang JK, Wolfe RS. 1998. An anaerobic,
intrachamber incubator for growth of Methanosarcina spp.
on methanol-containing solid media. Appl Environ Micro-
biol 64:768–770.
63. Novy FG. 1893. Die Kultur Anaeröber Bakterien. Cen-
tralbl Bakt 14:581–600.
64. Smillie WG. 1917. New anaerobic methods. J Exp Med 26:
59–66.
65. Brewer JH, Heer AA, Mclaughlin CB. 1955. The use of
sodium borohydride for producing hydrogen in an anaerobe
jar. Appl Microbiol 3:136–136.
66. Brewer JH, Allgeier DL. 1965. Disposable hydrogen gen-
erator. Science 147:1033–1034.
67. Brewer JH, Allgeier DL. 1966. Safe self-contained carbon
dioxide-hydrogen anaerobic system. Appl Microbiol 14:
985–988.
68. Brewer JH, Allgeier DL, McLaughlin CB. 1966. Improved
anaerobic indicator. Appl Microbiol 14:135–136.
69. Ferguson IR, Phillips KD, Tearle PV. 1975. An evalua-
tion of the carbon dioxide component in the GasPak anae-
robic system. J Appl Bacteriol 39:167–173.
70. Collee JG, Watt B, Fowler EB, Brown R. 1972. An eval-
uation of the Gaspak system in the culture of anaerobic bac-
teria. J Appl Bacteriol 35:71–82.
71. Seip WF, Evans GL. 1980. Atmospheric analysis and
redox potentials of culture media in the Gaspak System.
J Clin Microbiol 11:226–233.
72. Ferguson IR, Phillips KD, Willis AT. 1976. Evaluation of
“Gaskit” disposable hydrogen and carbon-dioxide genera-
tor for culture of anaerobic bacteria. J Appl Bacteriol 41:
433–437.
73. Miller PH, Wiggs LS, Miller JM. 1995. Evaluation of
Anaero Gen system for growth of anaerobic bacteria.
J Clin Microbiol 33:2388–2391.
74. Heizmann WR, Werner H. 1989. GasPak Plus versus
Anaerocult A—two carbon dioxide/hydrogen systems for
cultivation of anaerobes. Zentr Bakteriol Mikrobiol Hyg A
270:511–516.
75. Van Horn KG, Warren K, Baccaglini EJ. 1997. Evalua-
tion of the AnaeroPack system for growth of anaerobic bac-
teria. J Clin Microbiol 35:2170–2173.
2.1.2-10 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
76. Delaney ML, Onderdonk AB. 1997. Evaluation of the
AnaeroPack system for growth of clinically significant
anaerobes. J Clin Microbiol 35:558–562.
77. Collins C, Petts D. 2008. Culture methods for pathogenic
anaerobes: a brief history. Biomed. Scientist. (July) p.
599–601.
78. Brazier JS, Smith SA. 1989. Evaluation of the Anoxomat:
a new technique for anaerobic and microaerophilic clinical
bacteriology. J Clin Pathol 42:640–644.
79. Shahin M, Jamal W, Verghese T, Rotimi VO. 2003. Com-
parative evaluation of anoxomat and conventional anaero-
bic GasPak jar systems for the isolation of anaerobic
bacteria. Med Princ Pract 12:81–86.
80. Summanen PH, McTeague M, Vaisanen ML, Strong CA,
Finegold SM. 1999. Comparison of recovery of anaerobic
bacteria using the Anoxomat, anaerobic chamber, and
GasPak jar systems. Anaerobe 5:5–9.
81. Jencks WP, Regenstein J. 2010. Ionization constants of
acids and bases, p. 595–635. In Lundblad RL, Macdonald
F (eds), Handbook of Biochemistry and Molecular Biology.
4th ed. CRC Press, Boca Raton, FL.
82. Sander R. 2014. Compilation of Henry’s law constants,
version 3.99. Atmos Chem Phys Disc 14:29615–30521.
83. Widdel F, Pfennig N. 1981. Studies on dissimilatory
sulfate-reducing bacteria that decompose fatty-acids.1.
Isolation of new sulfate-reducing bacteria enriched with
acetate from saline environments - description of Desulfo-
bacter postgatei gen. nov., sp. nov. Arch Microbiol 129:
395–400.
84. Sekiguchi Y, Yamada T, Hanada S, Ohashi A, Harada H,
Kamagata Y. 2003. Anaerolinea thermophila gen. nov., sp
nov and Caldilinea aerophila gen. nov., sp nov., novel fila-
mentous thermophiles that represent a previously uncul-
tured lineage of the domain Bacteria at the subphylum
level. Int J Syst Evol Microbiol 53:1843–1851.
85. Yamada T, Sekiguchi Y, Hanada S, Imachi H, Ohashi A,
Harada H, Kamagata Y. 2006. Anaerolinea thermolimosa
sp. nov., Levilinea saccharolytica gen. nov., sp nov and
Leptolinea tardivitalis gen. nov., sp. nov., novel filamen-
tous anaerobes, and description of the new classes Anae-
rolineae classis nov. and Caldilineae classis nov. in the
bacterial phylum Chloroflexi. Int J Syst Evol Microbiol
56:1331–1340.
86. Narihiro T, Nobu MK, Kim NK, Kamagata Y, Liu WT.
2015. The nexus of syntrophy-associated microbiota in
anaerobic digestion revealed by long-term enrichment and
community survey. Environ Microbiol 17:1707–1720.
87. Iino T, Tamaki H, Tamazawa S, Ueno Y, Ohkuma M,
Suzuki K, Igarashi Y, Haruta S. 2013. Candidatus
Methanogranum caenicola: a novel methanogen from the
anaerobic digested sludge, and proposal of Methanomassilii-
coccaceae fam. nov. and Methanomassiliicoccales ord. nov.,
for a methanogenic lineage of the class Thermoplasmata.
Microbes Environ 28:244–250.
88. Delgado AG, Parameswaran P, Fajardo-Williams D, Hal-
den RU, Krajmalnik-Brown R. 2012. Role of bicarbonate
as a pH buffer and electron sink in microbial dechlorination
of chloroethenes. Microb Cell Fact 11:128.
89. Cervantes FJ, Dijksma W, Duong-Dac T, Ivanova A,
Lettinga G, Field JA. 2001. Anaerobic mineralization of
toluene by enriched sediments with quinones and humus
as terminal electron acceptors. Appl Environ Microbiol 67:
4471–4478.
90. Lee YJ, Romanek CS, Wiegel J. 2007. Clostridium acidito-
lerans sp. nov., an acid-tolerant spore-forming anaerobic
bacterium from constructed wetland sediment. Int J Syst
Evol Microbiol 57:311–315.
91. Gill JW, King KW. 1958. Nutritional characteristics of a
Butyrivibrio. J Bacteriol 75:666–673.
92. Atlas RM. 2010. Handbook of Microbiological Media, 4th ed.
CRC Press, Boca Raton, FL.
93. Hosoya S. 1925. A new method for the cultivation of anae-
robic bacilli. Tokyo Imp Univ, Sc Reps Gov Inst Infect Dise 4:
103.
94. Hazlewood GP, Nugent JHA. 1978. Leaf fraction 1 pro-
tein as a nitrogen-source for growth of a proteolytic rumen
bacterium. J Gen Microbiol 106:369–371.
95. Bryant MP. 1959. Bacterial species of the rumen. Bacteriol
Rev 23:125–153.
96. Bhatnagar L, Henriquet M, Zeikus JG, Aubert JP. 1984.
Utilization of mercapto-2-ethanol as medium reductant for
determination of the metabolic response of methanogens
towards inorganic sulfur-compounds. FEMS Microbiol Lett
22:155–158.
97. Bräuer S, Cadillo-Quiroz H, Ward RJ, Yavitt J, Zinder S.
2011. Methanoregula boonei gen. nov., sp. nov., an acidi-
philic methanogen isolated from an acidic peat bog.
Int J Syst Evol Microbiol 61:45–52.
98. Smith PH, Hungate RE. 1958. Isolation and characteriza-
tion of Methanobacterium ruminantium n. sp. J Bacteriol 75:
713–718.
99. Sijpesteijn AK. 1951. On Ruminococcus flavefaciens, a
cellulose-decomposing bacterium from the rumen of sheep
and cattle. J Gen Microbiol 5:869–879.
100. Bryant MP, Robinson IM. 1961. An improved nonselec-
tive culture medium for ruminal bacteria and its use in
determining diurnal variation in numbers of bacteria in
the rumen. J Dairy Sci 44:1446–1456.
101. Rothe O, Thomm M. 2000. A simplified method for the cul-
tivation of extreme anaerobic Archaea based on the use of
sodium sulfite as reducing agent. Extremophiles 4:247–252.
102. de Poorter LM, Geerts WJ, Keltjens JT. 2005. Hydrogen
concentrations in methane-forming cells probed by the
ratios of reduced and oxidized coenzyme F420. Microbiology
151:1697–1705.
103. Brock TD, Od’ea K. 1977. Amorphous ferrous sulfide as a
reducing agent for culture of anaerobes. Appl Environ Micro-
biol 33:254–256.
104. Zehnder AJ, Wuhrmann K. 1976. Titanium (III) citrate as
a nontoxic oxidation-reduction buffering system for the
culture of obligate anaerobes. Science 194:1165–1166.
105. Moench TT, Zeikus JG. 1983. An improved preparation
method for a titanium (III) media reductant. J Microbiol
Meth 1:199–202.
106. Wachenheim DE, Hespell RB. 1984. Inhibitory effects of
titanium (III) citrate on enumeration of bacteria from
rumen contents. Appl Environ Microbiol 48:444–445.
107. Bräuer SL, Yashiro E, Ueno NG, Yavitt JB, Zinder SH.
2006. Characterization of acid-tolerant H-2/CO2-utilizing
methanogenic enrichment cultures from an acidic peat
bog in New York State. FEMS Microbiol Ecol 57:206–216.
108. Carbonero F, Oakley BB, Purdy KJ. 2010. Improv-
ing the isolation of anaerobes on solid media: the example
of the fastidious Methanosaeta. J Microbiol Meth 80:
203–205.
109. Sekiguchi Y, Takahashi H, Kamagata Y, Ohashi A, Har-
ada H. 2001. In situ detection, isolation, and physiological
properties of a thin filamentous microorganism abundant in
methanogenic granular sludges: a novel isolate affiliated
with a clone cluster, the green non-sulfur bacteria, subdivi-
sion I. Appl Environ Microbiol 67:5740–5749.
110. Yamada T, Imachi H, Ohashi A, Harada H, Hanada S,
Kamagata Y, Sekiguchi Y. 2007. Bellilinea caldifistulae
gen. nov., sp nov and Longilinea arvoryzae gen. nov.,
sp nov., strictly anaerobic, filamentous bacteria of the phy-
lum Chloroflexi isolated from methanogenic propionate-
degrading consortia. Int J Syst Evol Microbiol 57:
2299–2306.

111. Gregoire P, Fardeau ML, Joseph M, Guasco S, Hamaide
F, Biasutti S, Michotey V, Bonin P, Ollivier B. 2011. Iso-
lation and characterization of Thermanaerothrix daxensis
gen. nov., sp nov., a thermophilic anaerobic bacterium per-
taining to the phylum “Chloroflexi”, isolated from a deep
hot aquifer in the Aquitaine Basin. Syst Appl Microbiol
34:494–497.
112. Nunoura T, Hirai M, Miyazaki M, Kazama H, Makita H,
Hirayama H, Furushima Y, Yamamoto H, Imachi H, Takai
K. 2013. Isolation and characterization of a thermophilic,
obligately anaerobic and heterotrophic marine Chloroflexi
bacterium from a Chloroflexi-dominated microbial com-
munity associated with a Japanese shallow hydrothermal sys-
tem, and proposal for Thermomarinilinea lacunofontalis gen.
nov., sp. nov. Microbes Environ 28:223–235.
113. Podosokorskaya OA, Bonch-Osmolovskaya EA,
Novikov AA, Kolganova TV, Kublanov IV. 2013. Orna-
tilinea apprima gen. nov., sp nov., a cellulolytic representa-
tive of the class Anaerolineae. Int J Syst Evol Microbiol 63:
86–92.
114. Yamada T, Sekiguchi Y. 2009. Cultivation of uncultured
Chloroflexi subphyla: significance and ecophysiology of
formary uncultured Chloroflexi “subphylum I” with natural
and biotechnological relevance. Microbes Environ 24:
205–216.
115. Barber MA. 1904. A new method of isolating microorgan-
isms. J Kansas Med Sot 4:489–494.
116. Barber MA. 1920. Use of the single cell method in obtain-
ing pure cultures of anaerobes. J Exp Med 32:295–311.
117. Chambers R. 1922. New micromanipulator and methods
for the isolation of a single bacterium and the manipulation
of living cells. J Infect Dis 31:334–343.
118. Reyniers JA. 1932. A new and simplified micrurgical appa-
ratus especially adapted to single cell isolation. J Bacteriol
23:183–192.
119. Skerman VB. 1968. A new type of micromanipulator and
microforge. J Gen Microbiol 54:287–297.
120. Fröhlich J, König H. 1999. Rapid isolation of single micro-
bial cells from mixed natural and laboratory populations
with the aid of a micromanipulator. Syst Appl Microbiol
22:249–257.
121. Fröhlich J, König H. 2000. New techniques for isolation
of single prokaryotic cells. FEMS Microbiol Rev 24:
567–572.
122. Hamilton-Brehm SD, Vishnivetskaya TA, Allman SL,
Mielenz JR, Elkins JG. 2012. Anaerobic high-throughput
cultivation method for isolation of thermophiles using
biomass-derived substrates. Meth Mol Biol 908:153–168.
123. Borner RA, Aliaga MTA, Mattiasson B. 2013. Microcul-
tivation of anaerobic bacteria single cells entrapped in algi-
nate microbeads. Biotechnol Lett 35:397–405.
124. Buerger S, Spoering A, Gavrish E, Leslin C, Ling L,
Epstein SS. 2012. Microbial scout hypothesis and micro-
bial discovery. Appl Environ Microbiol 78:3229–3233.
125. Kenters N, Henderson G, Jeyanathan J, Kittelmann S,
Janssen PH. 2011. Isolation of previously uncultured
rumen bacteria by dilution to extinction using a new liquid
culture medium. J Microbiol Meth 84:52–60.
126. Sizova MV, Hohmann T, Hazen A, Paster BJ, Halem SR,
Murphy CM, Panikov NS, Epstein SS. 2012. New
approaches for isolation of previously uncultivated oral bac-
teria. Appl Environ Microbiol 78:194–203.
127. Nakamura K, Tamaki H, Kang MS, Mochimaru H, Lee
ST, Kamagata Y. 2011. A six-well plate method: less labo-
rious and effective method for cultivation of obligate anae-
robic microorganisms. Microbes Environ 26:301–306.
128. Matthews AD, Karnauchow PN. 1961. A simple techni-
que for the cultivation of anaerobes. Can Med Assoc J 84:
793–794.
2.1.2. Anaerobic Cultivation ▪ 2.1.2-11
129. Rosenblatt JE, Stewart PR. 1975. Anaerobic bag culture
method. J Clin Microbiol 1:527–530.
130. Nakamura K, Takahashi A, Mori C, Tamaki H,
Mochimaru H, Takamizawa K, Kamagata Y. 2013. Meth-
anothermobacter tenebrarum sp. nov., a hydrogenotrophic,
thermophilic methanogen isolated from gas-associated for-
mation water of a natural gas field. Int J Syst Evol Microbiol
63:715–722.
131. Sakai S, Imachi H, Sekiguchi Y, Ohashi A, Harada H,
Kamagata Y. 2007. Isolation of key methanogens for global
methane emission from rice paddy fields: a novel isolate
affiliated with the clone cluster rice cluster I. Appl Environ
Microbiol 73:4326–4331.
132. Sakai S, Imachi H, Hanada S, Ohashi A, Harada H,
Kamagata Y. 2008. Methanocella paludicola gen. nov.,
sp nov., a methane-producing archaeon, the first isolate
of the lineage “Rice Cluster I”, and proposal of the new
archaeal order Methanocellales ord. nov. Int J Syst Evol
Microbiol 58:929–936.
133. Imachi H, Sakai S, Sekiguchi Y, Hanada S, Kamagata Y,
Ohashi A, Harada H. 2008. Methanolinea tarda gen. nov.,
sp nov., a methane-producing archaeon isolated from a
methanogenic digester sludge. Int J Syst Evol Microbiol 58:
294–301.
134. Tian JQ, Wang YF, Dong XZ. 2010. Methanoculleus
hydrogenitrophicus sp. nov., a methanogenic archaeon iso-
lated from wetland soil. Int J Syst Evol Microbiol 60:
2165–2169.
135. Kato S, Watanabe K. 2010. Ecological and evolution-
ary interactions in syntrophic methanogenic consortia.
Microbes Environ 25:145–151.
136. Kindaichi T, Awata T, Suzuki Y, Tanabe K, Hatamoto M,
Ozaki N, Ohashi A. 2011. Enrichment using an up-flow
column reactor and community structure of marine anam-
mox bacteria from coastal sediment. Microbes Environ 26:
67–73.
137. Oshiki M, Awata T, Kindaichi T, Satoh H, Okabe S.
2013. Cultivation of planktonic anaerobic ammonium oxi-
dation (anammox) bacteria using membrane bioreactor.
Microbes Environ 28:436–443.
138. Imachi H, Aoi K, Tasumi E, Saito Y, Yamanaka Y, Saito
Y, Yamaguchi T, Tomaru H, Takeuchi R, Morono Y, Ina-
gaki F, Takai K. 2011. Cultivation of methanogenic com-
munity from subseafloor sediments using a continuous-flow
bioreactor. ISME J 5:1913–1925.
139. Ettwig KF, van Alen T, van de Pas-Schoonen KT, Jetten
MS, Strous M. 2009. Enrichment and molecular detection
of denitrifying methanotrophic bacteria of the NC10 phy-
lum. Appl Environ Microbiol 75:3656–3662.
140. Ettwig KF, Shima S, van de Pas-Schoonen KT, Kahnt J,
Medema MH, Op den Camp HJ, Jetten MS, Strous M.
2008. Denitrifying bacteria anaerobically oxidize methane
in the absence of Archaea. Environ Microbiol 10:
3164–3173.
141. Ettwig KF, Butler MK, Le Paslier D, Pelletier E, Mange-
not S, Kuypers MM, Schreiber F, Dutilh BE, Zedelius J,
de Beer D, Gloerich J, Wessels HJ, van Alen T, Luesken
F, Wu ML, van de Pas-Schoonen KT, Op den Camp HJ,
Janssen-Megens EM, Francoijs KJ, Stunnenberg H,
Weissenbach J, Jetten MS, Strous M. 2010. Nitrite-driven
anaerobic methane oxidation by oxygenic bacteria. Nature
464:543–548.
142. Meulepas RJ, Jagersma CG, Gieteling J, Buisman CJ,
Stams AJ, Lens PN. 2009. Enrichment of anaerobic meth-
anotrophs in sulfate-reducing membrane bioreactors. Bio-
technol Bioeng 104:458–470.
143. Jagersma GC, Meulepas RJ, Heikamp-de Jong I, Gietel-
ing J, Klimiuk A, Schouten S, Damste JS, Lens PN,
Stams AJ. 2009. Microbial diversity and community
2.1.2-12 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
structure of a highly active anaerobic methane-oxidizing
sulfate-reducing enrichment. Environ Microbiol 11:
3223–3232.
144. Deusner C, Meyer V, Ferdelman TG. 2010. High-
pressure systems for gas-phase free continuous incubation
of enriched marine microbial communities performing
anaerobic oxidation of methane. Biotechnol Bioeng 105:
524–533.
145. Zhang Y, Maignien L, Zhao X, Wang F, Boon N. 2011.
Enrichment of a microbial community performing anaero-
bic oxidation of methane in a continuous high-pressure
bioreactor. BMC Microbiol 11:137.
146. Takai K, Inagaki F, Nakagawa S, Hirayama H, Nunoura
T, Sako Y, Nealson KH, Horikoshi K. 2003. Isolation and
phylogenetic diversity of members of previously unculti-
vated epsilon-Proteobacteria in deep-sea hydrothermal
fields. FEMS Microbiol Lett 218:167–174.
147. Takai K, Nealson KH, Horikoshi K. 2004. Hydrogenimo-
nas thermophila gen. nov., sp. nov., a novel thermophilic,
hydrogen-oxidizing chemolithoautotroph within the
epsilon-Proteobacteria, isolated from a black smoker in a
Central Indian Ridge hydrothermal field. Int J Syst Evol
Microbiol 54:25–32.
148. Takai K, Hirayama H, Nakagawa T, Suzuki Y, Nealson
KH, Horikoshi K. 2005. Lebetimonas acidiphila gen. nov.,
sp. nov., a novel thermophilic, acidophilic, hydrogen-
oxidizing chemolithoautotroph within the “Epsilonproteo-
bacteria”, isolated from a deep-sea hydrothermal fumarole
in the Mariana Arc. Int J Syst Evol Microbiol 55:183–189.
149. Silver BJ, Onstott TC, Rose G, Lin LH, Ralston C,
Sherwood-Lollar B, Pfiffner SM, Kieft TL, McCuddy
S. 2010. In situ cultivation of subsurface microorganisms
in a deep mafic sill: implications for SLiMEs. Geomicrobiol
J 27:329–348.
150. Shungu D, Valiant M, Tutlane V, Weinberg E, Weiss-
berger B, Koupal L, Gadebusch H, Stapley E. 1983. GEL-
RITE as an agar substitute in bacteriological media. Appl
Environ Microbiol 46:840–845.
151. Jeanthon C, L’Haridon S, Reysenbach AL, Corre E,
Vernet M, Messner P, Sleytr UB, Prieur D. 1999. Meth-
anococcus vulcanius sp. nov., a novel hyperthermophilic
methanogen isolated from East Pacific Rise, and identifica-
tion of Methanococcus sp. DSM 4213T as Methanococcus
fervens sp. nov. Int J Syst Bacteriol 49(2):583–589.
152. Jeanthon C, L’Haridon S, Reysenbach AL, Vernet M,
Messner P, Sleytr UB, Prieur D. 1998. Methanococcus
infernus sp. nov., a novel hyperthermophilic lithotrophic
methanogen isolated from a deep-sea hydrothermal vent.
Int J Syst Bacteriol 48(Pt 3):913–919.
153. L’Haridon S, Cilia V, Messner P, Raguenes G, Gamba-
corta A, Sleytr UB, Prieur D, Jeanthon C. 1998. Desulfur-
obacterium thermolithotrophum gen. nov., sp. nov., a novel
autotrophic, sulphur-reducing bacterium isolated from a
deep-sea hydrothermal vent. Int J Syst Bacteriol 48(3):
701–711.
154. Kaksonen AH, Spring S, Schumann P, Kroppenstedt
RM, Puhakka JA. 2007. Desulfovirgula thermocuniculi
gen. nov., sp. nov., a thermophilic sulfate-reducer isolated
from a geothermal underground mine in Japan. Int J Syst
Evol Microbiol 57:98–102.
155. L’Haridon S, Reysenbach AL, Tindall BJ, Schonheit P,
Banta A, Johnsen U, Schumann P, Gambacorta A, Stack-
ebrandt E, Jeanthon C. 2006. Desulfurobacterium atlanti-
cum sp. nov., Desulfurobacterium pacificum sp. nov. and
Thermovibrio guaymasensis sp. nov., three thermophilic
members of the Desulfurobacteriaceae fam. nov., a deep
branching lineage within the Bacteria. Int J Syst Evol Micro-
biol 56:2843–2852.
156. L’Haridon S, Miroshnichenko ML, Hippe H, Fardeau
ML, Bonch-Osmolovskaya EA, Stackebrandt E, Jean-
thon C. 2002. Petrotoga olearia sp. nov. and Petrotoga sibir-
ica sp. nov., two thermophilic bacteria isolated from a
continental petroleum reservoir in Western Siberia.
Int J Syst Evol Microbiol 52:1715–1722.
157. Prakash O, Gihring TM, Dalton DD, Chin KJ, Green SJ,
Akob DM, Wanger G, Kostka JE. 2010. Geobacter daltonii
sp. nov., an Fe(III)- and uranium(VI)-reducing bacterium
isolated from a shallow subsurface exposed to mixed heavy
metal and hydrocarbon contamination. Int J Syst Evol
Microbiol 60:546–553.
158. Viulu S, Nakamura K, Okada Y, Saitou S, Takamizawa
K. 2013. Geobacter luticola sp. nov., an Fe(III)-reducing
bacterium isolated from lotus field mud. Int J Syst Evol
Microbiol 63:442–448.
159. Pankratov TA, Kirsanova LA, Kaparullina EN, Kevbrin
VV, Dedysh SN. 2012. Telmatobacter bradus gen. nov., sp.
nov., a cellulolytic facultative anaerobe from subdivision 1
of the Acidobacteria, and emended description of Acidobac-
terium capsulatum Kishimoto et al. 1991. Int J Syst Evol
Microbiol 62:430–437.
160. Imachi H, Sakai S, Nagai H, Yamaguchi T, Takai K.
2009. Methanofollis ethanolicus sp nov., an ethanol-utilizing
methanogen isolated from a lotus field. Int J Syst Evol Micro-
biol 59:800–805.
161. Taupp M, Mewis K, Hallam SJ. 2011. The art and design
of functional metagenomic screens. Curr Opin Biotechnol
22:465–472.
162. Rinke C, Schwientek P, Sczyrba A, Ivanova NN, Ander-
son IJ, Cheng JF, Darling A, Malfatti S, Swan BK, Gies
EA, Dodsworth JA, Hedlund BP, Tsiamis G, Sievert
SM, Liu WT, Eisen JA, Hallam SJ, Kyrpides NC, Stepa-
nauskas R, Rubin EM, Hugenholtz P, Woyke T. 2013.
Insights into the phylogeny and coding potential of micro-
bial dark matter. Nature 499:431–437.
163. Nelson KE. 2013. Microbiomes. Microb. Ecol 65:916–919.
164. Keller M, Zengler K. 2004. Tapping into microbial diver-
sity. Nat Rev Microbiol 2:141–150.
165. Narihiro T, Kamagata Y. 2013. Cultivating yet-to-be cul-
tivated microbes: the challenge continues. Microbes Envi-
ron 28:163–165.
166. Vartoukian SR, Palmer RM, Wade WG. 2010. Strategies
for culture of “unculturable” bacteria. FEMS Microbiol Lett
309:1–7.
167. Puspita ID, Kamagata Y, Tanaka M, Asano K, Nakatsu
CH. 2012. Are uncultivated bacteria really uncultivable?
Microbes Environ 27:356–366.
168. Euzeby JP. 1997. List of bacterial names with standing
in nomenclature: a folder available on the Internet.
Int J Syst Bacteriol 47:590–592.
169. Pfennig N. 1993. Reflections of a microbiologist, or how to
learn from the microbes. Annu Rev Microbiol 47:1–29.
170. Plugge CM. 2005. Anoxic media design, preparation, and
considerations. Environ Microbiol 397:3–16.
171. Loach PA. 2010. Oxidation-reduction potentials, absorb-
ance bands and molar absorbance of compounds used in
biochemical studies, p. 557–563. In Lundblad RL, Mac-
donald F (eds), Handbook of Biochemistry and Molecular Biol-
ogy, 4th ed. CRC Press, Boca Raton, FL.
172. Mickey B, Howard J. 1995. Rigidity of microtu-
bules is increased by stabilizing agents. J Cell Biol 130:
909–917.
173. Thauer RK, Jungermann K, Decker K. 1977. Energy con-
servation in chemotrophic anaerobic bacteria. Bacteriol Rev
41:100–180.
174. Cleland WW. 1964. Dithiothreitol, a new protective
reagent for SH groups. Biochemistry 3:480–482.
 New Devices for Cultivation
YOSHITERU AOI AND SLAVA EPSTEIN
2.1.3
One of the most important observations in microbiology is
that the vast majority of microorganisms from most envi-
ronments on the planet do not grow on artificial media.
This is a significant impediment for both academic and
applied microbiology, necessitating innovations in cultiva-
tion technologies. Several recently advanced methodologies
offer a promise to close the gap between the high richness
of environmental species and low number of their cultiva-
ble representatives. This chapter describes the state of the
art in microbial cultivation methods, their principles, and
application. These methods are categorized into two types:
in situ cultivation whereby microbes are cultivated in situ,
and high-throughput cultivation, mostly in vitro. In the first
group, the following methods are described in detail, (a) dif-
fusion chamber, (b) i-chip, (c) microbial trap, and (d) hollow
fiber membrane chamber. In the second group, we focus on
gel micro-droplets (GMDs) based cultivation and micro-
fabrication-based technologies. The chapter also discusses
their relative merits and respective biases.
UNCULTURABLE MICROORGANISMS
The term “the great plate count anomaly” was coined by Sta-
ley and Konopka in 1985 (1) to describe a significant disparity
between the number of microbial cells in environmental sam-
ples and the number of colonies these cells form on artificial
media. The same gap is observed between the high microbial
richness in nature and relatively small number of species in
the existing culture collections, as evidenced by the rRNA
surveys of environmental habitats (2–9). This is arguably
the oldest unresolved microbiological phenomenon which
was first observed more than a century ago (9, 44). Its resolu-
tion remains critically important for the progress in microbi-
ology, biotechnology, pharmaceutical discovery, and human
health studies.
CONCEPT AND METHODS FOR ISOLATING
MICROORGANISMS
Cultivation of environmental microorganisms in pure culture
usually consists of the following steps: (a) isolation: separa-
tion of single cell from the other microbial cells coexisting
in the sample; (b) inoculation: inoculating single cell into a
nutrient medium; (c) incubation: allowing the cell to propa-
gate; (d) detection: visually recognizing growth; and (e)
doi:10.1128/9781555818821.ch2.1.3
2.1.3-1
recovery and archiving: replicating growth and storing thus
obtained biomass.
Conventional Methods
One of the simplest methodologies that incorporate the fore-
going steps is “limiting dilution” in liquid media. Diluting of
microbial cells such that specific aliquots of the cell suspen-
sion would contain on average less than one cell, and using
such aliquots as inocula brought a substantial number of spe-
cies into culture and remains an important tool today.
An alternative is cultivation on solid support in petri dishes,
which was invented almost 150 years ago and has long been
regarded as a principal cultivation technique. They remain
the primary choice of microbiologists and are widely used
for cultivation of environmental microorganisms, mainly
because these techniques are cost- and time-efficient, with
all the steps simplified for use in any laboratory setting. How-
ever, these methods share a critical limitation: they only pro-
vide access to less than 1% of species in the biosphere.
Recent Improvement in Methodologies
Over the past decade, there has been significant effort to over-
come this limitation. Some of the innovations are based on
modification of conventional methods, such as manipulat-
ing the nutrient composition and concentration, the use of
alternative gelling agents, experimenting with trace ele-
ments, and changing the length of incubation (10–20).
Other innovations include high-throughput variants of con-
ventional approaches, such as miniaturizing the size of the
growth compartment to a few microns, leading to a “million
well . . .” (whatever it is called). These modifications are out-
side of the scope of this chapter. Instead, it focuses on advan-
ces that have changed the very principle of cultivation by
emphasizing the importance of mimicking the natural milieu
of the species targeted for cultivation.
IN SITU CULTIVATION
It is indisputable that microbial species that do not grow in
the lab can and do grow in their natural environment. The
simplest explanation is that nature provides microbes with
growth components necessary and sufficient for their growth,
whereas artificial media do not. Therefore, replicating the
natural conditions as closely as possible may be of fundamen-
tal importance for cultivation of presently “uncultivable”
2.1.3-2 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
environmental microorganisms. One of the simplest ways to
achieve such replication is to cultivate microbes in a
membrane-bound device incubated in nature. The expecta-
tion is that diffusion will automatically provide the cells
with the necessary growth factors, allowing for their cultiva-
tion with no prior knowledge of their growth requirements.
Several recently introduced methods are based on this idea
(21–29).
Method 1: Diffusion Chamber
The originally proposed diffusion chamber for microbial cul-
tivation in situ consists of a metal washer or plastic ring and
two polycarbonate membranes with small pores (<0.1 μm)
(27). Polycarbonate membrane, such as, for example, 0.03
µm pore-size membrane (GE Osmonics Labstore, Minne-
tonka, MN), is glued to a stainless steel or plastic O-ring
with silicone glue (Fig. 1). The environmental microorgan-
isms are serially diluted, mixed with warm agar (0.7–1.5%,
at 40–45°C), and 3 ml of the suspension is poured on the
membrane. After the agar solidifies, the second membrane
is glued to the other side of the O-ring, sealing the agar
with embedded bacteria inside, thus forming a diffusion
chamber. The chamber is placed in the natural habitat of
the cells, such as the ocean floor, or its simulated version,
such as an aquarium (Fig. 1b). After a period of incubation,
FIGURE 1 Photographic (a) and schematic (b) image of the diffusion chamber. A polycarbonate membrane is attached to the bottom of the
metal disc, the other membrane is glued to the upper surface of disc. The inner space is filled with microbial cells mixed with agar. After the
inoculation and assembly the diffusion chamber is placed in the natural environment for incubation. (A part of the figure is reprinted from 23).
doi:10.1128/9781555818821.ch2.1.3.f1
the chamber is recovered, one membrane is peeled off, and
the grown colonies are sampled for purification by reinocula-
tion into new diffusion chambers. Alternatively, the grown
material can be homogenized by passing it through a syringe
equipped with a 25-gauge needle, and inoculated into a new
chamber for another round of in situ incubation (23). A ser-
endipitously made observation suggests that after a round of
in situ growth, up to a quarter of grown species acquire the abil-
ity to propagate on artificial media in the lab. Each successive
round of growth in in situ incubated chambers increases the
fraction of spontaneously domesticated species, opening a
way to grow a significant portion of otherwise uncultivable
species. Recent reports documented advantages of this
method over the conventional techniques because they led
to a significant increase in microbial recovery and isolation
of some uncultivated and phylogenetically new species previ-
ously known only by their molecular signatures (24).
Method 2: I-Chip
The follow-up studies simplified the diffusion chamber
approach and enabled a higher throughput isolation using a
modification of the technology termed the isolation chip
(i-chip) (28). The principle remained the same but was
reduced to practice differently by creating a device consisting
of hundreds of miniaturized diffusion chambers (Fig. 2). Each
 2.1.3. New Devices for Cultivation ▪ 2.1.3-3

FIGURE 2 Isolation chip, or i-chip, for high-throughput microbial cultivation in situ: (a) dipping a plate with multiple through-holes into a
suspension of cells leads to capturing (on average) single cell (b); (c) i-chip assembly; membranes cover arrays of through-holes from each side:
Upper and bottom plates with matching holes press the membranes against the central (loaded) plate. Screws provide sufficient pressure to seal
the content of individual through-holes, each becoming a miniature diffusion chamber containing (on average) a single cell. (Reprinted from
Manual of Industrial Microbiology and Biotechnology [2010], ed. R. H. Baltz, J. E. Davies, and A. Demain, Washington, DC, ASM Press.).
doi:10.1128/9781555818821.ch2.1.3.f2

diffusion mini-chamber is loaded with a single cell (on aver-
age), leading to a pure culture in most chambers after a single
round of in situ incubation. The i-chip is an assembly of flat
plastic ( polyoxymethylene) plates containing multiple regis-
tered through-holes (HI-TECH Manufacturing Schiller Park,
IL). The central plate (72 × 19 × 1 mm) and the two symmet-
rical top and bottom plates (72 × 19 × 6.5 mm each), have
multiple through-holes 1 mm in diameter, arranged in two
arrays with 192 through-holes per array. The area on the
chip having through-holes is completely covered by standard
25- or 47-mm diameter membranes. After the sterilization,
the central plate is dipped into a suspension of cells targeted
for cultivation. As a result, each through-hole captures gel-
microbial suspension containing (on average) one cell per
through-hole. Next, membranes are applied to each array of
through-holes from both sides of the central plate. Four mem-
branes, similar to the ones described above, are required for
the assembly. The top and bottom plates are applied and
aligned, and screws tighten the assembly to provide a seal.
The cells become individually “trapped” in their respective
through-holes and separated—physically but not chemi-
cally—from both each other and the outside environment.
Subsequent in situ incubation in the cell’s original environ-
mental habitat provides the immobilized cells with their
naturally occurring nutrients and growth factors. After incu-
bation, i-chips are washed from the outside and disassembled.
The central plate can then be examined under compound or
high-power dissecting microscope for colony count. Agar
plugs are extracted with thin, sterilized, stainless wire corre-
sponding to the diameter of the chamber for further analyses
such as phylogenetic analysis, genomic analysis, and subculti-
vation. The i-chip method has been extensively used in the
past to grow novel microorganisms from a variety of environ-
ments, including marine waters, soils, and the human micro-
biome (28, 29). Notably, the device was used in the industry
for drug discovery and served as a platform for identification of
2.1.3-4 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION

many new antibiotics, including a representative of a novel
class of antimicrobials (30, 31).
Method 3: Trap
Actinomycetes and fungi grow by forming filaments capable
of penetrating soft substrates and can pass through pores as
small as 0.2 μm. Therefore, a diffusion chamber or i-chip
loaded with sterile agar and membrane pores of larger than
0.2 μm could become a place for colonizing filamentous,
chain-forming, and possibly actively moving cells. Such a
device would allow these organisms to establish new growth
and “trap” it. The same design of i-chip or diffusion chamber
is used, loaded with sterile agar with medium, with a standard
0.2–0.6 μm pore size polycarbonate filter (26). The traps, with
sterile medium with or without addition of nutrients, are
placed in the environment such as soil, aquatic sediments,
and so on. After incubation, the traps are retrieved, the
membranes are peeled off, and the contents are examined
for microbial growth. Visible microcolonies can be sampled
and subcultured for purification on artificial media. In an ear-
lier study targeting soil microorganisms, the majority of
organisms grown in the traps proved to be actinomycetes,
unlike the situation in parallel conventional petri dishes.
Some of the isolates represented rare and unusual species
such as Streptacidiphilus, Catellatospora, Lentzea, and Catenulis-
pora. The trap methods thus appears a useful addition to a

FIGURE 3 Photographs of a 48-chamber HFMC showing (a) overall system, (b) membrane part, (c) injection part, and (d) cross-sectional
SEM image of a hollow fiber membrane. Bar represents 200 µm. (Reprinted from 21). doi:10.1128/9781555818821.ch2.1.3.f3

suite of in situ cultivation methodologies, leading to a selec-
tive capture of filamentous actinomycetes, enriched for new
and rare species.
Human microbiota can also be the target for in situ
cultivation methods. This application requires miniaturiza-
tion of the above-mentioned devices. An earlier study
reported the development and successful application of a
minitrap for isolation of microorganisms from the human
oral cavity (29).
Method 4: Hollow Fiber Membrane Chamber
Liquid culture is a key approach in microbial cultivation. Hol-
low fiber membrane chamber (HFMC) has been developed to
combine the liquid culture with the idea of in situ cultivation
to grow environmental microorganisms under conditions
mimicking their natural habitat (21). A piece of porous hol-
low fiber membrane is used to form a chamber for microbial
cultivation. One chamber unit consists of a porous hollow
fiber polyvinylidene fluoride (PVDF) or poly-sulfone mem-
brane (0.1 µm of mean pore size, 30 cm length, 1.2 mm out-
side diameter, 0.76 mm inside diameter) connected with
injection and sampling devices using syringes. The upper
part serves as the injection port to inoculate cells targeted
for cultivation and the sampling port to withdraw the grown
biomass; to maintain sterility, it is kept capped during cultiva-
tion. The HFMC system accommodates 48–96 such chamber
units, allowing for parallel processing of many cells (Fig. 3).
Microbial cells are sampled from the (aquatic) environ-
ment, serially diluted, and injected into the chambers. The
HFMC is then placed in the natural environment for incuba-
tion. Since the membrane part of the HFMC is immersed in a
liquid phase, the porous membrane allows exchange of
chemical compounds, such as nutrients, metabolites, and sig-
nal molecules, but restricts movement of microbial cells.
Consequently, if the initial inoculum consisted of a single
cell, each chamber will contain a pure culture, grown under
environment-simulating conditions. The chamber volume
used in the previous study was set to 130 µl. However, it is pos-
sible to increase or decrease the volume as needed using a lon-
ger or shorter piece of the hollow fiber membrane. An
important advantage of the hollow fiber membrane is the
contact area between cells inside and the environment out-
side, 5.2 mm2 · mm−3 in the case of a 130 µl chamber volume,
FIGURE 4 Microscopic image of GMDs before the incubation (a) and containing microcolony after the incubation (b); the bright area in the
droplet shown in (b) is a microbial colony. The microcolony containing GMDs was sorted by using cell sorter FACS Aria II (Becton Dickinson)
after the incubation. Bars represent 20 µm. doi:10.1128/9781555818821.ch2.1.3.f4
2.1.3. New Devices for Cultivation ▪ 2.1.3-5
which is approximately 20–40 times higher than that of a
multiwell plate with membranous bottom. Consequently,
molecular exchange is faster, and conditions inside equili-
brate with those outside better.
In the HFMC system, at least the membrane part should be
sterilized by electron beam sterilization (Nuclear Fuel Indus-
try, Osaka, Japan) or autoclaving. When hydrophobic mem-
branes such as PVDF are used, the membrane part must be
hydrophilized before use. This is necessary to remove the air
remaining in the membrane’s pores, which would block the
movement of the outside water into the chamber, as well as
diffusion. A glycerin-dried PVDF membrane (treated with
glycerin solution and dried) represents a practical solution.
Glycerin coated onto the membrane can be removed by
immersing HFMC in water or culture media solution immedi-
ately before cultivation. Ethanol or methanol treatment of
the membrane followed by replacement with water can also
be used for the purpose. According to an earlier study,
HFMC allows for cultivability, expressed as the ratio of the
number of chambers with growth to the number of cells ino-
culated, of 12.3%, 21.0%, and 9.2% of cells from the tidal flat,
activated sludge from sewage treatment plant, and enhanced
biological phosphorous removal (EBPR) process, respec-
tively. That is 6–45 times higher than cultivability in parallel
experiments using conventional approaches. The proportion
of novel phylotypes was also markedly higher in the HFMC
compared with both standard cultivation method as well as
other in situ cultivation techniques described already. Fur-
thermore, the phylogenetic diversity of HFMC-grown isolates
was also higher than that of petri dish isolates.
HIGH-THROUGHPUT CULTIVATION
Screening for rare types requires high-throughput cultivation.
This calls for simplification and optimization of inoculation,
incubation, detection, and selection steps. Standard dishes
and multiwell plates are impractical for processing 104 or
more of isolates. Gel microdroplets offer a potential solution,
as described below.
Method 5: GMDS
GMDs take advantages of the property of emulsion to self-
compartmentalize. Several methods can be used to generate
2.1.3-6 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
GMDs, with most based on water in oil (W/O) emulsions (see
below). If compartmentalization occurs in microbial suspen-
sion, individual cells become trapped inside microdroplets.
Manipulating the abundance of microbes in suspension can
lead to (on average) a single cell per microdroplet, typically
10–100 µm in diameter, which can be easily produced in
the hundreds of thousands or millions. If the cell grows, it
forms a microcolony inside (Fig. 4), and the empty micro-
droplets can be sorted out by a cell sorter. This idea was intro-
duced and tested for microbial cultivation in several reports
(32–38) showing its effectiveness for cultivation of environ-
mental microorganisms.
Generation of GMDs
Several methods to prepare GMDs as emulsion have been
proposed to enable microbial cultivation.
1. Mixing with oil: First, a diluted microbial cell suspen-
sion is mixed with preheated agarose containing a sur-
factant such as Pluronic. Low-temperature agarose is
recommended when using mesophilic microorganisms.
The suspension is then mixed with a certain amount
of mineral silicon oil and stirred to generate w/o emul-
sion. Agar is then solidified by lowering the temperature
down to bellow 5°C, with continuous stirring. GMDs
are retrieved from oil phase by centrifugation. More
specifically, a diluted cell suspension (100 µl) is mixed
with 0.5 ml of preheated (at 37°C) Seaprep agarose
(Lonza, USA), and 25 µl of Pluronic F-68. The mix is
added to 15 ml of Cell Mix Emulsion Matrix and emul-
sified by using a microdrop maker (One Cell Systems,
USA). In the microdrop maker, rapid vortexing contin-
ues for either (a) 1 min, 2,100 rpm, Room temperature,
(b) 1 min, 2,100 rpm, on ice, or (c) 8 min, 1,100 rpm,
on ice. This results in approximately 107 GMDs, with
approximately 10–20% of them containing single
cells. The mixture is centrifuged at 600 g for 10 min
to separate and remove the oil. The GMDs at the
bottom of the vial are then collected for cultivation
experiments.
2. Using membranes: Passing gel mixture through the
porous membrane to the oil phase is another method to
generate w/o emulsion. Agarose gel and microbial cell
mixture is prepared in the same way as above. Then the
mixture is slowly passed through a porous membrane
into the oil phase, with continuous stirring. Recovery
of GMDs follows the above protocol, with details given
in previous reports (33, 34).
3. Using microfluidic device: Recently, highly monodis-
perse emulsions with a narrow distribution of droplet sizes
can be produced by using microfluidic devices (i.e., less
than 3% difference in the diameter) (39, 40). This
approach has an advantage in unifying the size of droplets
more strictly than the method by simply mixing with oil
(method 1), or by using porous membrane (method 2).
However, the uniformity in droplet size produced by
methods 1 and 2 is enough in most cases of microbial cul-
tivation. Therefore, users can choose the appropriate
method based on their own experimental purpose and
lab settings.
Sorting GMDs Using Flow Cytometric Cell Sorter
After incubation, GMDs with microbial growth can be sorted
to retrieve a specific subset of GMDs. For example, GMDs
containing colonies can be separated, in a high-throughput
fashion, by flow cytometry. A study by Zengler et al. (38)
showed that such separation can be achieved because
GMDs with and without colonies have different light scatter-
ing properties in both forward scatter and side scatter. Addi-
tionally, GMDs containing microbial biomass can be
detected and sorted based on fluorescence if stained with a
nontoxic dye, such as CFDA and SYTOX green (34). This
method has a very low detection limit and as such may be
advantageous over sorting based on light scatter.
POTENTIAL OF MICRO-FABRICATION
TECHNOLOGIES
Micro-fabrication technologies, especially micro-engineered
mechanical systems methods, are promising to revolutionize
microbial cultivation (41). Ingham et al. (42) used these
technologies to build and test a miniaturized cultivation
chip, composed of a highly porous aluminum oxide sheet
with growth compartments a few microns each. This culture
format allowed cultivation of up to one million separate
microbial colonies on a single chip. The smallest compart-
ment size reported is 7 × 7 µm, allowing for up to 1 million
microwells on an 8 × 36 mm chip. A larger modification
(8 × 36 mm chip, 20 × 20 µm growth areas, 20 µm wide
walls) contained 180,000 individual growth compartments.
This chip can be used for several purpose as shown in the orig-
inal report, such as (a) monitoring of the growth, (b) high-
throughput screening targeting specific activity which can
be distinguished via fluorescent signals, and (3) isolation of
environmental microorganisms. Another method based on
similar technology has been proposed for a continuous,
chemostat-like cultivation, with an additional advantage of
an easy microscopy examination of growth (43). Micro-
fabrication technologies have an exciting potential to change
the way we cultivate microorganisms, principally because
they provide a significant increase in throughput. However,
significant challenges remain to be worked out.
CONCLUDING REMARKS
Cultivation-independent approaches such as rRNA-,
genomic-, or metagenomic-based methods significantly
impacted the field of environmental microbiology. Nonethe-
less, some of the key aspects of microbial biology can be best
studied using microbial cultures and cocultures, especially
understanding of physiology and biotechnological applica-
tions. Recent innovations in cultivation have brought to
the lab a wealth of new species, and prospective technologies
are promising to significantly improve the throughput.
REFERENCES
1. Staley JT, Konopka A. 1985. Measurement of in situ activ-
ities of nonphotosynthetic microorganisms in aquatic and
terrestrial habitats. Annu Rev Microbiol 39:321–346.
2. DeLong EF. 1992. Archaea in coastal marine environments.
Proc Natl Acad Sci USA 89:5685–5689.
3. Dojka MA, Harris JK, Pace NR. 2000. Expanding the
known diversity and environmental distribution of an uncul-
tured phylogenetic division of bacteria. Appl Environ Micro-
biol 66:1617–1621.
4. Fuhrman JA, McCallum K, Davis AA. 1992. Novel major
archaebacterial group from marine plankton. Nature 356:
148–149.
5. Giovannoni SJ, Britschgi TB, Moyer CL, Field KG. 1990.
Genetic diversity in Sargasso Sea bacterioplankton. Nature
345:60–63.

6. Hugenholtz P, Goebel BM, Pace NR. 1998. Impact of
culture-independent studies on the emerging phylogenetic
view of bacterial diversity. J Bacteriol 180:4765–4774.
7. Liesack W, Stackebrandt E. 1992. Occurrence of novel
groups of the domain Bacteria as revealed by analysis of
genetic material isolated from an Australian terrestrial envi-
ronment. J Bacteriol 174:5072–5078.
8. Ravenschlag K, Sahm K, Pernthaler J, Amann R. 1999.
High bacterial diversity in permanently cold marine sedi-
ments. Appl Environ Microbiol 65:3982–3989.
9. Ward DM, Weller R, Bateson MM. 1990. 16S rRNA
sequences reveal numerous uncultured microorganisms in a
natural community. Nature 345:63–65.
10. Bruns A, Cypionka H, Overmann J. 2002. Cyclic AMP
and acyl homoserine lactones increase the cultivation effi-
ciency of heterotrophic bacteria from the central Baltic
Sea. Appl Environ Microbiol 68:3978–3987.
11. Bruns A, Nubel U, Cypionka H, Overmann J. 2003. Effect
of signal compounds and incubation conditions on the cul-
turability of freshwater bacterioplankton. Appl Environ
Microbiol 69:1980–1989.
12. Connon SA, Giovannoni SJ. 2002. High-throughput meth-
ods for culturing microorganisms in very-low-nutrient media
yield diverse new marine isolates. Appl Environ Microbiol 68:
3878–3885.
13. Cho JC, Giovannoni SJ. 2004. Cultivation and growth
characteristics of a diverse group of oligotrophic marine gam-
maproteobacteria. Appl Environ Microbiol 70:432–440.
14. Davis KER, Joseph SJ, Janssen PH. 2005. Effects of growth
medium, inoculum size, and incubation time on culturability
and isolation of soil bacteria. Appl Environ Microbiol 71:
826–834.
15. Janssen PH, Yates PS, Grinton BE, Taylor PM, Sait M.
2002. Improved culturability of soil bacteria and isolation
in pure culture of novel members of the divisions Acidobac-
teria, Actinobacteria, Proteobacteria, and Verrucomicrobia.
Appl Environ Microbiol 68:2391–2396.
16. Joseph SJ, Hugenholtz P, Sangwan P, Osborne CA, Jans-
sen PH. 2003. Laboratory cultivation of widespread and pre-
viously uncultured soil bacteria. Appl Environ Microbiol 69:
7210–7215.
17. Sangwan P, Kovac S, Davis KER, Sait M, Janssen PH.
2005. Detection and cultivation of soil Verrucomicrobia.
Appl Environ Microbiol 71:8402–8410.
18. Sipkema D, Schippers K, Maalcke WJ, Yang Y, Salim S,
Blanch HW. 2011. Multiple approaches to enhance the
cultivability of bacteria associated with the marine
sponge Haliclona (gellius) sp. Appl Environ Microbiol 77:
2130–2140.
19. Stevenson BS, Eichorst SA, Wertz JT, Schmidt TM, Brez-
nak JA. 2004. New strategies for cultivation and detection of
previously uncultured microbes. Appl Environ Microbiol 70:
4748–4755.
20. Tamaki H, Sekiguchi Y, Hanada S, Nakamura K, Nomura
N, Matsumura M, Kamagata Y. 2005. Comparative analysis
of bacterial diversity in freshwater sediment of a shallow
eutrophic lake by molecular and improved cultivation-based
techniques. Appl Environ Microbiol 71:2162–2169.
21. Aoi Y, Kinoshita T, Hata T, Ohta H, Obokata H, Tsuneda
S. 2009. Hollow-fiber membrane chamber as a device for in
situ environmental cultivation. Appl Environ Microbiol 75:
3826–3833.
22. Ben-Dov E, Kramarsky-Winter E, Kushmaro A. 2009. An
in situ method for cultivating microorganisms using a double
encapsulation technique. FEMS Microbiol Ecol 68:363–371.
23. Bollmann A, Lewis K, Epstein SS. 2007. Incubation of
environmental samples in a diffusion chamber increases
the diversity of recovered isolates. Appl Environ Microbiol
73:6386–6390.
2.1.3. New Devices for Cultivation ▪ 2.1.3-7
24. Bollmann A, Palumbo AV, Lewis K, Epstein SS. 2010. Iso-
lation and physiology of bacteria from contaminated subsur-
face sediments. Appl Environ Microbiol 76:7413–7419.
25. Ferrari BC, Binnerup SJ, Gillings M. 2005. Microcolony
cultivation on a soil substrate membrane system selects for
previously uncultured soil bacteria. Appl Environ Microbiol
71:8714–8720.
26. Gavrish E, Bollmann A, Epstein S, Lewis K. 2008. A trap
for in situ cultivation of filamentous actinobacteria. J Micro-
biol Meth 72:257–262.
27. Kaeberlein T, Lewis K, Epstein SS. 2002. Isolating “uncul-
tivable” microorganisms in pure culture in a simulated natu-
ral environment. Science 296:1127–1129.
28. Nichols D, Cahoon N, Trakhtenberg EM, Pham L, Mehta
A, Belanger A, Kanigan T, Lewis K, Epstein SS. 2010. Use
of Ichip for high-throughput in situ cultivation of “uncultiv-
able” microbial species. Appl Environ Microbiol 76:2445–2450.
29. Sizova MV, Hohmann T, Hazen A, Paster BJ, Halem SR,
Murphy CM, Panikov NS, Epstein SS. 2012. Cultivability
of oral bacteria: new approaches for isolation of previously
uncultivated species. Appl Environ Microbiol 78:194–203.
30. Ling LL, Schneider T, Peoples AJ, Spoering AL, Engels I,
Conlon BP, Mueller A, Schäberle TF, Hughes DE, Epstein
S, Jones M, Lazarides L, Steadman VA, Cohen DR, Felix
CR, Fetterman KA, Millett WP, Nitti AG, Zullo AM,
Chen C, Lewis K. 2015. A new antibiotic kills pathogens
without detectable resistance. Nature 517:455–459.
31. Gavrish E, Sit CS, Cao S, Kandror O, Spoering A, Peoples
A, Ling L, Fetterman A, Hughes D, Bissell A, Torrey H,
Akopian T, Mueller A, Epstein S, Goldberg A, Clardy J,
Lewis K. 2014. Lassomycin, a ribosomally synthesized cyclic
peptide, kills Mycobacterium tuberculosis by targeting the
ATP-dependent protease ClpC1P1P2. Chem Biol 21:
509–518.
32. Akselband Y, Cabral C, Castor TP, Chikarmane HM,
McGrath P. 2006. Enrichment of slow-growing marine
microorganisms from mixed cultures using gel microdrop
(GMD) growth assay and fluorescence-activated cell sorting.
J Exp Mar Biol Ecol 329:196–205.
33. Katsuragi T, Tanaka S, Nagahiro S, Tani Y. 2000. Gel
microdroplet technique leaving microorganisms alive for
sorting by flow cytometry. J Microbiol Meth 42:81–86.
34. Manome A, Zhang H, Tani Y, Katsuragi T, Kurane R,
Tsuchida T. 2001. Application of gel microdroplet and
flow cytometry techniques to selective enrichment of non-
growing bacterial cells. FEMS Microbiol Lett 197:29–33.
35. Nir R, Lamed R, Gueta L, Sahar E. 1990. Single-cell
entrapment and microcolony development within uniform
microspheres amenable to flow cytometry. Appl Environ
Microbiol 56:2870–2875.
36. Park J, Kerner A, Burns MA, Lin XN. 2011.
Microdroplet-enabled highly parallel co-cultivation of
microbial communities. PLoS One 6:e17019.
37. Weaver JC, Williams GB, Klibanov A, Demain AL. 1988.
Gel microdroplets: rapid detection and enumeration of indi-
vidual microorganisms by their metabolic activity. Nat Bio-
technol 6:1084–1089.
38. Zengler K, Toledo G, Rappé M, Elkins J, Mathur EJ, Short
JM, Keller M. 2002. Cultivating the uncultured. Proc Natl
Acad Sci USA 99:15681–15686.
39. Eun YJ, Utada AS, Copeland MF, Takeuchi S, Weibel DB.
2011. Encapsulating bacteria in agarose microparticles using
microfluidics for high-throughput cell analysis and isolation.
ACS Chem Biol 6:260–266.
40. Joensson HN, Svahn HA. 2012. Droplet microfluidics—A
tool for single-cell analysis. Angew Chem Int Ed 51:
12176–12192.
41. Ingham CJ, van Hylckama Vlieg JET. 2008. MEMS and
the microbe. Lab Chip 8:1604–1616.
2.1.3-8 ▪ CULTURE-BASED AND PHYSIOLOGICAL DETECTION
42. Ingham CJ, Sprenkels A, Bomer J, Molenaar D, van den
Berg A, van Hylckama Vlieg JET, de Vos WM. 2007.
The micro-Petri dish, a million-well growth chip for the cul-
ture and high-throughput screening of microorganisms. Proc
Natl Acad Sci USA 104:18217–18222.
43. Hesselman MC, Odoni DI, Ryback BM, de Groot S, van
Heck RGA, Keijsers J, Kolkman, Nieuwenhuijse D, van
Nuland YM, Sebus E, Spee R, de Vries H, Wapenaar
MT, Ingham CJ, Schroën K, Martins dos Santos VAP,
Spaans SK, Hugenholtz F, van Passel MWJ. 2012. A multi-
platform flow device for microbial (Co-) cultivation and
microscopic analysis. PLoS One 7:e36982.
44. Winterberg H. 1898. Zur Methodik der Bakterienzahlung. Z
Hyg Infektionskr 29:75–93.
 Gold-Based In Situ Hybridization for Phylogenetic
Single-Cell Detection of Prokaryotes in
Environmental Samples
THILO EICKHORST AND HANNES SCHMIDT
2.2.1
IN SITU DETECTION OF MICROBIAL
CELLS IN ENVIRONMENTAL
MICROBIOLOGY
The molecular technique of in situ hybridization (ISH) was
originally developed in the field of cytogenetics to detect
specific genes on human chromosomes (1). During the past
decades, the use of fluorescently labeled oligonucleotide
probes in ISH has been proven most feasible for the detection
of specific target DNA sequences (2, 3), this is why the
method is commonly known as fluorescence in situ hybridiza-
tion (FISH). In environmental microbiology, FISH has been
introduced since the 1990s following its first application as
rRNA targeted phylogenetic stain for the detection of single
microbial cells (4, 5). It became a powerful method for the
identification, visualization, and enumeration of mainly pro-
karyotic cells either in combination with DNA sequencing
and probe design or phylogenetic differentiation by selecting
probes in hierarchical order (top-to-bottom approach; 6).
However, the application in various environments resulted
in a few limitations due to complexity of habitats consist-
ing of minerals and organic substances or reduced microbial
physiological activities (7, 8). The latter issue has been over-
come by the introduction of catalyzed reporter deposition
(CARD)-FISH (9) combining the ISH of oligonucleotide
probes to target rRNA with a subsequent signal amplification
of fluorescently labeled tyramides. This improvement resulted
in higher detection rates and introduced a new era for the
application of FISH in complex and less active environments
(10). Several studies show the applicability of CARD-
FISH for the quantification of absolute and relative abundan-
ces of target microorganisms (11–13). In recent years various
combinative FISH approaches have been developed (14),
including the single-cell detection of metabolic activity
(15), functional gene expression (16, 17), and nutrient
uptake (18, 19). FISH has also been extended to the tracking
of ribosomal molecules on the single cell level via superreso-
lution light microscopy (20).
With its potential for phylogenetic identification of
single microbial cells, FISH is one of the key methods for
the study of microbial ecology in environmental systems.
In this context, correlative analysis of microorganisms and
their microbial habitat is of great importance to investigate
microbe–habitat interactions. Sophisticated microscopic,
doi:10.1128/9781555818821.ch2.2.1
2.2.1-1
spectroscopic, and tomographic methods have recently
been introduced for the analysis of biogeochemical interfaces
and are useful to characterize the physiochemical properties
of microbial habitats in soil, biofilms, flocs, and slimes (e.g.,
21–23). For soils, the combination of FISH and micro-
morphological impregnation techniques allows the detection
of individual native microbial cells in the undisturbed soil
matrix (24). This approach would benefit from correlative
analyses characterizing the environmental living conditions
in terms of substrate availability or matter budget which
could be done with scanning electron microscopy and
energy-dispersive X-ray spectroscopy (SEM-EDS; 25), near-
edge X-ray fine structure spectroscopy (26) or nanoscale sec-
ondary ion mass spectrometry (nanoSIMS; 27). In terms of
the physical characterization of microbial environments
such as habitable pore space and availability of moisture,
X-ray-based computed tomography (X-ray µCT) is of high
potential in environmental systems such as sediments and
soils (22, 28).
The combination of those techniques with the identifica-
tion, visualization, and localization of individual microbial
cells on scales beyond light microscopy has great potential
in environmental microbiology. Especially in habitats such
as soils and sediments with complex matter composition,
the analysis of areas related to plant roots or material surfaces
would benefit from correlative approaches.
As most of the aforementioned techniques are based on
X rays, the detectability of microbial cells by electromagnetic
radiation harbors a huge potential for various analytical
combinations. Therefore the labeling of cells with elements
of high density is desirable for a phylogenetic identifica-
tion of single microbial cells on the same level as that com-
monly achieved for rRNA using FISH. This principal has
been utilized for combinations of FISH and electron micro-
scopy (EM) for the in situ identification and localization
of whole cells in environmental microbe–surface inter-
faces. Several ISH procedures have been developed to detect
microbial populations via the deposition of nanometer-
sized gold particles in and on the target organisms. Taking
advantage of the high atomic mass of gold, single microbial
cells have been analyzed by SEM-EDS (29), transmission
electron microscopy (TEM; 30), and X-ray fluorescence
microscopy (31).
2.2.1-2 ▪ MICROSCOPIC METHODS
NANOGOLD AS A MARKER FOR IN SITU
DETECTION OF MICROBES
The physicochemical characteristics of nanogold allow its
utilization as an optimal label for EM and X-ray-based techni-
ques. Due to their small, uniform size as well as the high
density and atomic mass number, gold particles yield strong
signals obtained by both secondary electron and back-
scattered electron detectors in EM (32). Commercially avail-
able reagents range from conjugates readily labeled with
0.8–1.4 nm nanogold particles (antibodies, lipids, proteins)
to functionalized gold particles of different size (0.8–5 nm;
Table 1). These surface-modified particles may be used either
to bind target molecules in the specimen of interest directly or
to prepare gold-labeled conjugates for individual appli-
cations. Conjugates such as Fab0 -/immunoglobulin G– or
streptavidin-nanogold may additionally be labeled with fluo-
rophores for correlative analysis of specimens using electron
and fluorescence microscopy (33).
Nanogold represents a widely applied and frequently used
marker for antibody labeling of specimens in immunocyto-
chemistry since its introduction in 1971 (32). In environ-
mental samples, immunogold labeling was also applied to
detect and visualize microbes within plant roots or biofilms
(34–36). With the advent of rRNA-based ISH, the idea of
using gold-labeled nucleotide probes for whole-cell detection
of microbes in environmental samples via EM aroused inter-
est (6). To our knowledge, the first successful application of
ISH applying nanogold- and fluorescently labeled nucleotide
probes for an identification of uncultivated environmental
microbes was reported by Spring et al. (29). Marine magnetic
bacteria were identified via standard FISH followed by a post-
embedding hybridization of thin sections using nanogold-
labeled polynucleotide probes to analyze cell morphology
and the presence of magnetosomes via TEM.
Almost a decade later, two working groups reported the
first applications of gold-based ISH techniques for the detec-
tion and identification of whole microbial cells via EM
(Table 2). Gérard et al. (30) applied fluorescein-labeled oli-
gonucleotide probes followed by an anti-FITC-immunogold
conjugate to detect pure cultures of Escherichia coli and Dein-
ococcus radiodurans via fluorescence microscopy, SEM, and
TEM. High detection efficiencies and the combination of
images obtained from secondary electrons as well as from
back-scattered electrons allowed for a clear identification of
positively labeled cells. In a follow-up study by Ménez et al.
(31), pure cultures were deposited on various substrates
and detected using scanning X-ray microscopy and synchro-
tron radiation. Cells of D. radiodurans were localized on
TABLE 1 Selection of commercially available nanogold particles
and conjugates
Modifications Particle size
Nanogold 0
Antibodies (Fab and IgG) 0.8, 1.4 nm
conjugates Lipids (Palmitoyl and DPPE) 0.8, 1.4 nm
Proteins (Albumin, streptavidin) 0.8, 1.4 nm
Nanogold Charged (+/−) or uncharged 0.8, 1.4 nm
particles Reactive groups 0.8, 1.4 nm
Amphilic, hydrophilic/-phobic 2–4, 3–5 nm
Fab0 : fragment antigen binding; IgG: immunoglobulin G; DPPE: dipalmi-
toyl phosphatidylethanolamine.
siderite based on specific elemental composition, showing
the possibility of linking phylogenetic identification with
complementary analyses of associated mineral phases. On
the surface of more complex substrates such as basaltic glass,
the discrimination of positively stained cells was hampered
due to weaker contrast of the resulting elemental maps.
Thus, an application of the technique as presented may
be insufficient for a reliable detection of uncultured micro-
bial cells in environmental samples such as sediments
or soil.
In a slightly different approach, Kenzaka et al. (37) ap-
plied biotin-labeled oligonucleotide probes followed by a
streptavidin-nanogold conjugate to analyze single microb-
ial cells in pure cultures, water samples, and sediments via
SEM. Cells belonging to the phylum Bacteroidetes were
detected on sediment particles in situ and microbial cell num-
bers obtained with standard FISH or SEM-ISH were well cor-
related for water samples. However, up to 50% of the total
cells were missed by either FISH or SEM-ISH, perhaps due
to low rRNA contents or insufficient penetration of microbial
cell walls by oligonucleotide probes, resulting in weak fluores-
cent and gold signals. The general applicability of the techni-
que was later confirmed by Jang et al. (38), who specifically
detected pure cultures of E. coli using biotinylated oligonu-
cleotide probes followed by the binding of a streptavidin-
nanogold conjugate and EM.
Recently, two methodological advancements of gold-
based applications were presented. Schmidt et al. (39) intro-
duced an enzymatic amplification of biotinylated binding
sites for nanogold-labeled conjugates to increase the number
of gold particles deposited per oligonucleotide probe, and
thus to enhance signal intensity as it is known for CARD pro-
cedures. Whole cells of pure strains as well as native bacteria
have been detected via fluorescence microcopy and SEM due
to the utilization of conjugates labeled with fluorophores
and nanogold particles. For the detection via SEM-EDS
enhancement of gold signals is required for this approach.
The gold-ISH method of Kubota et al. (40) is based on the
hybridization of oligonucleotide probes being conjugated
with one undecagold particle (0.8 nm) and one fluorophore
at each prime end, respectively. First assessments of the tech-
nical applicability in pure cultures and anaerobic sludge sam-
ples labeled with 13C were analyzed via nanoSIMS ion
imaging. Specific gold-ISH signals were detected as 197Au
counts but would still require signal enhancement. Gold-ISH
may be an alternative for ISH with halogenated oligonucleo-
tide probes commonly used for the specific detection of single
microbial cells in nanoSIMS analysis.
In a modified application of the gold-FISH protocol (39)
Almstrand et al. (41) used biotinylated probes targeting mul-
tiple ribosomal sites in order to increase signal intensities.
This method is referred to as polygold-FISH and utilizes
gold-labeling and enhancement analog to Schmidt et al.
(39) but has not been tested on environmental samples so far.
Mostly applied as 0.8 or 1.4 nm particles (Table 1),
nanogold-labeled conjugates are small enough to penetrate
microbial cell walls to specifically bind to the respective target
molecules (e.g., rRNA; 42). The detection of individual
nanogold particles on or within microbial cells, in turn, is lim-
ited by the small size of the gold particles and the resolving
power of EM or other element-specific detection techniques.
Consequently, a prerequisite for routine detection of posi-
tively stained cells is the enlargement of single nanogold par-
ticles by autometallography. The underlying principle of
autometallography is the reduction of silver or gold ions in
the vicinity of individual gold nanoparticles as catalyzed by
TABLE 2 Original approaches for gold-based in situ detection of single cells
Probe label Gold-labeled Size of gold Enhancement of
Designation Signal amplification Applied to Analyzed via Reference
(5'-end) substrate particles (nm) gold particles

2.2.1. Gold-Based In Situ Hybridization for Phylogenetic Single-Cell ▪ 2.2.1-3

– Digoxigenin/ – Anti-fluorescein- 6/10/15 – Enriched sediment, FM/SEM-EDS/ Spring et al., 1998
fluorescein nanogold water culture TEM (29)
– Fluorescein – Anti-FITC- 0.8 Silver Pure culture FM/SEM/TEM Gérard et al., 2005
immunogold (30)
SEM-ISH Biotin – Streptavidin- 1.4 Gold Pure culture, water, SEM Kenzaka et al.,
nanogold sediment 2005 (37)
– Fluorescein – Anti-FITC- 0.8 Silver Pure culture deposited SXM/XRF Ménez et al., 2007
immunogold on substrate (31)
Gold-FISH HRP Biotinylated tyramide Streptavidin-AF488- 1.4 Gold Pure culture, soil, FM/SEM-EDS Schmidt et al.,
nanogold inoculated rice root 2012 (39)
Gold-ISH Undecagold – – 0.8 – Pure culture, anaerobic FM/nanoSIMS Kubota et al., 2014
sludge (40)
Polygold- Biotin Multiple ribosomal Streptavidin-AF488- 1.4 Gold Pure culture FM/SEM-EDS Almstrand et al.,
FISH target sites nanogold 2015 (41)
FM: Fluorescence microscopy; SEM: Scanning electron microscopy; EDS: Energy-dispersive X-ray spectroscopy; TEM: Transmission electron microscopy; FITC: Fluorescein isothiocyanate; SXM: Scanning X-ray
microscopy; XRF: X-ray fluorescence; HRP: Horseradish peroxidase; AF488: Alexa Fluor 488; nanoSIMS: nano-scale secondary ion mass spectrometry.
2.2.1-4 ▪ MICROSCOPIC METHODS
reducing agents (43). Thus, small gold or silver particles
deposit on the surface of individual nanoparticles, resulting
in the growth of gold or silver grains, respectively. This
autometallographic enhancement largely increases the sensi-
tivity of the detection procedure. Gold is regarded as superior
to silver as it is less susceptible to low pH values and autonu-
cleation (less background) while giving better back-scatter
signals due to the higher electron density (44). In environ-
mental studies both enhancement techniques were necessa-
rily applied when nanogold particles of 1.4 nm or smaller
needed to be detected by EM (Table 2).
However, this postenlargement of nanogold particles rep-
resents the most crucial and error-prone step in studies on
gold-based detection of individual cells in environmental
samples. Background signals of randomly distributed metallic
particles were frequently observed via EM, which is a direct
result of the fusion and precipitation of gold and silver ions
in the enhancement solution (30, 31, 39, 45). To prevent
such extensive autonucleation of gold or silver ions, the incu-
bation time has to be reduced until the minimum size of
detectable gold or silver grains is reached. In addition, stop-
ping solutions such as sodium thiosulfate have been success-
fully applied to terminate the enhancement process (39,
46). Besides problems arising from autonucleation, gold or sil-
ver ions may bind nonspecifically to charged minerals such as
clay particles in soil and other environmental samples. Exten-
sive pre- and postwashing steps and a reduction of critical
reagents (oligonucleotide probes, nanogold conjugates)
FIGURE 1 Workflow diagram of the gold-FISH method. Modified from (39); HRP: horseradish peroxidase, Au: 1.4 nm nanogold particle,
AF488: Alexa Fluor 488. Reprinted from Systematic and Applied Microbiology (39) with permission from the publisher.
doi: 10.1128/9781555818821.ch2.2.1.f1
have been reported to minimize nonspecific background
and to facilitate the discrimination of positive signals from
the underlying specimen structure (30, 39). In general, the
enlargement process has to be adjusted for each sample of
interest to specifically detect whole cells with the best
signal-to-background ratio possible.
GOLD-FISH: SINGLE CELL DETECTION BY
FLUORESCENT AND NANOGOLD LABELING
The advantage of the recently developed gold-FISH protocol
is to detect single microbial cells in environmental samples on
different microscopic scales (39). Based on 16S rRNA tar-
geted ISH, a fluorescent dye and nanogold particles are intro-
duced into whole microbial cells, allowing for a correlative
identification and localization of positively stained microor-
ganisms via epifluorescence and SEM. The protocol has
been developed based on individual and mixed microbial
strains and applied to a variety of environmental samples.
As usual for whole cell hybridization procedures, cells are
fixed with formaldehyde and cell walls are permeabilized
by enzymatic lysis (lysozyme and achromopeptidase) to sup-
port the introduction of larger probe molecules and subse-
quent tyramide signal amplification. The ISH resembles the
basic step of CARD-FISH by using horseradish peroxidase–
labeled oligonucleotide probes. Subsequently, biotinylated
tyramides are activated through the horseradish peroxidase
 2.2.1. Gold-Based In Situ Hybridization for Phylogenetic Single-Cell ▪ 2.2.1-5

molecule of positively bound probes in the target cell resulting
in covalent binding of tyramide radicals (Fig. 1). These are the
binding sites for streptavidin molecules conjugated with a
fluorophore and a 1.4 nm nanogold particle (FluoroNano-
gold-Streptavidin, Nanoprobes). The detection of fluores-
cent signals can be done after this step directly, whereas for
the analysis via EM a subsequent enhancement of nanogold
particles is essential. This is done by autometallography,
resulting in an agglomeration of gold nanoparticles around
the previously introduced nanogold-streptavidin conjugate
(GoldEnhance EM, Nanoprobes).
As a proof of principle, the application of this protocol on
single strains of E. coli and Roseobacter sp. as well as a mixed
culture thereof is shown for the detection via fluorescence
and electron microscopy (Fig. 2). For the latter, enhanced
nanogold particles were detected by back-scattered electron
detector (BSE) but also in the secondary electron image itself.
The CARD-based ISH and signal amplification resulted in
distinct signals of both markers, the fluorescent dye and the
nanogold particles. The advantage of this approach over
ISHs without signal amplification has been proven by direct
comparisons with ISHs using monolabeled oligonucleotide
probes modified with biotin (Fig. 2a). In addition, the specif-
icity of the hybridization procedure was shown as no false sig-
nals were detected in negative controls on mixed cultures
(Fig. 2e).
Gold-FISH has been applied on a variety of environmen-
tal samples to date. Enumerations of single microbial cells
in soil and sediment samples showed comparable results to
those obtained with established CARD-FISH protocols
(12, 47). Fluorescent signals were only slightly weaker com-
pared to CARD-FISH (Fig. 3a), which was confirmed for
the visualization of single microbial cells of Rhizobium legumi-
nosarum on the rhizoplane of inoculated rice roots as well.
The corresponding EM images showed the potential of ISH-
detected cells on higher resolutions offering a more detailed
view on microbial rhizoplane colonization compared to the
light microscopic scale (Fig. 3b). In a recent application, sin-
gle cells of Bacteroidetes associated with marine diatoms (48)
were detected after gold-FISH analysis on different scales via
SEM and BSE. Merging both image data revealed a more dis-
tinct view of the association and interaction of specifically

FIGURE 2 Gold-FISH detected bacteria on polycarbonate filters. (a) E. coli hybridized with mono-gold-FISH, probe GAM42a-biotin and
AF 488 fluoro-nanogold–streptavidin conjugate. (b) E. coli hybridized with gold-FISH, probe GAM42a-HRP, biotinylated tyramide, and AF
488 fluoro-nanogold–streptavidin conjugate. (c) Mixed culture of E. coli and Roseobacter strain AK199 hybridized with gold-FISH, probe
ROS537-HRP, biotinylated tyramide, and AF 488 fluoro-nanogold–streptavidin conjugate. Numbers indicate corresponding images of (1) flu-
orescence microscopy, (2) SEM, and (3) BSE. (d) DAPI counterstain of image (c1). (E) Mixed culture of E. coli and Roseobacter strain AK199
hybridized with probe NONEUB. Scale bars: fluorescent micrographs 20 µm, SEM/BSE 2 µm. SEM and BSE images reprinted from Systematic
and Applied Microbiology (39) with permission from the publisher. doi: 10.1128/9781555818821.ch2.2.1.f2
2.2.1-6 ▪ MICROSCOPIC METHODS

FIGURE 3 Application of gold-FISH to environmental samples. (a) Bacteria in marine sediment, probe EUBI-III, scale bar 20 µm. (b) Rhi-
zobium leguminosarum on the root surface of Oryza sativa L. (wetland rice), probe: ALF968; (b1) fluorescence microscopy, scale bar 20 µm; (b2)
corresponding SEM-image, scale bar 10 µm; (b3) SEM image at higher magnification (red frame in image b2), scale bar 1 µm. (c) Bacteroidetes
associated with marine diatomes, probe: CF319a, (c1) SEM image, scale bar 5 µm; (c2) BSE-image of red frame in image (c1), scale bar 5 µm;
(c3) merged image of SEM (reddish color) and BSE (bluish color) of red frame in (c2), scale bar 1 µm. Parts a and b1–b3 reprinted from System-
atic and Applied Microbiology (39) with permission from the publisher. doi: 10.1128/9781555818821.ch2.2.1.f3

gold-labeled cells with their diatom microenvironment
including other nonlabeled microorganisms (Fig. 3c).
Nonspecific background of gold particles derived from
autometallographic gold enhancement was often observed
in environmental samples of complex organomineral compo-
sition. However, the density of gold particles associated with
target cells was much higher and therefore sufficient to reli-
ably localize these microorganisms via EM. This effect could
be used for the detection of gold-FISH labeled cells via BSE
electron microscopy and digital image analysis. Therefore in
BSE images a threshold was established where the resulting
binary images were analyzed considering only larger clusters
of BSE-detected particles to represent individual microbial
cells. These cells are highlighted in violet in the analyzed
BSE image while the unspecific detected particles are yellow
(Fig. 4a). This is a relatively easy way to evaluate EM images
according to the localization of successfully gold-FISH
labeled microbial cells.
Since the BSE signal is based on elements with higher
atomic numbers, it can be related to other substances in
mineral-containing environmental samples. The differen-
tiation of signal intensities is not very reliable, although
the introduced gold particles should result in the highest
BSE numbers in most samples. To demonstrate a specific
detection of gold, SEM with EDS has been utilized. Clear
gold signals were recorded when EDS spot analyses have
been applied on individual gold-FISH detected cells (Fig.
4b). With sophisticated EDS detectors of high sensitivity,
elemental mappings could be done. In the environmental
sample of Bacteroidetes being associated with marine dia-
toms, elemental gold could be mapped clearly on the corre-
sponding cells. These data can be easily combined with
elemental mapping of the environmental sample, such as
selected nutrients or other elements involved in microbial
processes (Fig. 4c).
In complex environmental samples, the autometallo-
graphic enhancement of nanogold particles frequently
resulted in the formation of relatively large and unspecific
cocci-shaped elemental gold structures. This was observed
especially in a clayey soil sample and could be the result of
negatively charged clay minerals (Fig. 4d). A similar effect
was observed on polycarbonate filters being processed in
plastic petri dishes while conducting the gold-FISH protocol
(Fig. 4e). The latter effect can be avoided by using glassware
throughout the entire procedure.
POTENTIALS FOR GOLD-BASED DETECTION
TECHNIQUES IN ENVIRONMENTAL
MICROBIOLOGY
Gold-based detection techniques are well suited when com-
plementary information on morphological, structural, and
chemical characteristics of the same region of a sample is
required. Such applications of correlative microscopy gener-
ally make use of a molecular biological labeling that is eval-
uated by fluorescence microscopy followed by a postlabeling
procedure for EM (49). Reagents conjugated with markers
for both fluorescence and EM (e.g., FluoroNanogold) can
be introduced into the sample in a single labeling procedure
and may hold an advantage over traditional approaches
with two or more labeling steps (33). To date, correlative
microscopy has been mainly applied to thin sections of
human, animal, and plant specimens, and yet has to be
 2.2.1. Gold-Based In Situ Hybridization for Phylogenetic Single-Cell ▪ 2.2.1-7

FIGURE 4 Digital image analysis, elemental mapping, and artifacts of gold-FISH applications. (a) Bacteroidetes associated with marine
diatoms, probe: CF319a, scale bar 5 µm; (a1) SEM image, (a2) result of digital image analysis, purple areas indicate detected bacterial cells
and yellow areas indicate non-specific gold deposition; (b) SEM-EDS analysis of a gold-FISH detected Bacteroidetes cell on a diatom, scale
bar 1 µm; (c) elemental mapping of silicon (Si-KA, c1), gold (Au-LA, c2), and merged data including SEM image (c3), scale bar 5 µm;
(d) SEM and BSE image of nonspecific gold deposition in soil samples hybridized with gold-FISH, scale bar 5 µm; (e) nonspecific gold dep-
osition on polycarbonate filters in the vicinity of gold-FISH detected E. coli cells, scale bar 5 µm. Parts d1 and d2 reprinted from Systematic
and Applied Microbiology (39, supplementary data) with permission from the publisher. doi: 10.1128/9781555818821.ch2.2.1.f4
2.2.1-8 ▪ MICROSCOPIC METHODS
routinely used to detect and analyze single microbial cells in
environmental samples on various microscopic scales.
A promising means of correlative analysis in environmen-
tal microbiology is the combination of gold-based ISH and
nanoSIMS. NanoSIMS provides information on the isotopic
composition of specimens on a biologically meaningful sub-
micron level and has been successfully applied to soil environ-
ments, demonstrating that assimilatory processes can be
detected and visualized in situ within soil and roots (50–52).
Combinatorial approaches with (CARD-)FISH have become
a state-of-the-art technique to investigate substrate uptake
and thereby activity of individual microorganisms in environ-
mental samples at the single-cell level (19, 53, 54). Such
FISH-NanoSIMS approaches link the identity and function
of native microorganisms (e.g., nitrogen-fixing bacteria)
and have been mainly applied in nonterrestrial environments
including marine water samples (55), microbial mats (56),
and deep-sea sediments (57). Very recently, a first combina-
tion of ISH with gold-conjugated oligonucleotide probes
and nanoSIMS was reported (40). Single microbial cells
labeled with 0.8 nm gold particles were successfully detected
via nanoSIMS ion imaging and subsequently analyzed regard-
ing their isotopic composition in pure cultures and anaerobic
sludge samples. This gold-ISH approach avoids a CARD-
based amplification of fluorescent or halogenated signals
and may be best suited where problems arise from cell wall
permeabilization, inactivation of endogenous peroxidases,
or the presence of halogenated minerals that may produce
high backgrounds in sediments and soils. However, the rou-
tine application of gold-ISH for an identification of single
cells via gold targets could be hampered by the fact that the
simultaneous detection of C/N- and Au-counts detectors
may not be possible (40). Furthermore, low rRNA levels of
environmental microbes could limit the detection of single
microbial cells due to weak signal intensities, which is a fre-
quently observed problem of monolabeled FISH approaches
in oligotrophic environments such as soil (6, 14, 58). A
more widely applicable means of combining nanoSIMS and
phylogenetic gold labeling of environmental microbes may
be the use of CARD-based ISH-techniques (e.g., gold-FISH;
39). Problems arising from low rRNA contents would be
overcome by an enzymatic amplification of nanogold par-
ticles. In addition, the increased number of nanogold par-
ticles deposited per microbial cell should allow their direct
visualization in SEM mode of the nanoSIMS instrument,
comparable to routine EM. Thus, the workflow of nanoSIMS
applications could be shortened significantly as the identifi-
cation of regions of interest via epifluorescence microscopy
before isotopic analysis (e.g., 56, 59) would no longer be
necessary.
Further potential for correlative applications of ISH-based
nanogold labeling exists related to the investigation of
microbe–habitat interaction in environmental systems.
Resin-impregnated sections prepared from soils or sediments
can be used to study microorganisms by fluorescence micro-
scopy and provide information on their spatial distribution
patterns related to pore structure and organic matter (60,
61), and furthermore allow for the formulation of hypotheses
on microbe–microbe interactions within the soil environ-
ment (62). As these studies rely on rather general DNA
stains, resin impregnation of soil samples has been combined
with FISH to specifically detect microbial phyla of interest
(24). This approach could be extended by ISH-based
nanogold labeling, allowing for correlative analyses of
microbial distribution patterns and morphological and chem-
ical characteristics of their respective microenvironments.
Furthermore, the abundance and spatial distribution of spe-
cifically labeled microbial cells could be automatically
mapped over larger areas by EDS as proven for marine samples
(Fig. 4).
The visualization of individual microorganisms and their
microhabitats in undisturbed environmental samples with
an intact structure has been limited to two dimensions.
Three-dimensional features of microbial habitats can be
reconstructed from combinatorial analyses with X-ray µCT,
a noninvasive imaging technique that provides spatial infor-
mation on a micrometer scale, and thus allows for a character-
ization of microbial habitats in terms of pore connectivity and
resource flow (63, 64). The direct observation of prokaryotic
cells within undisturbed environmental samples, however, is
not yet possible due to the limited spatial resolution of cur-
rently available systems (22). Nevertheless, with (a) the rapid
development of X-ray instruments (including feature recogni-
tion already down to 50 nm; 28), and (b) the possibility
of enhancing individual nanogold particles to sizes of 30–
100 nm (39, 46), it is not too far-fetched to imagine the
recognition of gold-labeled microbes within the three-
dimensional structure of soils and sediments.
As presented here, the nanogold-based detection of single
microbial cells in environmental samples represents a slowly
emerging field of research where methodological develop-
ments are still required. The potentials, however, are out-
standing regarding the phylogenetic identification and
analysis of single cells on submicroscopic scales, which could
fundamentally improve our understanding of interactions
between individual microorganisms and their associated
microenvironments in almost every environmental system
of interest.
We thank Rudi Amann, Christin Bennke, and Sten Littmann from the
MPI for Marine Microbiology Bremen for collaborative application of
Gold-FISH on marine diatoms as well as for excellent technical assistance
with SEM-EDS analyses.
REFERENCES
1. Gall JG, Pardue ML. 1969. Formation and detection of
RNA-DNA hybrid molecules in cytological preparations.
Proc Natl Acad Sci U.S.A. 63:378–383.
2. Bauman JG, Wiegant J, Borst P, van Duijn P. 1980. A new
method for fluorescence microscopical localization of spe-
cific DNA sequences by in situ hybridization of fluorochro-
melabelled RNA. Exp Cell Res 128:485–490.
3. Levsky JM, Singer RH. 2003. Fluorescence in situ hybridiza-
tion: past, present and future. J Cell Sci 116:2833–2838.
4. DeLong E, Wickham G, Pace N. 1989. Phylogenetic stains:
ribosomal RNA-based probes for the identification of single
cells. Science 243:1360–1363.
5. Amann RI, Krumholz L, Stahl DA. 1990. Fluorescent-
oligonucleotide probing of whole cells for determinative,
phylogenetic, and environmental studies in microbiology.
J Bacteriol 172:762–770.
6. Amann RI, Ludwig W, Schleifer KH. 1995. Phylogenetic
identification and in situ detection of individual microbial
cells without cultivation. Microbiol Rev 59:143–169.
7. Hahn D, Amann RI, Ludwig W, Akkermans ADL, Schlei-
fer K-H. 1992. Detection of micro-organisms in soil after
in situ hybridization with rRNA-targeted, fluorescently
labelled oligonucleotides. J Gen Microbiol 138:879–887.
8. Assmus B, Hutzler P, Kirchhof G, Amann R, Lawrence
JR, Hartmann A. 1995. In situ localization of Azospirillum
brasilense in the rhizosphere of wheat with fluorescently
labeled, rRNA-targeted oligonucleotide probes and
 2.2.1. Gold-Based In Situ Hybridization for Phylogenetic Single-Cell ▪ 2.2.1-9
scanning confocal laser microscopy. Appl Environ Microbiol
61:1013–1019.
9. Pernthaler A, Pernthaler J, Amann R. 2002. Fluorescence
in situ hybridization and catalyzed reporter deposition for the
identification of marine bacteria. Appl Environ Microbiol 68:
3094–3101.
10. Amann R, Fuchs BM. 2008. Single-cell identification in
microbial communities by improved fluorescence in situ
hybridization techniques. Nat Rev Micro 6:339–348.
11. Ishii K, Mußmann M, MacGregor B, Amann R. 2004. An
improved fluorescence in situ hybridization protocol for the
identification of bacteria and archaea in marine sediments.
FEMS Microbiol Ecol 50:203–212.
12. Woebken D, Fuchs BM, Kuypers MM, Amann R.
2007. Potential interactions of particle-associated anam-
mox bacteria with bacterial and archaeal partners in the
Namibian upwelling system. Appl Environ Microbiol 73:
4648–4657.
13. Schmidt H, Eickhorst T. 2013. Spatio-temporal variability
of microbial abundance and community structure in the
puddled layer of a paddy soil cultivated with wetland rice
(Oryza sativa L.). Appl Soil Ecol 72:93–102.
14. Wagner M, Horn M, Daims H. 2003. Fluorescence in situ
hybridisation for the identification and characterisation of
prokaryotes. Curr Opin Microbiol 6:302–309.
15. Ouverney CC, Fuhrman JA. 1999. Combined micro-
autoradiography-16S rRNA probe technique for determina-
tion of radioisotope uptake by specific microbial cell types in
situ. Appl Environ Microbiol 65:1746–1752.
16. Hoshino T, Schramm A. 2010. Detection of denitrifica-
tion genes by in situ rolling circle amplification-fluorescence
in situ hybridization to link metabolic potential with identity
inside bacterial cells. Environ Microbiol 12:2508–2517.
17. Moraru C, Lam P, Fuchs BM, Kuypers MMM, Amann R.
2010. GeneFISH—an in situ technique for linking gene pres-
ence and cell identity in environmental microorganisms.
Environ Microbiol 12:3057–3073.
18. Orphan VJ, House CH, Hinrichs K-U, McKeegan KD,
DeLong EF. 2001. Methane-consuming archaea revealed
by directly coupled isotopic and phylogenetic analysis.
Science 293:484–487.
19. Musat N, Halm H, Winterholler B, Hoppe P, Peduzzi S,
Hillion F, Horreard F, Amann R, Jørgensen BB, Kuypers
MMM. 2008. A single-cell view on the ecophysiology
of anaerobic phototrophic bacteria. Proc Natl Acad Sci
U.S.A. 105:17861–17866.
20. Moraru C, Amann R. 2012. Crystal ball: fluorescence in situ
hybridization in the age of super-resolution microscopy. Syst
Appl Microbio 35:549–552.
21. Chang W-S, Halverson LJ. 2003. Reduced water avail-
ability influences the dynamics, development, and ultra-
structural properties of pseudomonas putida biofilms.
J Bacteriol 185:6199–6204.
22. Hallett PD, Karim KH, Bengough AG, Otten W.
2013. Biophysics of the Vadose Zone: from reality to model
systems and back again. Vadose Zone J 12. doi: 10.2136/
vzj2013.05.0090
23. Rennert T, Totsche K, Heister K, Kersten M, Thieme J.
2012. Advanced spectroscopic, microscopic, and tomo-
graphic characterization techniques to study biogeochemical
interfaces in soil. J Soils Sed 12:3–23.
24. Eickhorst T, Tippkötter R. 2008a. Detection of microor-
ganisms in undisturbed soil by combining fluorescence in
situ hybridization (FISH) and micropedological methods.
Soil Biol Biochem 40:1284–1293.
25. Hapca SM, Wang ZX, Otten W, Wilson C, Baveye PC.
2011. Automated statistical method to align 2D chemical
maps with 3D X-ray computed micro-tomographic images
of soils. Geoderma 164:146–154.
26. Lehmann J, Solomon D, Kinyangi J, Dathe L, Wirick S,
Jacobsen C. 2008. Spatial complexity of soil organic matter
forms at nanometre scales. Nat Geosci 1:238–242.
27. Mueller CW, Weber PK, Kilburn MR, Hoeschen C,
Kleber M, Pett-Ridge J. 2013. Advances in the analysis
of biogeochemical interfaces: NanoSIMS to investigate
soil microenvironments, p 1–46. In Donald LS (ed),
Advances in Agronomy. Elsevier Academic Press, San
Diego, CA.
28. Helliwell JR, Sturrock CJ, Grayling KM, Tracy SR, Flavel
RJ, Young IM, Whalley WR, Mooney SJ. 2013. Applica-
tions of X-ray computed tomography for examining biophys-
ical interactions and structural development in soil systems: a
review. Europ J Soil Sci 64:279–297.
29. Spring S, Lins U, Amann R, Schleifer KH, Ferreira LCS,
Esquivel DMS, Farina M. 1998. Phylogenetic affiliation
and ultrastructure of uncultured magnetic bacteria
with unusually large magnetosomes. Arch Microbiol 169:
136–147.
30. Gérard E, Guyot F, Philippot P, López-García P. 2005.
Fluorescence in situ hybridisation coupled to ultra small
immunogold detection to identify prokaryotic cells using
transmission and scanning electron microscopy. J Microbiol
Meth 63:20–28.
31. Ménez B, Rommevaux-Jestin C, Salomé M, Wang Y, Phil-
ippot P, Bonneville A, Gérard E. 2007. Detection and phy-
logenetic identification of labeled prokaryotic cells on
mineral surfaces using scanning X-ray microscopy. Chem
Geol 240:182–192.
32. Hainfeld JF, Powell RD, Furuya FR. 2002. Microscopic
uses of Nanogold, p 85–106. In Hacker GW, Gu J (eds),
Gold and Silver Staining: Techniques in Molecular Morphology.
CRC Press, Boca Raton, FL.
33. Powell RD, Halsey CMR, Spector DL, Kaurin SL,
McCann J, Hainfeld JF. 1997. A covalent fluorescent-gold
immunoprobe: simultaneous detection of a pre-mRNA
splicing factor by light and electron microscopy. J Histochem
Cytochem 45:947–956.
34. Reinhold B, Hurek T. 1988. Localization of diazotrophs in
the root interior with special attention to the kallar grass
association. Plant Soil 110:259–268.
35. Levanony H, Bashan Y, Romano B, Klein E. 1989. Ultra-
strucural localization and identification of Azospirillum brasi-
lense Cd on and within wheat root by immuno-gold labeling.
Plant Soil 117:207–218.
36. Rogers J, Keevil CW. 1992. Immunogold and fluorescein
immunolabelling of Legionella pneumophila within an aquatic
biofilm visualized by using episcopic differential inter-
ference contrast microscopy. Appl Environ Microbiol 58:
2326–2330.
37. Kenzaka T, Ishidoshiro A, Yamaguchi N, Tani K, Nasu M.
2005. rRNA sequence-based scanning electron micro-
scopic detection of bacteria. Appl Environ Microbiol 71:
5523–5531.
38. Jang A, Tatavarty R, Kim IS. 2012. Bioconjugation of gold
nanoparticles with DNA for in situ hybridization. Desalin
Water Treatm 46:38–45.
39. Schmidt H, Eickhorst T, Mußmann M. 2012. Gold-FISH: a
new approach for the in situ detection of single microbial
cells combining fluorescence and scanning electron micro-
scopy. Syst Appl Microbio 35:518–525.
40. Kubota K, Morono Y, Ito M, Terada T, Itezono S, Harada
H, Inagaki F. 2014. Gold-ISH: A nano-size gold particle-
based phylogenetic identification compatible with Nano-
SIMS. Syst Appl Microbiol 37:261–266. doi: 10.1016/j.
syapm.2014.02.003.
41. Almstrand R, Drennan DM, Sharp JO. 2015. Polygold-
FISH for signal amplification of metallo-labeled microbial
cells. J Basic Microbiol 54:1–5.
2.2.1-10 ▪ MICROSCOPIC METHODS
42. Humbel BM, Sibon OCM, Stierhof YD, Schwarz H.
1995. Ultrasmall gold particles and silver enhancement
as a detection system in immunolabeling and in situ
hybridization experiments. J Histochem Cytochem 43:735–
737.
43. Danscher G, Hacker GW, Stoltenberg M. 2002. Autome-
tallographic tracing of gold, silver, bismuth, mercury, and
zinc, p 13–27. In Hacker GW, Gu J (eds), Gold and Silver
Staining: Techniques in Molecular Morphology. CRC Press,
Boca Raton, FL.
44. Hainfeld JF, Powell RD. 2002. Silver- and gold-based auto-
metallography of Nanogold, p 29–46. In Hacker GW, Gu J
(eds), Gold and Silver Staining: Techniques in Molecular Mor-
phology. CRC Press, Boca Raton, FL.
45. Ehrhardt CJ, Haymon RM, Sievert SM, Holden PA. 2009.
An improved method for nanogold in situ hybridiza-
tion visualized with environmental scanning electron micro-
scopy. J Microsc 236:5–10.
46. Weipoltshammer K, Schöfer C, Almeder M, Wachtler F.
2000. Signal enhancement at the electron microscopic level
using nanogold and gold-based autometallography. Histo-
chem Cell Biol 114:489–495.
47. Schmidt H, Eickhorst T, Tippkotter R. 2012. Evaluation of
tyramide solutions for an improved detection and enumera-
tion of single microbial cells in soil by CARD-FISH. J Micro-
biol Meth 91:399–405.
48. Bennke CM, Neu TR, Fuchs BM, Amann R. 2013.
Mapping glycoconjugate-mediated interactions of marine
Bacteroidetes with diatoms. Syst Appl Microbio 36:417–425.
49. Jahn KA, Barton DA, Kobayashi K, Ratinac KR, Overall
RL, Braet F. 2012. Correlative microscopy: Providing new
understanding in the biomedical and plant sciences. Micron
43:565–582.
50. Herrmann AM, Clode PL, Fletcher IR, Nunan N, Stock-
dale EA, O’Donnell AG, Murphy DV. 2007. A novel
method for the study of the biophysical interface in soils
using nano-scale secondary ion mass spectrometry. Rapid
Commun Mass Spectrom 21:29–34.
51. Clode PL, Kilburn MR, Jones DL, Stockdale EA,
Cliff JBIII, Herrmann AM, Murphy DV. 2009. In situ
mapping of nutrient uptake in the rhizosphere using nano-
scale secondary ion mass spectrometry. Plant Physiol 151:
1751–1757.
52. Nuccio EE, Hodge A, Pett-Ridge J, Herman DJ, Weber
PK, Firestone MK. 2013. An arbuscular mycorrhizal fungus
significantly modifies the soil bacterial community and nitro-
gen cycling during litter decomposition. Enivron Microbiol
15:1870–1881.
53. Behrens S, Lösekann T, Pett-Ridge J, Weber PK, Ng WO,
Stevenson BS, Hutcheon ID, Relman DA, Spormann AM.
2008. Linking microbial phylogeny to metabolic activity at
the single-cell level by using enhanced element labeling-
catalyzed reporter deposition fluorescence in situ hybridiza-
tion (EL-FISH) and Nano- SIMS. Appl Environ Microbiol
74:3143–3150.
54. Li T, Wu TD, Mazeas L, Toffin L, Guerquin-Kern JL,
Leblon G, Bouchez T. 2008. Simultaneous analysis of
microbial identity and function using NanoSIMS. Environ
Microbiol 10:580–588.
55. Thompson AW, Foster RA, Krupke A, Carter BJ, Musat
N, Vaulot D, Kuypers MMM, Zehr JP. 2012. Unicellular
cyanobacterium symbiotic with a single-celled eukaryotic
alga. Science 337:1546–1550.
56. Woebken D, Burow LC, Prufert-Bebout L, Bebout BM,
Hoehler TM, Pett-Ridge J, Singer SW, Spormann AM,
Weber PK. 2012. Identification of a novel cyanobacterial
group as active diazotrophs in a coastal microbial mat using
NanoSIMS analysis. ISME J 6:1427–1439.
57. Dekas AE, Poretsky RS, Orphan VJ. 2009. Deep-sea arch-
aea fix and share nitrogen in methane-consuming microbial
consortia. Science 326:422–426.
58. Eickhorst T, Tippkötter R. 2008b. Improved detection of
soil microorganisms using fluorescence in situ hybridization
(FISH) and catalyzed reporter deposition (CARD-FISH).
Soil Biol Biochem 40:1883–1891.
59. Foster RA, Kuypers MMM, Vagner T, Paerl RW, Musat N,
Zehr JP. 2011. Nitrogen fixation and transfer in open ocean
diatom–cyanobacterial symbioses. ISME J 5:1484–1493.
60. Nunan N, Wu KJ, Young IM, Crawford JW, Ritz K. 2003.
Spatial distribution of bacterial communities and their rela-
tionships with the micro-architecture of soil. FEMS Microl
Ecol 44:203–215.
61. Ritz K. 2011. Views of the underworld: in situ visualization of
soil biota, p 1–13. In Ritz K, Young I (eds), The Architecture
and Biology of Soils: Life in Inner Space. CABI Publishing,
Wallingford, UK.
62. Raynaud X, Nunan N. 2014. Spatial ecology of bacteria at
the microscale in Soil. PLoS One 9:e87217.
63. Young IM, Crawford JW, Nunan N, Otten W, Spiers A.
2008. Microbial distribution in soils: physics and scaling, p
81–121. In Sparks DL (ed), Advances in Agronomy. Elsevier
Academic Press, San Diego, CA.
64. Mooney SJ, Pridmore TP, Helliwell J, Bennett MJ. 2012.
Developing X-ray computed tomography to non-invasively
image 3-D root systems architecture in soil. Plant Soil 352:
1–22.
 Assessment of Prokaryotic Biological Activity at the
Single-Cell Level by Combining Microautoradiography
and Fluorescence in situ Hybridization
CLEBER C. OUVERNEY
2.2.2
ASSESSING MICROBIAL METABOLISM IN
COMPLEX MICROBIAL COMMUNITIES
Prokaryotes are highly genetically diverse, and their meta-
bolic activities have significant impacts in most biological
and abiotic systems. They make important ecological contri-
butions to the remineralization of nutrients, the recycling of
organic matter, and as symbionts in host organisms. Most
microbes, however, lack in unique morphological features,
exist in complex consortia composed of hundreds if not thou-
sands of other organisms, and remain uncultured. It has been
estimated that over 99% of environmental (1) and nearly 70–
80% of human-associated (2, 3) microbial consortia are com-
posed of uncultured microbes.
Even though many natural microbial communities are
abundant in number of cells and are composed of hundreds
to thousands of different microbes, at the genus or species level
most of those phylogenetic groups exist in fairly low abun-
dance. That is the case for soil, rhizosphere, the gut of mam-
mals and insects, eutrophic environments, activated sludge,
and coastal water bodies, which are some of the ecosystems
populated by the highest abundance and diversity of microbes.
Hence, mapping the ecophysiological contributions that spe-
cific types of microbes make to their communities constitute a
major hurdle to the field of microbial ecology today.
Historically, the first reports measuring in situ metabolic
activity of single taxonomic groups from natural habitats
relied on unique morphological features (4–8) or cultivation
(9) for species identification. Large-scale measurements of
prokaryotic role in biogeochemical cycles lumped all prokar-
yotes together, treating them as a single functional group, also
known as the microbial black box (10–15). This chapter
reviews the application of techniques that offer an approach
to measuring single-cell metabolic activity in situ without
the need of cultivation by combining autoradiography with
fluorescence in situ hybridization (FISH).
MEASURING SINGLE-CELL METABOLIC
ACTIVITY IN SITU WITHOUT CULTIVATION
Three main methods have been proposed to quantify single
cells based on their phylogenetic identification and their
ability to take up dissolved nutrients in a mixed natural micro-
bial community without the need of cultivation. All methods
have an overall similar concept and combine FISH with
microautoradiography (MAR or MICRO) to identify the
doi:10.1128/9781555818821.ch2.2.2
2.2.2-1
prokaryotic cell based on probe labeling and determine
whether that cell is capable of taking up and incorporating
the nutrient based on autoradiography. Those methods were
named MAR-FISH (16), MICRO-FISH (17) and substrate-
tracking autoradiographic FISH (STARFISH) (18). This
chapter focuses on the latter.
Other insightful cultivation-independent methods that
link function to taxonomic microbial groups, either without
the correlation to single-cells in mixed complex communities
(19–24) or not applying MAR (25, 26), are not discussed in
this chapter.
IMPORTANT CONSIDERATIONS FOR
STARFISH-LIKE METHODS
The overall end goal of STARFISH-like methods is to expose
live mixed microbial communities to a radioactively labeled
substrate (organic or inorganic) and later determine whether
specific cell types were capable of assimilating the substrate
into body mass. The method is quantitative at the single-
cell level; cells are labeled with up to three markers: a radio-
active substrate, a fluorescent probe, and a general DNA dye.
The protocol can be divided into these eight main steps in
chronological order (Fig. 1): (1) sample collection and prep-
aration; (2) exposure of cells to radioactively labeled substrate
(s); (3) performance of a nutrient uptake curve to determine
the maximum nutrient uptake; (4) concentration of cells by
filtration or centrifugation and transfer of cells onto a cover
glass; (5) whole-cell FISH; (6) coating of cover glass with
liquid nuclear emulsion; (7) emulsion exposure and develop-
ment; and (8) microscopic counts. It has also been demon-
strated that FISH can be done after MAR is completed
(18). For samples with weak FISH signal (slow-growing bac-
teria or microorganisms from low-nutrient habitats such as
oligotrophic waters) one can substitute FISH by catalyzed
reporter deposition–FISH (CARD-FISH) in combination
with MAR (27) since CARD-FISH enhances the probe flu-
orescence signal.
Because the three laboratories in which these methods
were developed focused their research in aquatic systems,
the protocols were originally optimized for marine and fresh-
water samples. Solid samples (soil, particulates, sludge solid
phase, feces, and oral plaque) have to be suspended in an
aqueous solution such as phosphate-buffered saline (PBS)
with the potential for a disruption of the original microbial
2.2.2-2 ▪ MICROSCOPIC METHODS

FIGURE 1 Protocol overview. Major steps involved in labeling cells with FISH and MAR for interrogation about the capability of cells
to be metabolic active when exposed to a generic (mixture of amino acids) or specific substrate in a natural mixed microbial community.
doi:10.1128/9781555818821.ch2.2.2.f1

community structure. If preservation of the original physical
structure is important, other protocols are available to pretreat
samples derived from a biofilm (28–30), suspended aquatic
aggregates (31, 32), or from soil (23, 33–37) prior to starting
the STARFISH-like protocol. Detailed protocols can be
found elsewhere (38), and it is worth mentioning some key
aspects of this technique (Fig. 1) to better understand its
applications and its pros and cons.
Sample Collection and Preparation
Since statistical analyses benefit from a larger cell count ( pop-
ulation number, n), the microscopic observations should
have 50–1,000 cells per microscopic field of view at
1,000 × total magnification. Hence, it is important to
determine the amount of sample to collect based on the con-
centration of prokaryotic cells (cells/ml). Typically, sample
size ranges from tens of milliliters for oligotrophic waters
(low cell concentration) such as the deep and open ocean
sites (39), to a couple of milliliters for coastal sea water
(18), to microliters for the human subgingival crevices (40,
41). For solid samples a few grams are normally used for soil
(23), concentrated activated sludge (42), human feces (43),
and human gut tissue (44). If needed, higher cell concentra-
tions can be achieved by centrifugation or filtration. The
ideal cell distribution is normally determined via epifluores-
cence microscopy, also known as total cell count (45–47)
that uses a DNA dye such as 40 ,6-diamidino-2-phenylindole
(DAPI) or SYBR green. Regardless of its source, samples
 2.2.2. Assessment of Prokaryotic Biological Activity at the Single-Cell Level ▪ 2.2.2-3
must be prepared immediately after collection and kept as
close to ambient conditions (temperature, light, turbulence)
as possible until cells are fixed. A fraction of the sample is
fixed prior to introduction of the radioactive substrate and
serves as a negative control (Fig. 1, step 1).
Incubation with a Radioactive Substrate
The length of time microbes are exposed to the radioactive
substrate changes based on the microbial growth rate. Rapidly
growing microbes (shorter doubling times) take up more of
the substrate as compared to slow growing microbes under
the same amount of time. A substrate uptake curve can be
generated to determine the optimum time to incubate the
sample with the radioactive nutrient (Fig. 1, steps 2 and 3).
Stopping the incubation too soon can lead to false negatives
by preventing the slowest-growing microorganisms capable of
taking up the introduced radioactive substrate from being
labeled (10). On the other hand, carrying out the incubation
too long can potentially lead to false positives, by allowing fast
growing cells to release secondary radioactive metabolites
( portions of the radioactively labeled substrate) into the envi-
ronment, and consequently lead to non-specific labeling. If
the growth rate of the microbe of interest is known, allow
enough incubation for at least one doubling time to have
most of those cells labeled with the substrate. To end the
nutrient uptake, samples are immediately fix in 10% formalin
for 30 min at ambient temperature (18) or in 4% paraformal-
dehyde for 2 h at 4°C (48).
Selection of the Radioactive Nuclide
Selection of the radioactive isotopes for in situ studies has
been extensively described elsewhere (49, 50). In general,
since naturally occurring microbes tend to measure only a
few micrometers in length (enriched environments excepted)
and exist in clusters (close proximity), 3H (tritium) is the pre-
ferred radioisotope due to its high resolution, low-energy beta
particle, and easy availability. Other isotopes such as 33Pi,
14
C, and 35S have also been used successfully (16, 51) as dis-
cussed later in this chapter. The concentration of the radio-
active substrate added to the sample should reflect that of
the ambient concentration (e.g., amino acids in coastal sam-
ples are in ∼2–10 nM concentration). The experiment aims
at tracing the natural flow of nutrients (the substrate) instead
of introducing disturbances to the system from the addition of
substrate above the natural concentration.
Cell Transfer to PTFE-Coated Cover Glass
After incubation with the radioisotope, cells are transferred to
a polytetrafluoroethylene (PTFE, or Teflon)-coated cover
glass (Thermo Scientific) (Fig. 1, step 4) containing various
etched wells. The cover glass is preferred over a glass slide
because you can visualize the cover glass from either side dur-
ing the microscopic observations under high magnification.
The cover glass is the size of a regular microscopic glass slide,
and the wells can be custom ordered in different diameters
and should preferably be pretreated with a cell adhesion
enhancer such as ADCell (Thermo Fisher). Others have
manually coated glass with other substances to improve cell
adhesion, including poly-l-lysine or gelatin (0.1% gelatin,
0.01% CrKSO4) (52). Once attached to the glass, cells are
dehydrated with an ethanol and formalin mixture (53).
FISH
The literature is abundant in FISH protocols (54), including
those reviewed by Eickhorst and Schmidt in this book. Some
crucial steps, however, have been modified from standard
FISH protocols that maximize fluorescence detection levels
while minimizing cell loss (52, 55, 56). Furthermore, detec-
tion of some microbes (especially Gram-positive bacteria
and slow-growing prokaryotes) labeled with FISH may require
additional sample treatment. For instance, chloramphenicol,
a protein inhibitor that prevents RNA degradation, can
increase FISH cell detection by as much as 25% of the total
DAPI-labeled cell count in coastal sea water (52) and in com-
bination with autoradiography in activated sludge (42). Cell
counts with FISH also show an increase after cell permeabili-
zation with either lysozyme (57, 58) or proteinase K (57) or
through acid hydrolysis (59). If samples can be processed
immediately on collection, fixation with 50% methanol dur-
ing FISH can enhance the fluorescence signal without loss of
cell morphology (42). Furthermore, combination of multiple
fluorescent specific probes targeting the same phylogenetic
group (60) and the addition of nonlabeled competitor oligo-
nucleotide probes in the hybridization buffer (61) may
improve cell detection without compromising probe specific-
ity. An extensive list of probes and respective hybridization
conditions are already available in public databases such as
ProbeBase (62). If new probes are designed, they must be first
validated for their specificity prior to usage (63–65). Follow-
ing probe hybridization, all cells are stained with a general
DNA dye with an emission spectrum that does not interfere
with the probe. Commonly used dyes are YO-PRO-1 (491
nm excitation maximum, 509 nm emission maximum) or
DAPI (358 nm Exmax, 461 nm Emmax). Last, the slide is
rinsed in 0.2-µm-filtered MilliQ water to diminish back-
ground fluorescence (16–18).
Photographic Emulsion
Radioactivity from the labeled substrate is recorded on a
liquid photographic nuclear emulsion. These light-sensitive
photographic emulsions are viscous and must be laid on the
cells in a thin layer (1–2 µm thick), which leads to brighter
fluorescent signals (less material blocking the probe and
DNA dye fluorescence) while minimizing secondary radioac-
tive background. Emulsions are manufactured with different
resolution and sensitivity levels as extensively reviewed by
Rogers (49). The cells are coated with the emulsion by care-
fully dipping the cover glass into the melted emulsion solu-
tion, covering all the wells and storing the cover glass in a
lightproof box at 4°C in the dark for ∼ 3 days for emulsion
exposure. In Lee et al. (16), emulsion LM-1 (Amersham)
was used and exposure time varied from 3 to 7 days at 4°C,
whereas in Cottrell and Kirchman (17) NTB-2 emulsion
was exposed for 2 days at −20°C. The emulsion is developed
following the manufacture recommendations. Then the cover
glass is dried and mounted with a low-fluorescence mounting
solution such as Vectashield (Vector). Data collection is
acquired either by epifluorescence or, better, by confocal laser
scanning microscopy (CLSM).
DATA COLLECTION AND INTERPRETATION
Four main types of quantitative data are generally gathered
from the sample.
From each microscopic field of view, the following main
counts are recorded (Figs 2 and 3): (A) the number of cells
within the phylogenetic group of interest, marked with the
fluorescent probe. Probe-labeled cell count is recorded under
the wavelength for the appropriate fluorochrome attached to
the probe (Figs 2A and 3a). Under a different wavelength, the
total cell count (B) tabulates all cells labeled with the general
DNA dye, either DAPI (UV light) or YO-PRO-1 (blue light)
2.2.2-4 ▪ MICROSCOPIC METHODS

FIGURE 2 Imaging and quantification of group-specific microbial cells using STARFISH. Samples treated for STARFISH on a Teflon-
coated cover glass can be analyzed by various fluorescence signals as well as transmitted light excitation. In this example, cells were tagged
with three labels, (A) Cy3-tagged oligonucleotide probe targeting a group-specific prokaryote, (B) YO-PRO-1 DNA staining dye, and (C)
3
H-substrate. Each label was detected and scored with a different light source. The Cy3-probe-labeled cells (Exmax 550 nm, Emmax 570 nm)
provide records of the total number of cells belonging to the phylogenetic group the probe targets. YO-PRO-1 is a DNA-binding dye (Exmax
491 nm, Emmax 509 nm) that labels most viable cells, providing the total cell count (usually reported as cells/ml). Finally, the transmitted light
shows the results from the radioactive nutrient incorporation into prokaryotic biomass, shown as the black markings on the photographic emul-
sion. By combining the three different microscopic field of views one can determine (i) (A + B) the percentage of a specific phylogenetic group
in the community (counts in field A divided by counts in field B over the same area), in other words, the # Cy3-probe-labeled cells / #
YO-PRO-1-labeled cells. In addition, (ii) (Panels B + C) the percentage of the total cells that take up the radioactive substrate can be quantified
by the # 3H-positive cells / # YO-PRO-1-labeled cells. Finally, (iii) the distribution of uptake within each phylogenetic group ((# probe-labeled
cells + # 3H-positive cells) / # YO-PRO-1-labeled cells). Notice that many of the radioactive marking (black dots) along the TM7 filaments are
outside the cells themselves (Panels B + C). Such phenomenon is due to the radionuclide (beta participle from tritium) ability to travel a few
microns away from the source. doi:10.1128/9781555818821.ch2.2.2.f2

(Figs 2B and 3b). The MAR counts (C) provide the number
of cells labeled with the radioactive nutrients under transmit-
ted light (Figs 2C and 3c). The silver grains on the emulsion
exposed to the radioactive material are visible as dark dots on
and next to the cells. The superimposition of two or all three
of the counts above (A, B, and C) allows the investigator to
determine (a) the percentage (relative abundance) the group-
specific cells represent in the total community (Fig. 2, panels
A and B); (b) the percentage of group-specific cells that are
capable to assimilate the radioactive substrate provided (Fig.
2, panels B and C); and (c) the total number of cells that
assimilate the substrate. In addition, any of these composite
counts can be transformed into total cell count (in cells/
ml) for a more absolute quantification of the total number
of cells in that environment carrying out the metabolic proc-
ess of interest.
Obviously current protocols combining FISH and MAR
are labor-intensive. In addition, it has been noted that some
sample artifacts (large salt crystals, fluorescent dye clumps,
and impurities) can interfere with FISH detection and/or
 2.2.2. Assessment of Prokaryotic Biological Activity at the Single-Cell Level ▪ 2.2.2-5

FIGURE 3 Triple-labeled uncultured TM7 bacteria using STARFISH. The same field of view is shown in Fig. 2 and description of the three
labels (a) Cy3-TM7905-probe targeting the 16S rRNA, (b) YO-PRO-1 DNA dye, and (b) trititated amino acids incorporated into bacterial
biomass are described in detail in Fig. 2 legend. The black dots in Panel C indicate exposed silver grains in the photographic emulsion. Some
silver grains do not seem to correlate to a bacterial cell labeled with either the YO-PRO-1 dye (b) or the Cy3-probe (a), possibly due to physical
stress against the Kodak NTB2 liquid emulsion during the procedure as discussed in the text, which should be avoided. However, it is possible
that the bacterial cells were small and not on the same focal plane as the silver grains since these clusters of cells form biofilms up to several
microns in diameter. Images were captured at 1,000 × total magnification with a Zeiss Axioscope-A 1 epifluorescence microscope equipped
with a Hamamatsu camera model ORCA R2 and AxioVision 4.7 image capturing software with frame averaging. Scale bar = 10 µm.
doi:10.1128/9781555818821.ch2.2.2.f3

autoradiography. To avoid such artifacts, all solutions, with
the exception of the liquid emulsion, are recommended to
be prefiltered through 0.2 µm pore size filters. In addition, buf-
fers and MilliQ water should be sterilized by autoclaving for
long-term storage.
ECOLOGICAL APPLICATIONS AND
SIGNIFICANCE
One powerful advantage of methods that combine FISH with
autoradiography as compared with other methods that have
also inquired about cell metabolism in situ is that the former
uses the same microscopic field of view to correlate the signal
from the metabolite with the same cell(s) carrying out the
metabolism. The combination of FISH and autoradiography
has been applied to samples from several habitats including
coastal surface (17, 18), 200-m deep sea (66), estuarine waters
(17), wastewater biofilms (34), and activated sludge (42, 48).
Examples of substrates used has ranged from single tritiated
(3H) amino acid (67) to a mixture of 3H-amino acids (42),
3
H-glucose (68, 69), 14C-acetate (69), and 14C-bicarbonate
(70, 71), 35S-dimethylsufoniopropionate (DMSP) (51).
Application of the method has produced some surprising
results. For instance, the first uncultured bacterium TM7
(Candidatus saccharibacteria) was detected as metabolically
active in a wastewater sample harboring a TM7 phylotype
99.6% similar to TM7 associated with human skin and oral
cavity (42) (Fig. 3). Members of the SAR11 division, an
abundant marine group found worldwide, were reported as
accountable for about 50% assimilation of dissolved amino
acids and a third of DMSP in Atlantic Ocean surface waters
(72). Algal DMSP, an important precursor of the greenhouse
gas dimethyl sulfide, plays an important role as a source of sul-
fur and carbon to heterotrophic bacteria and microzooplank-
ton herbivores in the marine food web. By tracing the fate of
35
S-labeled DMSP into various fractions of the plankton in
seawater, however, it was discovered that not only heterotro-
phic bacteria but also autotrophic phytoplankton (cyanobac-
teria Prochlorococcus and Synechococcus and diatoms)
assimilate sulfur from DMSP (51).
Other marine microbes belonging to the Cytophaga-
Flavobacter cluster and members of the α, β, and γ Proteobac-
teria were shown to play an important role in dissolved organic
matter (DOM) consumption (17). DOM resources were parti-
tioned among the various groups, without a single group dom-
inating, implying the importance of complex communities in
the degradation of DOM (17). Most members of the marine
Archaea lack cultivation and have been traditionally lumped
with the Bacteria domain in biogeochemical surveys. But
reports that 60% of Archaea are metabolically active in
200-m deep waters of the French Mediterranean Sea (66) raised
questions on previous thoughts that Archaea were restricted
to extreme environments for being outcompeted by Bacteria
in nonextreme environments (73, 74). That same report
by Ouverney and Fuhrman (66) detected Archaea in large
abundance (43% of the total prokaryote count) and suggested
Archaea were capable of heterotrophic metabolism (75).
Application with more specific nutrients allowed for fur-
ther “dissection” of the microbial black box into the ecologi-
cal roles of sulfate-reducing-bacteria and nitrite-oxidizing
bacteria in wastewater treatment plants (32), iron reducers
in activated sludge (76), and Microthrix parvicella in the uptake
2.2.2-6 ▪ MICROSCOPIC METHODS
of 14C-labeled long-chain fatty acids (77). Experiments apply-
ing two different radioisotopes (3H and 14C) simultaneously
were able to distinguish each isotope based on the respective
track marks they created on the emulsion (49). Multiple iso-
tope combinations coupled with FISH could help answer
even more complex questions concerning the flux of nutrients
and partitioning of resources among microbes.
It is true that STARFISH and similar techniques are labor-
intensive, opening space for a demand of high-throughput
or automated alternatives. Nonetheless, such approaches
have laid the groundwork for a better grasp of the enormous
ecophysiological complexity of many microbial commun-
ities. Technological advances in high-throughput single-cell
counting devices such as flow cytometry (78–81), microflui-
dics (82–85), microarrays (86, 87), automated confocal laser
scanning microscopy (88), automated time-lapse fluorescence
microscopy (89), and automated image analysis software (90,
91), among others, may greatly improve the efficiency and
convenience of applying STARFISH and similar methods,
while reducing subjective enumeration.
FUTURE ENDEAVORS
The advancement of genomic DNA sequencing technologies
such as next-generation sequencing has confirmed the magni-
tude of the prokaryotic genetic diversity, while unveiling
many novel genes with either low homology to existing
DNA sequences or with no assigned function. While
genomic sequencing analysis serves as a starting point from
which we can make inferences about the function of a
microbe (92, 93), confirmation of that function still requires
empirical validation. Techniques like STARFISH can be
applied to validate microbial metabolic function made by in
silico prediction from microbial genomes.
More recently, it has been proposed that the radioactive
label attached to the substrate in STARFISH methods should
be substituted by a fluorescent moiety. Such substitution is
in the works, where microbes are exposed to fluorescently
labeled amino acids and newly synthesized proteins are
detected via fluorescence microscopy (94). In that study,
FISH or nanoSIMS was used in combination for identifica-
tion of specific environmental microbes containing the newly
synthesized proteins.
This work was partly funded by NIH grant SC3GM082291 and NSF
MRI0923573.
REFERENCES
1. Schloss PD, Handelsman J. 2004. Status of the microbial
census. Microbiol Mol Biol Rev 68:686–691.
2. Kovatcheva-Datchary P, Zoetendal EG, Venema K, de Vos
WM, Smidt H. 2009. Tools for the tract: understanding the
functionality of the gastrointestinal tract. Ther Adv Gastroen-
terol 2:9–22.
3. Dewhirst FE, Chen T, Izard J, Paster BJ, Tanner AC, Yu
WH, Lakshmanan A, Wade WG. 2010. The human oral
microbiome. J Bacteriol 192:5002–5017.
4. Brock TD, Brock ML. 1966. Autoradiography as a tool in
microbial ecology. Nature 209:734–736.
5. Paerl HW. 1984. Alteration of microbial metabolic activ-
ities in association with detritus. Bull Marine Sciences 35:
393–408.
6. Brock TD. 1987. The study of microorganisms in situ: prog-
ress and problems, p 1–17. In Fletcher M, Gray TRG, Jones
JG (eds.), Ecology of microbial communities. Cambridge Uni-
versity Press, New York.
7. Carman KR. 1990. Radioactive labeling of a natural assem-
blage of marine sedimentary bacteria and microalgae for tro-
phic studies: an autoradiographic study. Microb Ecol 19:
279–290.
8. Andreasen K, Nielsen PH. 1997. Application of microau-
toradiography to the study of substrate uptake by filamentous
microorganisms in activated sludge. Appl Environ Microbiol
63:3662–3668.
9. Holt JG, Krieg NR, Sneath PHA, Staley JT, Williams ST.
1994. Bergey’s Manual of Determinative Bacteriology, 9th ed.
Williams & Wilkins, Baltimore, MD.
10. Fuhrman JA, Azam F. 1982. Thymidine incorporation as a
measure of heterotrophic bacterioplankton production in
marine surfaxe waters: evaluation and field results. Mar
Biol 66:109–120.
11. Azam F, Fenchel T, Field JG, Gray JS, Meyer-Reil LA,
Thingstad F. 1983. The ecological role of water-column
microbes in the sea. Mar Ecol Prog Ser 10:257–263.
12. Simon M. 1985. Specific uptake rates of amino acids by
attached and free-living bacteria in a mesotrophic lake.
Appl Environ Microbiol 49:1254–1259.
13. Fuhrman JA. 1992. Bacterioplankton roles in cycling
of organic matter: the microbial food web, p 361–383.
In Falkowski PG, Woodhead AD (eds.), Primary Productiv-
ity and Biogeochemical Cycles in the Sea. Plenum Press,
New York, NY.
14. Azam F, Smith DC, Steward GF, Hagström Å. 1993.
Bacteria-organic matter coupling and its significance for oce-
anic carbon cycling. Microb Ecol 28:167–179.
15. Karner M, Fuhrman JA. 1997. Determination of active
marine bacterioplankton: a comparison of universal 16 s
rRNA probes, autoradiography, and nucleoid staining.
Appl Environ Microbiol 63:1208–1213.
16. Lee N, Nielsen PH, Andreasen KH, Juretschko S, Nielsen
JL, Schleifer K-H, Wagner M. 1999. Combination of fluo-
rescent in situ hybridization and microautoradiography—a
new tool for structure-function analyses in microbial ecol-
ogy. Appl Environ Microbiol 65:1289–1297.
17. Cottrell MT, Kirchman DL. 2000. Natural assemblages of
marine proteobacteria and members of the Cytophaga-
Flavobacter cluster consuming low- and high-molecular-
weight dissolved organic matter. Appl Environ Microbiol
66:1692–1697.
18. Ouverney CC, Fuhrman JA. 1999. Combined
microautoradiography-16S rRNA probe technique for
determination of radioisotope uptake by specific microbial
cell types in situ. Appl Environ Microbiol 65:1746–1752.
19. Amann R, Kuhl M. 1998. In situ methods for assessment of
microorganisms and their activities. Curr Opin Microbiol 1:
352–358.
20. Gasol JM, Duarte CM. 2000. Comparative analyses in
aquatic microbial ecology: how far do they go? FEMS Micro-
biol Ecol 31:99–106.
21. Radajewski S, Ineson P, Parekh NR, Murrell JC. 2000.
Stable-isotope probing as a tool in microbial ecology. Nature
403:646–649.
22. Dahllof I. 2002. Molecular community analysis of microbial
diversity. Curr Opin Biotechnol 13:213–217.
23. Radajewski S, Webster G, Reay DS, Morris SA, Ineson P,
Nedwell DB, Prosser JI, Murrell JC. 2002. Identification of
active methylotroph populations in an acidic forest soil by
stable-isotope probing. Microbiology 148:2331–2342.
24. Boillot A, Demmer RT, Mallat Z, Sacco RL, Jacobs DR,
Benessiano J, Tedgui A, Rundek T, Papapanou PN,
Desvarieux M. 2015. Periodontal microbiota and phospho-
lipases: the Oral Infections and Vascular Disease Epidemiol-
ogy Study (INVEST). Atherosclerosis 242:418–423.
25. Orphan VJ, House CH, Hinrichs KU, McKeegan KD,
DeLong EF. 2001. Methane-consuming archaea revealed
 2.2.2. Assessment of Prokaryotic Biological Activity at the Single-Cell Level ▪ 2.2.2-7
by directly coupled isotopic and phylogenetic analysis. Sci-
ence 293:484–487.
26. Kantor RS, Wrighton KC, Handley KM, Sharon I, Hug
LA, Castelle CJ, Thomas BC, Banfield JF. 2013. Small
genomes and sparse metabolisms of sediment-associated bac-
teria from four candidate phyla. Mbio 4. e00708–13
27. Teira E, Reinthaler T, Pernthaler A, Pernthaler J,
Herndl GJ. 2004. Combining catalyzed reporter deposition-
fluorescence in situ hybridization and microautoradiography
to detect substrate utilization by bacteria and Archaea in the
deep ocean. Appl Environ Microbiol 70:4411–4414.
28. Poulsen LK, Ballard G, Stahl DA. 1993. Use of rRNA flu-
orescence in situ hybridization for measuring the activity of
single cells in young and established biofilms. Appl Environ
Microbiol 59:1354–1360.
29. Gieseke A, Purkhold U, Wagner M, Amann R, Schramm
A. 2001. Community structure and activity dynamics of
nitrifying bacteria in a phosphate-removing biofilm. Appl
Environ Microbiol 67:1351–1362.
30. Kindaichi T, Ito T, Okabe S. 2004. Ecophysiological inter-
action between nitrifying bacteria and heterotrophic bacte-
ria in autotrophic nitrifying biofilms as determined by
microautoradiography-fluorescence in situ hybridization.
Appl Environ Microbiol 70:1641–1650.
31. DeLong EF, Franks DG, Alldredge AL. 1993. Phylogenetic
diversity of aggregate-attached vs. free-living marine bacte-
rial assemblages. Limnology and Oceanography 38:924–934.
32. Daims H, Nielsen JL, Nielsen PH, Schleifer KH,
Wagner M. 2001. In situ characterization of Nitrospira-like
nitrite-oxidizing bacteria active in wastewater treatment
plants. Appl Environ Microbiol 67:5273–5284.
33. Rogers SW, Moorman TB, Ong SK. 2007. Fluorescent in
situ hybridization and micro-autoradiography applied to eco-
physiology in soil. Soil Sci Soc Am J 71:620–631.
34. Gray ND, Howarth R, Pickup RW, Jones JG, Head IM.
2000. Use of combined microautoradiography and fluores-
cence in situ hybridization to determine carbon metabolism
in mixed natural communities of uncultured bacteria from
the genus Achromatium. Appl Environ Microbiol 66:4518–4522.
35. Amann R, Snaidr J, Wagner M, Ludwig W, Schleifer K-H.
1996. In situ visualization of high genetic diversity in a nat-
ural microbial community. J Bacteriol 178:3496–3500.
36. Zarda B, han D, Chatzinotas A, Schönhuber W, Neef A,
Amann RI, Zeyer J. 1997. Analysis of bacterial community
structure in bulk soil by in situ hybridization. Arch Microbiol
168:185–192.
37. Blumes E, Bischoff M, Reichert J, Moorman T, Turco R.
2002. Surface and subsurface microbial biomass, community
structure and metabolic activity as a function of soil depth
and season. Appl Soil Ecol 20:171–181.
38. Ouverney CC, Fuhrman JA. 2004. Correlating single-cell
count with function in mixed natural microbial commun-
ities through STARFISH, p 1689–1710. In Kowalchuk
GA, de Bruijn FJ, Head IM, Akkermans AD, van Elsas JD
(eds.), Molecular Microbial Ecology Manual, 2nd ed. Kluwer
Academic Publishers, The Netherlands.
39. Teira E, van Aken H, Veth C, Herndl GJ. 2006. Archaeal
uptake of enantiomeric amino acids in the meso- and bath-
ypelagic waters of the North Atlantic. Limnol Oceanog 51:
60–69.
40. Brinig MM, Lepp PW, Ouverney CC, Armitage GC, Rel-
man DA. 2003. Prevalence of bacteria of division TM7 in
human subgingival plaque and their association with disease.
Appl Environ Microbiol 69:1687–1694.
41. Aas JA, Paster BJ, Stokes LN, Olsen I, Dewhirst FE. 2005.
Defining the normal bacterial flora of the oral cavity. J Clin
Microbiol 43:5721–5732.
42. Dinis JM, Barton DE, Ghadiri J, Surendar D, Reddy K,
Velasquez F, Chaffee CL, Lee MC, Gavrilova H, Ozuna
H, Smits SA, Ouverney CC. 2011. In search of an uncul-
tured human-associated TM7 bacterium in the environ-
ment. PLoS One 6:e21280.
43. Franks AH, Harmsen HJM, Raangs GC, Jansen GJ, Schut
F, Welling GW. 1998. Variations of bacterial populations
in human feces measured by fluorescent in situ hybridiza-
tion with group-specific 16S rRNA-target oligonucleotide
probes. Appl Environ Microbiol 64:3336–3345.
44. Zoetendal EG, von Wright A, Vilpponen-Salmela T,
Ben-Amor K, Akkermans AD, de Vos WM. 2002.
Mucosa-associated bacteria in the human gastrointestinal
tract are uniformly distributed along the colon and differ
from the community recovered from feces. Appl Environ
Microbiol 68:3401–3407.
45. Rodina AG. 1972. Methods in Aquatic Microbiology. Univer-
sity Park Press, Baltimore, MD.
46. Sherr B, Sherr E, Giorgio PD. 2001. Enumeration of Total
and Highly Active bBacteria, vol 30. Academic Press, San
Diego, CA.
47. Bloem J. 1995. Fluorescent Staining of Microbes for Total
Direct Counts. Kluwer Academic Publisher, Boston, MA.
48. Nielsen JL, Juretschko S, Wagner M, Nielsen PH. 2002.
Abundance and phylogenetic affiliation of iron reducers in
activated sludge as assessed by fluorescence in situ hybridiza-
tion and microautoradiography. Appl Environ Microbiol 68:
4629–4636.
49. Rogers AW. 1979. Techniques of Autoradiography. Elsevier/
North-Holland Biomedical Press, New York, NY.
50. Hudson AL. 1993. Autoradiography techniques, p 57–77.
In Wharton J, Polak JM (eds.), Receptor Autoradiography:
Principles and Practice. Oxford University Press, New York,
NY.
51. Vila-Costa M, Simo R, Harada H, Gasol JM, Slezak D,
Kiene RP. 2006. Dimethylsulfoniopropionate uptake by
marine phytoplankton. Science 314:652–654.
52. Ouverney CC, Fuhrman JA. 1997. Increase in fluorescence
intensity of 16S rRNA in situ hybridization in natural sam-
ples treated with chloramphenicol. Appl Environ Microbiol
63:2735–2740.
53. Braun-Howland EB, Danielsen SA, Nierzwicki-Bauer
SA. 1992. Development of a rapid method for detecting bac-
terial cells in situ using 16S rRNA-targeted probes. BioTech-
niques 13:928–933.
54. Bouvier T, Del Giorgio PA. 2003. Factors influencing the
detection of bacterial cells using fluorescence in situ hybrid-
ization (FISH): a quantitative review of published reports.
FEMS Microbiol Ecol 44:3–15.
55. Bond PL, Erhart R, Wagner M, Keller J, Blackall LL. 1999.
Identification of some of the major groups of bacteria in effi-
cient and nonefficient biological phosphorus removal acti-
vated sludge systems. Appl Environ Microbiol 65:4077–4084.
56. Hugenholtz P, Tyson GW, Blackall LL. 2002. Design and
evaluation of 16S rRNA-targeted oligonucleotide probes
for fluorescence in situ hybridization. Methods Mol Biol
179:29–42.
57. Wagner M, Schmid M, Juretschko S, Trebesius K-H,
Bubert A, Goebel W, Schleifer K-H. 1998. In situ detection
of a virulence factor mRNA and 16S rRNA in Listeria mono-
cytogenes. FEMS Microbiol Lett 160:159–168.
58. Thurnheer T, Gmur R, Giertsen E, Guggenheim B. 2001.
Automated fluorescent in situ hybridization for the specific
detection and quantification of oral streptococci in dental
plaque. J Microbiol Meth 44:39–47.
59. Macnaughton SJ, O’Donnell AG, Embley TM. 1994. Per-
meabilization of mycolic-acid-containing actinomycetes for
in situ hybridization with fluorescently labelled oligonucleo-
tide probes. Microbiology 140:2859–2865.
60. Lee S-H, Malone C, Kemp PF. 1993. Use of multiple 16S
rRNA-targeted fluorescent probes to increase signal strength
2.2.2-8 ▪ MICROSCOPIC METHODS
and measure cellular RNA from natural planktonic bacteria.
Mar Ecoly Prog Ser 101:193–201.
61. Manz W, Amann R, Ludwig W, Wagner M, Schleifer K-H.
1992. Phylogenetic oligodeoxynucleotide probes for the
major subclasses of proteobacteria: problems and solutions.
System Appl Microbiol 15:593–600.
62. Loy A, Maixner F, Wagner M, Horn M. 2007. probeBase—
an online resource for rRNA-targeted oligonucleotide
probes: new features 2007. Nucleic Acids Res 35:D800–804.
63. Schramm A, Fuchs BM, Nielsen JL, Tonolla M, Stahl DA.
2002. Fluorescence in situ hybridization of 16S rRNA gene
clones (Clone-FISH) for probe validation and screening of
clone libraries. Environ Microbiol 4:713–720.
64. Ouverney CC, Armitage GC, Relman DA. 2004. User of
cloned artificial targets for FISH (catFISH) for the optimiza-
tion of oligonucleotide probe hybridization conditions with
16S rRNA clones for in situ quantification of uncultivated
prokaryotic cells, p 727–742. In Kowalchuk GA, Bruijn
IM, Head IM, Akkermans AD, van Elsas JD (eds.), Molecular
Microbial Ecology Manual, 2nd ed. Kluwer Academic Pub-
lishers, Te Netherlands.
65. Ouverney CC, Armitage GC, Relman DA. 2003. Single-
cell enumeration of an uncultivated TM7 subgroup in the
human subgingival crevice. Appl Environ Microbiol 69:
6294–6298.
66. Ouverney CC, Fuhrman JA. 2000. Marine planktonic
archaea take up amino acids. Appl Environ Microbiol 66:
4829–4833.
67. Cottrell MT, Kirchman DL. 2003. Contribution of major
bacterial groups to bacterial biomass production (thymidine
and leucine incorportation) in the Delaware estuary. Limnol
Oceanogr 48:168–178.
68. Ouverney CC. 1999. Dissecting the marine bacterioplankton
“black box” by type and function through FISH and STARFISH.
Ph.D. thesis University of Southern California, Los Angeles.
69. Nielsen JL, Christensen D, Kloppenborg M, Nielsen PH.
2003. Quantification of cell-specific substrate uptake by
probe-defined bacteria under in situ conditions by microau-
toradiography and fluorescence in situ hybridization. Environ
Microbiol 5:202–211.
70. Okabe S, Kindaichi T, Ito T. 2005. Fate of 14C-labeled
microbial products derived from nitrifying bacteria in auto-
trophic nitrifying biofilms. Appl Environ Microbiol 71:
3987–3994.
71. Hornak K, Jezbera J, Simek K. 2010. Bacterial single-cell
activities along the nutrient availability gradient in a
canyon-shaped reservoir: a seasonal study. Aquat Microb
Ecol 60:215–225.
72. Malmstrom RR, Kiene RP, Cottrell MT, Kirchman DL.
2004. Contribution of SAR11 bacteria to dissolved dime-
thylsulfoniopropionate and amino acid uptake in the North
Atlantic ocean. Appl Environ Microbiol 70:4129–4135.
73. Olsen GJ. 1994. Archaea, Archaea, everywhere. Nature
371:657–658.
74. Stein JL, Simon MI. 1996. Archaeal ubiquity. Proc Natl
Acad Sci USA 93:6228–6230.
75. Furhman JA. 2002. Community structure and function in
prokaryotic marine plankton. Ant Van Leeuwen Int J Gen
Mol Microbiol 81:521–527.
76. Nielsen JL, Nielsen PH. 2002. Quantification of functional
groups in activated sludge by microautoradiography. Water
Sci Technol 46:389–395.
77. Nielsen PH, Roslev P, Dueholm TE, Nielsen JL.
2002. Microthrix parvicella, a specialized lipid consumer in
anaerobic-aerobic activated sludge plants. Water Sci Technol
46:73–80.
78. Monfort P, Baleux B. 1992. Comparison of flow cytometry
and epifluorescence microscopy for counting bacteria in
aquatic ecosystems. Cytometry 13:188–192.
79. Amann RI, Binder BJ, Olson RJ, Chisholm SW, Devereux
R, Stahl DA. 1990. Combination of 16S rRNA-targeted oli-
gonucleotide probes with flow cytometry for analyzing
mixed microbial populations. Appl Environ Microbiol 56:
1919–1925.
80. Winson MK, Davey HM. 2000. Flow cytometric analysis of
microorganisms. Methods 21:231–240.
81. Andreatta S, Wallinger MM, Posch T, Psenner R. 2001.
Detection of subgroups from flow cytometry measurements
of heterotrophic bacterioplankton by image analysis. Cytom-
etry 44:218–225.
82. Marcy Y, Ouverney C, Bik EM, Losekann T, Ivanova N,
Martin HG, Szeto E, Platt D, Hugenholtz P, Relman
DA, Quake SR. 2007. Dissecting biological “dark matter”
with single-cell genetic analysis of rare and uncultivated
TM7 microbes from the human mouth. Proc Natl Acad Sci
USA 104:11889–11894.
83. Gomez-Sjoberg R, Leyrat AA, Pirone DM, Chen CS,
Quake SR. 2007. Versatile, fully automated, microfluidic
cell culture system. Anal Chem 79:8557–8563.
84. Pratt ED, Huang C, Hawkins BG, Gleghorn JP, Kirby BJ.
2011. Rare cell capture in microfluidic devices. Chem Eng Sci
66:1508–1522.
85. Ma L, Kim J, Hatzenpichler R, Karymov MA, Hubert N,
Hanan IM, Chang EB, Ismagilov RF. 2014. Gene-targeted
microfluidic cultivation validated by isolation of a gut bacte-
rium listed in Human Microbiome Project’s most wanted
taxa. Proc Natl Acad Sci USA 111:9768–9773.
86. Stender H, Broomer AJ, Oliveira K, Perry-O’Keefe H,
Hyldig-Nielsen JJ, Sage A, Coull J. 2001. Rapid detection,
identification, and enumeration of Escherichia coli cells in
municipal water by chemiluminescent in situ hybridization.
Appl Environ Microbiol 67:142–147.
87. Loy A, Lehner A, Lee N, Adamczyk J, Meier H, Ernst J,
Schleifer KH, Wagner M. 2002. Oligonucleotide microar-
ray for 16S rRNA gene-based detection of all recognized lin-
eages of sulfate-reducing prokaryotes in the environment.
Appl Environ Microbiol 68:5064–5081.
88. Singleton S, Cahill JG, Watson GK, Allison C, Cummins
D, Thurnheer T, Guggenheim B, Gmur R. 2001. A fully
automated microscope bacterial enumeration system for
studies of oral microbial ecology. J Immunoassay Immuno-
chem 22:253–274.
89. Young JW, Locke JCW, Altinok A, Rosenfeld N, Bacarian
T, Swain PS, Mjolsness E, Elowitz MB. 2012. Measuring
single-cell gene expression dynamics in bacteria using fluo-
rescence time-lapse microscopy. Nat Protoc 7:80–88.
90. Schneider CA, Rasband WS, Eliceiri KW. 2012. NIH
Image to ImageJ: 25 years of image analysis. Nat Meth 9:
671–675.
91. Conrad C, Wunsche A, Tan TH, Bulkescher J, Sieck-
mann F, Verissimo F, Edelstein A, Walter T, Liebel U,
Pepperkok R, Ellenberg J. 2011. Micropilot: automation
of fluorescence microscopy-based imaging for systems biol-
ogy. Nat Meth 8:246–U289.
92. Wrighton KC, Thomas BC, Sharon I, Miller CS, Castelle
CJ, VerBerkmoes NC, Wilkins MJ, Hettich RL, Lipton
MS, Williams KH, Long PE, Banfield JF. 2012. Fermenta-
tion, hydrogen, and sulfur metabolism in multiple unculti-
vated bacterial phyla. Science 337:1661–1665.
93. Albertsen M, Hugenholtz P, Skarshewski A, Nielsen KL,
Tyson GW, Nielsen PH. 2013. Genome sequences of rare,
uncultured bacteria obtained by differential coverage bin-
ning of multiple metagenomes. Nat Biotechnol 31:533–538.
94. Hatzenpichler R, Scheller S, Tavormina PL, Babin BM,
Tirrell DA, Orphan VJ. 2014. In situ visualization of
newly synthesized proteins in environmental microbes using
amino acid tagging and click chemistry. Environ Microbiol
16:2568–2590.
 Antibody-Based Technologies for Environmental
Biodetection
CHERYL L. BAIRD AND SUSAN M. VARNUM
2.3.1
The ability to quantify trace amounts of small or large target
molecules is essential in the analysis of environmental sam-
ples. A typical environmental sample is a complex mixture
of materials; therefore, quantification of specific biological
materials from environmental samples requires reagents
with a high degree of specificity and sensitivity. For over 40
years, immunoassays have been a cornerstone for this applica-
tion with environmental samples. In general, an immunoas-
say has two components: a molecular recognition element
or affinity reagent (i.e., antibody) that forms a complex
with the target, and a detector that translates detector–target
complex formation into a quantifiable signal. In this chapter
we discuss these two components in the context of affi-
nity reagent engineering for specific biological materials
and their application to environmental samples. For presen-
tation purposes, the terms “antigen” and “analyte” are used
interchangeably.
BASIC CONCEPTS
Types of Immunoassays
Two categories of immunoassay designs have emerged over
time: the competitive and the noncompetitive immunoassay
(Fig. 1). The competitive immunoassay employs a labeled
antigen that competes with unlabeled antigen in the sample
for limited antibody binding sites. To quantify results, the
amount of labeled antigen that binds to the limited number
of antibody sites is measured; this binding is inversely propor-
tional to the concentration of unlabeled antigen in the
sample. Alternatively, the measured amount of unbound,
labeled antigen is directly proportional to the concentration
of antigen in the sample. Noncompetitive immunoassays,
also referred to as immunometric assays, employ antibodies
at concentrations that are in excess of the analyte. A “sand-
wich assay,” the most common noncompetitive immunoassay
format, employs two different antibodies that simultaneously
bind a single analyte molecule. One, the capture antibody, is
immobilized on a solid support and used to “capture” the ana-
lyte from the sample. The second antibody binds a different
epitope on the analyte, forming a sandwich. This second anti-
body is also labeled to permit signal generation (or detection).
In practice, antibodies are immobilized onto a plastic surface
(often a 96-well microtiter plate). The antigen-containing
doi:10.1128/9781555818821.ch2.3.1
2.3.1-1
sample is incubated with the immobilized antibody to
allow binding, followed by washing to remove unbound
antigen. The labeled second primary antibody, specific
for another part of the analyte, is incubated, followed by
washing to remove unbound antibodies. The label on the
antibody was traditionally a radioactive isotope, although
now it is far more common to use enzyme-conjugated
antibodies in what are called enzyme-linked immunosorbent
assays (ELISA). When an enzyme is used, the next step in the
assay is to add enzyme substrate. The bound enzyme cleaves
the substrate to form a colored (light-absorbing) product
where the intensity of the color (optical density) is propor-
tional to the amount of bound detection antibody.
Competitive immunoassay formats have been used to
quantify small molecules where there are steric limita-
tions to the binding of multiple antibodies. However,
noncompetitive immunometric assays are typically more
sensitive than competitive immunoassays because the
analyte is reacted with an excess concentration of the rel-
evant antibody. The excess antibody drives the reaction to
capture more antigen, thereby increasing the number
of antibody–antigen complexes and the subsequent signal
from these complexes. Additionally sandwich assays require
two antibody recognition events which improves the
specificity of the assay (1). Competitive immunoassays
also suffer from the fact that a response at low analyte
concentration is difficult to discern in comparison to a
sample without analyte (blank or zero) because the assay
is measuring a relatively small difference between two large
signals. In contrast, immunometric assays measure the dif-
ference between a small background signal in the absence
of analyte versus increasingly higher signal as analyte con-
centration is increased.
Most immunoassays require the separation of the anti-
body–antigen complex from unbound labeled antigen or
antibody before the signal is measured (i.e., a washing step).
This assay format is referred to as a heterogeneous immunoas-
say. In contrast, homogeneous immunoassays do not require a
separation step and hence offer advantages in simplicity and
speed. For this to work, some aspect of the binding step in
homogeneous immunoassays must change the label (or
signal) of the bound antibody–antigen complex allowing
binding events to be detected while free antibody or antigen
remains undetected.
2.3.1-2 ▪ TARGET-SPECIFIC DETECTION
(a)
+ +
(b)
+ +
FIGURE 1 Typical immunoassay formats. (a) competitive immunoassay, (b) noncompetitive or immunometric immunoassay.
doi:10.1128/9781555818821.ch2.3.1.f1
Immunoassay Performance Characteristics
Characteristics that describe the performance of an immu-
noassay include assay analytical sensitivity, specificity,
accuracy, precision, and interference (2). Assay sensitivity
can be defined as the minimal concentration that can be reli-
ably estimated. Analytical sensitivity can be expressed as the
limit of detection (LOD) or the limit of quantification
(LOQ). The LOD is the lowest analyte concentration that
can be detected in a sample but not necessarily quantified,
and is often calculated as two to three times the standard
deviation of the blank. The LOQ is the lowest analyte con-
centration that can be quantified with acceptable precision
and accuracy (3). When low levels of antigen are present in
a sample (i.e., antigen-limiting scenarios), the limit of detec-
tion of an immunoassay is determined primarily by the affinity
of the antibody; higher affinity results in detection that is
more sensitive. Similarly, assay specificity, the ability of the
antibody to detect a single analyte species to the exclusion
of others, is largely due to the antibody specificity. Assays
with low specificity use antibodies that are cross-reactive
with similar analytes. If an assay has low specificity, other
cross-reactive analytes with similar binding epitopes will be
measurable with the assay. Interference is any factor causing
bias (a difference between the measured and the true value)
in an assay, other than the presence of a true cross-reacting
substance. Interfering factors are also referred to as matrix
effects. Several mechanisms can contribute to interference,
including physical masking of the antibody, blocking of the
analyte, or physically altering the antigen structure. Accuracy
is defined by how close the average measured value is to the
true value and is typically measured by an assessment of recov-
ery. This is quantified by adding a predetermined quantity of
purified analyte to samples; the concentration measured
by the assay is then divided by the expected concentration.
If there is a background concentration of the analyte in
the matrix, then recovery can be determined by measuring
increasing concentrations of added analyte and comparing
the measured increase in concentration to the predicted
increase in concentration (2). Precision or the repeatability
of the immunoassay is expressed as the percentage coefficient
of variation (% CV). Precision is determined for both within-
run and interassay tests. Typically, an acceptable CV is less
Analyte
Labeled
analyte
Antibody
Labeled
antibody
than 20%; however, acceptable values vary depending on
the assay and application (4).
Affinity Reagents
Affinity reagents impart the specificity to an immunoassay.
Typically the affinity reagent for an immunoassay is an anti-
body. Vertebrate organisms naturally produce antibodies as
part of their adaptive immune response to counter pathogens
invading the body. An antibody has two roles in the immune
response: bind specifically to foreign pathogen molecules
and, once bound, attract other cells and molecules to destroy
the pathogen. It is the molecular binding properties of
antibodies that make them effective molecular recognition
elements—they can bind a wide range of molecules tightly
(i.e., high affinity) with exceptional specificity.
Antibody Structure
Antibodies are broadly characterized as immunoglobulins
(Igs). Igs consist of four polypeptide chains: two identical
heavy (H) chains of approximately 50 kDa, and two identical
light (L) chains of approximately 25 KDa. Each H chain is
linked to an L chain by a disulfide bond (see Fig. 2). Multiple
disulfide bonds further link the H chains to each other. The
basic structure of an immunoglobulin is a Y-shaped molecule,
with each arm containing an identical target- (antigen-)
binding interface. Igs are subdivided into classes or isotypes
(e.g., IgM, IgD, IgG, IgA, IgY) based on how they are
involved in the immune response (their “effector mecha-
nism”), which is dictated by the structure of the H chain.
These structural differences include variations in the number
of constant domains (C domains), numbers and locations of
disulfide bonds, and glycosylation patterns. IgGs are the
most commonly used isotype for immunoassays because of
their high abundance in the sera of immunized animals and
the generally higher specificity for their target.
Both the H and L chains of the IgG are divided into
structurally similar domains of about 110 amino acids. The
N-terminal domains of both chains (located at the tip of
each arm of the Y) are highly variable in amino acid sequence
and are thus called the variable or V domains (VH and VL, for
the H and L chains, respectively). Together, each VH−VL
pair makes up the target (or antigen) binding site. The
 2.3.1. Antibody-Based Technologies for Environmental Biodetection ▪ 2.3.1-3
Antibody
H H
VH
Fab
CH
region
VL
CL
Fc
region
H VH
L
Fab
FIGURE 2 Antibody structures including antibody molecule, Fab2, Fab, and ScFv molecules. doi:10.1128/9781555818821.ch2.3.1.f2
remaining domains of each chain (one for the L chain and
three for the H chain) are relatively constant in sequence
and are referred to as the constant or C domains.
The V and C domains associate to form larger globular
domains. An IgG consists of three equal-sized globular
domains—the two arms and stalk of the Y—that are joined
together by a flexible hinge region. The two globular domains
that bind the target/antigen are called Fabs, for fragment
antigen binding. The third globular domain does not bind
antigen, but in early studies was found to crystallize easily
and was thus termed the Fc fragment, for fragment crystalliz-
able. A singular Fab or conjoined Fab (F(ab’)2) can be pro-
duced by enzymatic digestion of an IgG. These molecules
can be used directly as an affinity reagent in situations where
the Fc may interfere with the assay or a smaller affinity reagent
is desired.
Both the Fab and Fab2 formats can also be produced using
recombinant techniques, as can a plethora of other formats
consisting of a variety of domain combinations (including
scFvs, diabodies, and minibodies) (reviewed in 5). Most of
these specialized antibody formats have been produced with
therapeutic applications in mind. Importantly, reducing anti-
bodies to smaller functional subunits allows them to be more
easily engineered using molecular approaches, which can
benefit both therapeutic and diagnostic applications. These
smaller entities also have the potential to improve the effi-
ciency and reduce the cost of antibody development and
production. Therefore, these alternative antibody formats
are also of interest for diagnostic applications.
Antibody Generation and Engineering
Generating antibodies for a specific target can be a challeng-
ing endeavor (6, 7). Even if a commercial antibody exists for
the target, variations in quality (sensitivity/affinity), specific-
ity (cross-reactivity with other proteins), and applicability
(i.e., an antibody that works for an ELISA assay may not
work for a Western blot analysis) can lead to time-consuming
screening experiments to find the optimal antibody for
the application. Informational databases such as Anti-
bodypedia (8, 9), and Abminer (10) collate published and
user-submitted information on the performance of research
Fab2
L L
VL
ScFv
antibodies and can help guide the antibody selection process.
Nevertheless, a significant amount of effort is usually required
to find high-quality antibodies for a specific biodetection
application.
Whether choosing a commercial antibody or developing
an antibody from scratch, it is important to consider the
end use of the antibody to determine the most suitable type
of antibody ( polyclonal, hybridoma-derived monoclonal, or
recombinant) that is best for the application. Antibodies
are naturally produced in animals as part of the immune
response when an organism encounters a foreign substance
or antigen. Typically, the antigen contains numerous regions
(epitopes) that are recognized by different antibodies. The
pools of antibodies produced by an organism are called poly-
clonal antibodies (PAbs). They target a variety of epitopes
and are a product of many different cell lines or clones. In
contrast, a monoclonal antibody (MAb) is a unique antibody
that recognizes a single epitope and is produced by a single cell
line. Traditional methods for generating antibodies involve
injecting a target protein into a mammalian host (such as a
mouse or rabbit) and letting the host’s immune system gener-
ate a panel of antibodies that recognize the target. If PAbs are
desired, the serum of the organism is harvested. If MAbs are
the goal, the splenocytes of the organism are harvested and
immortalized by hybridization with a myeloma cell line—
the resulting antibody-producing cell line is called a hybrid-
oma and it can be propagated using in vitro cell culture
methods.
Because PAbs typically recognize a range of epitopes on
the antigen (or target), they are advantageous for applica-
tions in which multiple antibodies binding to the target
are desired. For example, trying to detect a low-abundance
protein with a fluorescent antibody, a PAb could provide
more sensitivity than a MAb because multiple antibodies
could bind to a single protein molecule amplifying the signal.
In contrast, to quantify directly the target, such as in a flow
cytometry experiment, a MAb would be a better choice
because the fluorescence due to antibody binding would be
directly correlated to the amount of target present. PAbs are
also advantageous in applications where the target pro-
tein is heterogeneous, that is, containing small variations in
2.3.1-4 ▪ TARGET-SPECIFIC DETECTION
glycosylation, partial denaturation, or where polymorphisms
exist. Importantly, a critical limitation to PAbs is that they
are a finite resource—once the serum supply is depleted,
the immunization and characterization process must be per-
formed again and there is no guarantee that the new batch
will perform the same as the previous one.
In recent years, there has been a shift to employ recombi-
nant methods to develop antibodies, especially when con-
sidering the need for antibodies on a scale to parallel the
explosion of information being generated by systematic
analysis of genomes, transcriptomes, and proteomes (11,
12). Antibodies are a key tool for studies to help deconvolute
and validate the information about these systems. Recombi-
nant antibody technologies offer a high-throughput path
to generate renewable, highly uniform sources of affinity
reagents that, if required, can be easily manipulated or engi-
neered downstream for optimal performance in a given assay.
For example, producing a recombinant antibody fused with
an epitope tag allows the selective attachment of the protein
on a microarray chip (13).
The two components central to all recombinant antibody-
engineering approaches are (1) a nucleic acid (DNA or
RNA) library encoding antibody variants, and (2) a platform
for screening the library. Antibody libraries are typically
classified according to where the source antibody material
was derived. There are two broad classes of libraries: natural
and synthetic.
Natural libraries are derived from an organism and can be
further classified as naïve, nonimmune, or immune. A truly
naïve library is derived from an organism that has never
been exposed to antigen, more often, however, a naïve library
refers to a collection of antibodies derived from an organism
that was not specifically immunized against an identified anti-
gen. In actuality, “nonimmune” is a better descriptor of these
libraries, but they are often referred to as “naïve.” A non-
immune library is often derived from multiple individuals
and seeks to reflect the immune repertoire of an organism
so a single library can be mined to discover antibodies against
virtually any target. The larger and more diverse these libra-
ries are, the higher the likelihood that they will contain anti-
bodies to a given target (14). These libraries are advantageous
for toxic or poorly immunogenic antigens where immuniza-
tion could harm the host organism. Conversely, immune
libraries are derived from organisms that have been specifi-
cally immunized against an antigen. These libraries are heav-
ily biased toward the given antigen and usually contain high
affinity antibodies if the organism mounted strong immune
response against the antigen.
While natural libraries are created from the genetic mate-
rial of a host organism, synthetic libraries are created from
designed, synthesized DNA. Usually an antibody framework
that has been optimized for high stability or low immunoge-
nicity (if intended for a therapeutic application) is used as a
scaffold on to which chemical diversity is imparted in the
complementarity determining regions (CDRs) of the antigen
binding interface. Synthetic libraries allow precise control of
the library composition (reviewed in 15). Synthetic libraries
can be conceptually simple; functional antibodies have been
derived from synthetic libraries containing a single antibody
framework with only combinations of serine and tyrosine in
their CDRs (16). Semi-synthetic libraries are, as the name
implies, partially based on synthetic sequences. For example,
a natural library can be augmented to have more diversity by
the inclusion of synthetic CDRs (17, 18).
Regardless of the type, antibody libraries typically contain
millions to billions of variants. Therefore, effective screening
methods are needed to find the proverbial needle in the
haystack. Screening platforms are usually referred to as display
technologies because the antibody molecule(s) are displayed
or presented in a manner that allows for analysis of target
binding. A key feature of a display technology is that it also
provides a physical linkage between phenotype (antibody’s
target binding characteristics) and genotype (RNA/DNA
that encodes the antibody). This linkage allows amplifica-
tion of antibody clones after selection and for downstream
production (and/or subcloning) of the selected antibody.
There are a wide variety of in vitro display technologies avail-
able including phage, yeast, ribosome, mRNA, and bacterial
displays (reviewed in 19) Of these technologies, phage and
yeast display are the most widely used.
In filamentous phage display systems, antibodies are
encoded by the phage genome, and the expressed antibodies
are displayed on the phage surface via fusion to phage coat
proteins. To find the antibodies in a library that recognize a
given target, a phage library is “panned” against the target
immobilized on a surface (e.g., in the well of a microtiter
plate). Essentially, the phage-displayed antibody library is
mixed with the target and the phages that bind are retained
on the surface while the unbound phage is subsequently
washed away. The panning process enriches the library for
target-specific clones and several rounds of panning are
usually performed to increase the target specific clones in
the population. Typically, between each panning phase, the
phage are amplified by infecting bacteria with the eluted
phage. Very diverse phage libraries (containing >1010 mem-
bers) can be efficiently panned in a series of days.
In the yeast surface display system, antibodies are encoded
by plasmids transformed into the yeast and displayed as
fusions to a cell surface protein. Unlike a phage display library,
clones from the yeast library are typically selected using fluo-
rescently labeled, soluble target as opposed to an immobilized
target. This allows yeast display libraries to be screened by
fluorescence-activated cell sorting (FACS). With FACS,
the fluorescence intensity of individual cells can be quantita-
tively measured and the brightest cells (those displaying anti-
bodies with the highest affinity for the target) can be isolated
with high precision. The ability to rapidly score individual
yeast clones in a population for binding affinity is a major
advantage of this platform over phage display; however, this
precision comes at a cost, as it is impractical to screen large
libraries by FACS (for example, at a flow rate of 50,000 cells
per minute it would take roughly 2 weeks to screen a library of
109 cells by FACS).
Ferrara and colleagues have exploited the advantages of
both phage and yeast display by using them in combination
to select antibodies to a biomarker for tuberculosis, antigen
85 (20). In this approach, they used phage display initially
to screen a large nonimmune library for antibodies that bound
antigen 85. With two rounds of panning they reduced the
diversity of the library from ∼1011 different antibodies to
approximately 70,000. They then cloned the output antibod-
ies from the phage selection into a yeast display vector and
performed FACS-based selection to identify and characterize
hundreds of antigen 85–binding clones. In a standard phage-
based selection, only 384 individual clones would typically be
individually characterized for binding to the target; however,
by screening the phage output with yeast display they were
essentially able to analyze every clone individually. Accord-
ing to the authors, by combining the phage and yeast display
approaches, they were able to identify a greater variety of
antibodies than were usually found when screening by phage
display alone.
 2.3.1. Antibody-Based Technologies for Environmental Biodetection ▪ 2.3.1-5
In addition to identifying antibodies that bind to a given
target, when in vitro display approaches are coupled with
directed evolution techniques, this allows construction of
antibodies with unique characteristics that are difficult if
not impossible to develop with “natural” methods of antibody
development. For example, antibodies can be engineered
with affinities far superior to what is obtainable by the im-
mune system. A notable example is the affinity maturation of
the antifluorescein antibody, 4–4–20. The affinity of 4–4–20
(Kd = 0.7 ± 0.3 nM) is near the affinity ceiling of a natural
immune response. Through several rounds of mutagenesis
and selection, a 4–4–20 derivative scFv, 4M5.3, was isolated
with a monovalent Kd of 48 fM and an off-rate (detachment
rate) of greater than 5 days (slower than a streptavidin–biotin
complex) (21). Similar approaches can be used to engineer
reagents for applications in defined environments. Siegel
et al. improved the performance of an antibody in a diagnostic
immunoassay that involved samples containing organic sol-
vents (22). Using yeast surface display to screen mutagenic
antibody libraries in the presence of 10% methanol, the
authors isolated an antibody that, when used in the original
immunodiagnostic assay, reduced the limit of detection by
three- to sixfold. Strategies have also been developed for engi-
neering antibodies with increased stability by exposing anti-
bodies to denaturants and elevated temperatures to select
against unstable clones before final antibody selection (23).
Alternatives to Antibodies
Although antibodies are the gold standard affinity reagent,
they have shortcomings, including limited chemical and
physical stability in nonphysiological environments (e.g.,
those often encountered by environmental sensors) and rela-
tively high expense to develop and manufacture. To address
these limitations, synthetic or semisynthetic affinity reagents
based on alternative protein scaffolds (nucleic acids and
nonbiological polymers) are emerging as viable alternatives
to natural or recombinant antibodies.
Scaffolds composed of nonimmunoglobulin proteins have
been widely reported over the past decade in both academia
and industry (24–29). Engineering these scaffolds requires
the same basic approach used for an antibody scaffold, where
diversity is introduced into a region or regions of a protein and
then the desired molecular recognition characteristics are
selected using a molecular display approach. A wide variety
of proteins have been demonstrated as suitable scaffolds
(25), including a completely synthetic, computationally
designed protein, Top7 (30–32).
Aptamers are affinity reagents created from nucleic acid
oligomers (RNA or DNA) that adopt sequence-dependent
three-dimensional structures that enable selective binding
to molecular targets. Aptamers are typically generated in a
process called SELEX (systematic evolution of ligands by
exponential enrichment), where aptamers with the desired
binding characteristics are selectively enriched and amplified
by PCR from a large chemically synthesized combinatorial
library of oligonucleotides (33). The process was originally
described in two seminal publications by the Ellington and
Gold laboratories in 1990 (34, 35). Aptamers make desirable
affinity reagents because they can be directly produced and
modified with chemical reporters by chemical synthesis.
Nevertheless, their utility and application has lagged behind
protein-based affinity reagents (36). The reasons for this are
not widely documented in the peer-reviewed literature, how-
ever, anecdotal evidence suggests that because of the limited
diversity of nucleotides available to construct aptamers
(4 nucleotides compared to 20 amino acids), in general
they do not have the chemical diversity to reliably attain
the affinities and specificities commonly attained by anti-
bodies and alternative scaffold affinity reagents. Recently,
however, the Gold lab and SomaLogic have developed and
demonstrated a method to create aptamers containing
modified nucleotides that have protein-like functional groups
(37). This approach could enable aptamers to compete
broadly with protein-based affinity reagents.
Molecularly imprinted polymers (MIPs) are a completely
synthetic approach to designing robust affinity reagents.
MIPs are created by cross-linking organic polymers in the
presence of a template molecule that “imprints” its shape
into the polymer matrix. When the template molecule is
removed, a cavity is left that is complementary in size and
shape to the template. This technique is relatively routine
for small molecule chemical targets. However, despite signifi-
cant interest in the research community over the last decade,
few MIPs have been developed that recognize proteins
(38, 39).
ASSAY TYPES
ELISA
ELISA represent the most prevalent immunoassay utilized in
clinical and research laboratories. In its simplest form, anti-
gens from a sample are attached to a solid support, such as
the wells of a plastic multiwell plate, and then enzyme-linked
antibody is allowed to bind to the antigen. Antibody binding
is detected by a reaction of the enzyme-linked antibody that
converts a colorless substrate into a colored reaction product
that can be quantitatively measured. A variant of the ELISA
is the sandwich assay in which the analyte/antigen is bound
between two antibodies: a capture antibody and the enzyme-
linked detection antibody. As already discussed, the use of
two antibodies in the sandwich assay increases the analytical
sensitivity and specificity of the assay; however, this type of
ELISA requires the availability of paired antibodies. Because
the ELISA is a simple, rapid, robust, and sensitive assay
technique, it has been widely applied in the analysis of
environmental samples, including the detection of food-
and waterborne pathogens such as Escherichia coli, Salmonella
spp., and Listeria monocytogenes (40–42) and the detection of
cyanobacteria toxins (microcystins) (43, 44), mycotoxins
(45), and bacterial toxins (46). ELISAs and other immuno-
assays are especially useful for detecting toxins where the
organism is not present, which obviates the use of nucleic
acid–based methods. Some toxins, however, such as the
mycotoxins and microcystins, form a diverse group of com-
pounds and therefore require assays employing antibodies
with high specificity (45).
Lateral Flow Assays
Lateral flow assays (LFAs) and lateral flow devices (LFDs),
such as the dipstick assays, are simple devices enabling
fast and easy-to-perform immunoassays that can be either
qualitative ( presence/absence) or quantitative assays when
used with a photometric strip reader. LFDs are composed of
a porous membrane, typically nitrocellulose, which enables
the sample to flow across a surface via capillary action. The
sample is introduced onto a sample pad from which it migrates
through a region of gold nanoparticle–labeled antibodies that
are specific for the target antigen. Sample, with newly bound
antibody/antigen complexes, continues to migrate across the
membrane until reaching a region of antigen-specific anti-
bodies that are immobilized. The migrating antibody/antigen
2.3.1-6 ▪ TARGET-SPECIFIC DETECTION
complex binds the immobilized antibodies and accumulation
of gold-labeled complexes produce a visible line on the
membrane. For the test to be valid, a control line should
form in a second region of the membrane composed of an
immobilized line of antibodies that are specific to only the
gold nanoparticles. LFAs have been successfully used for
fast on-site pathogen and toxin detection (47–51). Overall,
however, LFAs have limited analytic sensitivity, they cannot
be “regenerated” (i.e., disposed after use), and they typically
provide only qualitative data. Quantitative analysis is possi-
ble from LFA data, but it is a time-consuming process and is
usually not suitable for field analysis.
Antibody-Based Planar Microarrays
Antibody-based microarrays are a powerful analytical tool for
the simultaneous detection of multiple analytes in a single
experiment. Antibody microarrays can be divided into two
types: planar microarrays and suspension arrays, also known
as bead-based arrays (discussed in the next section). Planar
microarrays consist of high-density microspots of capture
protein immobilized in an ordered pattern on a surface at spa-
tially discrete locations. Several antibody microarrays formats
are utilized. The predominant format, a microarray sandwich
ELISA, consists of an immobilized capture antibody that is
used to capture the target protein and a labeled second anti-
body that provides the signal for detection. In an alternative
format, the sample containing the target protein is labeled,
thus removing the need for a second antibody for detec-
tion purposes. A third antibody microarray format combines
immobilized cell lysates, or even intact tissue specimens, with
specific antibodies to identify and quantify the presence of
the target analyte in the immobilized sample (52, 53).
The basic principles of planar microarray technology
were originally described by Ekins (54). He theoretically
and empirically demonstrated that the miniaturization of
immunoassays improves lower limits of quantification due
to improved signal-to-noise ratios and decreased reaction
times because of shorter diffusion distances compared to tradi-
tional immunoassays (55). Signal detection from protein
microarrays is usually achieved through fluorescence detec-
tion using either charge coupled device (CCD) cameras or
microarray laser scanners.
Challenges and issues relevant to the production of pro-
tein microarrays (56) include determining the necessary
time for sample incubation (57, 58), washing and mixing
(59), eliminating cross-reactivity (60), and antibody immobi-
lization (13). Microarray surfaces commonly use planar glass,
plastic microwells, nitrocellulose, and three-dimensional
gels (52). Various strategies have been employed for antibody
immobilization on microarray surfaces. It is important that
the surface characteristics permit sufficient antibody move-
ment to retain function while optimizing antibody orienta-
tion on the array surface to maximize the formation of the
antibody/antigen complex. Antibodies can be immobilized
through direct physical adsorption, covalent coupling, or by
specific capture methods (61). Direct physical adsorption is
easy to perform, but antibodies will denature if the array
surface is hydrophobic. Additionally, antibody leaching
from the array surface can occur during the course of the assay,
thereby reducing both assay sensitivity and reproducibility.
These drawbacks can limit the usefulness of physical adsorp-
tion for antibody immobilization, although some studies
have shown that direct physical adsorption works as well or
better than covalent binding, possibly due to negative effects
of covalent interactions with antibody structure or function
(13). Covalent binding is often the preferred method for
antibody immobilization because covalently coupled anti-
bodies are more stable compared to physically adsorbed
antibodies and they are not subject to surface desorption.
Antibodies can be immobilized by forming a covalent bond
between the array surface and functional amino acid groups
on the antibody (e.g., amino, carboxyl, and sulfydryl groups).
In the precapture approach, a capturing molecule is first cova-
lently coupled to the array surface, and this molecule binds
the antibody that is subsequently used to capture the antigen.
Examples of specific precapture molecules include protein
A/G, nickel molecules that bind histidine tags, or streptavidin
molecules that bind biotin tags (62). These approaches re-
quire additional steps, although all of the antibodies are im-
mobilized in the same orientation and this helps standardize
assays (61, 63).
It is worth noting that there is considerable overlap
between the development of planar protein microarrays and
biosensor development. For instance, a planar waveguide
microarray has been developed at the Naval Research Labo-
ratory that employs an array of immobilized capture antibod-
ies (64). Because the signal is transduced by a waveguide
through total internal reflection fluorescence and because
of the flow-through nature of the assay, we have chosen to
discuss this microarray in the Optical Biosensors section of
this chapter.
Environmental monitoring protein microarray applica-
tions include the detection of pathogens and toxins in foods
and water supplies (65–69). Emerging trends in developing
microarray platforms include the integration of microfluidics
(65), fabricated nanostructures (70, 71), and electrochemical
detection systems (65).
Bead-Based Microarray Assays
An alternative to the planar microarray is the suspension
microsphere assay where antibodies are immobilized on
microspheres (beads) instead of spots on a planar surface.
This bead-based technology utilizes the basic sandwich assay
format—the antibody-coated beads are mixed with the sam-
ple to capture the antigen followed by addition of a second
labeled detection antibody that binds to a different epitope
on the antibody. The beads are spectrally unique (color-
coded) so that each set of beads with distinct capture antibod-
ies can be differentiated by flow cytometry, thereby allowing
multiplex detection. For example, the Luminex’s xMAP bead
array can simultaneously distinguish up to 100 analytes in a
sample (72). Bead-based assays have been developed to detect
pathogens and toxins in environmental samples (67, 73, 74).
Biosensors
Biosensors comprise two distinct components: a molecular
recognition element and a transducer element. The molecu-
lar recognition produces a signal such as a change in color, flu-
orescence, or electrical conductivity. The transducer detects
this signal and provides an indication of the magnitude of the
signal. In recent years, many types of biosensors have been
developed and applied to a wide range of analytical problems,
including environmental monitoring (75). Biosensors have
great potential to detect environmental pathogens and toxins
because they enable fast or real-time detection, portability,
and low-cost analysis of samples in the field (76–78).
A range of different biological recognition elements are
incorporated into biosensors for the specific binding of the
analyte, including cell receptors, nucleic acids, aptamers,
enzymes, or antibodies. Because antibodies provide high
affinity binding and specificity (as discussed in the Affinity
Reagents section), they are a widely chosen recognition
 2.3.1. Antibody-Based Technologies for Environmental Biodetection ▪ 2.3.1-7
element in advanced biosensors or immunosensors. Strategies
for antibody immobilization on biosensor surfaces are similar
to those discussed in the planar microarray section. The bio-
sensor transduction methodology (discussed later) can be
roughly categorized into three primary types: optical, electro-
chemical, and mechanical.
Biosensors can be divided into assays with signals that are
detected by label-based or label-free strategies. Label-based
detection requires additional steps and expense to generate
either a labeled antigen or antibody. Labeling, typically
with fluorescent dyes, enzymes, or radioisotopes, is often ran-
dom with respect to molecular position and thus can poten-
tially interfere with the antigen/antibody interaction (79).
Furthermore, label-based techniques do not provide real-time
detection capability. Nevertheless, label-free detection sys-
tems often suffer from decreased sensitivity compared with
label-based systems, which is a major limitation in the imple-
mentation of these assays. Thus, while there is great interest in
developing label-free methods for the detection of biomole-
cules, the majority of immunosensor platforms use a label-
based detection step. Label-free detection methods include
optical (surface plasmon resonance, waveguide-based and
ring resonators), electrochemical, and mechanical (quartz
crystal microbalance and cantilevers) detection methods
(80–82).
The incorporation of nanomaterials into biosensors has
been central to advancements in assay sensitivity and speed.
Nanomaterials possess several unique properties including
small size (1–100 nm), correspondingly large surface-to-
volume ratios, and unusual target binding properties. Because
of their high surface area to volume ratio, nanoparticles
serve as excellent scaffolds to immobilize large quantities of
antibodies, thereby increasing the chances of an interaction
with the target analyte (83). Issues that are important in
the design of biosensors (using nanomaterials) include nano-
material heterogeneity, ease of synthesis and purification,
stability, and reproducibility. Examples of nanomaterials
suitable for use with biosensor platforms include gold nano-
particles (84), carbon nanotubes (85), graphite nanosheets
(86), magnetic nanoparticles (87), and quantum dots (88).
Quantum dots (QD) are semiconductor nanoparticles that
fluoresce when stimulated by an excitation light source.
The small subwavelength size of QDs results in unique optical
properties that can be exploited for optically based biosensors.
QDs are more photostable and more resistant to photobleach-
ing compared with traditional organic fluorophores. Addi-
tionally, QDs have very narrow emission bands compared to
fluorescent dyes and this characteristic allows detection of
multiple QD labels without overlap of spectral emissions.
This property should facilitate multiplexed detection of a
greater number of target molecules.
Nanomaterial-based biosensors have been developed for
environmental monitoring, including pathogen monitoring
(89), toxin detection (90, 91), and food safety (92). Addi-
tionally, nanoparticles have been used to modify planar sens-
ing surfaces to generate increased signal strength. For
instance, a lawn of gold nanoparticles can enhance the signal
from a surface plasmon resonance biosensor (93).
Optical Biosensors
An optical biosensor uses light as a stimulus and is able to
detect changes in the intensity of light as it passes through
or refracts from a sample. Widely used optical immunosensor
examples include surface plasmon resonance (SPR)–based
sensors, optical fiber and waveguide-based biosensors, optical
ring resonator–based biosensors, and photonic crystal–based
biosensors (80).
SPR Biosensors
SPR-based immunosensors offer rapid, real-time monitor-
ing and label-free detection of antibody–antigen interactions.
SPR measures changes in the refractive index and thus in the
resonance angle at which polarized light is reflected from a
surface. An SPR immunosensor measures refractive index
changes caused by the formation of an antibody–antigen
complex on a thin metal (typically gold) surface. Several
recent reviews detail general progress in the SPR field (80,
94).
An emerging research area in surface plasmon detection is
based on using highly localized plasmon fields resulting from
nanostructured surfaces. By creating more localized plasmon
fields, the sensitivity of the device is greatly increased (81,
95). Additional recent trends in SPR biosensor development
include the use of nanoparticles and the integration of
microfluidics (96). Nevertheless, there are still limitations
with this technology, including occurrence of nonspecific
physical binding on the surface (80) and difficulty in the
detection of large targets, such as bacteria. Additionally,
the high cost of the instrument and the fact that it is difficult
to use in the field may further limit its use for environmental
monitoring.
Optical Waveguide–Based Biosensors
Optical fibers and evanescent wave fluorescence biosen-
sors are based on the principle of total internal reflection
and are another type of advanced optical biosensor. These
biosensors use narrow fibers to send excitation laser light to
the detection surface and receive emitted light. Propagation
of light through a fiber or waveguide can be very sensitive
to the surroundings, which makes the optical fibers excellent
detectors for a variety of environmental sensing applications.
Optical waveguides can be employed in either fluorescence-
based assays, through the efficient excitation of fluorophores
on the waveguide surface, or for label-free biosensing, through
recognition of surface coverage changes in the evanescent
field. Ligler et al. (97) at the Navy Research Laboratory
(NRL) developed the NRL Array Biosensor, a small, port-
able, automated system that has been successfully com-
mercialized and used for the fluorescent detection of a
variety of targets, including microorganisms and toxins,
in food and environmental matrices (64, 98, 99). This type
of biosensor typically consists of a patterned array of antibod-
ies that are immobilized on the surface of a planar waveguide.
The assay is performed on the patterned surface producing an
array of fluorescent spots. Signal transduction is achieved
using a diode laser for fluorescence excitation and a CCD
camera to capture the image.
Optical Ring Resonator–Based Biosensors
Optical microring resonators are an emerging label-free
sensing technology that also rely on evanescent waveguide–
based detection (100). In a ring resonator, light is confined
within an optical microcavity that has a curvature resulting
in the propagation of light in the form of whispering gallery
modes or circulating waveguide modes. As the optical field
recirculates, it extends into the surrounding environment
through evanescence and can interact with target molecules
as they bind to the surface. The recirculation of light allows
the light to interact with target molecules many more times,
resulting in increased sensitivity (80, 101). Indeed, single-
molecule detection has been achieved with whispering
2.3.1-8 ▪ TARGET-SPECIFIC DETECTION
gallery mode detectors (102). The biosensor is especially
useful for detection in small volume samples where the typical
volume sampled is in the range of tens of picoliters. While the
application of these sensors appears to have great potential,
they are in the early stages of development with very few
reports about their use with complex matrices such as envi-
ronmental samples.
Photonic Crystal Biosensors
Photonic crystal biosensors are another label-free detec-
tion method. One type of photonic biosensor is both sensitive
and effective for high-throughput multiplexed detection of
small molecules, proteins, and cells (103). Photonic crystals
are engineered to reflect selectively a narrow bandwidth of
light corresponding to emission from a specific fluorophore,
thereby selectively amplifying the signal sent to the detector.
The sensor is functionalized with antibodies and subse-
quently measures changes in the emitted fluorescent light
generated by the antigen assay (79, 80).
Electrochemical Biosensors
An electrochemical device couples a biological recognition-
binding event, such as an antibody–antigen interaction, to
an electrical signal. Amperometric and potentiometric trans-
ducers are the most common types of electrochemical biosen-
sors. Electrochemical immunosensors are popular because of
their low cost, high sensitivity, simplicity of construction,
ease of use and portability—properties that have spurred the
development of a large number of these sensors in the past
decade (104). In addition, because of these qualities, electro-
chemical immunosensors have great potential to be used as
field-deployable devices.
There are different electrochemical immunosensor for-
mats, although a common feature is a sandwich ELISA
format that employs detection antibodies tagged with redox
enzymes, redox-active probes, or other electrochemically
active compounds that generate the electrical signal. Several
suitable labels have been used including horseradish peroxi-
dase, alkaline phosphatase, and glucose oxidase (83, 105).
Reports detailing the fabrication of high-sensitivity
electrochemical biosensors have been described and reviewed
elsewhere (83, 106). Promising approaches often couple gold
nanostructured sensor surfaces, which are capable of binding
a large quantity of capture antibody, and detection nanopar-
ticles with large numbers of redox labels (83, 105). Detection
limits down to 1.0 fg/ml have been reported using these strat-
egies (83). Despite the many advantages of these biosensors,
there are some challenges, including limited reproducibility
and assay interference from biological matrices (107).
Mechanical Biosensors
Mechanical biosensors measure small changes in mass or force
on the sensor surface. Sensors in this category include piezo-
electric, magnetic, and cantilever designs (108). Improving
the throughput of samples analyzed by mechanical biosensors
remains a limiting factor in their development and commer-
cial implementation (108, 109).
Piezoelectric Biosensors
Piezoelectric biosensors are a label-free platform that offer
real-time output, simplicity of use, and cost-effectiveness.
The approach takes advantage of the piezoelectric effect
whereby an electrical charge accumulates in certain solid
materials in response to applied mechanical stress. A piezo-
electric material, such as quartz crystal, oscillates at a certain
frequency when a potential is applied across its surface.
Antibodies are immobilized on the piezoelectric crystal, and
when antigen binds to capture molecules on the surface,
this adds mass that is subsequently detected through a change
in the oscillation frequency.
Cantilever-Based Biosensors
Cantilever-based microbalances are a promising class of
label-free biosensors. This technology employs a small canti-
lever that has a capture antibody immobilized to the tip sur-
face. Two modes of operation are employed to detect the
binding of the target molecule to the tip: dynamic and static
mode (108, 109). In dynamic mode, the cantilever is oscil-
lated at its resonance frequency and on binding of the target
molecule, a shift in frequency is observed. In static mode, the
cantilever is stationary and deflection of the cantilever is
measured in response to the binding of the target molecule.
Cantilever transducers offer advantages in costs, assay speed,
sensitivity, reduced size, and increased reliability (108, 110).
Microfluidics Integrated with Biosensors
Critical to the development of a field-deployable biosensor
device is the integration of microfluidics for the efficient
handling of samples and reagents. Microfluidics, or “lab on
a chip” systems, can improve assay performance by reducing
the consumption of sample and other costly reagents, decreas-
ing assay time, and increasing reliability through automated
liquid handling of femto- to nanoliter volumes (111–113).
The reduced assay time is attributed to improved kinetics
compared to conventional nonmicrofluidic techniques and
is primarily a consequence of enhanced mass transport due
to reduced diffusional distances (114). By reducing the diffu-
sional distances, analysis time can be reduced from hours
down to seconds or minutes, making microfluidics a key com-
ponent in a field-deployable biosensor.
Biosensor Applications in Environmental Microbiology
The accurate detection of foodborne pathogens requires the
use of methods that are characterized by high sensitivity
and selectivity. Conventional methods rely on culture and
colony counting, PCR, and immunobased assays (51). The
U.S. Environmental Protection Agency (EPA) Method
1623: Cryptosporidium and Giardia in Water by Filtration/
IMS/FA is a good example.
This method begins with the concentration of the sample
from a large volume (>10 liters) of raw surface water by filtra-
tion and centrifugation. The concentrated sample is further
purified by immunomagnetic beads that specifically bind
Cryptosporidium oocysts and Giardia cysts. The oocysts and
cysts are then immobilized onto slides, stained with fluores-
cent antibodies, and enumerated by microscopy. Methods
such as these are routinely employed, but they are cumber-
some and do not provide real-time information. EPA Method
1623 exemplifies the challenges of immunoassays for envi-
ronmental detection in general. Namely, analytes of interest
(in this case pathogens) are usually extremely dilute and
require significant initial concentration for any assay, be it
an immunoassay, PCR, or culture method, to detect the few
pathogens that may be present. Although biosensors do not
obviate the presampling requirement, they can provide rapid
and “real-time” detection and are therefore increasingly
employed to detect foodborne pathogens in environmental
samples (42, 46, 62, 89, 92, 113, 115).
Optical techniques frequently employed for foodborne
pathogen detection include SPR and fiber optic biosensors
with fluorescence detection (42, 46, 92, 116). The sensiti-
vity limits of SPR in food safety applications ranges between
 2.3.1. Antibody-Based Technologies for Environmental Biodetection ▪ 2.3.1-9
102–104 CFU/ml (46, 92, 94). The use of SPR for the multi-
plex detection of foodborne pathogens has also been reviewed
(76, 94, 115). The utility of commercially available SPR bio-
sensors for the detection of food pathogens is extensively
reviewed by Homolo (94). The application of the NRL fiber
optic portable biosensor and other fiber optic assays for
the detection of foodborne pathogens have been reviewed
(46, 99, 113) with reported detection limits for these bio-
sensors ranging from 101 to 105 CFU/ml in food samples
such as beef and chicken carcass and resultant assay times
ranging from 30 min up to 24 h.
A number of electrochemical biosensors for the detection
of foodborne pathogens have been reported and are discussed
in recent reviews (46, 92, 113, 115). Electrochemical assays
are typically more rapid than optical detection with assay
times ranging from 6 min to several hours (46, 92). Detection
limits for these methods range from 101 to 103 CFU/ml in
buffer and food samples such as pork and milk (92).
Reviews detailing the use of mechanical biosensors (quartz
crystal microbalance and cantilever sensors) for the detection
of whole pathogenic cells have been published (46, 117).
Quartz crystal microbalance biosensors are potentially well
suited for the detection of pathogens because the sensing
principle is mass deposition-dependent (62). In general these
immunosensors can detect bacteria with a detection limit
ranging from 101 to 103 CFU/ml in buffer or food samples
with an estimated assay time ranging from 30 min up to 4 h
(117). Cantilever biosensors have detected E. coli O157:H7
in phosphate-buffered saline and ground beef with detection
limits of 10 cells/ml and 50–100 cells/ml, respectively (62).
Waterborne pathogens are a significant health problem
requiring rapid and sensitive detection methods for the
improvement of the health and welfare of millions of
people worldwide. Some waterborne pathogens are highly
infectious, for instance as few as 10 viable Cryptosporidium
parvum oocysts can cause infection in healthy adults. Conse-
quently, the U.S. EPA regulatory goal is the complete absence
of C. parvum oocysts from finished drinking water (118).
DNA biosensors are particularly effective for the detection
of low concentrations of pathogens such as required for
the detection of waterborne pathogens, but they typically
require time-consuming purification processes. Immunosen-
sors require relatively fewer sample processing steps resulting
in a simpler and faster assay. Nevertheless, improvements in
sensitivity are needed to fully develop the potential of bio-
sensors for the rapid and sensitive analysis of waterborne
pathogens. For instance, the use of an evanescent wave fiber
optic biosensor can yield results in as little as 10 min, but this
assay is not especially sensitive with limits of detection rang-
ing from 103 to 106 oocysts per ml (118). The NRL portable
evanescent waveguide biosensor has been used to detect
the waterborne pathogens G. lamblia, E. coli O157:H7, and
Salmonella enterica at the concentrations of 5 × 104 cysts/ml,
104 CFU/ml, and 103 CFU/ml, respectively (99). A cantile-
ver-based immunosensor detected as few as 10 G. lamblia cysts
per ml in as little as 15 min (117). Nevertheless, waterborne
pathogen detection in large sample volumes, such as drinking
water, still requires an initial filtration method or an immuno-
magnetic separation as discussed in detail in other chapters.
Biosensors have been applied to the detection of toxins
including biodefense toxins, mycotoxins, phycotoxins, and
cyanotoxins. PCR-based assays are not as useful for toxin
detection as for the detection of cellular pathogens because
often the toxin-encoding DNA is absent in environmental
samples. Therefore conventional detection methods for these
toxins frequently rely on animal bioassays or analytical
methods such as high-performance liquid chromatography
coupled with mass spectrometry (91). Biosensors have
emerged as reliable and promising alternative analysis tools
to these standard methods because of their high sensitivity,
ease of use, speed, and low cost. Several recent reviews have
detailed the development of optical biosensors (including
SPR and evanescent wave fiber optic immunosensors) and
electrochemical immunosensors for the detection of biode-
fense toxins (i.e. Staphylococcal enterotoxin B, ricin, and
botulinum neurotoxin), the cyanobacterial toxin micro-
cystin, mycotoxin aflatoxin B, trichothecene mycotoxins,
and shellfish poisoning toxins such as domoic acid, okadaic
acid, and saxitoxin (45, 76, 94, 115, 116, 119, 120). Despite
the promise of these immunosensors, there are challenges
in their application particularly in the development of
adequate antibody affinity reagents because of the increasing
number of toxins and their analogues that are being
recognized. For instance, there are more than 80 closely
related microcystin variants (121), leading to difficulties in
developing adequate numbers of antibodies with sufficient
specificity.
CONCLUSIONS
Immunoassay methods for environmental pathogen and
toxin detection are well established. The development of
recombinant antibody technology to produce antibodies
with enhanced binding affinities will lead to immunoassays
with better sensitivity, specificity, and reproducibility. The
development of alternative affinity reagents, such as aptamers,
engineered proteins and peptides, will provide a greater reper-
toire of affinity reagents to develop novel immunoassays.
Immunoassay formats for the detection of environmental
microbial targets range from lateral flow devices, standard
ELISAs, microarray platforms, and biosensor devices. In gen-
eral, immunoassays provide rapid assay times relative to other
conventional methods, such as colony counting and PCR
approaches. Recently, multiplexed assays for the detection
of foodborne and waterborne pathogens and toxins have
been developed using planar and bead-based microarray
approaches. Because environmental pathogens are mostly
present in very low numbers, a highly sensitive detection
method is necessary. In addition, real-time detection is requi-
site. Biosensors have the potential to address both of these
requirements. Indeed, biosensors are the fastest growing tech-
nology for pathogen detection. Integration of biosensors into
environmental and food safety monitoring systems is likely to
increase in the coming years, potentially leading to the devel-
opment of novel methods that are capable of providing the
necessary sensitivity and assay speed to replace the current
standards.
A portion of this effort was funded by the Department of Homeland
Security Science and Technology Directorate under Contract
HSHQDC-08-X-00843. Pacific Northwest National Laboratory is oper-
ated by Battelle for the U.S. Department of Energy under contract
DE-AC06-76RLO.
REFERENCES
1. Davies C. 2005. Introduction to immunoassay principles,
p 3–37. In Wild D (ed.), The immunoassay handbook. Elsev-
ier, Amsterdam.
2. Davies C. 2005. Concepts, p 103–135. In Wild D (ed.),
The immunoassay handbook, 3rd ed. Elsevier, Amsterdam.
2.3.1-10 ▪ TARGET-SPECIFIC DETECTION
3. Armbruster DA, Pry T. 2008. Limit of blank, limit of
detection and limit of quantitation. Clin Biochem Rev 29
(Suppl 1):S49–S52.
4. Braggio S, Barnaby RJ, Grossi P, Cugola M. 1996. A strat-
egy for validation of bioanalytical methods. J Pharm Biomed
Anal 14:375–388.
5. Holliger P, Hudson PJ. 2005. Engineered antibody frag-
ments and the rise of single domains. Nat Biotechnol 23:
1126–1136.
6. Taussig MJ, Stoevesandt O, Borrebaeck CA, Bradbury
AR, Cahill D, Cambillau C, de Daruvar A, Dubel S,
Eichler J, Frank R, Gibson TJ, Gloriam D, Gold L,
Herberg FW, Hermjakob H, Hoheisel JD, Joos TO,
Kallioniemi O, Koegl M, Konthur Z, Korn B, Kremmer
E, Krobitsch S, Landegren U, van der Maarel S, McCaff-
erty J, Muyldermans S, Nygren PA, Palcy S, Pluckthun
A, Polic B, Przybylski M, Saviranta P, Sawyer A, Sher-
man DJ, Skerra A, Templin M, Ueffing M, Uhlen M.
2007. ProteomeBinders: planning a European resource of
affinity reagents for analysis of the human proteome. Nat
Meth 4:13–17.
7. Colwill K, Graslund S. 2011. A roadmap to generate
renewable protein binders to the human proteome. Nat
Meth 8:551–558.
8. Cottingham K. 2008. Antibodypedia seeks to answer the
question: “how good is that antibody?” J Proteome Res 7:
4213.
9. Bjorling E, Uhlen M. 2008. Antibodypedia, a portal
for sharing antibody and antigen validation data. Mol Cell
Proteomics 7:2028–2037.
10. Major SM, Nishizuka S, Morita D, Rowland R, Sunshine
M, Shankavaram U, Washburn F, Asin D, Kouros-Mehr
H, Kane D, Weinstein JN. 2006. AbMiner: a bio-
informatic resource on available monoclonal anti-
bodies and corresponding gene identifiers for genomic,
proteomic, and immunologic studies. BMC Bioinformatics
7:192.
11. Stoevesandt O, Taussig MJ. 2007. Affinity reagent resour-
ces for human proteome detection: initiatives and perspec-
tives. Proteomics 7:2738–2750.
12. Nollau P. 2011. Generating high-quality protein binders: a
large screening effort pays off. Nat Meth 8:545–546.
13. Seurynck-Servoss SL, Baird CL, Miller KD, Pefaur NB,
Gonzalez RM, Apiyo DO, Engelmann HE, Srivastava S,
Kagan J, Rodland KD, Zangar RC. 2008. Immobilization
strategies for single-chain antibody microarrays. Proteomics
8:2199–2210.
14. Mondon P, Dubreuil O, Bouayadi K, Kharrat H. 2008.
Human antibody libraries: a race to engineer and explore
a larger diversity. Front Biosci 13:1117–1129.
15. Zhou HY, Zhang YL, Lu GD, Ji H, Rodi CP. 2011.
Recombinant antibody libraries and selection technolo-
gies. New Biotechnol 28:448–452.
16. Fellouse FA, Li B, Compaan DM, Peden AA, Hymowitz
SG, Sidhu SS. 2005. Molecular recognition by a binary
code. J Mol Biol 348:1153–1162.
17. Barbas CF, 3rd, Bain JD, Hoekstra DM, Lerner RA.
1992. Semisynthetic combinatorial antibody libraries: a
chemical solution to the diversity problem. Proc Natl
Acad Sci USA 89:4457–4461.
18. Hoet RM, Cohen EH, Kent RB, Rookey K, Schoon-
broodt S, Hogan S, Rem L, Frans N, Daukandt M, Pieters
H, van Hegelsom R, Neer NC, Nastri HG, Rondon IJ,
Leeds JA, Hufton SE, Huang L, Kashin I, Devlin M,
Kuang G, Steukers M, Viswanathan M, Nixon AE,
Sexton DJ, Hoogenboom HR, Ladner RC. 2005.
Generation of high-affinity human antibodies by com-
bining donor-derived and synthetic complementarity-
determining-region diversity. Nat Biotechnol 23:344–348.
19. Harel Inbar N, Benhar I. 2012. Selection of antibodies
from synthetic antibody libraries. Arch Biochem Biophys
526:87–98.
20. Ferrara F, Naranjo LA, Kumar S, Gaiotto T, Mukundan
H, Swanson B, Bradbury AR. 2012. Using phage and
yeast display to select hundreds of monoclonal antibodies:
application to antigen 85, a tuberculosis biomarker. PLoS
One 7:e49535.
21. Boder ET, Midelfort KS, Wittrup KD. 2000. Directed
evolution of antibody fragments with monovalent femto-
molar antigen-binding affinity. Proc Natl Acad Sci USA
97:10701–10705.
22. Siegel RW, Baugher W, Rahn T, Drengler S, Tyner J.
2008. Affinity maturation of tacrolimus antibody
for improved immunoassay performance. Clin Chem 54:
1008–1017.
23. Traxlmayr MW, Obinger C. 2012. Directed evolution of
proteins for increased stability and expression using yeast
display. Arch Biochem Biophys 526:174–180.
24. Gebauer M, Skerra A. 2009. Engineered protein scaffolds
as next-generation antibody therapeutics. Curr Opin Chem
Biol 13:245–255.
25. Skerra A. 2007. Alternative non-antibody scaffolds
for molecular recognition. Curr Opin Biotechnol 18:
295–304.
26. Nygren PA, Skerra A. 2004. Binding proteins from alter-
native scaffolds. J Immunol Methods 290:3–28.
27. Skerra A. 2000. Engineered protein scaffolds for molecular
recognition. J Mol Recognit 13:167–187.
28. Gronwall C, Stahl S. 2009. Engineered affinity proteins—
generation and applications. J Biotechnol 140:254–269.
29. Friedman M, Stahl S. 2009. Engineered affinity proteins
for tumour-targeting applications. Biotechnol Appl Biochem
53:1–29.
30. Soares TA, Boschek CB, Apiyo D, Baird C, Straatsma
TP. 2010. Molecular basis of the structural stability of a
Top7-based scaffold at extreme pH and temperature condi-
tions. J Mol Graph Model 28:755–765.
31. Boschek CB, Apiyo DO, Soares TA, Engelmann HE,
Pefaur NB, Straatsma TP, Baird CL. 2009. Engineering
an ultra-stable affinity reagent based on Top7. Protein Eng
Des Sel 22:325–332.
32. Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard
BL, Baker D. 2003. Design of a novel globular protein
fold with atomic-level accuracy. Science 302:1364–1368.
33. Stoltenburg R, Reinemann C, Strehlitz B. 2007.
SELEX—a (r)evolutionary method to generate high-
affinity nucleic acid ligands. Biomol Eng 24:381–403.
34. Ellington AD, Szostak JW. 1990. In vitro selection of
RNA molecules that bind specific ligands. Nature 346:
818–822.
35. Tuerk C, Gold L. 1990. Systematic evolution of ligands by
exponential enrichment: RNA ligands to bacteriophage
T4 DNA polymerase. Science 249:505–510.
36. Baird GS. 2010. Where are all the aptamers? Am J Clin
Pathol 134:529–531.
37. Gold L, Ayers D, Bertino J, Bock C, Bock A, Brody EN,
Carter J, Dalby AB, Eaton BE, Fitzwater T, Flather D,
Forbes A, Foreman T, Fowler C, Gawande B, Goss M,
Gunn M, Gupta S, Halladay D, Heil J, Heilig J, Hicke
B, Husar G, Janjic N, Jarvis T, Jennings S, Katilius E,
Keeney TR, Kim N, Koch TH, Kraemer S, Kroiss L,
Le N, Levine D, Lindsey W, Lollo B, Mayfield W, Mehan
M, Mehler R, Nelson SK, Nelson M, Nieuwlandt D,
Nikrad M, Ochsner U, Ostroff RM, Otis M, Parker T,
Pietrasiewicz S, Resnicow DI, Rohloff J, Sanders G,
Sattin S, Schneider D, Singer B, Stanton M, Sterkel A,
Stewart A, Stratford S, Vaught JD, Vrkljan M, Walker
JJ, Watrobka M, Waugh S, Weiss A, Wilcox SK, Wolfson
 2.3.1. Antibody-Based Technologies for Environmental Biodetection ▪ 2.3.1-11
A, Wolk SK, Zhang C, Zichi D. 2010. Aptamer-based
multiplexed proteomic technology for biomarker discovery.
PLoS One 5:e15004.
38. Kryscio DR, Peppas NA. 2012. Critical review and per-
spective of macromolecularly imprinted polymers. Acta
Biomaterialia 8:461–473.
39. Vasapollo G, Del Sole R, Mergola L, Lazzoi MR, Scardino
A, Scorrano S, Mele G. 2011. Molecularly imprinted pol-
ymers: present and future prospective. Int J Mol Sci 12:
5908–5945.
40. Biagini RE, Smith JP, Sammons DL, MacKenzie BA,
Striley CAF, Robertson SK, Snawder JE. 2008. Analyti-
cal performance criteria—the use of immunochemical
and biosensor methods for occupational and environmen-
tal monitoring. Part I: Introduction to immunoassays.
J Occup Environ Hyg 5:D25–D32.
41. Knopp D. 2006. Immunoassay development for environ-
mental analysis. Anal Bioanal Chem 385:425–427.
42. Lazcka O, Del Campo FJ, Munoz FX. 2007. Pathogen
detection: a perspective of traditional methods and biosen-
sors. Biosens Bioelectron 22:1205–1217.
43. van Apeldoorn ME, van Egmond HP, Speijers GJA, Bak-
ker GJI. 2007. Toxins of cyanobacteria. Mol Nutr Food Res
51:7–60.
44. McElhiney J, Lawton LA. 2005. Detection of the cyano-
bacterial hepatotoxins microcystins. Toxicol Appl Pharm
203:219–230.
45. Meneely JP, Ricci F, van Egmond HP, Elliott CT. 2011.
Current methods of analysis for the determination of tri-
chothecene mycotoxins in food. Trends Anal Chem 30:
192–203.
46. Velusamy V, Arshak K, Korostynska O, Oliwa K, Adley
C. 2010. An overview of foodborne pathogen detection: In
the perspective of biosensors. Biotechnol Adv 28:232–254.
47. Krska R, Molinelli A. 2009. Rapid test strips for analysis of
mycotoxins in food and feed. Anal Bioanal Chem 393:
67–71.
48. Posthuma-Trumpie GA, Korf J, van Amerongen A.
2009. Lateral flow (immuno) assay: its strengths, weak-
nesses, opportunities and threats. A literature survey.
Anal Bioanal Chem 393:569–582.
49. Zheng MZ, Richard JL, Binder J. 2006. A review of rapid
methods for the analysis of mycotoxins. Mycopathologia
161:261–273.
50. Mark D, Haeberle S, Roth G, von Stetten F, Zengerle R.
2010. Microfluidic lab-on-a-chip platforms: requirements,
characteristics and applications. Chem Soc Rev 39:
1153–1182.
51. Jasson V, Jacxsens L, Luning P, Rajkovic A, Uyttendaele
M. 2010. Alternative microbial methods: an overview and
selection criteria. Food Microbiol 27:710–730.
52. Kingsmore SF. 2006. Multiplexed protein measurement:
technologies and applications of protein and antibody
arrays. Nat Rev Drug Discov 5:310–320.
53. Zhu H, Snyder M. 2003. Protein chip technology. Curr
Opin Chem Biol 7:55–63.
54. Ekins RP, Chu FW. 1991. Multianalyte microspot immu-
noassay—microanalytical “compact disk” of the future.
Clin Chem 37:1955–1967.
55. Ekins RP. 1998. Ligand assays: from electrophoresis to
miniaturized microarrays. Clin Chem 44:2015–2030.
56. Ellington AA, Kullo IJ, Bailey KR, Klee GG. 2010.
Antibody-based protein multiplex platforms: technical
and operational challenges. Clin Chem 56:186–193.
57. Kusnezow W, Syagailo YV, Ruffer S, Baudenstiel N,
Gauer C, Hoheisel JD, Wild D, Goychuk I. 2006. Opti-
mal design of microarray immunoassays to compensate for
kinetic limitations—theory and experiment. Mol Cell Pro-
teomics 5:1681–1696.
58. Kusnezow W, Syagailo YV, Goychuk I, Hoheisel JD,
Wild DG. 2006. Antibody microarrays: the crucial impact
of mass transport on assay kinetics and sensitivity. Expert
Rev Mol Diagn 6:111–124.
59. Hartmann M, Toegl A, Kirchner R, Templin MF, Joos
TO. 2006. Increasing robustness and sensitivity of protein
microarrays through microagitation and automation. Anal
Chim Acta 564:66–73.
60. Gonzalez RM, Seurynck-Servoss SL, Crowley SA,
Brown M, Omenn GS, Hayes DF, Zangar RC. 2008.
Development and validation of sandwich ELISA microar-
rays with minimal assay interference. J Proteome Res 7:
2406–2414.
61. Cass T, Ligler FS. 1998. Immobilized biomolecules
in analysis: a practical approach. In Hames BD (ed.),
The practical approach series. Oxford University Press,
New York.
62. Skottrup PD, Nicolaisen M, Justesen AF. 2008. Towards
on-site pathogen detection using antibody-based sensors.
Biosens Bioelectron 24:339–348.
63. Henares TG, Mizutani F, Hisamoto H. 2008. Current
development in microfluidic immunosensing chip. Anal
Chim Acta 611:17–30.
64. Ligler FS, Sapsford KE, Golden JP, Shriver-Lake LC,
Taitt CR, Dyer MA, Barone S, Myatt CJ. 2007. The
array biosensor: portable, automated systems. Anal Sci 23:
5–10.
65. Seidel M, Niessner R. 2008. Automated analytical micro-
arrays: a critical review. Anal Bioanal Chem 391:
1521–1544.
66. Gehring AG, Albin DM, Reed SA, Tu SI, Brewster JD.
2008. An antibody microarray, in multiwell plate format,
for multiplex screening of foodborne pathogenic bacteria
and biomolecules. Anal Bioanal Chem 391:497–506.
67. Fischer NO, Tarasow TM, Tok JB. 2007. Heightened
sense for sensing: recent advances in pathogen immunoas-
say sensing platforms. Analyst 132:187–191.
68. Parro V. 2010. Biomonitoring by antibody microarrays. In
Timmis KN (ed.), Handbook of hydrocarbon and lipid micro-
biology. Springer, Berlin.
69. Zhang Y, Lou J, Jenko KL, Marks JD, Varnum SM. 2012.
Simultaneous and sensitive detection of six serotypes of
botulinum neurotoxin using enzyme-linked immuno-
sorbent assay-based protein antibody microarrays. Anal Bio-
chem 430:185–192.
70. Valiokas R. 2012. Nanobiochips. Cell Mol Life Sci 69:
347–356.
71. Hahm JI. 2011. Polymeric surface-mediated, high-density
nano-assembly of functional protein arrays. J Biomed Nano-
technol 7:731–742.
72. Houser B. 2012. Bio-Rad’s Bio-Plex(R) suspension array
system, xMAP technology overview. Arch Physiol Biochem
118:192–196.
73. Clotilde LM, Bernard Ct, Salvador A, Lin A, Lauzon CR,
Muldoon M, Xu Y, Lindpaintner K, Carter JM. 2013. A
7-plex microbead-based immunoassay for serotyping Shiga
toxin-producing Escherichia coli. J Microbiol Meth 92:
226–230.
74. Kim JS, Taitt CR, Ligler FS, Anderson GP. 2010. Multi-
plexed magnetic microsphere immunoassays for detection
of pathogens in foods. Sens Instrum Food Qual Saf 4:73–81.
75. Holford TR, Davis F, Higson SP. 2012. Recent trends in
antibody based sensors. Biosens Bioelectron 34:12–24.
76. Raz SR, Haasnoot W. 2011. Multiplex bioanalytical meth-
ods for food and environmental monitoring. Trends Anal
Chem 30:1526–1537.
77. Rodriguez-Mozaz S, de Alda MJL, Barcelo D. 2006.
Biosensors as useful tools for environmental analysis and
monitoring. Anal Bioanal Chem 386:1025–1041.
2.3.1-12 ▪ TARGET-SPECIFIC DETECTION
78. Rogers KR. 2006. Recent advances in biosensor techni-
ques for environmental monitoring. Anal Chim Acta 568:
222–231.
79. Qavi AJ, Washburn AL, Byeon JY, Bailey RC. 2009.
Label-free technologies for quantitative multiparameter
biological analysis. Anal Bioanal Chem 394:121–135.
80. Fan XD, White IM, Shopova SI, Zhu HY, Suter JD, Sun
YZ. 2008. Sensitive optical biosensors for unlabeled tar-
gets: a review. Anal Chim Acta 620:8–26.
81. Hunt HK, Armani AM. 2010. Label-free biological and
chemical sensors. Nanoscale 2:1544–1559.
82. Ray S, Mehta G, Srivastava S. 2010. Label-free detection
techniques for protein microarrays: prospects, merits and
challenges. Proteomics 10:731–748.
83. Rusling JF. 2012. Nanomaterials-based electrochemical
immunosensors for proteins. Chem Rec 12:164–176.
84. Jans H, Huo Q. 2012. Gold nanoparticle-enabled biolog-
ical and chemical detection and analysis. Chem Soc Rev 41:
2849–2866.
85. Balasubramanian K, Burghard M. 2006. Biosensors based
on carbon nanotubes. Anal Bioanal Chem 385:452–468.
86. Kuila T, Bose S, Khanra P, Mishra AK, Kim NH, Lee JH.
2011. Recent advances in graphene-based biosensors.
Biosens Bioelectron 26:4637–4648.
87. Koh I, Josephson L. 2009. Magnetic nanoparticle sensors.
Sensors 9:8130–8145.
88. Frasco MF, Chaniotakis N. 2009. Semiconductor quan-
tum dots in chemical sensors and biosensors. Sensors 9:
7266–7286.
89. Vikesland PJ, Wigginton KR. 2010. Nanomaterial
enabled biosensors for pathogen monitoring—a review.
Environ Sci Technol 44:3656–3669.
90. Upadhyayula VKK. 2012. Functionalized gold nanopar-
ticle supported sensory mechanisms applied in detection
of chemical and biological threat agents: a review. Anal
Chim Acta 715:1–18.
91. Campas M, Garibo D, Prieto-Simon B. 2012. Novel nano-
biotechnological concepts in electrochemical biosensors
for the analysis of toxins. Analyst 137:1055–67.
92. Perez-Lopez B, Merkoci A. 2011. Nanomaterials based
biosensors for food analysis applications. Trends Food Sci
Tech 22:625–639.
93. Ligler FS. 2009. Perspective on optical biosensors and inte-
grated sensor systems. Anal Chem 81:519–526.
94. Homola J. 2008. Surface plasmon resonance sensors for
detection of chemical and biological species. Chem Rev
108:462–493.
95. Mayer KM, Hafner JH. 2011. Localized surface plasmon
resonance sensors. Chem Rev 111:3828–3857.
96. Abbas A, Linman MJ, Cheng Q. 2011. New trends in
instrumental design for surface plasmon resonance-based
biosensors. Biosens Bioelectron 26:1815–1824.
97. Ligler FS, Taitt CR, Shriver-Lake LC, Sapsford KE, Shu-
bin Y, Golden JP. 2003. Array biosensor for detection of
toxins. Anal Bioanal Chem 377:469–477.
98. Golden JP, Taitt CR, Shriver-Lake LC, Shubin YS,
Ligler FS. 2005. A portable automated multianalyte bio-
sensor. Talanta 65:1078–1085.
99. Taitt CR, Anderson GP, Ligler FS. 2005. Evanescent
wave fluorescence biosensors. Biosens Bioelectron 20:
2470–2487.
100. Vollmer F, Arnold S. 2008. Whispering-gallery-mode bio-
sensing: label-free detection down to single molecules. Nat
Meth 5:591–596.
101. Luchansky MS, Bailey RC. 2012. High-Q optical sensors
for chemical and biological analysis. Anal Chem 84:
793–821.
102. Armani AM, Kulkarni RP, Fraser SE, Flagan RC, Vahala
KJ. 2007. Label-free, single-molecule detection with opti-
cal microcavities. Science 317:783–787.
103. Cunningham BT, Laing L. 2006. Microplate-based,
label-free detection of biomolecular interactions:
applications in proteomics. Exp Rev Proteomics 3:
271–281.
104. Ronkainen NJ, Halsall HB, Heineman WR.
2010. Electrochemical biosensors. Chem Soc Rev 39:
1747–1763.
105. Pingarron JM, Yanez-Sedeno P, Gonzalez-Cortes A.
2008. Gold nanoparticle-based electrochemical biosen-
sors. Electrochim Acta 53:5848–5866.
106. Wang J. 2005. Carbon-nanotube based electrochemical
biosensors: a review. Electroanal 17:7–14.
107. Tothill IE. 2009. Biosensors for cancer markers diagnosis.
Semin Cell Dev Biol 20:55–62.
108. Arlett JL, Myers EB, Roukes ML. 2011. Comparative
advantages of mechanical biosensors. Nat Nanotechnol 6:
203–215.
109. Calleja M, Kosaka PM, San Paulo A, Tamayo J. 2012.
Challenges for nanomechanical sensors in biological detec-
tion. Nanoscale 4:4925–4938.
110. Lavrik NV, Sepaniak MJ, Datskos PG. 2004. Cantilever
transducers as a platform for chemical and biological sen-
sors. Rev Sci Instrum 75:2229–2253.
111. Cheng SB, Skinner CD, Taylor J, Attiya S, Lee WE,
Picelli G, Harrison DJ. 2001. Development of a multi-
channel microfluidic analysis system employing affinity
capillary electrophoresis for immunoassay. Anal Chem 73:
1472–1479.
112. Ramsey JD, Collins GE. 2005. Integrated microfluidic
device for solid-phase extraction coupled to micellar elec-
trokinetic chromatography separation. Anal Chem 77:
6664–6670.
113. Yoon JY, Kim B. 2012. Lab-on-a-chip pathogen sensors for
food safety. Sensors 12:10713–10741.
114. Situma C, Hashimoto M, Soper SA. 2006. Merging micro-
fluidics with microarray-based bioassays. Biomol Eng 23:
213–231.
115. McGrath TF, Elliott CT, Fodey TL. 2012. Biosensors for
the analysis of microbiological and chemical contaminants
in food. Anal Bioanal Chem 403:75–92.
116. Borisov SM, Wolfbeis OS. 2008. Optical biosensors. Chem
Rev 108:423–461.
117. Xu S. 2012. Electromechanical biosensors for pathogen
detection. Microchim Acta 178:245–260.
118. Connelly JT, Baeumner AJ. 2012. Biosensors for the
detection of waterborne pathogens. Anal Bioanal Chem
402:117–127.
119. Vilarino N, Louzao MC, Vieytes MR, Botana LM. 2010.
Biological methods for marine toxin detection. Anal Bioa-
nal Chem 397:1673–1681.
120. Singh S, Srivastava A, Oh HM, Ahn CY, Choi GG,
Asthana RK. 2012. Recent trends in development of
biosensors for detection of microcystin. Toxicon 60:
878–894.
121. Westrick JA, Szlag DC, Southwell BJ, Sinclair J. 2010. A
review of cyanobacteria and cyanotoxins removal/inactiva-
tion in drinking water treatment. Anal Bioanal Chem 397:
1705–1014.
 PCR, Real-Time PCR, Digital PCR, and
Isothermal Amplification
RACHEL A. BARTHOLOMEW, JANINE R. HUTCHISON, TIMOTHY M. STRAUB,
AND DOUGLAS R. CALL
2.3.2
PCR was developed in 1983 by Kary Banks Mullis (1, 2).
What began as a very simple idea to apply a naturally occur-
ring enzyme to produce millions of copies of DNA fragments
has led to a complete revolution in the discipline of molecular
biology and a Nobel Prize in Chemistry for Mullis in 1993 (3).
Early uses of PCR included pathogen detection, cloning, and
sequencing (4–7). Currently, applications of nucleic acid
amplification methods in research and diagnostics are broad,
ranging from detection of rare single-nucleotide polymor-
phisms that are predictors for various human genetic diseases
to characterization of microbial communities and pathogen
detection in clinical and environmental samples (8–10).
For the purposes of this chapter, we focus on PCR and exten-
sions of PCR as they apply to environmental microbiology.
The goals of this chapter include providing (1) a back-
ground on the basics of PCR, (2) descriptions of variations
on PCR, (3) standards for experiments involving quantitative
applications of PCR, and (4) discussion of knowledge gaps in
the field and areas for further research. In our discussion of
these protocols, we highlight examples of these techniques
as a detection tool for scientists.
BACKGROUND
PCR Basics
PCR can be simply described as a means to amplify a specific
or many specific nucleotide sequence(s) from a biological
sample. The basic materials needed to conduct PCR are (1)
thermal cycler, (2) nucleic acid template (DNA or RNA),
(3) DNA polymerase, (4) primers (i.e., short DNA oligonu-
cleotides complementary to the sequence being amplified),
(5) deoxynucleotide triphosphates (dNTPs), and (6) buffer
(including appropriate salts). While it is possible to purchase
the reagents separately so that a user can optimize conditions,
most manufacturers sell premixed kits that include the DNA
polymerase, buffer with magnesium added, and dNTPs. Mas-
ter mixes simplify handling and consequently reduce oppor-
tunities for introducing contaminants or calculation errors
in the experiment.
PCR reaction components are combined and heated to
first separate the double-stranded DNA (denaturation). The
reaction temperature is then reduced so that primers can
anneal to the complementary target sequence. The DNA
doi:10.1128/9781555818821.ch2.3.2
2.3.2-1
polymerase then extends the primer through the free 30
hydroxyl group by incorporating dNTPs. The reagent and
cycling conditions can be modified to ensure that the target
amplicon is generated at maximal efficiency (11). The entire
process of denaturing, annealing, and extension completes
one cycle of the reaction (Fig. 1). The number of cycles in
a PCR is commonly 20–40 cycles (12–14). Theoretically,
there is an exponential amplification with every cycle com-
pletion (2n, where n = number of cycles) (13). In reality,
with increasing number of cycles, reagents become limited,
and the accumulation of product enters the plateau phase
where little if any additional product is formed. This topic
is discussed further in the section on real-time PCR. An
important exception to this rule is a newer technique called
linear after the exponential (LATE) PCR. This is a special-
ized asymmetric PCR technique where one of the primers is
added at a limiting concentration (e.g., 50 nM) with respect
to the other primer (e.g., 900 nM). The PCR starts out as an
exponential PCR reaction as already described, but when the
limiting primer is exhausted the PCR enters the linear ampli-
fication phase with the excess primer, generating single-
stranded products (15).
Thermal Stable Polymerases
Perhaps the two most important components of PCR involve
the choice of thermal stable DNA polymerase and design of
the primers for the reaction (discussed shortly). The first
PCR reactions utilized a DNA polymerase from Escherichia
coli that had poor thermostability and was destroyed during
the denaturing step, thereby requiring continual addition
during the reaction (13). An alternative was found with
Taq polymerase, an enzyme derived from the Thermus aquati-
cus bacteria, which was sufficiently stable for the duration of
the reaction (13, 16). Another enzyme, Pfu (source: Pyro-
coccus furiosus), is both stable at high temperatures and has
proof reading capability, lowering the overall error rate for
incorporation of the incorrect nucleotide (1 error in 1.3 mil-
lion bases versus 1 in 9,000 for Taq) (17, 18). There are other
PCR enzymes, including Pwo (source: Pyrococcus woese), rTth
(source: Thermus thermophilius), and Tfl (source: Thermus
flavus) (19). Benefits to using these alternative enzymes
include reduced spurious amplification of nontarget ampli-
cons and high fidelity to the original sequence (20).
2.3.2-2 ▪ TARGET-SPECIFIC DETECTION
FIGURE 1 Principles of PCR. doi:10.1128/9781555818821.ch2.3.2.f1
Primer Design
Practitioners often design sequence-specific primers for the
desired amplicon. Criteria for designing primers may include
number of bases (usually <25 base pairs in length), the num-
ber of guanine and cytosine bases in a given primer (usually
no more than 60%), minimizing duplication of a single base
or sets of bases in series, the temperature at which the primers
and the newly formed amplified product will melt (Tm), base
composition including inclusion (or not) of guanines and
cytosine at the 30 terminus, and the general composition to
avoid formation of hairpins or dimers (21). Additional con-
siderations are required when designing primers for a multi-
plex (e.g., multiple primer sets/targets per reaction) and
nested (multiple rounds of PCR from prior amplicon product)
PCR (21).
There are a variety of automated PCR primer design pro-
grams, including but not limited to PCR instrumentation
software (i.e., Primer Express, Applied Biosystems), freeware
programs available on the Internet (i.e., Primer3, BLAST-
Primers, etc.), or company-based programs (i.e., Operon,
https://www.lifetechnologies.com/us/en/home.html; http://
www.molbiol-tools.ca/PCR.htm; http://www.operon.com/
tools/oligo-design-tools.aspx; and http://www.ncbi.nlm.nih.
gov/tools/primer-blast/). Automated programs are mostly
used to select the primer sequences that pass basic design
criterion. Modifications to primers may include a single
base difference (e.g., single nucleotide polymorphisms),
degenerate base(s) (i.e., inosine, which may bind to more
than one base), or base modifications (e.g., phosphorylation,
quenchers, spacers, attachment chemistry/linkers) (22, 23).
Both the polymerase enzyme and the design of the primers
are driven by the requirements of the end user and are specific
to the downstream applications.
PCR revolutionized molecular biology; its initial limita-
tions were overcome with the advent of new approaches to
conventional PCR. Adaptations of the basic protocol include
incorporation of a reverse transcription (RT) step prior to
PCR, which allows the study of mRNA transcripts and viral
RNA via RT-PCR (24). Real-time PCR (qPCR) allows the
user to quantify DNA and RNA transcripts (RT-qPCR)
(24, 25). A sequence-specific approach for DNA amplifi-
cation, isothermal amplification, does not require tempera-
ture cycling and has been shown to be fast and efficient
(26, 27). Finally, with the advent of digital PCR, quantifica-
tion has been streamlined by replacing a conventional stand-
ard curve with direct estimation of copy number (28, 29).
Overviews of these methodologies are presented in remaining
sections of the chapter.
REVERSE TRANSCRIPTION-PCR
Reverse transcription-PCR involves amplification and detec-
tion of RNA targets (mRNA or viral RNA). A reverse tran-
scriptase enzyme (an RNA-dependent DNA polymerase) is
used to transcribe the RNA into single-stranded complemen-
tary DNA (cDNA) that is then used as template DNA for a
subsequent PCR reaction. The basic materials needed to
conduct reverse transcription are (1) mRNA or viral RNA,
(2) primers, (3) buffer, (4) dithiothreitol (DTT; used to limit
interference from secondary structures), (5) dNTPs, and (6)
enzyme (RNA-dependent DNA polymerase). The user may
also choose to include a reagent designed to protect the
RNA from digestion by contaminating RNAses (e.g., RNAse
inhibitor).
As with the PCR reaction, there are many options for
reagents in a RT reaction, including the enzyme and primer
type (i.e., target-specific versus random primers for the RT
step). Recombinant enzymes derived from the retroviruses
avian myeloblastosis virus or Moloney murine leukemia virus
serve as the basis for many commercial products for RT (30,
31). Others, such as rTth (source: Thermus thermophilius) gen-
erate both cDNA and DNA (32). This enzyme functions as
a reverse transcriptase in the presence of Mn2+ and functions
as a thermal stable DNA polymerase in the presence of Mg2+.
The recently isolated 3173 Pol (source: viral pyrophage)
appears to be at least comparable, if not superior, to com-
monly used RT enzymes (33). Users have many choices
for RT enzymes as a function of the length of the target ampli-
con, desired temperature of the RT reaction, or the ability of
the enzyme to amplify limited quantities (nanogram versus
picogram) of starting material (34).
Reverse transcription can be a stand-alone reaction and
the resultant cDNA can be added to the PCR reaction (two-
step), or the RT and PCR can be combined as a single reac-
tion (one-step) (35). In the one-step reaction, all the reagents
for the RT and the PCR are added to the reaction, including
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-3
the gene-specific primers only (i.e., no random primers) (35).
Because fewer manipulations are required with this approach,
there are fewer chances of introducing contamination to the
reaction (35). This one-step approach yields high specificity
for the target sequence, but typically lower analytical sensitiv-
ity (36, 37). This approach does not allow the user to optimize
the RT and the PCR independently (35).
In a two-step reaction, the RT and the PCR steps occur
independently of one another (35). Users can select either
nonspecific (e.g., oligo-dT or random hexamers) or the
gene-specific primers in the reaction, and the result is a either
a complementary cDNA library of the transcripts in the
original sample (e.g., random primers) or a library of cDNAs
generated from the gene-specific primers (35). User needs
will influence the selection of primer type for the RT reac-
tion, including amount of starting material, quality of starting
material, type of starting material (i.e., bacterial versus human
RNA), and the number of genes targeted in downstream
reactions.
There are trade-offs between generating cDNA using one-
or two- step approaches. While the one-step approach reduces
manipulations and reagents, the two-step can be more sen-
sitive (35, 38). Increased sensitivity is especially important
for environmental samples where target concentration is
usually limiting. Bustin compared one-step and two-step
approaches and found a 10-fold increase in RNA concen-
tration for real-time RT-PCR reactions (described below),
with the two-step reaction providing greater sensitivity
(39). Users must be cautious when proceeding from the RT
to the PCR step in a two-step reaction because the RT reac-
tion itself can be inhibitory to the PCR. Liss and others
demonstrated that it is necessary to only add a fraction of
the RT reaction to limit introduction of PCR inhibitors
while introducing sufficient cDNA to yield a product (40–
42). There are other ways to reduce the potential inhibitory
impact of the RT enzyme on the subsequent PCR step,
including ethanol and/or ethanol/phenol precipitation
between reactions, which removes DTT from the RT step,
and using heat to render the RT enzyme inactive (36, 40–
42). Other strategies for coping with PCR inhibition are
reviewed in chapter 2.3.4.
Many factors can affect the efficiency of the transcription,
including the conformation of the RNA molecule (e.g., sec-
ondary structure), the use of ancillary enzymes such as RNAse
H to degrade RNA from RNA:DNA duplexes, primer
annealing efficiency, and the specific reverse transcriptase
used (35, 43–46). Moreover, one must be cognizant of the
starting quality of the RNA in each experiment, whether
from a single source or performing multiple experiments
from the same source (38). RNA is less stable than DNA,
and care must be taken to preserve sample integrity for the
reverse transcription reaction because the quality of the
RNA is critical (38, 47–49). For transcripts derived from
living cells, removal of DNA prior to the RT reaction is
also important to ensure that any amplification is due to
the presence of the RNA rather than the DNA present in
the original sample. To confirm that DNA is removed, exper-
imental designs frequently incorporate a “no RT” PCR reac-
tion where successful amplification for this negative control
indicates presence of intact DNA (50).
REAL-TIME (FLUORESCENT) PCR
During a conventional PCR reaction, key reagents are
exhausted once the reaction nears completion of the linear
phase and the final quantification of amplified product can
be highly variable as the reaction approaches the plateau stage
(51). Consequently, quantification estimates based on total
product at the end of a conventional reaction are inherently
variable due to stochastic factors such as premature DNA pol-
ymerase termination. “Quantitative competitive” PCR was
an early attempt to provide more information about the con-
centration of template in a PCR reaction (52). Bustin noted
that this approach is expensive, time-consuming, and error-
prone; has poor reproducibility; and requires extensive and
complex protocols that lack a standard methodology (53).
Consequently, other techniques focused on estimating con-
centration early in the PCR amplification process where
the amount of PCR product is less variable for a given starting
template concentration. This drove development of fluores-
cent reporter systems and instrumentation needed to monitor
PCR product formation in “real time” (25, 54).
As with conventional PCR, fluorescent PCR incorporates
the same stages and reaction constituents, with the addition
of a nonspecific DNA intercalating (e.g., SYBR Green) dye
or specific oligonucleotide probe fluorescence resonance
energy transfer (FRET)–based chemistries (55–57). Both
strategies make it possible to detect amplified products via flu-
orescence during the reaction (e.g., real time, Fig. 2). As we
further discuss later, when specific oligonucleotide probes
are employed, it is possible to perform multiplexed (more
than one gene target) detection. The type of detection
method selected depends on desired specificity, cost, instru-
mentation, and overall project goals.
Fluorescent PCR Using Intercalating Dye Chemistry
Nonspecific intercalating dyes, including (but not limited to)
SYBR Green, SYBR Gold, and YoYo-1, provide a less expen-
sive alternative to probe-based systems. Intercalating dyes
do not, however, discriminate between expected and non-
specific amplification products and primer dimers (58). To
mitigate this problem, the user must perform a melt-curve
analysis after completion of the PCR reaction to determine
if nonspecific products are present. Alternatively, conven-
tional gel electrophoresis will also serve this function (58).
Melt-curve analysis works because the temperature at which
the two distinct strands of the DNA “melt” (Tm) differ due
to size and base composition. When the Tm is met, the strands
separate and the intercalating dye is no longer responsive.
Thus, by slowly ramping the temperature in a real-time ther-
mal cycler, the practitioner can record when peaks of fluores-
cence are observed and dissipated as a marker for the presence
of distinct PCR amplicons (36, 58).
Specific Fluorescent Oligonucleotide Probes for
Fluorescent PCR
Specific, fluorescent oligonucleotide probes are an alternative
to using intercalating dyes and are composed of two basic
types: hydrolysis probes and molecular beacons (Fig. 2)
(59). Both hydrolysis and molecular beacon type probes
typically consist of a covalently bound fluorophore at the 50
end of the molecule and a quencher at the 30 end (e.g.,
FRET) (55). For both types, probes bind directly to the target
sequence being amplified and the fluorescent products are
detected by the PCR instrument (55). Hydrolysis probes
(e.g., TaqMan, HyBeacons, and Eclipse) rely on the exonu-
clease activity of Taq to hydrolyze the probe, releasing the
fluorophore for detection (Fig. 2) (55). In contrast, a molec-
ular beacon begins as a self-hybridized hairpin structure with
the fluorophore and quencher located in close proximity.
In this hairpin conformation the beacon is self-quenched.
2.3.2-4 ▪ TARGET-SPECIFIC DETECTION
FIGURE 2 Oligonucleotide specific probes: hydrolysis probes (a) and molecular beacons (b). doi:10.1128/9781555818821.ch2.3.2.f2
With heat denaturation of the PCR product, the beacon also
denatures (but remains intact). As annealing and polymeri-
zation occurs, the beacon hybridizes to the target sequence,
and fluorescence is emitted because the fluorophore is now
distal to the quencher (Fig. 2).
Multiplexed Fluorescent PCR Techniques
Oligonucleotide probe-based detection methods can provide
multiplexed detection depending on the real-time PCR
instrument used. Many fluorescent reporters are available
for fluorescent background subtraction such as ROX, or
detectors such as FAM, Cy3, CAL Fluor Red 590, and
quenchers TAMRA or DABCYL. It is critical to select dye(s)
with distinct emission and excitation spectra to ensure dis-
crete signal from a given primer/probe set. Most real-time
thermal cyclers are capable of detecting signal from four chan-
nels simultaneously and can be configured to detect up to four
different gene targets in a single reaction.
As discussed previously, a new application of LATE-PCR
allows for more deeply multiplexed real-time detection. This
procedure combines the benefits of FRET-based probes (e.g.,
reaction followed in real time) with high-resolution melt-
curve analysis. Rice et al. reported that 17 different pathogens
associated with sepsis could be detected in a single reaction
(60). In the LATE-PCR technique, these authors designed
primers and probes that had different melting temperatures
within each fluorescent channel (example: FAM, CAL
Orange, CAL Red, and Quasar). Using LATE-PCR, they
were able to analyze results using melt-curve analysis to distin-
guish pathogens within channels and between channels,
allowing an optimized multiplex of 17 different gene targets
in a single qPCR reaction. Limits of detection varied from
10 to 1,000 genome equivalents.
ABSOLUTE AND RELATIVE PCR
QUANTIFICATION
One of the benefits of fluorescent PCR reactions that can be
followed in real time is the ability to generate quantitative
data. There are two general types of quantitative PCR meth-
ods: absolute and relative quantification. Absolute quanti-
fication provides information in relation to counts per unit
(i.e., copies per genome/DNA), whereas relative quantifica-
tion measures mRNA expression (38, 50). Examples of using
absolute quantification to detect pathogens in both food and
environmental matrices exist in the literature (61–64). In
contrast, relative quantification is typically used to assess
the cellular response to different treatment conditions by
examining the level of transcription of one or more target
genes relative to a control or “housekeeping” gene. This latter
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-5
type of quantification is often used in clinical and microbial
samples (65–68).
With quantitative PCR, the exponential phase of the
reaction is much less variable, and this feature allows quanti-
fication during a qPCR experiment (53, 69). The key variable
for quantification is the time when the reaction enters the
exponential phase and crosses a threshold fluorescent inten-
sity that is statistically distinguishable from background.
The cycle threshold (CT) also referred to as Cq (quantifica-
tion cycle) is defined as the PCR cycle number where the
accumulated fluorescence crosses a user defined threshold
beyond background levels (70, 71). The y axis is the delta
RN (change in fluorescence), and the x axis is the PCR cycle
number (Fig. 3). It is a typical practice to set threshold at the
approximate mid-point of the exponential amplification
phase. As can be seen in Fig. 3, the threshold was set at a
ΔRN = 0.1. With a logarithmic scale this means that the
threshold was set at approximately 10–100-fold greater than
background. The intersection of the cycle number (x axis)
and threshold is the CT or Cq value. A smaller CT value indi-
cates a higher copy number of the starting material, whereas a
larger CT value indicates less material (72). Because the
threshold is user defined, it is important that the threshold
for positive detection be clearly stated in the results.
In theory, the absolute quantification method provides
the user with a total count of DNA (e.g., gene copy number)
or RNA (e.g., amount of transcript) in a given sample,
whether performing qPCR or RT-qPCR, respectively. Abso-
lute quantification uses a standard curve to estimate the total
FIGURE 3 Typical plot of a real-time PCR reaction. (1) Plateau phase where critical reactants are exhausted, (2) late exponential/linear
phase where reactants become limited, (3) exponential phase, and (4) pre-exponential phase. doi:10.1128/9781555818821.ch2.3.2.f3
concentration. Standards are derived from total nucleic acid;
specific DNA sequences cloned into plasmids are typical
standards for qPCR, and in vitro RNA transcripts, generated
from cloned DNA, are often utilized as standards for
RT-qPCR (38, 49). In the case of in vitro transcripts, >2 kb
transcripts are recommended to provide more biological rele-
vance (50). Correctly applying a normalization strategy is the
main challenge in using qPCR quantification (49, 73).
Briefly, serial dilutions (10-fold) of the user-selected
standards are amplified; the resulting CT values are used
to generate a standard (or “calibration”) curve (Fig. 4a–b)
(38, 50). By graphing the log of the known standard
concentration/copy number against the resultant CT value,
a standard curve is generated, and the user can determine
the reaction efficiency for a given set of samples (49, 69).
Reaction efficiency is measured by the following equation:
 
Efficiency ¼ 10(1=slope line)  1,
where the slope is determined by linear regression (74). Effi-
ciencies between 90% and 110% are generally considered
acceptable, but this may not always be possible due to a num-
ber of factors beyond the researcher’s control (e.g., difficult to
amplify templates and incomplete removal of PCR inhibitors
from environmental samples). In this case, it is critical to
ensure that dilutions of the unknown are run with dilutions
of the standard, and that the PCR efficiency of the unknowns
is equivalent to the PCR efficiency of the standards. Other-
wise, quantification will be inaccurate.
 2.3.2-6 ▪ TARGET-SPECIFIC DETECTION
FIGURE 4 Calculation of the standard curve for absolute quantification. Dilutions of known concentration are run (a), and the CT values are plotted (b) as a function of quantity
(x axis, logarithmic) and CT value (y axis). doi:10.1128/9781555818821.ch2.3.2.f4
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-7
Real-time PCR will accommodate four to five orders of
magnitude differences in starting template concentration,
but the variance increases for lower starting concentrations
(e.g., <1,000 copies) (73). If plasmid or other DNA-based
standard are used to estimate mRNA copy number, it is
important to note that the standard has not undergone the
same sample preparation (i.e., reverse transcription) and
thus the standard is limited to semi-quantitative estimates
(38, 50). If the user decides to use an RNA standard for the
RT-qPCR reaction, the efficiency of the reaction must be
similar to that of the unknown to be considered a “valid
standard for mRNA quantification” (50). Interpreting “abso-
lute” quantification needs to be considered in terms of the
starting material (50).
There are two basic approaches to relative quantification,
the ΔΔCT and Pfaffl methods, that may be used to mea-
sure the amount of RNA transcript in a given sample.
When the gene amplification efficiencies between the
target and the calibrant vary widely, the Pfaffl method may
be more advisable, because the efficiencies of both target
and calibrant are assumed to be the same in the ΔΔCT
method (53).
The ΔΔCT method provides the relative change in expres-
sion between an experimental group and a control, where a
control group and expression of the selected housekeeping
gene in this group can serve as the calibrant (e.g., housekeep-
ing gene) for signal normalization (75). The first calculation
is the ΔCT, which is the difference in CT values between
the target and the calibrant gene in the control group (75).
The same calculation is performed for the difference in CT
between the target and the calibrant gene in the treated
group. The corresponding difference between these changes
in CT value (ΔΔCT) represents an approximate fold change
of the gene of interest. To compare experimental and control
groups, the difference (ΔΔCT) is determined using the CT
of the target and calibrant genes of both the control and
the treated groups. The following calculation is performed
in Fig. 5 (75):
DDCt ¼ (Ct(target)  Ct(calibrant) )experimental

ably minimizes the interaction of PCR reaction constituents
 Ct(target)  Ct(calibrant) control:
The Pfaffl method provides a relative comparison of the level
of transcript (mRNA) expression between the gene
of interest (target) in a given sample(s) and the calibrant
(housekeeping or standard gene) (38, 76). This method
assumes there are at least two comparison groups, a control
(nontreated) and a treated group (e.g., pharmaceutically
treated). The reaction efficiencies (Etarget or Ereference) of
both the reference and the target genes are calculated
in a similar manner as absolute quantification methods
(76). For relative quantification, the next step is to measure
expression of the target and reference gene between dif-
ferent groups (e.g., ΔCT of control versus sample) followed
by similar calculations for the reference gene. The ratio of
these calculations yields the relative efficiency, calculated
by (76):

DC Target ½gene (controlsample [groups])
Ratio ¼ Etarget t
 ðEreference ÞDC
t
Reference ½gene (controlsample [groups])
DIGITAL PCR
Digital PCR (dPCR), first described by Vogelstein and Kin-
zler, is an alternative qPCR method (29). Recent technology
developments have made this methodology commercially
available. The components and cycling conditions of digital
PCR are identical to qPCR and in theory, any assay that
has been optimized for a qPCR platform can be migrated
to dPCR (77). The primary difference between dPCR and
qPCR lies in the methods used to calculate copy number.
In dPCR, a sample is divided into hundreds to thousands of
subreactions ( picoliter to nanoliter volumes) where each sub-
reaction has, in theory, zero or one target molecules. PCR
reactions are carried out as normal, and if a target molecule
is present successful amplification is detected using a fluores-
cent probe. If no template is present, amplification does not
occur. The reactions are subsequently tallied as a binomial
result (yes/no) and the proportion of detection events is
evaluated relative to the Poisson distribution to estimate
the starting template concentration.
In practice, at very low target concentrations each positive
reaction is likely to represent a single starting molecule, but
at higher initial concentrations it is likely there are more
than one target molecules per reaction (28, 78). Conse-
quently, the original copy number can be estimated using
the Poisson distribution (28, 78): A = −ln(1 − P), where A
is the average number of molecules per subreaction, and P is
the proportion of all subreactions that are positive by PCR.
Most dPCR platforms perform this calculation for the user,
regardless of total number of target molecules, and report
the results per user preference.
To date, dPCR has been used primarily to estimate copy
number variation for cancer genes, sickle cell disease, and
detection of low concentrations of opportunistic pathogens
in clinical samples (79–81). Environmental applications of
dPCR are beginning to be described (77, 82–84). Hoshino
and Inagaki employed dPCR to quantify DNA in soils, and
from their studies they concluded that dPCR performed sig-
nificantly better than qPCR (84). There could be several rea-
sons we might expect this outcome. First, dPCR subdivides
the samples into hundreds to thousands of subreactions,
thus minimizing potential amplification bias and/or non-
target DNA competition in a sample. Second, dPCR prob-
with PCR inhibitors (chapter 2.3.4). Importantly, unlike
qPCR, the efficiency of dPCR does not influence the cal-
culation for target molecule concentration, provided that
amplification occurs, and that amplification event can be
detected (82, 83, 85). Digital PCR has the potential to be a
powerful tool for environmental microbiology, but its use
will have similar caveats to other applications of PCR, includ-
ing the need for careful standardization.
ISOTHERMAL AMPLIFICATION
Another modification of conventional PCR is isothermal
amplification. With the commercialization of thermal cyclers,
most scientists are conducting PCR daily across the world.
Not all research and diagnostic labs have this capability,
however, either due to budget, space constraints, or lack of
power during fieldwork. A thermal cycler is not needed
for isothermal amplification, and this approach permits
cost-effective and relatively rapid assays, which at times also
demonstrate increased sensitivity (26, 86). Isothermal
amplification has been implemented in the areas of forensics,
environmental detection, basic research, and point-of-care
clinical diagnostics (87–90).
There are many different isothermal methods, and this
continues to be a rapidly changing area of nucleic acid ampli-
fication (Table 1). First reported by Notomi et al., loop-
2.3.2-8 ▪ TARGET-SPECIFIC DETECTION

FIGURE 5 Delta-delta relative PCR quantitation method. A basic overview of the ΔΔCT is shown. The overall difference in CT between
expression within (ΔCT) and between (ΔΔCT) is shown above. In this example, the calibrant/reference gene is compared to the target gene of
interest in the control (a) and the experimental (b) groups. doi:10.1128/9781555818821.ch2.3.2.f5

mediated isothermal amplification (LAMP) is a highly
specific and efficient approach and is one of the most widely
used and published isothermal techniques for nucleic acid
detection (95). Specificity is accomplished using a set of
four primers targeting six unique sequences on the target
nucleic acid (95). A cauliflower-like structure is synthe-
sized through primer binding and DNA synthesis with a
polymerase that has strand displacement activity, typically
Bst polymerase (source: Bacillus stearothermophilus) (95).
For RNA, a reverse transcriptase is also required (96). The
cauliflower-like structure is made of inverted repeats and
loops that are further amplified, allowing rapid synthesis
of 109 copies of target DNA in 30–60 min (95). In the orig-
inal report, DNA analysis or detection was accomplished
using gel electrophoresis; advances in detector technology
and microfluidics now allow for the quantification of pyro-
phosphate ions and fluorescent probes (97–100). LAMP
has been used to detect DNA and RNA from various targets
such as HIV-1, aflatoxigenic molds in foods, and bacterial
pathogens (101–103).
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-9
TABLE 1 Isothermal amplification methods and reviews
Method name
Circular helicase dependent amplification
Exponential amplification reaction
Helicase-dependent amplification
Isothermal and chimeric primer-initiated amplification
Isothermal multiple displacement amplification
Loop-mediated isothermal amplification of DNA
Nucleic acid sequence-based amplification
Self-sustained sequence replication
Signal mediate amplification of RNA technology
Single primer isothermal amplification
Strand displacement amplification
Recombinase polymerase amplification
Rolling circle amplification
Transcription mediated amplification
The breakthrough in RNA isothermal amplification
was accomplished through a technique called nucleic acid
sequence-based amplification (NASBA) (27). NASBA
requires a single-stranded template, most often single-
stranded RNA, but double-stranded nucleic acids can be
used if they are converted to single-strand moiety (27). By
exploiting the retroviral strategy for RNA replication, three
enzymes are used to amplify target sequence: reverse tran-
scriptase, RNase H, and a DNA-dependent RNA polymerase
(27, 104). Amplification initiates with the binding of the
forward primer that incorporates a T7 promoter (a nucleic
acid sequence that has a high affinity for T7 RNA polymer-
ase) on the positive-strand template RNA. The reverse tran-
scriptase synthesizes a cDNA, forming a cDNA-RNA hybrid.
RNase H degrades the RNA from the RNA:DNA duplex,
and the second primer anneals to the cDNA strand. The
double-stranded cDNA is then used as a template for the
T7 DNA-dependent RNA polymerase, generating comple-
mentary RNA that can then be used as a template in sub-
sequent rounds of amplification (27, 92). The subsequent
rounds of amplification have been referred to as self-sustained
sequence replication (104). NASBA generates a pool of
single-stranded RNA (negative-sense) molecules that can
be used for probe-based hybridization. Target amplification
can be detected using gel electrophoresis or in real time with
molecular beacons (sequence-specific fluorescent probes)
(105). Several companies have commercialized detection
approaches using NASBA. BioMérieux has successfully
commercialized a NASBA platform coupled with molecular
beacons (NucliSENS) to detect hepatitis B virus, herpes sim-
plex virus, and methicillin-resistant Staphylococcus aureus
(89). The approach used to detect these organisms relies on
amplification of DNA targets that are digested by restriction
enzymes prior to amplification with NASBA (89). Other
targets include influenza A H5N1, severe acute respiratory
syndrome, and waterborne pathogens (106–108).
By exploiting the mechanism of in vivo rolling circle DNA
replication, the isothermal technique called rolling circle
amplification (RCA) was developed to detect DNA and
RNA targets (109, 110). Using bacteriophage ɸ29 DNA pol-
ymerase, a primer is extended continuously in the circle,
thereby generating multiple copies of the target sequence in
the form of a single-stranded concatemer. By using multiple
Acronym Reference(s)
cHDA 91
EXPAR 92
HDA 91, 93, 94
ICANs 92
IMDA 91, 92
LAMP 91, 92, 94
NASBA 91, 92, 94
3SR 91
SMART 91, 92
SPIA 91
SDA 91–94
RPA 93
RCA 91–94
TMA 91
primers in a technique called multiply primed RCA, greater
amplification of the target occurs (111). For maximum
amplification efficiency, the target molecule should be circu-
lar, single-stranded, and relatively small. A padlock probe is a
single-stranded oligonucleotide with complementary 50 and
30 ends to the target sequence and an intervening linker
sequence. When the padlock probe hybridizes to the target,
it is ligated to generate a single-stranded circular molecule
that is used as a substrate for RCA amplification (112).
The use of padlock probes increases specificity and reduces
background (112).
The major advantage of isothermal DNA amplification
is that there is no need for accurate and often expensive ther-
mal cyclers. Without this instrumentation constraint, small
field-portable isothermal amplification devices are being
developed that are fast, accurate, and increasingly sensitive.
Despite this advantage, isothermal amplification still requires
enzymes for the reaction and often has complicated primer
design. Through creative and thoughtful assay develop-
ment, however, scientists will continue to improve existing
approaches and develop novel isothermal techniques for
rapid, accurate, and low-cost detection of nucleic acids in a
variety of samples and environments.
STANDARD FOR REPORTING PCR RESULTS
When developing assays for qPCR and RT-qPCR, one of
the challenges has been the consistency of nomenclature,
protocol, and analysis parameters. As a result, Bustin et al.
compiled a set of guidelines, called the MIQE (Minimum
Information for Publication of Quantitative Real-Time
PCR Experiments) (113). MIQE is based on other well-
known standards such as Minimum Information About a
Microarray Experiment and Minimum Information About
a Proteomics Experiment. MIQE is a checklist of “expected
information” and “desired information” regarding the
experimental parameters of a qPCR experiment including
experimental design parameters, sample description and
preparation methods, nucleic acid extraction procedure,
PCR primer and probe sequence, PCR conditions, and how
the data analysis was performed. The MIQE checklist is
extensive, but its adoption has been limited (114–117), espe-
cially for the environmental microbiology community (113).
2.3.2-10 ▪ TARGET-SPECIFIC DETECTION
THE FUTURE OF PCR FOR ENVIRONMENTAL
MICROBIOLOGY: FINAL THOUGHTS
The development and use of PCR in research has transformed
how we approach experiments and question community
structure and diversity within environmental samples. The
ability to identify microbial members within a sample, their
relative ratio, and biological activity has transformed our
understanding of complex samples such as microbiomes,
biofilms, and microbial mats. Domain level identification of
bacterial members is possible with PCR using conserved
16S rRNA targets, and species-level or better identification
can be accomplished with hypervariable regions or alterna-
tive gene sequences. Quantification with qPCR provides
data on the relative change of community member number
and composition within diverse sample types across the
world. RT-qPCR can be used to demonstrate metabolic
activity and allow inferences to be made about community
member roles and function. Finally, PCR is proving to be
a powerful tool for the detection of dormant or novel
members in complex communities such as microbial mat
and biofilms. Areas of future improvement for PCR method-
ologies include development of efficient preparation tech-
niques from complex and inhibitor-rich environmental
samples. Additional areas of improvement include improved
sensitivity of rare or novel targets, rapid assay optimization,
and universal quantification approaches for direct com-
parisons of independent data sets. Continued advancement
and development of available reagents and instrumen-
tation will improve nucleic acid detection and quanti-
fication methods in the areas of speed, sensitivity, and
specificity, thus expanding our understanding of microbial
communities.
A portion of this effort was funded by the Department of Homeland
Security Science and Technology Directorate under Contract
HSHQDC-08-X-00843. Pacific Northwest National Laboratory is oper-
ated by Battelle for the U.S. Department of Energy under contract
DE-AC06–76RLO, and by the College of Veterinary Medicine, Wash-
ington State University.
REFERENCES
1. Mullis KB. 1990. The unusual origin of the polymerase
chain reaction. Sci Am 262:56–61, 64–65.
2. Mullis KB. 2009. Polymerase chain reaction. www.kary
mullis.com/pcr.shtml.
3. Mullis KB. 1997. Nobel lecture, December 8, 1993: the
polymerase chain reaction. In Malmström BG (ed), Nobel
Lectures, Chemistry 1991–1995. World Scientific Publish-
ing Co., Singapore.
4. Scharf SJ, Horn GT, Erlich HA. 1986. Direct cloning and
sequence analysis of enzymatically amplified genomic
sequences. Science 233:1076–1078.
5. Wrischnik LA, Higuchi RG, Stoneking M, Erlich HA,
Arnheim N, Wilson AC. 1987. Length mutations in
human mitochondrial DNA: direct sequencing of enzy-
matically amplified DNA. Nucleic Acids Res 15:529–542.
6. Wong C, Dowling CE, Saiki RK, Higuchi RG, Erlich
HA, Kazazian HH Jr. 1987. Characterization of beta-
thalassaemia mutations using direct genomic sequencing
of amplified single copy DNA. Nature 330:384–386.
7. Kwok S, Mack DH, Mullis KB, Poiesz B, Ehrlich G, Blair
D, Friedman-Kien A, Sninsky JJ. 1987. Identification of
human immunodeficiency virus sequences by using in vitro
enzymatic amplification and oligomer cleavage detection.
J Virol 61:1690–1694.
8. Darawi MN, Ai-Vyrn C, Ramasamy K, Hua PP, Pin TM,
Kamaruzzaman SB, Majeed AB. 2013. Allele-specific pol-
ymerase chain reaction for the detection of Alzheimer’s
disease-related single nucleotide polymorphisms. BMC
Med Genet 14:27.
9. Yin Q, Fu B, Li B, Shi X, Inagaki F, Zhang XH. 2013.
Spatial variations in microbial community composition
in surface seawater from the ultra-oligotrophic center to
rim of the South pacific gyre. PLoS One 8:e55148.
10. Pillai MM, Latha R, Sarkar G. 2012. Detection of methi-
cillin resistance in Staphylococcus aureus by polymerase
chain reaction and conventional methods: a comparative
study. J Lab Phys 4:83–88.
11. Roux KH. 2009. Optimization and troubleshooting in
PCR. Cold Spring Harb Protoc 2009:pdb ip66.
12. Mullis KB, Faloona F. 1987. Specific synthesis of DNA in
vitro via a polymerase-catalyzed chain reaction. Meth Enzy-
mol 155:335–50.
13. Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R,
Horn GT, Mullis KB, Erlich HA. 1988. Primer-directed
enzymatic amplification of DNA with a thermostable
DNA polymerase. Science 239:487–491.
14. Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H.
1986. Specific enzymatic amplification of DNA in vitro:
the polymerase chain reaction. Cold Spring Harb Symp
Quant Biol 51(Pt 1):263–273.
15. Sanchez JA, Pierce KE, Rice JE, Wangh LJ. 2004.
Linear-after-the-exponential (LATE)-PCR: an advanced
method of asymmetric PCR and its uses in quantitative
real-time analysis. Proc Natl Acad Sci USA 101:
1933–1938.
16. Lawyer FC, Stoffel S, Saiki RK, Myambo K,
Drummond R, Gelfand DH. 1989. Isolation, character-
ization, and expression in Escherichia coli of the DNA pol-
ymerase gene from Thermus aquaticus. J Biol Chem 264:
6427–6437.
17. Tindall KR, Kunkel TA. 1988. Fidelity of DNA synthesis
by the Thermus aquaticus DNA polymerase. Biochemistry
27:6008–6013.
18. Cline J, Braman JC, Hogrefe HH. 1996. PCR fidelity of
Pfu DNA polymerase and other thermostable DNA poly-
merases. Nucleic Acids Res 24:3546–3551.
19. Abu Al-Soud W, Radstrom P. 1998. Capacity of nine ther-
mostable DNA polymerases to mediate DNA amplification
in the presence of PCR-inhibiting samples. Appl Environ
Microbiol 64:3748–3753.
20. Rakshit S. 2010. Themostable enzymes used in polymerase
chain reaction. In Bustin S (ed), The PCR revolution: basic
techniques and applications. Cambridge University Press,
New York.
21. Apte A, Saurabha S. 2003. PCR primer design. In Dieffen-
bach C and Dveksler GS (eds), PCR Primer: a Laboratory
Manual. Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, NY.
22. Kwok S, Chang SY, Sninsky JJ, Wang A. 1994. A guide to
the design and use of mismatched and degenerate primers.
PCR Meth Appl 3:S39–S47.
23. Integrated DNA Technologies. Oligo modifications: sup-
port and educational content, accessed 1 October 2015.
http://www.idtdna.com/pages/decoded/decoded-articles/
oligo-modifications.
24. Gibson UE, Heid CA, Williams PM. 1996. A novel
method for real time quantitative RT-PCR. Genome Res
6:995–1001.
25. Heid CA, Stevens J, Livak KJ, Williams PM. 1996. Real
time quantitative PCR. Genome Res 6:986–994.
26. Boyanton BL, Sural P, Loomis CR, Pesta C, Gonzalez-
Krellwitz L, Robinson-Dunn B, Riska P. 2012. Loop-
mediated isothermal amplification compared to real-time
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-11
PCR and Enzyme immunoassay for toxigenic clostridium
difficile detection. J Clin Microbiol 50:640–645.
27. Compton J. 1991. Nucleic acid sequence-based amplifica-
tion Nature 350:91–92.
28. Dube S, Qin J, Ramakrishnan R. 2008. Mathematical
analysis of copy number variation in a DNA sample using
digital PCR on a nanofluidic device. PLoS ONE 3:e2876.
29. Vogelstein B, Kinzler KW. 1999. Digital PCR. Proc Natl
Acad Sci USA 96:9236–9241.
30. Kotewicz ML, D’Alessio JM, Driftmier KM, Blodgett KP,
Gerard GF. 1985. Cloning and overexpression of Moloney
murine leukemia virus reverse transcriptase in Escherichia
coli Gene 35:249–258.
31. Grandgenett DP, Gerard GF, Green M. 1973. A single
subunit from avian myeloblastosis virus with both RNA-
directed DNA polymerase and ribonuclease H activity.
Proc Natl Acad Sci USA 70:230–234.
32. Myers TW, Gelfand DH. 1991. Reverse transcription and
DNA amplification by a Thermus thermophilus DNA poly-
merase. Biochemistry 30:7661–7666.
33. Moser MJ, DiFrancesco RA, Gowda K, Klingele AJ,
Sugar DR, Stocki S, Mead DA, Schoenfeld TW. 2012.
Thermostable DNA polymerase from a viral metagenome
is a potent RT-PCR enzyme. PLoS One 7:e38371.
34. Life Technologies, RNA priming strategies, accessed 1
October 2015. http://www.invitrogen.com/site/us/en/home/
Products-and-Services/Applications/PCR/reverse-transcrip
tion/rna-priming-strategies.html
35. Nolan T, Hands RE, Bustin SA. 2006. Quantification of
mRNA using real-time RT-PCR. Nat Protoc 1:1559– 1582.
36. Deprez RHL, Fijnvandraat AC, Ruijter JM, Moorman
AFM. 2002. Sensitivity and accuracy of quantitative real-
time polymerase chain reaction using SYBR green I
depends on cDNA synthesis conditions. Anal Biochem
307:63–69.
37. Battaglia M, Pedrazzoli P, Palermo B, Lanza A, Bertolini
F, Gibelli N, Da Prada GA, Zambelli A, Perotti C, Robus-
telli della Cuna G. 1998. Epithelial tumour cell detection
and the unsolved problems of nested RT-PCR: a new
sensitive one step method without false positive results.
Bone Marrow Transplant 22:693–698.
38. Wong ML, Medrano JF. 2005. Real-time PCR for mRNA
quantitation. Biotechniques 39:75–85.
39. Bustin SA. 2002. Quantification of mRNA using real-time
reverse transcription PCR (RT-PCR): trends and prob-
lems. J Mol Endocrinol 29:23–29.
40. Chumakov KM. 1994. Reverse-transcriptase can inhibit
PCR and stimulate primer-dimer formation. PCR Meth
Appl 4:62–64.
41. Sellner LN, Coelen RJ, Mackenzie JS. 1992. Reverse tran-
scriptase inhibits Taq polymerase activity. Nucleic Acids Res
20:1487–1490.
42. Liss B. 2002. Improved quantitative real-time RT-PCR for
expression profiling of individual cells. Nucleic Acids
Research 30:e89.
43. Brooks EM, Sheflin LG, Spaulding SW. 1995. Secondary
structure in the 30 UTR of EGF and the choice of reverse
transcriptases affect the detection of message diversity by
RT-PCR. Biotechniques 19:806–815.
44. Stahlberg A, Kubista M, Pfaffl M. 2004. Comparison
of reverse transcriptases in gene expression analysis. Clin
Chem 50:1678–1680.
45. Stahlberg A, Hakansson J, Xian X, Semb H, Kubista M.
2004. Properties of the reverse transcription reaction in
mRNA quantification. Clin Chem 50:509–515.
46. Polumuri SK, Ruknudin A, Schulze DH. 2002. RNase H
and its effects on PCR. Biotechniques 32:1224–1225.
47. Swift GH, Peyton MJ, MacDonald RJ. 2000. Assessment
of RNA quality by semi-quantitative RT-PCR of multiple
regions of a long ubiquitous mRNA. Biotechniques 28:
524–531.
48. Mannhalter C, Koizar D, Mitterbauer G. 2000. Evalua-
tion of RNA isolation methods and reference genes for
RT-PCR analyses of rare target RNA. Clin Chem Lab
Med 38:171–177.
49. Kubista M, Andrade JM, Bengtsson M, Forootan A,
Jonak J, Lind K, Sindelka R, Sjoback R, Sjogreen B,
Strombom L, Stahlberg A, Zoric N. 2006. The real-time
polymerase chain reaction. Mol Aspects Med 27:
95–125.
50. Pfaffl M. 2004. Quantification strategies in real-time PCR.
In Bustin S (ed), A–Z of quantitative PCR. International
University Line, La Jolla, CA.
51. Applied Biosystems. Real-time PCR Vs. traditional PCR.
In Applied Biosystems (ed). http://www6.appliedbiosystems.
com/support/tutorials/pdf/rtpcr_vs_tradpcr.pdf.
52. Kappes JC, Saag MS, Shaw GM, Hahn BH, Chopra P,
Chen S, Emini EA, McFarland R, Yang LC, Piatak M
Jr., et al. 1995. Assessment of antiretroviral therapy by
plasma viral load testing: standard and ICD HIV-1 p24
antigen and viral RNA (QC-PCR) assays compared.
J Acquir Immune Defic Syndr Hum Retrovirol 10:139– 149.
53. Bustin S. 2004. Quantification of nucleic acids by PCR,
p. 3–46. In Bustin S (ed), A–Z of quantitative PCR. Interna-
tional University Line, La Jolla, CA.
54. Postollec F, Falentin H, Pavan S, Combrisson J, Sohier D.
2011. Recent advances in quantitative PCR (qPCR) appli-
cations in food microbiology. Food Microbiol 28:848–861.
55. Livak KJ, Flood SJ, Marmaro J, Giusti W, Deetz K. 1995.
Oligonucleotides with fluorescent dyes at opposite ends
provide a quenched probe system useful for detecting
PCR product and nucleic acid hybridization. PCR Meth
Appl 4:357–362.
56. Ishiguro T, Saitoh J, Yawata H, Yamagishi H, Iwasaki S,
Mitoma Y. 1995. Homogeneous quantitative assay of hep-
atitis C virus RNA by polymerase chain reaction in the
presence of a fluorescent intercalater. Anal Biochem 229:
207–213.
57. Holland PM, Abramson RD, Watson R, Gelfand DH.
1991. Detection of specific polymerase chain reaction
product by utilizing the 50 –30 exonuclease activity of
Thermus aquaticus DNA polymerase. Proc Natl Acad Sci
USA 88:7276–7280.
58. Ririe KM, Rasmussen RP, Wittwer CT. 1997. Product
differentiation by analysis of DNA melting curves during
the polymerase chain reaction. Anal Biochem 245:154–
160.
59. Bustin S, Nolan T. 2004. Primers and probes. In Bustin S
(ed), A–Z of quantitative PCR. International University
Line, La Jolla, CA.
60. Rice LM, Reis AH, Ronish B, Carver-Brown RK, Czajka
JW, Gentile N, Kost G, Wangh LJ. 2013. Design of a
single-tube, endpoint, linear-after-the-exponential-PCR
assay for 17 pathogens associated with sepsis. J Appl Micro-
biol 114:457–469.
61. Navas J, Ortiz S, Lopez P, Jantzen MM, Lopez V,
Martinez-Suarez JV. 2006. Evaluation of effects of primary
and secondary enrichment for the detection of Listeria
monocytogenes by real-time PCR in retail ground chicken
meat. Foodb Path Dis 3:347–354.
62. Mokhtari W, Nsaibia S, Gharbi A, Aouni M. 2013.
Real-time PCR using SYBR Green for the detection of Shi-
gella spp. in food and stool samples. Mol Cell Probes 27:
53–59.
63. Zhang L, Wang X, Zhang Y, Gong L, Mao H, Feng C,
Ojcius DM, Yan J. 2012. Rapid and sensitive identification
of RNA from the emerging pathogen, coxsackievirus A6.
Virol J 9:298.
2.3.2-12 ▪ TARGET-SPECIFIC DETECTION
64. Ma Y, Yang Y, Lv M, Yan Q, Zheng L, Ding F, Wu J, Tian
K, Zhang J. 2010. Real-time quantitative polymerase chain
reaction with SYBR green i detection for estimating copy
numbers of porcine endogenous retrovirus from Chinese
miniature pigs. Transpl Proc 42:1949–1952.
65. Anitha A, Nakamura K, Thanseem I, Yamada K,
Iwayama Y, Toyota T, Matsuzaki H, Miyachi T, Yamada
S, Tsujii M, Tsuchiya KJ, Matsumoto K, Iwata Y,
Suzuki K, Ichikawa H, Sugiyama T, Yoshikawa T, Mori
N. 2012. Brain region-specific altered expression and
association of mitochondria-related genes in autism. Mol
Autism 3:12.
66. Zhu YN, Lu SM, You JF, Zhu B, Yu MY. 2009. Novel
real-time PCR assay for rapid prenatal diagnosis of Down
syndrome: a prospective study of 563 amniocytes. Clin Bio-
chem 42:672–675.
67. Marco ML, Kleerebezem M. 2008. Assessment of real-time
RT-PCR for quantification of Lactobacillus plantarum gene
expression during stationary phase and nutrient starvation.
J Appl Microbiol 104:587–594.
68. Pholwat S, Heysell S, Stroup S, Foongladda S, Houpt E.
2011. Rapid first- and second-line drug susceptibility assay
for Mycobacterium tuberculosis isolates by use of quantitative
PCR. J Clin Microbiol 49:69–75.
69. Bustin S. 2000. Absolute quantification of mRNA using
real-time reverse transcription polymerase chain reaction
assays. J Mol Endocrinol 25:169–193.
70. Bustin SA, Nolan T. 2004. Pitfalls of quantitative real-
time reverse-transcription polymerase chain reaction.
J Biomol Tech 15:155–166.
71. D’Haene B, Vandesompele J, Hellemans J. 2010. Accu-
rate and objective copy number profiling using real-time
quantitative PCR. Methods 50:262–270.
72. Higuchi R, Fockler C, Dollinger G, Watson R.
1993. Kinetic PCR analysis: real-time monitoring of
DNA amplification reactions. Biotechnol (NY) 11:
1026–1030.
73. Bustin SA, Benes V, Nolan T, Pfaffl MW. 2005. Quanti-
tative real-time RT-PCR—a perspective. J Mol Endocrinol
34:597–601.
74. Rasmussen RP. 2001. Quantification on the lightcycler.
In: Meuer S, Wittwer CT, Nakagawara K (eds), Rapid
cycle real-time PCR, methods and applications. Spring Press,
Heidelberg.
75. Livak KJ, Schmittgen TD. 2001. Analysis of relative
gene expression data using real-time quantitative PCR
and the 2(-delta delta C(T)) method. Methods 25:
402–408.
76. Pfaffl MW. 2001. A new mathematical model for relative
quantification in real-time RT-PCR. Nucleic Acids Res
29:e45.
77. Straub T, Baird C, Bartholomew RA, Colburn H, Seiner
D, Victry K, Zhang L, Bruckner-Lea CJ. 2013. Estimated
copy number of Bacillus anthracis plasmids pXO1 and pXO2
using digital PCR. J Microbiol Meth 92:9–10.
78. Day E, Dear PH, McCaughan F. 2012. Digital PCR strat-
egies in the development and analysis of molecular bio-
markers for personalized medicine. Methods 59:101–107.
doi: 10.1016/j.ymeth.2012.08.001.
79. Azuara D, Ginesta MM, Gausachs M, Rodriguez-
Moranta F, Fabregat J, Busquets J, Pelaez N, Boadas J,
Galter S, Moreno V, Costa J, de Oca J, Capellá G.
2012. Nanofluidic digital PCR for KRAS mutation detec-
tion and quantification in gastrointestinal cancer. Clin
Chem 58:1332–1341.
80. Barrett AN, McDonnell TCR, Chan KCA, Chitty LS.
2012. Digital PCR analysis of maternal plasma for noninva-
sive detection of sickle cell anemia. Clin Chem 58:
1026–1032.
81. Schell W, Benton J, Smith P, Poore M, Rouse J, Boles D,
Johnson M, Alexander B, Pamula V, Eckhardt A, Pollack
M, Benjamin D, Perfect J, Mitchell T. 2012. Evaluation of
a digital microfluidic real-time PCR platform to detect
DNA of Candida albicans in blood. Eur J Clin Microbiol
Infect Diseases 31:2237–2245.
82. Burns M, Burrell A, Foy C. 2010. The applicability of dig-
ital PCR for the assessment of detection limits in GMO
analysis. Eur Food Res Technol 231:353–362.
83. Corbisier P, Bhat S, Partis L, Rui Dan Xie V, Emslie K.
2010. Absolute quantification of genetically modified
MON810 maize (Zea mays L.) by digital polymerase chain
reaction. Anal Bioanal Chem 396:2143–2150.
84. Hoshino T, Inagaki F. 2012. Molecular quantification
of environmental DNA using microfluidics and digital
PCR. Syst Appl Microbiol 35:390–395.
85. Bhat S, Herrmann J, Armishaw P, Corbisier P, Emslie K.
2009. Single molecule detection in nanofluidic digital
array enables accurate measurement of DNA copy number.
Anal Bioanal Chem 394:457–467.
86. Techathuvanan C, Draughon FA, D’Souza DH.
2011. Comparison of reverse transcriptase PCR, reverse
transcriptase loop-bediated isothermal amplification, and
culture-based assays for Salmonella detection from pork
processing environments. J Food Protect 74:294–301.
87. Khan M, Bhaskar K, Salam M, Akther T, Pluschke G,
Mondal D. 2012. Diagnostic accuracy of loop-mediated iso-
thermal amplification (LAMP) for detection of Leishmania
DNA in buffy coat from visceral leishmaniasis patients. Par-
asites Vectors 5:280.
88. Mori Y, Notomi T. 2009. Loop-mediated isothermal
amplification (LAMP): a rapid, accurate, and cost-effective
diagnostic method for infectious diseases. J Infect Chemo-
ther 15:62–69.
89. Deiman B, Jay C, Zintilini C, Vermeer S, van Strijp D,
Venema F, van de Wiel P. 2008. Efficient amplification
with NASBA® of hepatitis B virus, herpes simplex virus
and methicillin resistant Staphylococcus aureus DNA.
J Virol Meth 151:283–293.
90. Nakanishi H, Ohmori T, Hara M, Takada A, Shojo H,
Adachi N, Saito K. 2011. A simple identification method
of saliva by detecting Streptococcus salivarius using loop-
mediated isothermal amplification. J Forensic Sci 56:
S158–S161.
91. Gill P, Ghaemi A. 2008. Nucleic acid isothermal amplifi-
cation technologies-a review. Nucleos Nucleot Nucleic Acids
27:224–243.
92. Asiello PJ, Baeumner AJ. 2011. Miniaturized isothermal
nucleic acid amplification, a review. Lab on a Chip 11:
1420–1430.
93. Kim J, Easley CJ. 2011. Isothermal DNA amplification
in bioanalysis: strategies and applications. Bioanalysis 3:
227–239.
94. Chang C-C, Chen C-C, Wei S-C, Lu H-H, Liang Y-H,
Lin C-W. 2012. Diagnostic devices for isothermal nucleic
acid amplification. Sensors 12:8319–8337.
95. Notomi T, Okayama H, Masubuchi H, Yonekawa T,
Watanabe K, Amino N, Hase T. 2000. Loop-mediated
isothermal amplification of DNA. Nucleic Acids Res
28:e63.
96. Parida M, Posadas G, Inoue S, Hasebe F, Morita K. 2004.
Real-time reverse transcription loop-mediated isother-
mal amplification for rapid detection of West Nile virus.
J Clin Microbiol 42:257–263.
97. Mori Y, Kitao M, Tomita N, Notomi T. 2004. Real-time
turbidimetry of LAMP reaction for quantifying template
DNA. J Biochem Bioph Meth 59:145–157.
98. Mori Y, Nagamine K, Tomita N, Notomi T. 2001. Detec-
tion of loop-mediated isothermal amplification reaction
 2.3.2. PCR, Real-Time PCR, Digital PCR, and Isothermal Amplification ▪ 2.3.2-13
by turbidity derived from magnesium pyrophosphate for-
mation. Biochem Bioph Res Co 289:150–154.
99. Mori Y, Hirano T, Notomi T. 2006. Sequence specific
visual detection of LAMP reactions by addition of cationic
polymers. BMC Biotechnol 6:3.
100. Tomita N, Mori Y, Kanda H, Notomi T. 2008.
Loop-mediated isothermal amplification (LAMP) of gene
sequences and simple visual detection of products. Nat
Prot 3:877–882.
101. Curtis KA, Rudolph DL, Nejad I, Singleton J, Beddoe A,
Weigl B, LaBarre P, Owen SM. 2012. Isothermal amplifi-
cation using a chemical heating device for point-of-care
detection of HIV-1. PLoS One 7:e31432.
102. Luo J, Vogel RF, Niessen L. 2012. Development and
application of a loop-mediated isothermal amplification
assay for rapid identification of aflatoxigenic molds and
their detection in food samples. Int J Food Microbiol 159:
214–224.
103. Huy NT, Hang le TT, Boamah D, Lan NT, Van Thanh
P, Watanabe K, Huong VT, Kikuchi M, Ariyoshi K, Mor-
ita K, Hirayama K. 2012. Development of a single-tube
loop-mediated isothermal amplification assay for detection
of four pathogens of bacterial meningitis. FEMS Microbiol
Lett 337:25–30.
104. Guatelli JC, Whitfield KM, Kwoh DY, Barringer KJ,
Richman DD, Gingeras TR. 1990. Isothermal, in vitro
amplification of nucleic acids by a multienzyme reaction
modeled after retroviral replication. Proc Natl Acad Sci
87:1874–1878.
105. Leone G, van Gemen B, Schoen CD, van Schijndel H,
Kramer FR. 1998. Molecular beacon probes combined
with amplification by NASBA enable homogeneous,
real-time detection of RNA. Nucleic Acids Res 26:
2150–2155.
106. Moore C, Telles J-N, Corden S, Gao R-B, Vernet G, Van
Aarle P, Shu Y-L. 2010. Development and validation of a
commercial real-time NASBA assay for the rapid confir-
mation of influenza A H5N1 virus in clinical samples.
J Virol Meth 170:173–176.
107. Zhao X, Dong T, Yang Z, Pires N, Hoivik N. 2012. Com-
patible immuno-NASBA LOC device for quantitative
detection of waterborne pathogens: design and validation.
Lab on a Chip 12:602–612.
108. Keightley MC, Sillekens P, Schippers W, Rinaldo C,
George KS. 2005. Real-time NASBA detection of
SARS-associated coronavirus and comparison with
real-time reverse transcription-PCR. J Med Virol 77:
602–608.
109. Fire A, Xu SQ. 1995. Rolling replication of short DNA
circles. Proc Nat Acad Sci 92:4641–4645.
110. Lizardi PM, Huang X, Zhu Z, Bray-Ward P, Thomas DC,
Ward DC. 1998. Mutation detection and single-molecule
counting using isothermal rolling-circle amplification.
Nat Genet 19:225–232.
111. Dean FB, Nelson JR, Giesler TL, Lasken RS. 2001.
Rapid amplification of plasmid and phage DNA using
Phi29 DNA polymerase and multiply-primed rolling circle
amplification. Genome Res 11:1095–1099.
112. Nilsson M, Malmgren H, Samiotaki M, Kwiatkowski M,
Chowdhary B, Landegren U. 1994. Padlock probes:
circularizing oligonucleotides for localized DNA detection.
Science 265:2085–2088.
113. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett
J, Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley
GL, Vandesompele J, Wittwer CT. 2009. The MIQE
guidelines: minimum information for publication of quan-
titative real-time PCR experiments. Clin Chem 55:
611–622.
114. Gabert J, Beillard E, van der Velden VH, Bi W,
Grimwade D, Pallisgaard N, Barbany G, Cazzaniga G,
Cayuela JM, Cave H, Pane F, Aerts JL, De Micheli D,
Thirion X, Pradel V, Gonzalez M, Viehmann S,
Malec M, Saglio G, van Dongen JJ. 2003. Standardization
and quality control studies of “real-time” quantitative
reverse transcriptase polymerase chain reaction of
fusion gene transcripts for residual disease detection in leu-
kemia—a Europe Against Cancer program. Leukemia 17:
2318–2357.
115. Wolffs P, Grage H, Hagberg O, Radstrom P. 2004. Impact
of DNA polymerases and their buffer systems on quantita-
tive real-time PCR. J Clin Microbiol 42:408–411.
116. Yeung AT, Holloway BP, Adams PS, Shipley GL. 2004.
Evaluation of dual-labeled fluorescent DNA probe purity
versus performance in real-time PCR. Biotechniques 36:
266–275.
117. Whale AS, Huggett JF, Cowen S, Speirs V, Shaw J,
Ellison S, Foy CA, Scott DJ. 2012. Comparison of micro-
fluidic digital PCR and conventional quantitative PCR
for measuring copy number variation. Nucleic Acids Res
40:e82.
 Microarray-Based Environmental Diagnostics
DARRELL P. CHANDLER
2.3.3
A PROPOSED DEFINITION FOR AN
ENVIRONMENTAL DIAGNOSTIC
In science, definitions and terminology are important so that
information is clearly and precisely conveyed, and others are
able to repeat work and verify results. As concepts and ideas
from one discipline are incorporated into another, however,
the definition, meaning or use of certain terms can drift.
Such is the case for the term “diagnostic” and its derivatives.
According to Merriam-Webster’s dictionary (http://www.
merriam-webster.com, accessed 4 December 2012), defini-
tions for diagnose, diagnosis, and diagnostic are:
Diagnose:
• To recognize (as a disease) by signs and symptoms
• To diagnose a disease or condition in [diagnosed the
patient]
• To analyze the cause or nature of [diagnose the problem]
Diagnosis:
• The art or act of identifying a disease from its signs and
symptoms
• The decision reached by diagnosis
• A concise technical description of a taxon
• Investigation or analysis of the cause or nature of a condi-
tion, situation, or problem [diagnosis of engine trouble]
• A statement or conclusion from such an analysis
Diagnostic:
• Of, relating to, or used in diagnosis [a diagnostic tool]
• Using the methods of or yielding a diagnosis [diagnostic
tests]
• Serving to distinguish or identify [a diagnostic feature]
• The art or practice of diagnosis
• A distinguishing mark
For the most part, these terms and definitions are useful for
clinical and environmental microbiology, as well as in many
areas of technology development. Consider, however, the
U.S. Food and Drug Administration (FDA) definition for
an in vitro diagnostic (IVD) product:
In vitro diagnostic products are those reagents, instruments,
and systems intended for use in diagnosis of disease or other
doi:10.1128/9781555818821.ch2.3.3
2.3.3-1
conditions, including a determination of the state of health,
in order to cure, mitigate, treat, or prevent disease or its sequelae.
Such products are intended for use in the collection, prepara-
tion, and examination of specimens taken from the human
body. (21 CFR 809.3; emphasis added)
In brief, the FDA definition of a diagnostic focuses on the
use of the diagnostic or information to make a decision and
take action as it relates to the effect of a particular condition
on human health. Detection per se is necessary but not suffi-
cient to define an IVD. Similar distinctions between com-
mon and regulatory usage exist for terms such as verification
and validation (see definitions provided in 21 CFR 820.3).
The net effect of these subtle differences in both the concept
and definitions for a diagnostic translates into a host of federal
regulations and controls over the development, manufacture,
marketing, and use of IVDs for both the developer and user, to
ensure the safety and effectiveness of the device for the
patient. In environmental settings, however, how often are
the former diagnostic concept and terminology conflated
with those for an IVD?
Anyone who has studied FDA regulations in detail or who
has been involved in product development for regulated mar-
kets will recognize that regulations and standards are intended
to ensure that high-quality science and effective quality
controls prevail during product development, manufacture,
and use. On the other hand, if everyone in the diagnostic
community actually conducted high-quality science, imple-
mented current good manufacturing practice, and adhered
to objective and consistent quality control, there would be
no need for regulation! To be clear, this is not a call for
more government regulation, but an opportunity to more pre-
cisely define an environmental diagnostic (EVD) and convey
that there are significant differences between environmental
discovery, hypothesis testing, surveillance and monitoring,
and diagnostic uses of technology. For the purpose of
this chapter, then, a suggested definition is: EVD products
are those reagents, instruments, and systems intended for
use in diagnosis of environmental status or other conditions,
to mitigate, treat, or prevent human disease or its sequelae.
Such products are intended for use in the collection, prepara-
tion, and examination of samples taken external to the
human body.
Although this chapter is focused on nucleic acid microar-
rays, the proposed definition of an EVD is deliberately generic
2.3.3-2 ▪ TARGET-SPECIFIC DETECTION

and intended to encompass any technology or device that is

used in an environmental context to make a decision or take
action that may have an impact on human health. As with
IVDs, the potential applications of EVDs are likewise broad,
with the most common uses related to food safety, water qual-
ity, indoor and outdoor air quality, bioremediation, biosafety,
and biosecurity. The reason for limiting the definition of an
EVD to those situations where there is a potential impact
on human health is a consideration of risk. For example, there
is no risk for incorrectly or partially identifying the microbial
community composition of treated effluents as a matter of
understanding the effect of a new decontamination procedure
on pathogen abundance or viability, but there is significant
risk to the residential population for persistent false negative
detection and release of human pathogens from the waste-
water treatment plant once a new decontamination proce-
dure is implemented in practice.
Clinical diagnostic products are further separated into
Research Use Only (RUO), Investigational Use Only (IUO),
and IVD categories, concepts that likewise reflect subtle but
important distinctions between the common concepts of a
diagnostic and an IVD or EVD. These differences in labeling
and use have significant implications for manufacturing qual-
ity assurance and quality control, ease of use, total analysis
time, sensitivity and specificity, repeatability, per test cost,
and ultimately the preferred platform and approach for gener-
ating the information or diagnosis. Stated another way, and
nearly 25 years after their initial invention and description
(1–4), there are very concrete and justifiable reasons microar-
rays have still not been widely developed or adopted for IVD
or EVD intended uses (see 5, 6 as simple examples). The
objectives of this chapter are therefore to expand on the con- FIGURE 1 Microarray workflow simplification through inte-
cept of a microarray-based EVD (as defined here), describe grated biochemistry, resulting in an amplification microarray. Inte-
how the underlying microarray technologies and methods grating nucleic acid sample preparation into a sample-to-answer,
of manufacture and use may have a direct impact on the qual- fully contained consumable is discussed in detail elsewhere (e.g.,
ity and reliability of the resulting environmental diagnosis, [154], and chapter 2.3.4; Straub et al.), and is not considered
and illustrate areas of current and future research that may here. Automated microarray image and data analysis methods like-
facilitate the translation of microarray technologies from wise exist, and are expected to be a standard component of EVDs to
RUO to EVD intended uses. minimize or eliminate user subjectivity in the process and diagnosis.
doi:10.1128/9781555818821.ch2.3.3.f1
MICROARRAY TECHNOLOGIES,
REPRODUCIBILITY,
AND QUALITY CONTROL components, and analytical procedures comprising an EVD
There are numerous review articles that describe nucleic acid (Fig. 1), are these assumptions accurate? How many published
microarray technology and the environmental applications reports demonstrate that an environmental microarray fabri-
thereof (e.g., (7–27)). Typical protocols include manual steps cated and/or used by one research group generates the same
for nucleic acid purification, gene-specific or whole-genome results and conclusions when fabricated and/or used by
amplification, target fragmentation, target labeling, target another (see 30 as one example)?
purification, target concentration or quantitation, dilution Consider the components of a microarray itself and the var-
with hybridization buffer, target hybridization, and microar- ious methods for microarray manufacture. Any nucleic acid
ray washing prior to imaging and data analysis (Fig. 1) (see microarray may include a substrate, a surface coating or immo-
also [28, 29] and references therein). Even the imaging and bilization matrix, a chemical cross-linker, and a set of probes.
data analysis are subject to significant manual intervention Substrates may be solid glass or plastic surfaces, flow-through
and interpretation. To ensure not only the reproducibility or porous matrices of varied composition, membranes, micro-
of the system, but more important the safety and effectiveness spheres or microparticles, continuous hydrogels, or discrete gel
of an IVD, each one material, component, and process elements affixed to a support. Surface coatings may be specific
involved in the analysis are subject to very stringent quality functional chemicals, hydrogels, proteins, or combinations
control metrics and acceptance criteria that must be defined thereof. There are many methods to cross-link nucleic acids
during the design and manufacture of the product, prior to to surfaces and/or functionally reactive layers, either cova-
validation experiments intended to establish the safety and lently or noncovalently. Short oligonucleotide, long oligonu-
efficacy of the system. Even a cursory examination of environ- cleotide, gene fragment, or whole genome probes may be
mental microarray literature, however, reveals that the under- synthesized in situ, spotted, or mixed with reactive substrates
lying quality of microarrays themselves, let alone all other in bulk (e.g., coded microparticles). Describing each of these
materials or steps in the process, is assumed to be consistently microarray technologies and manufacturing methods would
high. Given the number and complexity of materials, require an entire book, but the salient question remains:
 2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-3
what controls ensure reproducible test performance and the
validity of the resulting diagnosis, independent of the user?
Some simple examples illustrate the number of variables
that may systematically impact the analytical or operational
performance of a microarray and highlight some of the dis-
tinctions being made between a common diagnostic and an
EVD. Early theoretical and experimental work with planar
microarrays, for example, described the effects of oligonucleo-
tide density, sequence, structure, probe length, linker chemis-
try, and linker length on the efficiency of solid-phase
hybridization (31–37). This was followed by experiments to
understand the behavior of mismatched probe:target pairs
(38–47), the effect of surface charge on hybridization effi-
ciency (48–50), impacts of modified nucleosides within
immobilized probes (51), and even the impact of feature
size on hybridization behavior (52). The intrinsic fluores-
cence of substrates, surface coatings, and print buffers has
also been addressed in the literature (53). Some of these basic
experiments and studies have likewise been extended to
microarrays developed for environmental research (e.g., 54–
57). Effort has also been devoted to correlating microarray
probe behavior based on thermodynamics and Gibbs free
energies (40–42, 58–64). Even with the continued develop-
ment of probe design software (65–71), of which the most
sophisticated commercial packages take into account nearest
neighbor and secondary structure effects of both the probe
and target on duplex formation, the underlying algorithms
do not take into account microarray surface charge or surface
energies and their influence on duplex formation and mis-
match discrimination because these effects are not under-
stood and are specific to each surface composition or layer.
Thus, different DNA probes and different arrays from the
same lot can and do hybridize with different intensities,
regardless of the extent of in silico probe design (see 72, 73
as recent examples).
While there remain significant gaps in our understanding
of microarray-based hybridizations, the prior literature does
describe some of the variables that may contribute to system-
atic variability in hybridization behavior, signal intensities,
mismatch discrimination, and ultimately test performance
(see examples in Table 1). Consequently, if each of these var-
iables has a direct effect on nucleic acid hybridization effi-
ciency, specificity, and sensitivity, then what type of
controls should be in place during manufacture and use to
ensure consistent quality of an EVD consumable and the
resulting test output? Even something as seemingly innocuous
as vapor deposition versus dip coating glass substrates with
silane can have a significant effect on nucleic acid binding
isotherms (47). Knowing that probe sequence and length
influences hybridization behaviors and test performance,
how are the sequences, length, and purity of probes for an
in situ synthesized or spotted array containing >104 discrete
elements objectively verified or validated? How much non-
target probe or failure sequences can be immobilized within
a microarray feature without impacting the diagnosis? If probe
design software cannot incorporate microarray surface effects
into thermodynamic calculations for probe:target behavior,
then is an in silico probe specificity evaluation (sometimes
termed probe validation) an adequate substitute for empirical
validation of probe behavior and repeatability?
While systematic noise and a lack of repeatability are very
common in microarray experiments (74–78), they are often
ignored, especially in RUO applications. Instead, microarray
users have tried to address microarray variability at the point
of data collection and analysis (e.g., 79–83). Even using well-
controlled nucleic acid standards, however, meta-analysis of
resulting datasets frequently shows significant variation
between technical replicates, amplification/labeling, and
hybridizations, with effects evident at the individual micro-
array, run, and experimental levels (see 80, 84 as some exam-
ples). With increased interest in using high-density arrays for
clinical diagnostics, the FDA established a Microarray Qual-
ity Control (MAQC) consortium to begin addressing some of
the inherent issues surrounding microarray repeatability and
provide specific recommendations to the microarray com-
munity (85–88). Others have likewise contributed to meth-
ods and techniques to ensure microarray quality and
repeatability (e.g., 89–92, among many others). As a conse-
quence, and as commercial manufacturers continue to imple-
ment more stringent quality control procedures and good
manufacturing practices to meet the demands of an IVD (as
defined by 21 CFR 820), the repeatability of (high density)
microarrays has certainly improved (e.g., 93, 94). Neverthe-
less, a recent review of Affymetrix data sets still showed that
up to 10% of released microarrays were of poor quality (72).
Therefore, to reduce the diagnostic impact of microarray var-
iability, a common recommendation is to perform true tech-
nical replicates (72). For an IVD or EVD, is a 10% product
“failure” rate acceptable? How many environmental microar-
ray studies perform true technical replicates to draw a conclu-
sion about microbial community composition, activity, and
fluctuations over time? If thousands of samples are needed
to generate a robust gene list for predicting outcome in cancer
(78), how many samples and replicates might be needed to
generate a list of gene signatures to predict or diagnose the
outcome of natural or attenuated bioremediation?
Those who develop, market, and sell microarray-based
IVDs are legally required to address the foregoing quality
and reproducibility issues and bring them under control,
otherwise the technology is not an IVD. If environmental
microarrays are likewise to be implicitly (by the research com-
munity) or explicitly (by commercial manufacturers) repre-
sented as EVDs, then it seems prudent to suggest that they
come with similar quality assurances and performance evalu-
ations as IVDs. Stated another way, it is very misleading to
develop an environmental microarray-based test under the
rubric of a common diagnostic and represent it as an EVD—
the former shows the potential of an environmental microar-
ray, the latter actually fulfills its promise for the end user.
SIMPLIFICATION CAN SIMPLIFY THE PATH
TO AN EVD
The previous section discusses some, but not all, of the com-
plexities associated with the microarray itself. As shown in
Fig. 1, however, the microarray is only one component of
an otherwise complex environmental test and set of equip-
ment. As for the microarray, we tend to make many assump-
tions regarding the quality and reproducibility of materials
and procedures for sample preparation, target amplification,
labeling, fragmentation, repurification, hybridization, wash-
ing, imaging, and data analysis, without explicit or objective
verification of the underlying assumptions. However, an IVD
needs to provide a level of quality assurance for each of these
steps in the process, to include the hardware and software
associated with each. Shouldn’t an EVD do the same?
Given the complexity of standard microarray workflows
relative to the number of factors that need to be controlled
to generate repeatable results, it is easy to become over-
whelmed and paralyzed by the sheer number of variables,
ignore the underlying challenges associated with developing
or using an EVD, and therefore constrain environmental
2.3.3-4 ▪ TARGET-SPECIFIC DETECTION

TABLE 1 Common areas and sources of systematic variability associated with environmental microarrays

Potential source(s)a Potential effect(s)

Raw Materials
Reagents and chemicals of improper purity or grade Nucleic acids and the function of microarrays can be degraded by chemical,
enzymatic, microbiological, or other impurities that are introduced as raw
chemicals/reagents, precursors, or during the manufacturing process.
Probe purity
Cross-contamination during probe synthesis or Nontarget probes or failure products alter probe immobilization efficiency,
purification immobilized probe concentration, hybridization kinetics, hybridization
specificity, and signal intensity.
Failure sequences (for in situ synthesis, this occurs during
microarray manufacture)
Substrate dimensions (flatness) Substrate tolerances are important for reproducible printing, microaray
alignment within the imager field of view and keeping arrays in focus.
Substrate fluorescence Substrates with unpredictable, nonuniform, or high global background may
lead to false positive or false negative test results.
Substrate artifacts (smears, scratches, lint, oils, dust)
Surface coating composition, purity, and thickness An improper or nonuniform surface modification will affect probe
immobilization efficiency and feature morphology, and may result in spot or
Surface coating uniformity over the surface
feature damage during use. Also a common cause of nonuniform and
nonspecific background artifacts.
Chemical compatibility of gaskets, adhesives, coverslips Chemical incompatibilities may interfere with hybridization, increase local
with samples, reagents, enzymes, buffers and global backgrounds, or lead to test failure.
Manufacture
Environmental controls Uncontrolled temperature and relative humidity lead to concentration
gradients and viscosity. High humidity can cause feature migration, depending
on the surface coating and attachment chemistry.
Pipetting accuracy Pipetting variability during the preparation of a source plate leads to
differences in inking depths during printing, which may lead to inconsistent
feature size or volume, morphology, and hybridization behavior.
Cross-contamination of PCR primers, microarray probes, or positive control
solutions could lead to false positives.
Nonuniform probe concentration (immobilization
efficiency)
Nonuniform probe concentration, feature size, shape, or volume may lead to
Feature size and shape variable hybridization kinetics, hybridization specificity, and signal intensity.
Deposition volume
Grid uniformity Nonuniform grids may not be recognized by image analysis software.
Instrument effects
Mechanical variations between instruments and pins can lead to variable
Inking depth
microarray and feature placement on the substrate, inking depths, feature size
Pin effects or shape, which may lead to variable hybridization properties.
Source plate effects Variability in probe homogeneity, the presence of concentration gradients,
and differences in solution volumes may impact inking depth, feature
composition, and (ultimately) hybridization behavior.
Splatter ( primarily for noncontact printers) Nonspecific splatter and printing artifacts can lead to false positive signals or
confound image segmentation and analysis algorithms.
Insufficient print head washes between probes Well-to-well contamination due to inefficient pin washing can lead to false
positives.
Insufficient washing of unbound probes from surface Could lead to nonspecific hybridization artifacts or altered hybridization
behavior during use.
Missing or damaged spots or features May lead to false negatives or confound image and data analysis algorithms.
Operator effects Subjective manufacturing and quality assurance/quality control practices lead
to variable microarray quality.

(Continued on next page)


TABLE 1 (Continued )
Potential source(s)a
Hybridization and Wash
True replicates versus pseudo replicates
Temperature nonuniformities
Inadequate mixing
Particulates and contaminants
Handling errors, microarray damage
Operator effects
Imaging and analysis
Autofluorescence (substrates, salt residue, handling
residue)
Fluorescent artifacts and debris
Nonuniform illumination
Stray light
Grid placement
Segmentation algorithm (spot or feature identification)
Signal saturation (or underexposure)
Definition and calculation of noise
Signal averaging, transformation
Feature exclusion and quality metrics
Decision algorithm, thresholds
a
The number of independent sources that may contribute to any one area of systematic variability is usually more extensive than what is listed, many of which are inter-
related. Biases and challenges associated with nucleic acid sample preparation and amplification are not considered here.
microarrays to the realm of RUO products. An alternative
(interim) approach used in the clinical realm is to establish
dedicated analysis laboratories and testing services that com-
ply with the Clinical Laboratory Improvement Amendments
to run high-complexity tests with some level of consistency
and quality. Similar concepts and practice do exist in envi-
ronmental circles (e.g., Environmental Protection Agency
Approved Standard Methods), and a similar infrastructure
could be developed for environmental microarrays. However,
one important requirement for an IVD or EVD, and especially
those being developed for the point of use, is that the test be
repeatable and reliable, independent of the user. The simpler
a test, the more robust it will be in the hands of multiple users.
Note that simple is not synonymous with simplistic—a Cep-
heid GeneXpert test is relatively simple, but the system and
technology is certainly not simplistic. How, then, can we sim-
plify an environmental microarray test?
Several commercial microarray manufacturers have engi-
neered fluidic hybridization, washing, and imaging stations,
2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-5
Potential effect(s)
End-point signal intensities and signal-to-noise thresholds have a bearing on
limits of detection. Pseudo replicates (replicated probes within a microarray)
are not a substitute for true replicates (from sample extraction through
microarray analysis).
Within and between arrays. Can affect hybridization specificity and
sensitivity.
Within and between arrays, during hybridization and wash. May lead to
differential hybridization behavior, and/or differential release of
nonspecifically hybridized targets.
Increase local and global background, potentially resulting in false negatives or
false positives.
May lead to false negatives or confound image and data analysis algorithms.
Nonstandardized use and subjective practices lead to variable hybridization
behavior, artifacts, noise, damage, and failures.
May lead to false positives or negatives or confound image and data analysis
algorithms.
Within and between imagers. Leads to inaccurate estimates of signal and
noise, regardless of segmentation or analysis algorithms.
Errors in grid placement, segmentation, exposure time affect measures of signal
and noise. May lead to false positives or false negatives.
Specific decisions and algorithms affect final calculations and test outputs, and
estimates of sensitivity or specificity. Subjectivity in feature exclusion and
quality metrics leads to nonrepeatable results.
and provide semi-automated software and data analysis as
one means to simplify microarray workflow for the user. Sam-
ple preparation and amplification chemistries are also incor-
porated into microarray workflows either through robotic
(95, 96) or microfluidic transfer steps (97–103). Some micro-
array methods have been fully automated as prototype instru-
ments and commercial products (96, 102–107; see also
products from Autogenomics, Nanosphere, ClonDiag, and
Luminex), and there are now examples of fully integrated,
field-deployable environmental sensors that incorporate
microarray technologies and principles (108–111). While
there is tremendous promise in simplifying environmental
microarrays through engineering, the manufacturing and
quality control challenges for the instruments and consum-
ables scale with the complexity of the microfluidic and hard-
ware architecture. Robotic systems must also generally
contend with an open-amplicon workflow that plagued the
adoption of conventional PCR technologies as an environ-
mental diagnostic tool. For both microfluidics or robotics,
2.3.3-6 ▪ TARGET-SPECIFIC DETECTION
the quality control challenges for converting fully automated
microarrays into quality-assured EVDs are significant and
extend far beyond the relatively short list of factors described
in Table 1.
An alternative approach to simplifying microarray work-
flow, making the tests more robust, and increasing microarray
adoption as an EVD is to simplify the biochemistry and ana-
lytical steps in the process. Simplifying the procedure itself
provides a corollary opportunity to simplify the complexity
of the consumables and instrumentation and reduce the
global quality control and validation burden for manufac-
turers and users alike. One example (illustrated in Fig. 1) is
to combine target amplification, labeling, and microarray
hybridization into a single-solution, closed-amplicon reaction
chamber. Early attempts involved amplifying target nucleic
acids by PCR on a solid support, where the amplification
primers were cross-linked to the surface (112–119). In gen-
eral, these studies showed limits of detection at 105–106
genomes per reaction, which is insufficient to meet the user
needs for many EVD applications, especially those in food
safety, water quality, and biodefense. Supplementing the reac-
tion mixture with unbound, gene-specific primers and allow-
ing the PCR to simultaneously proceed in the liquid and solid
phases was one approach to increase product yield and analyt-
ical sensitivity (120–125). As an analog of nested PCR and
where the products of the first amplification phase are not
purified before the second amplification phase, however, pri-
mer artifacts and primer interference restrict multiplexing
capacity and amplification efficiency of these solid-phase
amplification methods.
Other variants of microarray-based PCR are also de-
scribed. In an attempt to overcome primer interference,
for example, Tillib et al. (123) created a microarray of PCR
chambers separated from each other by mineral oil, with
each chamber containing a unique primer set. Pemov et al.
developed a gel element array where multiplex PCR occurs
on and within gel elements and is enhanced by pseudo-
monoplex PCR in solution (126). Sun et al. described an
approach to influenza RNA amplification and detection,
where RNA is reverse transcribed in solution over the micro-
array, PCR is initiated in solution with free-floating reverse
primers, and the resulting cDNA is then extended from
nested, immobilized primers on the microarray (127). A
related method used gene-specific, immobilized reverse pri-
mers to directly interrogate an mRNA transcriptome (128).
Isothermal, helicase-dependent, solid-phase amplification
has also been demonstrated, which provides an opportunity
to simplify EVD instrumentation by eliminating the need
for a thermal cycler (129). Unfortunately, solid-phase ampli-
fication is limited by the kinetics of low-copy target hybridiza-
tion to the microarray surface during the initial rounds of the
amplification reaction, which translates either into extended
analysis times, complex workflow (e.g., repeated addition of
enzymes or reagents), relatively low limits of detection, or
some combination thereof.
Highly multiplexed, solution-phase amplification techni-
ques that precede microarray detection are becoming more
common in microarray experiments (e.g., 130, 131), and
may provide an alternative path toward simplifying microar-
ray workflow. The underlying principles rely on driving the
amplification reaction to the plateau phase and using the
microarray to discriminate targets from amplification artifacts
that may arise as a consequence of the high multiplexing. A
clever simplification of these approaches is to print microar-
rays within a gel-lined cap of a microcentrifuge tube. Nucleic
acid amplification occurs within the bottom of the tube as
a solution-phase PCR. After amplification, the tube is
inverted, a hybridization buffer is released from an internal
chamber, and microarray hybridization occurs within the
PCR vessel (132). The general concept of integrating quan-
titative, real-time PCR with microarray readouts has likewise
been reported. Khodakov et al. used a first-stage, multiplex
PCR in solution, and then performed a real-time, quantita-
tive PCR amplification on gel element arrays (133). SYBR
Green I dye intercalation was used to detect target nucleic
acids that were extended from gel-immobilized primers during
the final stages of the elongation step. Pierik et al. (134)
described a similar system and approach, except that a Cy5
fluorescent tag was incorporated into one of the solution-
phase amplification primers, and real-time quantitation was
based on target hybridization to the array rather than chain
extension from the array. In both cases, customized instru-
ments that integrate a thermal cycler with optical detection
around a planar substrate are required, the amplification effi-
ciency from the array is rather low, and a relatively large num-
ber of cycles is required to achieve useful limits of detection.
Given these examples of integrated and simplified bio-
chemistry associated with a microarray test, it is therefore
reasonable to ask whether the complexity of current environ-
mental microarray procedures is really needed to satisfy the
environmental diagnostic objective, let alone user needs
related to ease of use, total analysis time, and per test cost.
Could it be that the development of both IVDs and EVDs
is being unintentionally constrained by biochemical and
enzymatic methods dating from the advent of molecular biol-
ogy in the 1970s and 1980s?
AMPLIFICATION MICROARRAYS
AS POTENTIAL EVDs
From the foregoing literature, it is clear that the development
of a simple, microarray-based EVD requires additional advan-
ces in integrated biochemistry, consumables, hardware, and
software. Toward that end, we are exploiting advances in
solution-phase multiplexed amplification chemistry, the
solution-phase properties and high probe immobilization
capacity of gel element arrays (e.g., 135), valveless microflui-
dic consumables (136), and portable microarray imagers
(137) in an attempt to create EVDs for the point of use.
The principles for simplifying the biochemical steps and
microarray consumable are to utilize multiplex, asymmetric
PCR or reverse-transcriptase PCR in solution with fluores-
cently labeled reverse primers, perform thermal cycling in
the presence of the gel element array, and allow the predom-
inantly single-stranded, labeled amplicons to hybridize to the
array during or after on-chip thermal cycling, all within the
confines of a single reaction chamber. This type of integrated
biochemistry effectively converts up to seven manual steps
and processes into a single step in a single microfluidic cham-
ber (Fig. 1). Microarrays are washed after thermal cycling in
bulk solution or with a bolus of self-imbibing wash solution
pipetted into a flow cell that contains an integrated waste
chamber, similar to the action of lateral flow test strips.
The resulting amplification microarray differs from other
PCR array technologies such as those from BioTrove, Idaho
Technologies, SABiosciences, Lonza, Perkin Elmer, BioRad,
and others (reviewed in chapter 2.3.4) because an amplifica-
tion microarray is a homogeneous reaction whereas PCR
arrays utilize spatially isolated reaction wells, droplets, or
channels to achieve multiplexed amplification and detection.
A homogeneous reaction is important when the diagnostic
objective requires a very stringent limit of detection for
 2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-7
multiple targets simultaneously, and it is important to avoid
splitting the sample to the point where there are not enough
nucleic acid targets in each reaction well for repeatable detec-
tion. In this vein, a prototype 16S rDNA amplification micro-
array and its use for environmental monitoring has been
recently described (138). Analytical limits of detection
were between 2 and 200 cell equivalents of purified DNA
and amplification microarray signatures were well correlated
with 16S-targeted qPCR results and hybridization microarray
signatures. The succession of the microbial community was
evident with and consistent between the amplification
microarray and conventional hybridization microarray plat-
forms and was empirically consistent with in situ processes
occurring at the site. While the study demonstrated the
potential to greatly simplify environmental microarray work-
flow with instruments and consumables that can also be
deployed in a field trailer (137), it revealed a number of tech-
nical and process-level variables over and above the amplifi-
cation microarray consumable that must still be addressed and
controlled before the platform could be considered as an
EVD, as opposed to an RUO test. For example, amplification
microarray efficiency is significantly impacted by the effi-
ciency of heat transfer from an in situ thermal block, through
the microarray substrate, to the reaction solution. We have
observed that end-point signal intensities in an amplification
microarray can vary as a function of “where” the amplification
microarray is physically located on the in situ block. Using in
situ temperature probes, we have also measured significant
variations in block and solution temperatures, within and
between instruments, all running identical thermal cycling
programs. These observations are not new and were identified
at the very advent of the PCR (139–141), yet thermal cycler
nonuniformities are still a cause of PCR variability almost 30
years later (142–147)! This simple example again illustrates
that the microarray consumable itself is only one part of the
EVD, and ensuring repeatable results for any and all users
will require standardized controls and objective acceptance
criteria for the corollary EVD reagents, methods, and
instruments.
CURRENT AND FUTURE CHALLENGES
In the same way that PCR advanced from an open-amplicon,
gel-based analysis technique to fully integrated, sample-
to-answer cartridges and instrumentation that has been
cleared by the FDA as an IVD for specific intended uses, it
is very likely that continued advances in automated sample
preparation, integrated biochemistry, and engineering will
likewise make microarrays much more user friendly and sim-
ple for environmental users. An argument can therefore be
made that the future of microarrays as EVDs is no longer
limited by technology per se. This is not to say that there
are no technology limitations with microarrays, amplification
microarrays, or integrated systems. For example, for many of
the underlying quality reasons already discussed, microarrays
are still a qualitative test and only enable relative quantitation
above some detection threshold. The lack of absolute quanti-
tation is not necessarily a fatal flaw, but it brings the discussion
back to one of the distinctions between a common diagnostic
and EVD—that of intended use.
The FDA’s definition of intended use relates to product
labeling and representations made about the product, not
the technology used to generate the information:
The words intended uses or words of similar import in §§
801.5, 801.119, and 801.122 refer to the objective intent of
the persons legally responsible for the labeling of devices.
The intent is determined by such persons’ expressions or
may be shown by the circumstances surrounding the distri-
bution of the article. This objective intent may, for example,
be shown by labeling claims, advertising matter, or oral or
written statements by such persons or their representatives.
(21 CFR 801.4)
As such, the FDA definition is not very informative or
instructive for the environmental microbiologist or EVD
developer. Nevertheless, a clinical intended use statement
typically includes a general description of the diseases or con-
ditions that the device will diagnose, treat, prevent, cure, or
mitigate. It will define and describe the patient population
and indications for use for which the device is intended, spec-
ify the type of clinical specimen or sample, clinical setting and
use environment, and expected level of operator training or
competency. The intended use must address the needs of users
and patients, and it must be explicit in the labeling claims. In
short, the intended use is a summary of the claims that one
wants to make about the product.
What differentiates the intended use statement of an IVD
from an RUO product is its specificity—it is very explicit.
For a nucleic acid microarray-based EVD, elements of an
intended use statement may include:
• The biochemical parameter(s) or analyte(s) being meas-
ured (e.g., DNA, rRNA, mRNA);
• Target microorganisms (taxonomic or functional identifi-
cation) and specific activities (e.g., Fe reduction, dehalo-
genators; organisms carrying the nirS gene);
• Limitations of the microarray (e.g., in coverage or known
sequence specificity);
• Environmental sample (e.g., indoor or outdoor air; filters;
surface water, finished water; waste water, groundwater,
salt water, process water; soils; sediment, sludge, sewage,
effluents; food matrices or surfaces) and specific context
(e.g., radionuclide or hydrocarbon-contaminated samples;
human pathogens; poultry processing);
• User (e.g., reference laboratory, regulatory agency, munic-
ipality, food processors, bioremediation service industry);
• Application environment (e.g., centralized or core molec-
ular biology laboratory; field trailers; processing site);
• Conditions of use (e.g., before, during, or after in situ
biostimulation);
• Purpose of the measurement (screening, diagnosis, sur-
veillance, process control);
• Identification of or implication of an effect on human
health (e.g., heavy metals cause neurological damage; Sal-
monella and Escherichia coli cause food poisoning and may
lead to death).
From the foregoing discussion about systematic sources of
environmental microarray variability, one should now be able
to see or understand why being very specific and explicit
about the intended use matters a priori—the product re-
quirements, specific technical solutions for meeting those
requirements, and implied quality control practices and pro-
cedures are all derived from the intended use. For example,
using an amplification microarray to monitor microbial com-
munity structure during in situ bioremediation leads to a
requirement that the system detect ≤103 gene copies per
test, a limit of detection that is consistent with in situ biomass
of subsurface environments. A zero tolerance for Salmonella
leads to a requirement that the EVD detect one organism
per sample—the derivative requirements of zero tolerance
for sample preparation, microarray workflow, and equipment
are numerous indeed.
2.3.3-8 ▪ TARGET-SPECIFIC DETECTION
Failure to have an accurate and specific intended use will
ultimately result in an incomplete product design, and failure
to objectively demonstrate the adequacy of the product for
its intended use. Simply providing a large number of test pro-
cedures and results for an environmental microarray does
not constitute adequate evidence of device performance or
safety—these experiments and results must be explicitly
related to the product requirements and specific intended
use. Thus, the intended use statement leads to a product def-
inition and set of requirements that are complete, consistent,
understandable, and (most important) objectively testable
and verifiable. The more one wants to say or claim about a
microarray-based EVD, the more unwieldy and exponentially
larger becomes the verification and validation burden. For
example, it is one thing to claim and validate that a microar-
ray can be used to “monitor microbial community structure in
the subsurface.” It is quite another to claim that the microar-
ray is intended to be used “in a field trailer setting to monitor
the occurrence and relative abundance of iron-reducing bac-
teria from genera A–Q present in filtered subsurface ground-
water containing <R ppm dissolved organic carbon and an Eh
between S and T before, during, and after in situ treatment
with organic electron donors.” The validation burden for
the former claim is immense; the latter claim is better defined
and therefore more tractable. One of the primary challenges
in developing an EVD now and in the future, then, is for envi-
ronmental microbiologists to be precise about the intended
use, claims, and representations of any given microarray.
Only then can technology gaps and solutions be objectively
identified and technically addressed.
A second challenge for microarray-based EVDs is the
absence of regulatory standards or guidelines governing their
development, validation, or use, which is both a blessing and
a curse. Stated another way, RUO microarray manufacturers
(including university core laboratories) need not even be cer-
tified by the International Organization for Standardization,
let alone be compliant with an FDA-type Quality System
Regulation (which also governs the actual design and devel-
opment of the test). From the perspective of an EVD, this lack
of oversight or certification may be construed as a blessing
because it significantly reduces the overhead burden and costs
associated with product development and commercialization.
On the other hand, the absence of standards or independent
review and approval is a curse because everyone has different
ideas and definitions for quality, verification, validation,
proof, and good laboratory practice, which leads directly to
nonrepeatable test performance in the hands of independent
users and a slower adoption of microarray technology by those
environmental users who could most benefit from it.
Finally, the development and adoption of nucleic acid
microarray-based EVDs are and will continue to be chal-
lenged by historical and current concepts of biology and
how various user groups perceive and understand nucleic
acid signatures relative to the intended use statement or
(in some respects) existing legislation and environmental reg-
ulations. For example, what does it mean for an EVD to detect
microbial virulence factors in environmental samples (26,
148–152)? Many virulence, antibiotic resistance, and func-
tional genes are mobile. Extracellular nucleic acids can be
very stable when adsorbed to environmental particulates.
The presence of DNA is not synonymous with viability or
activity. Detecting mRNA is not synonymous with the pres-
ence of functional proteins or toxins. Metabolic activity is not
synonymous with infectivity. Therefore, even though envi-
ronmental microarrays can detect many different nucleic
acid signatures that are associated with human pathogens or
environmental processes of interest, whether or not those sig-
natures contain actionable information (as required by the
definition of an EVD) is a fundamental biology, intended
use, and regulatory quandary, not a technology limitation
or dilemma.
Similar challenges and questions arise regarding absolute
quantitation in environmental samples. Assuming that
microarrays will eventually be able to quantify or count indi-
vidual targets with the sensitivity of real-time PCR (153),
what does it mean to detect 2.38 × 102 genes? Is doubling
of target genes (e.g., from 4 × 103 to 8 × 103) meaningful?
Does it contribute anything to the diagnosis or the way
the user will take action? What about a 10-fold increase in
target abundance? Given the prior discussion about system-
atic variability, how does the user know that these quantita-
tive outputs are not simply reflecting the inherent noise of
the analytical process, or systematic variability built into
the EVD itself? Again, these are no longer technology chal-
lenges per se.
CONCLUSION
The preceding decade certainly established the potential of
environmental microarrays as diagnostic tools. To fulfill
that promise, however, it is necessary to become more precise
in our definition of an environmental diagnostic. One possi-
ble definition of an EVD is therefore provided, borrowing
concepts and ideas from the FDA and the regulation of
IVD devices, the salient point being that an EVD and the
resulting diagnosis is used to “mitigate, treat, or prevent
human disease or its sequelae.” In this context, future tech-
nology challenges are primarily associated with reducing
the variability of environmental microarrays during manufac-
ture and use so that repeatable results can be obtained inde-
pendent of the user. Process simplification, perhaps through
amplification microarrays or other technologies described in
this book, may help achieve the objective of repeatability
for independent users, but technology per se will not substi-
tute for a clearly defined intended use, effective product
design, and objective verification and validation for that spe-
cific intended use. The absence of regulatory oversight for
EVDs is both a blessing and a curse, the solution to which
will only come with a consensus biological and regulatory
opinion regarding the meaning of environmental nucleic
acid signatures relative to the real or perceived risks associated
with each intended use.
REFERENCES
1. Khrapko KR, Lysov YP, Khorlyn AA, Shick VV, Floren-
tiev VL, Mirzabekov AD. 1989. An oligonucleotide
hybridization approach to DNA sequencing. FEBS Lett
256:118–122.
2. Southern EM, Maskos U, Elder JK. 1992. Analyzing and
comparing nucleic acid sequences by hybridization to arrays
of oligonucleotides: evaluation using experimental models.
Genomics 13:1008–1017.
3. Maskos U, Southern EM. 1992. Oligonucleotide hybrid-
isations on glass supports: a novel linker for oligonucleotide
synthesis and hybridisation properties of oligonucleotides
synthesized in situ. Nucl Acids Res 20:1679–1684.
4. Fodor SP, Read JL, Pirrung MC, Stryer L, Lu AT, Solas
D. 1991. Light-directed, spatially addressable parallel
chemical synthesis. Science 251:767–773.
5. Loy A, Bodrossy L. 2006. Highly parallel microbial diag-
nostics using oligonucleotide microarrays. Clin Chim Acta
363:106–119.
 2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-9
6. Petricoin EF, Hackett JL, Lesko LJ, Puri RK, Gutman SI,
Chumakov K, Woodcock J, Feigal DW, Zoon KC, Sistare
FD. 2002. Medical applications of microarray tech-
nologies: a regulatory science perspective. Nat Genet
32:474–479.
7. Kostić T, Sessitsch A. 2012. Microbial diagnostic microar-
rays for the detection and typing of food- and water-borne
(bacterial) pathogens. Microarrays 1:3–24.
8. McLoughlin KS. 2011. Microarrays for pathogen detection
and analysis. Brief Funct Genomics 10:342–353.
9. Lauri A, Mariani PO. 2009. Potentials and limitations of
molecular diagnostic methods in food safety. Genes Nutr
4:1–12.
10. Aw TG, Rose JB. 2012. Detection of pathogens in water:
from phylochips to qPCR to pyrosequencing. Curr Opin
Biotechnol 23:422–430.
11. Weinberg S. 2004. Emergent FDA biodefense issues
for microarray technology: process analytical technology.
Exp Rev Mol Diagn 4:779–781.
12. Call DR. 2005. Challenges and opportunities for pathogen
detection using DNA microarrays. Crit Rev Microbiol 31:
91–99.
13. Hazen TC, Rocha AM, Techtmann SM. 2012. Advances
in monitoring environmental microbes. Curr Opin Biotech-
nol 23:S0958–S1669.
14. Leski TA, Lin B, Malanoski AP, Stenger DA. 2012.
Application of resequencing microarrays in microbial
detection and characterization. Fut Microbiol 7:625–637.
15. Piña B, Barata C. 2011. A genomic and ecotoxicological
perspective of DNA array studies in aquatic environmental
risk assessment. Aquat Toxicol 105:40–49.
16. Dugat-Bony E, Peyretaillade E, Parisot N, Biderre-Petit
C, Jaziri F, Hill D, Rimour S, Peyret P. 2011. Detecting
unknown sequences with DNA microarrays: explorative
probe design strategies. Environ Microbiol 14:356–371.
17. Roh SW, Abell GC, Kim KH, Nam YD, Bae JW. 2010.
Comparing microarrays and next-generation sequencing
technologies for microbial ecology research. Trends Biotech-
nol 28:291–299.
18. Wagner M, Smidt H, Loy A, Zhou J. 2001. Unravelling
microbial communities with DNA-microarrays: challenges
and future directions. Microb Ecol 53:498–506.
19. Stahl DA. 2004. High-throughput techniques for analyz-
ing complex bacterial communities. Adv Exp Med Biol
547:5–17.
20. Uttamchandani M, Neo JL, Ong BN, Moochhala S. 2009.
Applications of microarrays in pathogen detection and bio-
defence. Trends Biotechnol 27:53–61.
21. Steinberg CE, Stürzenbaum SR, Menzel R. 2008. Genes
and environment—striking the fine balance between
sophisticated biomonitoring and true functional environ-
mental genomics. Sci Total Environ 400:142–161.
22. Gentry TJ, Wickham GS, Schadt CW, He Z, Zhou J.
2006. Microarray applications in microbial ecology re-
search. Microb Ecol 52:159–175.
23. Lemarchand K, Masson L, Brousseau R. 2004. Molecular
biology and DNA microarray technology for microbial
quality monitoring of water. Crit Rev Biotechnol 30:
145–172.
24. Ye RW, Wang TH, Bedzyk L, Croker KM. 2001. Appli-
cations of DNA microarrays in microbial systems.
J Microbiol Meth 47:257–272.
25. Al-Khaldi SF, Martin SA, Rasooly A, Evans JD. 2002.
DNA microarray technology used for studying foodborne
pathogens and microbial habitats: minireview. J AOAC
Int 85:906–910.
26. Rasooly A, Herold KE. 2008. Food microbial pathogen
detection and analysis using DNA microarray technologies.
Foodborne Pathog Dis 5:531–550.
27. Kostrzynska M, Bachand A. 2006. Application of DNA
microarray technology for detection, identification, and
characterization of food-borne pathogens. Can J Microbiol
52:1–8.
28. Waldron PJ, Wu L, Nostrand JDV, Schadt CW, He Z,
Watson DB, Jardine PM, Palumbo AV, Hazen TC,
Zhou J. 2009. Functional gene array-based analysis of
microbial community structure in groundwaters with a gra-
dient of contaminant levels. Environ Sci Technol 43:
3529–3534.
29. Handley KM, Wrighton KC, Piceno YM, Andersen GL,
DeSantis TZ, Williams KH, Wilkins MJ, N’Guessan AL,
Peacock A, Bargar J, Long PE, Banfield JF. 2012. High-
density PhyloChip profiling of stimulated aquifer microbial
communities reveals a complex response to acetate amend-
ment. FEMS Microbiol Ecol 81:188–204.
30. Davis MA, Lim JY, Soyer Y, Harbottle H, Chang YF,
New D, Orfe LH, Besser TE, Call DR. 2010. Develop-
ment and validation of a resistance and virulence gene
microarray targeting Escherichia coli and Salmonella enterica.
J Microbiol Meth 82:36–41.
31. Maskos U, Southern EM. 1992. Parallel analysis of oligo-
deoxyribonucleotide (oligonucleotide) interactions. I.
Analysis of factors influencing oligonucleotide duplex for-
mation. Nucl Acids Res 20:1675–1678.
32. Shchepinov MS, Case-Green SC, Southern EM.
1997. Steric factors influencing hybridisation of nucleic
acids to oligonucleotide arrays. Nucl Acids Res 25:
1155–1161.
33. Williams JC, Case-Green SC, Mir KU, Southern EM.
1994. Studies of oligonucleotide interactions by hybridisa-
tion to arrays: the influence of dangling ends on duplex
yield. Nucl Acids Res 22:1365–1367.
34. Chan V, Graves DJ, Fortina P, McKenzie SE. 1997.
Adsorption and surface diffusion of DNA oligonucleotides
at liquid/solid interfaces. Langmuir 13:320–329.
35. Chan V, Graves DJ, McKenzie SE. 1995. The biophysics
of DNA hybridization with immobilized oligonucleotide
probes. Biophys J 69:2243–2255.
36. Peterson AW, Heaton RJ, Georgiadis RM. 2001. The
effect of surface probe density on DNA hybridization.
Nucl Acids Res 29:5163–5168.
37. Chou C-C, Chen C-H, Lee T-T, Peck K. 2004. Optimiza-
tion of probe length and the number of probes per gene for
optimal microarray analysis of gene expression. Nucl Acids
Res 32:e99.
38. Binder H, Preibisch S. 2005. Specific and nonspecific
hybridization of oligonucleotide probes on microarrays.
Biophys J 89:337–352.
39. Binder H, Preibisch S, Kirsten T. 2005. Base pair interac-
tions and hybridization isotherms of matched and mis-
matched oligonucleotide probes on microarrays. Langmuir
21:9287–9302.
40. Forman JE, Walton ID, Stern D, Rava RP, Trulson
MO. 1998. Thermodynamics of duplex formation and
mismatch discrimination on photolithographically syn-
thesized oligonucleotide arrays. ACS Symp Ser 682:206–
228.
41. Hooyberghs J, Van Hummelen P, Carlon E. 2009. The
effects of mismatches on hybridization in DNA microar-
rays: determination of nearest neighbor parameters. Nucl
Acids Res 37:e53.
42. Horne MT, Fish DJ, Benight AS. 2006. Statistical ther-
modynamics and kinetics of DNA multiplex hybridizaton
reactions. Biophys J 91:4133–4153.
43. Li ES, Ng JK, Wu JH, Liu W-T. 2004. Evaluating
single-base-pair discriminating capability of planar oligo-
nucleotide microchips using a non-equilibrium dissocia-
tion approach. Environ Microbiol 6:1197–1202.
2.3.3-10 ▪ TARGET-SPECIFIC DETECTION
44. Peterson AW, Wolf LK, Georgiadis RM. 2002. Hybridiza-
tion of mismatched or partially matched DNA at surfaces.
J Am Chem Soc 124:14601–14607.
45. Sorokin NV, Chechetkin VR, Livshits MA, Pan’kov SV,
Donnikov MY, Gryadunov DA, Lapa SA, Zasedatelev
AS. 2005. Discrimination between perfect and mis-
matched duplexes with oligonucleotide gel microchips:
role of thermodynamic and kinetic effects during hybridiza-
tion. J Biomol Struct Dyn 22:725–734.
46. Urakawa H, El Fantroussi S, Smidt H, Smoot JC, Tribou
EH, Kelly JJ, Noble PA, Stahl DA. 2003. Optimization of
single-base-pair mismatch discrimination in oligonucleo-
tide arrays. Appl Environ Microbiol 69:2848–2856.
47. Pozhitkov AE, Boube I, Brouwer MH, Noble PA. 2009.
Beyond Affymetrix arrays: expanding the set of known
hybridization isotherms and observing pre-wash signal
intensities. Nucl Acids Res 38:e28.
48. Belosludtsev Y, Belosludtsev I, Iverson B, Lemeshko S,
Wiese R, Hogan M, Powdrill T. 2001. Nearly instantane-
ous, cation-independent, high selectivity nucleic acid
hybridization to DNA microarrays. Biochem Biophys Res
Comm 282:1263–1267.
49. Lemeshko SV, Powdrill T, Belosludtsev YY, Hogan M.
2001. Oligonucleotides form a duplex with non-helical
properties on a positively charged surface. Nucl Acids Res
29:3051–3058.
50. Piunno PAE, Watterson JH, Kotoris CC, Krull UJ.
2005. Alteration of the selectivity of hybridization of
immobilized oligonucleotide probes by co-immobilization
with charged oligomers of ethylene glycol. Anal Chim
Acta 534:53–61.
51. Seela F, Budow S. 2008. Mismatch formation in solution
and on DNA microarrays: how modified nucleosides can
overcome shortcomings of imperfect hybridization caused
by oligonucleotide composition and base pairing. Mol Bio-
syst 4:232–245.
52. Dandy DS, Wu P, Grainger DW. 2007. Array feature size
influences nucleic acid surface capture in DNA microar-
rays. Proc Natl Acad Sci USA 104:8223–8228.
53. Diehl F, Grahlmann S, Beier M, Hoheisel JD. 2001.
Manufacturing DNA microarrays of high spot homogeneity
and reduced background signal. Nucl Acids Res 29:e38.
54. He Z, Wu L, Fields MW, Zhou J. 2005. Use of microarrays
with different probe sizes for monitoring gene expression.
Appl Environ Microbiol 71:5154–5162.
55. Chandler DP, Newton GJ, Small JA, Daly DS. 2003.
Sequence vs. structure for the direct detection of 16S
rRNA on planar oligonucleotide microarrays Appl Environ
Microbiol 69:2950–2958.
56. Urakawa H, Noble PA, Fantroussi SE, Kelly JJ, Stahl
DA. 2002. Single-base-pair discrimination of terminal
mismatches by using oligonucleotide microarrays and
neural network analyses. Appl Environ Microbiol 68:
235–244.
57. Liu WT, Mirzabekov AD, Stahl DA. 2001. Optimization
of an oligonucleotide microchip for microbial identifica-
tion studies: a non-equilibrium dissociation approach.
Environ Microbiol 3:619–629.
58. Santa Lucia Jr. J. 1998. A unified view of polymer, dumb-
bell, and oligonucleotide DNA nearest-neighbor thermo-
dynamics. Proc Natl Acad Sci USA 95:1460–1465.
59. Vasiliskov VA, Prokopenko DV, Mirzabekov AD. 2001.
Parallel multiplex thermodynamic analysis of coaxial base
stacking in DNA duplexes by oligodeoxyribonucleotide
microchips. Nucl Acids Res 29:2303–2313.
60. Kunitsyn A, Kochetkova S, Timofeev E, Florentiev V.
1996. Partial thermodynamic parameters for prediction of
stability and washing behavior of DNA duplexes immobi-
lized on gel matrix. J Biomol Struct Dynam 14:239–244.
61. Lesnik EA, Freier SM. 1995. Relative thermodynamic
stability of DNA, RNA, and DNA:RNA hybrid duplexes:
relationship with base composition and structure. Biochem-
istry 34:10807–10815.
62. Wu P, Nakano S-i, Sugimoto N. 2002. Temperature
dependence of thermodynamic properties for DNA/DNA
and RNA/DNA duplex formation. Eur J Biochem 269:
2821–2830.
63. Schmitt TJ, Knotts TA, IV. 2011. Thermodynamics
of DNA hybridization on surfaces. J Chem Phys 134:
205105.
64. Wick LM, Rouillard JM, Whittam TS, Gulari E, Tiedje
JM, Hashsham SA. 2006. On-chip non-equilibrium disso-
ciation curves and dissociation rate constants as methods to
assess specificity of oligonucleotide probes. Nucl Acids Res
34:e26.
65. Xu D, Li G, Wu L, Zhou J, Xu Y. 2002. PRIMEGENS:
robust and efficient design of gene-specific probes for
microarray analysis. Bioinformatics 18:1432–1437.
66. Reymond N, Charles H, Duret L, Calevro F, Beslon G,
Fayard J-M. 2004. ROSO: optimizing oligonucleotide
probes for microarrays. Bioinformatics 20:271–273.
67. DeSantis TZ, Dubosarskiy I, Murray SR, Andersen GL.
2003. Comprehensive aligned sequence construction for
automated design of effective probes (CASCADE-P) using
16S rDNA. Bioinformatics 19:1461–1468.
68. Ashelford KE, Weightman AJ, Fry JC. 2002. PRIM-
ROSE: a computer program for generating and estimating
the phylogenetic range of 16S rRNA oligonucleotide
probes and primers in conjunction with the RDP-II data-
base. Nucl Acids Res 30:3481–3489.
69. Li W, Huang J, Fan M, Wang S. 2002. MProbe: a computer
aided probe design for oligonucleotide microarrays. Appl
Bioinformatics 1:163–166.
70. Tolstrup N, Nielsen PS, Kolberg JG, Frankel AM, Viss-
ing H, Kauppinen S. 2003. OligoDesign: optimal design of
LNA (locked nucleic acid) oligonucleotide capture probes
for gene expression profiling. Nucl Acids Res 31:
3758–3762.
71. Rouillard J-M, Zuker M, Gulari E. 2003. OligoArray
2.0: design of oligonucleotide probes for DNA microarrays
using a thermodynamic approach. Nucl Acids Res 31:3057–
3062.
72. McCall MN, Murakami PN, Lukk M, Huber W, Irizarry
RA. 2011. Assessing Affymetrix GeneChip microarray
quality. BMC Bioinformatics 12:137.
73. Wei T, Pearson MN, Armstrong K, Blohm D, Liu J. 2012.
Analysis of crucial factors resulting in microarray hybridiza-
tion failure. Mol Biosyst 8:1325–1338.
74. Marshall E. 2004. Getting the noise out of gene arrays.
Science 306:630–631.
75. Frantz S. 2005. An array of problems. Nat Rev Drug Discov
4:362–363.
76. Ioannidis JP, Allison DB, Ball CA, Coulibaly I, Cui X,
Culhane AC, Falchi M, Furlanello C, Game L, Jurman
G, Mangion J, Mehta T, Nitzberg M, Page GP, Petretto
E, van Noort V. 2009. Repeatability of published microar-
ray gene expression analyses. Nat Genet 41:149–155.
77. Ioannidis JP. 2005. Microarrays and molecular research:
noise discovery? Lancet 365:454–455.
78. Ein-Dor L, Zuk O, Domany E. 2006. Thousands of sam-
ples are needed to generate a robust gene list for predict-
ing outcome in cancer. Proc Natl Acad Sci USA 103:
5923–5928.
79. Draghici S, Khatri P, Eklund AC, Szallasi Z. 2006. Reli-
ability and reproducibility issues in DNA microarray meas-
urements. Trends Genet 22:101–109.
80. Kitchen RR, Sabine VS, Simen AA, Dixon JM, Bartlett
JM, Sims AH. 2011. Relative impact of key sources of
 2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-11
systematic noise in Affymetrix and Illumina gene-expres-
sion microarray experiments. BMC Genomics 12:589.
81. Ambroise J, Bearzatto B, Robert A, Govaerts B, Macq B,
Gala JL. 2011. Impact of the spotted microarray preprocess-
ing method on fold-change compression and variance
stability. BMC Bioinformatics 12:413.
82. Hsiao L-L, Jensen RV, Yoshida T, Clark KE, Blumen-
stock JE, Gullans SR. 2002. Correcting for signal satura-
tion errors in the analysis of microarray data.
BioTechniques 32:330–336.
83. Kauffmann A, Huber W. 2010. Microarray data quality
control improves the detection of differentially expressed
genes. Genomics 95:138–142.
84. Fujita A, Sato JR, Da Silva FH, Galvão MC. 2009. Quality
control and reproducibility in DNA microarray experi-
ments. Genome Inform 23:21–31.
85. Shi L, Tong W, Goodsaid FM, Frueh FW, Fang H, Han T,
Fuscoe JC, Casciano DA. 2004. QA/QC: challenges and
pitfalls facing the microarray community and regulatory
agencies. Exp Rev Mol Diagn 4:761–777.
86. Shi L, Jones WD, Jensen RV, Harris SC, Perkins RG,
Goodsaid FM, Guo L, Croner LJ, Boysen C, Fang H,
Qian F, Amur S, Bao W, Barbacioru CC, Bertholet V,
Cao XM, Chu TM, Collins PJ, Fan XH, Frueh FW, Fus-
coe JC, Guo X, Han J, Herman D, Hong H, Kawasaki ES,
Li QZ, Luo Y, Ma Y, Mei N, Peterson RL, Puri RK,
Shippy R, Su Z, Sun YA, Sun H, Thorn B, Turpaz Y,
Wang C, Wang SJ, Warrington JA, Willey JC, Wu J,
Xie Q, Zhang L, Zhang L, Zhong S, Wolfinger RD,
Tong W. 2008. The balance of reproducibility, sensitivity,
and specificity of lists of differentially expressed genes in
microarray studies. BMC Bioinformatics 9:S10.
87. Shi L, Perkins RG, Fang H, Tong W. 2008. Reproducible
and reliable microarray results through quality control:
good laboratory proficiency and appropriate data analysis
practices are essential. Curr Opin Biotechnol 19:10–18.
88. MAQC Consortium: Shi L, Campbell G, Jones WD,
Campagne F, Wen Z, Walker SJ, Su Z, Chu TM, Good-
said FM, Pusztai L, Shaughnessy JDJ, Oberthuer A, Tho-
mas RS, Paules RS, Fielden M, Barlogie B, Chen W, Du
P, Fischer M, Furlanello C, Gallas BD, Ge X, Megherbi
DB, Symmans WF, Wang MD, Zhang J, Bitter H, Brors
B, Bushel PR, Bylesjo M, Chen M, Cheng J, Cheng J,
Chou J, Davison TS, Delorenzi M, Deng Y, Devanarayan
V, Dix DJ, Dopazo J, Dorff KC, Elloumi F, Fan J, Fan S,
Fan X, Fang H, Gonzaludo N, Hess KR, Hong H, Huan J,
et al. 2010. The MicroArray Quality Control (MAQC)-II
study of common practices for the development and valida-
tion of microarray-based predictive models. Nat Biotechnol
28:827–838.
89. Lippa KA, Duewer DL, Salit ML, Game L, Causton HC.
2010. Exploring the use of internal and externalcontrols for
assessing microarray technical performance. BMC Res
Notes 3:349.
90. Ritari J, Paulin L, Hultman J, Auvinen P. 2009. Applica-
tion of hybridization control probe to increase accuracy on
ligation detection or minisequencing diagnostic microar-
rays. BMC Res Notes 2:249.
91. Hsiao CR, Chen CH. 2009. Characterization of DNA
chips by nanogold staining. Anal Biochem 389:118–123.
92. Peterson G, Bai J, Narayanan S. 2009. A co-printed
oligomer to enhance reliability of spotted microarrays.
J Microbiol Meth 77:261–266.
93. Scott DJ, Devonshire AS, Adeleye YA, Schutte ME,
Rodrigues MR, Wilkes TM, Sacco MG, Gribaldo L, Fab-
bri M, Coecke S, Whelan M, Skinner N, Bennett A,
White A, Foy CA. 2011. Inter- and intra-laboratory study
to determine the reproducibility of toxicogenomics data-
sets. Toxicology 290:50–58.
94. Trachtenberg AJ, Robert JH, Abdalla AE, Fraser A, He
SY, Lacy JN, Rivas-Morello C, Truong A, Hardiman
G, Ohno-Machado L, Liu F, Hovig E, Kuo WP. 2012.
A primer on the current state of microarray technologies.
Methods Mol Biol 802:3–17.
95. Dorris DR, Ramakrishnan R, Trakas D, Dudzik F, Belval
R, Zhao C, Nguyen A, Domanus M, Mazumder A. 2002.
A highly reproducible, linear, and automated sample prep-
aration method for DNA microarrays. Genome Res 12:
976–984.
96. Regan J, Létant S, Adams K, Nguyen N, Derlet R,
Cohen S, Vitalis E, Tammero L, Ortiz J, McBride M, Birch
J. 2010. A sample-in-answer-out instrument for the
detection of multiple respiratory pathogens in unprepared
nasopharyngeal swab samples. Analyst 135:2316–2322.
97. Liu RH, Dill K, Fuji HS, McShea A. 2006. Integrated
microfluidic biochips for DNA microarray analysis. Expert
Rev Mol Diagn 6:253–261.
98. Raymond F, Carbonneau J, Boucher N, Robitaille L,
Boisvert S, Wu W-K, De Serres G, Boivin G, Corbeil J.
2009. Comparison of automated microarray detection with
real-time PCR assays for detection of respiratory viruses in
specimens obtained from children. J Clin Microbiol 47:
743–750.
99. Kumar S, Wang L, Fan J, Kraft A, Bose ME, Tiwari S,
Van Dyke M, Haigis R, Luo T, Ghosh M, Tang H, Hagh-
nia M, Mather EL, Weisburg WG, Henrickson KJ.
2008. Detection of 11 common viral and bacterial
pathogens causing community-acquired pneumonia or
sepsis in asymptomatic patients by using a multiplex
reverse transcription-PCR assay with manual (enzyme
hybridization) or automated (electronic microarray) detec-
tion. J Clin Microbiol 46:3063–3072.
100. Foglieni B, Brisci A, San Biagio F, Di Pietro P, Petralia S,
Conoci S, Ferrari M, Cremonesi L. 2010. Integrated PCR
amplification and detection processes on a Lab-on-Chip
platform: a new advanced solution for molecular diagnos-
tics. Clin Chem Lab Med 48:329–336.
101. Njoroge SK, Chen HW, Witek MA, Soper SA. 2011.
Integrated microfluidic systems for DNA analysis. Topics
Curr Chem 304:203–260.
102. Teo J, Pietro PD, Biagio FS, Capozzoli M, Deng YM, Barr
I, Caldwell N, Ong KL, Sato M, Tan R, Lin R. 2011. Ver-
eFlu™: an integrated multiplex RT-PCR and microarray
assay for rapid detection and identification of human influ-
enza A and B viruses using lab-on-chip technology. Arch
Virol 156:1371–1378.
103. Summerer D, Hevroni D, Jain A, Oldenburger O, Parker
J, Caruso A, Stähler CF, Stähler PF, Beier M. 2010. A
flexible and fully integrated system for amplification, detec-
tion and genotyping of genomic DNA targets based on
microfluidic oligonucleotide arrays. New Biotechnol 27:
149–155.
104. Liu RH, Yang J, Lenigk R, Bonanno J, Grodzinski P.
2004. Self-contained, fully integrated biochip for sample
preparation, polymerase chain reaction amplification, and
DNA microarray detection. Anal Chem 76:1824–1831.
105. Liu RH, Lodes MJ, Nguyen T, Siuda T, Slota M, Fuji
HS, McShea A. 2006. Validation of a fully integrated
microfluidic array device for influenza A subtype identifica-
tion and sequencing. Anal Chem 78:4184–4193.
106. Trau D, Lee TMH, Lao AIK, Lenigk R, Hsing I-M, Ip
NY, Carles MC, Sucher NJ. 2002. Genotyping on a com-
plementary metal oxide semiconductor silicon polymerase
chain reaction chip with integrated DNA microarray. Anal
Chem 74:3168–3173.
107. Yeung S-W, Lee TM-H, Cai H, Hsing IM. 2006. A DNA
biochip for on-the-spot multiplexed pathogen identifica-
tion. Nucl Acids Res 34:e118.
2.3.3-12 ▪ TARGET-SPECIFIC DETECTION
108. Regan JF, Makarewicz AJ, Hindson BJ, Metz TR, Gutier-
rez DM, Corzett TH, Hadley DR, Mahnke RC, Henderer
BD, Breneman JW, Weisgraber TH, Dzenitis JM. 2008.
Environmental monitoring for biological threat agents
using the autonomous pathogen detection system with
multiplexed polymerase chain reaction. Anal Chem 80:
7422–7429.
109. Hindson BJ, Makarewicz AJ, Setlur US, Henderer BD,
McBride MT, Dzenitis JM. 2005. APDS: the autonomous
pathogen detection system. Biosens Bioelectron 20:1925–
1931.
110. Preston CM, Marin 3rd R, Jensen SD, Feldman J, Birch
JM, Massion EI, Delong EF, Suzuki M, Wheeler K, Scho-
lin CA. 2009. Near real-time, autonomous detection of
marine bacterioplankton on a coastal mooring in Monterey
Bay, California, using rRNA-targeted DNA probes. Envi-
ron Microbiol 11:1168–1180.
111. Scholin C, Doucette G, Jensen S, Roman B, Pargett D,
Marin 3rd R, Preston C, Jones W, Feldman J, Everlove
C, Harris A, Alvarado N, Massion E, Birch J, Greenfield
D, Vrijenhoek R, Mikulski C, Jones K. 2009. Remote
detection of marine microbes, small invertebrates, harmful
algae and biotoxins using the Environmental Sample Pro-
cessor (ESP). Oceanography 22:158–167.
112. Stamm S, Brosius J. 1991. Sanchored PCR: PCR with
cDNA coupled to a solid phase. Nucl Acids Res 19:1350.
113. Erdogan F, Kirchner R, Mann W, Ropers H-H, Nuber
UA. 2001. Detection of mitochondrial single nucleotide
polymorphisms using a primer elongation reaction on oli-
gonucleotide microarrays. Nucl Acids Res 29:e36.
114. Shapero MH, Leuther KK, Nguyen A, Scott M, Jones
KW. 2001. SNP genotyping by multiplexed solid-phase
amplification and fluorescent minisequencing. Genome
Res 11:1926–1934.
115. Lockley AK, Jones CG, Bruce JS, Franklin SJ, Bardsley
RG. 1997. Colorimetric detection of immobilised PCR
products generated on a solid support. Nucl Acids Res 25:
1313–1314.
116. Adessi C, Matton G, Ayala G, Turcatti G, Mermod J-J,
Mayer P, Kawashima E. 2000. Solid phase DNA amplifica-
tion: characterisation of primer attachment and amplifica-
tion mechanisms. Nucl Acids Res 28:e87.
117. Adams CP, Kron SJ. June 24, 1997. Method for perform-
ing amplification of nucleic acid with two primers bound to
a single solid support. US patent 5,641,658.
118. Onodera K, d’Offay J, Melcher U. 2002. Nylon mem-
brane-immobilized PCR for detection of bovine viruses.
Biotechniques 32:74–80.
119. Westin L, Xu X, Miller C, Wang L, Edman CF, Nerenberg
M. 2000. Anchored multiplex amplification on a micro-
electronic chip array. Nat Biotechnol 18:199–204.
120. Strizhkov BN, Drobyshev AL, Mikhailovich VM, Mirza-
bekov AD. 2000. PCR amplification on a microarray of
gel-immobilized oligonucleotides: detection of bacterial
toxin- and drug-resistant genes and their mutations. Bio-
Techniques 29:844–857.
121. Turner MS, Penning S, Sharp A, Hyland VJ, Harris R,
Morris CP, van Daal A. 2001. Solid-phase amplification
for detection of C282y and H63D hemochromatosis (HFE)
gene mutations. Clin Chem 47:1384–1389.
122. Huber M, Losert D, Hiller R, Harwanegg C, Mueller
MW, Schmidt WM. 2001. Detection of single base altera-
tions in genomic DNA by solid phase polymerase chain
reaction on oligonucleotide microarrays. Anal Biochem
299:24–30.
123. Tillib SV, Strizhkov BN, Mirzabekov AD. 2001. Integra-
tion of multiple PCR amplifications and DNA mutation
analyses by using oligonucleotide microchip. Anal Biochem
292:155–160.
124. Mitterer G, Huber M, Leidinger E, Kirisits C, Lubitz W,
Mueller MW, Schmidt WM. 2004. Microarray-based iden-
tification of bacteria in clinical samples by solid-phase PCR
amplification of 23S ribosomal DNA sequences. J Clin
Microbiol 42:1048–1057.
125. Mitterer G, Schmidt WM. 2006. Microarray-based
detection of bacteria by on-chip PCR. Meth Mol Biol
345:37–51.
126. Pemov A, Modi H, Chandler DP, Bavykin S. 2005. DNA
analysis with multiplex microarray-enhanced PCR. Nucl
Acids Res 33:e11.
127. Sun Y, Dhumpa R, Bang DD, Handberg K, Wolff A.
2011. DNA microarray-based solid-phase RT-PCR for
rapid detection and identification of influenza virus type A
and subtypes H5 and H7. Diagn Microbiol Infect Dis 69:
432–439.
128. von Nickisch-Rosenegk M, Marschan X, Andresen D,
Bier FF. 2008. Reverse transcription-polymerase chain
reaction on a microarray: the integrating concept of “active
arrays.” Anal Bioanal Chem 391:1671–1678.
129. Andresen D, Von Nickisch-Rosenegk M, Bier FF. 2009.
Helicase-dependent amplification: use in OnChip amplifi-
cation and potential for point-of-care diagnostics. Expert
Rev Mol Diagn 9:645–650.
130. Vora GJ, Meador CE, Stenger DA, Andreadis JD. 2004.
Nucleic acid amplification strategies for DNA microarray-
based pathogen detection. Appl Environ Microbiol 70:
3047–3054.
131. Fredriksson S, Banér J, Dahl F, Chu A, Ji H, Welch K,
Davis RW. 2007. Multiplex amplification of all coding
sequences within 10 cancer genes by Gene-Collector.
Nucl Acids Res 35:e47.
132. Borel N, Kempf E, Hotzel H, Schubert E, Torgerson P,
Slickers P, Ehricht R, Tasara T, Pospischil A, Sachse
K. 2008. Direct identification of chlamydiae from clinical
samples using a DNA microarray assay: a validation study.
Mol Cell Probes 22:55–64.
133. Khodakov DA, Zakharova NV, Gryadunov DA, Filatov
FP, Zasedatelev AS, Mikhailovich VM. 2008. An oligo-
nucleotide microarray for multiplex real-time PCR identi-
fication of HIV-1, HBV, and HCV. BioTechniques 44:
241–248.
134. Pierik A, Boamfa M, van Zelst M, Clout D, Stapert H,
Dijksman F, Broer D, Wimberger-Friedl R. 2012. Real
time quantitative amplification detection on a microarray:
towards high multiplex quantitative PCR. Lab Chip 12:
1897–1902.
135. Yershov G, Barsky V, Belgovskiy A, Kirillov E, Krein-
dlin E, Ivanov I, Parinov S, Guschin D, Drobishev A,
Dubiley S, Mirzabekov A. 1996. DNA analysis and diag-
nostics on oligonucleotide microchips. Proc Natl Acad Sci
USA 93:4913–4918.
136. Cooney CG, Sipes D, Thakore N, Holmberg R, Belgrader
P. 2012. A plastic, disposable microfluidic flow cell for
coupled on-chip PCR and microarray detection of infec-
tious agents. Biomed Microdev 14:45–53.
137. Chandler DP, Kukhtin A, Mokhiber R, Knickerbocker
C, Ogles D, Rudy G, Golova J, Long P, Peacock A.
2010. Monitoring microbial community structure and
dynamics during in situ U(VI) bioremediation with a field-
portable microarray analysis system. Environ Sci Technol 44:
5516–5522.
138. Chandler DP, Knickerbocker C, Bryant L, Golova J,
Wiles C, Williams KH, Peacock AD, Long PE. 2013.
Profiling in situ microbial community structure with an
amplification microarray. Appl. Environ. Microbiol. 79:
799–807.
139. Linz U. 1990. Thermocycler temperature variation inva-
lidates PCR results. Biotechniques 9:286–293.
 2.3.3. Microarray-Based Environmental Diagnostics ▪ 2.3.3-13
140. Kopp DW, Sansieri CA, Mifflin TE. 1994. Routine mon-
itoring of temperatures inside thermal cycler blocks for
quality control. Clin Chem 40:2117–2119.
141. Tweed GP, Whitney J, Bloch PL. 1991. Temperature
cycler evaluation: what do you need to know? Biotechniques
10:526–532.
142. Chen PC, Nikitopoulos DE, Soper SA, Murphy MC.
2008. Temperature distribution effects on micro-CFPCR
performance. Biomed Microdev 10:141–152.
143. Saunders GC, Dukes J, Parkes HC, Cornett JH. 2001.
Interlaboratory study on thermal cycler performance in
controlled PCR and random amplified polymorphic
DNA analyses. Clin Chem 47:47–55.
144. Kim YH, Yang I, Bae YS, Park SR. 2008. Performance
evaluation of thermal cyclers for PCR in a rapid cycling
condition. Biotechniques 44:495–500.
145. Yang I, Kim YH, Byun JY, Park SR. 2005. Use of multi-
plex polymerase chain reactions to indicate the accuracy of
the annealing temperature of thermal cycling. Anal Bio-
chem 338:192–200.
146. Von Keyserling H, Bergmann T, Wiesel M, Kaufmann
AM. 2011. The use of melting curves as a novel approach
for validation of real-time PCR instruments. Biotechniques
51:179–184.
147. Zuna J, Muzikova K, Madzo J, Krejci O, Trka J. 2002. Tem-
perature non-homogeneity in rapid airflow-based cycler sig-
nificantly affects real-time PCR. Biotechniques 33:508–512.
148. Unc A, Zurek L, Peterson G, Narayanan S, Springthorpe
SV, Sattar SA. 2012. Microarray assessment of virulence,
antibiotic, and heavy metal resistance in an agricultural
watershed creek. J Environ Qual 41:534–543.
149. Volokhov DV, Kong H, Herold KE, Chizhikov VE,
Rasooly A. 2011. Oligonucleotide microarrays for identifi-
cation of microbial pathogens and detection of their viru-
lence-associated or drug-resistance determinants. Meth
Mol Biol 671:55–94.
150. Martinez G, Bruant G, Brousseau R, Masson L, Harel J.
2006. Development of a new integrated diagnostic test for
identification and characterization of pathogens. Dev Biol
(Basel) 126:213–218.
151. Peterson G, Bai J, Nagaraja TG, Narayanan S. 2010.
Diagnostic microarray for human and animal bacterial dis-
eases and their virulence and antimicrobial resistance
genes. J Microbiol Meth 80:223–230.
152. Dubois JW, Hill S, England LS, Edge T, Masson L, Tre-
vors JT, Brousseau R. 2004. The development of a DNA
microarray-based assay for the characterization of commer-
cially formulated microbial products. J Microbiol Meth 58:
251–262.
153. Fu GK, Hu J, Wang PH, Fodor SPA. 2011. Counting
individual DNA molecules by the stochastic attach-
ment of diverse labels. Proc Natl Acad Sci USA 108:
9027–9031.
154. Delattre C, Allier CP, Fouillet Y, Jary D, Bottausci F,
Bouvier D, Delapierre G, Quinaud M, Rival A, Davoust
L, Peponnet C. 2012. Macro to microfluidics system for
biological environmental monitoring. Biosens Bioelectron
36:230–235.
 Field Application of Pathogen Detection
Technologies
TIMOTHY M. STRAUB, DOUGLAS R. CALL, CINDY BRUCKNER-LEA,
HEATHER COLBURN, CHERYL L. BAIRD, RACHEL A. BARTHOLOMEW,
RICHARD OZANICH, AND KRISTIN JARMAN
2.3.4
Over the past 10 years there has been a significant increase
in commercial products designed for field-based detection
of microbial pathogens. This is due in part to the anthrax
attacks in the United States in 2001 and the need for first
responders to quickly identify the composition of suspected
white powders and other potential biothreats. Demand for
rapid detection is also driven by the need to ensure safe
food, water, and environmental systems. From a technology
perspective, rapid identification methods have largely capi-
talized on PCR and other molecular recognition techniques
that can be deployed as robust field instrumentation. Exam-
ples of the relevant needs include the ability to (1) declare
a water distribution system free of microbial pathogens after
a pipe/main break repair; (2) assess risks of contamination,
such as when produce production and processing plants are
located near concentrated animal feeding operations; (3)
evaluate the safety of ready-to-eat products; (4) determine
the extent of potential serious disease outbreaks in remote
or disaster-stricken areas where access to clinical laboratories
is not an immediate option; and (5) quickly assess credible
biological terrorism events.
Many of the principles underlying rapid detection meth-
ods are derived from methods for environmental micro-
biology, but there is a dearth of literature describing and
evaluating field-based detection systems. Thus, the aims of
this chapter are to (1) summarize the different kinds of com-
mercially available sampling kits and field-based biological
detectors; (2) highlight some of the continued challenges
of sample preparation to stimulate new research toward min-
imizing the impact of inhibitors on PCR-based detection sys-
tems; (3) describe our general rationale and statistically based
approach for instrument evaluation; (4) provide statistical
and spatial guidelines for developing valid sampling plans;
and (5) summarize some current needs and emerging tech-
nologies. This information is presented to highlight the state
of the field and major questions that students may wish to
consider investigating further. Where possible we cite studies
that have been conducted and published either in traditional
peer-reviewed or other literature (e.g., AOAC International
Methods).
doi:10.1128/9781555818821.ch2.3.4
2.3.4-1
COMMERCIAL OFF-THE-SHELF FIELD-BASED
BIODETECTION TOOLS
Field-based biodetection tools are varied, and in this sec-
tion we describe commercially available sampling systems
and portable, field-based systems that can be used to screen
samples for biological materials and pathogens. Screening
approaches range from general methods that detect the pres-
ence of proteins, DNA, or ATP, to specific pathogen identi-
fication based on immunoassays or DNA sequence analysis
using PCR. In general, the first responder community does
not have the financial resources to buy sophisticated equip-
ment, and the high cost and limited shelf life of reagents
make ownership and use of many systems cost prohibitive
for many jurisdictions. In addition, there is little peer-
reviewed information regarding the performance (sensitivity
and specificity) of instruments and test kits, and training
for use of these instruments and kits is lacking. The following
discussion illustrates the current state of knowledge regard-
ing sample collection systems and hand-held/field-based
detectors.
Sampling Kits Used in Conjunction with
Field-Based Detection Systems
The sampling kits listed in Table 1 are designed primarily to
sample “white powders.” With a few exceptions (e.g., the Bio-
Fire Film Array that incorporates sample processing to remove
inhibitors prior to the PCR, and the Lincoln Nucleic Acid
Kit [LiNK 2.0] [1] that was evaluated in the Lincoln Labora-
tory’s evaluation of BioFire’s RAZOR system), the kits do not
include any removal of potential assay inhibitors. The kits are
designed to suspend suspect material in a buffered solution for
downstream analysis. Whereas the kits themselves have not
been formally evaluated, principal components of the kits
(e.g., swabs, wipes, and sponges) have been evaluated for
their ability to recover Bacillus spp. spores from various surfa-
ces (2–7). In some cases, the protocols and components
provided in the kits are designed to meet standards for law
enforcement and chain-of-custody requirements for the
criminal investigation of a biological attack (8–10).
 2.3.4-2 ▪ TARGET-SPECIFIC DETECTION
TABLE 1 Overview of products for sampling and detection of potential biothreats (12)
Instrument/kit Manufacturer Company website Type of assay Cost Notes
Alexeter Collection Swab Alexeter http://www.alexeter.com/biow/index.asp Sampling kit $75/box of 25 Compatible with PCR and immunoassay
Sample Collection and ASD BioSystems http://www.asdbiosystems.com/p1.htm Sampling kit $12 each Compatible with PCR and immunoassay
Recovery Device (SCRD)
SWIPE Kits New Horizons http://www.nhdiag.com/profile_one.shtml Sampling kit Kit dependent Compatible with PCR and immunoassay
Diagnostics $13.20–$23.10
B2C Bulk Sample QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $26.71 each Meets ASTM Standard E 2458-10 for bulk
Collection Kit sample collection
Biological Sampling Kit QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $31.50 each Large area sampling kit compatible with PCR
(BiSKit)–Large Area and immunoassay
QSA Model 102 Full QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $2961.80 each Large area sampling kit developed and used by
Forensic Analytical U.S. Army Mobile Labs and Kits Team
Center (FAC)
Mini Push Pack Kits QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $36.70 each Single-use kits
Wipe/Bio/Solid/ Liquid
Small Area Sampling (SAS) QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $17.96 each Designed for small area sampling; 4-year
Kit shelf-life
S3 Bio Sampler QuickSilver Analytics http://www.chembiokits.com/ Sampling kit $22.56 each PBS buffer-moistened sponge for sampling;
compatible with PCR and immunoassay
LiNK2 Lincoln Nucleic Lincoln Laboratory Sampling kit unknown Removes some PCR inhibitors; not compatible
Acid Kit with immunoassay
HAZCAT WMD Kit HazTech Systems http://www.hazcat.com Assorted $4,910.17 An all-in-one chemical/biological/
radiological/nuclear/explosive detection and
classification kit
Prime Alert GenPrime http://www.genprime.com/ Nonspecific biological $12,000 reader; Nonspecific DNA colorimetric test with a
(DNA) $200/test separate immunoassay toxin test
BioCheck Powder Screening 20/20 Bioresponse http://biocheckinfo.com/ Nonspecific biological $687.50/25 kits Nonspecific colorimetric protein detection and
Kit ( protein & pH) pH test
Indipro Macherey-Nagel http://www.mn-net.com/tabid/10441/default. Nonspecific biological $103/50 strips Nonspecific protein detection
aspx ( protein)
Profile 1 New Horizons http://www.nhdiag.com/profile_one.shtml Nonspecific biological $5,000 reader: Nonspecific test for live bacterial cells using
Diagnostics (ATP) $450/100 tests ATP
HazMatID 360 Smiths Detection http://www.smithsdetection.com Nonspecific biological $53,000 Nonspecific test for biological material
(FT-IR/protein) ( protein-based) using FT-IR spectroscopy
HazMatID Ranger Smiths Detection http://www.smithsdetection.com Nonspecific biological $35,000 Nonspecific test for biological material
(FT-IR/protein) ( protein-based) using FT-IR spectroscopy
HazMatID Elite Smiths Detection http://www.smithsdetection.com Nonspecific biological $50,000 Nonspecific test for biological material
(FT-IR/protein) ( protein-based) using FT-IR spectroscopy
TruDefender FT/FTi Thermo Scientific http://www.thermoscientific.com Nonspecific biological $45,000–$46,500 Nonspecific test for biological material
(formerly Ahura (FT-IR/protein) ( protein-based) using FT-IR spectroscopy
Scientific)
BADD Biowarfare Agent Advnt http://www.advnt.org Immunoassay $24.50/test Single-agent lateral flow immunoassay
Detection Devices
Pro Strips Advnt http://www.advnt.org Immunoassay $699.50/10 tests Five-agent lateral flow immunoassay
BioDetect Test Strips Alexeter http://www.alexeter.com/biow/products.asp Immunoassay $685/25 tests Single-agent lateral flow immunoassay
Defender TSR reader for use Alexeter http://www.alexeter.com/biow/products.asp Immunoassay reader $6,500 reader Handheld optical reader for single-agent
with BioDetect Test BioDetect test strips using Windows Mobile
Strips interface
Guardian Reader for use Alexeter http://www.alexeter.com/biow/products.asp Immunoassay reader $9,995 reader Optical reader for single-agent BioDetect
with BioDetect Test Alexeter test strips
Strips
RAID Multi-Test Strips Alexeter http://www.alexeter.com/biow/products.asp Immunoassay $695–$995/10 Multiplexed lateral flow immunoassay available
tests as either five-plex or eight-plex
ENVI Assay System Environics http://www.environicsusa.com Immunoassay $400–$450/10 Toxin assays can be read visually or with a
tests reader ($3,452), while the ENVI-FL
pathogen assays require a reader
BIOSENSOR 2200R MSA/QTL http://www.msafire.com/ Immunoassay, $16,450 Fluorescence-based automated immunoassay
automated $80/single- detection system
agent test

2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-3

Smart II New Horizons http://www.nhdiag.com/profile_one.shtml Immunoassay $550/25 tests Single-agent lateral flow immunoassay strips
Diagnostics
RAMP Response Biomedical http://www.responsebio.com Immunoassay and $11,333 Fluorescence-based lateral flow immunoassay
reader $27/test strips and reader
(anthrax)
RAPTOR: Portable, Research International http://resrchintl.com/ Immunoassay, $50,000 Portable automated immunoassay system
Multianalyte Bioassay automated $200/test designed for the U.S. military
Detection System
NIDS ANP Technologies/ http://www.smithsdetection.com Immunoassay and $7,500, Multiplexed lateral flow immunoassay with
Smiths Detection reader $45–$75/assay optical reader
BioThreat Alert and Reader Tetracore http://www.tetracore.com/bio-warfare/ Immunoassay and $5,500, Single-agent lateral flow immunoassay with
reader $605/box of 25 optical reader
FilmArray BioFire Diagnostics http://www.biofiredx.com PCR $49,500, Sample-to-answer PCR system, detects 21
(formerly Idaho $180/pouch targets per pouch
Technology)
R.A.P.I.D. (JBAIDS) BioFire Diagnostics http://www.biofiredx.com PCR $55,000, Portable PCR system designed for the U.S.
(formerly Idaho $452/48 assays military; multiple assays available
Technology)
RAZOR EX BioFire Diagnostics http://www.biofiredx.com PCR $38,500, Portable PCR system detects 10 targets per
(formerly Idaho $200/pouch pouch
Technology)
Bio-Seeq PLUS Smiths Detection http://www.smithsdetection.com PCR $45,000, Portable PCR system, uses single-agent assays
$30/assay
T-COR 4 Tetracore http://www.tetracore.com/t-cor/index.html PCR $16,000, Portable PCR system, uses single-agent assays
$12/cartridge
2.3.4-4 ▪ TARGET-SPECIFIC DETECTION
Most of the available literature on sampling materials
for rapid detection concerns recovery efficiency. Recovery
efficiency is affected by a number of factors, including the
sampling materials, surface area covered (2), type of inani-
mate surface (stainless steel, tile, carpet, drywall, etc.), how
samples will be assayed, and even spore deposition (2–7).
Due to the multifactorial nature of the sampling and preser-
vation issues, these studies offer little consensus on best prac-
tices for sampling white powders from suspected incidents.
For example, Edmonds et al. (6) compared liquid deposition
followed by drying to a dry aerosol with gravity settling on
multiple surfaces (glass, chemical agent resistant coated
[CARC] steel, polycarbonate, and vinyl tile, and four differ-
ent swab types: cotton, rayon, Dacron, and macrofoam). In
general, there appeared to be few differences between differ-
ent types of swabs. As expected, recovery varied between
different materials. Recovery was greater for liquid deposition
on glass and polycarbonate, but less for CARC-painted steel
and vinyl tile. Recoveries were generally >50% regardless of
deposition method, but the tested surfaces were nonporous.
Valentine et al. (5) employed an ink-jet-type aerosol deposi-
tion to various surfaces. In general, recovery of the Bacillus
subtilis spores was <10%, and several of their matrices (carpet
and upholstery) were extremely porous.
With these studies it is important to emphasize that there
is an underlying assumption that high spore/agent concen-
trations will be present in a small area. Buttner et al. (2) eval-
uated the BiSKit system for a larger area (1 m2) and compared
results to swabs. Ten times more spores were recovered using
the BiSKit system, with detection sensitivity of 42 ± 5.8
CFU/m2 for wet sampling, 100.5 ± 10.2 CFU/m2 for dry sam-
pling, and a recovery efficiency ranging from 11% to 18%.
When visible powder is present, the spore levels will be
many orders of magnitude higher and much easier to sample.
Regardless of the sampling kit employed, practitioners are
cautioned to make sure that when buffers are used the final
sample is compatible with downstream detection methods.
When the information is available, we have indicated the
compatibility of sampling kits with downstream analyses
(Table 1).
Nonspecific Biological Tests
Some jurisdictions in the first responder community use non-
specific biological tests as their first screen in a white powder
incident (11). In general these tests can be categorized as
detectors for protein, DNA (non-PCR based), and detection
of adenosine triphosphate (ATP) (12). In some cases, juris-
dictions use Fourier transform infrared (FT-IR) spectroscopy,
which not only tests for the presence of proteins but can also
identify the type of protein in the sample (12). An overview
of the key methods is provided next. References to detection
of biological agents merely refers to the ability to detect pro-
tein, DNA, or ATP derived from the organism, not the spe-
cific detection and identification of an organism of interest.
Protein
Assays such as the BioCheck (20/20 GeneSystems, Rockville,
MD) are designed to detect any type of protein in a sample,
including milk proteins in coffee creamer and powdered
infant formula. Because biological material is typically neutral
in pH, these assays may also include a pH test. Samples are
added and mixed, and the operator reads color change as
indicators of protein and pH after approximately 5 min. Poore
et al. (11) studied the BioCheck system and compared it
with the Prime Alert (GenPrime, Spokane, WA) for total
DNA analysis and the Profile-1 (New Horizons Diagnostics,
Columbia, MD) for ATP analysis. The BioCheck system was
capable of detecting B. anthracis spores at a concentration
of 1 × 107 to 1 × 108 spores per ml depending on the number
of spore washes (2× and 4×, respectively). This detection
limit is about 1,000 to 10,000 times the estimated infectious
dose and would barely be visible as a white powder to the
naked eye (<1 mg). An attenuated Yersinia pestis strain could
be detected at 1 × 107 CFU/ml, and ricin toxin could be
detected at 100 μg. As expected, the BioCheck cross-reacted
with white powders containing various proteins (e.g., yeast,
coffee creamer). Nevertheless, the test also worked when
the spores, cells, or toxin were spiked into a variety of powders
at 1× to 5× limit of detection (LOD) levels (LOD is defined
here as the detection limit of the assay determined when test-
ing with pure spores, cells, and toxin). While lacking specif-
icity, the BioCheck protein test has a relatively low price per
analysis ($40 per assay and shelf life of 12 months) making it
attractive to the first responder community.
Total DNA Test
Non-PCR DNA tests employ DNA intercalating dyes to
detect the presence of double-stranded DNA (11, 12) where
a battery-powered fluorescence reader is required. The time
needed to go from sample to answer for most tests is 5 min.
Poore et al. (11) evaluated the Prime Alert system (Gen-
Prime) for the “detection” of B. anthracis, Y. pestis, and ricin.
Although the Prime Alert system also contains immunoassay
test strips to detect ricin and botulinum toxin, Poore’s study
focused on the DNA detection component. With 4× washed
B. anthracis spores, this system was unable to detect 1 × 109
spores/ml, but in a less purified preparation of spores (2×
wash), spores were minimally detected at a 1 × 109 spores/
ml concentration, which is approximately 1–10 mg of a puri-
fied white powder. Y. pestis could be detected at a concentra-
tion of 1 × 108 CFU/ml, and as expected, ricin could not be
detected using the DNA test. For unspiked white powders,
yeast and Dipel could be detected because these both contain
DNA. When B. anthracis spores (either 4× or 2× washed) or
Y. pestis cells were spiked into the white powders panel, DNA
was nominally detected in all matrices at 5× LOD where the
LOD was determined from tests with purified spores or cells.
DNA based systems typically cost more per assay than protein
tests. The cost of the fluorometer for sample readout is
approximately $12,000, and cost per assay is about $200.
Shelf life of the reagents is approximately 12 months.
ATP Test
ATP tests work by detecting metabolic activity of living cells.
These tests typically employ a control sample that has not
been incubated. The test sample is the same sample that
has been incubated for a short period of time (approximately
20 min; 11) where the relative light units (RLUs) for unincu-
bated sample are subtracted from the RLU of the incubated
sample to obtain an RLU value. Poore et al. (11) evaluated
the Profile-1 system (New Horizons Diagnostics) as a com-
parison to the Prime Alert and BioCheck systems. Of the
three tests—protein, DNA, and ATP—Poore et al. (11)
found that the ATP test was the most sensitive of the three
tests. For both 2× and 4× washed spores, the ATP test could
reliably detect 100,000 spores/ml. Lower concentrations
were not measured, but the ATP test most likely has a
nominal detection limit of approximately 10,000 spores/ml.
If B. anthracis was a pure white powder, 10,000/ml spores
would be equal to approximately one infectious dose, and
this amount of spores would not be visible to the unaided
eye. The detection limit for Y. pestis was 1 × 106 CFU/ml,
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-5
and purified ricin protein was not detected (as expected)
because it is not a metabolically active product. When tested
against unspiked powders, yeast provided an intense positive
result, but the RLU for the unincubated control and the RLU
of the incubated test sample were too close to declare a true
positive result. Dipel, which is a biological insecticide con-
taining Bacillus thuringiensis, was also positive. When tested
with other unspiked powders, there was some ATP presence
noted with flour and talcum powder, indicating some cross-
reactivity and/or nonspecificity. When spores or cells were
spiked at their LOD into the panel of 16 different powders,
all produced a positive result at LOD except for samples spiked
into “Powder C” and bentonite. For Powder C, the 2× washed
spores could only be detected in the 5× LOD spike, but
Y. pestis could not be detected at all in this sample. In the
case of bentonite, the samples clogged the filter and a result
could not be determined. Startup costs for the system include
the luminometer to read fluorescence ($5,000), and the cost
of the assay itself is approximately $5, and the reagents have a
shelf life of 12 months.
FT-IR
FT-IR is a reagent-free detection system for chemicals in
liquid and solid samples. Some first responder jurisdictions
can afford this instrument (approximately $35,000 to
$90,000, depending on the unit), and they use this system
to screen potentially hazardous samples (e.g., chemicals,
explosives). In addition, because it is a spectroscopic method,
there are no concerns for reagent shelf life. When a sample is
analyzed, its spectral trace is compared to a stored library
of spectra to putatively identify the chemical of concern.
Protein content must be at least 10%, and this would be
important in terms of identifying a biological sample
(B. anthracis, etc.) or toxin (ricin). Unfortunately, there are
no published data evaluating these instruments for biological
agent or toxin analysis, and the FT-IR spectral libraries are
not specifically developed for the positive identification of
biological agents or toxins of concern (13).
SPECIFIC BIOLOGICAL TESTS
Immunoassays
The two formats used for specific detection of biological
agents are immunoassays and PCR. Some immunoassays are
truly hand-held and portable, such as lateral flow assays
(LFAs). Like home pregnancy test kits where the readout con-
sists of a sample line and control line, a positive reaction for
the biological agent is declared when the sample line and con-
trol line are both positive. The second advantage of LFAs is
that sample-to-answer time is relatively short (approximately
5–15 min). This is a potential consideration for first res-
ponder/hazmat teams in Level A suits where the amount of
breathable air is typically ≤45 min. As discussed shortly, the
AOAC has established criteria for the performance of hand-
held immunoassays that are used to investigate white powder
incidents. As defined by the AOAC, the acceptable mini-
mum detection limit (AMDL) for B. anthracis is currently
107 spores/ml and for ricin is 25 ng (RCA 60)/ml (14, 15).
Several different LFAs have been evaluated with the
results published in the peer-reviewed literature. For example,
King et al. (16) evaluated three products for the sensitive and
specific detection of B. anthracis. These were the Anthrax
Bio-Threat Alert (BTA) test strips (Tetracore, Gaithersburg,
MD), BioWarfare Agent Detection Devices (BADD)
(Advent Biotechnologies, Phoenix, AZ), and SMART-II
(New Horizons Diagnostics). Spore concentrations ranged
from 102 to 106 spores/ml. Cross-reactivity with B. cereus and
B. thuringiensis was evaluated using 106 spores for each near
neighbor. Sample size at each concentration was small: only
two to three replicates per concentration, and only two repli-
cates for near neighbor testing. BADD and SMART-II could
reliably detect 105 spores/ml, whereas the Anthrax BTA
could only reliably detect 106 spores/ml. In terms of cross-
reactivity, the Anthrax BTA and BADD did not cross-react
with B. cereus and B. thuringiensis, but the SMART-II assay
cross reacted in one out of two tests with B. thuringiensis.
Some immunoassay platforms are combined with optical
readers or fiber optic sensor arrays, and this can improve
assay sensitivity by 1–2 logs (17–19). Other enhancements
include localization of the target with the affinity reagent to
achieve sensitivity gains. The Rapid Analyte Measurement
Platform (RAMP) (Response Biomedical, Burnaby, BC) sys-
tem uses LFA test strips combined with an optical reader.
There are several published evaluations of RAMP (17, 18).
For B. anthracis the LOD was 6,200 spores (10 μl sample
with a concentration of 620,000 CFU/ml), and there was
no cross-reactivity with closely related near neighbor spores
(B. thuringiensis and B. cereus tested at 100 million spores
per 10 μl sample), and there was no interference observed
with samples containing common household powders (18).
RAMP was also evaluated to detect botulinum neurotoxin
A (17) and was compared to other LFA vendors (BTA,
Tetracore, BADD, Advnt, and SMART-II, New Horizons
Diagnostics). The best sensitivity was achieved with the
RAMP at 50 ng/ml using purified toxin. In addition, the
RAMP assay only detected botulinum neurotoxin A and
did not cross-react with botulinum neurotoxin B. Another
immunoassay-based system is the RAPTOR (Research Inter-
national, Monroe, WA), which is a fluoroimmunoassay-based
fiber optic sensor system that can detect a number of threat
agents including B. anthracis, Y. pestis, Brucella abortus, Fran-
cisella tularensis, Staphylococcus enterotoxin type B (SEB), and
E. coli O157:H7 (19). We found only one peer-reviewed
publication (19) for the RAPTOR, and in that study the
LOD for B. anthracis was 100,000 spores/ml, but cross-
reactivity was not assessed.
PCR
Most field-based PCR systems consist of a disposable assay
cartridge containing all of the consumable reagents (includ-
ing the polymerase), an instrument that integrates the ther-
mal components required to perform the heating and
cooling cycles required for PCR, and the optical components
required to assess the amplified DNA products (12). PCR-
based assays are advantageous because they are generally
very sensitive and specific, but few field-based systems have
integrated sample preparation to remove PCR inhibitors.
Approaches to mitigate PCR inhibition are reviewed later
in this chapter, but in the context of a white powder, one
solution would be to dilute the sample in PCR-compatible
buffer. PCR assays, however, are expensive (instrumentation
ranges from $6K–$50K and assays cost $12–$200 per sample)
and require relatively long assay times (30–60 min). PCR
assays will not detect toxins, unless the toxin preparation
contains DNA from the source organism (e.g., castor bean
DNA in a preparation of ricin toxin).
For field-based detection, the two most thoroughly
reviewed instruments come from Biofire Diagnostics (for-
merly Idaho Technology, Salt Lake City, UT). The RAZOR
EX has undergone a formal AOAC evaluation process as out-
lined by the Stakeholder Panel on Agent Detection Assays
2.3.4-6 ▪ TARGET-SPECIFIC DETECTION
(SPADA) (20) and the evaluation focused on the analysis of
potential B. anthracis spores recovered from aerosol samples
(e.g., filter units). This evaluation was completed for a
previous version of the RAZOR EX B. anthracis assay. Other
published applications for the RAZOR include new assays
(Phymatotrichopsis omnivore, a fungal pathogen of crops; 21)
or detection of other pathogens like influenza A [22] and
Vibrio cholera [23]).
A second unit from Biofire Diagnostics, the FilmArray, has
been evaluated by multiple investigators (24–31). The Fil-
mArray is the only hand-held and portable (less than 5
pounds) PCR system currently on the market that includes
an automated sample preparation step to remove potential
PCR inhibitors prior to PCR as well as all the reagents needed
for sample preparation and PCR detection, although several
other systems with integrated sample preparation are under
development. Different “pouches” can be used for different
panels of targets, thereby allowing detection of multiple
pathogens. These include a respiratory panel (24, 26–30)
that detects major respiratory viral and bacterial pathogens,
a blood culture panel (25), and a biothreat panel (31). In
Seiner et al. (31), B. anthracis, Francisella tularensis, and
Y. pestis DNA could be detected reliably at 250 genome
equivalents (GEs) per sample of purified DNA. In addition,
DNA for these three agents was detected at an equivalent
of 25 GEs for 63 of 72 samples. No cross-reactivity with
near neighbor DNA was observed. The system also detected
25 CFU of B. anthracis Sterne spores.
Smith’s Detection BioSeeq Plus is another field-
deployable PCR assay that is based on linear after the expo-
nential amplification (LATE) PCR technology from the
Wangh laboratory at Brandeis University (32). LATE-PCR
is a specially designed asymmetric PCR protocol where
the primer system Tm (Tm Limiting primer–Tm Excess primer) is
>0 (32). When primers are designed in this manner, the
asymmetric PCR is more efficient than its symmetric counter-
part, and fluorescence continues to increase through the lin-
ear phase of the amplification. In theory this leads to greater
sensitivity or significant improvement in signal:noise ratio.
Rice et al. (33) developed a single-tube 17-plex LATE-PCR
sepsis panel that could potentially be adapted to future plat-
forms (33). Deeper multiplexing can be achieved in this sys-
tem by taking advantage of four different reporter channels.
Because each reporter channel can distinguish multiple prod-
ucts with different melting temperatures, the assay has the
potential to incorporate higher levels of multiplex testing.
Assay sensitivity was reported to be 10–1,000 copies of puri-
fied spiked DNA per reaction, depending on the gene target.
It should be noted that the field-based detection systems
we have reviewed here are not capable of processing multiple
samples simultaneously. Both the immunoassay and PCR
systems have the capability to provide multiple target (e.g.,
multiplex) pathogen detection, but these systems are not
capable of handling many samples at once (e.g., surge capa-
city). For hand-held assays, this may not be as critical because
sample-to-answer time is usually within a few minutes.
PCR-based systems, in contrast, typically require at least
45–60 min from sample input to answer. The user’s applica-
tion will dictate the suitability of any field-based system
they would use.
SAMPLE PREPARATION AND
ASSAY INHIBITORS
Because most applications in field-based detection include
field-deployable PCR, this section examines how inhibitors
limit the application of PCR assays for detection systems
and in the context of a variety of other environmental micro-
biology applications. Environmental sample testing typically
requires separating the sample material from unwanted debris,
concentrating material from large volumes, and mitigating
PCR inhibitors. In our experience, without sample purifica-
tion and concentration, the minimum analytic sensitivity
of a PCR assay will be ∼103 copies, which may exceed the
infectious dose of many pathogens (34). Whereas this may
not be a problem where pathogens are extremely concen-
trated (e.g., white powder), this would be critical in other
environmental applications where pathogens are expected
to occur in limited abundance. PCR inhibition can be partic-
ularly problematic when samples are concentrated because
this concurrently concentrates compounds that interfere
with PCR amplification.
PCR inhibition is best illustrated using an example from
real-time PCR where an uninhibited reaction might yield a
CT value of 21.5, but inhibition with the same sample might
produce a final CT value of 32 or completely block PCR
and lead to a false negative result (Fig. 1). In cases where
inhibition does not completely block the PCR reaction, a
multiplexed assay might be used where a positive control
template is added to the reaction to normalize the fluorescent
signal (35). Under this scenario the fluorescent intensity of
the sample can be divided by the fluorescent intensity of
the control product ( produced in the same reaction well)
to produce a ratio that is corrected for inhibition. This proce-
dure assumes that each primer set is equally affected by any
PCR inhibitors that are present in the reaction. If they are
not, PCR normalization is decoupled between the control
and sample reactions; consequently there is no way to accu-
rately estimate the starting copy number of target molecules.
Unfortunately, PCR inhibitors can differentially affect differ-
ent primer sets (36).
Mechanisms of Inhibition
PCR inhibition is primarily caused by (a) direct binding
or other physical interference with the Taq polymerase, (b)
indirect interference with polymerase activity via chelation
or competition with Mg2+ by other ions, and (c) direct bind-
ing or interference with the DNA template or primers. The
degree that these mechanisms interfere with PCR amplifica-
tion can vary with the sample matrix and sequence-specific
differences between different PCR products and primers. In
the case of fluorescent detection methods, inhibitors can
also squelch fluorescent signal by absorbing emitted photons
or by producing nonspecific background that masks the signal
from the PCR product (37–39).
Opel et al. (36) investigated the probable mechanisms
by which a variety of inhibitors impact PCR amplification.
The experimental design included PCR products of different
size (100, 200, and 300 bp), different annealing temperature
(53°C, 55°C, and 58°C), and sequence differences. They
evaluated defined inhibitors using real-time PCR and exam-
ined the effects on CT, fluorescent intensity, amplification
efficiency, and melt curves (Table 2). Notably, collagen,
humic acid, and melanin appeared to bind DNA template
directly, thereby extending the CT. This type of binding is
most likely sequence-specific, meaning that copy number
estimates will be confounded by differential impacts that are
sequence-specific, and conventional normalization strategies
will not resolve this problem. In most cases practitioners
do not know what inhibitors are coextracted with template
DNA or RNA. Rock et al. (40) employed fluorescence
excitation-emission matrix profiling to estimate what
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-7

FIGURE 1 Example of a real-time PCR reaction where PCR inhibitors either block product detection completely, or cause the CT estimate to
be extended and thereby underestimate the starting copy number for the template DNA. doi:10.1128/9781555818821.ch2.3.4.f1

inhibitors are present in a sample and predict the degree of
PCR inhibition that might arise from DNA extractions
from biosolids. This approach could be used to evaluate lim-
ited samples before they are used in a reaction and evaluate
various strategies that have been employed to mitigate PCR
inhibition.
An often overlooked aspect of PCR inhibition can be the
presence of too much template in the PCR. Whereas there is
very limited information on this phenomenon (41), most
manufacturers of PCR kits cite this as one problem resulting
in PCR failure. Whereas too much template would not be a
major concern for environmental samples, in limited cases
(e.g., concentrated white powders), this could be a cause for
concern. For real-time PCR, excess template is manifested
by an amplification curve that may or may not cross the
detection threshold or the shape of the curve becomes con-
cave as excess DNA interferes with fluorescent intensity
(Fig. 2).
Methods to Mitigate PCR Inhibition
Many potential solutions have been applied to mitigate PCR
inhibition and they have been employed either before or after

TABLE 2 Predicted mechanisms of inhibition and outcomes for seven inhibitors added to a standardized qPCR reaction
Sensitive to
Inhibitor Mechanism Extended Reduced Reduced Impacts amplicon
CT a fluorescence efficiency melt curve size and/or Tm b
Calcium Inhibits Taq activity X X
Collagen Inhibits Taq activity X X X X X
Binds DNA template
Hematin Inhibits Taq activity X X X X
Humic acid Binds DNA template X X X X
Indigo Absorbs fluorescent signal n.d.c n.d. n.d. n.d. n.d.
Blocks assay
Melanin Binds DNA template X X X X
Tannic acid Inhibits Taq activity X X

Adapted from Tables 3 and 4, Opel et al. (36); permission granted.

a
CT = cycle threshold for qPCR.
b
Tm = melt temperature.
c
n.d. = not determined.
2.3.4-8 ▪ TARGET-SPECIFIC DETECTION
FIGURE 2 Example of a real-time PCR reaction where too much template was added to the sample. Initially fluorescence increases as
expected, but both the initial quantity and amount of accumulating template squelches fluorescence. doi:10.1128/9781555818821.ch2.3.4.f2
DNA extraction. The outcome of these methods depends on
the sample type, concentration of inhibiting substances, and
abundance of the template DNA. The degree of inhibition
can also vary depending on the type of organism being
targeted (34). In this section we consider environmental
(soil, water, air), food and physiological samples.
Pre-Extraction Procedures
Pre-extraction procedures range from incorporating extra
washing steps to physically separating target organisms from
sample matrices or using methods to precipitate or otherwise
degrade and remove inhibiting compounds. For example,
Fortin et al. (42) reported that washing sediment samples
multiple times reduced the concentration of inhibitory
compounds before DNA extraction, although it is clear that
this strategy is not universally applicable (43). Jacobson
et al. (44), used density gradient centrifugation to separate
Lawsonia intracellularis from PCR inhibitors and fecal particles
prior to extracting PCR-ready DNA template. Other physical
separation methods including sorted flow cytometry (45)
and magnetic beads conjugated to antibodies for pathogen
capture (46). Persoh et al. (43) used Al2(SO4)3 to precipitate
humic compounds from soil samples prior to cell lysis
and DNA extraction. While Al2(SO4)3 precipitation can
be effective, it may be necessary to vary the concentration
of Al2(SO4)3 with each sample.
Concurrent and Postextraction Procedures
Inhibiting compounds can be removed concurrent with
nucleic acid extraction or postextraction. This could be as
simple as repeating the extraction procedure more than
once. While this can undoubtedly reduce the total yield of
nucleic acids, Kemp et al. (35) found that repeated extraction
was sufficient to remove PCR inhibitors for work with ancient
DNA from human skeletal remains. Akane (47) reported that
inhibiting heme compounds (from blood) could be degraded
after nucleic acid extraction by adding hydrogen peroxide as
a final step before ethanol precipitation of extracted DNA.
Techer et al. (48) combined either cetyl trimethylammonium
bromide or KCl with a vitamin mixture ( pyridoxal and
thiamine hydrochloride) during a bead beating procedure.
This produced PCR-ready DNA from a diversity of soil types,
including samples with a high humic acid concentration
(100 mg/g). This purification procedure is sensitive to the
pH of the lysis buffer.
Another strategy for separating nucleic acids concurrently
(or in tandem) with the extraction procedure is to employ
silica-based techniques (49–51). For these methods the
nucleic acid is adsorbed to a silica surface after which other
compounds are washed away, and then the nucleic acid is
eluted from the silica surface. This works because nucleic
acids adsorb to silica in the presence of a high ionic strength
buffer. According to Melzak et al. (52), the mechanisms for
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-9
adsorption are a combination of (a) weakened electrostatic
repulsion in the presence of a high salt concentration (both
nucleic acids and glass are negatively charged); (b) in the
presence of chaotrophic salts (e.g., guanidine-HCl or guani-
dine thioisocyanate) both the nucleic acid and silica surface
are dehydrated, thereby allowing the positively charged ions
from the nucleic acid to bind the negatively charged silica;
and (c) intramolecular hydrogen bonding between the
nucleic acid and silica surface. Addition of guanidine thio-
isocyanate also helps lyse cells and inactivates nucleases
(guanidine-HCl does the same but to a lesser extent; 49).
The basic procedure involves lysing cells, nucleic acid adsorp-
tion to silica surfaces, centrifugation and washing of silica par-
ticles, and elution of the nucleic acid with a low ionic strength
buffer, such as EDTA. Adsorption to silica is the basis for a
number of commercially available kits, and silica-based meth-
ods have been applied to a variety of sample types including
skeletal samples (53, 54), clinical samples (55), formalin-
fixed and paraffin-embedded tissue (56), and soils and sedi-
ments (57, 58). Because it is possible to use a variety of silica
configurations, silica-based methods are also conducive for
use in micro- and nanochip technologies (59, 60).
Probably the most common method for relieving the
effects of PCR inhibition is by physically separating extracted
nucleic acids from inhibiting compounds. Three primary
strategies have been employed.
1. Extracted materials can be passed through an agarose gel
by electrophoresis, thereby separating nucleic acids from
contaminating inhibitors (61). The purified nucleic acid
can be recovered directly from the gel using an agarase
digestion protocol (62) or by electroelution. Chandler
et al. (38) found that electroelution improved PCR
detection sensitivity from soil samples by 1,000- to
10,000-fold compared to the control extractions without
electroelution. Engel et al. (63) reported that both linear
and nonlinear electrophoresis yielded DNA of sufficient
quality for construction of metagenomic libraries.
2. Size exclusion chromatography has been widely used to
relieve PCR inhibition. Columns are made by adding
the separation matrix to a tube (column). Nucleic acid
extracts are applied to the top of the matrix and allowed
to pass through the matrix passively or with the aid of
centrifugation. Hydroxyapatite columns are an example
of a method that binds DNA allowing contaminates to
be washed through the column while the DNA is subse-
quently eluted using a different buffer condition (64).
Most other size exclusion methods rely on capturing
smaller molecules (e.g., humic acids) in the pores of a
gel or bead material such as Sepharose, Sephadex,
Sephacryl, polyvinylpolypyrrolidone, or polyacrylamide.
Because larger molecules cannot enter the pores, they
elute more rapidly. Columns have also been used sequen-
tially to produce a more purified product (65). Size exclu-
sion chromatography has been applied successfully
to reduce PCR inhibition for nucleic acids extracted
from a wide variety of samples including soils (66–68),
compost materials (61, 69), biogas fermenters (70), man-
ure (71), and forensic materials (72). While size exclu-
sion chromatography has been used very successfully,
these methods can result in reduced yield of nucleic
acids (70).
3. A variety of selective capture molecules (e.g., antibodies,
oligonucleotides) can be conjugated to magnetic
beads, allowing the practitioner to capture target mole-
cules and remove them from contaminating materials.
Oligonucleotide-conjugated magnetic beads are typi-
cally mixed with a nucleic acid extract, after which the
beads are held to the side of a microcentrifuge tube while
contaminants are washed away. Maher et al. (73) used
this approach with an oligonucleotide probe specific for
the 16S rRNA sequence of a fungus, Pneumocystis
carinii. When applied to crude DNA extracts, this mag-
netic bead capture system made it possible to detect fun-
gal DNA in 28 out of 30 otherwise PCR-inhibited
samples. Gopal et al. (74) used a similar approach, mak-
ing it possible to pool up to 200 black flies for PCR
screening to detect Onchocerca volvulus. This resulted
in a fourfold increase in the efficiency of sample pro-
cessing. Opsteegh et al. (75) employed a long (70-mer)
oligonucleotide for a defined repeat region found in Tox-
oplasma gondii. This procedure made it possible to detect
as few as 230 of the parasites in 100 g of meat product.
Other Strategies for Overcoming PCR Inhibition
One novel way to circumvent PCR inhibition is to produce
a more tolerant Taq polymerase. Kermekchieve et al. (39)
succeeded in producing a mutagenized polymerase that
improved PCR performance 10–100-fold in the presence
of humic acids and interference from coextracted compounds
from blood. Zhang et al. (76) combined an “inhibitor-
resistant” Taq mutant with a nonionic detergent, L-carnitine,
D-(+)-trehalose, and heparin, making it possible to amplify
DNA targets directly from human plasma, serum, or
whole blood without any nucleic acid extraction. Digital
PCR methods are also likely to gain increasing prominence
in the future as a means to mitigate PCR inhibition for
pathogen detection. Digital PCR (see chapter X) involves
separating the PCR reaction mixture into a large number of
separate reaction volumes as small as 1 nl. Given a large num-
ber of these reactions (700–20,000), inhibiting compounds
are essentially diluted and there is an increasing probability
that the PCR reaction will be successful in several reactions.
Hoshino and Inagaki (77) tested this idea with soil samples
and known concentrations of humic acids. In their study
quantitative PCR (qPCR) underestimated copy number
by 1/4,400 in the presence of 6.6 ng/liter humic acids while
digital PCR accurately predicted copy number with up to
9.3 ng/liter humic acids. Presumably, even if digital PCR can-
not be used to accurately predict copy number in these cases,
it should be more sensitive for determining the presence
or absence of a specific pathogen marker because of this
dilution effect.
The impact of inhibitors can also be mitigated at least
partially by varying the components of the PCR reaction
itself. For example, Opel et al. (36) found that if an inhibitor
directly affected the Taq polymerase or chelated MgCl2, add-
ing additional enzyme or magnesium could partially circum-
vent these mechanisms. Adding bovine serum albumen has
been a widely used strategy that can help limit inhibition
(58, 78–81), presumably by binding inhibitors that would
otherwise interfere directly with the Taq polymerase. Boom
et al. (55) combined alpha-casein with a silica extraction pro-
cedure to overcome PCR inhibitors found in cerebrospinal
fluid and urine samples. Provided that the target template is
sufficiently abundant, it is also possible to dilute the nucleic
acid extraction and thereby limit the effect of inhibitors
(34, 37). For example, Chen et al. (82) showed that dilution
by itself reduced iron and humic acid contaminates in water
sufficiently to relieve qPCR inhibition 10–100-fold for a
Legionella pneumophila assay. Finally, altering the PCR assay
2.3.4-10 ▪ TARGET-SPECIFIC DETECTION
design can also help circumvent inhibition problems by pro-
ducing shorter PCR products and using primers that function
at higher annealing temperatures (36).
STATISTICAL DESIGN FOR DETECTOR
EVALUATION
Performance requirements for a detection system depend on
the application. Most of the assays and detectors highlighted
here were developed with bioterrorism in mind, where a likely
scenario consists of a white powder containing an abundance
of pathogenic material. Unfortunately, this scenario might
not fit other applications where the pathogen may be present
in dilute quantities. When a detector developed for one
purpose is used for another, the device must be tested to deter-
mine how well it performs with respect to sensitivity, specif-
icity, and susceptibility to application-specific background
conditions. We recommend a graded approach to instrument
testing, one that takes the intended use of the device into
account. This can be a cost-effective and efficient way to val-
idate an existing technology that is applied to a new appli-
cation. Once completed, the results of such testing should
be peer-reviewed, published, and made widely available for
the benefit the instrument developers and potential users.
Identifying Standards for Field-Based
Detection Systems
Field-based biodetection systems were originally developed
to detect biological warfare agents. In recent years, the use
of these devices has spread to other applications, and there
have been concurrent efforts to develop testing standards to
define the performance characteristics and potential limita-
tions of these technologies (83). The U.S. Department of
Homeland Security funded the AOAC to establish standards
for PCR-based assays (84–86) and hand-held assays (14, 15).
Additionally, the AOAC formed the SPADA. Tasked with
establishing protocols to evaluate the performance of assays
and equipment, this panel is composed of a number of stake-
holders from 15 different participating organizations in the
government, academic, public health, and first responder
sectors (87). Among its other duties, SPADA is responsible
for setting the AMDL that assays should reliably achieve
and recommending panels of organisms that should be tested
to determine the false positive and false negative rates for a
given system. We use the AOAC guidelines as an example
of the statistical considerations and design for testing candi-
date assays and field-based biodetection systems. As already
stated, it is important to emphasize that if the application
for an assay or instrument changes, the framework for the
evaluation would remain the same although the AMDL
and/or tolerance for false positives may change, depending
on the new application.
Statistical Considerations
The AOAC/SPADA standards provide us with concrete,
testable performance metrics for biodetection systems. Based
on the AMDL for a system, AOAC/SPADA guidelines
require that the 95% lower confidence bound on the pro-
bability of detection be 0.95. This statistical requirement
provides the foundation on which evaluation studies can be
built. In particular, for a given study, the estimated probability
of detection (POD) is simply the fraction of samples that
the detection system correctly determines to be positive for
the target pathogen. Because the POD is calculated from
experimental data, there is some uncertainty attached to it.
This uncertainty can be quantified using a score confidence
interval (88). The score confidence interval depends on
the number of pathogenic samples tested, the number of false
negatives observed throughout the study, and the desired
confidence level. The 95% lower confidence bound on the
POD can be calculated as
PODLower
0 vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u" #, 1,
1:64 2 u 1:64 2
¼ @^p þ  1:64t ^pð1  ^pÞ þ NA
2N 4N
 
 1 þ 1:642 =N
where
#TP
^p ¼
N
is the experimental probability of detection calculated
by dividing the number of true positives (#TP) by the
total number of pathogenic samples tested (N ). The
AOAC-required lower confidence bound is 0.95, so a given
detection system meets the performance requirements if
PODLower ≥ 0.95. Fig. 3 relates PODLower to experimental
results. In particular, the 95% lower confidence bound is
plotted as a function of the total number of pathogenic sam-
ples tested for different numbers of observed true positives.
The AOAC-required lower limit of 0.95 is highlighted in
bold for reference.
Fig. 3 can be used to specify the total number of samples
that need to be tested and establish the criteria for passing
or failing a given detection system. For example, when fewer
than 47 samples are tested, the lower confidence bound
PODLower will be less than 0.95, regardless of how many
false negatives are observed. Therefore, to meet the AOAC/
SPADA requirements, a minimum of 47 samples must be
tested. If 47 samples are tested, then to meet the required
lower confidence bound of PODLower ≥ 0.95, there must be
no false negatives observed throughout the course of the
experiment. As the number of samples increases, so do the
number of acceptable false negatives. For example, when 79
or more samples are tested, 1 false negative is allowed.
When 107 or more samples are tested, 2 false negatives are
acceptable.
The AOAC/SPADA requirements for the probability of
detection refer specifically to the sensitivity of the assay,
that is, the ability to detect a pathogen target when it is
present in a sample. It is widely understood that the POD of
a detection system is dependent on concentration and, for
many systems, can drop sharply when a pathogen is present
in amounts approaching the detection limit of the device.
To keep the testing process relevant to the intended appli-
cation, it is therefore recommended that testing be performed
at the AMDL for the assay.
The statistical reasoning outlined here can also be used to
evaluate the specificity of a detection system. “Specificity”
refers to an assay’s ability to correctly identify a negative
sample. Specificity is often characterized by the number of
false positives, which in turn may be caused by experimental
error, background material, or nontarget near neighbors of
pathogen in a sample. The first step in designing a specificity
study is to list biological organisms or materials that are likely
to cause such erroneous results. These materials should be
incorporated into the testing process.
Fig. 3 can also be used to determine the number of samples
and evaluation criteria for specificity testing. In particular,
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-11
FIGURE 3 Number of independent tests required to meet test performance criteria. ** A one-sided score confidence interval.
doi:10.1128/9781555818821.ch2.3.4.f3
the true negative probability (one minus the false positive
probability) is the probability a detection system will correctly
determine a sample to be negative for a specific target.
Experimentally, this value is calculated as the number of non-
target samples that an assay determines to be nontarget (the
true negatives) divided by the total number of nonpositive
samples tested. To achieve a 95% lower confidence bound
of 0.95 on the true negative probability (which is the same
as a 95% upper confidence bound of 0.05 on the false positive
probability), at least 47 samples should be tested. Out of 47
samples, no false positives are acceptable. Out of 79 samples,
1 false positive is acceptable. Out of 107 samples, 2 false
positives are acceptable.
Example of Current and Proposed Evaluation of
PCR-Based Detection Systems
As of this writing, the RAZOR EX in combination with
manual filter extraction is the only PCR-based assay that
has been subjected to a formal AOAC testing process outlined
by SPADA (20). Specifically, the unit was evaluated for its
ability to detect B. anthracis after extraction from a collection
filter. This process serves as an initial proof-of-concept study;
it is not meant to be a comprehensive evaluation protocol
for all detection systems under all circumstances. Deviations
from this protocol can and should occur when appropriate.
In the RAZOR EX example, the AOAC/SPADA testing
protocol outlines two procedures, one for inclusivity strains
(true positives, or independent strains that should be positive
by the assay) and exclusivity strains (true negatives, or
independent strains that are near neighbors but should test
negative by the assay) strains. The protocol can be summar-
ized as follows:
1. 15 inclusivity strains are tested once each. Inclusivity
strains are defined as those pathogenic strains that should
be detected by the assay.
2. 20 exclusivity strains are tested once each. Exclusivity
strains are defined as near neighbors of the pathogen
target that should not be detected in the assay.
3. If up to 5 failed results are observed then the strains
that produced the failed result are tested an additional
96 times.
4. In none of the additional tests produce a failed result,
then the device passes; otherwise it fails.
Under the best of circumstances, where no false positives
and no false negatives are observed in steps 1 and 2, the device
will “pass,” but as outlined this testing design does not meet
the requisite statistical criteria of 95% confidence/0.95 lower
bound. This requires testing at least 47 samples. Furthermore,
if there are several assay failures, the number of samples
needed to meet the 0.95 POD criteria increases significantly,
and so does the cost of the evaluation. This unpredictability is
one of the biggest weaknesses of the protocol. The number
of inclusivity samples could be anywhere between 15 and
2.3.4-12 ▪ TARGET-SPECIFIC DETECTION
495 depending on the results of steps 1 and 2. Similarly, the
number of exclusivity strains tested could be anywhere
between 20 samples and 500 samples. Such a protocol can
wreak havoc on experimental budgets and end up costing a
lot of money.
A more practical approach to evaluation testing addresses
the requirements directly while taking into account the
intended use of the instrument. For example, consider the
case where there are 15 inclusivity strains. For inclusivity test-
ing, three replicates from each strain are evaluated. Two
strains are then selected at random, and a third replicate of
each is tested. This gives a total number of pathogenic (inclu-
sivity) strains of 47. If none of those samples produces a false
negative, then the testing terminates with the determination
that the detection system meets the AOAC/SPADA require-
ments. If two or more samples produce a false negative result,
then testing terminates with the determination that the sys-
tems fails to meet the performance requirements. If only
one sample produces a false negative, then the system is given
a second chance to meet the requirements. In this case, 32
more replicates are selected for testing, resulting in a total of
79 inclusivity samples. If none of these additional samples
produce a false negative, then the system passes the evalua-
tion. Otherwise, it fails.
An analogous testing protocol can be constructed for
exclusivity testing. For 20 exclusivity strains, two replicates
of each are prepared for testing. Seven of the strains are
then selected at random, and a third replicate of each is tested.
Out of these 47 samples, if none produce a false positive, then
the device is determined to meet the AOAC/SPADA criteria
for POD≥0.95. If two or more samples produce a false posi-
tive, it fails. If one produces a false negative, then 32 more rep-
licates are selected for testing, and if none of these produce a
false positive, the system is determined to meet the perform-
ance requirements.
Unlike the original AOAC/SPADA procedure, this test-
ing protocol ensures that if a detection system meets the eval-
uation criteria, it also meets the AOAC/SPADA criteria of a
95% lower confidence bound of 0.95 for the POD (or true
negative probability). Perhaps more important, it provides a
more precise indication of the number of samples needed
for testing. In particular, for both inclusivity and exclusivity
testing, the determination can be made with a minimum of
47 or 79 samples. With the original AOAC/SPADA proto-
col, the sample numbers could be anywhere between 15
and 500. In other words, this new experimental design pro-
vides us with an efficient protocol that minimizes the number
of samples and the cost while producing results that directly
address the requirements at hand.
WHERE TO SAMPLE?
Assay development, validation, and application are key issues
in environmental microbiology, but the conclusions we draw
from these assays are only as useful as the sampling design per-
mits. Every student and practitioner working at the field level
is faced with questions about how to design their sampling
frame to address specific goals and achieve a desired level of
confidence with unbiased estimators. Even with a suitable
sampling design, it is also useful to have tools that permit
rapid identification of spatially explicit sampling coordinates.
From a logistics and cost perspective, it is also useful to con-
sider if it is advisable to pool samples for a single analysis.
In this section we briefly describe the Visual Sample Plan
(VSP) software (89), which provides a powerful toolbox to
address some of these concerns. A number of other extensive
tools and references are also available (90–93), including
modules that can be executed using the R language (94, 95)
As an example of a VSP application, assume that a student
needs to estimate a mean and confidence interval for the
log CFU of E. coli for two plots of land that have received
an application of biosolids. Samples will be collected and
brought to the lab for plating on selective media. Sources of
variance include intra- and interassay variation, sample
collection variation, and natural variation due to a heteroge-
neous distribution of bacteria. Research funds are tight and
the student needs to optimize the sampling design to mini-
mize total expenditures. Consequently, she elects to use the
confidence interval estimation module with incremental
sampling.
Incremental sampling is equivalent to composite sampling
whereby a discrete number of samples (increment samples) is
mixed before analysis of the composite mixture. The number
of increment samples and number of composite samples can
be optimized to minimize cost. In this example the student
selected a 95% confidence interval spanning 1 log CFU
above and below the estimated mean. A pilot study demon-
strated that the standard deviation for increment to incre-
ment sampling was 1 log CFU and the standard deviation
for the analytic method was 0.5 log CFU. Assuming that
the cost of sample collection was $0.50/increment and $5
for the composite sample, and the goal was to limit cost
then the recommended sample design combined three incre-
ment samples (nine total) each into three composite sample
for analysis (Fig. 4). The VSP software also provides spatially
explicit coordinates for sample collection.
Sampling design needs clearly depend on the goals of the
study. The VSP software is oriented toward environmental
cleanup problems, but provided that underlying assumptions
can be met, the tools available in this package can be adapted
for a variety of applications in environmental microbiology
(Table 3). VSP also incorporates sophisticated mapping tools
that allow the user to create maps of rooms and buildings for
contaminant sampling. The user can also import a variety of
maps, quickly identify sampling polygons, and develop a list
of sampling locations (Fig. 5).
CURRENT NEEDS AND EMERGING
TECHNOLOGIES
This chapter examined several important topics regarding
the current capabilities of field-based biodetection systems,
challenges for using these instruments for environmental
samples, rational approaches for evaluating instrumentation,
and finally how to develop a site sampling plan to address bio-
detection needs. At least for the near future, field-portable
PCR based biodetection systems are likely to be the instru-
ment of choice based on (1) sensitivity and specificity of
the assays, (2) the dynamic range of concentrations allows
PCR to be used when samples are very concentrated (e.g.,
“white powder”) or very dilute ( pathogens in water), and
(3) assays can be redesigned or even new assays added to
replace cross-reactive assays or if there is a need to detect
emerging pathogens from the environment. The packet and
lyophilized reagent systems allow the assays to be stored at
room temperature. This may be an important consideration
in austere environments where electricity is sporadic or not
available and cold storage may not be possible.
Current needs for field-based systems fall into two broad
categories: (1) the need to incorporate sample processing into
the biodetection system, and (2) peer-reviewed evaluation
and publication of testing results of emerging technologies
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-13

FIGURE 4 Symbols show distribution of “multiple increment” samples where sample location is randomly allocated (Visual Sample Planner
software; 89) and samples for each symbol within a plot are pooled for analysis. Inset table is an example of the coordinate output, where coor-
dinates can be arranged randomly or systematically for the confidence interval estimation module. doi:10.1128/9781555818821.ch2.3.4.f4

TABLE 3 Sampling goals and sample selection strategies with to determine suitability of the instrument for the specified
some options that are available in Visual Sample Planner (89) uses for the instrumentation. For pathogenic organisms that
are widely dispersed in the environment, there is also a
Sampling goals: need to incorporate a sample concentration step with the
Estimate mean and confidence interval caveat that coinhibitors may also be concentrated.
An additional area needing further attention to develop-
Compare mean to background mean or fixed threshold
ment of effective and statistically robust test plans that
Estimate a proportion or compare to a fixed threshold or reference provide confidence in the results. Currently the AOAC/
proportion SPADA approach is a good attempt at addressing this and
Locate hot spots several issues that surround the intended use of the instru-
Trend detection ment or assay. Specifically, the most effective testing plans
Detect sample redundancy are based on the requirements of the application rather
than the abilities of the system being tested. For example,
Item sampling the AOAC-led SPADA committee (83) determined that
Sampling within rooms and buildings hand-held immunological-based assays should be able to
Estimate boundaries of contamination detect 107 B. anthracis spores/ml in a white powder (14).
Radiological transect surveys This was based on a graded approach that took into account
the scenario that first responders would likely encounter
Sample selection strategies and options: where the standard was loosened to the current standard of
Judgmental sampling 107 spores/ml to accommodate the fact that most immuno-
assays could not achieve the AOAC’s original standard of
Incorporation of historical samples 106 spores/ml (83) (note that 107 spores is approximately
Random, stratified, or systematic sampling 0.1 mg and is barely visible as a white powder). Based on
Random within systematic sampling the current AMDL recommendation, it is reasonable to
Multiple increment sampling (composite sampling) require other technologies, such as PCR-based assays, to do
Collaborative sampling (combine data from two analytic methods) the same. The PCR standard for white powder monitoring
(86) is 2,000/ml. Notably, even though this standard is
Sequential sampling much lower than required for immunoassays, it is probably
Transect (specific for unexploded ordinance) lower than is needed for typical emergency responder needs
Adaptive cluster sampling (for hotspot detection) with respect to white powder applications. In other words,
Room and building sampling 2,000/ml may be the standard for this type of detection sys-
tem, but it should not be the standard for this particular
2.3.4-14 ▪ TARGET-SPECIFIC DETECTION
FIGURE 5 Left panel illustrates an example of how a room can be mapped and shown in three dimensions with sample locations (white
circles; n = 10). Right panel shows how sample area polygons can be overlaid on Google Earth imagery and sample sites subsequently mapped
to location (n = 50 random points within systematic grids in this case). Images such as that shown here can be easily calibrated for scale.
Produced using VSP software (89). doi:10.1128/9781555818821.ch2.3.4.f5
application. Keeping this in mind, a testing plan can be
devised that directly addresses the requirements and mini-
mizes the number of samples needed for validation of the
technology. If the application should change to address other
requirements for environmental sampling (e.g., pathogens
diluted in the environment), then both the instrumentation
and test plan can be evaluated in advance to determine if
the detection system used can even meet a significantly lower
AMDL.
A potential need for field-based detection that has largely
been overlooked is the ability to evaluate many samples
simultaneously (i.e., surge capacity). Most of the units on
the market today perform multiplexed detection of different
pathogen targets, but each unique sample must be evaluated
one at a time. Thus their use could be considered point-of-
care diagnostics. Some environmental applications may
benefit from being able to process multiple samples simulta-
neously. For example, consider a food processing plant where
Listeria monocytogenes has been introduced from a contami-
nated food product at some point in the production process.
The ability to sample multiple sites in the production train
and perform a single assay run would allow management
to more quickly identify where in the process the contamina-
tion was introduced and then implement disinfection plans
to address the issue. Multiple sample capability may also be
useful to screen animal herds to identify exposed or sick ani-
mals that would allow more rapid and effective control and
treatment.
With respect to emerging technologies there is typically
very little published literature available for assessment,
but what is available indicates that the industry is working
to improve PCR sensitivity, specificity, and especially speed
for detection. Of the systems on the market today, speed is
the second major impediment to the PCR. Assays typically
take 45 min to 1 h to complete, and this may be too long
for emergency situations (hazmat teams). Pipper et al. (96)
reported on a system that uses magnets in a clockwork type
system to provide sample separation, washing, and then
PCR. They demonstrated detection of 30 cells in approxi-
mately 17 min. Core Life Sciences (Laguna Nigel, CA) has
advertised a PCR machine that can complete the reaction
in as little as 8 min. The key to the technology is optimization
of the heating and cooling process (up to 15°C per second)
and optimization of heat transfer using super-thin-walled
tubes. While this appears to be a laboratory-based instrument,
optimization of components could make it possible for field
deployment.
In conclusion, field-based assays and equipment to detect
pathogens in environmental samples is an emerging field.
Depending on the application, the relatively rapid nonspe-
cific tests and immunoassays may provide a very quick answer
to an emergency situation. Nevertheless, if the pathogens
are very diluted in the environment, then more sensitive tests
are needed, which may require significantly more time both in
terms of sample processing and the actual assay to detect the
pathogen.
A portion of this effort was funded by the Department of Home-
land Security Science and Technology Directorate under Contract
HSHQDC-08-X-00843. Pacific Northwest National Laboratory is
operated by Battelle for the U.S. Department of Energy under contract
DE-AC06-76RLO.
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-15
REFERENCES
1. Parameswaran L, Bortolin L, Cabrera C, Rudzinski C.
2007. Recovery of organisms and nucleic acids from complex
samples. Lincoln Lab J 17:167–186.
2. Buttner MP, Cruz P, Stevens VL, Emanuel PA. 2004.
Evaluation of the Biological Sampling Kit (BiSKit) for
large-area surface sampling. Appl Environ Microbiol 70:
7040–7045.
3. Hodges LR, Rose LJ, Peterson A, Nobel-Wang J, Arduino
MJ. 2006. Evaluation of a macrofoam swab protocol for
the recovery of Bacillus anthracis spores from a steel surface.
Appl Environ Microbiol 72:4429–4430.
4. Rose LJ, Hodges L, O’Connell HO, Nobel-Wang J. 2011.
National validation study of a cellulose sponge wipe-
processing method for use after sampling Bacillus anthracis
spores from surfaces. Appl Environ Microbiol 77:8355–8359.
5. Valentine NB, Butcher MG, Su YF, Jarman KH, Matzke
M, Webb-Robertson BJ, Panisko EA, Seiders BAB,
Wahl KL. 2008. Evaluation of sampling tools for environ-
mental sampling of bacterial endospores from porous and
nonporous surfaces. J Appl Microbiol 105:1107–1113.
6. Edmonds JM, Collett PJ, Valdes ER, Skowronski EW,
Pellar GJ, Emanuel PA. 2009. Surface sampling of
spores in dry-deposition aerosols. Appl Environ Microbiol
75:39–44.
7. Rose L, Jensen B, Peterson A, Banerjee SN, Srduino MJ.
2004. Swab materials and Bacillus anthracis spore recovery
from nonporous surfaces. Emerg Infect Dis 10:1023–1029.
8. Budowle B, Schutzer SE, Ascher MS, Atlas RM, Burans
JP, Chakraborty R, Dunn JJ, Fraser CM, Franz DR,
Leighton TJ, Morse SA, Murch RS, Ravel J, Rock DL, Sle-
zak TR, Velsko SP, Walsh AC, Walters RA. 2005. Toward
a system of microbial forensics: from sample collection
to interpretation of evidence. Appl Environ Microbiol 71:
2209–2213.
9. Budowle B, Schutzer SE, Burans JP, Beecher DJ, Cebula
TA, Chakraborty R, Cobb WT, Fletcher J, Hale ML,
Harris RB, Heitkamp MA, Keller FP, Kuske C, Leclerc
JE, Marrone BL, McKenna TS, Morse SA, Rodriguez
LL, Valentine NB, Yadev J. 2006. Quality sample collec-
tion, handling, and preservation for an effective microbial
forensics program. Appl Environ Microbiol 72:6431–6438.
10. Budowle B, Schutzer SE, Einseln A, Kelley LC, Walsh
AC, Smith JA, Marrone BL, Robertson J, Campos J.
2003. Public health. Building microbial forensics as a
response to bioterrorism. Science 301:1852–1853.
11. Poore C, Clark P, Emanuel PA. 2009. An evaluation of sus-
picious powder screening tools for first responders. J Haz
Mater 172:559–565.
12. Baird C, Seiner D, Ozanich R, Bartholomew R, Colburn
H, Straub T, Bruckner-Lea CJ. 2012. Biodetection technolo-
gies for first responders. Pacific Northwest National Labora-
tory, Richland, WA.
13. Arnó J, Cardillo L, Judge K, Frayer M, Frunzi M, Hether-
ington P, Levy D, Oberndorfer K, Perec W, Sauer T, Stein
J, Zuidema E. 2012. Advances in handheld FT-IR instru-
mentation. Proc SPIE 8374:837405.
14. AOAC. 2011. AOAC SMPR 2010.004 Standard method
performance requirements for immunological-based hand
held assays (HHAs) for detection of Bacillus anthracis spores
in visible powders. J AOAC Int 94:1352–1355.
15. AOAC. 2011. AOAC SMPR 2010.005 Standard method
performance requirements for immunological-based hand
held assays (HHAs) for detection of Ricin in visible powders.
J AOAC Int 94:1356–1358.
16. King D, Luna V, Cannons A, Cattani J, Amuso P. 2003.
Performance assessment of three commercial assays for
direct detection of Bacillus anthracis spores. J Clin Microbiol
41:3454–3455.
17. Gessler F, Pagel-Wieder S, Avondet M-A, Böhnel H. 2007.
Evaluation of lateral flow assays for the detection
of botulinum neurotoxin type A and their application in
laboratory diagnosis of botulism. Diag Microbiol Infect Dis
57:243–249.
18. Hoile R, Yuen M, James G, Gilbert GL. 2007. Evaluation
of the rapid analyte measurement platform (RAMP) for
the detection of Bacillus anthracis at a crime scene. Forensic
Sci Int 171:1–4.
19. Jung CC, Saaski EW, McCrae DA, Lingerfelt BM,
Anderson GP. 2003. RAPTOR: a fluoroimmunoassay-
based fiber optic sensor for detection of biological threats.
IEEE Sensors J 3:352–360.
20. Spaulding UK, Christensen CJ, Crisp RJ, Vaughn MB,
Trauscht RC, Gardner JR, Thatcher SA, Clemens KM,
Teng DHF, Bird A, Ota IM, Hadfield T, Ryan V, Brunelle
SI. 2012. RAZOR® EX Anthrax air detection system.
J AOAC Int 95:860–891.
21. Arif M, Fletcher J, Marek SM, Melcher U, Ochoa-Corona
FM. 2013. Development of a rapid, sensitive, and field-
deployable Razor Ex biodetection system and quantitative
PCR assay for detection of Phymatotrichopsis omnivora
using multiple gene targets. Appl Environ Microbiol 79:
2312–2320.
22. Mölsä M, Koskela KA, Rönkkö E, Ikonen N, Ziegler T,
Nikkari S. 2012. Detection of influenza A viruses with a
portable real-time PCR instrument. J Virol Meth 181:
188–191.
23. Koskela KA, Matero P, Blatny JM, Fykse EM, Olsen JS,
Nuotio LO, Nikkari S. 2009. A multiplatform real-time
polymerase chain reaction detection assay for Vibrio cholerae.
Diag Microbiol Infect Dis 65:339–344.
24. Babady NE, Mead P, Stiles J, Brennan C, Li H, Shuptar S,
Stratton CW, Tang Y-W, Kamboj M. 2012. Comparison of
the Luminex xTAG RVP Fast assay and the Idaho Technol-
ogy FilmArray RP assay for detection of respiratory viruses in
pediatric patients. J Clin Microbiol 50:2282–2288.
25. Blaschke AJ, Heyrend C, Byington CL, Fisher MA,
Barker E, Garrone NF, Thatcher SA, Pavia AT, Barney
T, Alger GD, Daly JA, Ririe KM, Ota I, Poritz MA.
2012. Rapid identification of pathogens from positive
blood cultures by multiplex polymerase chain reaction
using the FilmArray system. Diag Microbiol Infect Dis 74:
349–355.
26. Loeffelholz MJ, Pong DL, Pyles RB, Xiong Y, Miller AL,
Bufton KK, Chonmaitree T. 2011. Comparison of the
FilmArray Respiratory Panel and Prodesse real-time PCR
assays for detection of respiratory pathogens. J Clin Microbiol
49:4083–4088.
27. Pierce VM, Elkan M, Leet M, McGowan KL, Hodinka RL.
2012. Comparison of the Idaho Technology FilmArray sys-
tem to real-time PCR for detection of respiratory pathogens
in children. J Clin Microbiol 50:364–371.
28. Poritz MA, Blaschke AJ, Byington CL, Meyers L, Nilsson
K, Jones DE, Thatcher SA, Robbins T, Lingenfelter B,
Amiott E, Herbener A, Daly J, Dobrowolski SF, Teng
DH, Ririe KM. 2011. FilmArray, an automated nested
multiplex PCR system for multi-pathogen detection: devel-
opment and application to respiratory tract infection. PLoS
One 6:e26047.
29. Rand KH, Rampersaud H, Houck HJ. 2011. Comparison
of two multiplex methods for detection of respiratory viruses:
FilmArray RP and xTAG RVP. J Clin Microbiol 49:
2449–2453.
30. Renaud C, Crowley J, Jerome KR, Kuypers J. 2012. Com-
parison of FilmArray Respiratory Panel and laboratory-
developed real-time reverse transcription-polymerase chain
reaction assays for respiratory virus detection. Diag Microbiol
Infect Dis 74:379–383.
2.3.4-16 ▪ TARGET-SPECIFIC DETECTION
31. Seiner DR, Colburn HA, Baird C, Bartholomew RA,
Straub T, Victry K, Hutchison JR, Valentine N,
Bruckner-Lea CJ. 2013. Evaluation of the FilmArray
((R)) system for detection of Bacillus anthracis, Francisella
tularensis and Yersinia pestis. J Appl Microbiol 114:
992–1000.
32. Sanchez JA, Pierce KE, Rice JE, Wangh LJ. 2004.
Linear-after-the-exponential (LATE)-PCR: an advanced
method of asymmetric PCR and its uses in quantitative
real-time analysis. Proc Natl Acad Sci USA 101:1933–1938.
33. Rice LM, Reis AH Jr., Ronish B, Carver-Brown RK,
Czajka JW, Gentile N, Kost G, Wangh LJ. 2013. Design
of a single-tube, endpoint, linear-after-the-exponential-
PCR assay for 17 pathogens associated with sepsis. J Appl
Microbiol 114:457–469.
34. Loge FJ, Thompson DE, Call DR. 2002. PCR detection of
specific pathogens in water: a risk-based analysis. Environ Sci
Technol 36:2754–2759.
35. Kemp B, Monroe C, Smith D. 2006. Repeat silica extrac-
tion: a simple technique for the removal of PCR inhibitors
from DNA extracts. J Archaeol Sci 33:1680–1689.
36. Opel KL, Chung D, McCord BR. 2010. A study of PCR
inhibition mechanisms using real time PCR. J Forensic Sci
55:25–33.
37. Bachoon DS, Otero E, Hodson RE. 2001. Effects of humic
substances on fluorometric DNA quantification and DNA
hybridization. J Microbiol Meth 47:73–82.
38. Chandler D, Schreckhise R, Smith J, Bolton HJ. 1997.
Electroelution to remove humic compounds from soil
DNA and RNA extracts. J Microbiol Meth 28:11–19.
39. Kermekchiev MB, Kirilova LI, Vail EE, Barnes WM.
2009. Mutants of Taq DNA polymerase resistant to PCR
inhibitors allow DNA amplification from whole blood and
crude soil samples. Nucleic Acids Res 37:e40.
40. Rock C, Alum A, Abbaszadegan M. 2010. PCR inhibitor
levels in concentrates of biosolid samples predicted by a
new method based on excitation-emission matrix spectro-
scopy. Appl Environ Microbiol 76:8102–8109.
41. Ahmed I, Islam M, Mannan A, Naeem R, Mirza B. 2009.
Optimization of conditions for assessment of genetic diver-
sity in barley (Hordeum vulgare L.) using microsatellite
markers. Barley Gen Newsl 39:5–12.
42. Fortin N, Beaumier D, Lee K, Greer CW. 2004. Soil wash-
ing improves the recovery of total community DNA from
polluted and high organic content sediments. J Microbiol
Meth 56:181–191.
43. Persoh D, Theuerl S, Buscot F, Rambold G. 2008. Towards
a universally adaptable method for quantitative extraction of
high-purity nucleic acids from soil. J Microbiol Meth 75:
19–24.
44. Jacobson M, Norling B, Gunnarson A, Aspan A. 2009.
Flotation—a new method to circumvent PCR inhibitors
in the diagnosis of Lawsonia intracellularis. Int J Microbiol
2009:410945.
45. Guillebault D, Laghdass M, Catala P, Obernosterer I, Leb-
aron P. 2010. Improved method for bacterial cell capture
after flow cytometry cell sorting. Appl Environ Microbiol
76:7352–7355.
46. Waller DF, Ogata SA. 2000. Quantitative immunocapture
PCR assay for detection of Campylobacter jejuni in foods.
Appl Environ Microbiol 66:4115–4118.
47. Akane A. 1996. Hydrogen peroxide decomposes the heme
compound in forensic specimens and improves the efficiency
of PCR. Biotechniques 21:392–394.
48. Techer D, Martinez-Chois C, D’Innocenzo M, Laval-Gilly
P, Bennasroune A, Foucaud L, Falla J. 2010. Novel
perspectives to purify genomic DNA from high humic
acid content and contaminated soils. Sep Purif Technol 75:
81–86.
49. Boom R, Sol CJ, Salimans MM, Jansen CL, Wertheim-van
Dillen PM, van der Noordaa J. 1990. Rapid and simple
method for purification of nucleic acids. J Clin Microbiol
28:495–503.
50. Chen CW, Thomas CA Jr. 1980. Recovery of DNA seg-
ments from agarose gels. Anal Biochem 101:339–341.
51. Marko MA, Chipperfield R, Birnboim HC. 1982. A proce-
dure for the large-scale isolation of highly purified plasmid
DNA using alkaline extraction and binding to glass powder.
Anal Biochem 121:382–387.
52. Melzak KA, Sherwood CS, Turner RFB, Haynes CA.
1996. Driving forces for DNA adsorption to silica in perchlo-
rate solutions. J Colloid Interf Sci 181:635–644.
53. Hoss M, Paabo S. 1993. DNA extraction from Pleistocene
bones by a silica-based purification method. Nucleic Acids
Res 21:3913–3914.
54. Lee HY, Kim NY, Park MJ, Sim JE, Yang WI, Shin KJ.
2010. DNA typing for the identification of old skeletal
remains from Korean War victims. J Forensic Sci 55:
1422–1429.
55. Boom R, Sol C, Beld M, Weel J, Goudsmit J,
Wertheim-van Dillen P. 1999. Improved silica-guani-
diniumthiocyanate DNA isolation procedure based on selec-
tive binding of bovine alpha-casein to silica particles. J Clin
Microbiol 37:615–619.
56. Farrugia A, Keyser C, Ludes B. 2010. Efficiency evaluation
of a DNA extraction and purification protocol on archival
formalin-fixed and paraffin-embedded tissue. Forensic Sci
Int 194:e25–e28.
57. Miller DN, Bryant JE, Madsen EL, Ghiorse WC. 1999.
Evaluation and optimization of DNA extraction and purifi-
cation procedures for soil and sediment samples. Appl Envi-
ron Microbiol 65:4715–4724.
58. Rojas-Herrera R, Narvaez-Zapata J, Zamudio-Maya M,
Mena-Martinez ME. 2008. A simple silica-based method
for metagenomic DNA extraction from soil and sediments.
Mol Biotechnol 40:13–17.
59. Cady NC, Stelick S, Batt CA. 2003. Nucleic acid purifica-
tion using microfabricated silicon structures. Biosens Bioelec-
tron 19:59–66.
60. Wolfe KA, Breadmore MC, Ferrance JP, Power ME, Con-
roy JF, Norris PM, Landers JP. 2002. Toward a microchip-
based solid-phase extraction method for isolation of nucleic
acids. Electrophoresis 23:727–733.
61. LaMontagne MG, Michel FC Jr., Holden PA, Reddy CA.
2002. Evaluation of extraction and purification methods for
obtaining PCR-amplifiable DNA from compost for micro-
bial community analysis. J Microbiol Methods 49:255–264.
62. Zhou J, Bruns MA, Tiedje JM. 1996. DNA recovery from
soils of diverse composition. Appl Environ Microbiol 62:
316–322.
63. Engel K, Pinnell L, Cheng J, Charles TC, Neufeld JD.
2012. Nonlinear electrophoresis for purification of soil
DNA for metagenomics. J Microbiol Meth 88:35–40.
64. Purdy KJ, Embley TM, Takii S, Nedwell DB. 1996.
Rapid Extraction of DNA and rRNA from sediments by a
novel hydroxyapatite spin-column method. Appl Environ
Microbiol 62:3905–3907.
65. Petric I, Philippot L, Abbate C, Bispo A, Chesnot T,
Hallin S, Laval K, Lebeau T, Lemanceau P, Leyval C,
Lindstrom K, Pandard P, Romero E, Sarr A, Schloter M,
Simonet P, Smalla K, Wilke BM, Martin-Laurent F.
2011. Inter-laboratory evaluation of the ISO standard
11063 Soil quality—method to directly extract DNA from
soil samples. J Microbiol Meth 84:454–460.
66. Lakay FM, Botha A, Prior BA. 2007. Comparative analysis
of environmental DNA extraction and purification methods
from different humic acid-rich soils. J Appl Microbiol 102:
265–273.
 2.3.4. Field Application of Pathogen Detection Technologies ▪ 2.3.4-17
67. Matheson C, Gurney C, Esau N, Lehto R. 2010. Assessing
PCR inhibition from humic substances. Open Enzyme Inhib J
3:38–45.
68. Miller DN. 2001. Evaluation of gel filtration resins for the
removal of PCR-inhibitory substances from soils and sedi-
ments. J Microbiol Meth 44:49–58.
69. Reuter T, Xu W, Alexander TW, Stanford K, Xu Y,
McAllister TA. 2009. Purification of polymerase chain reac-
tion (PCR)-amplifiable DNA from compost piles containing
bovine mortalities. Bioresour Technol 100:3343–3349.
70. Bergmann I, Mundt K, Sontag M, Baumstark I, Nettmann
E, Klocke M. 2010. Influence of DNA isolation on
Q-PCR-based quantification of methanogenic Archaea in
biogas fermenters. Syst Appl Microbiol 33:78–84.
71. Trochimchuk T, Fotheringham J, Topp E, Schraft H,
Leung KT. 2003. A comparison of DNA extraction and
purification methods to detect Escherichia coli O157:H7 in
cattle manure. J Microbiol Meth 54:165–175.
72. Shutler GG, Gagnon P, Verret G, Kalyn H, Korkosh S,
Johnston E, Halverson J. 1999. Removal of a PCR
inhibitor and resolution of DNA STR types in mixed
human-canine stains from a five year old case. J Forensic Sci
44:623–626.
73. Maher N, Dillon HK, Vermund SH, Unnasch TR. 2001.
Magnetic bead capture eliminates PCR inhibitors in samples
collected from the airborne environment, permitting detec-
tion of Pneumocystis carinii DNA. Appl Environ Microbiol 67:
449–452.
74. Gopal H, Hassan HK, Rodriguez-Perez MA, Toe LD, Lus-
tigman S, Unnasch TR. 2012. Oligonucleotide based mag-
netic bead capture of Onchocerca volvulus DNA for PCR pool
screening of vector black flies. PLoS Negl Trop Dis 6:e1712.
75. Opsteegh M, Langelaar M, Sprong H, denHartog L, De
Craeye S, Bokken G, Ajzenberg D, Kijlstra A, van der
Giessen J. 2010. Direct detection and genotyping of Toxo-
plasma gondii in meat samples using magnetic capture and
PCR. Int J Food Microbiol 139:193–201.
76. Zhang Z, Kermekchiev MB, Barnes WM. 2010. Direct
DNA amplification from crude clinical samples using a
PCR enhancer cocktail and novel mutants of Taq. J Mol
Diagn 12:152–161.
77. Hoshino T, Inagaki F. 2012. Molecular quantification of
environmental DNA using microfluidics and digital PCR.
Syst Appl Microbiol 35:390–395.
78. Akane A, Matsubara K, Nakamura H, Takahashi S,
Kimura K. 1994. Identification of the heme compound cop-
urified with deoxyribonucleic acid (DNA) from bloodstains,
a major inhibitor of polymerase chain reaction (PCR) ampli-
fication. J Forensic Sci 39:362–372.
79. Akane A, Shiono H, Matsubara K, Nakamura H, Hase-
gawa M, Kagawa M. 1993. Purification of forensic specimens
for the polymerase chain reaction (PCR) analysis. J Forensic
Sci 38:691–701.
80. Comey C, Koons B, Presly K, Smerick J, Sobieralski C,
Sanley D, Baechtel F. 1994. DNA extraction strategies for
amplified fragment length polymorphism analysis.
J Forensic Sci 39:1254–1269.
81. Eilert KD, Foran DR. 2009. Polymerase resistance to poly-
merase chain reaction inhibitors in bone. J Forensic Sci 54:
1001–1007.
82. Chen NT, Chang CW, Wu JD. 2012. Evaluation of DNA
isolation procedures for detecting and quantifying environ-
mental Legionella by real-time quantitative PCR. Water Sci
Technol 65:989–997.
83. Coates SG, Brunelle SL, Davenport MG. 2011. Develop-
ment of standard method performance requirements for
biological threat agent detection methods. J AOAC Int 94:
1328–1337.
84. AOAC. 2011. AOAC SMPR 2010.001: Standard method
performance requirements for polymerase chain reaction
(PCR) methods for detection of Francisella tularensis in
aerosol collection filters and/or liquids. J AOAC Int 94:
1338–1341.
85. AOAC. 2011. AOAC SMPR 2010.002: Standard method
performance requirements for polymerase chain reaction
(PCR) methods for detection of Yersinia pestis in aerosol col-
lection filters and/or liquids. J AOAC Int 94:1342–1346.
86. AOAC. 2011. AOAC SMPR 2010.003: Standard method
performance requirements for polymerase chain reac-
tion (PCR) methods for detection of Bacillus anthracis in
aerosol collection filters and/or liquids. J AOAC Int 94:
1347–1351.
87. AOAC. 2011. AOAC International Methods Committee
guidelines for validation of biological threat agent methods
and/or procedures. J AOAC Int 94:1359–1381.
88. Jarman KH. 2013. The Art of Data Analysis: How to Answer
Almost Any Question Using Basic Statistics. John Wiley &
Sons, New York.
89. PNNL. 2011. Visual sample plan. http://vsp.pnnl.gov.
90. Barnett F. 2010. ProUCL Version 4.1 User Guide: statistical
software for environmental applications for data sets with
and without nondetect observations. EPA/600/R-07/41,
http://www.epa.gov/osp/hstl/tsc/software.htm.
91. Biggerstaff BJ. 2008. Confidence intervals or the difference
of two proportions estimated from pooled samples. J Agric
Biol Environ Stat 13:478–496.
92. Biggerstaff BJ. 2009. PooledInRate, version 40: a Microsoft
Office Excel add in to compute prevalence estimates from pooled
samples. CDC, Fort Collins, CO.
93. Gerlach RW, Nocerino JM. 2003. Guidance for obtaining
representative laboratory analytic subsamples from particu-
late laboratory samples. EPA/600/R-03/27.
94. Baddeley A, Turner R. 2005. Spastat: an R package for ana-
lyzing spatial point patterns. J Stat Softw 12:1–42.
95. Zhang B, Bilder C, Biggerstaff B, Schaarschmidt F. 2012.
Package “binGroup.” http://cran.r-project.org/web/pack
ages/binGroup/index.html.
96. Pipper J, Zhang Y, Neuzil P, Hsieh T-M. 2008. Clockwork
PCR including sample preparation. Ang Chem Int Ed 47:
3900–3904.
 Introduction to Microbial Community Analysis of
Environmental Samples with Next-Generation Sequencing
STEFAN J. GREEN AND JOSH D. NEUFELD
2.4.1
This section of the MEM4 examines the addition of NGS
to the toolbox of environmental microbiologists and micro-
bial ecologists. High-throughput sequencing technologies
have been embraced as a welcome extension of more tradi-
tional molecular methods that have been used for nearly three
decades. The application of NGS tools to the study of micro-
bial communities has progressed rapidly due to reduced labor
and costs associated with NGS data acquisition, compared
with cloning and capillary electrophoresis (Sanger) sequenc-
ing, and due to recent dramatic increases in read length
and sequence quality. NGS approaches have opened the
door on microbial community surveys and experiments that
were unimaginable a few years ago. High-throughput, deep
sequencing of gene amplicons, for example, has brought
much-needed robustness to the characterization of microbial
community structure, has provided for substantial advances in
the field with regard to the understanding of microbial diver-
sity, and has driven the development of novel bioinformatics
analysis pipelines to streamline the analysis of these large data
sets. NGS has also opened new avenues for the exploration of
microbial communities: shotgun DNA sequencing (metage-
nomics) and shotgun mRNA sequencing (metatranscriptom-
ics) have further generated substantial advances in the field,
across many different environmental systems (1–3) and offer
additional insight into the structure and function of microbial
communities. New molecular tools have also allowed the gen-
eration of NGS data from minute quantities of nucleic acids
(on the scale of 1 ng DNA or RNA; (e.g., 4), and these tools
are a boon to environmental microbiologists working with
limited biomass or with enrichment strategies such as com-
bined stable isotope probing and metagenomics (5). Reduced
costs and high data yields from NGS approaches have enabled
greater replication during sampling and analysis and have pro-
vided for more sophisticated surveys of microbial community
diversity, composition, and structure, leading to robust
hypothesis testing (e.g., 6). The high data yield from shotgun
sequencing allows scientists to investigate microbial ecosys-
tems without a priori information regarding gene content
and sequences, without resorting to selective enrichment of
specific taxa or functional groups, and without known biases
associated with primer-mediated sequencing approaches. Fur-
thermore, in shotgun sequencing approaches the high data
yield can provide for partial or even complete assembly of
genomes of dominant microorganisms (e.g., 7). The increase
in data output has occurred in tandem with the development
doi:10.1128/9781555818821.ch2.4.1
2.4.1-1
of algorithms and databases for analyzing and storing large
data sets. One concern is that computational capacity may
be outstripped by the production of data (e.g., 8).
In this introduction, we briefly survey developments in the
field of molecular biology that have been applied to environ-
mental microbiology, consider major advances and areas
where progress has been slow, and highlight chapters in this
and other sections of the MEM4 relevant to nucleic acid–
based molecular analyses of microbial communities. Experts
on various aspects of NGS have contributed to this section,
including chapters addressing the analysis of microbial
communities through the study of (1) ribosomal RNA
gene amplicons (Chapter 2.4.2); (2) shotgun metagenomic
sequence data (Chapter 2.4.4); and (3) shotgun metatran-
scriptomic sequence data (Chapter 2.4.5). In addition, this
section includes a chapter reviewing functional metagenom-
ics (Chapter 2.4.3).
DEVELOPMENT OF MOLECULAR BIOLOGY
TOOLS FOR ENVIRONMENTAL
MICROBIOLOGY
In the MEM3, Wen-Tso Liu and David A. Stahl wrote
an excellent review of molecular tools for microbial commun-
ity analyses. This prior review included discussions of the lim-
itations on recovery of nucleic acids from environmental
samples and the implications for the interpretation of molec-
ular data. In addition, they reviewed standard approaches
in the field, including the design of probes for performing
quantification through Northern hybridization (RNA tar-
geted with DNA probes), restriction enzyme digestion and
electrophoresis of nucleic acids (restriction fragment length
polymorphism, RFLP), DNA reassociation studies for esti-
mating genome-genome similarity (DNA-DNA hybridiza-
tion, DDH), community fingerprinting techniques, and
cloning and sequencing of small subunit (SSU) rRNA genes.
Although these techniques are still conducted in many
laboratories, remarkable technological developments have
resulted in a dramatic shift in the dominant approaches to
microbial community analysis. For example, hybridization
approaches for quantification at the domain- and lower taxo-
nomic levels have been largely replaced with quantitative
PCR (qPCR) approaches. The field appears to be on the
verge of a further revolution in qPCR techniques, from
real-time PCR to digital PCR (9). Likewise, the suite of
2.4.1-2 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
fingerprinting techniques such as denaturing gradient gel
electrophoresis (DGGE), terminal RFLP (T-RFLP), and
automated ribosomal intergenic spacer analysis (ARISA),
although still extant, have gradually decreased in frequency
of use for microbial community analysis. This trend has
been fueled by the rapidly decreasing cost in gene amplicon
sequencing (10–12), the availability of methods, and core
facilities and commercial enterprises capable of processing
large numbers of samples using multiplex identifier (barcode)
sequences. Thus, the cost differential between a fingerprint-
ing technique and an NGS multiplexed amplification and
sequencing approach is narrowing. This is particularly true
if the cost associated with recovering sequence data from
select bands in fingerprinting approaches is considered, as is
commonly practiced with DGGE analysis (13). Fingerprint-
ing techniques were particularly useful when sequence data
were expensive and time-consuming to acquire, but the rap-
idity of data acquisition, the reduced labor (particularly if
sequencing is outsourced to commercial enterprises or core
facilities), and the modest cost now favor NGS approaches
for most microbial community profiling. Furthermore, the
advent of moderately priced next-generation sequencers
can bring high-throughput analyses to the laboratories of
individual researchers.
A more fundamental difference underlies the shift toward
NGS data as well. Community fingerprinting techniques
such as T-RFLP or DGGE are data reduction strategies; the
whole of a community is reduced to a relatively simple band-
ing pattern representing the most abundant taxa. These tech-
niques are explicitly a proxy for community composition, and
although some sequence information can be extracted, most
of this information is lost in the analysis. Conversely, NGS
sequence data starts with high sequence information, and
these data are progressively reduced to increase ease of analysis
and facilitate application of multisample analytical strategies
(Chapter 2.4.2). In these situations, however, the original
data remain and can be readily extracted. Finally, we note that
NGS techniques are much more amenable to scale-up of sam-
ple number. Fingerprinting techniques such as DGGE can be
technically challenging and are not easily expanded to large
datasets due to inter-gel variability. Although run-to-run var-
iability exists with NGS as well, sequence data generated in
one run can be compared or pooled with prior data without
difficulty.
Other common techniques are also being replaced by
sequence data acquisition. For example, DDH techniques
have been the standard for characterizing novel microbial
species and are still used, especially when describing novel
species with rRNA genes with greater than 97% sequence
similarity across the full SSU rRNA gene (14–16). As with
DGGE, T-RFLP, or ARISA, DDH is an approximation and
is rapidly being replaced by whole genome de novo sequencing
(16). Whole genome sequences are readily acquired, and
novel methods for analysis have been shown to be consistent
with DDH, including calculations of average nucleotide iden-
tity and conserved DNA percentage (17). Furthermore, DDH
techniques require substantial quantities of DNA, partic-
ularly considering that a separate DDH analysis must be
conducted for each pair of organisms. These concerns are cir-
cumvented by whole genome sequencing. Fairly limited
genomic DNA (as low as 1 ng) is necessary for the generation
of a reasonable draft genome, though much more DNA is
required for the closing of a draft genome. In addition, pair-
wise comparison of whole genome sequence data can be per-
formed computationally with the addition of each new
genome sequence and does not require additional wet-lab
work, and the whole genome sequence is far more valuable
than DDH data. As such, genome-genome comparison for
the purposes of defining a novel species is only a very small
part of the utility of such data. We note that the isolation
and physiological and genetic characterization of microor-
ganisms is still a critical component of research in microbial
ecology and environmental microbiology. Isolation of micro-
organisms is necessary for experimental testing of physiologi-
cal potential and for validating predictions of microbial
behavior under varying environmental conditions. Con-
versely, metagenomic data can be used to guide isolation
efforts.
The isolation of environmental microorganisms serves
another role: de novo genome sequencing from an isolate pro-
vides key genetic data that can be used for interpretation of
shotgun metagenomic and metatranscriptomic sequence
data (e.g., 18). Reference genomes can also, for example, pro-
vide the basis for understanding genetic variation among
closely related organisms and for identifying novel genetic
material acquired via lateral gene transfer. These assembled
genomes can further be used to develop metabolic models and
generate hypotheses about metabolic functional potential of
organisms (e.g., 19, 20). Hypotheses then can be tested using
in vivo experiments with microbial isolates, and transcript
profiling (transcriptome sequencing) of isolates grown under
experimental conditions. Such data are used to validate or
improve metabolic models, leading to a more profound under-
standing of the in situ activity of microbial taxa (e.g., 21).
The most striking development in the field of molecular
microbial ecology and environmental microbiology is that
of shotgun sequencing, which is the sequencing of genetic
material without a priori knowledge (e.g., such as conserved
primer sites used for PCR). This is achieved, for example,
through sequence-independent ligation of sequencing adapt-
ers to fragmented genomic DNA or to double-stranded
complementary DNA (ds-cDNA) generated by reverse tran-
scription (and other sequence-independent molecular mani-
pulations) of RNA. The ligation of a known DNA fragment
(i.e., sequencing adapters) to environmental sequences
allows for the initiation of polymerase copying, which under-
lies most current sequencing technologies. The diversity of
many environmental samples is such that a shotgun sequenc-
ing approach employing cloning and capillary electrophoresis
sequencing was, and still is, costly enough to limit metage-
nomic analyses to certain high-profile studies and represen-
tative samples (e.g., 22, 23). Nonetheless, innovative
approaches were used to identify metagenomic DNA associ-
ated with metabolic functions of interest through the gener-
ation and screening and sequencing of plasmids, bacterial
artificial chromosomes, and cosmids containing large metage-
nomic DNA fragments (reviewed in 24). These screening
approaches can be effective, but with the advent of NGS,
function-independent analysis of total microbial genetic
material can be performed; selection for genes of interest is
then performed post-sequencing. The high sequence output
from shotgun sequencing approaches provides for deep char-
acterization of environmental microbial communities,
although this varies with diversity and genetic complexity
in each sample. In systems with more limited microbial rich-
ness or highly skewed abundances/low diversity, substantial
assembly and even closure of individual microbial genomes
can be obtained (e.g., 7, 25).
Shotgun sequencing of mRNAs isolated from individual
organisms (reviewed in 26) and environmental systems
(e.g., 27, 28) has also benefited from NGS approaches. Previ-
ously, shotgun mRNA sequencing or RNA-seq was
 2.4.1. Microbial Community Analysis of Environmental Samples with Next-Generation Sequencing ▪ 2.4.1-3
performed using methods such as differential display (29) and
serial analysis of gene expression (SAGE; 30). The limita-
tions of these approaches include relatively short sequences,
in the case of SAGE, and the requirement for a reference
genome, the low resolution of the method, and dependence
on electrophoretic separation of randomly amplified frag-
ments (differential display). For RNA-seq, full transcriptional
profiles of individual organisms demonstrate treatment-
specific shifts in expression patterns. In the future, increasing
sequence read length will allow for more accurate determina-
tion of genes and taxonomies of the organisms from which the
mRNAs were derived. Furthermore, in conjunction, shotgun
sequencing of genomic DNA and mRNA-derived cDNA can
provide for a more holistic understanding of microbial ecosys-
tems. Simultaneous analysis of rRNA and mRNAs has also
been advocated (31). Another development from shotgun
approaches ( particularly metagenomic sequencing) includes
the routine detection of nonbacterial microorganisms,
including archaea, fungi, micro-eukaryotes, and viruses. In
bulk analyses, the relative genetic contribution of these vari-
ous groups of organisms to the system as a whole can be
assessed (32–34).
It should be noted that although shotgun metagenome
and metatranscriptome sequencing are powerful techniques,
many limitations remain. The high diversity of some micro-
bial ecosystems, particularly soil, and the lack of cultured rep-
resentatives of many taxa contained therein, limit the utility
of metagenomic sequencing. Under these conditions, short
individual reads may not provide adequate taxonomic resolu-
tion, leaving many sequences of undetermined affiliation.
Although de novo assembly of short reads into longer contigs
can be performed to improve sequence identification, shot-
gun metagenome sequencing of genomic DNA extracted
from high-diversity environments with relatively even distri-
bution of taxa may only yield limited assembly. In shotgun
metatranscriptome sequencing, the mRNA pool contains
abundant transcripts related to housekeeping functions. Fur-
thermore, genes may be expressed constitutively and can
obfuscate shifts in expression. Various options are available
to researchers for reducing system complexity and addressing
scientific questions of interest. These include, but are not
limited to, experimental enrichments (35), fluorescence-
activated cell sorting and genome sequencing (36), and stable
isotope probing (5). So-called targeted metagenomics has
been reviewed elsewhere (37). These techniques can be
used prior to shotgun sequencing approaches, and through
reduction in community diversity and expression diversity
produce more targeted results, leading to identification of
key taxa involved in environmental processes and the gene
expression patterns of specific organisms within a complex
community.
MICROBIAL ECOLOGY IS A FULL-FLEDGED
FIELD OF ECOLOGY
The field of microbial ecology is now a full member of the field
of ecology, and standard ecological statistical approaches
have been fully assimilated and expanded on (6, 38, 39).
This development has been achieved in part because of the
dramatic increase in sequence data acquisition; these data
are more amenable to statistical analysis than fingerprinting
approaches due to rigorous data reduction strategies. Deep
sampling that is now commonly achieved with NGS, partic-
ularly with respect to single gene analyses such as rRNA
genes, is more robust, and a more complete image of the
microbial community is obtained. Interestingly, such deep
coverage can enable the testing of ecological hypothesis
regarding the diversity and distribution of microbial com-
munities (e.g., 40), by sequencing to a depth approaching
the full coverage of a community (e.g., sequencing of 108
rRNA gene amplicon sequences from a community extract
could nearly approach the full number of genes within the
sample), although the measured diversity can be confounded
by the error rate of PCR amplification and sequencing (41).
Furthermore, due to the relative ease and modest cost, repli-
cation is more prevalent, and such data are adding to the
robustness of the field of microbial ecology and environ-
mental microbiology. Whereas extensive replication was
prohibitively costly previously, such replication—powerfully
advocated by Prosser (42)—can now be readily incorporated
into sequencing strategies with limited financial burden.
Although most current studies in the field are correlation
analyses, and experimental work is necessary to establish a
causal link to these findings (e.g., 39), we nonetheless prog-
ress toward of our goal of linking microbial community struc-
ture and system function. Through this understanding we aim
to predict the impact of changes to environmental systems
from natural and anthropogenic events and identify
approaches to manipulate microorganisms and microbial eco-
systems for the benefit of humanity.
CONTINUING OBSTACLES
Relative to traditional capillary sequencing, the most com-
monly applied NGS platforms produce short reads, on the
order of 100–500 bases. These short reads can be problematic
for identification purposes and certainly for de novo assembly.
In the absence of closed reference genomes, much genome
context is lost with shotgun metagenome sequence data.
Read length is increasing on the dominant sequencing plat-
forms (Illumina, Ion Torrent), and is already longer on other
platforms such as the Roche 454 pyrosequencer and Pacific
Biosystems instruments. These platforms are less common
due to high cost per base and, as for PacBio, have high error
rates (∼10%). Nonetheless, new technologies in develop-
ment promise dramatic increases in read length (43), and
this will be a tremendous boon for genomic and metagenomic
sequencing. For de novo bacterial genome sequencing, draft
genome assemblies are relatively easy, but full closure of bac-
terial chromosomes and large plasmids are often exceedingly
difficult, and this is in part due to repetitive sequences,
including ribosomal RNA gene operons, present within
genomes that confound assembly of short sequences (44).
As reads of high accuracy exceed the length of the ribosomal
RNA genes, the complete de novo assembly of bacterial
genomes will be greatly simplified. Likewise, long shotgun
reads will improve gene identification, gene context, and
assembly of partial genomes from environmental systems in
metagenomic and metatranscriptomic studies.
Community structure analyses have greatly benefited from
high-throughput analysis of rRNA genes. The adoption of
primers targeting the V4 region of bacterial and archaeal
SSU rRNA genes by the Earth Microbiome Project (http://
www.earthmicrobiome.org/) has helped encourage compari-
son of data sets between different researchers. Although the
advantages of rRNAs and rRNA genes for community anal-
yses are well known, the limitations of this molecule are
also becoming clearer as samples are more thoroughly charac-
terized. One limitation includes the known variability in
copy number of the rRNA gene operon between different lin-
eages of organisms (45, 46). This variability confounds com-
munity analyses based on SSU rRNA genes because relative
2.4.1-4 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
abundance in these libraries does not correlate directly with
relative abundance of cells in the sample of interest. Several
novel approaches have been developed to address this limita-
tion, through a correction based on known gene copy number
in closely related taxa (47). Furthermore, the short length of
most amplicon sequencing approaches and the 16S rRNA
gene itself, limit the taxonomic resolution of the approach.
Increases in sequencing length will generate more reliable
data for classification purposes, and eventually read lengths
may be long enough to include rRNA genes and internal
transcribed spacer regions, providing for classification of
sequences across a wide range of taxonomic levels. As with
any PCR-based approach, amplicon sequencing lends itself
to bias associated with primer design (e.g., lower amplifica-
tion efficiency of templates with mismatches) and with GC
content (even more relevant for shotgun sequencing). Most
sequence platforms, except for Pacific Biosciences (i.e., Pac-
Bio RS II), utilize PCR amplification at some stage during
the library preparation or template preparation stages, and
thus are also susceptible to loss of very low and very high
GC content templates (e.g., 48).
Although shotgun sequencing approaches are favored
because no a priori selection is performed, low-frequency genes
of interest may be rare enough that even deep (and costly)
sequencing efforts do not provide enough data for compara-
tive purposes. Novel strategies are needed to retain the advan-
tages of shotgun sequencing to capture unknown gene
variants with selection for genes or organisms of interest.
Finally, although sequencing costs are decreasing rapidly,
the preparation of high-quality libraries is still relatively
high. Thus, sequencing of biological replicates is often dispro-
portionally expensive due to the need to produce multiple
libraries from replicate samples. As library preparation and
equimolar pooling strategies becomes more automated, these
costs are expected to decrease. This may allow, for example,
rapid analysis of hundreds of genomes or metagenomes in a
single sequencing run.
The vast amounts of data generated with NGS approaches
should also be considered as a continuing obstacle. Data files
are often so large (e.g., ∼40 Gb zipped files from individual
paired-end Illumina HiSeq2000 lanes) that data transfer is a
significant bottleneck. Subsequently, it can be computation-
ally demanding to perform quality control checks on large
data sets, much less perform de novo assembly. Thus, central-
ized computer clusters are often necessary to process and store
NGS data. As a corollary, the tremendous boom in data pro-
duction has led to significant development of novel algo-
rithms and software packages for processing and analyzing
amplicon (12, 49) and shotgun NGS data (reviewed recently
in 8).
Finally, it should be noted that NGS technologies are
open platforms that favor discovery of novel transcripts,
genes, and organisms. Samples are prepared for sequencing
using sequence-independent strategies (metagenome and
metranscriptome sequencing) or extremely broad PCR-based
amplification approaches (domain-level amplicon sequenc-
ing of rRNA genes). In these cases, where no references are
used (e.g., de novo assembly of raw reads, sequence similar-
ity–based clustering of rRNA gene amplicons), the computa-
tional burden is much greater in comparison to reference
mapping strategies. In certain cases, alternate, lower informa-
tion strategies may be valid, produce less data, and have a
much reduced bioinformatic burden. These strategies include
microarray approaches with a fixed set of targets (e.g., Geo-
chip [50], Phylochip [51]) and quantitative PCR arrays target-
ing narrow sets of microbial taxa or functional genes of
interest (e.g., Qiagen Microbial DNA qPCR Arrays). These
narrower approaches have been enabled through discovery
using open platforms, including NGS, and in some cases, par-
ticularly in human-associated microbial communities, we can
reap the benefits of prior large and open data sets by develop-
ing targeted and closed data sets that are easier to analyze.
SUMMARY
The field of microbial ecology has a suite of technologies
to enable thorough characterization of the structure of micro-
bial communities and the activity of microorganisms in the
environment. In particular, NGS technologies can be used
in a number of areas: community structure analyses based
on rRNA gene sequencing, functional gene surveys using
amplicon sequencing with gene-specific and degenerate
primers, de novo bacterial genome sequencing (isolates, cocul-
tures, enrichments), bacterial isolate resequencing (mutation
detection, rearrangement detection, lateral gene transfer
detection), bacterial transcriptome sequencing, metagenome
sequencing, and metatranscriptome sequencing. The choice
of NGS platform is heavily dependent on the specific appli-
cation, available resources (financial and bioinformatics),
and patience of the investigator. In addition to monumental
advances in the field of NGS, which garners most of the
attention, there have been ancillary technological advances
that either directly enable NGS applications, validate NGS
data, or serve as critical components of complete data sets
(i.e., more than just NGS data). As NGS technologies mature
and become routine in the field of microbial ecology and
environmental microbiology, sequence data will be supple-
mented with a much broader spectrum of data. These types
of accessory data may include quantitative data such as gene
abundance, activity data based on transcript sequencing
and proteomics, isolation and genomic characterization of
multiple isolates of relevance, and natural isotopic composi-
tion measurements. The tools we have dreamed of are here,
and we only have to seize them to address critical scientific
goals regarding microbial ecosystem function and the manip-
ulation of microbial ecosystems for beneficial uses. We antici-
pate that the chapters in this manual will serve as a guide for
researchers to use these techniques to enable scientific study
of microbial communities.
REFERENCES
1. Brisson VL, West KA, Lee PK, Tringe SG, Brodie EL,
Alvarez-Cohen L. 2012. Metagenomic analysis of a stable
trichloroethene-degrading microbial community. ISME J
6:1702–1714.
2. Mackelprang R, Waldrop MP, DeAngelis KM, David MM,
Chavarria KL, Blazewicz SJ, Rubin EM, Jansson JK. 2011.
Metagenomic analysis of a permafrost microbial community
reveals a rapid response to thaw. Nature 480:368–371.
3. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Mani-
chanh C, Nielsen T, Pons N, Levenez F, Yamada T, Mende
DR, Li J, Xu J, Li S, Li D, Cao J, Wang B, Liang H, Zheng
H, Xie Y, Tap J, Lepage P, Bertalan M, Batto JM, Hansen
T, Le Paslier D, Linneberg A, Nielsen HB, Pelletier E,
Renault P, Sicheritz-Ponten T, Turner K, Zhu H, Yu C,
Jian M, Zhou Y, Li Y, Zhang X, Qin N, Yang H, Wang J,
Brunak S, Dore J, Guarner F, Kristiansen K, Pedersen
O, Parkhill J, Weissenbach J, Bork P, Ehrlich SD. 2010.
A human gut microbial gene catalogue established by meta-
genomic sequencing. Nature 464:59–65.
4. Marine R, Polson SW, Ravel J, Hatfull G, Russell D, Sul-
livan M, Syed F, Dumas M, Wommack KE. 2011.
 2.4.1. Microbial Community Analysis of Environmental Samples with Next-Generation Sequencing ▪ 2.4.1-5
Evaluation of a transposase protocol for rapid generation of
shotgun high-throughput sequencing libraries from nano-
gram quantities of DNA. Appl Environ Microbiol 77:
8071–8079.
5. Neufeld JD, Dumont MG, Vohra J, Murrell JC. 2007.
Methodological considerations for the use of stable
isotope probing in microbial ecology. Microb Ecol 53:
435–442.
6. Ramette A. 2007. Multivariate analyses in microbial ecol-
ogy. FEMS Microbiol Ecol 62:142–160.
7. Iverson V, Morris RM, Frazar CD, Berthiaume CT,
Morales RL, Armbrust EV. 2012. Untangling genomes
from metagenomes: revealing an uncultured class of marine
Euryarchaeota. Science 335:587–590.
8. Scholz MB, Lo CC, Chain PS. 2012. Next generation
sequencing and bioinformatic bottlenecks: the current
state of metagenomic data analysis. Curr Opin Biotechnol
23:9–15.
9. Hoshino T, Inagaki F. 2012. Molecular quantification of
environmental DNA using microfluidics and digital PCR.
Syst Appl Microbiol 35:390–395.
10. Bartram AK, Lynch MD, Stearns JC, Moreno-Hagelsieb
G, Neufeld JD. 2011. Generation of multimillion-sequence
16S rRNA gene libraries from complex microbial commun-
ities by assembling paired-end Illumina reads. Appl Environ
Microbiol 77:3846–3852.
11. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D,
Lozupone CA, Turnbaugh PJ, Fierer N, Knight R. 2011.
Global patterns of 16S rRNA diversity at a depth of millions
of sequences per sample. Proc Natl Acad Sci USA 108:
4516–4522.
12. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K,
Bushman FD, Costello EK, Fierer N, Pena AG, Goodrich
JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koenig
JE, Ley RE, Lozupone CA, McDonald D, Muegge BD,
Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ, Walters
WA, Widmann J, Yatsunenko T, Zaneveld J, Knight R.
2010. QIIME allows analysis of high-throughput community
sequencing data. Nat Meth 7:335–336.
13. Green SJ, Leigh MB, Neufeld JD. 2009. Denaturing gra-
dient gel electrophoresis (DGGE) for microbial community
analysis, p 4137–4158. In Timmis KN (ed.), Handbook of
Hydrocarbon and Lipid Microbiology. Springer, Heidelberg.
14. Stackebrandt E, Frederiksen W, Garrity GM, Grimont
PAD, Kampfer P, Maiden MCJ, Nesme X, Rossello-Mora
R, Swings J, Truper HG, Vauterin L, Ward AC, Whitman
WB. 2002. Report of the ad hoc committee for the re-
evaluation of the species definition in bacteriology. Int J
Syst Evol Micr 52:1043–1047.
15. Stackebrandt E, Goebel BM. 1994. A place for DNA-DNA
reassociation and 16s ribosomal RNA sequence analysis in
the present species definition in bacteriology. Int J Syst Bac-
teriol 44:846–849.
16. Tindall BJ, Rossello-Mora R, Busse HJ, Ludwig W, Kamp-
fer P. 2010. Notes on the characterization of prokaryote
strains for taxonomic purposes. Int J Syst Evol Microbiol 60:
249–266.
17. Goris J, Konstantinidis KT, Klappenbach JA, Coenye T,
Vandamme P, Tiedje JM. 2007. DNA-DNA hybridization
values and their relationship to whole-genome sequence
similarities. Int J Syst Evol Microbiol 57:81–91.
18. Oh S, Caro-Quintero A, Tsementzi D, DeLeon-Rodriguez
N, Luo C, Poretsky R, Konstantinidis KT. 2011. Metage-
nomic insights into the evolution, function, and complexity
of the planktonic microbial community of Lake Lanier, a
temperate freshwater ecosystem. Appl Environ Microbiol 77:
6000–6011.
19. Zomorrodi AR, Maranas CD. 2012. OptCom: a multi-level
optimization framework for the metabolic modeling and
analysis of microbial communities. PLoS Comput Biol 8:
e1002363.
20. Edwards JS, Covert M, Palsson B. 2002. Metabolic model-
ling of microbes: the flux-balance approach. Environ Micro-
biol 4:133–140.
21. Park JH, Lee KH, Kim TY, Lee SY. 2007. Metabolic engi-
neering of Escherichia coli for the production of L-valine
based on transcriptome analysis and in silico gene knockout
simulation. Proc Natl Acad Sci USA 104:7797–7802.
22. Nealson KH, Venter JC. 2007. Metagenomics and the
global ocean survey: what’s in it for us, and why should we
care? ISME J 1:185–187.
23. Gill SR, Pop M, Deboy RT, Eckburg PB, Turnbaugh PJ,
Samuel BS, Gordon JI, Relman DA, Fraser-Liggett CM,
Nelson KE. 2006. Metagenomic analysis of the human distal
gut microbiome. Science 312:1355–1359.
24. Mocali S, Benedetti A. 2010. Exploring research frontiers in
microbiology: the challenge of metagenomics in soil micro-
biology. Res Microbiol 161:497–505.
25. Schloss PD, Handelsman J. 2005. Metagenomics for study-
ing unculturable microorganisms: cutting the Gordian knot.
Genome Biol 6:229.
26. Croucher NJ, Thomson NR. 2010. Studying bacterial tran-
scriptomes using RNA-seq. Curr Opin Microbiol 13:
619–624.
27. Gilbert JA, Field D, Huang Y, Edwards R, Li W, Gilna P,
Joint I. 2008. Detection of large numbers of novel sequences
in the metatranscriptomes of complex marine microbial
communities. PLoS One 3:e3042.
28. Poretsky RS, Bano N, Buchan A, LeCleir G, Kleikemper
J, Pickering M, Pate WM, Moran MA, Hollibaugh JT.
2005. Analysis of microbial gene transcripts in environmen-
tal samples. Appl Environ Microbiol 71:4121–4126.
29. Liang P, Pardee AB. 1992. Differential display of eukaryotic
messenger RNA by means of the polymerase chain reaction.
Science 257:967–971.
30. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. 1995.
Serial analysis of gene expression. Science 270:484–487.
31. Urich T, Lanzen A, Qi J, Huson DH, Schleper C, Schuster
SC. 2008. Simultaneous assessment of soil microbial com-
munity structure and function through analysis of the meta-
transcriptome. PLOS ONE 3:e2527.
32. Fierer N, Breitbart M, Nulton J, Salamon P, Lozupone C,
Jones R, Robeson M, Edwards RA, Felts B, Rayhawk S,
Knight R, Rohwer F, Jackson RB. 2007. Metagenomic
and small-subunit rRNA analyses reveal the genetic diver-
sity of bacteria, archaea, fungi, and viruses in soil. Appl Envi-
ron Microbiol 73:7059–7066.
33. Glass EM, Wilkening J, Wilke A, Antonopoulos D, Meyer
F. 2010. Using the metagenomics RAST server
(MG-RAST) for analyzing shotgun metagenomes. Cold
Spring Harb Protoc 2010:pdb prot5368.
34. Willner D, Hugenholtz P. 2013. Metagenomics and com-
munity profiling: culture-independent techniques in the
clinical laboratory. Clin Microbiol Newsl 35:1–9.
35. Knietsch A, Waschkowitz T, Bowien S, Henne A, Daniel
R. 2003. Construction and screening of metagenomic libra-
ries derived from enrichment cultures: generation of a gene
bank for genes conferring alcohol oxidoreductase activity
on Escherichia coli. Appl Environ Microbiol 69:1408–1416.
36. Woyke T, Xie G, Copeland A, Gonzalez JM, Han C, Kiss
H, Saw JH, Senin P, Yang C, Chatterji S, Cheng JF, Eisen
JA, Sieracki ME, Stepanauskas R. 2009. Assembling
the marine metagenome, one cell at a time. PLOS ONE 4:
e5299.
37. Suenaga H. 2012. Targeted metagenomics: a high-
resolution metagenomics approach for specific gene clusters
in complex microbial communities. Environ Microbiol 14:
13–22.
2.4.1-6 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
38. Lozupone C, Lladser ME, Knights D, Stombaugh J, Knight
R. 2011. UniFrac: an effective distance metric for microbial
community comparison. ISME J 5:169–172.
39. Warwick RM, Clarke KR. 2006. Primer 6. Primer-E,
Plymouth.
40. Prosser JI, Bohannan BJ, Curtis TP, Ellis RJ, Firestone
MK, Freckleton RP, Green JL, Green LE, Killham K, Len-
non JJ, Osborn AM, Solan M, van der Gast CJ, Young JP.
2007. The role of ecological theory in microbial ecology. Nat
Rev Microbiol 5:384–392.
41. Kunin V, Engelbrektson A, Ochman H, Hugenholtz P.
2010. Wrinkles in the rare biosphere: pyrosequencing errors
can lead to artificial inflation of diversity estimates. Environ
Microbiol 12:118–123.
42. Prosser JI. 2010. Replicate or lie. Environ Microbiol 12:
1806–1810.
43. Eisenstein M. 2012. Oxford Nanopore announcement sets
sequencing sector abuzz. Nat Biotechnol 30:295–296.
44. Bashir A, Klammer AA, Robins WP, Chin CS, Webster D,
Paxinos E, Hsu D, Ashby M, Wang S, Peluso P, Sebra R,
Sorenson J, Bullard J, Yen J, Valdovino M, Mollova E,
Luong K, Lin S, LaMay B, Joshi A, Rowe L, Frace M,
Tarr CL, Turnsek M, Davis BM, Kasarskis A, Mekalanos
JJ, Waldor MK, Schadt EE. 2012. A hybrid approach for the
automated finishing of bacterial genomes. Nat Biotechnol 30:
701–707.
45. Case RJ, Boucher Y, Dahllof I, Holmstrom C, Doolittle
WF, Kjelleberg S. 2007. Use of 16S rRNA and rpoB genes
as molecular markers for microbial ecology studies. Appl
Environ Microbiol 73:278–288.
46. Klappenbach JA, Saxman PR, Cole JR, Schmidt TM.
2001. rrndb: the ribosomal RNA operon copy number data-
base. Nucleic Acids Res 29:181–184.
47. Kembel SW, Wu M, Eisen JA, Green JL. 2012. Incorporat-
ing 16S gene copy number information improves estimates of
microbial diversity and abundance. PLoS Comput Biol 8:
e1002743.
48. Oyola SO, Otto TD, Gu Y, Maslen G, Manske M,
Campino S, Turner DJ, Macinnis B, Kwiatkowski DP,
Swerdlow HP, Quail MA. 2012. Optimizing Illumina next-
generation sequencing library preparation for extremely
AT-biased genomes. BMC Genomics 13:1.
49. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann
M, Hollister EB, Lesniewski RA, Oakley BB, Parks DH,
Robinson CJ, Sahl JW, Stres B, Thallinger GG, Van
Horn DJ, Weber CF. 2009. Introducing mothur: open-
source, platform-independent, community-supported soft-
ware for describing and comparing microbial communities.
Appl Environ Microbiol 75:7537–7541.
50. He Z, Gentry TJ, Schadt CW, Wu L, Liebich J, Chong SC,
Huang Z, Wu W, Gu B, Jardine P, Criddle C, Zhou J. 2007.
GeoChip: a comprehensive microarray for investigating bio-
geochemical, ecological and environmental processes. ISME
J 1:67–77.
51. DeAngelis KM, Wu CH, Beller HR, Brodie EL, Chakra-
borty R, DeSantis TZ, Fortney JL, Hazen TC, Osman
SR, Singer ME, Tom LM, Andersen GL. 2011. PCR
amplification-independent methods for detection of micro-
bial communities by the high-density microarray Phylo-
Chip. Appl Environ Microbiol 77:6313–6322.
 Microbial Community Analysis Using High-Throughput
Amplicon Sequencing
DANNY IONESCU, WILL A. OVERHOLT, MICHAEL D. J. LYNCH, JOSH D. NEUFELD,
ANKUR NAQIB, AND STEFAN J. GREEN
2.4.2
PCR is one of the fundamental underpinnings of modern
molecular biology. The reaction is an in vitro manipulation
of cellular DNA replication machinery and serves as a tar-
geted copying mechanism. Through polymerase-mediated
copying, large quantities of a specific piece of DNA can be
made, and these copies are amenable to subsequent molecular
analyses. Microbial ecologists employ PCR amplification and
sequencing to screen environmental nucleic acid extracts for
characterizing microbial community composition (e.g., 1–3)
because many microorganisms are difficult to cultivate (4)
and suitable alternative morphological or physiological met-
rics are lacking. Ribosomal RNA (rRNA) and genome-
encoded rRNA genes are suitable informational molecules
for phylogenetic analyses (5) because they (a) are present in
every known organism, (b) contain consecutive highly con-
served bases critical for ribosome functionality that can be
used for oligonucleotide primer sites, (c) contain highly var-
iable regions that provide lineage-specific phylogenetic infor-
mation, and (d) are converted directly to ribosomes without
undergoing translation, thereby avoiding codon degeneracy
issues. Despite these obvious benefits, analysis of rRNA genes
for community composition and taxonomy have detrimental
attributes as well, and these have come into clearer focus as
the use of rRNA genes has become more routine. These
attributes include variable gene copy number between bacte-
rial lineages (6); substantial variability of within-organism
rRNA gene sequences (7); inadequate informational content
to assign species and strain-level taxonomy consistently, par-
ticularly when smaller fragments of the full gene are analyzed
(8); and insufficiently long regions that are perfectly con-
served among all members of microbial domains (e.g., Bacte-
ria and Archaea) to allow targeting of organisms from broad
taxonomic groups using a single, nondegenerate primer set.
Finally, microbial taxonomy does not consistently provide
an indication of the corresponding physiology or genetic
potential. In part to address some of these limitations of
rRNA genes, a host of microbial functional genes have
been targeted for microbial community analysis of specific
functional groups, including (but by no means limited to)
sulfate-reducing microorganisms (dissimilatory sulfite reduc-
tase, dsrAB; 9), methanogens (methyl coenzyme reductase
A, mcrA, 10), methanotrophs (e.g., particulate methane
monooxygenase gene, pmoA; 11), ammonia-oxidizing bacte-
ria (ammonia monooxygenase subunit A, amoA; 12). As a
result, microbial ecologists can use PCR-based approaches
doi:10.1128/9781555818821.ch2.4.2
2.4.2-1
to target multiple genes depending on organisms and func-
tions of interest, which complement many other tools and
biomarkers available for microbial community exploration.
Increasingly, shotgun sequencing of genomic DNA
(gDNA) is performed to characterize microbial communities
(“metagenome” sequencing), and the utility of this method
has been well validated in many studies. However, amplicon
sequencing can still serve an important role in microbial
community characterization due to its lower overall cost
and the production of robust data sets that are highly amena-
ble to replication and statistical testing. Also, because many
organisms and functions of interest will be present at lower
relative abundance within a microbial community, targeted
approaches that leverage PCR may be more suitable than
bulk DNA sequencing for studying these microorganisms.
Furthermore, recent bioinformatics innovations have pro-
vided new mechanisms to interpret community structure
and function on the basis of rRNA genes alone (13, 14).
This approach has been most successful in human-associated
microbiome studies and currently is unlikely to be suitable for
more diverse ecosystems such as soil, where greater microbial
diversity, fewer relevant reference genomes, and closely
related taxa with differing functional capabilities, lead to
less robust predictive capabilities. Nonetheless, we anticipate
that amplicon sequencing, which requires a priori knowledge
of target gene sequences, will continue to be performed by
itself and in tandem with shotgun metagenome sequencing.
Shotgun metagenome sequencing, which does not require a
priori knowledge of targets, can provide novel data for discov-
ery and for informing future PCR-based studies.
In this chapter we specifically address the use of PCR
amplification coupled with high-throughput sequencing
(HTS) for the analysis of microbial community composition
and structure and for subsequent statistical analyses of these
communities. Although the field of microbial amplicon
sequencing is quite well established, changes in next-
generation sequencing (NGS) platforms have brought about
dramatic changes in the manner in which sequencing is con-
ducted and how data are analyzed. We present an overview of
the overall trends in the field as well as information to help
the beginning and moderately experienced user conduct their
own PCR-based HTS studies. The field of HTS is highly vol-
atile and specific details—such as sequence read length, total
read number, and total data output—change rapidly. The
reader is warned! There are many other excellent reviews of
2.4.2-2 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
this subject, and we suggest that any user examine multiple
sources of information before deciding on any single workflow
(e.g., 15–20).
WHY HTS?
For decades, microbial ecologists relied on clone libraries to
generate sequence data from pools of PCR-amplified genes.
However, the number of sequences generated in a study was
often dictated by cost and labor limitations rather than opti-
mal experimental design. These limitations often precluded
replication, leading to a general lack of statistical rigor in
the field (21). The development of new NGS technologies
around 2005 helped “democratize” sequence-based analyses
by allowing the generation of large data sets from multiple
samples at increasingly affordable costs. The Roche 454 pyro-
sequencing method dominated microbial amplicon sequenc-
ing field for several years; the read length of the individual
sequences from the pyrosequencer outpaced those of compet-
itors. More recently, other sequencing platforms such as the
Illumina MiSeq and the Ion Torrent Personal Genome
Machine (PGM) produce equivalent data, with somewhat
shorter read lengths, but with greater numbers of sequences
and substantially lower costs (e.g., 18, 22, 23). In 2014, the
Illumina platform outperformed other sequencing platforms
in number of reads, read length, error rates, cost per base,
and ease of use. In particular, the Illumina MiSeq instrument
has become the default sequencer for amplicon sequencing.
Nonetheless, the Roche 454 pyrosequencer and the Ion Tor-
rent PGM are still viable for HTS of PCR amplicons, and this
chapter is applicable to sequence data collected independent
of the specific sequencing platform.
Regardless of sequencing platform, HTS of amplicons has
reduced costs and labor involved in microbial community
profiling. This is because of two primary developments: clon-
ing of PCR amplicons is no longer necessary for sequencing of
complex pools of amplicons, and miniaturization and auto-
mation of the sequencing process. The cloning step was
labor-intensive and expensive, and introduced biases into
the data through negative interactions of the host with the
cloned DNA sequences. Miniaturization and automation
has meant that much less labor is currently involved in
sequencing 10 million amplicons on an Illumina MiSeq
than 100 amplicons on a capillary electrophoresis sequencer.
As a result, the research bottleneck has gone from data acquis-
ition to data analysis.
STEPS IN THE NGS AMPLICON WORKFLOW
There are multiple steps in an HTS workflow, and this chap-
ter is not intended to cover all aspects of microbial ecology
sequencing comprehensively. However, we discuss many
considerations of HTS workflows, including experimental
or sampling design (including methodological approach
and choice of sequencing platform), sample acquisition and
storage, nucleic acid extraction and storage, nucleic acid qual-
ity control, PCR reagent and primer set selection, PCR
amplification for HTS (one-step, two-step, or three-step),
sample-specific barcoding, and amplicon quality control.
With respect to sequencing and analysis, we explore pooling,
purification and quality control of samples for single HTS
runs, sequencing and quality control of raw data output, stor-
age and sharing of raw sequence data, processing of raw data,
secondary bioinformatics, statistical tests, and data visualiza-
tion. Because a detailed review of all these aspects is beyond
the scope of this chapter, we consider these aspects in general
and provide suitable references wherever recent detailed
reviews are available. We provide the user with the know-how
to generate and analyze HTS amplicon sequence data and
highlight additional resources that can assist with the process.
Experimental or Sampling Design
Experimental design is the most important consideration for
any experimental work in microbial ecology. We recommend
that any experimental design include suitable replication of
sampling and technical manipulations to enable the use of
appropriate statistical analyses to identify significant trends
and avoid generating solely descriptive and unreproducible
data sets. For this reason, we recommend that the reader con-
sult a recent publication by Goodrich et al. (15) in addition to
a commentary on the pitfalls of unreplicated analyses in
microbial ecology (21). Due to the decreased cost of amplicon
sequencing, fully replicated experimental analyses are now
essential and affordable. In general, at least three replicates
must be employed in experimental studies, and five or more
replicates are recommended whenever possible.
Nucleic Acid Extraction
DNA extraction is a critical component of HTS-based anal-
yses of microbial communities (e.g., 24) and this process is
now largely performed using off-the-shelf extraction kits.
Many options for nucleic acid extraction are available
from established companies, such as Mo Bio Laboratories
(www.mobio.com), Zymo (www.zymoresearch.com), Qiagen
(www.qiagen.com), MP Biomedicals (www.mpbio.com),
Promega (www.promega.com), and others. A wide variety
of kits are available for extraction of soil, feces, water (filters),
blood, microbial mats, or biofilms, and most involve physical
shearing as well as chemical lysis of cells. Thus, gDNA from a
wide variety of cells, including Gram-positive bacteria, can
be recovered quantitatively. However, gDNA recovered
by physical shearing is generally lower molecular weight
(10–40 kb) relative to nucleic acid recovered by procedures
that use only chemical lysis. For PCR-based sequencing
applications, low molecular weight fragments are suitable.
Higher molecular weight fragments are needed for some
applications, such as functional metagenomics screening
(see chapter 2.4.3), optical mapping (www.opgen.com),
and de novo genome sequencing with mate-pair or Pacific Bio-
systems platform (www.pacificbiosciences.com). For such
applications, alternate extraction methods are needed, but
these methods may not be suitable for recovery of gDNA
from organisms in complex environments such as soil. In
addition, in microbiome studies associated with eukaryotic
hosts, gDNA yields can frequently be dominated by host
DNA; therefore, large yields do not necessarily indicate a large
recovery of microbial DNA.
At the time of nucleic acid extraction, negative controls
should be introduced into the workflow. The presence of
DNA contamination, especially from microbial sources, has
been observed in various laboratory reagents that can affect
the final product (e.g., 25). Most commonly, polymerase
enzymes are contaminated (e.g., 26, 27–31) but many other
sources of contamination are possible, including water, other
reagents, disposables, and general laboratory contamination
(32, 33). The contamination is not solely limited to bacteria;
for example, viral DNA contamination has also been
observed (34). Furthermore, as stated by Salter et al. (35),
“The presence of contaminating DNA is a particular chal-
lenge for researchers working with samples containing a low
microbial biomass.” Thus, in extreme environmental systems
with low total microbial biomass, or in medical samples such
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-3
as serum or skin swabs, extra precautions need to be taken and
negative controls included through all steps of the HTS
workflow.
Although knowledge of reagent and laboratory contami-
nation has been known for decades, NGS technologies
have increased our ability to detect low levels of DNA. In
our experience, negative control PCRs, which may appear
blank by way of agarose gel electrophoresis, still contain
some low level of amplification, which can be detected using
HTS platforms. Therefore, we recommend that a series of
negative controls and background controls be performed at
all stages of sample processing. This includes performing
blank DNA extractions and negative PCR controls, and these
blanks should be propagated throughout the workflow and
sequenced as unique samples. Sequencing facilities should
be made aware of which samples are blanks, so as not to
attempt extreme measures to generate amplicons (e.g., addi-
tional PCR cycles) from these samples. Furthermore, some
tools to identify background contaminants have been devel-
oped (36), and a “low-biomass” contaminant database was
previously established (37). Finally, the reader should be
aware of commercially available polymerases with advertised
low microbial contamination (e.g., MTP Taq DNA Polymer-
ase, Sigma-Aldrich; DFS-Taq DNA polymerase, BIORON
International).
Nucleic Acid Quality Control
Nucleic acid quality can be checked for DNA concentra-
tion, molecular weight, and purity using multiple approaches,
including agarose gel electrophoresis (<1% agarose), auto-
mated electrophoresis (e.g., Agilent TapeStation2200;
Advanced Analytical Fragment Analyzer), and quantifica-
tion with NanoDrop (www.nanodrop.com), PicoGreen (38),
or Qubit (https://www.lifetechnologies.com/qubit.html).
There is minimal interest in performing systematic gDNA
quality analysis because these data do not guarantee PCR fail-
ure or success. Specifically, gDNA quantities below detection
of these instruments can still be successfully amplified by
PCR; conversely, samples in which plenty of gDNA is present
can still produce poor PCR amplicon yields due to the pres-
ence of inhibitors. Although the types and scale of PCR
inhibitors is beyond the scope of this chapter, the presence
of PCR inhibitors can be assessed using standard PCR and
through the use of spike-in tests in which the gDNA of inter-
est is mixed with an inhibitor-free control DNA (39). Aside
from inhibitors, nucleic acid quality tests can be important in
the overall sequencing workflow. Recently, Kennedy et al.
(40) identified input gDNA template concentration as a
major variable impacting the observed microbial profile using
a PCR-based HTS workflow. Therefore, accurate quantifica-
tion of DNA and validation that the DNA sample is free of
inhibitors, is important in the overall workflow to reduce
workflow variability in sample analysis. Concentration meas-
urements based on double-stranded DNA-binding dyes such
as PicoGreen (e.g., using the Qubit) should be employed
(41). Although NanoDrop readings (i.e., 260/280 and 260/
230 nm absorbance ratios) are often used to determine
DNA purity, the NanoDrop is unreliable at low DNA con-
centrations (<10 ng/µl) and all reported values (including
ratios) should not be used to make any specific recommenda-
tion for quality or quantity of DNA. Typically, 260/280 ratios
should be approximately 1.8 and 260/230 ratios should be
in the range of 1.8–2.2. However, some DNA samples may
amplify suitably with ratios outside of this range, while con-
versely, some inhibitors may not absorb in any of these
tested wavelengths but are still capable of inhibiting PCR.
Ultimately, empirical tests are often needed to determine if
DNA is suitably free of inhibitors for amplification.
PCR Amplification for NGS (One-Step, Two-Step,
or Three-Step)
Several PCR strategies are available for HTS of amplicons.
Unlike more traditional capillary electrophoresis (Sanger)
sequencing, current NGS technologies (with certain excep-
tions; see mention of Pacific Biosciences) tend to produce
fairly short individual sequences (reads). This short output
has constrained sequencing approaches, though there has
been a robust history, predating HTS, of deep sequencing of
short amplicons (e.g., 42). However, the NGS instruments
have developed rapidly, and amplicon lengths of 400–500
bases can be feasibly sequenced in their entirety on multiple
sequencing platforms. PCR amplicons must be prepared for
sequencing on NGS sequencers by the incorporation of for-
ward and reverse sequencing adapters on all fragments. In
addition, if multiple amplicon pools are to be sequenced on
a single sequencer, a sample-specific identifier called a multi-
plex identifier or barcode (typically 6–12 bases long) is incor-
porated as well. There are a number of strategies that can be
employed to incorporate the sequence adapters, including
(a) PCR amplification with template-specific primers,
coupled with a subsequent sequencing adapter ligation step;
(b) PCR amplification with primers containing sequencing
adapters, barcodes, and template-specific primers (e.g., 43,
44); (c) two-step amplification reactions in which tem-
plate-specific primers containing 50 linker sequences are
used to generate amplicons from gDNA and, subsequently,
primers containing sequence adapter sequences, sample-
specific barcodes, and the same linker sequences ( positioned
at the 30 end) are used (45, 46); and (d) a multistage protocol
in which gDNA is amplified with template-specific primers,
the PCR products are fragmented into small pieces, and the
small pieces are prepared for sequencing by sequence adapter
ligation (e.g., 47). It should be noted that each sequencing
instrument has unique sequence adapters, and often barcodes,
but the barcodes can be shared between different instruments.
Furthermore, the sequence adapters are essential for sequenc-
ing because they are used to clonally amplify PCR fragments
for sequencing, and these adapters are also generally used as
common sequences among all fragments from which to ini-
tiate the sequencing reactions.
There are different reasons for using each of the four
approaches. The first approach is rarely used due to the
high per sample costs (money and labor) of ligation. How-
ever, this approach avoids additional issues of PCR bias asso-
ciated with the more complex primer design used in the
second and third approaches. The second approach (Fig. 1)
is favored when a single primer set is used routinely. This
approach requires only a single PCR amplification to generate
amplicons ready for sequencing and is well established in
the field. Such an approach is employed by the Earth Micro-
biome Project (EMP; http://www.earthmicrobiome.org/emp-
standard-protocols/16s/), and multiple protocols have been
published (e.g., 43, 44). For the Ion Torrent platform, this
approach is called the “Fusion” method (e.g., 48). The major
caveat with this method is the cost associated with synthesiz-
ing unique, barcoded primers. In particular, the large primers
required for this approach (generally 60 + bases) may require
additional purification (e.g., high-performance liquid chro-
matography purification), and can be quite costly, even as a
one-time cost. Companies such as Integrated DNA Technol-
ogies have special synthesis and quality control techniques for
2.4.2-4 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
FIGURE 1 PCR and sequencing workflow for a one-step PCR amplification protocol to generate sequence-ready amplicon libraries.
Shown are the Earth Microbiome Project (EMP) primers for amplification and sequencing of the V4 variable region of microbial 16S
rRNA genes on the Illumina sequencing platform. doi:10.1128/9781555818821.ch2.4.2.f1
long oligonucleotides (i.e., ultramer oligonucleotides), but
charge an additional premium for special processing of oligo-
nucleotides specifically synthesized for multiplexing (e.g.,
TruGrade Processing Service for Multiplexed NGS) to avoid
barcode cross-contamination.
The third approach (Fig. 2) is favored when multiple pri-
mer sets are used frequently. In this case, a single set of primers
containing sequencing adapters, barcodes, and linker sequen-
ces are synthesized or purchased (e.g., Fluidigm Access Array
Barcode Library). In addition, template-specific PCR primers
targeting 16S rRNA gene (or others) are synthesized with
linker sequences at the 50 end of each primer. Two PCR stages
are then used to generate sequencer-ready amplicons, includ-
ing a first stage in which template-specific gene amplification
is performed, and a second stage in which barcoded adapter
sequences are introduced into each amplicon pool (Fig. 2).
This approach greatly increases flexibility and limits the
cost associated with designing or utilizing new primers and
can lead to substantially lower library preparation costs.
This approach has been described previously by Bybee et al.
(46) and de Cárcer et al. (45), and is implemented in human
microbiome studies (49). Illumina has developed a spe-
cific workflow based on this targeted amplicon sequencing
or modular tagged high-throughput amplicon sequencing
approach (http://support.illumina.com/content/dam/illumina-
support/documents/documentation/chemistry_documentation/
16s/16s-metagenomic-library-prep-guide-15044223-b.pdf ).
Here, Illumina uses linker sequences matching Nextera XT
Index primers at both ends as a dual barcoding strategy. Inter-
estingly, this approach can be performed in a single reaction
tube by including both internal and external primers at differ-
ing concentrations (e.g., 50). This strategy is employed in other
commercial settings as well. For example, BioO Scientific
produces an amplicon sequencing kit to generate amplicon
libraries spanning the eukaryotic internal transcribed spacer
(ITS) region between the small and large subunit ribosomal
RNA genes (NEXTflex 18S ITS Amplicon-Seq Kit).
The fourth approach is favored when large amplicons are
to be sequenced. Such an approach should be taken when no
other suitable primers generating smaller amplicons are avail-
able. This approach may require postsequencing de novo
assembly, depending on the bioinformatic application.
Such de novo assembly can be confounded by sequences
with highly similar regions, such as those found in conserved
regions of the 16S rRNA gene. However, any amplicon can
be sequenced on any NGS sequencer using this approach,
regardless of read length, by shearing the total amplicon to
the appropriate size for the sequencer of interest. Recently,
Burke and Darling (51) demonstrated how a modification
of this approach can still yield full length amplicon sequences
without confounded assembly through multiple sequencing
efforts.
Amplicon Quality Control
PCR amplification yield should be verified using agarose gel
electrophoresis or a similar strategy to verify robust amplifica-
tion, amplicons of the expected size, and the absence of
amplification in negative (water blank) control reactions.
The presence of nonspecific amplicons, both larger and
smaller, can create difficulties, but smaller nontarget ampli-
cons are particularly troublesome due to preferential amplifi-
cation and sequencing on NGS platforms. The presence of
negative control amplification is a frequent occurrence, and
a separate sequencing barcode should be used for such reac-
tions. Even in the absence of a visible negative control ampli-
con using agarose gel electrophoresis, the negative control
sample should be included in the final sequencing pool to
detect background contaminants. In our experience, the out-
put (in terms of number of sequences) is generally two orders
of magnitude lower for negative controls than for standard
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-5

FIGURE 2 PCR and sequencing workflow for a two-step PCR amplification protocol to generate sequence-ready amplicon libraries.
Shown are the Earth Microbiome Project (EMP) primers with Fluidigm common sequence linkers for amplification and sequencing of the
V4 variable region of microbial 16S rRNA genes on the Illumina sequencing platform. Two separate PCR stages are required to generate
template-specific amplicons, and to attach sample-specific barcodes and sequencing adapters. doi:10.1128/9781555818821.ch2.4.2.f2

samples, even when no band is visible. This is a function of
the low sensitivity of agarose gel electrophoresis, the high sen-
sitivity of NGS amplicon sequencing, and the widespread
presence of microbial contaminants in laboratories and in
general reagents. Low-level contamination, detectable by
NGS sequencing of negative controls, is a highly regrettable
but consistent feature of microbial rRNA gene amplification
processes and does not invalidate data from other samples.
However, extraordinary caution should be taken in systems
with extremely low biomass, as the background contamina-
tion can exceed the signal from the sample itself. In addition,
positive control reactions, of either mock community DNA
or of a complex microbial community, should be included
in each reaction setup to assess PCR and sequencing
workflows.
Primer Set Selection
The most important decision in the NGS amplicon sequenc-
ing workflow is that of primer set. There is a staggering num-
ber of primers and primer sets for the microbial small subunit
(SSU) ribosomal RNA gene, and each primer and primer set
has its own limitations (e.g., 52–54) (Table 1). These primers
have different target ranges and target different variable
regions of the gene; such differences can impact the observed
 2.4.2-6 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
TABLE 1 Representative oligonucleotide sequences
Primer Details Target Primer Name Sequences (50 -30 )a References
Standard Primer, Earth Bacteria + Archaea, 16S, V4 515F GTGCCAGCMGCCGCGGTAA 43
Microbiome Project
Standard Primer, Earth Bacteria + Archaea, 16S, V4 806R GGACTACHVGGGTWTCTAAT 43
Microbiome Project
Standard Primer, Human Bacteria, 16S, V1-3 8F AGAGTTTGATCCTGGCTCAG 57b
Microbiome Project
Standard Primer, Human Bacteria, 16S, V1-3 534R ATTACCGCGGCTGCTGG 58
Microbiome Project
Standard Primer, Human Bacteria, 16S, V3-5 357F CCTACGGGAGGCAGCAG 58
Microbiome Project
Standard Primer, Human Bacteria, 16S, V3-5 926R CCGTCAATTCMTTTRAGT 58
Microbiome Project
Standard Primer Bacteria, 16S, V3-4 341F CCTACGGGAGGCAGCAG 58
Standard Primer Bacteria, 16S, V3-4 806R GGACTACHVGGGTWTCTAAT 43
Standard Primer Fungi, 18S FF390 CGATAACGAACGAGACCT 59
Standard Primer Fungi, 18S FR1 AICCATTCAATCGGTAIT 59
Standard Primer Archaea, 16S Ar915aF AGGAATTGGCGGGGGAGCAC 60, 61
Standard Primer Archaea, 16S Ar1386R GCGGTGTGTGCAAGGAGC 61, 62
Standard Primer Fungi, ITS, ITS1 ITS1F CTTGGTCATTTAGAGGAAGTAA 63
Standard Primer Fungi, ITS, ITS1 ITS2R GCTGCGTTCTTCATCGATGC 64
Fluidigm Linker and Bacteria + Archaea, 16S, V4 CS1_515F ACACTGACGACATGGTTCTACAGTGCCAGCM 43; Fluidigm
Standard Primer GCCGCGGTAA
Fluidigm Linker and Bacteria + Archaea, 16S, V4 CS2_806R TACGGTAGCAGAGACTTGGTCTGGACTACHV 43; Fluidigm
Standard Primer GGGTWTCTAAT
Single Step NGS Amplicon Bacteria, 16S, V3-4 Illumina_V3_F AATGATACGGCGACCACCGAGATCTACACTCTTT 44
Primer CCCTACACGACGCTCTTCCGATCTNNNNCCTACGGGAGGCAGCAG
Single Step NGS Amplicon Bacteria, 16S, V3-4 Illumina_V4_R CAAGCAGAAGACGGCATACGAGATXXXXXXGT 44
Primer GACTGGAGTTCAGACGTGTGCTCTTCCGATCTGGA
CTACHVGGGTWTCTAAT
Single Step NGS Amplicon Bacteria + Archaea, 16S, V4 Illumina_EMP_515F AATGATACGGCGACCACCGAGATCTACACTATGG 43
Primer TAATTGTGTGCCAGCMGCCGCGGTAA
Single Step NGS Amplicon Bacteria + Archaea, 16S, V4 Illumina_EMP_806R CAAGCAGAAGACGGCATACGAGATXXXXXXXXXXXX 43
Primer AGTCAGTCAGCCGGACTACHVGGGTWTCTAAT
Single Step NGS Amplicon Eukaryotes, 18S, V9 Illumina_Euk_1391F AATGATACGGCGACCACCGAGATCTACACTATCGC 65
Primer CGTTCGGTACACACCGCCCGTC
Single Step NGS Amplicon Eukaryotes, 18S, V9 Illumina_EukBr CAAGCAGAAGACGGCATACGAGATXXXXXXXXXXXX 65
Primer AGTCAGTCAGCATGAT
CCTTCTGCAGGTTCACCTAC
Single Step NGS Amplicon
Primer
Single Step NGS Amplicon
Primer
Sequencing Primer for EMP
Bacterial Primer Set
Sequencing Primer for EMP
Bacterial Primer Set
Sequencing Primer for EMP
Bacterial Primer Set
Sequencing Primer for EMP
Eukaryote Primer Set
Sequencing Primer for EMP
Eukaryote Primer Set
Sequencing Primer for EMP
Eukaryote Primer Set
Sequencing Primer for
Fungal TIS Primer Set
ITS1F/ITS2
Sequencing Primer for
Fungal TIS Primer Set
ITS1F/ITS2
Sequencing Primer for
Fungal TIS Primer Set
ITS1F/ITS2
Fluidigm Sequencing Primer
Fluidigm Sequencing Primer
Fluidigm Sequencing Primer
a
Underscored sequences represent the common sequence linkers. M = A or C; H = A or C or T; V = A or C or G; W = A or T; N = any base; I = inosine substitution; + = locked nucleic acid substitution. [XXXXX] =
Barcode unique to each sample.
b
See variants suggested by (191).
Fungi, ITS, ITS1 Illumina_Fun_ITS1F
Fungi, ITS, ITS1 Illumina_Fun_ITS2R
Bac_Illumina, Read 1 Primer
Bac_Illumina, Read 2 Primer
Bac_Illumina, Index Primer
Euk_Illumina, Read 1
Primer
Euk_Illumina, Read 2
Primer
Euk_Illumina, Index Primer
ITS_Illumina, Read 1
Primer
ITS_Illumina, Read 2
Primer
ITS_Illumina, Index Primer
Fluidigm_CS1, Read 1
Primer
Fluidigm_CS2, Read 2
Primer
Fluidigm_CS2rc, Index
Primer
AATGATACGGCGACCACCGAGATCTACACGGC 66
TTGGTCATTTAGAGGAAGTAA
CAAGCAGAAGACGGCATACGAGATXXXXXXXXXXXXCGGC 66
TGCGTTCTTCATCGATGC
TATGGTAATTGTGTGCCAGCMGCCGCGGTAA 43
2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-7
AGTCAGTCAGCCGGACTACHVGGGTWTCTAAT 43
ATTAGAWACCCBDGTAGTCCGGCTGACTGACT 43
TATCGCCGTTCGGTACACACCGCCCGTC 43
AGTCAGTCAGCATGATCCTTCTGCAGGTTCACCTAC 43
GTAGGTGAACCTGCAGAAGGATCATGCTGACTGACT 43
TTGGTCATTTAGAGGAAGTAAAAGTCGTAACAAGGTTTCC 66
CGTTCTTCATCGATGCVAGARCCAAGAGATC 66
TCTCGCATCGATGAAGAACGCAGCCG 66
A + CA + CTG + ACGACATGGTTCTACA Fluidigm
T + AC + GGT + AGCAGAGACTTGGTCT Fluidigm
A + GAC + CA + AGTCTCTGCTACCGTA Fluidigm
2.4.2-8 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
microbial community diversity (e.g., 55, 56). Importantly,
the amplicons are of different lengths as well, and the choice
of sequencing platform may constrain the choice of primer
set. Depending on the sequencing strategy employed (see
below), different length amplicons can be tolerated.
Before choosing a primer set and sequencing strategy for a
PCR-based microbial community study, several questions
should be answered. What is the objective of the study
(e.g., diversity survey; identifying specific/novel taxa; measur-
ing shifts in community structure between sample groups;
semi-quantitative analysis of sample community structure)?
Which organisms are the focus of the study (e.g., all bacteria,
all bacteria and archaea, fungi, all eukaryotes, specific taxo-
nomic lineage, specific functional group)? What gene or
genetic region is to be targeted? What is the amplicon length?
In general, most microbial community studies aim for a broad
analysis, typically targeting the entire bacterial domain using
universal primers. Identification at species or strain levels is
less frequent in such studies, and this is in part due to the rel-
atively small fragments that can be sequenced using NGS
instruments. For fungi, the 18S rRNA (SSU rRNA) can
also be readily targeted (59), but the internal transcribed
spacer (ITS) region between the small and large subunit
rRNA genes is also frequently targeted (63, 64, 66). The
amplification and sequencing of the full ITS region (includ-
ing ITS1, 5.8S rRNA, and ITS2) can be problematic for NGS
approaches due to the size variation in the ITS region.
Smaller amplicons are favored at all stages of PCR amplifica-
tion and sequencing. Therefore, organisms with larger ITS
regions are likely to be underestimated in final analyses.
Nonhost (e.g., human/mouse/rat) eukaryotes can also be
characterized using PCR-based HTS approaches (e.g., 65;
http://www.earthmicrobiome.org/emp-standard-protocols/
18s/). However, in host-associated studies (e.g., microeu-
karyotes in mammalian feces), primers can also anneal to
host gDNA and dominate much or most of the sequencing
capacity. This situation is not unique to mammalian hosts,
but also frequently occurs in plant-microbe systems as well.
Multiple strategies have been employed to avoid amplifica-
tion of unwanted templates, such as those from organelles,
including the design and use of primer sets containing mis-
matches with the unwanted templates (e.g., 67), selective
restriction digestion of unwanted templates (e.g., 68),
and the use of blocking primers that prevent the annealing
of the standard primers but are modified to prevent poly-
merase elongation (e.g., 69, 70). If the burden of the
unwanted template is not too high (e.g., less than 50% of
output), these modified PCR approaches are probably not
essential, and the problem can be overcome through a
deeper sequencing effort. We have observed that in some
cases, PCR amplifications that include blocking primers
can produce very weak amplification, and this may be par-
ticularly acute in systems where the unwanted host DNA is
extremely high.
For ribosomal RNA gene amplicon sequencing, the final
critical consideration (aside from amplicon length) is that
of which variable region to sequence. As read length
increases, this consideration will become less relevant as
nearly the entire SSU rRNA gene will be sequenced in
each read or paired reads (e.g., Pacific Biosciences; see below).
For the moment, the importance of the targeted region is still
high. For broad surveys of microbial communities at the taxo-
nomic level of genus or higher, short amplicons can be
favored. This is because short reads are less amenable to chi-
mera formation during PCR, are more likely to be sequenced
completely regardless of platform, and generate smaller files
and have a smaller data footprint. None of these factors
should prevent users from migrating to longer amplicons;
however, one should anticipate that the relative abundance
of detected chimeras will increase. However, the commonly
used EMP primer set 515F/806R is excellent for broad-
spectrum amplification of most bacteria and archaea. After
removal of PCR primer sequences, sequenced amplicons are
approximately 250 bases in length, and the taxonomic resolu-
tion is largely limited to the genus level and higher. Two cav-
eats should be considered here. First, if analysis of the Archaea
is desired, it is generally advisable to target this domain sepa-
rately with an independent primer set that enriches for most
archaea. In our experience, relatively few microbial commun-
ities are dominated by archaea, and archaeal sequences gener-
ally represent less than 1% of all sequences recovered from
many sequencing efforts. Second, the choice of variable
region can be highly lineage-specific. Some closely related
bacteria may be easily distinguished based on one variable
region, but indistinguishable in another (Fig. 3). Thus, in sil-
ico bioinformatic analyses should be performed to aid in the
determination of the most appropriate primer set for any
given study. The Human Microbiome Project (HMP; 71–
73) uses primers targeting the V1–V3 and V3–V5 variable
regions of the bacterial 16S rRNA gene, and these regions
are frequently more sensitive for genus- and species-level
taxonomy (Fig. 3). Although the resulting amplicon frag-
ment length is relatively long, it can now be sequenced on
both Illumina MiSeq and Roche 454 pyrosequencing plat-
forms. Similarly, another primer set (341F/806R; Table 1)
generating a fragment covering the V3–V4 regions was shown
to be effective for microbial community analysis, with
sequencing performed on an Illumina MiSeq (44).
What Sequencing Platform and Workflow
Should Be Employed?
Ion Torrent, Illumina, and Roche 454 pyrosequencing plat-
forms each accommodate different amplicon lengths. The
Ion Torrent PGM platform is the most restrictive and ampli-
cons greater than 400 bp are generally not recommended,
unless shearing is employed. However, new developments
on the Ion Torrent PGM platform, particularly the introduc-
tion of the Hi-Q sequencing chemistry, may allow for longer
reads on the PGM. Similarly, Ion Torrent is said to be devel-
oping an emulsion PCR-independent sequencing technology
(Ion Isothermal Amplification Chemistry), but this is not yet
available. Such a development will likely allow for longer
fragments to be sequenced, in addition to longer reads. Illu-
mina, conversely, has the shortest single read lengths (up to
300 bases), but paired-end sequencing can be used to
sequence fragments of approximately 500 bases (40, 44). Fur-
thermore, fragments up to approximately 800 bp, and possi-
bly larger, can be handled by the MiSeq sequencer, though
the middle of the amplicon will not be sequenced. Although
sequence quality decreases with the fragment length, this is
partially mitigated by assembly when reads overlap. As with
the Ion Torrent, large fragments can be sequenced using a
three-step protocol, but this limits the types of analyses that
can be performed and increases the complexity of subsequent
bioinformatics analyses. Roche 454 pyrosequencing can
accommodate large fragments in excess of 1 kb, and read
lengths up to 1 kb have been observed, though usually reads
from amplicons are shorter than those for shotgun genomic
sequencing. We recommend that for routine analyses, users
only employ reactions in which the entire region between
the forward and reverse primers can be sequenced, either in
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-9

FIGURE 3 Comparison of 16S rRNA gene of Staphylococcus aureus (top strand; accession number L37597) and S. epidermidis strain
DSM20044 T (bottom strand; accession number LN681574). A basic local alignment search tool (BLAST) analysis was performed to com-
pare the two gene sequences. Mismatches are highlighted in red. The region of comparison begins at the 50 end of the molecule, directly after the
commonly used domain-level primer 27F, and ends just before the commonly used domain-level primer 1492R. Common primer sequences,
highlighted in green, include 341F, 534R, 806R, 907R, 1114R, and 1392R (from left to right). Approximate locations of the nine microbial
variable regions are indicated. Note that no differences exist between the two sequences in the V4 region, commonly targeted by the 515F/806R
primer set. doi:10.1128/9781555818821.ch2.4.2.f3

a single read or through the assembly of paired-end reads.
When this is not possible, the standard approach is to employ
only a single read. Typically, this is the forward read on Illu-
mina sequencers, due to higher sequence quality in the for-
ward read. However, this can increase the complexity of the
bioinformatics analyses because reads will likely be of differ-
ent lengths and may represent slightly different regions of
the targeted gene. For example, some regions of the 16S
rRNA gene have size variability, and thus 200 bases of one
organism may not represent the exact same portion of the
ribosome as 200 bases in another organism. This may affect
the analysis of operational taxonomic units (OTUs) by con-
tributing to the production of spurious OTUs. For some
amplicons, there may be no suitable primers generating
appropriate-length amplicons, and such single-read analyses
can still be viable. Finally, we note that sequencing of full-
length or near-full-length amplicons may be viable using
the Pacific Biosciences single molecule, real-time sequencing
platform (74, 75), but this platform has not been used widely
due to high error rate and sequencing cost.
Barcode Design
A number of barcode strategies have been developed, includ-
ing single-index and dual-index strategies. In single-index, a
single unique identifier sequence, generally 6–12 nucleotides
in length, is added to the amplicon. Dual-index sequencing
utilizes two such identifiers, thereby reducing the number of
primers to synthesize through combinatorial use of each index
pool. Dual indexing (e.g., 76) is easier to implement on the
Illumina sequencers because unique sequencing reads of the
exact length of the barcode are used to generate indices.
Ion Torrent and Roche 454 pyrosequencing platforms do
not perform paired-end sequencing, and dual barcoding,
although possible, is not recommended for these platforms.
Many different barcode sets are available, both from sequenc-
ing companies themselves, but have also been designed else-
where. A set of 2,168 barcodes are provided in a seminal
publication (Supplemental File 1 within reference 18); these
barcodes are so-called error-correcting Golay codes, and can-
not be easily mistaken for each other, even in the presence of
sequencing errors. For Illumina sequencers, barcode length is
not particularly relevant because separate sequencing reads
are performed to identify the barcode. For Roche 454 and
Ion Torrent sequencing platforms, the barcodes “consume”
part of the sequence length and are typically located at the
beginning of the sequencing reaction.
Sources of Sequencing Error
Although technically straightforward, sequencing of PCR
amplicons can produce data containing many artifacts. The
most common of these artifacts include (a) background con-
tamination from reagents and laboratory (discussed already);
(b) errors introduced into amplicons due to PCR error;
(c) chimera formation during PCR; and (d) sequencing error
introduced by the sequencing instrument itself (e.g., 77).
In general, it is preferable to use proofreading polymerases
during PCR amplification. However, these enzymes are
generally more sensitive to PCR inhibitors than standard
2.4.2-10 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
Taq polymerases, or enhanced “robust” polymerases suitable
for difficult samples (e.g., KAPA2G Robust DNA poly-
merase, KAPA Biosystems; Herculase II Fusion DNA poly-
merase, Agilent). Coupled with both nonsystematic and
systematic (e.g., homopolymer errors on Ion Torrent and
Roche 454 sequencers) instrument errors, a number of
sequence errors accumulate in the final data output. For appli-
cations at the OTU level (described in more detail later),
where discriminating between closely related taxa is critical,
or when alpha diversity indexes (such as taxon richness
and estimated richness, e.g., Chao1 indexes) are calculated,
such errors can produce spurious results. Conceptually, OTU-
based dissimilarity analyses are more sensitive to sequencing
error than measuring sample dissimilarities based on phyloge-
netic distances (e.g., Unifrac; 78). This is because many
OTUs in the final data set will be spurious, a result of sequenc-
ing errors (79). Soon after HTS analyses were introduced,
estimates of microbial diversity in complex environments
were revised downward or disregarded because of the inflated
nature of amplicon sequence diversity (80, 81). For this
reason, in data analysis strategies that align OTU sequences
into a phylogenetic tree or place sequences into establi-
shed taxonomic groups (e.g., phylum to genus), these errors
contribute little to the overall distortion of microbial
community data for sample comparison. Novel strategies to
reduce sequencing error have been developed, though such
strategies have a more complex library preparation strategy,
higher sequencing demand, and higher bioinformatics
demand (82, 83).
Setting Up a PCR
Setting up amplification reactions for NGS is fundamentally
no different than other PCRs. Among the most important
factors are master mix/enzyme selection, primer and primer
concentration, input gDNA concentration, reaction volume
and technical replicates, and the number of PCR cycles.
These are discussed in more detail below.
Enzyme/Master Mix Selection
Amplification reactions for NGS are similar to other PCR,
and all appropriate precautions for PCR amplifying microbial
genes (especially rRNA genes) must be taken. We recom-
mend premade master mixes that incorporate buffer, polymer-
ase, dNTPs, and MgCl2 together. The user adds gDNA (or
cDNA), primers, and water to a final volume. Such reagents
mixes reduce the burden of setting up PCR amplifications but
are also susceptible to contamination. Reagents should be
prepared in aliquots under aseptic conditions so as to limit
reagent contamination. Best practices dictate that reagents
for microbial rRNA gene amplification should be prepared
in single-usage aliquots whenever possible. Many polymer-
ases and master mixes are available, and there is no dominant
enzyme used in the field. As discussed previously, proofread-
ing enzymes are preferred, but these enzymes do not always
produce robust yield in reactions with environmental
gDNA. Some additives to PCR can be used to overcome
inhibition (e.g., 84–86). In other cases, modified or engi-
neered enzymes can be used to overcome inhibition (e.g., 87).
Primer and Primer Concentration
Short primers (<50 bases) do not have to be purified other
than by standard desalting. When synthesizing long primers
containing barcodes and sequencing adapters, higher purifi-
cation (such as HPLC purification) is often used but may
not be necessary if ordered as ultramers (Integrated DNA
Technologies), for example. Additional levels of purity are
available for barcoded primers to ensure the absence of cross-
over contamination. Primer concentrations are not different
than standard PCR, and users should employ concentrations
described in reference manuscripts. High primer concentra-
tion can, under some conditions, create nonspecific amplifi-
cation. In silico prediction of dimer formation through
self-dimerization or heterodimerization of primers can be
performed using software programs such as OligoAnalyzer
(88). Nonspecific amplification and generation of dimers is
often highly deleterious to the overall sequencing effort,
as these fragments will generate sequencer-ready amplicons
that “consume” sequencing capacity. Often, high primer
concentrations will lead to strong primer dimer formation;
such small fragments can be removed through size-specific
purification of the band of interest. Commonly used puri-
fication reagents such as AMPure XP reagents (Beckman
Coulter) can be adjusted to remove fragments smaller than
200 or 300 bp. Therefore, there is an advantage in generating
PCR amplicons of at least 250 bp (including adapters, barco-
des, linkers and the fragment of interest), so as to allow for
purification reactions to remove primer dimers that can be as
large as 150 bp. Other techniques—such as gel purification
or automated size selection (e.g., Pippin Prep, Sage Science,
Beverly, MA; E-Gel SizeSelect Gels, Life Technologies,
Grand Island, NY)—can be used, but these are more laborious
and expensive.
DNA Concentration
DNA concentration is not well controlled in PCR amplifica-
tions. In general, users add a set volume of gDNA to a PCR
(e.g., 1 µl gDNA into a 20 µl reaction). In systems where
DNA recovery is fairly consistent, this is unlikely to be a
major concern. However, other systems (e.g., across an envi-
ronmental gradient), large differences in DNA recovery can
be expected. Recently, Kennedy et al. (40) examined the
role of various factors in PCR-NGS yield and found input
gDNA concentration to be the dominant factor affecting
observed bacterial 16S rRNA gene amplicon sequence pro-
files. Other tested factors, including pooling of PCR ampli-
cons, sample preparation, and interlane variability were not
significant in contributing to variability (40). Therefore, if
possible, gDNA concentrations should be equalized before
loading into a PCR. However, this is also not problem-free;
if there are PCR inhibitors in samples, different dilutions
may differently impact amplification.
Reaction Volume and Technical Replicates
No systematic study of reaction volumes has been conducted
for PCR amplification. In general, 10–20 µl reaction volumes
are used. The total reaction volume does not need to be large
for NGS. This is in part due to the high level of multiplexing
for NGS runs (usually >48 samples simultaneously); final
libraries from each sample therefore only need a small amount
of PCR product. The critical factor is the sum total DNA
concentration before sequencing. Such a consideration may
ultimately allow NGS to be conducted on PCR amplicons
generated with many fewer PCR cycles (see below). This
may reduce PCR artifacts such as chimeras and heterodu-
plexes (89).
No specific recommendations have been established for
technical replication of NGS amplicon sequencing. In
some cases, independent PCR amplification reactions are
performed on the same sample and then pooled. Several stud-
ies have suggested that pooling replicates can decrease varia-
tion during amplification, and this has particular relevance
for low-abundance templates (90, 91). Overall, a wide range
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-11
of independent technical reactions have been used in the lit-
erature, ranging from a single replicate (many studies) to as
many as 10 (e.g., 66, 92, 93). Smith and Peay (66) observed
that increasing the number of PCR replicates did not alter
observed ecological measures, whereas sequencing depth
improved beta-diversity analyses. Nonetheless, some reports
in the literature have identified significant variation at the
OTU level in technical replicates (94). Within-sample varia-
tion can be measured through the generation of independent
technical replicates using separate barcodes for each sam-
ple. However, this is not typically done due to the extra
cost on the library preparation, increased bioinformatics,
and decreased yield per amplicon pool when run on a
sequencer with the same total output. If amplicon sequencing
shifts to larger output sequencers (e.g., HiSeq2500), the total
output may be so high that the average user can easily afford to
run technical replicates and still produce deep coverage of
microbial communities. Zhou et al. (94) suggest that high
observed variability in technical replicates is likely due to
sampling artifacts associated with random sampling that
occurs in PCR-NGS protocols in which microbial diversity
greatly exceeds the sequencing output. Here, Zhou et al.
(94) recommend that at least three technical replicates
should be run for each sample. For analyses performed at
higher taxonomic levels (e.g., family level), technical varia-
bility within samples is unlikely to play a major role; however,
this has not been explicitly tested. Other studies have sug-
gested that technical reproducibility for low-abundance
taxa is particularly suspect, and that only taxa with a set mini-
mum number of recovered sequences should be examined
(66, 95). Thus, while technical replication may or may not
improve data, the consensus in the field appears to be that
deeper sequencing of each sample is important.
Number of PCR Cycles
In general, it is advantageous to use the fewest number of PCR
cycles possible. Fewer PCR cycles have been shown to reduce
PCR bias (91). Running fewer PCR cycles decreases the yield
from negative control amplification reactions. In addition,
the rate of chimera formation and other PCR artifacts (89,
96) has been shown to increase with increasing PCR cycles.
We have observed this in two-stage PCR amplifications in
which the number of first-stage and second-stage cycles
were systematically varied (Fig. 4). We observed a strong cor-
relation between chimera formation and total number of
cycles, regardless of whether they were primarily in the first
or second stage of amplification. Furthermore, Qui et al.
(89) have suggested that in reactions where one of the
reagents become depleted (e.g., dNTPs), the error rate can
increase, even when a proofreading enzyme is used. When
running relatively few PCR cycles (e.g., 20 cycles or fewer),
detection of amplification by agarose gel electrophoresis can
be difficult. Nonetheless, due to the lower DNA demand
for NGS amplicons, the field of microbial ecology should
aim to reduce PCR cycles to the absolute minimum necessary.
This may necessitate, for example, quantitative testing of
input gDNA prior to amplification for NGS. Ultimately,
separate qPCR runs may be performed before determining
how many amplification cycles should be utilized on a
sample-by-sample basis. This is not done routinely yet.
In a seminal manuscript, Polz and Cavanaugh (91) made
recommendations to reduce PCR bias that are still useful,
regardless of the changes in PCR and sequencing technology.
These recommendations included (a) reduce primer degener-
acy when possible, to reduce possibilities of primer-template
mismatches generating PCR bias; (b) use high-input gDNA
FIGURE 4 Relationship between total PCR cycles and presence
of putative chimeric sequences in the output amplicon sequence
data. A single gDNA sample from mammalian feces (chinchilla)
was PCR amplified using a two-step PCR protocol, utilizing the
EMP primers and Fluidigm common sequence linkers. A systematic
variation of first-stage and second-stage PCR cycles was performed,
with a range of 20–36 cycles total. After sequencing on an Ion
Torrent PGM, sequence data were demultiplexed and chimeras
were identified using usearch61. A linear relationship between total
cycle number and chimera formation (R 2 = 0.65) was observed.
doi:10.1128/9781555818821.ch2.4.2.f4
to reduce variability; (c) perform multiple replicates and
pool; and (d) use few PCR cycles. In that study, as few as
five amplification cycles were deemed suitable, provided
that a high-input template was used. Currently, because hun-
dreds of reactions can be pooled for NGS, such a low cycle
number may also be achievable without the use of hundreds
of nanograms of input gDNA.
Pooling, Purification, and Quality Control of
Samples for Single NGS Runs
A variety of pooling strategies can be employed, but initially,
the user should determine how many reads per sample are
desired. From here, assuming equimolar pooling, a simple cal-
culation of total reads anticipated per run is divided by the
total number of barcodes (samples) used. Due to variation
between different samples, deriving from varying PCR yields,
PCR DNA input into pools, and differential sequencing, a
fair amount of room should be left as a buffer. Many factors
decrease the per sample yield—including PhiX spike-in (Illu-
mina platform only), less than expected run yields, and
uneven coverage. Therefore, fewer samples than the absolute
maximum should be pooled. The number of reads necessary
per sample is variable from project to project and between
sample types. Momozawa et al. (97), for example, suggested
that in studies of colon biopsies and fecal samples, at least
1,000 reads per sample were required, and that below this
level, beta-diversity measurements were highly variable.
Extremely deep sequencing of microbial communities can
be performed relatively inexpensively, as studies with
100,000 s or millions of reads per sample can be found in
the literature (18, 98). Such depth may be relevant for iden-
tifying members of the rare biosphere, but is typically not
2.4.2-12 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
needed for studies in which shifts in microbial community
structure are identified. Statistically significant effects of
treatment or environmental gradients can usually be observed
at relatively shallow sequencing efforts. Pinto and Raskin
(99) observed that greater sequencing depth typically
improved variability between replicates but that did not nec-
essarily improve accuracy of beta-diversity estimates. Thus,
there is no absolute minimum number of reads needed for
any single study, and the objectives of the project will ulti-
mately determine the depth of sequencing necessary. It is rea-
sonable, however, to aim for 20,000–50,000 reads per sample,
with the assumption that under standard amplification
regimes, 20% of the reads will be chimeric and removed
from the analysis. Furthermore, some samples will produce
relatively few reads, and these samples create the greatest trou-
ble for large-scale studies, as various bioinformatics strategies
can be utilized to rigorously control for these low-coverage
samples (see below). Thus, it is often appropriate to overshoot
the number of sequences per sample, to bring up the level of
the lowest coverage samples, even at the expense of wasted
sequencing effort. For Ion Torrent PGM, utilizing the 318
chip, somewhere between 4 and 6 million reads are generally
obtained. Therefore, with appropriate pooling, 96 samples
can generally be sequenced using this platform. For the ampli-
con sequencing on the Illumina MiSeq, utilizing V2 chemis-
try (2 × 250 paired-end reads) can generate 5–10 million
clusters (10–20 million total reads) and is generally suitable
for up to 192 samples. Similarly, the V3 chemistry on the Illu-
mina MiSeq (2 × 300 paired-end reads) can generate 15–25
million clusters and could be suitable for up to 384 samples.
Often, equimolar pooling of so many samples is difficult,
and various pooling strategies can help decrease the sequence
output variability between samples. In sequencing runs where
amplicons of different size are to be used, equimolar ratios of
fragments will generally lead to lower sequencing output of
larger fragments, and greater amounts of large fragments are
typically added to the total pool (61).
Larger reaction volumes may be necessary if high-
throughput purification strategies are used. For example, the
SequalPrep 96-well normalization plate kit (Life Technolo-
gies) provides a convenient mechanism for mixing PCR
amplicons in roughly equimolar ratio. This kit performs
PCR purification; however, each well has a limited capacity
to bind DNA (∼25 ng DNA total). When at least 250 ng
of PCR product are added to each well, the binding capacity
is saturated, and the 25 ng of DNA from each reaction can be
eluted in the same volume. This can greatly decrease the var-
iability between samples in a pooled reaction. Alternatively,
reactions can be pooled in approximate equimolar ratio based
on observation in agarose gel electrophoresis, and then a sin-
gle purification can be performed. This approach is less costly
due to reduced reagent usage. However, severely unbalanced
amplicon pools can occur due to the limited sensitivity of
agarose gel electrophoresis. Care should also be taken when
combining amplicons of different size together; in general,
large amplicons sequence less robustly, and small amplicons
can dominate.
Prior to sequencing, pools of amplicons with sequencing
adapters at each end (libraries) should be purified using col-
umn purification or magnetic bead–based purifications. Mag-
netic bead purifications (e.g., AMPURE XP beads, Beckman
Coulter) may be preferable, due to the ability to use the puri-
fication protocol to avoid recovery of small DNA fragments,
such as primer dimers. Column-based cleanup approaches
are also available for dimer removal (e.g., GeneJET NGS
Cleanup Kit, ThermoScientific; E.Z.N.A. NGS Clean-IT
Kit; Omega bio-tek). The inclusion of primer dimers in a final
library pool is hugely detrimental because these double-
stranded DNA fragments are short and can contain both
forward and reverse sequence adapters. Thus, they can be
amplified preferentially using either emulsion PCR or bridge
amplification, and consume a large portion of the sequencing
output on a NGS sequencer. If the amplicon of interest is rel-
atively large (>200 bp), and the nonspecific DNA of interest
is smaller, AMPure XP beads (and similar “home-brew” alter-
natives), can be used to recover fragments of specified size
(100, 101). All libraries, before sequencing, should be visual-
ized using agarose gel electrophoresis (or other strategy, such
as the Agilent Bioanalyzer) to verify that no spurious bands or
primer dimers are present in the final pool. We have observed
that nonspecific bands that are of minor abundance in any
one sample can aggregate and contribute significantly in a
final pool of hundreds of samples.
Sequencing
Sequencing for amplicons on NGS sequencers, while gener-
ally similar to sequencing any other library type, is often more
troublesome than shotgun sequencing. In general, sequence
output metrics for NGS platforms are lower for amplicon
sequencing—including read length and total yield. The
primary reason for this, specifically on the Roche 454 and
Illumina platforms, is the so-called low nucleotide diversity
of bases in highly conserved regions of gene amplicons (see
the Illumina bulletin “Low-Diversity Sequencing on the Illu-
mina MiSeq Platform,” http://www.illumina.com/documents/
products/technotes/technote_low_diversity_rta.pdf). In con-
served regions, many wells (Roche 454) or clusters (Illumina)
will either produce luminescence at the same time (Roche
454) or have the same color fluorescence signal (Illumina).
Such features can confound the camera, and, for example,
make it difficult for the system to identify specific clusters
(Illumina). This problem is well known and is particularly
problematic during the initial cycles of the Illumina sequenc-
ing platform, where the coordinates of the individual clusters
are identified (102). Several strategies have been developed
to address this problem. These include (a) mixing in a high-
diversity DNA sample into the pool (i.e., library generated
from PhiX viral reference genome), (b) diluting the input
library to reduce the number of clusters (Illumina), (c) intro-
ducing diversity into the library during the PCR stage,
through the incorporation of spacers of different length at
the 50 ends of primers (83) or with barcodes of different
length (103). The most common approach for Illumina
sequencers is to spike-in 5–20% PhiX library ( purchased
from Illumina). This represents a loss of sequencing capacity,
but greatly improves data quality. Interestingly, the Ion
Torrent sequencers are not subject to this bias, as signal detec-
tion is performed on chip using electrochemical methods.
Each well in an Ion Torrent chip represents a separate reactor
and is not influenced by reactions in adjacent wells.
Once pooled libraries have been created and purified, they
must be quantified before sequencing. This quantification
step is essential, as both emulsion PCR (Ion Torrent, Roche
454) and bridge amplification (Illumina) are highly sensitive
to library quantity. Thus, too little input library will lead to
low yield, and too high library input will generate poor quality
data and for emulsion PCR, will create a large number of poly-
clonal sequences that will be discarded by the sequencer. Ini-
tially, libraries should be quantified using qPCR-based library
quantification kits for the appropriate platform (e.g., Illumina
and Ion Torrent library quantification kits, KAPA Biosys-
tems; Ion Library Quantitation Kit, Life Technologies).
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-13
Subsequently, users may find that standard quantification of
the library (e.g., using a Qubit fluorometer [Life Technolo-
gies], Quantus fluorometer [Promega], or similar) is suitable,
provided the same amplicon is being sequenced. It may be
useful to pool like-sized amplicons together, perform quanti-
tative PCR quantification of each pool and then mix together
in appropriate ratio for the final library.
The final concentration for the input library prior to load-
ing for either emulsion PCR or bridge amplification, should
be determined based on reports in the literature and on
instructions from individual manufacturer. For Illumina
sequencers, amplicon libraries of 4–8 pM are typically loaded
onto sequencers (e.g., 43, 104). A range of concentrations
from 6.5 (internal, at UIC DNA services facility) to 26 pM
have been used for emulsion PCR on Ion Torrent PGM sys-
tems (105, 106).
For final sequencing, a single primer (Ion Torrent, Roche
454) or multiple primers (Illumina) are needed to initiate
sequencing. For Ion Torrent and Roche 454 platforms, stand-
ard primers from each company are provided for sequencing.
For Illumina, more flexibility is provided due to the range of
different library kits that are supported (standard kits and
transposase-mediated library preparation kits or Nextera
kits), the common use of paired-end reads, and the use of sep-
arate index reads. In sequencing efforts on Illumina instru-
ments, as many as four individual reads can be performed,
including two reads for each of the paired-end sequencing
efforts, and two individual index reads (when dual indexing
is performed). This flexibility has some implications for
amplicon sequencing, particularly when paired-end reads
are expected to overlap. In all Illumina library preparations,
final library fragments contain standard (TruSeq) or Nextera
adapter sequences at the distal ends of the molecules. These
are required for bridge amplification, and a portion of these
sequences are bound to the flow cell and should not be
used for initiating sequencing reactions. Sequencing can be
initiated by any primer annealing to template molecules,
even if the primer does not anneal to the extreme distal
end of the fragment. Sequencing primers for the HiSeq and
MiSeq instruments should have melting temperatures (Tm)
of at least 55°C (HiSeq) and 60–65°C (MiSeq). However,
there does not appear to be any difficulty if the primers
have a substantially higher Tm (e.g., 70–75°C).
In the one-step amplification procedure utilized by
the Earth Microbiome Project (44; http://www.earthmicro
biome.org/emp-standard-protocols/16s/), the forward and
reverse sequencing primers and the index primer all have
Tm values ranging from 69.7°C to 74.9°C (using the Oligo-
Analyzer 3.1, with Mg2+ concentration of 2 mM and
dNTP concentration of 0.2 mM; 88). As described by the
Earth Microbiome Project, the sequencing primers include
the actual primer sequence used for PCR amplification, a
2-base linker, and a primer pad sequence of 10 bases (Fig. 2;
Table 1). The primer pad sequence is used to extend the
region over which the sequencing primer anneals, and
increases the Tm of the sequencing primer to fit that of the
Illumina platform. In the two-step amplification procedure
utilized by Fluidigm, modified primers, containing three
locked nucleic acid (LNA) substitutions per primer, are
used (i.e., CS1, CS2, and CS2rc) (Fig. 3). The LNA substi-
tutions elevate the Tm of each primer. The LNA primers
anneal to the linker sequences only and can be used for
sequencing any amplicon. Without LNA substitutions, the
CS1 and CS2 primers have Tm values of 63.3–64.2°C.
Two basic strategies for sequencing are used for the Illu-
mina platform. In the first strategy, the sequencing primers
contain the template-specific primer sequence (43). In such
a strategy, the sequencing begins directly after the primer
site and the primer sequence is never seen in the final
sequence data. The advantages of this strategy include:
expensive primer modifications are not needed, and larger
insert sizes can be sequenced because the template-specific
primer regions are not sequenced. In some cases, this can
mean that smaller sequencing kits can be used (e.g., 2 ×
150 base reads instead of 2 × 250 base reads). In most cases,
primer sequences are removed from any sequence data output,
as sequence data in these positions are generally not viable
sequence data. If multiple amplicons are used, multiple pri-
mers can be combined for sequencing on the Illumina instru-
ments. Although no specific guidelines have been set, at a low
number of different amplicons (e.g., two or three different pri-
mer sets), a pool of the primers may be made and added at
standard loading concentrations. For larger numbers of pri-
mer sets, an increasing amount of total primer concentration
can be added; such combinations may have to be empirically
validated. This strategy could introduce bias as different
sequencing primers may have different annealing behaviors
on the sequencer, thereby altering the distribution of clusters
sequenced. This has not yet been experimentally verified. A
second strategy, as utilized by the Fluidigm sequencing
approach (Fig. 2; Table 1), initiates sequencing by annealing
sequencing primers to linker sequences present in all template
molecules. However, the sequencing reaction is then initiated
at the beginning of the template-specific primer sequence. In
any case, the sequence data covering the region of the primers
are typically discarded from final output. For reactions in
which very little overlap is expected between forward and
reverse reads, the first strategy should be employed. When
excess read length is available (e.g., a 2 × 250 paired-end Illu-
mina MiSeq kit is used, but the insert size of the amplicon is
only 300 bp), the second strategy is preferred. Likewise, when
many different amplicons are to be sequenced, the use of a sin-
gle, template-independent sequencing primer is preferred. It
should be also noted that for Illumina sequencing runs, the
custom primers are added in addition to the standard Illumina
sequencing primers. The standard sequencing primers are
required for sequencing of the PhiX DNA spiked into the
final sample.
Sequence Data Output and Preliminary
Data Processing
Raw Data
Raw sequence data are delivered from each sequencer in var-
ious formats and the raw files should be retained and submit-
ted to online archives when submitting manuscripts based on
the sequence data (see below). Two common forms of raw
sequence data are generated: standard flowgram format
(SFF) files, which are produced by Roche 454 pyrosequencers
and Ion Torrent sequencers, and FASTQ files (FASTA files
with quality scores for each base included; produced by Ion
Torrent and Illumina sequencers). Both file types can be sub-
mitted to online databases such as the sequence read archive
(http://www.ncbi.nlm.nih.gov/sra/) and the European
Nucleotide Archive (http://www.ebi.ac.uk/ena). On the Ion
Torrent servers, SFF and FASTQ files can be generated as
the output from sequencing reactions. After quality trim-
ming, SFF and FASTQ files are often converted to FASTA
files, but this conversion removes quality data associated
with each base, and such information may be needed by other
downstream software. FASTA files are, however, substantially
smaller than FASTQ files of the same sequence data.
2.4.2-14 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
Demultiplexing
Raw data from sequencers may or may not be demultiplexed
(i.e., sequences separated into individual files according to
barcode). Both Ion Torrent and Illumina sequencing instru-
ments are capable of generating demultiplexed data, but
some users prefer to receive nondemultiplexed data and uti-
lize other software packages to perform this procedure (e.g.,
QIIME, 107–109). In Ion Torrent and Roche 454 platforms,
the barcode is part of the single sequencing read generated by
the instrument. Therefore, if demultiplexing is not performed
on the instrument, the sequencing read usually begins with
the barcode itself. Dual barcodes are not recommended on
single-read platforms, because only reads in which the entire
sequence is covered at high quality can be sorted into the
appropriate sample bin (i.e., the sequence read must go
through the forward barcode, the template-specific forward
primer, the sequence, the template-specific reverse primer,
and then the reverse barcode). Only a single file for each
sample is generated as Roche 454 and Ion Torrent data.
For Illumina sequencers, a separate file for “forward” and
“reverse” reads are generated (generally labeled R1 and R2,
respectively). If demultiplexing is not performed on instru-
ment, a separate index read is generated; if dual indexing is
performed, two index reads are generated. Each read will
produce a separate FASTQ file on the Illumina sequencer,
and four FASTQ files are generated when dual indexing
is used.
Paired-End Read Merging
When amplicons are short enough, the paired Illumina reads
can be merged into a single contiguous read. There are many
software packages that can perform this, including (but not
limited to) QIIME (“join_paired_ends.py”), CLC Genomics
Workbench (CLC bio), PANDAseq (110; https://github.
com/neufeld/pandaseq), and PEAR (111; http://sco.h-its.
org/exelixis/web/software/pear/doc.html). Each assembler
has different underlying algorithms, and can be “fooled”
under different conditions. For example, the merging func-
tion in CLC Genomics Workbench does not perform well
when the insert size is small and read lengths are large. We
have observed that it can be necessary to trim some fixed
number of bases from the end of each read to improve assem-
bly. The software package PEAR is designed to assemble even
under these conditions, as well as when fragment insert sizes
vary (111). Quality trimming is generally not performed
before merging; in this manner, low-quality bases may over-
lap, and an improved quality score for a single position is gen-
erated when paired reads match. Subsequently, quality
trimming can be performed with limited loss. Aggressive
trimming prior to merging may result in a lower percentage
of forward and reverse reads merging.
Quality Trimming
Quality trimming is performed to remove poor quality bases
from the 50 and 30 ends of sequences. Simultaneously, primer
sequences and ambiguous bases can be removed. It may be
appropriate to remove sequences that do not contain forward
and reverse primer sequences from the data set (depending on
sequencing strategy, amplicon length, and sequence read
length). For standard PCR and sequencing, template-specific
primer sites should be removed as the data in these regions are
not reliable because they are derived from the synthesized oli-
gonucleotide primers themselves, and not from copying of
gDNA. As mentioned, depending on sequencing strategy,
these primer sequences may or may not be included in the
final data output. For users employing the Fluidigm two-stage
sequencing strategy on the Ion Torrent sequencing platform,
the linker sequences will also be detectable in the sequence
output, after demultiplexing, when using the standard Ion
Torrent primers to initiate sequencing. Such a strategy
decreases the total output on the Ion Torrent platform (i.e.,
the sequencer must read through the barcode, CS1 linker,
and template-specific forward primer before reaching the
amplicon sequence). In general, quality trimming is per-
formed at a level of 0.01 (Q20, or expected error of 1 in
100), though there is flexibility to increase or decrease strin-
gency, depending on specific application. Ion Torrent
sequence data have traditionally had lower quality scores,
and trimming criteria may be relaxed. This does not appear
to result in poor quality data, and Ion Torrent sequence
data have been used in many microbiome studies (e.g., 22,
112, 113). Salipante et al. (23) recently performed a compar-
ison of the Ion Torrent PGM and Illumina MiSeq for 16S
rRNA gene amplicon sequencing and found generally good
agreement between the two platforms. In late 2014, Ion Tor-
rent released the Ion PGM Hi-Q Sequencing Kit, which is
anticipated to generate higher quality sequence data on the
PGM instrument. At the stage of quality trimming, length
trimming may be performed as well. For Illumina paired-end
reads which are expected to overlap, stringent size trimming
can be performed, preserving only reads that merge and are
within the expected size range. For Roche 454 and Ion Tor-
rent sequencers, individual read lengths may vary due to
the mechanism of sequencing, and typically a lower size limit
is used as a threshold. With some genes, including the micro-
bial rRNA genes, there is some size variability, and these par-
tial fragments that are not bounded by primer sites at either
end may introduce additional (artifactual) diversity into
the subsequent bioinformatics analysis.
Removal of PhiX Sequence Data
Although PhiX sequence data are typically removed from
Illumina sequence data output on instrument, some residual
sequence generally remains. This can be removed at any point
during the preliminary data processing. Within the software
package CLC genomics, for example, default stringency map-
ping of reads to the PhiX reference genome sequence (Gen-
Bank accession number NC_001422) is effective for
removing unwanted nontarget DNA. The unmapped reads
are retained and used for downstream processing.
Denoising
Amplicon sequence data generated using the Roche 454
pyrosequencing platform is typically treated with a so-called
denoising algorithm, after studies indicated that initial micro-
bial diversity estimates based on untreated data were greatly
inflating the true abundance of microbial taxa (80, 81).
This denoising approach is intended to reduce errors intro-
duced by PCR and sequencing (56, 77, 114). Such strategies
have not been implemented for Illumina or Ion Torrent
sequence data. Protocols for denoising Roche 454 pyrose-
quence data across multiple samples and sequencing runs
varies by pipeline/algorithm (115). The current QIIME
denoising protocol (http://qiime.org/tutorials/denoising_
454_data.html), for example, recommends independently
denoising each sample. Additionally, most such algorithms
are computationally expensive, requiring independent analy-
sis of multiple runs, although newer, more efficient algorithms
are available (116). In general, processing all samples or runs
together may be preferable, but is not always be feasible.
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-15
Chimera Removal
The formation of spurious amplification products during PCR
is well known and is not unique to HTS. Chimeras are PCR
amplicons that are artifacts generated during the coamplifica-
tion of homologous genes and do not represent true underly-
ing sequences (117). Wang and Wang (117) observed a
relationship between number of PCR cycles and chimera for-
mation, with more cycles generating a higher proportion of
chimeras. Longer elongation stages during PCR were found
to reduce chimera frequency. In addition, although this is
not always mentioned, the more similar two sequences are,
the higher the propensity for chimera formation, and the
smaller the chance of detecting the chimera. Nonetheless, a
number of software packages have been developed to detect
(and remove) likely chimeric sequences. These include Bel-
lerophon (118) and Chimera Slayer (119). Chimera detec-
tion and removal can be performed within the software
package QIIME, implementing the script identify_chimer-
ic_seqs.py and then filtering the chimeras from the original
sample using a QIIME script (filter_fasta.py script). Within
QIIME, there are three ways in which chimeras can be
identified, including (a) a BLAST_fragments approach, (b)
ChimeraSlayer, and (c) the usearch61 algorithm. In the
BLAST_fragments approach, the input sequence is split
into equal-sized fragments (default is three fragments) that
do not overlap. Subsequently, the individual fragments are
assigned taxonomy, and these taxonomic assignments are
compared to each other. When the taxonomy results between
different fragments do not match, the sequence is flagged as
chimeric. Reads must be in the “forward” orientation. The
ChimeraSlayer utilizes a reference database and requires
that sequences to be screened are aligned (i.e., PyNAST
aligned input; 120). Both putative chimeric sequences and
likely “parent” gene sequences are identified. Both the
BLAST_fragments and ChimeraSlayer algorithms are com-
putationally intensive and cannot be readily used for unclus-
tered sequence data.
Conversely, the usearch61 algorithm utilizes both
reference-based and de novo chimera detection algorithms
and is more rapid for analysis of unclustered sequence data.
The input sequences are screened for potential chimeras
against a user-defined reference database (e.g., Greengenes).
The process is computationally parallelizable and can identify
chimeras rapidly in raw demultiplexed sequences. When
implemented within QIIME, these methods generate a .txt
file containing flagged chimeric sequences that can be
removed from the input sequence or input representative
sequence using the filter_fasta.py script. As a note of warning,
these algorithms can fail by misidentifying chimeric sequen-
ces and by failing to identify chimeric sequences.
The SILVA-NGS pipeline (121) (https://www.arb-silva.
de/ngs/) adopts a different approach to denoising and chimera
removal similar to the closed reference approach described
below. In the SILVA_NGS pipeline, sequences are aligned
to curated databases such as that available through Green-
genes or the SILVA project (122, 123). Chimeric sequences
are discarded at this stage as they will have a low alignment
quality. There are numerous multiple sequence aligners avail-
able to date (reviewed in 124, 125). Many of the most com-
monly used aligners do not perform de novo alignments but
use manually curated reference multiple sequence alignments
to guide alignment of unknown sequences. Such aligners
include PyNast (120) frequently implemented within the
QIIME pipeline, the MOTHUR aligner (NAST-based as
well) and the SINA aligner created as part of the SILVA
project (124). The latter has shown the highest accuracy and
is adapted to work with the SILVA SSU and LSU REF data-
bases (124). Aligning the query sequences against a validated
set of sequences rather than de novo aligning allows for a great
increase in the speed, throughput and accuracy of the process.
An important aspect of 16S rRNA gene sequence align-
ment is the consideration of the secondary structure of the
rRNA molecule; however, this is not fully implemented in
all multiple sequence aligners (124). Some aligners, such as
the “Infernal” aligner implemented within the RDP pyrose-
quencing pipeline, do take into account secondary structure
during alignment procedures, but retain the capability of per-
forming rapid alignment of large sequence data sets (126,
127). Sequences with a low alignment score are typically
removed from the data set, and although these unaligned
sequences frequently represent erroneous sequences, some
environments may harbor uncharacterized organisms with
divergent gene sequences and these can be lost with such
an approach (128).
The underlying assumption of the SILVA pipeline is that
variability within clustered OTUs (typically at 98% similar-
ity) is primarily the result of sequencing errors, and that after
the removal of sequences with poor alignment score, the
remaining sequences will be clustered into OTUs (see below).
Within each obtained cluster (OTU), the sequence with the
highest alignment score will be used as representative for any
downstream analysis.
Clustering Sequences into OTUs
Clustering of sequences is an explicit data reduction strategy
to alleviate the computational burden associated with mil-
lions of sequences by grouping together highly similar sequen-
ces. Prior to clustering, sequence data are dereplicated,
particularly if the data set is large. The dereplication process
leaves in the working data set only unique sequences while
retaining the original abundance of each of these sequences.
Two types of dereplication are typically performed, including
full-length (identical sequences of identical length) and pre-
fix (sequences of different length, beginning at the same posi-
tion and with identical sequences over the shared sequence
length) dereplication. Scanning for replicates in the data
set can be done by clustering with a sequence similarity crite-
ria of 100% but is also automatically performed as part of the
usearch with quality filtering procedure within the software
package QIIME.
During clustering, sequence data are grouped into sets of
sequences with high similarity, generated using a user-defined
minimum similarity threshold. Ninety-seven percent similar-
ity is often chosen as a minimum exclusion threshold for 16S
rRNA gene sequences (i.e., organisms with sequences less
than 97% similar cannot be grouped together as the same spe-
cies) (129). This assumption has been shown to be incorrect
in some situations, due to a larger than anticipated intrage-
nome (multiple rRNA operons within a single organism)
and intragenus variability (7, 130). Furthermore, it should
be noted that the 97% similarity threshold refers to similarity
across the entire 16S rRNA gene, but is not consistent across
all lineages (e.g., 131, 132). A common misconception is to
equate clustering at 97% sequence similarity to clustering of
sequences into bins of species. OTUs composed of multi-
ple sequences with >97% similarity can artificially bring
together sequences from organisms with substantially diver-
gent genomes and physiological capabilities. One approach
to address this uncertainty it to perform clustering using dif-
ferent levels of similarity for OTU creation; by this approach,
2.4.2-16 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

of the target gene have been sequenced and are to be analyzed

simultaneously. A third clustering approach, open-reference,
is a hybrid between de novo and closed reference mapping
( pick_open_reference_otus.py). In this hybrid strategy, input
sequences are initially clustered against the reference database
as in closed-reference mapping. A subset (typically 0.1–1%)
of the remaining sequences are subject to de novo clustering.
The protocol is then repeated, using both the source reference
database and newly described taxa generated during de novo
clustering. Such as strategy can be effective but time-
consuming if a high proportion of sequences are not clustered
during the closed-reference mapping stage. New algorithms
are available for fast open-reference OTU picking as well,
which may be used in place of closed-reference picking (135).

One Final Note

Although clustering all sequences from all samples within a
study at the same time is essential for many projects, such
an approach can lead to an underestimation of the total num-
ber of OTUs within a sample. Clusters defined in individual
FIGURE 5 Relationship between OTU creation and similarity samples may be redistributed among other clusters when glob-
threshold. The number of identified OTUs generated at each simi- ally analyzed with sequences from multiple samples, thus
larity threshold is standardized by the number of OTUs identified leading to the loss of clusters in global analyses (Fig. 6).
at 100% similarity (solid line). The percent change in number of Therefore, clustering each sample by itself may better reflect
OTU at each threshold was also calculated (dashed line) and shows the true sequence diversity within each sample.
that the greatest change occurs from a 98% to 97% threshold cutoff.
The data are derived from six amplicon data sets generated on a Secondary Bioinformatics and Statistics
Roche 454 pyrosequencer. doi:10.1128/9781555818821.ch2.4.2.f5 Currently, there is a wide range of freely available or inexpen-
sive software packages that provide the user with a high
level of functionality to perform basic alpha diversity analyses
it is also possible to explore microbial diversity at different
(i.e., richness and evenness of each sample), compositional
phylogenetic levels with the same data set (133) (Fig. 5).
characterization (e.g., taxonomy), beta-diversity (e.g., differ-
Thus, 97% similarity is used merely as an informatics tools
entiation in microbial community composition among sam-
as opposed to one consistently and rigorously founded in
ples), statistical tests, and an ever-expanding range of data
microbial biology. Readers may wish to explore, in any case,
visualizations. The most frequently used software packages
many excellent reviews of the microbial species concept
include QIIME (107; http://qiime.org/, 108, 109), mothur
and linking single gene analyses with whole genome sequenc-
(136; www.mothur.org/), Ribosomal Database Project
ing (129, 132, 134).
(137; http://rdp.cme.msu.edu/index.jsp), phyloseq (138),
As implemented within QIIME ( pick_otus.py), a wide
SILVA-NGS (123), and AXIOME (139), which manages
variety of tools are available (e.g., cd-hit, BLAST, mother,
uclust, and usearch). Three basic strategies for OTU picking
are employed, including de novo, closed-reference, and open-
reference. In de novo clustering ( pick_de_novo_otus.py),
input sequences are aligned to each other, and sequences
with similarity above the threshold are clustered. Sequences
failing to align to established clusters become the seeds for
novel OTU. No external reference databases are required,
but the method is computationally intensive and cannot
be easily distributed among different computer nodes. Such
a method must be performed if no reference databases
are available for the gene of interest. Conversely, closed-
reference OTU picking ( pick_closed_reference_otus.py)
aligns sequences to a reference database (e.g., Greengenes;
122), reduces computational demand, and is easily distributed
among multiple nodes. However, such an algorithm cannot
detect novel taxa. If such an approach is used, the proportion
of reads that do not map to the reference database should be
calculated to determine if the data loss (i.e., unclustered FIGURE 6 Effect of single-sample versus complete study clus-
sequences) is substantial enough to preclude further analysis. tering on observed microbial diversity. A consistent decrease in
Closed-reference mapping is likely to be suitable for deeply the number of measured OTUs was observed when sample diversity
characterized environments such as human-associated micro- was estimated based on total data set clustering as opposed to
biomes, but is less likely to be suitable for highly diverse envi- one-by-one sample clustering and diversity calculations. The four
ronments such as soil. Closed-reference mapping is also data sets, from freshwater (DB1), seawater (DB2), marine sediments
required for functional prediction analysis using the software (DB3), and microbial mats (DB4), consisted of 10–40 samples each,
package PICRUSt (14). Finally, closed-reference mapping with amplicons generated on a Roche 454 pyrosequencer.
can also be employed when sequences from different regions doi:10.1128/9781555818821.ch2.4.2.f6
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-17
the other analysis pipelines. In particular, mothur and QIIME
provide a wide suite of applications from raw data processing
to visualization and statistical testing. In this chapter, several
QIIME scripts have been mentioned as examples, and the
reader should consider that mothur can be a suitable alterna-
tive to QIIME. Also viable but less heavily used for amplicon
sequencing are the online metagenomics rapid annotation
using subsystem technology (MG-RAST) software (140; pri-
marily for shotgun metagenomic data) and MEtaGenome
ANalyzer MEGAN (141; after high-throughput BLAST
analysis of amplicon sequence data). Statistical analyses can
be performed with a variety of software packages, including
STAMP (142), Primer6 (Primer-E; www.primer-e.com),
Canoco (www.canoco5.com; 143), as well as many algo-
rithms implemented within the R environment (e.g., 138).
Two broad mechanisms for amplicon sequence data anal-
ysis are commonly employed. These include, for lack of better
terms, “discrete” analyses and “continuous” analyses. In the
continuous approach, multivariate microbial community
data is converted to univariate data using the UniFrac metric
(78). Briefly, sequence data are used to generate a phyloge-
netic tree, and in a pair-wise fashion, the UniFrac metric is
used to generate a univariate value representing the sum of
phylogenetic tree branch length unique to each sample.
Both unweighted ( presence or absence of taxa) and weighted
( presence and abundance) UniFrac values have been used,
though a generalized UniFrac measure has been developed
to downweight the impact of dominant lineages (144). As
the number of sequences generated in routine microbiome
studies can easily exceed 1 million sequences, the production
of phylogenetic trees can be computationally impractical. For
such approaches, the Fast UniFrac tool has been developed
(145). Here, a fixed tree is used, and amplicon sequences
are mapped against the reference tree. This method circum-
vents having to generate a de novo phylogenetic tree, but
retains the phylogenetic structure of the original tree. Ulti-
mately, a pairwise distance matrix is generated from such
analyses, and these distance matrices are used for downstream
statistical analyses and visualization. Visualization of UniFrac
matrices is generally performed using principal coordinates
analysis (PCoA) (78), and more recently with edge principal
components analysis (146).
In the discrete approach, sequence data are combined into
groups of highly similar sequences based on a user-defined
similarity threshold, as described above (Clustering Sequen-
ces into OTUs). These groups of sequences, typically clus-
tered at 97% sequence similarity when using furthest
neighbor clustering algorithm, which is the most common,
prevents sequences within each group being less than 97%
similar to each other. These are termed OTUs and do not
necessarily represent single bacterial species (147). In OTU-
based methods, clustering is dependent on the other sequen-
ces within the data set (8). Due to the possibility for artifacts
from the clustering algorithm, highly similar OTUs can be
generated, and the exact sequences clustered into each
OTU can vary if clustering is performed iteratively, random-
izing the order of sequences each time. In any case, the clus-
tering (for example, implemented in QIIME using the
pick_otus.py function), serves as a data reduction strategy
and as a discrete binning strategy for sequences. Typically, a
single representative sequence (e.g., the most abundant
sequence) from each OTU is extracted, the sequence classi-
fied, and the classification propagated for all sequences in
the OTU. In general, all sequences from a given study are
pooled together for such a clustering-based approach. As a
result, the software generates a OTU table, also called a
biological observation matrix (BIOM; 148), which contains
a taxon-by-sample grid, with numbers of sequences per sam-
ple belonging to each OTU or higher level taxon. Because
each OTU is classified (annotated), higher taxonomic level
BIOMs can be generated. Some caveats should be noted.
Microbial taxonomy is fraught with difficulties, and in some
cases sequences classified as belonging to the specific genus
may actually represent very different taxa (15). Advantages
of “discrete” analysis include the following. (a) The data
reduction strategy is rapid, as a small subset of representative
sequences are classified. Clustering is the single most compu-
tationally demanding aspect of this approach. (b) BIOM
data can be easily analyzed using standard ecological meas-
ures, including alpha and beta diversity (149, 150). In beta-
diversity analyses, samples are frequently compared in a
pair-wise fashion and converted to univariate data using the
Bray-Curtis dissimilarity metric or similar (see below). (c)
Taxon-by-taxon statistical tests can be performed on BIOM
files, provided that robust replication has been performed
(e.g., Kruskal-Wallis nonparametric analysis of variance,
implemented within QIIME using the group_significance.
py function). (d) Correlational analyses (e.g., Pearson corre-
lation, implemented within QIIME using the otu_category_
significance.py function) can be readily performed using
BIOM data at any taxonomic level. (e) Clustering at higher
taxonomic levels (e.g., family) can reduce the impact of spu-
rious diversity developing as the result of PCR or sequencing
error.
Considerations for High-Throughput Amplicon
Sequence Data Processing
Normalizing Sequence Counts
Following sequence quality control and OTU clustering, it is
appropriate to standardize the sample library sizes so they are
on a common scale. Comparing samples with similar library
sizes is required to avoid differences in diversity being
explained solely by sequencing effort or if planning on testing
hypotheses about differential OTU abundance. There are
three commonly seen methods for normalizing library sizes:
(a) rarefying sequence data sets from each sample such that
each sample has the same number of sequences (151), (b)
converting OTU counts to a proportion (dividing by total
library size), or (c) converting libraries to a common scale
by statistically estimating the relative depths of the sample
libraries and then standardizing OTU variances using a mix-
ture model (152). When the microbial communities being
compared come from very different environments, any of
the normalizing methods typically produces useable results.
The first technique is the most widespread but may be stat-
istically inadmissible since valid data are being discarded.
This serves to increase Type II error (decreased sensitivity),
and the random subsampling artificially increases uncertainty
(152). McMurdie and Holmes (152) indicate that rarefying
to an even depth should never be used under any circumstan-
ces; however, subsampling can reduce bias associated with
highly uneven library sizes that are frequent in NGS amplicon
sequencing approaches (151). Likewise, converting OTU
counts to a proportion can inflate the role of the most abun-
dant OTU when libraries sizes are substantially different
between samples. Uneven library sizes can affect alpha diver-
sity calculations strongly and beta-diversity calculations
when strong transformations, such as log or binary, are
employed. If subsampling is performed, it is appropriate to
perform multiple resamplings from the same data set to gen-
erate average values.
2.4.2-18 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
For the best results in downstream analysis, particularly
with respect to determining differentially abundant taxa
between samples, we recommend converting sequence libra-
ries to a common scale and modeling counts using a negative
binomial distribution (DESeq2) or zero-inflated Gaussian
distribution (metagenomeSeq) (152–155). Through both
experimental validation and computer simulations, standard-
izing libraries through these methods maximized the number
of samples, eliminated arbitrary sample size cutoffs, preserved
true pairwise sample similarities, and allowed for detection of
differential abundance for moderately abundant OTU (152).
The newest release of QIIME (1.9.0) includes scripts that uti-
lize these methods (see, for example, normalize_table.py and
differential_abundance.py).
Analyses of OTU Tables
Once gene amplicon data have been converted to a species
count table (BIOM; 148), many common ecological analyt-
ical tools can be applied. A typical goal in microbial ecology is
to visualize how similar samples or groups of samples are to
each other, and then to test a priori hypotheses to determine
if samples or groups of samples are significantly different from
each other. Such tests can be performed at the whole com-
munity level, whereas other tests attempt to identify signifi-
cant shifts in relative abundance at a taxon-by-taxon level.
At the community level, complex multivariate microbial
data (i.e., abundance of many taxa per sample) are generally
converted to univariate data through the calculation of a sim-
ilarity coefficient for each pair of samples. This pairwise com-
parison leads to the generation of a resemblance matrix,
which is further exploited for statistical tests and visualization
(described below). Thus, when approaching complex micro-
bial community analysis in such a manner, the choice of an
appropriate coefficient (156–158) is crucial. Online resources
for choosing the most appropriate coefficients are available
(157). The Bray-Curtis dissimilarity metric (159) is most
commonly employed, because it is not affected by changes
in scale or shared absences (taxa that are missing in both sam-
ples), and generates a value from 0 (no species in common) to
100 (identical samples) (158, 160). However, many other
metrics are available and may be more appropriate for certain
data sets.
Transformations can be applied to common-scaled
counts. Some techniques ( particularly those grounded in
Euclidean distances) require specific assumptions to be met
and transforming the raw data can be used to meet these
assumptions. The Guide to Statistical Analysis in Microbial
Ecology (GUSTA ME; 157) details how transformations
can be employed to produce greater insight into microbial
communities. Often, when using the Bray-Curtis metric,
common transformations are applied to downweight the
most abundant OTU so that moderate to rare OTUs can con-
tribute to dissimilarity values. Ordered from least to most
extreme, these include square root, fourth root, log(X + 1),
and presence/absence (161). Transformation is generally rec-
ommended, but not required, and nonparametric ordination
techniques such as nonmetric multidimensional scaling do
not require normally distributed data. However, analyses of
communities in which rare taxa are important, or in which
relatively few taxa dominate all samples, may benefit from
increasingly extreme transformation.
Generating Resemblance Matrices
As described already, a series of data reduction strategies are
employed to convert multivariate microbial community
sequence data to a resemblance matrix that can be used for
hypothesis testing and visualization. These strategies include
clustering of sequence data into an OTU table, generation of
sample-by-taxon matrices (BIOMs) from the OTU data, and
finally the generation of a resemblance matrix by employing
pair-wise calculation of a coefficient such as Bray-Curtis sim-
ilarity. Such analyses have been reviewed, and the reader is
encouraged to review this extensive literature (156–158,
162). Some considerations are provided here.
Although Bray-Curtis similarity is used commonly, Kuc-
zynski et al. (156) have suggested that the chi-squared family
of distance metrics better represented simulated microbial
samples taken across a one-dimensional environmental gra-
dient whereas Bray-Curtis dissimilarities struggled to accu-
rately represent a gradient response. Chi-squared distance is
the underlying distance used in correspondence analysis
and related methods such as detrended correspondence anal-
ysis and canonical correspondence analysis. Conversely, Faith
and colleagues (163), using simulated species distributions,
found that the chi-squared distance performed poorly in
recognizing a two-dimensional environmental gradient.
Chi-squared distances can be conceptually harder to under-
stand since maximum distance is not constant (when no spe-
cies are shared); instead, chi-squared distances are influenced
by relative differences in species abundances, even when no
species are shared.
When the phylogenetic relationship between OTUs is
known, resemblance coefficients that use phylogenetic infor-
mation can be used. Such information can improve character-
ization of microbial community structure because similarity
thresholds—such as the commonly used threshold of 97%
similarity—are essentially arbitrary (78). Arbitrary cutoffs
such as those used for clustering are confounded by the
absence of scientific consensus for the term “species” within
microorganisms (e.g., 132, 164, 165), by variability between
lineages in the correlation of 16S rRNA gene sequence and
whole genome similarity, and by variability between different
regions of the 16S rRNA gene (i.e., 97% similarity in V1–V3
region may represent a different group of organisms than 97%
similarity in the V4 region). In particular, when OTU picking
is performed using furthest neighbor algorithms (i.e., all
sequences clustered into a single OTU must be at least as sim-
ilar as the threshold similarity with every other sequence
within the OTU), highly similar sequences can be split into
different OTUs. Such spurious OTUs can be addressed by fur-
ther data reduction and grouping of OTU into higher taxo-
nomic groups (e.g., species, genus), through continuous
phylogenetic methods such as UniFrac, or through coupled
OTU-phylogenetic approaches. As reviewed by Parks and
Beiko (166), there are three main ways to incorporate branch
lengths from a phylogenetic tree into a resemblance analysis:
(a) most recent common ancestor, (b) complete lineage, and
(c) complete tree methods. Parks and Beiko (166) recom-
mended the phylogenetic Gower measure to place heavy
emphasis on rare OTUs, the Soergel measure (very similar
to phylogenetic Bray-Curtis and weighed UniFrac) that
incorporates both abundant and rare OTU in a more bal-
anced fashion, and the Morisita-Horn measure (similar to
Rao’s Hp and FST) which heavily emphasizes abundant
OTUs.
As a final guide, we recommend evaluating a priori what
factors are important to the researcher for comparing their
samples. Studies in which highly similar samples are exam-
ined can benefit from metrics that favor lower abundance
OTUs (after applying a low cutoff to reduce noise) such the
Canberra or Gower metrics. Studies in which microbial com-
munities are more divergent are likely to employ more
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-19
balanced metrics like Bray-Curtis, Kuczynski, or Soegel meas-
ures. Including phylogenetic information can be powerful,
but care should be given to minimize long artifactual branches
in the phylogenetic tree.
Statistical Testing
A Priori Hypothesis Testing
For many studies, it is important to statistically test whether
two or more a priori defined groups (e.g., treatments) of sam-
ples are significantly different. Some of the more common
methods used are ANOSIM (analysis of similarity), adonis,
MRPP (Multi-Response Permutation Procedure) (167) and
PERMANOVA ( permutational multivariate analysis of var-
iance; synonymous with NPMANOVA) (161, 168). In
some instances, multivariate ANOVA (MANOVA) can be
used, but in our experience it is challenging to meet all the
underlying assumptions. Functionally, adonis and PERMA-
NOVA are nearly identical and when implemented within
the software package R, the only difference is that adonis is
more robust to sample group names. All these tests are permu-
tative procedures that operate on similarity or distance matri-
ces, although ANOSIM is based on rank similarities, whereas
PERMANOVA and adonis are not. ANOSIM is also more
sensitive to differences in dispersion (within group spread)
than PERMANOVA and adonis (168). All methods can be
calculated using QIIME (compare_categories.py), which
uses the underlying R package vegan for all calculations,
and within the software package Primer6 with Permanova.
MOTHUR can calculate ANOSIM and AMOVA (synony-
mous to PERMANOVA).
Results from ANOSIM give a test statistic (R) and associ-
ated significance value (p-value). A test statistic close to 1
with a p-value <0.05 indicates groups are significantly differ-
ent. A test statistic close to 0 indicates samples within a group
are equally similar to samples in other groups (null hypothe-
sis). Care should be taken when comparing two groups made
up of many samples, since a low test statistic (close to 0) could
still be significant (p < 0.05) but not biologically relevant.
We have observed a wide range of significant R statistic values
from studies conducted on environmental samples, where
variation within groups can be high, as compared to highly
controlled experimental studies, where within group simi-
larity is generally high. Adonis results in an ANOVA-like
table including a pseudo F-statistic, an R 2 value, and a signifi-
cance level (p-value), while PERMANOVA provides a
pseudo F-statistic and associated p-value. When calculating
adonis or PERMANOVA, PERMDISP should also be per-
formed to determine whether significance is solely due to
group dispersions.
Detecting Differentially Abundant Taxa
After verifying that treatments/groupings are significantly dif-
ferent, a common goal is to identify those taxa that contribute
most to the differences. A common approach is to use the sim-
ilarity of percentages test (SIMPER) (161). SIMPER deter-
mines how much each taxon within an OTU table
contributes to pairwise Bray-Curtis dissimilarity values
between two defined groups of samples, and provides contri-
butions of each OTU to the average dissimilarity between
groups and average contribution to group similarity. Such
analyses can be confounded by differences in OTU means
between groups with intragroup OTU variation. Therefore,
some identified OTUs may not be truly differentially abun-
dant, but simply highly variable within groups (169).
SIMPER can only be used with the Bray-Curtis dissimilarity
coefficient.
Other approaches for detecting differently abundant taxa
include using one-way ANOVA and with corrections for mul-
tiple testing ( popular corrections include Benjamini-
Hochberg false discovery rate, FDR; Bonferroni [extremely
conservative], and Storey’s FDR). ANOVA assumes OTU
distributions are normal (rarely valid, and often requires trans-
formations to meet this assumption). The Kruskal-Wallis test,
a nonparametric ANOVA, is also often used, and does not
require OTU abundances to be normally distributed (within
QIIME, group_significance.py). Such methods can identify
significantly differently abundant taxa even within the con-
text of communities that are highly similar as suggested by
total community analyses such as ANOSIM. More recently,
fitting count data to mixture models like the negative bino-
mial model or zero-inflated Gaussian models have been sug-
gested (152), and these have been shown to be sensitive
and specific in detecting differentially abundant OTUs.
These inference models have been shown to handle a large
range of effect sizes, number of biological replicates, and
library sizes (152).
Additional Tools for Analysis of Complex Microbial
Community Data Sets
A wide array of tools are available for additional analyses of
microbial communities, and the reader is guided towards
these tools. One of the most exciting new tools is that of pre-
dictive functional profiling of microbial communities based
on 16S rRNA gene amplicon sequence data alone. Recently
published, the Phylogenetic Investigation of Communities by
Reconstruction of Unobserved States (PICRUSt) bioinfor-
matics software package allows the user to predict functional
content from marker gene surveys through comparison of
marker gene sequences to a database of genome-sequenced
microorganisms (14). Such an approach has been shown to
be highly effective in predicting metagenome gene content
as generated through shotgun metagenome sequencing
(13). In addition, this software adjusts 16S rRNA gene ampli-
con abundances by known rRNA gene operon abundances in
reference genomes. The fundamental distortion of microbial
community analysis by 16S rRNA gene copy number varia-
tion across microbial genomes has also been addressed in a
number of recent publications (170), and can also be imple-
mented within the Ribosomal Database Project (171). A
number of caveats apply to this predictive bioinformatics
strategy: (a) only closed-reference OTU-picking can be per-
formed, since the approach depends on similarity of sample
data to gene sequences from genome-sequenced organisms;
(b) this strategy is only effective for environments with suffi-
cient reference genomes in the reference database (13), and
(c) this strategy can be confounded when shifts in community
composition occur at too fine a taxonomic level to be distin-
guished by such predictive analyses.
For identifying significant trends in microbial community
structure across large data sets, several software packages are
available, including the LDA Effect Size (LEfSe) algorithm
(172), Microbiomes: Picking Interesting Taxonomic Abun-
dance tool (microPITA) (173), and Multivariate Analysis
by Linear Models (MaAsLin) package (174), all available
through a Galaxy instance maintained by the Hutten-
hower laboratory (http://huttenhower.sph.harvard.edu/gal
axy/). The MaAsLin package can be used to find associations
between metadata and microbial community abundance,
whereas microPITA surveys microbiome data to identify
2.4.2-20 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
samples with unique features, such as maximum diversity.
The LEfSe tool can be used for rapid discovery and visualiza-
tion of differentially abundant taxa between a priori defined
groups.
Isolating specific species that may have significant contri-
butions or associations with metadata categories is also an
important step toward interpreting ecological dynamics in
sequencing libraries. Indicator species analysis (175), imple-
mented as indval in labdsv (176), can identify specific
OTUs statistically associated with metadata classes. For exam-
ple, indicator analysis was used to screen less invasive oral
sampling sites for organisms implicated in chronic periodon-
titis (177). Rare biosphere species can sometimes have
dynamics associated with specific events, and therefore trac-
ing cyclical abundance patterns can provide significant
insight. Identifying these so-called conditionally rare taxa
(178) in time series data sets can provide substantial insight
into ecosystem dynamics. Experimental design is imperative
when investigating these cyclical patterns as time series sam-
pling with sufficient frequency and duration are required to
observe patterns.
DATA VISUALIZATION AND NETWORK
ANALYSIS
A wide range of options for visualization of complex micro-
biome data sets are available. The most common are ordina-
tion plots derived from similarity (or dissimilarity) matrices
generated by UniFrac or metrics such as Bray-Curtis. Thus,
many manuscripts will present a nonmetric multidimensional
scaling plot for discrete analyses, whereas PCoA plots are typ-
ically used for visualizing data sets based on UniFrac distan-
ces. We recommend the GUSTA ME (http://mb3is.megx.
net/gustame) site as a starting place for identifying the appro-
priate ordination strategy for each data set. In addition to ordi-
nation plots, heat maps are frequently used in the analysis of
microbial community HTS data sets (179). The software
package EMPeror is well integrated with the software package
QIIME and provides a wide range of visualization tools (180).
One of the most important strategies for using microbiome
studies to identify microbial associations is network analysis.
Network analysis of microbial community amplicon data is
easily implemented as a visualization tool and is an effective
way to explore high-dimensional data (181). Networks
have been used to visualize broad taxonomic patterns in
HTS data (e.g., 182–184), as well as investigating deeper
community structure through analyzing network topology.
The most common such applications are co-occurrence net-
works based on various metrics, where co-occurrence or
exclusion patterns among elements of a complex system,
such as taxa or ecosystem processes, can be identified from
network structure. Co-occurrence networks have successfully
identified such patterns as niche specialization in the human
microbiome (185) and correlating changes in Bacteroides and
Firmicutes levels with carbohydrate and fat oxidation, respec-
tively (186). Biological information can also be inferred from
network motifs. For example, a denser network (nodes tend-
ing to have more edges) tends to be more robust and more
easily adapted to environmental change. This is consistent
with correlations between higher diversity and ecosystem
services.
Network analysis is also widely used in the analysis of
metagenomes and genomes, where topological differences
in representations such as metabolic networks are useful
for identifying unique characteristics of an experimental
condition or contrasting conditions or sampling locations
(e.g., 187, 188). Networks provide insight into metage-
nomic/genomic capacity, visualizing a complex web of
interactions. Network visualizations allow for massive data
to be explored quickly, and since the underlying structure
of communities is network-like, developing network analy-
sis tools tailored to microbial ecosystems is an active area of
research. One such algorithm is Local Similarity Analysis
(189), which looks for time-dependent relationships, that
is, correlations at identical and adjacent time points, between
microbes and between microbes and environmental factors.
Software packages Gephi (http://gephi.github.io/) and Cyto-
scape (www.cytoscape.org) are the most prevalent for net-
work visualization and analysis, but libraries for common
programming languages such as R, Python, and C/C++ also
exist (e.g., igraph, www.igraph.org; NetworkX: http://net
workx.github.io/).
Data Storage, Workflow Tracking, and Data and
Metadata Submission
Data Storage
Storage of NGS data is one of the most vexing aspects of the
HTS workflow. A variety of sequence files are generated by
each sequencer, but in general, users retain only the SFF or
FASTQ files generated by each instrument. These files should
be ideally stored in multiple locations as a long-term back-up.
For studies to be published in the scientific literature, such
files (either SFF or FASTQ) should be uploaded to publicly
accessible databases such as the SRA (http://www.ncbi.nlm.
nih.gov/sra/) or European Nucleotide Archive (http://www.
ebi.ac.uk/ena). Other options include the environmentally
oriented databases such as PANGAEA (http://www.pan
gaea.de/) or MEGX (http://mb3is.megx.net/). In addition,
submission of sequence data should be accompanied with
sample-specific metadata. Metadata submission has been
standardized under the minimum information about a marker
gene sequence (MIMARKS) and minimum information
about any (x) sequence (MIxS) specifications (190).
Raw data files, even when zipped, generally exceed 5 GB,
and from larger sequencers such as the HiSeq, files can be on
the scale of 40 GB per lane. Thus, data storage and transfer
can become limiting factors, and in some cases, data may be
downloaded to external hard drives and shipped via over-
night courier rather than transferred using FTP or similar.
Increasingly, however, cloud computing can be employed to
transfer, store, and process sequence data. For example, Illu-
mina maintains the BaseSpace cloud server that is freely (cur-
rently) available to new customers. Sequence data generated
from Illumina instruments can be directly transferred to Base-
Space accounts, and this data can be shared or transferred to
other users. In addition, some limited bioinformatics analyses
can be performed in the cloud through this interface. For
example, Illumina maintains a 16S rRNA gene amplicon
analysis tool (16S Metagenomics App) that performs a rapid
classification of sequences in a manner akin to the Ribosomal
Database Project (RDP) classifier (137), and produces an
informative report. Although not appropriate for larger scale
studies in which samples are to be compared, this can provide
a rapid means to determine if amplicon sequencing was suc-
cessful. Ion Torrent has a competitor product called Ion Tor-
rent Suite, with similar functionality. Other computational
functionality is available in the cloud as well. An instance
of the software package QIIME is maintained on the Amazon
cloud, and data can be uploaded, stored, and processed there.
This provides users without access to computers with high
memory capacity to process large amplicon data sets. In any
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-21
case, users are encouraged to store raw data in multiple loca-
tions and upload raw data to data archives.
HTS projects generate relatively large amounts of data and
the analysis techniques are becoming progressively more com-
plicated. Additionally, versions of software packages and
reference libraries change frequently and can significantly
influence analysis results. We therefore suggest that efforts
should be made to provide a complete record of bioinformatic
pipelines in publications. This can be as minimal as shell
scripts reproducing the analysis or as thorough as recorded
analysis sessions using environments such as iPython or con-
trol files such as AXIOME (139). Another emerging alterna-
tive is packaging the analysis environment, data, and
commands through instances of Docker or Amazon Elastic
Compute Cloud (EC2), thereby encapsulating the entire
analysis workflow. Efforts to make each research study repro-
ducible has the added benefit of organizing data and analysis
locally, using tools such as git/GitHub, iPython, Sweave, and
knitr and automating local pipelines.
We kindly than Dr. Alvaro Hernandez and Chris Wright of the Uni-
versity of Illinois at Urbana-Champaign (High-Throughput Sequencing
and Genotyping Unit) and Sarah Owens (IGSB-NGS Core Facility at
the Argonne National Laboratory) for discussions regarding sequencing
on the Illumina platform. Weihua Wang and Raghavee Venkatramanan
(DNA Services Facility, UIC) are thanked for their contributions to
sequencing on the Ion Torrent PGM.
REFERENCES
1. Pace NR, Stahl DA, Lane DJ, Olsen GJ. 1986. The anal-
ysis of natural microbial populations by ribosomal RNA
sequences, Adv Microb Ecol 9:1–55.
2. Schmidt TM, DeLong E, Pace N. 1991. Analysis of a
marine picoplankton community by 16S rRNA gene clon-
ing and sequencing. J Bacteriol 173:4371–4378.
3. Reysenbach A-L, Wickham GS, Pace NR. 1994. Phyloge-
netic analysis of the hyperthermophilic pink filament com-
munity in Octopus Spring, Yellowstone National Park.
Appl Environ Microbiol 60:2113–2119.
4. Amann RI, Ludwig W, Schleifer K-H. 1995. Phyloge-
netic identification and in situ detection of individual
microbial cells without cultivation. Microbiol Rev 59:
143–169.
5. Woese CR, Gutell R, Gupta R, Noller H. 1983. Detailed
analysis of the higher-order structure of 16S-like ribosomal
ribonucleic acids. Microbiol Rev 47:621.
6. Lee ZM-P, Bussema C, Schmidt TM. 2009. rrnDB: docu-
menting the number of rRNA and tRNA genes in bacteria
and archaea. Nucleic Acids Res 37:D489–D493.
7. Větrovský T, Baldrian P. 2013. The variability of the 16S
rRNA gene in bacterial genomes and its consequences for
bacterial community analyses. PLoS One 8:e57923.
8. Schloss PD, Westcott SL. 2011. Assessing and improving
methods used in operational taxonomic unit-based
approaches for 16S rRNA gene sequence analysis. Appl
Environ Microbiol 77:3219–3226.
9. Frank KL, Rogers DR, Olins HC, Vidoudez C, Girguis
PR. 2013. Characterizing the distribution and rates of
microbial sulfate reduction at Middle Valley hydrothermal
vents. ISME J 7:1391–1401.
10. Snelling TJ, Genç B, McKain N, Watson M, Waters SM,
Creevey CJ, Wallace RJ. 2014. diversity and community
composition of methanogenic archaea in the rumen of
Scottish upland sheep assessed by different methods.
PLoS One 9:e106491.
11. Dumont MG, Lüke C, Deng Y, Frenzel P. 2014. Classifi-
cation of pmoA amplicon pyrosequences using BLAST and
the lowest common ancestor method in MEGAN. Front
Microbiol 5:34.
12. Sánchez O, Ferrera I, González JM, Mas J. 2013. Assessing
bacterial diversity in a seawater-processing wastewater
treatment plant by 454-pyrosequencing of the 16S rRNA
and amoA genes. Microb Biotechnol 6:435–442.
13. Xu Z, Malmer D, Langille MG, Way SF, Knight R. 2014.
Which is more important for classifying microbial com-
munities: who’s there or what they can do? ISME J 8:
2357–2359.
14. Langille MG, Zaneveld J, Caporaso JG, McDonald D,
Knights D, Reyes JA, Clemente JC, Burkepile DE,
Thurber RLV, Knight R. 2013. Predictive functional
profiling of microbial communities using 16S rRNA
marker gene sequences. Nat Biotechnol 31:814–821.
15. Goodrich JK, Di Rienzi SC, Poole AC, Koren O, Walters
WA, Caporaso JG, Knight R, Ley RE. 2014. Conducting a
microbiome study. Cell 158:250–262.
16. Hamady M, Knight R. 2009. Microbial community profil-
ing for human microbiome projects: tools, techniques, and
challenges. Genome Res 19:1141–1152.
17. Kuczynski J, Lauber CL, Walters WA, Parfrey LW,
Clemente JC, Gevers D, Knight R. 2011. Experimental
and analytical tools for studying the human microbiome.
Nat Rev Genet 13:47–58.
18. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D,
Lozupone CA, Turnbaugh PJ, Fierer N, Knight R.
2011. Global patterns of 16S rRNA diversity at a depth
of millions of sequences per sample. Proc Natl Acad Sci
108:4516–4522.
19. Nguyen NH, Smith D, Peay K, Kennedy P. 2014. Parsing
ecological signal from noise in next generation amplicon
sequencing. New Phytol 205:1389–1393.
20. Scholz MB, Lo C-C, Chain PS. 2012. Next generation
sequencing and bioinformatic bottlenecks: the current
state of metagenomic data analysis. Curr Opin Biotechnol
23:9–15.
21. Prosser JI. 2010. Replicate or lie. Environ Microbiol 12:
1806–1810.
22. Yergeau E, Lawrence JR, Sanschagrin S, Waiser MJ,
Korber DR, Greer CW. 2012. Next-generation sequenc-
ing of microbial communities in the Athabasca River and
its tributaries in relation to oil sands mining activities.
Appl Environ Microbol 78:7626–7637.
23. Salipante SJ, Kawashima T, Rosenthal C, Hoogestraat
DR, Cummings LA, Sengupta DJ, Harkins TT, Cookson
BT, Hoffman NG. 2014. Performance comparison of Illu-
mina and Ion Torrent next-generation sequencing plat-
forms for 16S rRNA-based bacterial community profiling.
Appl Environ Microbiol 80:7583–7591.
24. Starke IC, Vahjen W, Pieper R, Zentek J. 2014. The
influence of DNA extraction procedure and primer set on
the bacterial community analysis by pyrosequencing of bar-
coded 16S rRNA gene amplicons. Mol Biol Int 2014:
548643.
25. Rand KH, Houck H. 1990. Taq polymerase contains bac-
terial DNA of unknown origin. Mol Cell Probes 4:445–450.
26. Corless CE, Guiver M, Borrow R, Edwards-Jones V,
Kaczmarski EB, Fox AJ. 2000. Contamination and sensi-
tivity issues with a real-time universal 16S rRNA PCR. J
Clin Microbiol 38:1747–1752.
27. Meier A, Persing D, Finken M, Böttger E. 1993. Elimina-
tion of contaminating DNA within polymerase chain reac-
tion reagents: implications for a general approach to
detection of uncultured pathogens. J Clin Microbiol 31:
646–652.
28. Hughes M, Beck L, Skuce R. 1994. Identification and
elimination of DNA sequences in Taq DNA polymerase.
J Clin Microbiol 32:2007–2008.
2.4.2-22 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
29. Carroll NM, Adamson P, Okhravi N. 1999. Elimination
of bacterial DNA from TaqDNA polymerases by restriction
endonuclease digestion. J Clin Microbiol 37:3402–3404.
30. Maiwald M, Ditton H-J, Sonntag H-G, von Knebel Doe-
beritz M. 1994. Characterization of contaminating DNA
in Taq polymerase which occurs during amplification
with a primer set for Legionella 5S ribosomal RNA. Mol
Cell Probes 8:11–14.
31. Klaschik S, Lehmann LE, Raadts A, Hoeft A, Stuber F.
2002. Comparison of different decontamination methods
for reagents to detect low concentrations of bacterial 16S
DNA by real-time-PCR. Mol Biotechnol 22:231–242.
32. Borst A, Box A, Fluit A. 2004. False-positive results and
contamination in nucleic acid amplification assays: sugges-
tions for a prevent and destroy strategy. Eur J Clin Microbiol
Infect Dis 23:289–299.
33. Bartram AK, Poon C, Neufeld JD. 2009. Short technical
reports. Biotechniques 47:1019–1022.
34. Newsome T, Li B-J, Zou N, Lo S-C. 2004. Presence of
bacterial phage-like DNA sequences in commercial Taq
DNA polymerase reagents. J Clin Microbiol 42:2264–2267.
35. Salter SJ, Cox MJ, Turek EM, Calus ST, Cookson WO,
Moffatt MF, Turner P, Parkhill J, Loman NJ, Walker
AW. 2014. Reagent and laboratory contamination can crit-
ically impact sequence-based microbiome analyses. BMC
Biol 12:87.
36. Knights D, Kuczynski J, Charlson ES, Zaneveld J, Mozer
MC, Collman RG, Bushman FD, Knight R, Kelley ST.
2011. Bayesian community-wide culture-independent
microbial source tracking. Nat Meth 8:761–763.
37. Barton H, Taylor N, Lubbers B, Pemberton A. 2006.
DNA extraction from low-biomass carbonate rock: an
improved method with reduced contamination and the
low-biomass contaminant database. J Microbiol Meth 66:
21–31.
38. Ahn SJ, Costa J, Emanuel JR. 1996. PicoGreen quantita-
tion of DNA: effective evaluation of samples pre-or
post-PCR. Nucleic Acids Res 24:2623–2625.
39. Kontanis EJ, Reed FA. 2006. Evaluation of real-time PCR
amplification efficiencies to detect PCR inhibitors. J Foren-
sic Sci 51:795–804.
40. Kennedy K, Hall MW, Lynch MD, Moreno-Hagelsieb G,
Neufeld JD. 2014. Evaluating bias of Illumina-based bacte-
rial 16S rRNA gene profiles. Appl Environ Microbol 80:
5717–5722.
41. Simbolo M, Gottardi M, Corbo V, Fassan M, Mafficini A,
Malpeli G, Lawlor RT, Scarpa A. 2013. DNA qualifica-
tion workflow for next generation sequencing of histopa-
thological samples. PLoS One 8:e62692.
42. Neufeld JD, Yu Z, Lam W, Mohn WW. 2004. Serial anal-
ysis of ribosomal sequence tags (SARST): a high-
throughput method for profiling complex microbial com-
munities. Environ Microbiol 6:131–144.
43. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D,
Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer
M. 2012. Ultra-high-throughput microbial community
analysis on the Illumina HiSeq and MiSeq platforms.
ISME J 6:1621–1624.
44. Bartram AK, Lynch MD, Stearns JC, Moreno-Hagelsieb
G, Neufeld JD. 2011. Generation of multimillion-
sequence 16S rRNA gene libraries from complex microbial
communities by assembling paired-end illumina reads. Appl
Environ Microbiol 77:3846–3852.
45. de Cárcer DA, Denman SE, McSweeney C, Morrison M.
2011. Strategy for modular tagged high-throughput ampli-
con sequencing. Appl Environ Microbol 77:6310–6312.
46. Bybee SM, Bracken-Grissom H, Haynes BD, Herman-
sen RA, Byers RL, Clement MJ, Udall JA, Wilcox ER,
Crandall KA. 2011. Targeted amplicon sequencing
(TAS): a scalable next-gen approach to multilocus, multi-
taxa phylogenetics. Genome Biol Evol 3:1312–1323.
47. Miller CS, Handley KM, Wrighton KC, Frischkorn KR,
Thomas BC, Banfield JF. 2013. Short-read assembly of
full-length 16S amplicons reveals bacterial diversity in sub-
surface sediments. PLoS One 8:e56018.
48. Piñol J, San Andrés V, Clare E, Mir G, Symondson W.
2014. A pragmatic approach to the analysis of diets of gen-
eralist predators: the use of next-generation sequencing
with no blocking probes. Mol Ecol Resour 14:18–26.
49. Pearce MM, Hilt EE, Rosenfeld AB, Zilliox MJ, Thomas-
White K, Fok C, Kliethermes S, Schreckenberger PC,
Brubaker L, Gai X. 2014. The female urinary microbiome:
a comparison of women with and without urgency urinary
incontinence. MBio 5:e01283–e01214.
50. Menke S, Wasimuddin MM, Melzheimer J, Mfune JK,
Heinrich S, Thalwitzer S, Wachter B, Sommer S. 2014.
Oligotyping reveals differences between gut microbiomes
of free-ranging sympatric Namibian carnivores (Acinonyx
jubatus, Canis mesomelas) on a bacterial species-like level.
Frontiers Microbiol 5:526.
51. Burke C, Darling AE. 2014. Resolving microbial microdi-
versity with high accuracy full length 16S rRNA Illumina
sequencing. bioRxiv. 010967.
52. Mao D-P, Zhou Q, Chen C-Y, Quan Z-X. 2012. Coverage
evaluation of universal bacterial primers using the metage-
nomic datasets. BMC Microbiol 12:66.
53. Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C,
Horn M, Glöckner FO. 2013. Evaluation of general 16S
ribosomal RNA gene PCR primers for classical and next-
generation sequencing-based diversity studies. Nucleic
Acids Res 41:e1.
54. Youssef N, Sheik CS, Krumholz LR, Najar FZ, Roe BA,
Elshahed MS. 2009. Comparison of species richness esti-
mates obtained using nearly complete fragments and simu-
lated pyrosequencing-generated fragments in 16S rRNA
gene-based environmental surveys. Appl Environ Microbol
75:5227–5236.
55. Engelbrektson A, Kunin V, Wrighton KC, Zvenigorod-
sky N, Chen F, Ochman H, Hugenholtz P. 2010. Experi-
mental factors affecting PCR-based estimates of microbial
species richness and evenness. ISME J 4:642–647.
56. Schloss PD, Gevers D, Westcott SL. 2011. Reducing the
effects of PCR amplification and sequencing artifacts on
16S rRNA-based studies. PloS one 6:e27310.
57. Lane DJ. 1991. 16S/23S rRNA sequencing, p. 115–147. In
Stackebrandt E, Goodfellow M (eds), Nucleic acid techni-
ques. John Wiley & Sons, Chichester, UK.
58. Muyzer G, De Waal EC, Uitterlinden AG. 1993. Profil-
ing of complex microbial populations by denaturing gra-
dient gel electrophoresis analysis of polymerase chain
reaction-amplified genes coding for 16S rRNA. Appl Envi-
ron Microbol 59:695–700.
59. Vainio EJ, Hantula J. 2000. Direct analysis of wood-
inhabiting fungi using denaturing gradient gel electro-
phoresis of amplified ribosomal DNA. Mycol Res 104:
927–936.
60. Watanabe T, Asakawa S, Nakamura A, Nagaoka K,
Kimura M. 2004. DGGE method for analyzing 16S
rDNA of methanogenic archaeal community in paddy field
soil. FEMS Microbiol Lett 232:153–163.
61. Kittelmann S, Seedorf H, Walters WA, Clemente JC,
Knight R, Gordon JI, Janssen PH. 2013. Simultaneous
amplicon sequencing to explore co-occurrence patterns of
bacterial, archaeal and eukaryotic microorganisms in
rumen microbial communities. PLoS One 8:e47879.
62. Skillman LC, Evans PN, Naylor GE, Morvan B, Jarvis
GN, Joblin KN. 2004. 16S ribosomal DNA-directed
PCR primers for ruminal methanogens and identification
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-23
of methanogens colonising young lambs. Anaerobe 10:
277–285.
63. Gardes M, Bruns TD. 1993. ITS primers with enhanced
specificity for basidiomycetes—application to the identifi-
cation of mycorrhizae and rusts. Mol Ecol 2:113–118.
64. White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification
and direct sequencing of fungal ribosomal RNA genes for
phylogenetics. PCR Prot Meth Appl 18:315–322.
65. Amaral-Zettler LA, McCliment EA, Ducklow HW, Huse
SM. 2009. A method for studying protistan diversity using
massively parallel sequencing of V9 hypervariable regions
of small-subunit ribosomal RNA genes. PLoS One 4:e6372.
66. Smith DP, Peay KG. 2014. Sequence depth, not PCR rep-
lication, improves ecological inference from next genera-
tion DNA sequencing. PLoS One 9:e90234.
67. Edwards JE, Huws SA, Kim EJ, Kingston-Smith AH.
2007. Characterization of the dynamics of initial bacterial
colonization of nonconserved forage in the bovine rumen.
FEMS Microbiol Ecol 62:323–335.
68. Green SJ, Minz D. 2005. Suicide polymerase endonuclease
restriction, a novel technique for enhancing PCR amplifi-
cation of minor DNA templates. Appl Environ Microbiol 71:
4721–4727.
69. Vestheim H, Jarman SN. 2008. Blocking primers to
enhance PCR amplification of rare sequences in mixed
samples–a case study on prey DNA in Antarctic krill stom-
achs. Fronti Zool 5:12.
70. Liles MR, Manske BF, Bintrim SB, Handelsman J, Good-
man RM. 2003. A census of rRNA genes and linked
genomic sequences within a soil metagenomic library.
Appl Environ Microbiol 69:2684–2691.
71. Human Microbiome Project Consortium. 2012. Struc-
ture, function and diversity of the healthy human micro-
biome. Nature. 486:207–214.
72. Peterson J, Garges S, Giovanni M, McInnes P, Wang L,
Schloss JA, Bonazzi V, McEwen JE, Wetterstrand KA,
Deal C. 2009. The NIH human microbiome project.
Genome Res 19:2317–2323.
73. Turnbaugh PJ, Ley RE, Hamady M, Fraser-Liggett C,
Knight R, Gordon JI. 2007. The human microbiome proj-
ect: exploring the microbial part of ourselves in a changing
world. Nature 449:804.
74. Mosher JJ, Bowman B, Bernberg EL, Shevchenko O, Kan
J, Korlach J, Kaplan LA. 2014. Improved performance of
the PacBio SMRT technology for 16S rDNA sequencing.
J Microbiol Meth 104:59–60.
75. Fichot EB, Norman RS. 2013. Microbial phylogenetic
profiling with the Pacific Biosciences sequencing platform.
Microbiome 1:10.
76. Kozich JJ, Westcott SL, Baxter NT, Highlander SK,
Schloss PD. 2013. Development of a dual-index sequenc-
ing strategy and curation pipeline for analyzing amplicon
sequence data on the MiSeq Illumina sequencing platform.
Appl Environ Microbiol 79:5112–5120.
77. Quince C, Lanzen A, Davenport RJ, Turnbaugh PJ.
2011. Removing noise from pyrosequenced amplicons.
BMC Bioinform 12:38.
78. Lozupone C, Knight R. 2005. UniFrac: a new phylogenetic
method for comparing microbial communities. Appl Envi-
ron Microbiol 71:8228–8235.
79. Huse SM, Welch DM, Morrison HG, Sogin ML. 2010.
Ironing out the wrinkles in the rare biosphere through im-
proved OTU clustering. Environ Microbiol 12:1889–1898.
80. Kunin V, Engelbrektson A, Ochman H, Hugenholtz P.
2010. Wrinkles in the rare biosphere: pyrosequencing
errors can lead to artificial inflation of diversity estimates.
Environ Microbiol 12:118–123.
81. Quince C, Lanzén A, Curtis TP, Davenport RJ, Hall N,
Head IM, Read LF, Sloan WT. 2009. Accurate
determination of microbial diversity from 454 pyrose-
quencing data. Nat Meth 6:639.
82. Faith JJ, Guruge JL, Charbonneau M, Subramanian S,
Seedorf H, Goodman AL, Clemente JC, Knight R, Heath
AC, Leibel RL. 2013. The long-term stability of the human
gut microbiota. Science 341:1237439.
83. Lundberg DS, Yourstone S, Mieczkowski P, Jones CD,
Dangl JL. 2013. Practical innovations for high-throughput
amplicon sequencing. Nat Meth 10:999–1002.
84. Kreader CA. 1996. Relief of amplification inhibition in
PCR with bovine serum albumin or T4 gene 32 protein.
Appl Environ Microbiol 62:1102–1106.
85. Arbeli Z, Fuentes CL. 2007. Improved purification and
PCR amplification of DNA from environmental samples.
FEMS Microbiol Lett 272:269–275.
86. Opel KL, Chung D, McCord BR. 2010. A study of PCR
inhibition mechanisms using real time PCR. J Forensic Sci
55:25–33.
87. Kermekchiev MB, Kirilova LI, Vail EE, Barnes WM.
2009. Mutants of Taq DNA polymerase resistant to PCR
inhibitors allow DNA amplification from whole blood
and crude soil samples. Nucleic Acids Res 37:e40.
88. Owczarzy R, Tataurov AV, Wu Y, Manthey JA, McQuis-
ten KA, Almabrazi HG, Pedersen KF, Lin Y, Garretson J,
McEntaggart NO. 2008. IDT SciTools: a suite for analysis
and design of nucleic acid oligomers. Nucleic Acids Res 36:
W163–W169.
89. Qiu X, Wu L, Huang H, McDonel PE, Palumbo AV,
Tiedje JM, Zhou J. 2001. Evaluation of PCR-generated
chimeras, mutations, and heteroduplexes with 16S rRNA
gene-based cloning. Appl Environ Microbiol 67:880–887.
90. Chandler D, Fredrickson J, Brockman F. 1997. Effect of
PCR template concentration on the composition and dis-
tribution of total community 16S rDNA clone libraries.
Mol Ecol 6:475–482.
91. Polz MF, Cavanaugh CM. 1998. Bias in template-to-prod-
uct ratios in multitemplate PCR. Appl Environ Microbiol 64:
3724–3730.
92. Orgiazzi A, Lumini E, Nilsson RH, Girlanda M, Vizzini
A, Bonfante P, Bianciotto V. 2012. Unravelling soil fungal
communities from different Mediterranean land-use back-
grounds. PLoS One 7:e34847.
93. Medinger R, Nolte V, Pandey RV, Jost S, Ottenwaelder
B, Schloetterer C, Boenigk J. 2010. Diversity in a hidden
world: potential and limitation of next-generation
sequencing for surveys of molecular diversity of eukaryotic
microorganisms. Mol Ecol 19:32–40.
94. Zhou J, Wu L, Deng Y, Zhi X, Jiang Y-H, Tu Q, Xie J,
Van Nostrand JD, He Z, Yang Y. 2011. Reproducibility
and quantitation of amplicon sequencing-based detection.
ISME J 5:1303–1313.
95. Ge Y, Schimel JP, Holden PA. 2014. Analysis of
run-to-run variation of bar-coded pyrosequencing for eval-
uating bacterial community shifts and individual taxa
dynamics. PLoS One 9:e99414.
96. Liu J, Song H, Liu D, Zuo T, Lu F, Zhuang H, Gao F.
2014. Extensive recombination due to heteroduplexes gen-
erates large amounts of artificial gene fragments during
PCR. PLoS One 9:e106658.
97. Momozawa Y, Deffontaine V, Louis E, Medrano JF.
2011. Characterization of bacteria in biopsies of colon
and stools by high throughput sequencing of the V2 region
of bacterial 16S rRNA gene in human. PLoS One 6:
e16952.
98. Sogin ML, Morrison HG, Huber JA, Mark Welch D,
Huse SM, Neal PR, Arrieta JM, Herndl GJ. 2006. Micro-
bial diversity in the deep sea and the underexplored
“rare biosphere”. Proc Natl Acad Sci USA 103:
12115–12120.
2.4.2-24 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
99. Pinto AJ, Raskin L. 2012. PCR biases distort bacterial and
archaeal community structure in pyrosequencing datasets.
PLoS One 7:e43093.
100. Quail MA, Kozarewa I, Smith F, Scally A, Stephens PJ,
Durbin R, Swerdlow H, Turner DJ. 2008. A large genome
center’s improvements to the Illumina sequencing system.
Nat Meth 5:1005–1010.
101. Rohland N, Reich D. 2012. Cost-effective, high-
throughput DNA sequencing libraries for multiplexed tar-
get capture. Genome Res 22:939–946.
102. Krueger F, Andrews SR, Osborne CS. 2011. Large scale
loss of data in low-diversity illumina sequencing libraries
can be recovered by deferred cluster calling. PLoS One 6:
e16607.
103. Hummelen R, Fernandes AD, Macklaim JM, Dickson
RJ, Changalucha J, Gloor GB, Reid G. 2010. Deep
sequencing of the vaginal microbiota of women with
HIV. PLoS One 5:e12078.
104. Lange V, Böhme I, Hofmann J, Lang K, Sauter J, Schöne
B, Paul P, Albrecht V, Andreas JM, Baier DM. 2014.
Cost-efficient high-throughput HLA typing by MiSeq
amplicon sequencing. BMC Genomics 15:63.
105. Fujimoto M, Moyerbrailean GA, Noman S, Gizicki JP,
Ram ML, Green PA, Ram JL. 2014. Application of Ion
Torrent sequencing to the assessment of the effect of alkali
ballast water treatment on microbial community diversity.
PLoS One 9:e107534.
106. Tonge DP, Pashley CH, Gant TW. 2014. Amplicon-
based metagenomic analysis of mixed fungal samples using
proton release amplicon sequencing. PLoS One 9:e93849.
107. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K,
Bushman FD, Costello EK, Fierer N, Pena AG, Goodrich
JK, Gordon JI. 2010. QIIME allows analysis of high-
throughput community sequencing data. Nat Meth 7:
335–336.
108. Kuczynski J, Stombaugh J, Walters WA, González A,
Caporaso JG, Knight R. 2012. Using QIIME to analyze
16S rRNA gene sequences from microbial communities.
Curr Protoc Microbiol oUnit 1E.5.
109. Navas-Molina JA, Peralta-Sánchez JM, González A,
McMurdie PJ, Vázquez-Baeza Y, Xu Z, Ursell LK,
Lauber C, Zhou H, Song SJ. 2013. Advancing our under-
standing of the human microbiome using QIIME. Meth
Enzymol 531:371–444.
110. Masella AP, Bartram AK, Truszkowski JM, Brown DG,
Neufeld JD. 2012. PANDAseq: paired-end assembler for
illumina sequences. BMC Bioinformatics 13:31.
111. Zhang J, Kobert K, Flouri T, Stamatakis A. 2014. PEAR:
a fast and accurate Illumina paired-end read merger. Bioin-
formatics 30:614–620.
112. Whiteley AS, Jenkins S, Waite I, Kresoje N, Payne H,
Mullan B, Allcock R, O’Donnell A. 2012. Microbial
16S rRNA Ion Tag and community metagenome sequenc-
ing using the Ion Torrent (PGM) platform. J Microbiol Meth
91:80–88.
113. Milani C, Hevia A, Foroni E, Duranti S, Turroni F, Lugli
GA, Sanchez B, Martín R, Gueimonde M, van Sinderen
D. 2013. Assessing the fecal microbiota: an optimized ion
torrent 16S rRNA gene-based analysis protocol. PLoS
One 8:e68739.
114. Reeder J, Knight R. 2010. Rapid denoising of pyrose-
quencing amplicon data: exploiting the rank-abundance
distribution. Nat Meth 7:668.
115. Gaspar JM, Thomas WK. 2013. Assessing the consequen-
ces of denoising marker-based metagenomic data. PLoS
One 8:e60458.
116. Gaspar JM, Thomas WK. 2015. FlowClus: efficiently fil-
tering and denoising pyrosequenced amplicons. BMC Bio-
informatics 16:105.
117. Wang GC, Wang Y. 1996. The frequency of chimeric mol-
ecules as a consequence of PCR co-amplification of 16S
rRNA genes from different bacterial species. Microbiology
142:1107–1114.
118. Huber T, Faulkner G, Hugenholtz P. 2004. Bellerophon:
a program to detect chimeric sequences in multiple
sequence alignments. Bioinformatics 20:2317–2319.
119. Haas BJ, Gevers D, Earl AM, Feldgarden M, Ward DV,
Giannoukos G, Ciulla D, Tabbaa D, Highlander SK,
Sodergren E. 2011. Chimeric 16S rRNA sequence forma-
tion and detection in Sanger and 454-pyrosequenced PCR
amplicons. Genome Res 21:494–504.
120. Caporaso JG, Bittinger K, Bushman FD, DeSantis TZ,
Andersen GL, Knight R. 2010. PyNAST: a flexible tool
for aligning sequences to a template alignment. Bioinfor-
matics 26:266–267.
121. Ionescu D, Siebert C, Polerecky L, Munwes YY, Lott C,
Häusler S, Bižić-Ionescu M, Quast C, Peplies J, Glöckner
FO. 2012. Microbial and chemical characterization of
underwater fresh water springs in the Dead Sea. PLoS
One 7:e38319.
122. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie
EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL.
2006. Greengenes, a chimera-checked 16S rRNA gene
database and workbench compatible with ARB. Appl Envi-
ron Microbiol 72:5069–5072.
123. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer
T, Yarza P, Peplies J, Glöckner FO. 2012. The SILVA
ribosomal RNA gene database project: improved data proc-
essing and web-based tools. Nucleic Acids Res 41:
D590–D596.
124. Pruesse E, Peplies J, Glöckner FO. 2012. SINA: accurate
high-throughput multiple sequence alignment of riboso-
mal RNA genes. Bioinformatics 28:1823–1829.
125. Zinger L, Gobet A, Pommier T. 2012. Two decades of
describing the unseen majority of aquatic microbial diver-
sity. Mol Ecol 21:1878–1896.
126. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ,
Kulam-Syed-Mohideen A, McGarrell DM, Marsh T,
Garrity GM. 2009. The Ribosomal Database Project:
improved alignments and new tools for rRNA analysis.
Nucleic Acids Res 37:D141–D145.
127. Nawrocki EP, Eddy SR. 2007. Query-dependent banding
(QDB) for faster RNA similarity searches. PLoS Comp Biol
3:e56.
128. Salman V, Amann R, Shub DA, Schulz-Vogt HN.
2012. Multiple self-splicing introns in the 16S rRNA genes
of giant sulfur bacteria. Proc Nat Acad Sci 109:4203–4208.
129. Stackebrandt E, Goebel B. 1994. Taxonomic note: a place
for DNA-DNA reassociation and 16S rRNA sequence
analysis in the present species definition in bacteriology.
Int J Syst Bacteriol 44:846–849.
130. Acinas SG, Marcelino LA, Klepac-Ceraj V, Polz MF.
2004. Divergence and redundancy of 16S rRNA sequences
in genomes with multiple rrn operons. J Bacteriol 186:
2629–2635.
131. Adékambi T, Shinnick TM, Raoult D, Drancourt M.
2008. Complete rpoB gene sequencing as a suitable sup-
plement to DNA–DNA hybridization for bacterial species
and genus delineation. Int J Syst Evol Microbiol 58:
1807–1814.
132. Rosselló-Mora R, Amann R. 2001. The species concept
for prokaryotes. FEMS Microbiol Rev 25:39–67.
133. Claire Horner-Devine M, Leibold MA, Smith VH,
Bohannan BJ. 2003. Bacterial diversity patterns along a
gradient of primary productivity. Ecol Lett 6:613–622.
134. Konstantinidis KT, Ramette A, Tiedje JM. 2006. The
bacterial species definition in the genomic era. Phil
Trans R Soc B Biol Sci 361:1929–1940.
 2.4.2. Microbial Community Analysis Using High-Throughput Amplicon Sequencing ▪ 2.4.2-25
135. Rideout JR, He Y, Navas-Molina JA, Walters WA,
Ursell LK, Gibbons SM, Chase J, McDonald D, Gonza-
lez A, Robbins-Pianka A. 2014. Subsampled open-
reference clustering creates consistent, comprehensive
OTU definitions and scales to billions of sequences. PeerJ
2:e545.
136. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann
M, Hollister EB, Lesniewski RA, Oakley BB, Parks DH,
Robinson CJ. 2009. Introducing mothur: open-source,
platform-independent, community-supported software for
describing and comparing microbial communities. Appl
Environ Microbiol 75:7537–7541.
137. Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun
Y, Brown CT, Porras-Alfaro A, Kuske CR, Tiedje JM.
2013. Ribosomal Database Project: data and tools for
high throughput rRNA analysis. Nucleic Acids Res 42:
D633–D642.
138. McMurdie PJ, Holmes S. 2013. phyloseq: an R package for
reproducible interactive analysis and graphics of micro-
biome census data. PLoS One 8:e61217.
139. Lynch MD, Masella AP, Hall MW, Bartram AK, Neufeld
JD. 2013. AXIOME: automated exploration of microbial
diversity. GigaSci 2:3.
140. Meyer F, Paarmann D, D’Souza M, Olson R, Glass EM,
Kubal M, Paczian T, Rodriguez A, Stevens R, Wilke A.
2008. The metagenomics RAST server–a public resource
for the automatic phylogenetic and functional analysis of
metagenomes. BMC Bioinformatics 9:386.
141. Huson DH, Mitra S, Ruscheweyh H-J, Weber N, Schus-
ter SC. 2011. Integrative analysis of environmental
sequences using MEGAN4. Genome Res 21:1552–1560.
142. Parks DH, Tyson GW, Hugenholtz P, Beiko RG. 2014.
STAMP: statistical analysis of taxonomic and functional
profiles. Bioinformatics 30:3123–3124.
143. Braak Ct, Šmilauer P. 2002. CANOCO reference manual
and CanoDraw for Windows user’s guide: software for canon-
ical community ordination (version 4.5). Section on Permuta-
tion Methods. Microcomputer Power, Ithaca, NY.
144. Chen J, Bittinger K, Charlson ES, Hoffmann C, Lewis J,
Wu GD, Collman RG, Bushman FD, Li H. 2012. Associ-
ating microbiome composition with environmental covari-
ates using generalized UniFrac distances. Bioinformatics 28:
2106–2113.
145. Hamady M, Lozupone C, Knight R. 2009. Fast UniFrac:
facilitating high-throughput phylogenetic analyses of
microbial communities including analysis of pyrosequenc-
ing and PhyloChip data. ISME J 4:17–27.
146. Matsen IVFA, Evans SN. 2013. Edge principal compo-
nents and squash clustering: using the special structure
of phylogenetic placement data for sample comparison.
PLoS One 8:e56859.
147. Schloss PD, Handelsman J. 2005. Introducing DOTUR, a
computer program for defining operational taxonomic
units and estimating species richness. Appl Environ Micro-
biol 71:1501–1506.
148. McDonald D, Clemente JC, Kuczynski J, Rideout JR,
Stombaugh J, Wendel D, Wilke A, Huse S, Hufnagle J,
Meyer F. 2012. The Biological Observation Matrix
(BIOM) format or: how I learned to stop worrying and
love the ome-ome. GigaSci 1:7.
149. Hill TC, Walsh KA, Harris JA, Moffett BF. 2003. Using
ecological diversity measures with bacterial communities.
FEMS Microbiol Ecol 43:1–11.
150. Jost L. 2007. Partitioning diversity into independent alpha
and beta components. Ecology 88:2427–2439.
151. Gihring TM, Green SJ, Schadt CW. 2012. Massively par-
allel rRNA gene sequencing exacerbates the potential for
biased community diversity comparisons due to variable
library sizes. Environ Microbiol 14:285–290.
152. McMurdie PJ, Holmes S. 2014. Waste not, want not: why
rarefying microbiome data is inadmissible. PLoS Comput
Biol 10:e1003531.
153. Paulson JN, Stine OC, Bravo HC, Pop M. 2013. Differen-
tial abundance analysis for microbial marker-gene surveys.
Nat Meth 10:1200–1202.
154. Love MI, Huber W, Anders S. 2014. Moderated estima-
tion of fold change and dispersion for RNA-Seq data
with DESeq2. Genome Biol 15:550
155. Anders S, Huber W. 2010. Differential expression analysis
for sequence count data. Genome Biol 11:R106.
156. Kuczynski J, Liu Z, Lozupone C, Mcdonald D, Fierer N,
Knight R. 2010. Microbial community resemblance meth-
ods differ in their ability to detect biologically relevant pat-
terns. Nat Meth 7:813–819.
157. Buttigieg PL, Ramette A. 2014. A guide to statistical anal-
ysis in microbial ecology: a community-focused, living
review of multivariate data analyses. FEMS Microbiol Ecol
90:543–550.
158. Clarke KR, Somerfield PJ, Chapman MG. 2006. On
resemblance measures for ecological studies, including
taxonomic dissimilarities and a zero-adjusted Bray-Curtis
coefficient for denuded assemblages. JExp Mar Biol Ecol
330:55–80.
159. Bray JR, Curtis JT. 1957. An ordination of the upland for-
est communities of southern Wisconsin. Ecol Monographs
27:325–349.
160. Clarke K, Warwick R. 2001. Changes in Marine Commun-
ities: An Approach to Statistical Analysis and Interpretation.
PRIMER-E, Plymouth.
161. Clarke K. 1993. Non-parametric multivariate analyses of
changes in community structure. Aus J Ecol 18:117–143.
162. Ramette A. 2007. Multivariate analyses in microbial ecol-
ogy. FEMS Microbiol Ecol 62:142–160.
163. Faith DP, Minchin PR, Belbin L, Water D, Resources L,
Box GPO. 1987. Compositional dissimilarity as a robust
measure of ecological distance. Vegetatio 69:57–68.
164. Staley JT. 2009. The phylogenomic species concept for
bacteria and archaea. Microbe August.
165. Cole JR, Konstantinidis K, Farris RJ, Tiedje JM. 2010.
Microbial diversity and phylogeny. Environ Mol Microbiol
17:1339–1346.
166. Parks DH, Beiko RG. 2013. Measures of phylogenetic dif-
ferentiation provide robust and complementary insights
into microbial communities. ISME J 7:173–183.
167. Mielke PJ. 1984. Meteorological applications of permuta-
tion techniques based on distance functions, p. 813–830.
In Krishnaiah P, Sen P (eds), Handbook of Statistics, volume
4: Nonparametric Methods. Elsevier, Amsterdam.
168. Anderson MJ. 2001. A new method for non-parametric
multivariate analysis of variance. Aus Ecol 26:32–46.
169. Warton DI, Wright ST, Wang Y. 2012. Distance-based
multivariate analyses confound location and dispersion
effects. Meth Ecol Evol 3:89–101.
170. Kembel SW, Wu M, Eisen JA, Green JL. 2012. Incorpo-
rating 16S gene copy number information improves esti-
mates of microbial diversity and abundance. PLoS
Comput Biol 8:e1002743.
171. Stoddard SF, Smith BJ, Hein R, Roller BR, Schmidt TM.
2015. rrnDB: improved tools for interpreting rRNA
gene abundance in bacteria and archaea and a new founda-
tion for future development. Nucl Acids Res 43:
D593–D598.
172. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L,
Garrett WS, Huttenhower C. 2011. Metagenomic bio-
marker discovery and explanation. Genome Biol 12:R60.
173. Tickle TL, Segata N, Waldron L, Weingart U, Hutten-
hower C. 2013. Two-stage microbial community experi-
mental design. ISME J 7:2330–2339.
2.4.2-26 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
174. Haberman Y, Tickle TL, Dexheimer PJ, Kim M-O, Tang
D, Karns R, Baldassano RN, Noe JD, Rosh J, Markowitz
J. 2014. Pediatric Crohn disease patients exhibit specific
ileal transcriptome and microbiome signature. J Clin Invest
124:3617.
175. Dufrêne M, Legendre P. 1997. Species assemblages and
indicator species: the need for a flexible asymmetrical
approach. Ecol Monographs 67:345–366.
176. Roberts DW. 2007. “labdsv: Ordination and multivariate
analysis for ecology.” R package version 1, no. 1.
177. Galimanas V, Hall MW, Singh N, Lynch MDJ,
Goldberg M, Tenenbaum H, Cvitkovitch DG, Neufeld
JD, Senadheera DB. 2014. Bacterial community com-
position of chronic periodontitis and novel oral sampl-
ing sites for detecting disease indicators. Microbiome
2:32.
178. Shade A, Jones SE, Caporaso JG, Handelsman J,
Knight R, Fierer N, Gilbert JA. 2014. Conditio-
nally rare taxa disproportionately contribute to temporal
changes in microbial diversity. mBio 5:e01371–
01314.
179. Hoffmann C, Dollive S, Grunberg S, Chen J, Li H, Wu
GD, Lewis JD, Bushman FD. 2013. Archaea and fungi
of the human gut microbiome: correlations with diet and
bacterial residents. PLoS One 8:e66019.
180. Vázquez-Baeza Y, Pirrung M, Gonzalez A, Knight R.
2013. EMPeror: a tool for visualizing high-throughput
microbial community data. Structure 585:20.
181. Kelder T, Conklin BR, Evelo CT, Pico AR. 2010.
Finding the right questions: exploratory pathway analysis
to enhance biological discovery in large datasets. PLoS
Biol 8:5.
182. Bartram AK, Jiang X, Lynch MD, Masella AP, Nicol
GW, Dushoff J, Neufeld JD. 2014. Exploring links
between pH and bacterial community composition in soils
from the Craibstone Experimental Farm. FEMS Microbiol
Ecol 87:403–415.
183. Muegge BD, Kuczynski J, Knights D, Clemente JC, Gon-
zález A, Fontana L, Henrissat B, Knight R, Gordon JI.
2011. Diet drives convergence in gut microbiome functions
across mammalian phylogeny and within humans. Science
332:970–974.
184. Ley RE, Hamady M, Lozupone C, Turnbaugh PJ, Ramey
RR, Bircher JS, Schlegel ML, Tucker TA, Schrenzel
MD, Knight R. 2008. Evolution of mammals and their
gut microbes. Science 320:1647–1651.
185. Faust K, Sathirapongsasuti JF, Izard J, Segata N, Gevers
D, Raes J, Huttenhower C. 2012. Microbial co-occurrence
relationships in the human microbiome. PLoS Comp Biol 8:
e1002606.
186. Kelder T, Stroeve JHM, Bijlsma S, Radonjic M, Roeselers
G. 2014. Correlation network analysis reveals relationships
between diet-induced changes in human gut microbiota
and metabolic health. Nut Diab 4:e122.
187. Greenblum S, Turnbaugh PJ, Borenstein E. 2012. Meta-
genomic systems biology of the human gut microbiome
reveals topological shifts associated with obesity and
inflammatory bowel disease. Proc Natl Acad Sci USA
109:594–599.
188. McHardy IH, Goudarzi M, Tong M, Ruegger PM,
Schwager E, Weger JR, Graeber TG, Sonnenburg JL,
Horvath S, Huttenhower C. 2013. Integrative analysis of
the microbiome and metabolome of the human intestinal
mucosal surface reveals exquisite inter-relationships. Micro-
biome 1:17.
189. Ruan Q, Dutta D, Schwalbach MS, Steele JA, Fuhr-
man JA, Sun F. 2006. Local similarity analysis reveals
unique associations among marine bacterioplankton
species and environmental factors. Bioinformatics 22:
2532–2538.
190. Yilmaz P, Kottmann R, Field D, Knight R, Cole JR,
Amaral-Zettler L, Gilbert JA, Karsch-Mizrachi I, John-
ston A, Cochrane G. 2011. Minimum information about
a marker gene sequence (MIMARKS) and minimum infor-
mation about any (x) sequence (MIxS) specifications. Nat
Biotechnol 29:415–420.
191. Frank JA, Claudia IR, Shobha S, Jon SW, Brenda AW,
Gary JO. 2008. Critical evaluation of two primers com-
monly used for amplification of bacterial 16S rRNA genes.
Appl Environ Microbiol 74:2461–2470.
 Functional Metagenomics: Procedures and Progress
LAURA S. MORRIS AND JULIAN R. MARCHESI
2.4.3
Cultivation-independent molecular methods have been
essential for the field of environmental microbiology over
the past several decades. Using such approaches, the compo-
sition, diversity, structure, activity, and genetic makeup of
microbial communities have been explored. Using shotgun
metagenomics (SM) (and associated bioinformatics analy-
ses), cultivation-independent analyses of microbial com-
munities have yielded de novo assembly of entire microbial
genomes from mixed environmental samples (e.g., 1) and
the identification of a tremendous diversity of sequences
from within a single family of functional genes (2). Sequence-
driven metagenomics has become routine, driven by the
reduced costs of sequencing and improvements in automated
annotation pipelines (e.g., 3). SM has become a baseline
technology for microbiological investigations into microbial
ecosystems, but this approach has a number of limitations
(Table 1). First, sequencing metagenomics relies entirely on
identification of comparison of sequences with previously
annotated sequences uploaded to public databases. Sequen-
ces with low similarity to other sequences in public data-
bases are classified as “unidentified”; as a result, novel genes
are often disregarded as ambiguous data or unknowns. In
some instances these unknowns can account for >70% of
the data generated (4). This is even the case for single-
organism genomes which are much simpler systems than
metagenomes. For example, the most well-studied microor-
ganism E. coli K12-W3110 and the considerably smaller
genome of Mycoplasma genitalium still have around 10%
unannotated genes. The continued implementation of large-
scale metagenomics projects such as the Human Microbiome
Project (5, 6) has considerably aided the functional annota-
tion of many metagenome-derived genes, but the level of
annotation remains only at ∼70% compared to the human
genome, which is at 82% (7). The level of annotation is
even lower in environments that have received less attention
and had consequently fewer relevant microbial genomes
sequenced. For example, the soil and ocean microbiomes
are estimated to have >55% unannotated genes, and this per-
centage increases in more obscure environments such as the
cow rumen (∼70% unknown genes) (7). Massive sequencing
efforts, such as the Earth Microbiome Project (8) shotgun
sequencing studies conducted by Venter and colleagues (2),
generate billions of bases of sequence data that are uploaded
onto public databases. Therefore the level of unknown
genetic information may continue to decrease as more
doi:10.1128/9781555818821.ch2.4.3
2.4.3-1
projects elucidate and define genes from such data. However,
currently the resources and tools available to aid in the func-
tional annotation process are insufficient to keep up with the
demands of the huge amount of metagenomic data being pro-
duced (7).
Another significantly debilitating aspect of sequencing
metagenomics is that entire gene sequences are rarely
revealed using the shotgun sequencing approach, leaving an
aspect of ambiguity surrounding deduced gene products and
preventing comprehensive biochemical and functional anno-
tation. Even for gene sequences with clear analogues in gene
databases, annotation is only putative and based on sequence
similarity. Validation of gene annotation requires additional
experimentation, including expression of genes in a suitable
host. Conversely, sequence-based screening methods imple-
ment gene-targeting methods to mine genomes for novel
genes, such as those using PCR or DNA probe hybridization.
Here primers or probes are designed from conserved regions of
multiple alignments of the sequences encoding the particular
gene of interest (9) and are used to amplify genes of interest
from the metagenomic sample. Genome walking techniques
such as inverse PCR can then be implemented if needed to
capture the entire functional gene (10). However, this meth-
odology is restricted in its requirement of a priori knowledge of
sequence data as the basis for designing primers and probes
(11). Very rarely can novel, nonhomologous genes be identi-
fied this way.
To circumvent the limitations of shotgun metagenome
sequencing and directed PCR amplification a form of pheno-
typic screening of metagenomic DNA is required to charac-
terize their functional capacity. Functional metagenomics
(FM) is a methodological strategy to perform such metage-
nomic DNA screening and relies on screening clones harbor-
ing random fragments of metagenomic DNA (mgDNA).
Clones are transformed or transfected with plasmids contain-
ing mgDNA. Individual clones are screened for chosen func-
tional capabilities; this screening relies on the heterologous
gene expression of the randomly inserted DNA and a method
for high-throughput screening of clones. For example, screen-
ings are often conducted in colorimetric media, allowing for
rapid detection of enzymatic activity (e.g., 5-bromo-4-
chloro-3-indolyl phosphate [BCIP; phosphatases] or X-gal
[galactosidases]). Alternatively, though in lower throughput,
halo formation around colonies can be used to indicate deg-
radation of the target compound (e.g., halo formation on
2.4.3-2 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

TABLE 1 Advantages and disadvantages of the two main types of metagenomic approaches
Function based Sequence based
Advantages Screen large contiguous fragments of DNA Qualitative/quantitative analysis of whole metagenome
Provides functional annotation of novel genes Detects toxic genes
Provides DNA of interest in a usable construct Library preparation and sequencing is rapid and
Provides genomic context relatively inexpensive
Disadvantages Potentially no expression of toxic genes Annotation is not definitive
Mismatch between gene and host can prevent proper Genomic context is lost
expression Computationally expensive
Storage issues for the number of libraries needed for
coverage
“Wet lab” labour is extremely intensive

lactose-free skimmed milk agar is indicative of protease
activity; 12). In some cases, colony pigmentation can be diag-
nostic for clones with the ability to degrade a component
of the media (13). Conceptually, the best phenotypic screens
allow for growth in the select medium only with appropriate
exogenous genes; for example, screening for antibiotic resist-
ance can be performed in media containing the antibiotic of
interest. In the absence of resistance genes, clones cannot
grow and are removed from the functional screen. Similarly,
genes and gene pathways involved in antibiotic production
can be identified by culturing clones on an overlay of a
particular target organism and looking for clearing zones.
Screening methods are becoming increasingly sophisticated
and advanced, enabling the discovery of an ever more
diverse range of molecules (see Table 2 for a more compre-
hensive list).
FM provides a means of accessing phenotypic characteris-
tics of “the uncultivated majority” and is currently the only
method that facilitates the discovery of novel genes (23).
For example, using a functional metagenomic approach, Li
and colleagues (6) isolated β-lactamase genes from a remote
Alaskan soil site that had no known exposure to exogenous
antibiotics and hypothesized that antibiotic resistance mech-
anisms were not driven by human-induced selective pressure.
Using the FM approach, Allen et al. (18) were also able to
identify the first example of a bifunctional β-lactamase
enzyme in that a class D β-lactamase gene and a class C
β-lactamase gene were linked on the same open reading
frame encoding a protein twice the size of most reported
β-lactamases (19).
In an era where antibiotic resistance is a major concern,
novel drug discovery is of interest to most stakeholders. FM
screening has proved to be a promising approach for new
drug discovery, as exemplified by Courtois et al. (24), who
created and screened a soil metagenomic library in E. coli
and Streptomyces lividans to identify antibiotic synthesis genes.
They discovered novel polyketide synthase genes, thus paving
the way for use of FM as a means of identifying antimicrobial
genes from natural environments (24).
Understanding the functions of microorganisms leads to
understanding how selection pressures drive adaptation of
bacteria to the particular environment they inhabit. Research
by Jones and colleagues (25) used FM to screen the human
gut metagenome for bile salt hydrolases (BSHs). Functional
BSHs were identified in all major divisions of bacteria, and
it was suggested that BSH activity is a conserved adaptation
to life in the human gut with significant genetic redundancy
(25). This research emphasized how FM can be used to iden-
tify novel genes and understand functional adaptations rele-
vant for microbial survival and success in particular
environments.
Although FM approaches are extremely powerful, they are
far more labor-intensive than SM sequencing approaches,
and the frequency of positive results (such as an active clone)

TABLE 2 List of common screens used in functional metagenomics

Function Screen Reference
DNAse Agar supplemented with DNA and methyl green dye. DNAse activity is indicated by 14
formation of halos surrounding colonies.
Lipases Agar supplemented with 1% tributyrin. Lipase activity is indicated by clear halos forming 15
around colonies due to hydrolysis of tributyrin.
Osmotolerance Agar supplemented with NaCl of varying concentrations. 16
Galactosidase 5-Bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) supplemented media, 17
galactosidase activity liberates 5-bromo-4-chloro-3-hydroyindole, which forms a blue dye.
Proteases Lactose-free skimmed milk agar. Protease activity is indicated by formation of clear halos 12
around a colony due to degradation of protein in the milk powder.
Antibiotic resistance Agar supplemented with varying concentrations of an antibiotic of interest. 18, 19
Esterases Agar supplemented with 1% tributyrin or α-naphthyl ester substrates where hydrolysis activity 20
is indicated by appearance of a brown color around colonies.
Amidases Detection of a fluorescent dye after amidase activity hydrolyzes N-benzyl-2-chloroacetamine. 21
Addition of 4-nitro-7-chloro-benzo-2-oxa-1,3-diazole reacts with the benzylamine product
of amide hydrolysis to yield fluorescence.
Quorum sensing METREX system. 22
 2.4.3. Functional Metagenomics: Procedures and Progress ▪ 2.4.3-3

is quite low. However, the development of high-throughput
sequencing of metagenomic clones offers access to huge
amounts of genetic information for which there are increas-
ingly sophisticated bioinformatic tools for analysis. Similarly,
the development of high-throughput wet-lab instrumenta-
tion, particularly colony-picking robots, which identify,
pick, and duplicate clones in multiwell plates, enables
researchers to screen very large numbers of clones. Such high-
throughput systems reduce the cost and labor associated with
FM screening, thereby increasing the likelihood that a
broader spectrum of laboratories will perform such screens.
The increased automation and novel screening strategies
are essential to ensure that large FM screens are performed,
generating adequate coverage of complex microbial commun-
ities. Examples of such high-throughput screening strategies
are discussed further on (see section on Library Screening).
One of the major technical challenges associated with FM
is acquiring adequate depth of screening from environments
with high microbial and genetic diversity (26). In many cases,
libraries of tens of thousands of clones must be screened to
identify clones with a positive phenotype. For a successful
FM project, the number of clones to be screened is dependent
on the screened function, the size of the DNA fragments, and
the number of genes required for expression (see Table 3 for
examples). Overall, a number of steps must be undertaken
for a successful metagenomic screen; large DNA fragments
must be successfully ligated into a vector, transformed into
a surrogate host, and then translated into the fully function-
ing protein or product which must demonstrate a function
(23). The methodologies used for FM are discussed next,
including suggestions as to how to adapt these methods to
remove certain limitations to increase the efficiency of FM
screens.
DNA EXTRACTION
The extraction of total DNA is relevant to both sequencing
and functional metagenomics. As highlighted in a review
by Ekkers et al. (27), DNA extraction protocols ideally should
yield genomic DNA of high yield, high quality/purity, large
fragment size, and fair or unbiased representation of the
microbial community in question. Large fragment size is crit-
ical for FM, while SM sequencing can tolerate more frag-
mented DNA. Unfortunately, the enhancement of one of
these criteria often has a deleterious effect on one of the other
factors (27). In addition, some bacterial taxa, such as Gram-
positive organisms, require harsher lysis methods to recover
high yields of DNA, although this can result in heavily frag-
mented DNA.
Genomic DNA can be extracted from any sample type for
use in FM analyses. Furthermore, cell lysis, and consequently
the DNA extraction of the entire microbial community, can
be carried out either directly from the environmental sample
in question or by indirect means where cells are first isolated
from the sample (for example, by Nycodenz density gradient;
28) and subsequently subject to cell lysis and DNA extrac-
tion. Cell lysis may be carried out similarly for both direct
and indirect methods and can include mechanical (e.g.,
bead beating), enzymatic, or chemical-based cell breakage.
Although bead beating can yield heavily sheared DNA,
which can be challenging to generate “sticky ends” for subse-
quent ligation, blunt-end cloning is usually more successful
when working with heavily sheared DNA (29). A number
of studies have been conducted to determine optimum
DNA extraction methods for different environments. Salo-
nen and colleagues compared four methods for isolating
DNA from fecal samples for metagenomic studies of human
gut microbiota. The four different methods entailed differen-
tial centrifugation and enzymatic lysis; the Promega Wizard
DNA extraction kit relying purely on enzymatic lysis;
repeated bead beating rounds with SDS, salt, and EDTA;
and last, the QiaAmp DNA stool mini kit coupled with a pre-
ceding bead beating step (30). They found the extraction
methods to give highly variable DNA yields with as much
as 35-fold variations among the methods. However, their
results demonstrated that optimal DNA recovery from human
fecal samples was obtained by coupling mechanical DNA
extraction by repeated bead beating followed by precipitation

TABLE 3 List of screens used on a functional metagenomic fosmid library from the
distal human gut and examples of hit rate

Clones
Screen +ve Hit/Mb
screened
Proteasesa 120,576 231 1/20.8
Lipases/esterases 83,712 240 1/14
Cellulase 46,080 64 1/28.8
Oxalate degradation 69,080 40 1/69.1
Acid pH tolerance 46,080 135 1/13.7
Bile salt hydrolase 89,856 142 1/25.3
DNase 46,080 175 1/10.5
Osmotolerance 23,040 52 1/17.2
Antibiotic resistance 23,040 45 1/20
Xylanase 46,080 0 –
Quorum sensing 46,080 0 –
Chlorhexidine resistance 36,864 0 –
Bacteriocin production 34,560 0 –
Cholesterol degradation 40,000 21 1/76
Phosphatases 18,000 32 1.40
a
All 231 positive clones were β-galactosidases although the screen was designed for proteases (16).
2.4.3-4 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
(30). Purohit and Singh (31) assessed “soft” lysis (SDS and
enzymatic extraction), “harsh” lysis (bead beating and sonica-
tion), and soft and harsh together extraction methods for
isolation of DNA from saline soil habitats. They found that
bead beating and soft lysis together provided high-quality,
high-molecular-weight DNA (31). Furthermore, Liles and
colleagues (32) aimed to recover high-molecular-weight, con-
tiguous genomic DNA to preserve the genes involved in bio-
synthetic pathways. High-molecular-weight DNA (>1 Mbp)
was recovered successfully using an agarose plug. In this
method, microbial cells are recovered by differential cen-
trifugation from the soil sample, and subsequently embedded
in low-melting-point agarose within a syringe. The cells
are then lysed using a lysis buffers and washed (33). DNA
extracted using this method exceeded the size of the DNA iso-
lated by direct extraction methods (20 kb–1 Mb), although
DNA concentration was significantly lower that what was
achieved by direct extraction methods (32).
DNA yield is also critical. Most FM protocols require a
large initial input of DNA as most proceeding processing steps
(which involve fragmentation of DNA in some way) such as
sonication, needle shearing, and enzyme-based treatments
often require an initial input concentration of at the very least
0.1–1 µg DNA. Further processing includes electrophoresis of
the DNA on a gel, purification of DNA fragments of the
appropriate size, and end repair. Often, these steps result in
a loss in DNA yield. Therefore, the greater the starting con-
centration of DNA on extraction, the more robust the suc-
ceeding ligation and library construction steps will be.
DNA quality should be tested using an enzymatic test such
as PCR to verify that no inhibitors have been coextracted.
Typically, DNA is typically tested using PCR amplification
with primers targeting conserved regions of bacterial riboso-
mal RNA genes. Extracted DNA is also run on a gel to ensure
high-molecular-weight DNA has been extracted. DNA is
subsequently manipulated through a series of steps for FM
analysis, including ligation of large fragments into an appro-
priate vector (size of DNA is also dependent on the type of
vector being used), transformation of host (surrogate) cells
with the vectors containing environmental DNA, and finally
a target-specific functional screening of surrogate cells.
INSERT SIZE AND VECTOR CHOICE
The choice of vector and insert size is dependent on the type
of study and biological questions posed. Small-insert libraries
(i.e., fragments smaller than 15 kb) are prepared in plasmid
vectors. These plasmids are typically present in high copy
number within the cytoplasm of the host (e.g., competent
E. coli strains) and have strong constitutive or inducible
promoters. In some cases, vectors without promoters are
used, and in such approaches, the exogenous DNA must pro-
vide its own promoter for expression. Plasmids are commonly
used for FM screens in which single genes shorter than 10 kb
in length are the intended target, and numerous successful
studies have been achieved (34–36).
Large-insert libraries—those with exogenous DNA frag-
ments of greater than 15 kb—can be processed using vectors
such as cosmids (15–35 kb), fosmids (25–45 kb), or bacterial
artificial chromosomes (BACs; 100–200 kb). Large insert
libraries are favored for screens because genomic DNA frag-
ments of such length can contain multiple genes, including
genes from entire biochemical pathways. Larger fragments
also provide greater genomic context to functional genes of
interest. The genome context is revealed by full vector insert
sequencing of clones that are positive during functional
screening. In addition, the larger the insert fragment size,
the greater the likelihood that the insert will contain a phy-
logenetic marker gene in addition to functional genes of
interest (27). A further benefit of using large-insert libraries
is that fewer clones need to be screened to achieve coverage
of an equivalent amount of genomic DNA.
Although very large fragments are useful in FM screening,
in practice, small- and medium-insert libraries (e.g., fosmids)
are most commonly used. Cloning very large DNA fragments
into BACs is challenging and time-consuming. Moreover,
the frequency of detecting clones with the phenotype of
interest (hit rate) remains highest for single-gene products
in small- and large-insert methods (11). This is thought to
be a result of truncated pathways associated with fragmented
genomic DNA and of host incompatibility for complex met-
abolic pathways requiring multiple enzyme products (29).
A number of methods have improved the success (hit rate)
of FM screens; some of these developments have involved the
optimization of expression systems of the cloning vector. A
key example is the work of Lämmle and colleagues (34),
who used a high copy number plasmid ( pJOE930) capable
of bidirectional transcription due to the insertion of two indu-
cible lac promoters on either side of a symmetrical multiple
cloning site (37). As a result, the functional screens were
not dependent on the orientation of the insert, thereby effec-
tively increasing the amount of screened DNA. Using this
FM method, Lämmle et al. (34) successfully isolated 14
lipase/esterases with a hit rate of 1:1,500, 13 phosphotases
with a hit rate of 1:2,500, and 38 amylases at a hit rate of
1:800 (34). This incidence of positive clones was higher
than previous FM studies dedicated to screening for these
enzymes, emphasizing the usefulness of this dual-orientation
expression system (34).
CHOICE OF SURROGATE CLONING HOST
FOR MAINTENANCE OF mgDNA
Strains of E. coli are frequently employed for metagenomic
screening, as these organisms are relatively easy to manipulate
genetically and are the basis of many commercial cloning kits.
However, as a primary objective of FM is to clone DNA from
exotic and uncultured microorganisms, the use of E. coli
strains can be limiting as some novel sequences cannot be
expressed within this host (23). The inability of a host organ-
ism to express target genes can be due to a number of factors.
For example, different microorganisms have different codon
usage preferences; such preferences lead to codon usage
bias, where organisms use specific codons more frequently
for specific amino acids (38). As a result, codons present in
an inserted gene during FM may not be recognized by the
host, resulting in incorrect translation and in some instances
no expression. Low adaptation of the host to codons of
inserted DNA could lower cellular fitness of the cloning
host, thus inhibiting screens. The impact of codon usage
bias has been highlighted by Kudla and colleagues (39).
Other features of heterologous DNA can cause host trans-
lation difficulties. To ensure that genes are expressed in the
host, the presence of cis acting DNA sequences such as the
correct promoter sequence is required, as is the presence of
host-derived trans acting factors such as certain cofactors,
and signal peptides for sufficient secretion of the product by
E. coli (27, 40). Last, any host may be capable of heterolo-
gously expressing a gene product; however, the product may
exert a toxic effect on such a host, killing or rendering the
cells insufficiently active for screening purposes.
 2.4.3. Functional Metagenomics: Procedures and Progress ▪ 2.4.3-5
Research by Gabor and colleagues (41) quantified the
accessibility of metagenomic DNA with E. coli as a host using
the complete genome sequences of 32 diverse prokaryotes.
Their studies concluded that there are significant differences
in expression mechanisms between distinct taxonomic groups
and that approximately 40% of the enzymatic activities
encoded for by metagenomic DNA can be expressed in
E. coli by random cloning methods (42). There are different
methods of accessing the rest of this DNA. One approach is
to maximize expression in E. coli by genetically engineering
the relevant transcription and translation systems. For exam-
ple, Bernstein and colleagues subjected the S1 ribosomal pro-
tein of E. coli to site-directed mutagenesis to enable the
organism to efficiently express genes from high GC genomes
(43). Also, Leggewie and colleagues (44) were able to miti-
gate the problem of insufficient promoter recognition by con-
structing a transposon MUExpress system; this recombinant
transposon with dual inducible promoters was inserted ran-
domly into metagenomic DNA fragments, enabling expres-
sion of flanking DNA regions in both directions (44).
An alternative method to improve heterologous gene
expression is to broaden the surrogate host range for cloning.
In such an approach, different bacterial hosts are used for
maintenance and expression of the mgDNA, and some stud-
ies implement multiple expression hosts, including E. coli, for
screening of a single mgDNA library. In general, when two or
more hosts are employed, E. coli serves as the host of choice for
mgDNA library storage and maintenance, though the E. coli
library will likely be screened, too. When different hosts are
used for screening the library, the use of multiple gene expres-
sion systems increases the chance of a positive functional
screen. Such multiple-host approaches require a replicating
shuttle vector that can be maintained in both hosts. For
example, Dobrijevic and colleagues developed a high-
throughput system for analysis of surface-exposed proteins
in bacteria using a Gram-negative/Gram-positive shuttle vec-
tor system with E. coli and Bacillus subtilis for optimal expres-
sion (45). Likewise, Courtois and colleagues (24) also
implemented a shuttle vector system using E. coli and Strepto-
myces lividans (a genus from which most of the commercially
successful antibiotics were isolated) as hosts to mine the soil
metagenome for novel antibiotics. They successfully identi-
fied novel polyketide synthase genes (24). Development of
alternative host systems is critical to improving functional
screens. This is exemplified by the work of McMahon and col-
leagues, who demonstrated that certain mgDNA fragments
could be expressed in S. lividans but not in E. coli (46). Craig
et al. (47) used six different proteobacterial hosts to increase
the discovery rate of molecules responsible for altered colony
morphology and antibiosis. They were successful in isolating
more than 35 clones exhibiting alterations in colony mor-
phology and 8 clones exhibiting antibiosis with only one of
the molecules identified twice with the same host species.
This further demonstrates the usefulness of broad host range
for FM projects (47).
LIBRARY SCREENING
A measurable phenotype is needed to identify novel products
by FM. There are three broad approaches for screening an FM
library. The first is a specific, identifiable phenotype of an
individual clone. The majority of successful metagenomic
screens have identified hydrolytic enzymes such as lipases,
esterases, and glycoside hydrolases, and this is likely due to
the relative ease of the simple plate assays used to detect
them (48, 49). Identification of such enzymes relies on “low-
tech” detection of observable phenotypic changes, such as
colony morphology or pigmentation that are indicative of
the expression of a particular gene within the metagenome.
Similarly, indicator dyes are incorporated into growth media
to screen clones for gene expression, and these dyes facilitate
high-throughput screening. For example, the incorpora-
tion of BCIP into growth media allows for colorimetric detec-
tion of alkaline phosphatases which hydrolyze BCIP to
5-bromo-4-chloro-3-indole; this compound is then oxidized
to form a dark blue dye (see Table 2).
The second main strategy employed for screening of meta-
genomic libraries is selective. In this approach, media and
incubation conditions do not allow the host to grow without
the expression of heterologous genes obtained from environ-
mental mgDNA. A simple, yet effective and selective screen-
ing method employing this heterologous complementation
method was conducted by Culligan et al. (16). Agar supple-
mented with a high concentration of NaCl was used to select
clones harboring a gene conferring salt tolerance, since the
E. coli host could not grow at the salt concentrations used
(16). Similarly, Simon and Daniel (49) used a mutant
E. coli with a lethal mutation in the 50 -30 exonuclease domain
of DNA polymerase I (50). Only clones harboring an insert
encoding a functional DNA polymerase were capable of
growth.
The development of high-throughput screening tech-
nology has aided direct detection and heterologous com-
plementation-based strategies for screening metagenomic
libraries (49). In recent studies discussed here, for example,
from 23,000 to 230,000 clones were screened (Table 4).
Such throughput necessitates the use of high-throughput
technologies, including growth of clones and phenotypic
expression detection in 96- and 384-well plates, microplate
readers, and colony picking robots such as the Genetix range,
Pickolo, or the RapidPick, all of which aid the development
of reliable, automated, and efficient library screening.
Despite great success, some assays used in the identi-
fication of certain enzymes such as proteases have proven inef-
ficient (12, 51). In addition, many enzymes require more
sensitive or advanced screening means, including high-
performance liquid chromotography or reactive chromogenic
or fluorescent substrates (48). The field is progressing toward a
third method for high-throughput FM library screen in which
intracellular gene expression is induced by a particular sub-
strate or product. Williamson and colleagues (22) have devel-
oped METREX, a novel intracellular system used to screen
metagenomic clones. The host contains a biosensor for
detecting quorum-inducing molecules. Upon sensing, the
cell produces a green fluorescent protein (GFP), which is
detectable by fluorescence microscopy or fluorescence-based
cell sorting. Similarly, a substrate-induced gene expression
system was also developed. With this method, an operon-trap
GFP expression vector is used to construct a metagenomic
library. mgDNA is inserted upstream of the GFP gene with
the notion that catabolic genes are often induced in the pres-
ence of a particular substrate and that these genes are usually
located close together as an operon. The entire library is
subject to a substrate-specific gene induction. Upon expres-
sion of substrate-specific genes, the GFP is also expressed.
High-throughput detection of GFP and consequently sub-
strate-specific genes is carried out with fluorescence-activated
cell sorting, allowing for detection of novel catabolic operons
(52). Uchiyama and Miyazaki (53) also developed the
product-induced gene expression system that is conceptually
similar to substrate-induced gene expression except the GFP
reporter gene is located on a product of enzymatic (gene of
 2.4.3-6 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
TABLE 4 Recent (2010-present) successful functional metagenomic studies summarizing host organism(s) for library construction, types of genes targeted, diverse environments DNA was isolated
from for metagenomic screening, type of vector used (with average metagenomic insert size), hit rate, and type of screening method implementeda

Vector/average
Environment Target gene Host(s) # positives/#screened clones Assay type Ref
insert size
Human gut Dietary fiber catabolic E. coli Fosmid (30–40 kb) 310/704,000 Agar plate assay supplemented [69]
enzymes with polysaccharides
Soil (deciduous forest, Antimicrobial Agrobacterium tumefacietins, Cosmid (n/ab) 170,000(forest); Primary agar plate assay screen [47]
creek bed, and cold Burkholderia graminis, Caulobacter 450,000(creek); looking for alterations in
desert) vibrioides, E. coli, Pseudomonas 130,000(desert) colony morphology and
putida, Ralstonia metallidurans pigmentation, then overlay
with B. subtilis to screen for
growth inhibition
Activated sludge from a Amidase E. coli harbouring pCmGFPbenR Fosmid (33 kb) 4/96,000 PIGEX [53]
coke waste treatment (benzoate-responsive sensor
plant plasmid)
Soil 40 -phosphoantetheinyl E. coli Plasmid (2–6 kb) 4/3 × 106 Agar plate assay: blue coloration [70]
Glacier ice DNA polymerase E. coli Plasmid (4 kb), Plasmid = 230,00; Agar plate assay: ability to grow [50]
fosmid (36 kb) fosmid = 4000 at 18°C
Compost soil Cellulase E. coli Cosmid (33 kb) 4/100,000 Plate assay [71]
Forest soil and grassland Lipase E. coli Plasmid, 28/2217648 ( plasmid); Agar plate assay: agar [72]
soil fosmid 9/711794 (fosmid) supplemented with 1%
tributyrin
Soil from an urban Antibiotic resistance E. coli Plasmid (2 kb) 39/n/a Agar plate assay: using [73]
environment inhibitory concentrations of
antibiotics
Soil; agricultural field Antibiotic resistance E. coli Plasmid (6.5–7 kb) 11/550,000 Agar plate assay: using [74]
soil and soil in vicinity inhibitory concentrations of
of Mammillaria carnea antibiotics
Human gut Salt tolerance E. coli Fosmid (∼40 kb) 47c/23,040 Agar plate assay: inhibitory [16][75]d
concentrations of NaCl
Marine sponge Antimicrobial E. coli Fosmid (∼40 kb) 1/250,000 Agar plate assay: overlay with [76]
B. cereus
Alaskan soil Antibiotic resistance E. coli Plasmid, 1/(13,201 Mb DNA) Agar plate assay: inhibitory [77]
fosmid (∼40 kb) concentrations of antibiotics
Deep-sea sediment Lipase E. coli Fosmid (15–33 kb) /681,100 Agar plate assay: agar [78]
supplemented with 1%
tributyrin
Unvegetated Antarctic Cellulase E. coli BAC (5.1 kb) 11/124,000 Agar plate assay: agar [79]
soil supplemented with cellulose
Human gut (fecal and Hydrolytic E. coli Fosmid Fecal = 11 hits; Agar plate assay: supplemented [55]
ileum mucosa) enzymes-( probiotic Ileum = 49 with carbon source
breakdown)
Activated sludge from a Esterase E. coli Plasmid (3 kb) 2/40,000 Agar plate assay: agar [80]
paper mill supplemented with 1%
tributyrin
Forest soil Protease E. coli Plasmid (7–12 kb) 1 positive clone Agar plate assay: agar [81]
supplemented with
AZCL-casein
Human gut Antibiotic resistance E. coli Fosmid (30 kb) 17/415,000 Agar plate assay: using [82]
inhibitory concentrations of
antibiotics
Biomass from sequencing Lipase, E. coli, Streptomyces lividans Cosmid 17/2000 (esterase/lipase) Agar plate assay: agar [83]
fed-batch reactor esterase, protease supplemented with 1%
tributyrin or skimmed milk
Desert sand (Death Protease E. coli Plasmid (6 kb), Plasmid 1/30,000; Agar plate assay: agar [84]
Valley and Gobi) fosmid (32 kb) Fosmid 4/17,000 supplemented with skimmed

2.4.3. Functional Metagenomics: Procedures and Progress ▪ 2.4.3-7

milk
Apple orchard soil Antibiotic resistance E. coli Fosmid (30 kb) 13/446,000 Agar plate assay: using [85]
inhibitory concentrations of
antibiotics
Elephant feces and river Flavonoid-modifying E. coli Fosmid (35 kb) 1/50,000 Extract thin layer [48]
sediment enzymes chromatography
Cotton field Tannase E. coli Plasmid (3.5 kb) 1/92,000 Agar plate assay: agar [86]
supplemented with
X-caprylate
Dairy cow manure Antibiotic resistance E. coli Plasmid (2.1–3.2 kb), Plasmid 5,684,580; Agar plate assay: using [57]
fosmid (20–40 kb) Fosmid 432,800; inhibitory concentrations of
80 positive clones in total antibiotics
a
For earlier studies see ref 11.
b
N/A; the relevant data could not be found in the publication.
c
47 clones were found but not all clones were deemed suitable for further characterisation.
d
Metagenomic library used was constructed in a previous study.
2.4.3-8 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
interest) activity. In their study, they aimed to isolate ami-
dase enzymes that convert benzamide to benzoate. Here, the
GFP reporter gene was downstream of the BenR gene, a gene
encoding benzoate transcriptional precursor. In the presence
of a substrate (benzamide) if the host cells had an enzyme
capable of transforming benzamide to benzoate, then on
conversion of the molecule, the GFP and transcriptional pre-
cursor would be expressed, indicating enzymatic activity.
These studies suggest that such strategies can be similarly
implemented for identifying other biologically active small
molecules in metagenomic libraries (22).
WHAT SIZE FM SCREEN IS NEEDED?
The determination of the appropriate size of an FM screen is
dependent on the genetic diversity of the system being
studied, the evenness of the distribution of microbial taxa
in the system, the average size of mgDNA inserted into the
vector, the type and complexity of the functional target,
the type of screen, and many other factors. Preliminary meta-
genomic screening using rRNA gene amplicon sequencing is
used to estimate the microbial diversity with a sample and
identify the most abundant organisms (54). Similarly, an
SM sequencing approach can be employed to estimate the
abundance of genes of known function in the sample (55).
Such an approach will underestimate novel functional genes
due to the lack of appropriate annotation, but can provide
some insight into the approximate abundance of some genes.
Furthermore, the type of screen has a major effect on how
much mgDNA can be screened. Negative screens, in which
only organisms expressing genes of interest are capable of
growth, the entire transformation yield is screened. This
equates to several orders of magnitude more DNA than is cov-
ered in most positive screens. Table 3 shows the wide spread of
frequency with which a function can be found. In our experi-
ence, this is an empirical factor that needs to be established
with a small functional metagenomic library first, and from
this an approximate hit rate per Mb of DNA can be estimated.
Small screens of 20,000 clones are appropriate for such empir-
ical tests, though the final libraries are likely to be much
larger.
For example, in our functional metagenomic study of bile
salt hydrolases (25), a library of 250,000 clones (approxi-
mately 10 billion bp of mgDNA) was screened, yielding
101 positive clones.
SEQUENCE ANALYSIS AND
CHARACTERIZATION
Following the identification of a positive hit (clone) from an
FM screen, the clone must be isolated, and the presence of an
insert must be confirmed to ensure that the positive pheno-
type was not due to a host-genome mutation or contaminant.
The presence of an insert in vectors can be validated by means
of blue-white screening techniques, PCR, restriction analysis
of the vector, or retransformation of the putative positive vec-
tor to ensure recapitulation of the phenotype. After the pres-
ence of an insert is verified, the insert is sequenced. Most FM
projects published to date have sequenced the insert DNA of
the recombinant clones by primer walking (56) and transpo-
son mutagenesis (25). More recently, studies incorporate
deep sequencing approaches to either determine the insert
sequence or complement FM screens. This is particularly use-
ful for larger inserts in fosmid and cosmid vectors. For exam-
ple, in a recent study by Wichmann and colleagues (57), 80
different antibiotic resistance genes were isolated after
functional screening of both small-insert libraries and larger,
fosmid libraries. Coupling of FM screening with third-
generation PacBio sequencing of fosmid clones enabled a
comprehensive characterization of the “resistome” of the
environment they were studying (57).
After the sequence of the insert has been determined, it is
subject to bioinformatic analysis and activity assays to charac-
terize the expressed molecule. Initially, the sequence data
must be annotated, and this can be conducted using sequence
editing software packages such as the commercially available
Lasergene99 core suite (DNASTAR, Madison, WI) or
sequence editor modules within the CodonCode Alignment
Software (CodonCode, Dedham, MA) or free software,
including BioEdit (58) or MEGA (Molecular Evolutionary
Genetics Analysis) (59). Predictions of open reading frames,
structures, promoters, and homology searches can be con-
ducted using software such as the NCBI Open Reading Frame
finder (60). There are databases that specialize in specific
classes of enzymes and are used for homology searches and
structural and biochemical properties predictions. For exam-
ple, the MEROPS database for proteolytic enzymes (http://
merops.sanger.ac.uk/index.shtml) can be used for homol-
ogy searches and to determine substrates and inhibitors of
proteases (61), the Carbohydrate-Active Enzyme Data-
base can be used to analyze carbohyrdate-active enzymes
(CAZymes) such as glycoside hydrolases (http://www.cazy.
org) (62), and the Lipase Engineering Database (http://
www.led.uni-stuttgart.de) (63) can be used for similar analy-
ses on putative lipases. For annotation of antibiotic resistance
genes, the Antibiotic Resistance Genes Database (ARGD:
http://www.ardb.cbcb.umd.edu/) (64), the Comprehensive
Antibiotic Resistance Database (http://arpcard.mcmaster.ca/
) (65), and the Antibiotic and Secondary Metabolites Shell
(known as antiSMASH: http://www.antismash.secondaryme
tabolites.org/) (66, 67) are useful tools. AntiSMASH can also
be used to search for homologous secondary metabolites. A
comprehensive list of enzyme databases are discussed in the
work of Schomburg and Schomburg (68).
Subsequent to bioinformatic analyses, activity-based
characterization of the insert DNA is conducted. For exam-
ple, primers can be designed to span the predicted coding
region of DNA, and the PCR product generated with these
primers can be subcloned using standard vector systems
such a pBluescript variations or pET vector expression
systems, and screened for recapitulation of the primary phe-
notype. The expressed protein characteristics can be meas-
ured in vitro, including temperature optima, pH optima,
the effect of certain inhibitors, and substrate specificity.
Coupling of bioinformatic analysis and activity-based assays
is the purpose of the FM survey, and these combined data
may have clinical, biotechnological, and environmental
applications.
CONCLUSIONS
The development of high-throughput technology and its
implementation in both functional- and sequencing-based
metagenomic processes have enriched our understanding of
the uncultured world, provided insight into the true extent
of taxonomic and functional diversity of virtually any habitat
on Earth. FM screening provides a mechanism to generate
accurate functional annotation of genetic material from
organisms that cannot currently be cultivated. Although
more demanding at the wet-lab level, FM addresses a criti-
cal knowledge gap that limits cultivation-independent
metagenomic screens: specifically, the inability to annotate
 2.4.3. Functional Metagenomics: Procedures and Progress ▪ 2.4.3-9
novel genes and the misannotation or putative nature of bio-
informatics annotations. FM screening and shotgun metage-
nomic screening are both required to improve functional
characterization of microbial communities. The field of FM
is still developing, and major hurdles remain to improve
throughput and improve the expression of foreign DNA in
a domesticated host. Future development will broaden the
range of screening strategies, provide for means to identify
ever more enzymes and molecules, and recover genes from
low-abundance species within complex communities. The
recent research mentioned herein offers promising results,
such as the exploration of novel host systems for optimal
expression of heterologous DNA and subsequent novel
gene discovery and the development of high-throughput
screening techniques. We emphasize the notion that FM
remains the only method of detecting novel biocatalysts
and biomolecules, enabling a direct phenotypic investigation
of the physiology and biochemistry of microorganisms in their
habitat. When implementing SM sequencing approaches,
functional annotation is only putative; with FM screens,
function can be definitively demonstrated.
REFERENCES
1. Herlemann DPR, Lundin D, Labrenz M, Jurgens K, Zheng
Z, Aspeborg H, Andersson AF. 2013. Metagenomic de novo
assembly of an aquatic representative of the verrucomicro-
bial class spartobacteria. mBio 4:e00569-12.
2. Venter JC, Remington K, Heidelberg JF, Halpern AL,
Rusch D, Eisen JA, Wu D, Paulsen I, Nelson KE,
Nelson W, Fouts DE, Levy S, Knap AH, Lomas MW,
Nealson K, White O, Peterson J, Hoffman J, Parsons R,
Baden-Tillson H, Pfannkoch C, Rogers Y-H, Smith HO.
2004. Environmental genome shotgun sequencing of the
Sargasso Sea. Science 304:66–74.
3. Meyer F, Paarmann D, D'Souza M, Olson R, Glass EM,
Kubal M, Paczian T, Rodriguez A, Stevens R, Wilke A,
Wilkening J, Edwards RA. 2008. The metagenomics
RAST server—a public resource for the automatic phyloge-
netic and functional analysis of metagenomes. BMC Bioin-
form 9:386.
4. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, Mani-
chanh C, Nielsen T, Pons N, Levenez F, Yamada T, Mende
DR, Li J, Xu J, Li S, Li D, Cao J, Wang B, Liang H, Zheng
H, Xie Y, Tap J, Lepage P, Bertalan M, Batto J-M, Hansen
T, Le Paslier D, Linneberg A, Nielsen HB, Pelletier E,
Renault P, Sicheritz-Ponten T, Turner K, Zhu H, Yu C,
Li S, Jian M, Zhou Y, Li Y, Zhang X, Li S, Qin N, Yang
H, Wang J, Brunak S, Dore J, Guarner F, Kristiansen K,
Pedersen O, Parkhill J, Weissenbach J. 2010. A human
gut microbial gene catalogue established by metagenomic
sequencing. Nature 464:59–65.
5. Peterson J, Garges S, Giovanni M, McInnes P, Wang L,
Schloss JA, Bonazzi V, McEwen JE, Wetterstrand KA,
Deal C, Baker CC, Di Francesco V, Howcroft TK, Karp
RW, Lunsford RD, Wellington CR, Belachew T, Wright
M, Giblin C, David H, Mills M, Salomon R, Mullins C,
Akolkar B, Begg L, Davis C, Grandison L, Humble M,
Khalsa J, Little AR, Peavy H, Pontzer C, Portnoy M, Sayre
MH, Starke-Reed P, Zakhari S, Read J, Watson B, Guyer
M. 2009. The NIH Human Microbiome Project. Genome
Res 19:2317–2323.
6. Li J, Jia H, Cai X, Zhong H, Feng Q, Sunagawa S, Arumu-
gam M, Kultima JR, Prifti E, Nielsen T, Juncker AS, Man-
ichanh C, Chen B, Zhang W, Levenez F, Wang J, Xu X,
Xiao L, Liang S, Zhang D, Zhang Z, Chen W, Zhao H,
Al-Aama JY, Edris S. 2014. An integrated catalog of
reference genes in the human gut microbiome. Nat Biotech-
nol 32:834–841.
7. Prakash T, Taylor TD. 2012. Functional assignment of
metagenomic data: challenges and applications. Brief Bioin-
form 13:711–727.
8. Gilbert J, Meyer F, Jansson J, Gordon J, Pace N, Tiedje J,
Ley R, Fierer N, Field D, Kyrpides NC, Glöckner FO,
Klenk H-P, Wommack KE, Glass E, Docherty K, Gallery
R, Stevens R, Knight R. 2010. The Earth Microbiome
Project. Meeting report of the 1st EMP meeting on sample
selection and acquisition. Argonne National Laboratory,
Chicago, IL, 6 October 2010, Vol. 3.
9. Iwai S, Chai B, Sul WJ, Cole JR, Hashsham SA, Tiedje
JM. 2009. Gene-targeted-metagenomics reveals extensive
diversity of aromatic dioxygenase genes in the environment.
ISME J 4:279–285.
10. Kotik M. 2009. Novel genes retrieved from environmental
DNA by polymerase chain reaction: current genome-
walking techniques for future metagenome applications.
J Biotechnol 144:75–82.
11. Uchiyama T, Miyazaki K. 2009. Functional metagenomics
for enzyme discovery: challenges to efficient screening. Curr
Opin Biotechnol 20:616–622.
12. Morris LS, Evans J, Marchesi JR. 2012. A robust plate assay
for detection of extracellular microbial protease activity in
metagenomic screens and pure cultures. J Microbiol Meth
91:144–146.
13. Brady SF. 2007. Construction of soil environmental DNA
cosmid libraries and screening for clones that produce
biologically active small molecules. Nat Protocols 2:
1297–1305.
14. Smith PB, Hancock GA, Rhoden DL. 1969. Improved
medium for detecting deoxyribonuclease-producing bacte-
ria. Appl Microbiol 18:991–993.
15. Henne A, Schmitz RA, Bömeke M, Gottschalk G, Daniel
R. 2000. Screening of environmental DNA libraries for the
presence of genes conferring lipolytic activity on Escherichia
coli. Appl Environ Microbiol 66:3113–3116.
16. Culligan EP, Sleator RD, Marchesi JR, Hill C. 2012. Func-
tional metagenomics reveals novel salt tolerance loci from
the human gut microbiome. ISME J 6:1916–1925.
17. Shuman HA, Silhavy TJ. 2003. The art and design of
genetic screens: Escherichia coli. Nat Rev Genet 4:419–431.
18. Allen HK, Donato J, Wang HH, Cloud-Hansen KA,
Davies J, Handelsman J. 2010. Call of the wild: antibiotic
resistance genes in natural environments. Nat Rev Microbiol
8:251–259.
19. Allen HK, Moe LA, Rodbumrer J, Gaarder A, Handels-
man J. 2008. Functional metagenomics reveals diverse
[beta]-lactamases in a remote Alaskan soil. ISME J 3:
243–251.
20. Kim Y-J, Choi G-S, Kim S-B, Yoon G-S, Kim Y-S, Ryu
Y-W. 2006. Screening and characterization of a novel ester-
ase from a metagenomic library. Prot Expr Purif 45:315–323.
21. Henke E, Bornscheuer UT. 2002. Fluorophoric assay for
the high-throughput determination of amidase activity.
Anal Chem 75:255–260.
22. Williamson LL, Borlee BR, Schloss PD, Guan C, Allen HK,
Handelsman J. 2005. Intracellular screen to identify metage-
nomic clones that induce or inhibit a quorum-sensing biosen-
sor. Appl Environ Microbiol 71:6335–6344.
23. Handelsman J. 2004. Metagenomics: application of
genomics to uncultured microorganisms. Microbiol Mol
Biol Rev 68:669–685.
24. Courtois S, Cappellano CM, Ball M, Francou F-X, Nor-
mand P, Helynck G, Martinez A, Kolvek SJ, Hopke J,
Osburne MS, August PR, Nalin R, Guérineau M, Jeannin
P, Simonet P, Pernodet J-L. 2003. Recombinant environ-
mental libraries provide access to microbial diversity for
2.4.3-10 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
drug discovery from natural products. Appl Environ Microbiol
69:49–55.
25. Jones BV, Begley MI, Hill C, Gahan CGM, Marchesi JR.
2008. Functional and comparative metagenomic analysis of
bile salt hydrolase activity in the human gut microbiome.
Proc Natl Acad Sci 105:13580–13585.
26. Ni J, Yan Q, Yu Y. 2013. How much metagenomic sequenc-
ing is enough to achieve a given goal? Sci Rep 3:1968.
27. Ekkers DM, Cretoiu MS, Kielak AM, Elsas JD. 2012. The
great screen anomaly—a new frontier in product discovery
through functional metagenomics. Appl Microbiol Biotechnol
93:1005–1020.
28. Bertrand HLN, Poly F, Van VT, Lombard N, Nalin R,
Vogel TM, Simonet P. 2005. High molecular weight
DNA recovery from soils prerequisite for biotechnological
metagenomic library construction. J Microbiol Meth 62:
1–11.
29. Cowan D, Meyer Q, Stafford W, Muyanga S, Cameron R,
Wittwer P. 2005. Metagenomic gene discovery: past, present
and future. Trends Biotechnol 23:321–329.
30. Salonen A, Nikkila J, Jalanka-Tuovinen J, Immonen O,
Rajilic-Stojanovic M, Kekkonen RA, Palva A, de Vos
WM. 2010. Comparative analysis of fecal DNA extraction
methods with phylogenetic microarray: effective recovery
of bacterial and archaeal DNA using mechanical cell lysis.
J Microbiol Metho 81:127–134.
31. Purohit MK, Singh SP. 2009. Assessment of various meth-
ods for extraction of metagenomic DNA from saline habitats
of coastal Gujarat (India) to explore molecular diversity. Lett
Appl Microbiol 49:338–344.
32. Liles MR, Williamson LL, Rodbumrer J, Torsvik V, Good-
man RM, Handelsman J. 2008. Recovery, purification, and
cloning of high-molecular-weight DNA from soil microor-
ganisms. Appl Environ Microbiol 74:3302–3305.
33. Liles M, Williamson L, Handelsman J, Goodman R. 2004.
Section 3 update: isolation of high molecular weight
genomic DNA from soil bacteria for genomic library con-
struction, p. 2741–2751. In Kowalchuk GA, de Bruijn FJ,
Head IM, Akkermans AD, van Elsas JD (eds), Molecular
Microbial Ecology Manual. Springer, Dordrecht.
34. Lämmle K, Zipper H, Breuer M, Hauer B, Buta C, Brun-
ner H, Rupp S. 2007. Identification of novel enzymes with
different hydrolytic activities by metagenome expression
cloning. J Biotechnol 127:575–592.
35. Hu Y, Zhang G, Li A, Chen J, Ma L. 2008. Cloning and
enzymatic characterization of a xylanase gene from a soil-
derived metagenomic library with an efficient approach.
Appl Microbiol Biotechnol 80:823–830.
36. Waschkowitz T, Rockstroh S, Daniel R. 2009. Isolation
and characterization of metalloproteases with a novel
domain structure by construction and screening of metage-
nomic libraries. Appl Environ Microbiol 75:2506–2516.
37. Altenbuchner J, Viell P, Pelletier I. 1992. Positive selection
vectors based on palindromic DNA sequences. Meth Enzy-
molo 216:457–466.
38. Behura SK, Severson DW. 2013. Codon usage bias: causa-
tive factors, quantification methods and genome-wide pat-
terns: with emphasis on insect genomes. Biol Rev 88:49–61.
39. Kudla G, Murray AW, Tollervey D, Plotkin JB. 2009.
Coding-sequence determinants of gene expression in Escher-
ichia coli. Science 324:255–258.
40. Angelov A, Mientus M, Liebl S, Liebl W. 2009. A two-host
fosmid system for functional screening of (meta)genomic
libraries from extreme thermophiles. Syst Appl Microbiol
32:177–185.
41. Gabor EM, De Vries EJ, Janssen DB. 2004. Construction,
characterization, and use of small-insert gene banks of DNA
isolated from soil and enrichment cultures for the recovery of
novel amidases. Environ Microbiol 6:948–958.
42. Gabor EM, Alkema WBL, Janssen DB. 2004. Quantifying
the accessibility of the metagenome by random expression
cloning techniques. Environ Microbiol 6:879–886.
43. Bernstein JR, Bulter T, Shen CR, Liao JC. 2007. Directed
evolution of ribosomal protein S1 for enhanced translational
efficiency of high GC Rhodopseudomonas palustris DNA in
Escherichia coli. J Biol Chem 282:18929–18936.
44. Leggewie C, Henning H, Schmeisser C, Streit WR, Jaeger
K-E. 2006. A novel transposon for functional expression of
DNA libraries. J Biotechnol 123:281–287.
45. Dobrijevic D, Di Liberto G, Tanaka K, de Wouters T, Der-
vyn R, Boudebbouze S, Binesse J, Blottiere HM, Jamet A,
Maguin E, van de Guchte M. 2013. High-throughput system
for the presentation of secreted and surface-exposed proteins
from Gram-positive bacteria in functional metagenomics
studies. PLOS One 8:e65956.
46. McMahon MD, Guan C, Handelsman J, Thomas MG.
2012. Metagenomic analysis of Streptomyces lividans reveals
host-dependent functional expression. Appl Environ Micro-
biol 78:3622–3629.
47. Craig JW, Chang F-Y, Kim JH, Obiajulu SC, Brady SF.
2010. Expanding small-molecule functional metagenomics
through parallel screening of broad-host-range cosmid envi-
ronmental DNA libraries in diverse proteobacteria. Appl
Environ Microbiol 76:1633–1641.
48. Rabausch U, Juergensen J, Ilmberger N, Böhnke S,
Fischer S, Schubach B, Schulte M, Streit WR. 2013. Func-
tional screening of metagenome and genome libraries for
detection of novel flavonoid-modifying enzymes. Appl Envi-
ron Microbiol 79:4551–4563
49. Simon C, Daniel R. 2009. Achievements and new knowl-
edge unraveled by metagenomic approaches. Appl Microbiol
Biotechnol 85:265–276.
50. Simon C, Herath J, Rockstroh S, Daniel R. 2009. Rapid
identification of genes encoding DNA polymerases by
function-based screening of metagenomic libraries derived
from glacial ice. Appl Environ Microbiol 75:2964–2968.
51. Jones BV, Sun F, Marchesi JR. 2007. Using skimmed milk
agar to functionally screen a gut metagenomic library for pro-
teases may lead to false positives. Lett Appl Microbiol 45:
418–420.
52. Uchiyama T, Abe T, Ikemura T, Watanabe K. 2005.
Substrate-induced gene-expression screening of environ-
mental metagenome libraries for isolation of catabolic genes.
Nat Biotech 23:88–93.
53. Uchiyama T, Miyazaki K. 2010. Product-induced gene
expression, a product-responsive reporter assay used to screen
metagenomic libraries for enzyme-encoding genes. Appl
Environ Microbiol 76:7029–7035.
54. Rondon MR, August PR, Bettermann AD, Brady SF,
Grossman TH, Liles MR, Loiacono KA, Lynch BA, Mac-
Neil IA, Minor C, Tiong CL, Gilman M, Osburne MS,
Clardy J, Handelsman J, Goodman RM. 2000. Cloning
the soil metagenome: a strategy for accessing the genetic
and functional diversity of uncultured microorganisms.
Appl Environ Microbiol 66:2541–2547.
55. Cecchini DA, Laville E, Laguerre S, Robe P, Leclerc M,
Doré JL, Henrissat B, Remaud-Siméon M, Monsan P,
Potocki-Véronése G. 2013. Functional metagenomics
reveals novel pathways of prebiotic breakdown by human
gut bacteria. PLoS One 8:e72766.
56. Sanger F, Nicklen S, Coulson AR. 1977. DNA sequencing
with chain-terminating inhibitors. Proc Natl Acad Sci USA
74:5463–5467.
57. Wichmann F, Udikovic-Kolic N, Andrew S, Handelsman
J. 2014. Diverse antibiotic resistance genes in dairy cow
manure. mBio 5:e01017-13.
58. Hall T. 2005. BioEdit. Biological sequence alignment editor
for Win95. http://www.mbio.ncsu.edu/bioedit/bioedit.html
 2.4.3. Functional Metagenomics: Procedures and Progress ▪ 2.4.3-11
59. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S.
MEGA6: molecular evolutionary genetics analysis version
6.0. Mol Biol Evol 30:2725–2729
60. Wheeler DL, Church DM, Federhen S, Lash AE, Madden
TL, Pontius JU, Schuler GD, Schriml LM, Sequeira E,
Tatusova TA, Wagner L. 2003. Database resources of the
National Center for Biotechnology. Nucleic Acids Res 31:
28–33.
61. Rawlings ND, Barrett AJ, Bateman A. 2012. MEROPS:
the database of proteolytic enzymes, their substrates and
inhibitors. Nucleic Acids Res 40:D343–D350.
62. Lombard V, Golaconda Ramulu H, Drula E, Coutinho
PM, Henrissat B. 2013. The carbohydrate-active enzymes
database (CAZy) in 2013. Nucleic Acids Res 42:
D490–D495.
63. Fischer M, Pleiss J. 2003. The Lipase Engineering Database:
a navigation and analysis tool for protein families. Nucleic
Acids Res 31:319–321.
64. Liu B, Pop M. 2009. ARDB—Antibiotic Resistance Genes
Database. Nucleic Acids Res 37:D443–D447.
65. McArthur AG, Waglechner N, Nizam F, Yan A, Azad
MA, Baylay AJ, Bhullar K, Canova MJ, De Pascale G,
Ejim L, Kalan L, King AM, Koteva K, Morar M, Mulvey
MR, O’Brien JS, Pawlowski AC, Piddock LJV, Spanogian-
nopoulos P, Sutherland AD, Irene Tang I, Taylor PL,
Thaker M, Wang W, Yan M, Yu T, Wright GD. 2013.
The comprehensive antibiotic resistance database. Antimi-
crob Agents Chemother 57:3348–3357.
66. Blin K, Medema MH, Kazempour D, Fischbach MA,
Breitling R, Takano E, Weber T. 2013. antiSMASH
2.0—a versatile platform for genome mining of secon-
dary metabolite producers. Nucleic Acids Res 41:W204–
W212.
67. Medema MH, Blin K, Cimermancic P, de Jager V,
Zakrzewski P, Fischbach MA, Weber T, Takano E,
Breitling R. 2011. antiSMASH: rapid identification, anno-
tation and analysis of secondary metabolite biosynthesis
gene clusters in bacterial and fungal genome sequences.
Nucleic Acids Res 39:W339–W346.
68. Schomburg D, Schomburg I. 2010. Enzyme databases.
Meth Mol Biol 609:113–128.
69. Tasse L, Bercovici J, Pizzut-Serin S, Robe P, Tap J, Klopp
C, Cantarel BL, Coutinho PM, Henrissat B, Leclerc M,
Dore J, Monsan P, Remaud-Simeon M, Potocki-Veronese
G. 2010. Functional metagenomics to mine the human gut
microbiome for dietary fiber catabolic enzymes. Genome
Res 20:1605–1612.
70. Owen JG, Robins KJ, Parachin NS, Ackerley DF. 2012. A
functional screen for recovery of 40 -phosphopantetheinyl
transferase and associated natural product biosynthesis
genes from metagenome libraries. Environ Microbiol 14:
1198–1209.
71. Pang H, Zhang P, Duan C-J, Mo X-C, Tang J-L, Feng J-X.
2009. Identification of cellulase genes from the metage-
nomes of compost soils and functional characterization
of one novel endoglucanase. Current Microbiology 58:
404–408.
72. Nacke H, Will C, Herzog S, Nowka B, Engelhaupt M,
Daniel R. 2011. Identification of novel lipolytic genes and
gene families by screening of metagenomic libraries derived
from soil samples of the German Biodiversity Exploratories.
FEMS Microbiol Ecol 78:188–201.
73. McGarvey KM, Queitsch K, Fields S. 2012. Wide variation
in antibiotic resistance proteins identified by functional
metagenomic screening of a soil DNA library. Appl Environ
Microbiol 78:1708–1714.
74. Torres-Cortes G, Millan V, Ramirez-Saad HC, Nisa-
Martinez R, Toro N, Martinez-Abarca F. 2011. Character-
ization of novel antibiotic resistance genes identified by
functional metagenomics on soil samples. Environ Microbiol
13:1101–1114.
75. Jones BV, Marchesi JR. 2007. Transposon-aided capture
(TRACA) of plasmids resident in the human gut mobile
metagenome. Nat Meth 4:55–61.
76. Rui He BW, Toshiyuki Wakimoto, Manyuan Wang,
Liancai Zhu and Ikuro Abe. 2013. Cyclodipeptides from
metagenomic library of a japanese marine sponge. J Braz
Chem Soc 24:1926–1932.
77. Lang KS, Anderson JM, Schwarz S, Williamson L,
Handelsman J, Singer RS. 2010. Novel florfenicol and
chloramphenicol resistance gene discovered in alaskan soil
by using functional metagenomics. Appl Environ Microbiol
76:5321–5326.
78. Jeon JH, Kim JT, Lee HS, Kim S-J, Kang SG, Choi SH,
Lee J-H. 2011. Novel lipolytic enzymes identified from
metagenomic library of deep-sea sediment. Evid Based
Complement Alternat Med doi:
79. Berlemont R, Delsaute M, Pipers D, D’Amico S, Feller G,
Galleni M, Power P. 2009. Insights into bacterial cellulose
biosynthesis by functional metagenomics on Antarctic soil
samples. ISME J 3:1070–1081.
80. Shao H, Xu L, Yan Y. 2013. Isolation and characterization
of a thermostable esterase from a metagenomic library. J Ind
Microbiol Biotechnol 40:1211–1222.
81. Biver S, Portetelle D, Vandenbol M. 2013. Charac-
terization of a new oxidant-stable serine protease isolated
by functional metagenomics. Springerplus 28:410. doi:
10.1186/2193-1801-2-410.
82. Cheng G, Hu Y, Yin Y, Yang X, Xiang C, Wang B, Chen Y,
Yang F, Lei F, Wu N, Lu N, Li J, Chen Q, Li L, Zhu B.
2012. Functional screening of antibiotic resistance genes
from human gut microbiota reveals a novel gene fusion.
FEMS Microbiol Lett 336:11–16.
83. Lussier F-X, Chambenoit O, Côté A, Hupé J-F, Denis F,
Juteau P, Beaudet R, Shareck F. 2010. Construction and
functional screening of a metagenomic library using a T7
RNA polymerase-based expression cosmid vector. J Ind
Microbiol Biotechnol 38:1321–1328.
84. Neveu J, Regeard C, DuBow M. 2011. Isolation and char-
acterization of two serine proteases from metagenomic libra-
ries of the Gobi and Death Valley deserts. Appl Microbiol
Biotechnol 91:635–644.
85. Donato JJ, Moe LA, Converse BJ, Smart KD, Berklein
EC, McManus PS, Handelsman J. 2010. Metagenomic
analysis of apple orchard soil reveals antibiotic resistance
genes encoding predicted bifunctional proteins. Appl Envi-
ron Microbiol 76:4396–4401.
86. Yao J, Fan XJ, Lu Y, Liu YH. 2011. Isolation and Character-
ization of a Novel Tannase from a Metagenomic Library.
J Agric Food Chem 59:3812–3818.
 Metagenomics: Assigning Functional Status to
Community Gene Content
NASEER SANGWAN AND RUP LAL
2.4.4
TRACING FUNCTIONAL REPERTOIRE OF AS
YET UNCULTIVATED MICROBIAL
ASSEMBLAGES
Metabolic capabilities of microbes and their inherent genetic
entities (i.e., phages, plasmids) are the major drivers of bio-
sphere processes (1–3). Precise functional annotation of the
community gene content and further downstream analyses
(e.g., richness and evenness) can generate important infor-
mation about the total community metabolism. Gene calling
or prediction is the first step in the annotation of shotgun
sequence data (Fig. 1) derived from genomes or metage-
nomes. Despite advancement in gene prediction algorithms,
the fragmentary nature and short read lengths of the environ-
mental DNA sequences and the genetic complexity of micro-
bial communities are the major difficulties in accurate gene
prediction and providing genomic context to the recovered
sequences. Gene calling can be performed on assembled
(de novo or reference assisted) contigs, individual metage-
nome reads, or a combination of both. Similarity searching
against known databases, commonly known as “evidence-
based” gene calling (4), is the most popular gene prediction
approach and has been used widely in metagenomic sur-
veys (e.g., 5, 6). The Basic Local Alignment Search Tool
(BLAST; 7) is the prototype algorithm used for evidence-
based gene calling and annotation, but like other similarity
or pattern search algorithms, it is limited due to its inability
to identify new protein families and genes that have no homo-
logs in reference databases. In addition, evidence-based gene
calling is also limited by two major biases: (a) sequences with
no affiliation or functional homologues and (b) the inability
to classify short reads and sequences of low quality or chimeric
sequences, particularly in highly diverse environments. Var-
ious other motif-finding algorithms, implemented within
software tools such as CRITICA (8) and Orpheus (9), use
similar approaches for gene calling.
A second major approach for gene prediction, known as ab
initio (“from the beginning”), uses DNA composition-based
features to discriminate between coding and noncoding
sequences using gene prediction parameters trained over
reference sequences. Although such algorithms can identify
individual genes from environmental DNA fragments using
supervised learning (i.e., machine learning based on inferring
information from data using already trained (set) parameters)
doi:10.1128/9781555818821.ch2.4.4
2.4.4-1
and pattern recognition methods, most ab initio pipelines use
only Markov models or hidden Markov models (10). Tools
like Genemark.hmm (11) use a reference-trained Markov
model framework for gene calling, whereas Genemark,
GLIMMER (12), Prodigal (13), and Metagene (14) are self-
trained algorithms (i.e., those that make use of an existing
model to extract information from data) and use generic mod-
els for gene prediction. To avoid the potential bias caused by
their low sensitivity to fragmentary metagenomic open read-
ing frames (ORFs), unsupervised methods (i.e., extracting
information from unknown data without using any preset
parameters) are required; various algorithms like SCIMM
(Sequence Clustering with Interpolated Markov Models)
(15) have been introduced to perform such analyses, but are
still in development. A combined approach can be adopted
for efficient gene calling and annotation; evidence-based
gene calling can be used to identify sequence homologues,
and composition-based approaches (e.g., MetaCV) can
resolve as-yet-uncharacterized (after evidence-based analysis)
community protein sequences.
ASSEMBLING COMMUNITY GENOTYPES: AN
IMPETUS TO RESOLVE THE COMMUNITY
METABOLIC POTENTIAL
Metagenomics provides a new route to analyze microbial
processes and their dynamics in situ. These processes include
biogeochemical transformations (including bioremediation)
and evolutionary process such as lateral gene transfer. For
any environment, by correctly organizing the community
nucleotide pool (by generating long nucleotide stretches),
we can analyze the as yet unknown evolutionary dynamics
that generate and maintain genetic diversity.
Although shotgun sequencing of microbial isolates is not
trivial, assembling community genotypes from diverse and
unknown mixtures of genomic DNA is a daunting challenge.
The difficulty of assembly increases with increasing diversity
and species evenness. Such evenness is a major problem, as in
undersampled systems, every individual read could theoreti-
cally represent an individual genotype (strain). Such diversity
can confuse the assembly algorithms and preclude assembly of
large contigs. Recently there have been efforts to address the
limitations of existing genome assembly algorithms and make
2.4.4-2 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
FIGURE 1 Flow chart for the downstream analysis of metagenome sequences, with a focus on functional analysis. Processes described here are
sequence technology independent and can be used for assembled and individual (unassembled) reads. doi:10.1128/9781555818821.ch2.4.4.f1
them efficient to process the “voluminous” community geno-
mics data (16). Remarkably, draft and/or complete genomes
of uncultivated groups have been recovered from the natu-
ral environments with relatively even species abundance
(17–19). Increasingly, attempts to improve assembly from
highly complex environments have aimed to separate contigs
through a process called binning, in which contigs or reads
with similar GC contents, oligonucleotide frequencies,
putative taxonomy, or coverage are pooled together in groups
(e.g., 4, 17, 20–24). Here, we center our discussion on the
levels of resolution obtainable by comparing the “raw” and/
or assembled sequence data to resolve the community metab-
olism (i.e., how functions are distributed and identification
of strain and/or environment-specific community traits).
Since most community genotypes are unknown (25), de
novo assembly is often employed to assemble the community
gene content prior to annotation. Depending on the average
length of the metagenome contigs, full-length or partial pro-
tein coding regions can be called for selected (as per length
and coverage cut-off ) contigs using various gene calling algo-
rithms. Predicted protein coding sequences can be compared
using “best BLAST hit” or pattern matching methodology
against various reference protein databases to type (classify)
the community metabolism at various hierarchy levels from
enzymes to pathways. Genome assembly algorithms are
designed to generate long nucleotide stretches based on
inherited characteristics of “input” raw sequences; therefore,
genetically divergent (below the set cut-off ) individual
metagenome reads (singletons) do not take part in assembly
and finally in community functional typing. Unlike low-
abundance ecosystems (i.e., those representing community
genotypes with low genetic diversity and evenness (minimum
genetic rearrangements) (5), the percentage of reads that par-
ticipate in genome assembly is much lower in complex envi-
ronments such as soil and sediments (26). However, an
estimate based on pre–shotgun sequencing surveys of marker
genes such as the small subunit (SSU or 16S) rRNA genes (5)
and/or large-scale sequencing efforts (27) can result into a
“good” assembly and can provide statistically reliable taxo-
nomical and functional insights even from the “rare
biosphere” (i.e., less abundant but highly diverse taxa) of
the studied ecosystem. Assembly of shotgun metagenomic
sequence data derived from complex ecosystems yields a
high proportion of individual reads that do not assemble
into contigs (singletons). In such common situations, ORF-
based functional annotation will fail for many short sequen-
ces, leading to the underrepresentation of the full community
metabolic potential. In complex environments, individual
metagenome reads can be compared against reference pro-
tein databases using iteratively selected and “moderately
stringent” alignment parameters (e.g., iteratively selected
BLASTX parameters; e-value, percentage identity, and query
length cut-offs).
A combined approach can be employed to provide im-
proved coverage of the genetic components of a microbial
community and community metabolism. Such an approach
includes classifying protein coding sequences derived from
de novo assembled contigs, corrected for misassembly, using
BLASTX or tBLASTX. Subsequently, singletons are then
screened against reference protein databases using iteratively
stringent parameters. Results from direct functional typing of
individual reads must be standardized by coverage parameters
(i.e., metagenome data set must be nonredundant). Using the
combined approach outlined here, it is possible to deal with
differences originating from “individual metagenome reads”
and “postassembly ORF-based functional typing” independ-
ently (Fig. 2).
REFERENCE DATABASES: INDIVIDUAL
FUNCTIONAL UNITS TO PATHWAYS AND
SUBSYSTEMS
“Similarity-based” gene finding and annotation is the most
common approach for the analysis of shotgun metagenomic
sequence data and has been implemented in most metage-
nomic surveys (e.g., 2, 5, 6, 26, 27). The core of the similarity-
based functional annotation is to compare the individual
reads, assembled contigs, or translated amino acid sequences
(in all six frames) from protein coding sequences, against
existing reference protein databases. Metagenomic gene
 2.4.4. Metagenomics: Assigning Functional Status to Community Gene Content ▪ 2.4.4-3
FIGURE 2 Relative potential of similarity-based methods used in
metagenome functional annotation. Method_A: BLASTP-based
comparison of protein sequences ( predicted from contigs generated
by de novo assembly of raw reads) against a protein database.
Method_B: Direct comparison (BLASTX; e value = 10−5) of indi-
vidual metagenome reads against protein database. Ten thousand
randomly selected metagenome reads (average read length = 300
bases) from hexachlorocyclohexane-contaminated soil (acces-
sion no: SRX0964712; [21]) and ORFs predicted (minimum read
length = 90 amino acid) from the metagenomic contigs were
compared (BLASTX; e value = 10–5) against the STRING (33)
database using methods B and A, respectively. The number of unique
protein families were compared against metagenome reads and ORFs
and Pearson correlation coefficient (PCC) was calculated. Owing to
its high coverage based characteristic (variation in each read gets
compared individually) Method_B provides better annotation
results with more number of unique protein families per read.
doi:10.1128/9781555818821.ch2.4.4.f2
content can be classified into a wide range of functional cat-
egories using these comparative approaches, and the effec-
tiveness of such an approach will improve with regular
updating of the reference databases, increased computational
resources, and efficient algorithms (e.g., 28).
Metagenomic sequence functional typing can follow the
hierarchical order from individual units (enzymes, gene
ontology and domains), protein families and profiles, protein
interactions to pathways and subsystems (Fig. 1). For example,
the NCBI nonredundant, RefSeq (29), SMART (30),
M5NR (31), and UniProt/UniRef (32) databases provide a
comprehensive collection of reference protein sequences,
including sequences from environmental studies for similar-
ity-based annotation. Web-based servers such as CAMERA
(33), MG-RAST (34), and IMG/M (35) also provide
sequence data sets and annotation results from metagenomic
surveys representing a diverse array of environments. Individ-
ual metagenomic reads, assembled contigs, or inferred amino
acid sequences derived from protein coding sequences can be
assigned to individual functional units (enzymes, gene ontol-
ogy [GO] and domains) by comparing sequences against
databases of clusters of orthologous groups (COGs [36]) and
GO databases. Similarity to various protein families can be
determined by comparing metagenomic data sets (i.e., pre-
dicted amino acids) against Pfam (37) and TIGRFAM (38)
databases. Annotation using the Pfam database generally pro-
vides better resolution than TIGRFAM, due to the large dif-
ference in the number of protein families in each database
(13,672 in the Pfam 26.0 release, 4,209 in the TIGRFAM
12.0 release). Protein–protein interactions and metabolic
pathways can be predicted for any data set via comparison
against STRINGS (39), SEED (40), and KEGG (41) data-
bases. Functional characterization at each level of this hier-
archy, however, suffers from the inherent complexities of
the metagenome sequences (Fig. 1). These complexities
include (a) information is derived from short sequence read
lengths, (b) high genetic diversity resulting in incomplete
coverage of individual organism genomes, and (c) limited
sequencing depth (42). Functional typing of the metagenome
data sets with shorter read length (Roche 454 pyrosequenc-
ing data, 400–800 bases), Illumina (100–500 bases, if paired
reads are merged), Ion Torrent (100–400 bases)) is more
problematic than longer reads generated by capillary electro-
phoresis (Sanger) sequencing. Nonetheless, short read data
are almost exclusively used in rapid annotation analyses,
such as those performed by the MG-RAST server (34).
This is particularly relevant for metagenomic analysis of
highly complex microbial communities (e.g., soil and sedi-
ment [26]) where de novo assemblies generate short contigs
and the majority of sequences remain unassembled).
The inability of the current gene prediction algorithms
to process short sequences (4) makes it more difficult to cor-
rectly assign putative function to fragmentary environmental
sequences; individual reads or assembled contigs shorter than
the average length of the prokaryotic protein coding sequence
(∼1,000 bp) can cause under- and overrepresentation of spe-
cific components of the community functional potential.
Improvements in next-generation sequencing technology
will likely result in improved identification of unassembled
reads. Therefore, instead of predicting protein coding sequen-
ces by comparing individual metagenome reads to reference
databases (Fig. 2), iteratively optimized (stringency of the
parameters, e.g., BLASTX; e-value iterations) similarity-
based comparison tools can be a more efficient (higher
number of classified sequences) way to annotate and infer
functional potential of complex microbial communities
(26, 43).
RECONSTRUCTION OF COMMUNITY
METABOLISM AND CORRELATING
METABOLISM WITH GENETIC DIVERSITY
Classic metagenomic studies (e.g., 5, 6, 26, 27) revealing the
microbial potential of various ecological niches have demon-
strated the role of microbial functional dynamics in shaping
the biogeochemistry of the habitat. Current community-wide
metabolic reconstruction approaches primarily include soft-
ware tools and platforms such as MG-RAST (34), IMG/M
(35), MEGAN (44), metaSHARK (45), CAMERA (33),
and HUMAnN (46). The first four of these rely on a BLAST-
based methodology; that is, similarity searches of individual
metagenomic reads or of ORFs from assembled contigs,
against reference protein databases. The BLAST output is
parsed according to various parameters, and taxonomic
and metabolic information can be assigned to each sequence
at various functional and taxonomic levels. CAMERA
(33), MG-RAST (34), and IMG/M (35) are freely available,
access regularly updated and comprehensive databases, and
2.4.4-4 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

provide cloud computing–based automated analysis and
basic visualization of data. M5NR (31), COG (36), SEED
(39), and KEGG (41) are the main reference protein data-
bases included in these public resources (33, 34, 35).
MEGAN (44) is a stand-alone program that inputs the
translated BLAST results and provides functional annotation
results depicted via rooted tree with internal nodes represent-
ing the subsystems and leaves as pathways (functional roles).
Currently, MEGAN (version 5) can be used with SEED- and
KEGG-based classifications with over 13,000 nodes in each
database. Additionally, MEGAN provides statistical analysis
for performing comparative analysis. The functional poten-
tial of multiple microbiomes can be compared by calculating
ecological indices (e.g., Goodall’s index, 47) calculated for
every individual microbiome separately. Results can be
visualized using “split-network” or “neighbor-joining” algo-
rithms within MEGAN, and can also be analyzed using stand-
ard hierarchical clustering and ordination strategies such as
nonmetric multidimensional scaling. To correlate the com-
munity functional potential with phylogenetic diversity,
MEGAN implements the lowest common ancestor (LCA)
algorithm, in which each metagenome read is assigned to a
taxon based on its gene conservation inferred from homology
search analysis. The BLAST2LCA program (https://github.
com/emepyc/Blast2lca) also uses the LCA algorithm and
can provide both functional and taxonomical identities to
the fragmented environmental DNA sequences.
Despite recent advances in computing, evidence-based
approaches (Table 1) are unable to accurately quantify
community-wide functional traits using deeply sequenced
but unassembled data sets composed only of short sequences.
In addition, these methods lack rigorous statistical correc-
tions: coverage ( presence/absence) and abundance of func-
tional pathways, especially over taxonomical correlations
(i.e., reconstructed traits linked to individual microbial gene
families). The software package HUMAnN (46) provides a
statistically corrected and quantified overview of community
metabolism. HUMAnN bypasses the need for assembly and
utilizes translated BLAST (BLASTX) results as the primary
data input. The software identifies genes (orthologous fami-
lies) via weighted sum of hits selected over BLAST parame-
ters (i.e., e-value, query coverage). Relative abundance and
coverage ( presence or absence) can be calculated for each
identified module and pathway. HUMAnN can process
output from various translated BLAST implementations
such as NCBI-BLASTX (7), MAPX (Real Time Geno-
mics, San Francisco, CA), and USEARCH (48). There are
two major advantages of HUMAnN over “best BLAST
hit” approaches. First, the software removes taxonomically
divergent traits (false positives) and provides a gap-filling
(false negative) step before the final output. Second, the
software facilitates diversity index calculation to estimate
the in situ functional diversity (evenness and richness).
Although all of the current community-level functional
reconstruction approaches provide qualitative and semi-
quantitative overviews of community metabolism, the major
limitation of these approaches is their dependency on existing
reference protein databases. This limitation prevents the
identification and characterization of new protein families
and may fail to identify novel genes and organisms within
samples.
CROSS-VALIDATION (REPLICATES) VERSUS
DEEP COVERAGE
Classic metagenome surveys describing the microbial poten-
tial of various environments such as acid mine drainage
(5), soil/permafrost (27), cow rumen (49), and marine global
ocean sampling (50) have provided insight into the func-
tional potential of the organisms within these complex
microbial communities. Due to the absence of sufficient num-
bers of replicates, these studies have been described as “obser-
vational” and without rigorous statistical analysis (51). The
field of microbial ecology has been recently chided for

TABLE 1 Computational resources for community metabolism reconstruction

S.no Name Availability Accessibility Resourcesa Input Toolsb
1. MG-RAST http://metagenomics.anl. Public/Web server M5nr, SEED, KEGG, Raw reads/ BLAT, Perl, BioPython
gov/ PATRIC, RefSeq contigs
etc.
2. CAMERA http://camera.calit2.net/ Public/Web server KEGG, Pfam, Raw reads/ BLAST, HMMER
TIGARfams contigs
3. IMG/M https://img.jgi.doe.gov/ Public/Web server COGs, KEGG, Pfam, Raw reads BLAST, HMMER
cgi-bin/m/main.cgi TIGARfams
4. GOmixer http://www.raeslab.org/ Public/Web server Bio-Cyc, KEGG Annotation Mapping of enzymes
gomixer/ file to reference metabolic
libraries
5. metaSHARK http://metadatabase.org/ Public/Web server PRIAM Annotation Enzyme profile search
wiki/MetaSHARK file
6. MEGAN http://ab.inf.uni- Stand-alone KEGG, SEED, GO Raw reads/ BLASTx
tuebingen.de/software/ ontology, nr contigs
megan5/
7. HUMAnN http://huttenhower.sph. Stand-alone KEGG, COGs, NOGs BLAST BLASTX, MAPX,
harvard.edu/humann output USEARCH
8. MinPath http://omics.informatics. Stand-alone and KEGG, SEED BLAST Integer programming
indiana.edu/MinPath/ Web-server output
a
Reference databases that can be used for comparison.
b
Algorithms and methods used to generate and parse the results.
 2.4.4. Metagenomics: Assigning Functional Status to Community Gene Content ▪ 2.4.4-5
insufficient replication in microbial compositional assays,
with particular respect to rRNA gene amplicon libraries
(51). As library preparation and sequencing prices drop,
increased replication will likely follow. Currently, library
preparation and deep sequencing can be costly, and there is
still considerable debate in the field as how to best spend
sequencing dollars. The options include greater replication
(independent sequencing libraries with unique barcodes)
and shallower sequencing depth in each replicate, or fewer
(or no) replicates with deeper overall coverage. Gilbert and
collaborators (52) sampled the microbial community genetic
potential across the western English Channel over a
time-series gradient using shallow metagenomic sequencing
and have proven that replicated samples (sampled across gra-
dient and time series) can increase our understanding to infer
the microbial functional dynamics, supported with rigorous
statistical analysis.
Ultimately, the choice between “deeply sequence few
samples” or “shallow sequencing with multiple replicates”
depends on the aim of the survey and the nature of the ecosys-
tem, but we should heed our previous warnings and ensure
at least a minimum level of replication when attempting
to compare community composition or genetic composition
between samples (53). Furthermore, this choice may also
be considered a false dichotomy—namely, one can generate
independent biological replicates for metagenomic sequenc-
ing to determine if statistically significant differences exist
between different sample types or environments. These bio-
logical replicates can then be bioinformatically pooled for
combined sequence analysis. The limitation of this approach
is the cost in generating and sequencing the independent
libraries for each replicate, but the cost of libraries is expected
to continue to drop.
Deep sequencing is required to assemble and unravel the
genetic composition of uncharacterized microorganisms.
Such sequence data can be used to reconstruct dominant or
key genotypes playing important roles in the ecosystem
biogeochemistry (27, 49). To infer the topology of the ecolog-
ical structure of any environment, one can opt for the shallow
sequencing with numerous replicates but to increase the
resolution of taxonomical and functional dynamics, sampl-
ing must be broad and selected over time and/or all other local
variables (e.g., physiochemical). However, using SSU rRNA
gene distribution patterns corrected over microbial genome
size and SSU rRNA gene copies per genome variations
(53), one can certainly determine the sequencing depth
require to represent the actual in situ microbial dynamics.
High sequence coverage and low genetic diversity provide
better assembly results and as-yet-uncultivated genotypes
can be reconstructed both from dominant population and
the “rare biosphere” (18, 19).
METADATA AND EXPERIMENTAL DESIGN
The value of sequence data increases with associated meta-
data acquired during sampling. In every metagenome analysis
targeted to reconstruct the community metabolism and in situ
microbial dynamics, there are variables needed to decide sam-
pling breadth, sequence data generation (sequencing plat-
form and depth), and are also essential for interpretation of
the results. These metadata include environmental (biotic
and abiotic) and experimental variables and can include
(but are not limited to) sampling features such as temperature,
latitude/longitude, humidity, physicochemical data, sample
collection/handing, and methods of nucleic acid extraction
and sequencing. Recently, a report was published detailing
recommendations regarding the minimum information that
should be reported along with sequence data (54). The mini-
mum information about any (x) sequence (54) standards pro-
vide a robust framework of essential technical information
that can play a key role in better experimental design and
interpretation of results and allow for more robust cross-study
analyses.
The work was supported by Grants from Department of Biotechnol-
ogy (DBT), Government of India under project BT/PR3301/BCE/8/
875/11.
REFERENCES
1. Whitman WB, Coleman DC, Wiebe WJ. 1998. Prokar-
yotes: the unseen majority. Proc Natl Acad Sci USA 95:
6578–6583.
2. Dinsdale EA, Edwards RA, Hall D, Angly F, Breitbart M,
Brulc JM, Furlan M, Desnues C, Haynes M, Li L, McDa-
niel L, Moran MA, Nelson KE, Nilsson C, Olson R, Paul J,
Brito BR, Ruan Y, Swan BK, Stevens R, Valentine DL,
Thurber RV, Wegley L, White BA, Rohwer F. 2008. Func-
tional metagenomic profiling of nine biomes. Nature 452:
629–632.
3. Pace NR. 2009. Mapping the tree of life: progress and pros-
pects. Microbiol Mol Biol Rev 73:565–576.
4. Kunin V, Copeland A, Lapidus A, Mavromatis K, Hugen-
holtz P. 2008. A bioinformatician’s guide to metagenomics.
Microbiol Mol Biol Rev 72:557–578.
5. Tyson GW, Chapman J, Hugenholtz P, Allen EE, Ram RJ,
Richardson PM, Solovyev VV, Rubin EM, Rokhsar DS,
Banfield JF. 2004. Community structure and metabolism
through reconstruction of microbial genomes from the envi-
ronment. Nature 428:37–43.
6. Sangwan N, Verma H, Kumar R, Negi V, Lax S, Khurana
P, Khurana JP, Gilbert JA, Lal R. 2014. Reconstructing an
ancestral genotype of two hexachlorocyclohexane-degrad-
ing Sphingobium species using metagenomic sequence
data. ISME J 8:398–408.
7. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ.
1990. Basic local alignment search tool. J Mol Biol 215:
403–410.
8. Badger JH, Olsen GJ. 1999. CRITICA: coding region iden-
tification tool invoking comparative analysis. Mol Biol Evol
16:512–524.
9. Frishman D, Mironov A, Mewes HW, Gelfand M. 1998.
Combining diverse evidence for gene recognition in com-
pletely sequenced bacterial genomes. Nucleic Acids Res 26:
2941–2947.
10. Kaelbling LP, Littman ML, Cassandra AR. 1998. Planning
and acting in partially observable stochastic domains. Artif
Intell 101:99–134.
11. Borodovsky M, Lomsadze A, Ivanov N, Mills R. 2003.
Eukaryotic gene prediction using GeneMark.hmm. Curr
Protoc Bioinformatics 1:4.6:4.6.1–4.6.12.
12. Kelley DR, Liu B, Delcher AL, Pop M, Salzberg SL. 2012.
Gene prediction with Glimmer for metagenomic sequences
augmented by classification and clustering. Nucleic Acids Res
40:e9.
13. Hyatt D, Chen GL, Locascio PF, Land ML, Larimer FW,
Hauser LJ. 2010. Prodigal: prokaryotic gene recognition
and translation initiation site identification. BMC Bioinfor-
matics 11:119.
14. Noguchi H, Park J, Takagi T. 2006. MetaGene: prokaryotic
gene finding from environmental genome shotgun sequen-
ces. Nucleic Acids Res 34:5623–5630.
15. Kelley DR, Salzberg SL. 2010. Clustering metagenomic
sequences with interpolated Markov models. BMC Bioinfor-
matics 11:544.
2.4.4-6 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
16. Boisvert S, Raymond F, Godzaridis E, Laviolette F, Corbeil
J. 2012. Ray Meta: scalable de novo metagenome assembly
and profiling. Genome Biol 13:R122.
17. Albertsen M, Hugenholtz P, Skarshewski A, Nielsen KL,
Tyson GW, Nielsen PH. 2013. Genome sequences of
rare, uncultured bacteria obtained by differential cover-
age binning of multiple metagenomes. Nat Biotechnol 31:
533–538.
18. Wrighton KC, Thomas BC, Sharon I, Miller CS, Castelle
CJ, VerBerkmoes NC, Wilkins MJ, Hettich RL, Lipton
MS, Williams KH, Long PE, Banfield JF. 2012. Fermenta-
tion, hydrogen, and sulfur metabolism in multiple unculti-
vated bacterial phyla. Science 337:1661–1665.
19. Sharon I, Morowitz MJ, Thomas BC, Costello EK, Relman
DA, Banfield JF. 2013. Time series community genomics
analysis reveals rapid shifts in bacterial species, strains, and
phage during infant gut colonization. Genome Res 23:
111–120.
20. Mohammed MH, Ghosh TS, Singh NK, Mande SS. 2011.
SPHINX—an algorithm for taxonomic binning of metage-
nomic sequences. Bioinformatics 27:22–30.
21. Johannes A, Bjarnason BS, de Bruijn I, Schirmer M,
Quick J, Ijaz UZ, Lahti L, Loman NJ, Andersson AF,
Quince C. 2014. Binning metagenomic contigs by coverage
and composition. Nat Meth doi: 10.1038/nmeth.3103.
22. Stamps BW, Corsetti FA, Spear JR, Stevenson BS. 2014.
Draft genome of a novel Chlorobi member assembled by tet-
ranucleotide binning of a hot spring metagenome. Genome
Announce 2:e00897–e00914.
23. Wu YW, Tang YH, Tringe SG, Simmons BA, Singer
SW. 2014. MaxBin: an automated binning method to
recover individual genomes from metagenomes using an
expectation-maximization algorithm.Microbiome 2:1–18.
24. Wang Y, Leung H, Yiu S, Chin F. 2014. MetaCluster-TA:
taxonomic annotation for metagenomic data based on
assembly-assisted binning. BMC Genomics 15:S12.
25. Wooley JC, Godzik A, Friedberg I. 2010. A primer on
metagenomics. PLoS Comput Biol 6:e1000667.
26. Sangwan N, Lata P, Dwivedi V, Singh A, Niharika N,
Kaur J, Anand S, Malhotra J, Jindal S, Nigam A, Lal D,
Dua A, Saxena A, Garg N, Verma M, Kaur J, Mukherjee
U, Gilbert JA, Dowd SE, Raman R, Khurana P, Khurana
JP, Lal R. 2012. Comparative metagenomic analysis of soil
microbial communities across three hexachlorocyclohexane
contamination levels. PLoS One 7:e46219.
27. Mackelprang R, Waldrop MP, DeAngelis KM, David MM,
Chavarria KL, Blazewicz SJ, Rubin EM, Jansson JK. 2012.
Metagenomic analysis of a permafrost microbial community
reveals a rapid response to thaw. Nature 480:368–71.
28. Huson DH, Xie C. 2013. A poor man’s BLASTX-high-
throughput metagenomic protein database search using
PAUDA. Bioinformatics 30:38–39.
29. Pruitt KD, Tatusova T, Brown GR, Maglott DR. 2012.
NCBI Reference Sequences (RefSeq): current status, new
features and genome annotation policy. Nucleic Acids Res
40:D130–D135.
30. Letunic I, Doerks T, Bork P. 2012. SMART 7: recent
updates to the protein domain annotation resource. Nucleic
Acids Res 40:D302–D3025.
31. Wilke A, Harrison T, Wilkening J, Field D, Glass EM,
Kyrpides N, Mavrommatis K, Meyer F. 2012. The M5nr:
a novel non-redundant database containing protein sequen-
ces and annotations from multiple sources and associated
tools. BMC Bioinformatics 13:141.
32. Suzek BE, Huang H, McGarvey P, Mazumder R, Wu CH.
2007. UniRef: comprehensive and non-redundant UniProt
reference clusters. Bioinformatics 23:1282–1288.
33. Sun S, Chen J, Li W, Altintas I, Lin A, Peltier S, Stocks K,
Allen EE, Ellisman M, Grethe J, Wooley J. 2011.
Community cyberinfrastructure for Advanced Microbial
Ecology Research and Analysis: the CAMERA resource.
Nucleic Acids Res 39:D546–D451.
34. Meyer F, Paarmann D, D’Souza M, Olson R, Glass EM,
Kubal M, Paczian T, Rodriguez A, Stevens R, Wilke A,
Wilkening J, Edwards RA. 2008. The metagenomics
RAST server—a public resource for the automatic phyloge-
netic and functional analysis of metagenomes. BMC Bioin-
formatics 9:386.
35. Markowitz VM, Chen IM, Chu K, Szeto E, Palaniappan K,
Jacob B, Ratner A, Liolios K, Pagani I, Huntemann M,
Mavromatis K, Ivanova NN, Kyrpides NC. 2012. IMG/
M-HMP: a metagenome comparative analysis system for
the Human Microbiome Project. PLoS One 7:e40151.
36. Tatusov RL, Galperin MY, Natale DA, Koonin EV.
2000. The COG database: a tool for genome-scale analysis
of protein functions and evolution. Nucleic Acids Res 28:
33–36.
37. Bateman A, Coin L, Durbin R, Finn RD, Hollich V,
Griffiths-Jones S, Khanna A, Marshall M, Moxon S, Sonn-
hammer EL, Studholme DJ, Yeats C, Eddy SR. 2004.
The Pfam protein families database. Nucleic Acids Res 32:
D138–D141.
38. Haft DH, Loftus BJ, Richardson DL, Yang F, Eisen JA,
Paulsen IT, White O. 2001. TIGRFAMs: a protein family
resource for the functional identification of proteins. Nucleic
Acids Res 29:41–43.
39. Szklarczyk D, Franceschini A, Kuhn M, Simonovic M,
Roth A, Minguez P, Doerks T, Stark M, Muller J, Bork
P, Jensen LJ, von Mering C. 2011. The STRING database
in 2011: functional interaction networks of proteins, glob-
ally integrated and scored. Nucleic Acids Res 39:D561–
D568.
40. Overbeek R, Begley T, Butler RM, Choudhuri JV, Chuang
HY, Cohoon M, de Crécy-Lagard V, Diaz N, Disz T,
Edwards R, Fonstein M, Frank ED, Gerdes S, Glass EM,
Goesmann A, Hanson A, Iwata-Reuyl D, Jensen R,
Jamshidi N, Krause L, Kubal M, Larsen N, Linke B,
McHardy AC, Meyer F, Neuweger H, Olsen G, Olson R,
Osterman A, Portnoy V, Pusch GD, Rodionov DA, Rück-
ert C, Steiner J, Stevens R, Thiele I, Vassieva O, Ye Y,
Zagnitko O, Vonstein V. 2005. The subsystems approach to
genome annotation and its use in the project to annotate 1000
genomes. Nucleic Acids Res 33:5691–5702.
41. Ogata H, Goto S, Sato K, Fujibuchi W, Bono H, Kanehisa
M. 1999. KEGG: Kyoto Encyclopedia of Genes and
Genomes. Nucleic Acids Res 27:29–34.
42. Mende DR, Waller AS, Sunagawa S, Järvelin AI, Chan
MM, Arumugam M, Raes J, Bork P. 2012. Assessment of
metagenomic assembly using simulated next generation
sequencing data. PLoS One 7:e31386.
43. Prakash T, Taylor TD. 2012. Functional assignment of
metagenomic data: challenges and applications. Brief Bioin-
form 13:711–727.
44. Huson DH, Mitra S, Weber N, Ruscheweyh H, Schuster
SC. 2011. Integrative analysis of environmental sequences
using MEGAN4. Genome Res 21:1552–1560.
45. Pinney JW, Shirley MW, McConkey GA, Westhead DR.
2005. metaSHARK: software for automated metabolic net-
work prediction from DNA sequence and its application to
the genomes of Plasmodium falciparum and Eimeria tenella.
Nucleic Acids Res 33:1399–1409.
46. Abubucker S, Segata N, Goll J, Schubert AM, Izard J,
Cantarel BL, Rodriguez-Mueller B, Zucker J, Thiagarajan
M, Henrissat B, White O, Kelley ST, Methé B, Schloss
PD, Gevers D, Mitreva M, Huttenhower C. 2012.
Metabolic reconstruction for metagenomic data and its
application to the human microbiome. PLoS Comput Biol
8:e1002358.
 2.4.4. Metagenomics: Assigning Functional Status to Community Gene Content ▪ 2.4.4-7
47. Goodall DW. 1966. A new similarity index based on proba-
bility. Biometrics 22:882–907.
48. Edgar RC. 2010. Search and clustering orders of magnitude
faster than BLAST. Bioinformatics 26:2460–2461.
49. Hess M, Sczyrba A, Egan R, Kim TW, Chokhawala H,
Schroth G, Luo S, Clark DS, Chen F, Zhang T, Mackie
RI, Pennacchio LA, Tringe SG, Visel A, Woyke T,
Wang Z, Rubin EM. 2011. Metagenomic discovery of
biomass-degrading genes and genomes from cow rumen.
Science 331:463–467.
50. Williamson SJ, Rusch DB, Yooseph S, Halpern AL,
Heidelberg KB, Glass JI, Andrews-Pfannkoch C, Fadrosh
D, Miller CS, Sutton G, Frazier M, Venter JC. 2008. The
Sorcerer II Global Ocean Sampling Expedition: metage-
nomic characterization of viruses within aquatic microbial
samples. PLoS One 3:e1456.
51. Prosser JI. 2010. Replicate or lie. Environ Microbiol 12:
1806–1810.
52. Gilbert JA, Field D, Swift P, Thomas S, Cummings D,
Temperton B, Weynberg K, Huse S, Hughes M, Joint I,
Somerfield PJ, Mühling M. 2010. The taxonomic and func-
tional diversity of microbes at a temperate coastal site: a
“multi-omic” study of seasonal and diel temporal variation.
PLoS One 5:e15545.
53. Ni J, Yan Q, Yu Y. 2013. How much metagenomic sequenc-
ing is enough to achieve a given goal? Sci Rep 3:1968.
54. Glass EM, Dribinsky Y, Yilmaz P, Levin H, Van Pelt R,
Wendel D, Wilke A, Eisen JA, Huse S, Shipanova A,
Sogin M, Stajich J, Knight R, Meyer F, Schriml LM.
2014. MIxS-BE: a MIxS extension defining a minimum
information standard for sequence data from the built envi-
ronment. ISME J 8:1–3.
 Generation and Analysis of Microbial Metatranscriptomes
NEHA SARODE, DARREN J. PARRIS, SANGITA GANESH,
SHERRY L. SESTON, AND FRANK J. STEWART
2.4.5
INTRODUCTION
Analysis of the collective RNA pool from a microbial
community—the metatranscriptome—yields valuable infor-
mation on microbial gene expression patterns and biogeo-
chemical processes in natural environments. Molecular
and analytical tools for analyzing metatranscriptomes using
high-throughput sequencing have advanced rapidly in recent
years and continue to evolve and expand (see Chapter 2.4.1).
The technique is increasingly available to individual research
projects, even those with a modest budget or lacking an exten-
sive bioinformatics toolkit. However, key considerations
should be addressed before beginning a metatranscriptomic
study. This chapter addresses these considerations and then
describes in detail the major steps of a sequencing-based meta-
transcriptomic analysis, from RNA isolation to sequence
analysis.
Metatranscriptomics, sometimes known as RNA-Seq, has
been applied successfully to microbial communities from
diverse natural environments. The method, using high-
throughput shotgun sequencing to identify transcripts, was
first applied to microbial communities from the oligotrophic
surface ocean (1). In contrast to array-based methods, this
shotgun sequencing approach does not require a priori knowl-
edge of the community sequence pool, and is therefore able to
identify novel transcripts. This method also provides an
important contrast to prior techniques for interrogating com-
munity transcription, namely, microarrays. Metatranscrip-
tome data sets, by yielding counts of discrete genes, provide
a quantitative comparison of the relative transcription of dif-
ferent genes, typically highlighting a core set of genes whose
transcripts dominate a data set and an exceedingly long tail of
genes represented only sparingly (fewer than five sequences).
This general pattern, with fluctuation in the overall domi-
nance of the most highly transcribed genes, has been observed
repeatedly in environments ranging from soils to the human
gut (2, 3). The nature of this distribution poses challenges,
notably with regard to ensuring sequence coverage levels
sufficient for detecting the differential expression of low-
abundance transcripts.
Despite this challenge, metatranscriptomics has been
widely adopted, and a core set of practices can now be identi-
fied. Key steps include RNA extraction, messenger RNA
(mRNA) enrichment, synthesis of complementary DNA
(cDNA), shotgun sequencing of cDNA, and bioinformatic
doi:10.1128/9781555818821.ch2.4.5
2.4.5-1
analysis of sequence data. Embedded within this general
framework are a number of important questions that should
be considered before beginning an analysis.
Do I Need Metatranscriptome Data to Explore
Community Function?
This question should be addressed when considering experi-
mental design and research objectives. In some cases where
expression of a small set of specific genes is of the most inter-
est, gene-targeted approaches such as reverse transcription
quantitative PCR may be more cost-effective and efficient,
especially if the experimental design calls for large numbers
of samples and replicates. Gene expression microarrays can
also facilitate the analysis of a large number of samples if there
is enough information on the gene content of the microbial
community to create an array. However, if the research objec-
tive is to describe and quantify community function based on
gene content and expression, then metagenomic, metatran-
scriptomic, or metaproteomic analysis is appropriate.
The question then becomes which “omic” method should
be applied. The metagenome represents the metabolic poten-
tial of a community, whereas the metatranscriptome repre-
sents metabolic activity. In many cases, metagenome and
metatranscriptome data are correlated (2). This is logical—
transcript abundance depends on the number of copies of a
gene (either within a genome or across the genomes of multi-
ple individuals) available to be transcribed.
The community pool of proteins—the metaproteome—
also provides a snapshot of community functional activity,
but this is integrated over a different time scale. Protein con-
tent per cell turns over on the scale of days, whereas mRNA
can be degraded in minutes (4–6). The metatranscriptome
therefore is presumably a more accurate proxy of instantane-
ous physiological state. It is, however, an imperfect proxy.
mRNA turnover rates vary among cells depending on activity
levels and also among genes within a genome (5, 7, 8).
Furthermore, mRNA levels may be decoupled from those
of other RNA pools, notably ribosomal RNA (rRNA),
which can be refractory over hours or days partly due to its
association with proteins (9, 10). Therefore, inferences of
physiological activity based on mRNA abundances should
be interpreted as hypotheses. Some of these hypotheses
can be validated by coupled biochemical rate measurements
2.4.5-2 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
(e.g., using mesocosm experiments), although this practice is
surprisingly rare. Metatranscriptomes can also be analyzed
over temporal or environmental gradients to determine base-
line mRNA abundances and rates of change. Concurrent
sampling of other “omic pools” (genes, proteins, metabolites;
e.g., 11) is necessary to clarify the dynamics and relative infor-
mation content of each pool. Such multitiered “omic” studies
can be incredibly powerful but are expensive and analytically
challenging, and therefore rare.
How Much Data Do I Need?
Determining an appropriate level of sequencing is not
straightforward and is contingent on the biological ques-
tion, budgetary considerations, the relative activity levels
and taxonomic complexity of the sampled communities,
and the sequence preparation protocol (e.g., whether rRNA
has been removed from the sample). Studies of marine micro-
bial metatranscriptomes over the past six years have generated
an average of ∼120 Mb of cDNA sequence data per sample
(6). Following methods to reduce total rRNA content (see
below), on average approximately >50% of the sequences in
these studies corresponded to rRNA (6), the amount of
data representing the protein-coding transcript pool per sam-
ple was thus <∼60 Mb. This depth of sequencing is almost
certainly inadequate for detecting transcriptional patterns
in all but the most abundant taxa or the most actively tran-
scribed genes. Indeed, even at relatively high sequencing
levels (>1 million reads per sample), the metatranscriptome
will likely be dominated by a handful of taxa or genes, with
the majority of transcripts instead represented by only a
few copies (12). For most genes, increasing the sequence
count per gene (coverage) is necessary to increase the power
of statistical methods to detect differences in transcription.
Fortunately, improvements in sequencing technology now
enable the generation of multiple gigabases (Gb) of data,
corresponding to millions to tens of millions of sequences,
in a matter of hours for less than $0.1 per Mb. High data yields
are especially important for transcriptomic samples where
rRNA is not subtracted (see later discussion). To detect
meaningful differences in transcription among many taxa
and genes, we suggest that millions or tens of millions of reads
per sample are necessary.
Can I Afford It?
Increasingly, the answer is yes. The majority of microbial
metatranscriptome studies over the past six years have used
Roche 454 technology for sequencing (6), due in part to
the longer read lengths afforded by this platform (now upward
of 1,000 bp). At current specifications, generating 1 Gb of
data using 454 costs ∼$10,000, or $1,200 for 120 Mb (13).
This cost is prohibitive if analyzing multiple samples, or
if deeper coverage levels (read counts per transcript) are
required, as is likely the case to statistically validate abun-
dance differences in low-frequency transcripts. Sequencing
in-house with the Roche 454 FLX instrument may also
be impractical for most labs due to high upfront costs
(∼$500,000), as well as the relatively high level of training
required to prepare samples and run the instrument. It is to
be noted that in October 2013, Roche announced the shut-
down of 454 Life Sciences and will discontinue support
for existing instruments by 2016.
Alternative sequencing technologies, notably the Illu-
mina platform, are increasing in popularity and now generate
sequence reads with sufficient length for accurate gene predic-
tion, at costs per base that are orders of magnitude lower
than that of the 454. The Illumina HiSeq 2500 instrument
can generate 150 × 150 bp paired end sequences as a cost of
∼$40–70 per Gb. The Illumina MiSeq, a benchtop instru-
ment designed for use by individual labs, costs ∼$300–400
per Gb but supports 300 × 300 paired-end sequencing, which
yields merged sequences with lengths rivaling those of the
454. A MiSeq can be purchased for ∼$125,000 with per
run costs of ∼$1,000 to $1,400. Such benchtop techno-
logies, and those of competing manufacturers, make in-house
sequencing a tangible option for individual labs. In October
2014 Illumina also launched two new instruments: the HiSeq
X Ten (only available as a set of 10 sequencers or more) cater-
ing exclusively to sequencing facilities (cost ∼$1 million
per instrument) and the benchtop NextSeq 500 (cost
∼$250,000), both currently supporting 150 × 150 paired-end
sequencing (∼$4,000 per run cost) and upward of >100 mil-
lion reads per sequencing run (http://www.illumina.com/sys
tems/sequencing.ilmn). However, for most researchers, the
most cost-effective option for metatranscriptome sequenc-
ing will involve sending cDNA samples to institutional
or external genomic facilities, which offer a range of sequenc-
ing platforms to accommodate specific goals and budgets.
Sequencing will become increasingly affordable as improve-
ments in technology and competition continue to drive rapid
changes in the market.
How Much Starting Material Is Required and Is RNA
Amplification Necessary?
Although the actual quantity of DNA/cDNA loaded onto
next-generation sequencers is often in the picogram or nano-
gram range, most methods for preparing sequencing libraries
advise starting with micrograms of double-stranded DNA.
Indeed, library preparation generally is more efficient and
less susceptible to error when input quantities are maximized.
Fortunately, new library preparation protocols, including the
Nextera DNA XT kit (Illumina) or the ScriptSeq RNA-Seq
kit (Epicentre) for generating libraries for Illumina sequenc-
ing, require as little as 1 ng (or less) of input template. How-
ever, some of these library preparation methods, notably the
Nextera kit, have been shown to impose a bias toward certain
sequences (typically related to guanine plus cytosine content)
and should be used with caution until further vetted for meta-
transcriptomics (14).
RNA yield is often variable and low in environmen-
tal samples. A typical marine bacterioplankton sample
(from ∼5–10 liters of seawater) yields 250 ng to 5 μg of total
RNA (often closer to the lower end of the range), depending
on extraction efficiency. In environments where microbial
abundances or activities are even lower, such as the deep-
subsurface or nutrient-starved regions of the ocean, recover-
able RNA quantities may be in the tens of nanograms or
less (15). Although yield can sometimes be improved by
increasing sampling volume, doing so may extend the time
required for sample collection, increasing the risk that tran-
scription changes during collection itself. Downstream proc-
essing prior to cDNA synthesis may involve a step to deplete
rRNA (see below). Since the total RNA fraction contains
on average >50% rRNA, rRNA depletion from total RNA
significantly reduces final RNA yield. Consequently, meet-
ing target input requirements for cDNA synthesis and
sequencing may require amplifying total RNA, typically via
an in vitro transcription reaction or as part of the cDNA
synthesis step. RNA amplification methods are often quite
efficient, yielding tens of micrograms of antisense RNA or
cDNA, which can serve as a surplus for multiple sequencing
runs (if required).
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-3
Should rRNA Be Removed?
The primary goal of most metatranscriptomic studies is to
examine patterns in the abundances of protein-coding
mRNA transcripts. To maximize data recovery, it is therefore
often desirable to enrich for mRNA, often by selectively
removing rRNA, which typically accounts for >80% of
total RNA in natural microbial communities (16). How-
ever, removing rRNA involves loss of information, as
rRNA sequences can provide valuable information on the
taxonomic composition of the transcriptionally active com-
munity (17). rRNA depletion also requires additional sample
manipulation, involving both time and money, as well as
some amount of RNA degradation, resulting in reduced total
RNA concentrations. Minor degradation will likely not affect
sequencing, as fragmentation is a component of sequenc-
ing library preparation. However, reduction of total RNA
amounts may pose a greater risk, depending on the starting
template requirements for downstream steps (e.g., cDNA
synthesis).
The cost effectiveness of rRNA depletion will vary
depending on choice of sequencing technology and depletion
method. A wide variety of custom or commercial methods are
available for rRNA depletion, most of which cost between
$15 and $100 per sample and differ in their removal efficiency
(16, and discussed later). Investing in a depletion step saves
money when using a relatively pricey sequencing technology,
such as Roche 454 pyrosequencing (see above). In contrast,
when using methods such as Illumina that cost only dollars
per gigabase, it is likely not cost effective to deplete rRNA,
as coverage can simply be increased by investing in additional
sequencing. However, choice of sequencing method may be
motivated by factors other than cost. Notably, pyrosequenc-
ing generates single-direction reads up to 1,000 bases in
length (mode: 450–700), compared to <300 for Illumina
technology. Longer reads offer a considerable advantage for
accurate gene prediction (18), potentially eliminating a
requirement for contig assembly. Depletion of rRNA will
likely remain a priority for pyrosequencing-based metatran-
scriptomic studies.
I Have Limited Bioinformatics Resources or
Experience. Will I Be Able to Analyze the Data?
Yes, you will. However, the efficiency and scope of the anal-
ysis will depend on the size of the sequence data sets, the com-
putational resources available, and the bioinformatics goals
and skills of the user. Raw sequence files may contain tens
of thousands to tens of millions of sequences and be several
gigabytes in size. Processing of these data sets, notably compu-
tationally intensive tasks of assembly or homology searching,
typically requires access to cluster-based computing, which
may not be available in individual environmental microbiol-
ogy labs. In lieu of personal/private hardware, external com-
putational resources such as high-performance computing
resources (e.g., cost-shared institutional clusters) and cloud
computing may be attractive options as they provide off-site
computing power and are typically maintained by a dedicated
support staff. The details of computational infrastructure and
logistics are beyond the scope of this chapter, and the reader is
directed to recent reviews on these topics (19–22).
The analysis of metatranscriptomes involves a diverse set
of tasks, ranging from simple text file management to statisti-
cal analysis of differential gene expression. Many analytical
tools are available to support analysis workflows—these tools
include custom (lab-specific) scripts run via the command
line as well as commercial or community resources. Many
custom scripts for managing large sequence data files are pub-
licly available as open-source programs. However, these appli-
cations typically perform only a limited set of data analysis
steps and lack user-friendly graphical user interfaces (GUI).
Running such scripts requires familiarity with scripting,
Unix and the command-line environment. Excellent books
and tutorials on using UNIX and Perl scripts for bioinfor-
matics are available for biologists who are unfamiliar with
these tools (23, 24). While a basic working knowledge of
scripting (e.g., in Perl, Python, or C++) can be very useful
for meta-omics analyses, the lack of this skill set will not pre-
vent users from processing sequence data. Indeed, several
web-based resources have been developed to enable analysis
of meta-omic data sets by nonspecialist users. Resources
such as Galaxy (25), MG-RAST (26), CAMERA (27, now
discontinued), and IMG/M (28) allow users to upload their
data and access a suite of bioinformatics tools for sequence
analysis and visualization. Compared to custom resources,
these online tools offer limited opportunities for specialized
sequence analysis tasks. Nonetheless, these tools are power-
ful and provide an attractive entry point for noncomputa-
tional biologists. Users are encouraged to consult the many
reviews on next-generation sequence analysis, as well as to
experiment with open-source custom scripts, which typically
contain step-by-step instructions and allow for easy modifica-
tion (29–32). In addition, commercial software packages for
integrated sequence analysis (e.g., CLC Genomics Work-
bench, JMP Genomics) offer intuitive processing via graphi-
cal user interfaces. However, these platforms are expensive
and were not developed for use with multispecies community
sequence data.
How Are Results Expressed?
The currency in metatranscriptomic analysis, as in metage-
nomics, is the “hit count”, the count of sequence reads in a
data set that fall within a given category, such as a type of
gene or the genes of a type of organism. To allow for compar-
isons across samples, hit counts are normalized to account for
uneven sequencing depth, often by expressing a gene’s hit
count as a proportion of the total number of protein-coding
reads in a data set, or the total number of reads mapping to
a reference sequence (33–36). Choice of normalization
method (discussed in more detail later) should be considered
carefully, as methods can differ in the relative weights
assigned to low- versus high-abundance transcripts. Normal-
ized hit count data are then used to statistically test for differ-
ential transcription patterns between samples, typically
by comparing either individual genes or groupings of genes
(e.g., functional pathways, genes of a specific organism).
It is important to recognize that normalized meta-omic
data sets allow for comparisons of the proportional abundan-
ces of genes within the sequence pool, but do not provide
absolute measurements of sequence numbers per sample vol-
ume. Shifts in the proportional abundance of a gene’s tran-
scripts should therefore be interpreted with caution, as they
can be influenced by expression changes in other genes
or organisms in the sample, as well as by shifts in absolute
abundance. This limitation has been partly overcome by
standardization methods, through which a known quantity
of exogenous RNA sequence is added to each sample (before
RNA isolation) as a reference, thereby enabling the calcula-
tion of transcript numbers per sample based on the percent
recovery of the reference in the sequenced data set (12).
This technique, although not without potential biases, can
provide valuable insight into the absolute transcriptional
activity of a community.
2.4.5-4 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
What’s Next for Metatranscriptomics?
Metatranscriptomics is a young and rapidly evolving field
with tremendous potential for advancements in all aspects
of the methodology, from sampling to data analysis. For cer-
tain environments, such as the deep ocean or subsurface,
new sampling techniques are necessary to enable in situ pres-
ervation of microbial RNA so as to minimize changes in com-
munity transcription during collection (37, 38). Obtaining
sufficient RNA amounts remains a primary challenge for
other environments, such as surface-associated microbiome
communties of the skin or on plants, where microbial cell
numbers are low or contribute a minor fraction to the bulk
RNA pool. Further declines in sequencing cost and improve-
ments in cDNA library preparation methods will likely
continue to enhance data recovery from low-abundance
organisms or communities.
Importantly, lowered sequencing costs will also help
broaden the use of metatranscriptomics as an experimental
tool. Manipulative experiments involving diverse treatments,
biological or technical replicates, or sampling across time
points have been largely impractical for metatranscriptomics
due to requirements for high sample numbers and the associ-
ated costs of sequencing and analysis. The field is now poised
for an exciting transition from its early descriptive stage to a
focus on more targeted ecological questions that can be
addressed through robust experimental designs and repeat
sampling. Such studies could benefit substantially from the
joint application of other rapidly developing meta-omic
methods, including metaproteomics and meta-metabolomics
(39, 40). Integrated, multi-omic studies, though still rare,
have the potential to yield unparalleled insight into complex
patterns of community metabolic regulation at multiple levels
of organization.
Increased sample numbers and sequencing depths require
concurrent improvements in bioinformatics infrastructure.
Sequencing advancements, spurred by private and public
investment and clear end goals (e.g., the $1,000 genome proj-
ect; 41), have outpaced advances in data analysis, sharing,
visualization, and integration with metadata (19). Meeting
this need requires enhancements in computational infra-
structure as well as education resources to train environmen-
tal microbiologists for key informatics tasks. The former can
be addressed by the continued development, optimization,
and distribution of federated tools (e.g., scripts, online ana-
lysis platforms) by the bioinformatics community, as well as
by investments in computing power through institutional
high-performance clusters or cloud computing. Cloud
computing has been touted as a solution for labs lacking
computational resources (20, 21, 42). However, valid con-
cerns about data security, privacy, and sharing logistics have
prevented many in the community from wholeheartedly uti-
lizing this resource. Training microbiologists to be proficient
in meta-omic analysis is equally challenging, due partly to a
paucity of formal training options (e.g., bioinformatics course
requirements for graduate students, training workshops).
Investments in such resources would help demystify basic
computational tasks that may appear daunting to nonspecial-
ists, and thereby broaden the use of meta-omics in environ-
mental microbiology.
DATA GENERATION
The following sections describe key steps in the generation of
microbial metatranscriptome data sets, from RNA preserva-
tion to sequencing. The methods described here were honed
primarily through analyses of aquatic bacterioplankton
samples but are widely applicable to analyses of community
RNA from diverse microbial habitats. Bioinformatic analysis
of sequence data is covered in the final section.
Sample Collection and RNA Preservation
Gene regulation responds near instantaneously to envi-
ronmental cues and mRNA half-lives can be as short as
1–2 min (5, 43). Hence, strategies for sampling micro-
bial community RNA should minimize the potential for
transcriptional changes during the collection procedure.
Sampling-induced environmental changes (e.g., in light, dis-
solved oxygen concentration, or temperature) bias commun-
ity transcriptional signals, making it critical to keep the time
between sample collection and RNA preservation as short
as possible.
Rapid RNA preservation may be especially challenging
when sampling deep water columns or sediments. Water
column sampling typically involves the collection of water
into bottles attached to a wire. Samples are winched to the
surface, and suspended biomass is concentrated onto filters
(e.g., Sterivex cartridge filters [0.22 μm pore size; Millipore]
with prefiltration through glass fiber filters [e.g., Whatman])
and preserved by flash freezing, addition of an RNA stabi-
lization solution, or a combination of both. The collection
period can be minimized by sampling only one depth at a
time and reducing filtering time by distributing sample water
over multiple replicate filters. The majority of aquatic meta-
transcriptome studies have focused on surface environments
where in situ conditions do not differ appreciably from those
on deck and for which the collection procedure can be kept
relatively short (typically less than 20 min). However, sam-
pling deeper in the water column, notably in regions with
sharp environmental gradients (e.g., oxygen minimum
zones), poses a significant risk of community change during
collection (37).
Methods for in situ RNA preservation are being explored
to address the possibility of transcriptional changes during
sample collection. Technologies such as the wire-mounted
Submersible Incubation Device–In situ Microbial Sampler
(Woods Hole Oceanographic Institution) or the autonomous
Environmental Sample Processor (44) have the capacity to
both filter and preserve multiple water column samples in
situ. Such custom platforms require considerable expertise
and expense for development and are not available for most
researchers, suggesting a need for smaller-scale technologies
for use by individual labs. Indeed, Feike et al. (37) recently
tested a wire-mounted in situ sampler that consists of a sample
bottle and a syringe that mixes fixative with the sample on
bottle closure. Use of this device in the suboxic zone of the
Baltic Sea showed that transcripts encoding ammonium oxi-
dation genes were up to 30-fold higher in samples obtained
using in situ preservation, compared to those preserved aboard
ship. Such discrepancies highlight the importance of rapid
collection and preservation methods for sampling environ-
mental RNA, notably from challenging environments such
as the deep sea.
For most microbial samples, preservation involves either
flash freezing in liquid nitrogen without the addition of an
RNA stabilization reagent, or the addition of a stabilization
reagent with or without freezing (45–47). Flash freezing is
commonly employed for soil samples, although commercial
preservation kits for soil RNA isolation are also available
(Lifeguard by Mo Bio Laboratories; 48–50). RNAlater, an
ammonium sulfate–based buffer originally marketed by
Ambion, is perhaps the most widely used stabilization reagent
in metatranscriptome studies. This preservative rapidly
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-5
penetrates cell walls and stabilizes RNA by precipitating pro-
teins, including RNA-degrading RNases. Simister et al. (51)
demonstrated slightly lower yield of total RNA for samples
preserved in RNAlater compared with direct freezing but
found no difference in the recovery of targeted mRNAs based
on preservation method. Ottesen et al. (45) showed that
marine microbial RNA remained stable without obvious deg-
radation for 4 weeks during storage in RNAlater. Other
reagents, including organic-based reagents such as Trizol
that both solubilize organic material and denature proteins,
have also been used successfully for RNA preservation in
the field. However, some of these reagents are relatively toxic.
While total RNA yield and degradation rate have been
shown to vary by RNA preservation method (45–47, 51),
the effect of preservative choice on metatranscriptome pro-
files has not been systematically investigated. We suggest an
optimal strategy of stabilization using a rapidly penetrat-
ing reagent such as RNAlater, followed by flash freezing as
soon as possible.
RNA Extraction
A wide variety of protocols have been used successfully for
total RNA extraction from microorganisms. These methods
vary but typically involve steps of cell lysis and protein diges-
tion, separation of nucleic acids from proteins, RNA recovery,
DNase digestion, and a final purification and concentration
procedure. Lysis and protein digestion may be done enzymati-
cally (i.e., lysozyme + proteinase K) or via a combination of
mechanical disruption in denaturing lysis reagent (e.g., phe-
nol and guanidine-thiocyanate; Tri Reagent, Molecular
Research Center, also sold as TRIzol, Life Technologies). Iso-
lation of RNA from proteins is achieved by phase separation
using acid phenol:chloroform ( pH 4.5–4.7), since DNA par-
titions into the organic phase at acidic pH, leaving the RNA
in the aqueous supernatant. Following phenol:chloroform
addition and centrifugation, RNA is recovered from the aque-
ous fraction by alcohol precipitation and rehydration or by
binding to glass fiber, silica, or ion exchange membranes in
centrifugation spin columns (e.g., the Tempus Spin RNA iso-
lation kit, Ambion).
Commercial kits are widely used for the lysis and RNA
separation steps. Many of these kits are hybrid methods that
combine the effectiveness of organic extraction with the
ease of spin column RNA recovery. Kits vary depending on
sample type (i.e., water versus soil) and on the targeted size-
fraction of the RNA pool. Most column methods discrimi-
nate against RNA molecules smaller than 200 bp, a size range
including tRNAs, 5S rRNA, and most small regulatory RNAs
(sRNAs). sRNAs in particular have been overlooked in
most metatranscriptome analyses but have been recently
identified as important signals in many regulatory pathways
(52–54) and shown to be abundant and diverse in natural
microbial communities (55). Popular kits employed for meta-
transcriptome studies include the mirVana isolation kit
(Ambion) and RNeasy kit (Qiagen). The mirVana kit retains
total RNA, including both sRNAs and longer mRNA and
rRNA molecules, whereas the RNeasy kit discriminates
against low molecular weight RNAs. For soils and sediments,
total RNA is often isolated via physical extraction using phe-
nol:chloroform followed by precipitation (48, 56). Commer-
cial kits for isolating total soil microbial RNA, such as the
Powersoil kit by MoBio, are also widely used (2, 57).
Extracted total RNA should be eluted and stored in
RNase-free water rather than Tris-EDTA buffer to avoid
inhibition of DNase or other downstream enzymes that
require Mg2+ or Mn2+. After incubation with DNase (e.g.,
TURBO DNase, Ambion) to remove contaminating DNA,
the RNA should be purified to remove DNase enzymes and
incubation buffer, and then quantified and assessed for qual-
ity. Post-DNase treatment, purification, and concentration
can be done using a glass fiber–based spin column, although
caution should be used in choosing a kit that retains the
desired size range of RNA fragments. Popular kits such as
the RiboPure-Bacteria Kit (Ambion) or the RNeasy MinElute
Cleanup Kit (Qiagen) may discriminate against RNA mole-
cules at the shortest end of the size spectrum (<70 bp),
whereas kits such as RNA Clean & Concentrator-5 Kit
(Zymo Research) or the SurePrep RNA Cleanup and Con-
centration Kit (Fisher Scientific), allow recovery across
all RNA size ranges. DNase-treated and purified RNA can
be quantitated spectrophotometrically or fluorometrically.
Spectrophotometric quantitation, for example, using Nano-
Drop (Thermo Scientific), is rapid but overestimates nucleic
acid concentrations. Fluorescence-based methods such as
Quant-iT RiboGreen RNA Assay (Invitrogen) are highly
sensitive but require longer preparation times. Fluorescence
quantitation on small, benchtop fluorometers such as the
Qubit 2.0 (Life Technologies) offers an alternative requiring
minimal processing time, although the effective concentra-
tion range is higher for this instrument (250 pg/µl for the
Qubit versus 1 pg/µl for the Quant-iT). When possible, quan-
titated RNA should be assessed for RNA quality (i.e., level of
degradation), for example, using gel electrophoresis, includ-
ing automated electrophoretic platforms such as the Agilent
2100 Bioanalyzer, Agilent TapeStation 2200, or Bio-Rad
Experion. These instruments provide a numerical measure
of RNA degradation, the RNA Integrity Number or equiva-
lent, and a direct visualization of the relative contributions
of different RNA fractions, including rRNA and smaller
RNAs. Such knowledge can be useful for evaluating the
need for downstream procedures, such as rRNA depletion.
Ribosomal RNA Depletion
The detection and comparison of rare transcripts may require
steps that reduce the total amount of rRNA prior to se-
quencing, thereby increasing the proportional abundance of
mRNA and regulatory small RNA transcript pools. An alter-
native to rRNA depletion is to selectively enrich for mRNA
during cDNA synthesis. Armour et al. (58) first employed this
approach using a set of hexameric “not-so-random” (NSR)
primers targeting human mRNA. The NuGEN Ovation
Prokaryotic RNA-Seq Kit uses a modification of this princi-
ple, with NSR primers designed based on relatedness to 50
bacterial and archaeal genomes from major phylogenetic
groups (59). This method requires a priori knowledge of
the target mRNA sequences and therefore may have limited
use for diverse environmental communities. Amplification of
cDNA using multiple displacement amplification (MDA)
has also been used to increase template concentrations in
metatranscriptome studies. Gilbert et al. (60) successfully
amplified total cDNA from a coastal marine community
using the Illustra GenomiPhi V2 DNA Amplification kit
(GE Healthcare), which enables isothermal MDA using ran-
dom hexamer priming and a high-fidelity DNA polymerase
(Phi29 DNA polymerase). This study also reported a reduc-
tion of rRNA to 0.08% of the total RNA, potentially because
of discrimination against rRNA due to secondary structure.
However, MDA may amplify unevenly across genomes (61,
62) and its use for quantitative analysis of gene expression is
not recommended.
The most commonly used rRNA depletion methods
involve rRNA removal by binding to sequence-specific
2.4.5-6 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
probes (subtractive hybridization), rRNA digestion by an
exonuclease, poly(A) capture (for eukaryotes), or some com-
bination. Subtractive hybridization can be done using com-
mercial kits or custom protocols. Several metatranscriptome
studies have used the MICROBExpress Bacterial mRNA
Enrichment kit (Ambion), which removes rRNA by hybrid-
ization to 16S and 23S rRNA-targeting oligonucleotides
bound to magnetic beads. Capture oligonucleotides in the
MICROBExpress kit are reported to be compatible with
a wide range of Gram-negative and Gram-positive bacteria,
although other microorganisms such as Archaea, Mollicutes,
and potentially many other environmental species are
not compatible (Life Technologies documentation). The
RiboMinus kit (Life Technologies) uses a modification of
the selective hybridization method, in which 16S and
23S rRNA probes contain locked nucleic acid monomers
(LNA—ribonucleoside sugar backbone modified to increase
Tm) that provide stability to the probe–rRNA duplex to
increase hybridization efficiencies. These probes are biotiny-
lated, enabling easy removal of the probe–rRNA complex by
streptavidin binding (63). Unlike the MICROBExpress and
RiboMinus kits, the Ribo-Zero rRNA-Removal kit (Bacte-
ria) (Epicenter), which also uses magnetic capture and a pro-
prietary probe set, targets 5S rRNA in addition to 16S and
23S rRNA of diverse bacterial clades.
rRNA removal efficiency varies depending on the taxo-
nomic composition of the sampled community and may be
low for communities whose members are distantly related to
the rRNA probe sequences (59). Subtraction efficiency
may be increased by using custom probes designed to match
the organisms in the sample (16). In this method, sample-
specific rRNA probes are prepared based on rRNA gene
sequences extracted from a coupled DNA sample from the
same community. rRNA genes are amplified by PCR using
universal primers, and the amplicons are used as template
for in vitro transcription with biotinylated nucleotides, yield-
ing biotin-labeled antisense rRNA. Labeled rRNA is hybri-
dized to complementary molecules in the total RNA pool,
over a step-down procedure of decreasing temperature. The
double-stranded probe-template hybrid is then removed by
binding to streptavidin-conjugated magnetic beads. This
method can be used with probes targeting small and large sub-
unit rRNA sequences, from both prokaryotes (5S, 16S, 23S
rRNA) and eukaryotes (5/5.8S, 18S, 28S rRNA).
rRNA by exonuclease digestion can be done inde-
pendently or in conjunction with subtractive hybridization,
although combining multiple methods may pose a risk to
RNA fidelity (64). The exonuclease-based mRNA-ONLY
Prokaryotic mRNA Isolation Kit (Epicentre; now discontin-
ued) has been used in several metatranscriptome studies, with
mixed results (6). This method uses a 50 → 30 exonuclease
that digests RNA with a 50 monophosphate, which is present
on processed bacterial rRNAs but absent from primary
mRNAs transcripts carrying a 50 triphosphate. This method
is sensitive to the integrity of the RNA pool, as a degraded
rRNA pool will contain 50 hydroxyl ends that escape diges-
tion by the exonuclease. An alternative nuclease-based
methods utilizes a duplex-specific nuclease to target double-
stranded DNA (65). This method is a modification of the
eukaryotic cDNA normalization protocol developed by Zhu-
lidov et al. (66) to enhance the detection of low-abundance
transcripts. The protocol, which requires that total RNA first
be converted to cDNA, relies on the principle that abundant
sequences (e.g., rRNA) form double strands more effectively
(67) and are therefore preferentially removed. Yi et al. (65)
demonstrated that duplex-specific nuclease treatment with
the Trimmer-Direct cDNA Normalization Kit (Evrogen)
removes rRNA sequences while retaining the original relative
abundances of mRNA transcripts from Escherichia coli cul-
tures. However, the efficiency of this method for species-rich
environmental metatranscriptomes is uncertain.
A small number of studies have evaluated the relative
effectiveness of different rRNA depletion methods (64, 68).
He et al. (64) compared two popular methods of rRNA deple-
tion, subtractive hybridization with the MICROBExpress
Bacterial mRNA Enrichment kit (Ambion) and exonuclease
digestion with the mRNA-ONLY Prokaryotic mRNA Isola-
tion kit (Epicenter), using RNA from synthetic five-member
microbial communities. Removal efficiencies were shown to
depend on community composition and RNA integrity,
and subtractive hybridization resulted in greater rRNA reduc-
tions compared to exonuclease digestion, although rRNA
percentages following depletion were still >77%. This study
suggested that depletion levels would be maximized by a sub-
tractive hybridization strategy using custom probes (e.g., as
in 16).
Achieving consistent depletion of rRNA from diverse
environmental metatranscriptomes has proved challenging.
A recent survey of published marine metatranscriptome stud-
ies revealed an average of >50% rRNA following the applica-
tion of commercial or custom rRNA depletion methods (6).
Depletion efficiencies varied considerably, both among
methods and among studies using the same method, ranging
from no depletion (69; I. Hewson, personal communication)
to nearly total depletion (68). Of the commercial methods
available, the Ribo-Zero kit (Epicenter) appears especially
promising, having been shown to reduce rRNA to less than
5% of total RNA in a species-rich gut microbial metatran-
scriptome. This method has been incorporated into the meta-
transcriptome analysis pipeline used by the DOE Community
Science Program. However, this kit does not target Archaea
and has not yet been extensively vetted using metatranscrip-
tomes from diverse natural habitats. For environmental com-
munities with representatives from all three domains,
significant rRNA drawdown will likely require subtractive
hybridization with sample-specific probes (16).
Amplification of mRNA/cDNA for
Metatranscriptomes
RNA yields from environmental samples may be too low to
generate sufficient cDNA for sequencing, therefore requiring
amplification of total RNA, or cDNA. RNA amplification
commonly involves in vitro transcription (IVT). The Messa-
geAmp II Bacteria RNA Amplification kit (Ambion) has
been used extensively for IVT-based amplification prior to
metatranscriptome sequencing (6). This method involves
an initial step during which all transcripts are polyadenylated
using poly(A) polymerase. Poly(A) RNA is then converted
to cDNA by reverse transcription, with priming by oligo
(dT) primers appended with the promoter sequence for T7
RNA polymerase. IVT with the T7 polymerase results in a
linear amplification of RNA, yielding microgram quantities
from nanograms of starting material over the course of the
incubation (typically 6–16 h). Amplified RNA is then used
for a second cDNA synthesis step. The MessageAmp II kit
amplifies total RNA from as little as 100 ng input RNA (rec-
ommended), although successful amplifications have been
performed in our lab with less than 10 ng. However, RNA
degradation may occur during the IVT reaction (70)—we
have observed a decline in RNA integrity when using the
MessageAmp kit for incubations longer than 12 h.
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-7
Furthermore, initiation of first strand cDNA synthesis by
oligo(dT) priming also may result in an overrepresentation
of fragments mapping to the 30 end of genes (71). The effect
of this bias on measurements of differential transcript abun-
dance is not well characterized, but it may be minimal if the
bias is uniform across all transcripts.
Amplification also can be achieved through the activ-
ity of DNA polymerase during the cDNA synthesis step.
For example, the WT-Ovation series of kits (Nugen Technol-
ogies) utilizes a unique priming strategy involving 30 random
hexameric DNA fragments ligated in a chimera to a 50 RNA
fragment that serves as a primer site for linear amplifica-
tion via an RNA-dependent DNA polymerase (72). This
method can generate micrograms of cDNA from as lit-
tle as 500 pg to 50 ng of total RNA in a matter of hours.
Amplification of cDNA using MDA has also been used to
increase template concentrations in metatranscriptome
studies (60).
Several studies have compared amplification efficiencies
among commercial kits. Using human RNA, Clement-Ziza
et al. (73) compared the WT-Ovation cDNA amplification
method to three kits utilizing T7-based IVT amplification.
The WT-Ovation Pico system was shown to produce consis-
tent, reproducible results following a single round of amplifi-
cation, although expression profiles (measured by microarray
hybridization) varied among amplification protocols (cor-
relations of 0.67–0.8 in comparisons between IVT and WT-
Ovation methods). Amplification methods have also been
compared using RNA from environmental communities. Wu
et al. (74) used deep-sea bacterioplankton RNA (input: 4–
20 ng) to compare two MDA-based methods (the Illustra
GenomiPhi V2 DNA Amplification and QuantiTect Whole-
Transcriptome [Qiagen] kits) and the NuGEN WT-Ovation
Pico RNA Amplification System. The Illustra GenomePhi
kit failed to amplify total RNA from deep-sea samples with
low microbial abundance, whereas the NuGEN WT-Ovation
provided the highest overall amplification yields and a rela-
tively higher technical reproducibility. This study did not com-
pare the reproducibility of gene expression profiles among
methods. Indeed, while prior studies indicate that IVT-based
methods, including polyadenylation-dependent methods
such as the MessageAmp kit, introduce minimal bias into
array-based expression profiles (typically based on RNA
from a single taxon; 75–77), a systematic comparison of
amplification bias across diverse methods has not been
done using diverse environmental metatransciptomes. Test-
ing for differential gene expression between samples should
therefore only be done using samples prepared with the
same protocol.
Improvements in sequencing library preparation and
reductions in sequencing input requirements may soon
make RNA amplification unnecessary for most metatran-
scriptome studies. Notably, the ScriptSeq v2 RNA-Seq
Library Preparation kit (Illumina) can be used to prepare
cDNA tagged with Illumina adaptors starting from as little
as 500 pg of total RNA or rRNA-depleted RNA (rRNA
depletion is recommended). This method initiates cDNA
synthesis by priming with random hexamers appended with
a unique 50 tag. A second unique tag is then annealed to
the 30 end of the cDNA, yielding ditagged cDNA that is
then amplified by PCR (10–15 cycles). This method gener-
ates sequencing-ready cDNA in as little as 4 h, while also pre-
serving strand directionality. However, the ScriptSeq kit
generates higher levels of variation in sequence coverage
compared to other cDNA library preparation methods for
Illumina sequencing (e.g., TruSeq RNA-Seq) and is not
recommended for the sequencing of small RNAs. Nonethe-
less, this method has been used effectively, including in our
laboratory, to generate high-quality cDNA data sets, and
offers a considerable reduction in time and personnel costs
relative to methods involving a separate RNA amplification
step.
cDNA Synthesis from RNA
Total RNA must be converted to cDNA prior to sequenc-
ing. Conversion to cDNA also helps ensure long-term stabil-
ity of the samples, as DNA is less susceptible to degradation
compared to RNA. Many commercial kits are available for
cDNA synthesis, including some of the previously described
kits for RNA amplification (e.g., WT-Ovation Pico) or kits
that combine cDNA synthesis with sequencing-adaptor
incorporation (ScriptSeq). cDNA synthesis methods are typ-
ically differentiated based on priming strategy, either via ran-
dom hexamers or poly(T) oligonucleotides. Use of the latter
first requires the addition of 30 poly(A) stretches to microb-
ial total RNA, which is typically not polyadenylated. Sev-
eral hexamer-based kits, including the Superscript cDNA
synthesis system (Invitrogen) and the Quantitect Reverse
Transcription kit (Qiagen), have been used successfully for
metatranscriptome samples. These kits use random hexamer
primers for cDNA synthesis from RNA template, by reverse
transcription, followed by second strand synthesis. The
RNA template may or may not have the rRNA fraction sub-
tracted. A modification of this technique to avoid rRNA con-
tamination in the cDNA pool is the use of computationally
selected NSR primers, as described previously (58). All pri-
ming strategies for cDNA synthesis, including the use of ran-
dom hexamers and oligo(dT) primers, have the potential
to bias sequence coverage via selective priming (78, 79).
The effect of such bias on comparative analyses of differential
transcript abundance in metatranscriptomes has not been
explored.
Environmental metatranscriptome studies typically do not
present information regarding the strand orientation of tran-
scripts. However, knowing the DNA strand from which the
RNA molecule originates can be important in resolving
annotation ambiguities for known and novel genes, provide
hints to the function of the studied RNA, identify exact
boundaries of genes on opposite strands, and help correctly
predict the expression levels of a given transcript (80, 81).
It is also important for identifying antisense RNAs, which
are common in bacterial transcriptomes, but whose func-
tional significance is not well understood (82). Certain
strand-specific cDNA synthesis methods developed for
microarray studies can also be applied for metatranscriptom-
ics. One of the simplest and most effective methods of pre-
serving strand orientation during cDNA synthesis is the
incorporation of deoxy-UTP, rather than dTTP, during first-
or second-strand cDNA synthesis. The uridine-containing
strand is then selectively degraded by digestion with
uracil-N-glycosylase during preparation of sequencing libra-
ries (83). This method has been applied to determine strand
specificity in transcriptomes of cultured bacteria (68). For
Illumina-sequencing applications, the library preparation
protocol can be altered to enable cDNA synthesis directly
on the sequencing flowcell surface. This approach, called
flowcell reverse transcription, also allows determination of
strandedness (84). As discussed already, the ScriptSeq v2
RNA-Seq Library Preparation kit (Epicentre) incorporates
30 and 50 tagging strategies for cDNA to allow detection of
sense and antisense RNA sequences. However, because
interpreting patterns of strand orientation requires knowledge
2.4.5-8 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

FIGURE 1 A typical workflow for metatranscriptome sequence analysis. Raw reads are the input for the workflow. The analysis can be
divided into five subsections (enclosed within dotted boxes): pre-processing, assembly (optional), annotation, statistical analysis, and analysis
(See legend on following page)
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-9

of gene orientation, determining strandedness in metatran-
scriptome studies may be most useful when reference genomes
are available for transcript read mapping.
Sequencing
High-throughput sequencing technologies have revolution-
ized the study of microbial gene expression in the environ-
ment. Most researchers opt to outsource their sequencing
library preparation and cDNA sequencing to an institutional
or commercial core facility. Such facilities typically offer a
variety of sequencing, data handling, and data analysis
options. Competing platforms utilize diverse molecular meth-
ods for base detection, including fluorescence of reversible
terminator bases (Illumina technology), light emission on
base incorporation ( pyrosequencing) (Roche 454), ligation
of fixed length oligonucleotides (SOLiD), and hydrogen
ion release during DNA polymerization (Ion Torrent).
All of these methods exploit engineering advancements
(e.g., nanofluidic separation) that allow the sequencing proc-
ess to be run in parallel for millions to billions of individual
DNA fragments, but differ in terms of sequence yield, cost
per base, read length, and sequencing error rate. See the
Chapter 20 for a comparison of Ion Torrent, Roche 454,
and Illumina technologies for microbial-oriented sequencing
applications. Most published environmental metatranscrip-
tome studies have utilized the Roche 454 pyrosequencing
platform (6), due to the early availability of this technology
and the relatively long read lengths enabled by pyrosequenc-
ing compared to competing platforms (e.g., mode 454 read
lengths of 450–700 bp, compared to 50–300 bp for Illumina).
Longer read lengths provide an advantage for accurate gene
prediction, phylogenetic analysis, and assembly. How-
ever, compared to the 454, Illumina technology offers a sub-
stantially lower cost per base (<$0.10 per Mb compared to
∼$10 per Mb, when using the Illumina HiSeq2000 and
Roche GS FLX+) and a significantly lower error rate (13),
notably in homopolymer regions (85, 86). Furthermore, the
Illumina MiSeq instrument now supports 300 × 300 base
paired-end sequencing, which enables merged read lengths
within a range comparable to that of the Roche 454 platform.
While Illumina has replaced the Roche 454 platform as
the standard for microbial metatranscriptomics, other plat-
forms such as the Ion Torrent Proton also offer cost-effective
sequencing but are not yet widely available and have not been
used extensively for environmental metatranscriptomics.
Most sequencing manufacturers now also offer benchtop
versions (e.g., the Illumina MiSeq) of their hallmark
sequencing instruments, enabling generation of meta-omic
data sets by individual investigators in a matter of days.
Indeed, the use of Illumina-specific kits such as the ScriptSeq
v2 RNA-Seq Library Preparation kit (discussed above) can
generate sequencing-ready cDNA from total RNA in as little
as 4 h (skipping rRNA depletion and not including time for
assessing final cDNA yield, e.g., by Bioanalyzer). Labs with
direct access to benchtop instruments (no queue time) can
now go from raw sample to metatranscriptome sequence
data in 3–4 days. However, maintaining and operating a
benchtop sequencer requires significant recurring costs and
labor, and high-end sequencers may be more cost-effective
for large output and heavily replicated sample sets. Labs
should take this into consideration when deciding between
outsourcing and purchasing their own instrument. Ulti-
mately, choice of sequencing platform will depend on cover-
age requirements (sequencing depth), balanced against
considerations of cost, sequence length and quality, platform
availability, and compatibility with downstream informatics
steps.
DATA ANALYSIS
Thoroughly mining the rich information content of
metatranscriptome data sets requires multiple analytical
approaches, which will likely vary depending on the experi-
mental goals of the study. The majority of metatranscriptome
studies have two major analytical goals: identification and
annotation of functional gene (mRNA) or small RNA tran-
scripts, followed by a statistical analysis to detect transcripts
that are differentially expressed (DE) between treatments or
samples. Meeting these goals requires a series of data process-
ing steps, ranging from quality assurance filtering to homol-
ogy-based database searches. Fig. 1 and the discussion that
follows provide an overview of the major analytical steps of
metatranscriptome data analysis, along with recommenda-
tions for informatics resources to support the analysis.
Preprocessing: Sequence Formats, Quality Assurance,
Duplicate and rRNA Removal
Assuring the quality of sequence data is a critical first step in
metatranscriptome analysis. Different sequencing technolo-
gies generate different sequence output file formats, including
FASTQ (Illumina, Ion Torrent), SFF (Roche 454, Ion Tor-
rent), CFASTQ (AB SOLiD), and FNA/FASTA (Roche
454), along with a suite of additional files containing run
log/statistics, images, base quality scores, and sequence data.
The following description pertains to data generated using
Illumina technology, although many of the same steps also
are required for processing data from other platforms. To max-
imize read length, and therefore the accuracy of gene predic-
tion and assembly, environmental transcriptomics analyses
using Illumina should aim to obtain reads of the greatest
length offered by the technology (currently 150 × 150 bp
for HiSeq, 300 × 300 bp for MiSeq).
An Illumina sequencing run will generate millions
(MiSeq) to billions (HiSeq) of individual sequence reads,
typically binned according to sample identifier (barcode) if
multiple samples are combined (multiplexed) on a single
run. Demultiplexing of raw data can also be done using per-
sonal scripts or public resources such as the FASTX-Toolkit
(http://hannonlab.cshl.edu/fastx_toolkit/) or Trimmomatic
(87). The FASTQ data format used by Illumina integrates

FIGURE 1 (Continued.) of biological patterns. Some of the most commonly used tools and a rough estimate of the time required per step are
listed. The time estimates will vary considerably depending on the computational resources available and the size of the input data set. The
estimates shown here are based on a 64-bit Linux machine with 12 CPUs and 50 Gb RAM, and a data set of ∼1 million 300 × 300 paired-end
Illumina reads (MiSeq). It is important to stress that the final step, the conversion of transcriptome patterns into meaningful inferences about
microbial physiology, is the most labor-intensive. This process may require months of manual exploration and validation of the sequence data,
often using task-specific scripts beyond the scope of the standard analysis pipeline. Multiple steps in this workflow can be achieved using a single
resource (see Table 3 and text), although such platforms offer limited flexibility for adding custom modules to the pipeline.
doi:10.1128/9781555818821.ch2.4.5.f1
2.4.5-10 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

both the sequence and Sanger or Phred (88) quality scores
(QPHRED) in a single file. QPHRED scores (Q-scores) estimate,
for each base, the probability of the base call being incorrect.
For example, a Q-score of 30 (Q30) represents an error rate of
1 in 1,000, that is, a base call accuracy of 99.9%. Thus, a
higher quality score indicates a smaller probability of error.
Although most analysis tools accept multiple sequence and
Q-score formats, some require Q-score format as an input
parameter, making it important to know the data format gen-
erated by the current version of the sequencing technology.
Removing sequences with potential base-call errors or
library preparation artifacts is critical for preventing erroneous
conclusions during downstream analysis. Table 1 lists some of
the tools available for assessing sequence quality and filtering
out low-quality reads. Outputs from these tools may include
plots of per base quality, sequence length distributions, counts
or estimates of sequence duplicates, and lists of overrepre-
sented or possible contaminant sequences. Q-scores typically
drop at the ends of reads, requiring that reads be trimmed to
remove low-quality regions. Although a standard threshold
for base trimming does not exist, a Phred33 score cutoff
between Q20 (lenient) and Q30 (stringent) is considered rea-
sonable (the Q-score is logarithmically related to the base
calling error probability; scores of Q20 and Q30 therefore
indicate that the probability of the base-call being incorrect
is 1/100 and 1/1,000, respectively; 88). Library preparation
artifacts may include remnants of PCR primers, poly(A) tails
introduced during cDNA synthesis, or sequencing adapters,
all of which should be removed before further analysis.
Preprocessing of paired-end data sets typically involves
an additional step in which forward and reverse reads (read
partners) are merged to generate single contiguous sequences.
Generation of longer merged reads increases the informa-
tion content per transcript, which increases the accuracy of
gene assignment (18). Tools for paired-end merging are listed
in Table 1 and allow for user-defined adjustment of allow-
able mismatches and overlap criteria. Depending on read
length, an overlap of 10–20% with zero to two mismatches
is recommended.
Preprocessing often also involves the identification and
removal of duplicate reads, which potentially reflect errors
introduced during library preparation or sequencing. The def-
inition of duplicate varies among studies, however, and the
field has not converged on a standard practice for duplicate
removal. Working with metagenomic data, Gomez-Alvarez
et al. (100) proposed that sequences sharing the same start
site (within 3 bp) and greater than 90% nucleotide iden-
tity likely represent methodological artifacts, given the
exceedingly low probability of sampling near-identical
DNA fragments from a multispecies genome pool. However,

TABLE 1 Popular sequence processing tools

Local or GUI or command
Software Utility Citation
web line (CL)
FASTQC Local Both QA Websitea
HTQC Local CL QA, preprocessing 89
PRINSEQ Both CL QA, preprocessing 90
FASTX-Toolkit Local CL Multiple, including QA, preprocessing, Websiteb
duplicate removal
Trimmomatic Local CL Preprocessing 87
SHERA Local CL Read merging 153
FLASH Local CL Read merging 154
PANDASeq Local CL Read merging 155
CD-HIT Local CL Duplicate removal 91
Ribopicker Both CL rRNA removal 92
rRNASelector Local GUI rRNA removal 93
SortMeRNA Local CL rRNA removal 94
Cufflinks Local CL ab initio assembly 95
Velvet Local CL ab initio assembly 156
Scripture Local CL ab initio assembly 157
MetaVelvet Local CL de novo metagenome assembly 96
IDBA-UD Local CL de novo metagenome assembly 158
Genovo Local CL de novo metagenome assembly 159
RayMeta Local CL de novo metagenome assembly 160
SOAPdenovo-Trans Local CL de novo metatranscriptome assembly Websitec
Trans-Abyss Local CL de novo metatranscriptome assembly 97
Rnnotator Local CL de novo metatranscriptome assembly 161
QIIME Local CL rRNA-based taxonomic classification and 98
diversity analysis
Mothur Local CL rRNA-based taxonomic classification and 99
diversity analysis
a
http://www.bioinformatics.babraham.ac.uk/projects/fastqc/.
b
http://hannonlab.cshl.edu/fastx_toolkit/.
c
http://soap.genomics.org.cn/SOAPdenovo-Trans.html.
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-11
this definition may be overly rigid for sequences generated
from discrete transcripts that occur in multiple copies per
cell. Based on comparisons of technical replicates analyzed
on the Roche 454 platform, Stewart et al. (16) defined dupli-
cates as sequences differing in length by no more than 1 bp
and sharing greater than 99% identity, with a requirement
that the first 3 bp of the sequences be identical. In the absence
of additional knowledge about the frequency of true (nonar-
tifact) duplicates in a sample, we recommend (at the least) the
removal of sequences sharing 100% identity and length. Pro-
grams such as CD-HIT (91) and FASTX-Toolkit (http://han
nonlab.cshl.edu/fastx_toolkit/) are commonly used for dupli-
cate identification and removal (101, 102).
Preprocessing may also involve detection and removal of
rRNA transcripts. rRNA will be present in metatranscrip-
tome data sets, even in studies employing an rRNA depletion
step (see above) prior to sequencing. Residual rRNA frag-
ments in data sets generated after rRNA depletion should
not be used for taxonomic analysis, as this transcript pool is
likely highly biased by selective rRNA removal (i.e., by non-
uniform binding to rRNA probes used for subtractive hybrid-
ization). Identification and removal of these sequences will
reduce total data set size and therefore increase the computa-
tional efficiency of downstream analysis steps (e.g., assembly,
annotation). For metatranscriptomes generated without
rRNA depletion, rRNA fragments may provide valuable
information about the taxonomic composition of the tran-
scriptionally active gene pool (17). Bioinformatics tools
such as RiboPicker (92), SortMeRNA (94), rRNASelector
(93), and Parallel-Meta (103) can be used to identify, extract,
and taxonomically characterize rRNA fragments from total
RNA data sets. Similarity-based searches (e.g., BLASTN)
against custom or existing rRNA databases such as SILVA
(http://www.arb-silva.de), Greengenes (104), or the Riboso-
mal Database Project (RDP) database (105) also can be
used for rRNA identification. Some rRNA databases, such
as RDP, are integrated into tools for rRNA classification, vis-
ualization, and statistical comparisons of taxonomic composi-
tion among samples. Alternatively, the taxonomic structure
of the rRNA data pool can be visualized in programs such
as MEGAN (106), which maps identified reads onto a micro-
bial phylogeny based on the taxonomic annotations of data-
base sequences identified by BLAST. rRNA sequence
fragments also can be used as input for classification by soft-
ware packages such as Mothur (99) or QIIME (98). These
tools were designed primarily for comparison of 16S rRNA
data generated by amplicon sequencing but also support anal-
ysis of rRNA fragments recovered from meta-omic data sets
( please see Table 1).
Assembly
Assembly is the recursive process of linking overlapping short
reads (or merged pairs of reads) to generate longer contiguous
sequences (contigs). While assembly is a critical step of most
genomic and metagenomic analyses, the majority of environ-
mental metatranscriptome studies have been based on unas-
sembled mRNA sequence fragments, typically 100–500 bp
in length (6). These sequences are queried against nucleotide
or protein databases to draw inferences about functional gene
and taxonomic identity. The obvious advantage of assembly
is that it generates longer sequences with greater information
content, which can substantially improve the accuracy of
homology-based gene assignment (18). Transcript assemblies
may also enable the detection of cotranscribed genes on poly-
cistronic mRNA, a common feature of prokaryotic transcrip-
tomes (107).
The trade-off is that assembly adds a computationally
intensive step to the analysis pipeline and can introduce inac-
curacies into reconstructed sequences, including small
nucleotide changes or the misjoining of transcripts from unre-
lated genes (chimera formation). Assemblies of meta-omic
data may be especially prone to generating chimeric contigs,
as these analyses inherently involve data from mixed-species
communities. The potential for misassembly depends on a
range of factors, including sequencing read length and error
rate, data set size (number of sequences), the complexity
and composition of the transcriptome (e.g., number of taxa
and genes represented, their relative abundances, and their
phylogenetic relatedness), and the choice and parameteriza-
tion of the assembly software (108).
Several assembly programs for next generation sequencing
data have been designed to take into account the unique
properties of transcriptomes (see review by Martin and
Wang [109] and Table 1). Compared to genomic sequence
data, microbial transcriptome-derived reads encompass few
repeat regions and typically exhibit nonuniform coverage
across transcripts (due to differential abundance of organisms
in a community and differential expression of genes), as well
as within the same transcript (due to priming biases dur-
ing cDNA synthesis or RNA amplification or to directional
biases during transcript degradation). Although isoforms due
to alternative splicing are rare in prokaryotic transcriptomes,
transcript variants can be generated by alternate promoters
and terminators (110). Transcriptome-specific assembly tools
are designed to help resolve different gene isoforms. Programs
are available for both de novo assembly (e.g., Oases [111],
Trinity [112]) as well as assembly based on alignment to a
reference sequence (e.g., Cufflinks [95]). For many environ-
ments, reference genomes from transcriptionally active taxa
are not available. Alternatively, metagenomic contigs,
assembled from sequence reads from a co-collected DNA sam-
ple, can be used to guide metatranscriptome assembly. Popu-
lar programs optimized for metagenomic assembly from short
reads (e.g., Meta-Velvet [96]) may also be used effectively
for metatranscriptome assemblies, as both metagenomes and
transcriptome data sets share similar properties (i.e., non-
uniform coverage). An optimal metatranscriptome assembly
strategy to maximize both accuracy and contiguity (contig
length) will likely require de novo assembly, followed by refer-
ence-based mapping to a coupled metagenome. Sequence
reads can then be mapped back onto the assembled contigs
(or contigs from reference assemblies) using short read align-
ment tools like Bowtie 2 (113), Burrows-Wheeler alignment
(114), SHRiMP (115), or Stampy (116), whereas programs
such as Trans-Abyss (97) can carry out both assembly and
read mapping.
Assembly is expected to become an increasingly impor-
tant tool in metatranscriptome analysis as transcriptome-
specific assemblers continue to improve and the availability
of reference sequences increases. Transcript abundance pat-
terns inferred from nonassembled short reads have been vali-
dated in several metatranscriptome studies using quantitative
PCR (1, 55, 117, 118). Although increasing sequence frag-
ment length is known to improve the likelihood of homolog
detection (18), a direct comparison of transcript abundances
between assembled and nonassembled metatranscriptome
data has not been done (to the best of our knowledge).
It therefore remains unclear to what extent metatranscrip-
tome assembly improves inferences of community function.
Assembly may be most important for metatranscriptomes
generated with short reads (<100 bp), for example, on the
Illumina HiSeq platform.
2.4.5-12 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES

Functional and Taxonomic Annotation
of Transcripts
A primary objective of metatranscriptome analysis is the
annotation of recovered transcripts, from either assembled
or nonassembled data. As with gene annotation in microbial
genome analysis, transcript annotation primarily involves
comparisons, using BLAST or alternative search algorithms,
of query sequences (transcripts) against nucleotide or amino
acid databases to detect related sequences. Query sequences
may be either nonassembled short reads, assembled contigs,
or open reading frames identified on contigs using gene pre-
diction tools (e.g., GeneMarkS [119], Glimmer [120], Prodi-
gal [121]). The functional and taxonomic identities of
database sequences that match the query can then be used
to characterize (annotate) the query sequence.
A wide variety of nucleotide or protein databases are
available for similarity searches and sequence annota-
tion (Table 2). These differ in content (taxonomic and func-
tional gene representation), size (number of reference entries),
level of sequence redundancy, extent and details of annota-
tion, and state of curation and maintenance. It is beyond the
scope of this chapter to contrast these databases, and it is rec-
ommended that users familiarize themselves with the proper-
ties and limitations of these resources before undertaking
an analysis (128). Indeed, use of outdated, poorly annotated,
or incomplete databases can mask or obscure the detection
of biological signals from sequence data (e.g., see 102).
NCBI’s nonredundant (nr) database offers perhaps the
most comprehensive collection of amino acid sequence data
for sequence annotation. A BLASTX search against the nr
database identifies reference sequences that align with part
or all of the query sequence, in addition to associated expect
(E)-values and bit scores that are used to assess the reliability
of the alignment. The E-value is a measure of statistical signif-
icance that estimates the probability of a match occurring by
random taking into account the size of the reference database.
The smaller the E-value, the greater the significance. In con-
trast to E-values, which are challenging to compare across
studies that use different databases, bit scores are independent
of database size. The bit score is proportional to the quality
of alignment and is often used as a threshold for gene assign-
ment—a significance threshold of bit score 50 has been used
routinely in metatranscriptome analyses with data from
diverse sequencing platforms (e.g., Shi et al. [55], Stewart
et al. [102], Xiong et al. [129]). Significance thresholds can
be imposed either at the start of a BLAST search or while
parsing BLAST result files.
Databases other than NCBI-nr may offer advantages in
terms of sequence content, smaller size, or data organiza-
tion. For example, the iMicrobe Project (http://data.imic
robe.us) houses a wealth of metagenomic or transcriptomic
data from environmental microbial communities, much
of which was formerly deposited in the Cyberinfrastructure
for Advanced Microbial Ecology Research and Analysis
(CAMERA; 27) database (now discontinued). Smaller
custom databases, for example, containing a subset of cura-
ted whole genome sequences (e.g., David et al. [130]) or
a subdatabases of nr (125), may have sufficient information
content for metatranscriptome annotation depending on
the biological question and the genetic diversity of the
sample and may enable a significant reduction in data pro-
cessing time (BLAST searches) compared to larger data-
bases. Other databases, such as KEGG (122, 126), COG
(127), SEED (124), or eggNOG (131) (Table 2), are struc-
tured into hierarchical groupings that contain orthologous
sequences with related biochemical function or sequences
linked into metabolic pathways. These databases, in contrast
to the nr database for example, offer a convenient frame-
work for clustering metatranscriptome data according to
functional category, and have been integrated into several
popular platforms for meta-omic data analysis (Table 3).

TABLE 2 A subset of nucleotide or protein databases for sequence annotation

Database Description Citation
NCBI-nr Primary repository of nonredundant protein sequences; comprehensive, with functional and taxonomic 162, 163
information linked to sequence data from multiple sources
UniProt Annotated protein database; composed of two sections: (a) Swiss-Prot, which is manually annotated, and 164
(b) TrEMBL, which is automatically annotated
COG Manually curated database organized into functional categories, i.e., clusters of orthologous groups 122
Pfam Protein families defined by multiple sequence alignment and hidden Markov models (HMMs); composed 165
of Pfam-A, containing high-quality curated families, and Pfam-B, containing automatically generated
and uncurated protein families of lower quality
Rfam Database of noncoding RNA, including microbial small rRNAs; can be queried for homology based on a 123
combination of primary sequence and secondary structure
TIGRFAMS Complementary to Pfam; curated protein families represented by HMMs, multiple sequence alignments 166
and extensive annotation information, including gene ontology assignments, information on related
models in Pfam, and literature references
eggNOG Evolutionary genealogy of genes: nonsupervised orthologous groups database; similar to COG; generates 124
orthologous groups, predicts protein domains, and assigns functional descriptions using automated
annotation methods
KEGG Kyoto encyclopedia of genes and genomes; protein sequence data organized into a hierarchy of orthologs 125, 126
and functional pathways; requires paid subscription
iMicrobe Repository for microbial meta-omic data, including data from the discontinued CAMERA database (27) Websitea
SEED Protein database based on curated genomic data; data organized into subsystems 127
(functional categories)
a
http://imicrobe.us/
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-13

TABLE 3 Public resources for meta-omic data management and analysis

Resource Description Citation
MG-RAST Web-based; repository for microbial metagenomic and metatranscriptomic data; provides tools for read preprocessing, 26
rRNA prediction, gene prediction, BLAT searching (against M5NR, a nonredundant integration of diverse
databases) and classification into functional categories, data set comparison, and visualization; requires user
registration
IMG/M Web-based; provides tools for annotation, comparative analysis, and visualization of data; includes draft and microbial 28
genomes sequenced by the Joint Genome Institute; requires user registration
GALAXY Web-based, open-source platform for data analysis; includes modules for read preprocessing, taxonomic assignment of 25
meta-omic sequences based on BLAST results; allows construction of reusable multistep analysis workflows; requires
user registration
MEGAN Locally installed; uses BLAST results as input; enables taxonomic and functional classification, interactive analysis, 106
and visualization of samples
HUMAnN Locally installed; uses tabular translated BLAST (blastx) results (e.g., against KEGG or MetaCyc) as input; pipeline for 132
identifying and estimating abundance of metabolic pathways; enables visualization and comparison of relative
abundance of pathways

Annotation by BLAST-based similarity searching is one
of the most computationally intensive steps in metatranscrip-
tome analysis, particularly when searching against large data-
bases like nr. Similarity searches involving hundreds of
thousands to millions of transcripts typically require access
to cluster computing, either locally or via servers accessed
through public web-based platforms such as MG-RAST (26)
(Table 3). Computational and time requirements for searches
can be reduced by using smaller reference databases or faster
search algorithms including the BLAST-Like Alignment
Tool (BLAT; 133), RAPSearch (134), or LAST (135). For
some aligners, notably BLAT, increases in speed may come
at a cost of diminished sensitivity for detecting matches
with lower sequence similarity (133).
Analysis of Annotated Transcripts
Output files from similarity searches are parsed to generate
counts of sequences per unit of biological information (hit
counts). These units range from individual reference genes
(accession numbers) recovered as top matches during
searches, clusters of orthologous sequences (genes grouped
by shared evolutionary ancestry and encoding functionally
equivalent proteins), functional categories (the metabolic
pathway or biochemical function associated with a gene prod-
uct), or taxonomic groups (determined by the taxonomic
annotation of top matching reference sequences). Generating
hit counts can be done manually using custom scripts. Alter-
natively, search output can be imported to programs such as
MEGAN (106) or HUMAnN (132) to classify, tabulate,
and display hit count data according to function or taxo-
nomic affiliation. MEGAN, for example, extracts the annota-
tion information associated with top BLAST matches,
grouping sequences according to NCBI taxon identifications
and the metabolic categories in the SEED (124) and KEGG
(122,126) hierarchies (Tables 2 and 3).
Analysis and comparison of hit count data requires nor-
malization to adjust for variation both between and within
samples, notably variation in sequencing depth and gene
length. Count normalization can be done for individual hit
count files but is also included as a key step in most sequence
analysis platforms and programs for detecting differential
gene expression (Tables 2 and 3). Normalization is often
accomplished by total sum scaling, which involves dividing
counts of genes (or taxa) by the total number of sequences
in a sample, thereby converting raw counts to ratios. Counts
may also be normalized to account for differences in gene
length (long transcripts contribute more reads to the
sequenced pool compared to short transcripts). This assumes
that gene lengths are known, for example, based on refer-
ence genes or predicted open reading frames from coupled
metagenomic data sets. However, total sum scaling, though
widely used, can introduce bias as high-frequency genes/
transcripts exert a disproportionate influence over counts
(34, 35). Alternative methods that account for undersam-
pling of low-frequency transcripts (e.g., cumulative-sum
scaling; 136) offer an attractive alternative but have not yet
been widely used in metatranscriptomics. Detailed discus-
sion of normalization methods is beyond the scope of this
section; these techniques are covered in depth in recent
reviews (34–36).
An important end goal of metatranscriptome analysis is
the detection of genes (or gene categories) that are dif-
ferentially expressed (DE) among samples. Transcript hit
counts are proxies of gene expression levels and serve as
input data for DE detection programs. Commonly used DE
programs include baySeq (137), DESeq (138), edgeR (139),
and NOIseq (140). These tools differ primarily in their
assumption regarding the probability distribution of tran-
script count data. DE methods may also vary depending
on whether they accept complex experimental designs and
replicate structures, conduct pairwise versus nonpairwise
sample comparisons, adjust for data overdispersion, or offer
diverse methods for multiple test correction. A thorough
comparison of DE methods is beyond the scope of this chap-
ter. However, we have observed robust results employing
the software program baySeq (137), which uses an empirical
Bayesian approach that incorporates information from the
sampled data structure to improve model predictions (e.g.,
regarding data dispersion) and allows testing of multi-
ple, complex sample groupings (models). This method is
easily implemented in the R computing environment
(141) and enables parallel processing of thousands of genes.
The development and optimization of DE detection meth-
ods remains an active field of research, and we direct users
to recent excellent reviews for a comprehensive evaluation
of the strengths and pitfalls of different methods (33, 34,
142). As with statistical analyses in general, the power for
2.4.5-14 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
detecting DE increases with sample size and the number of
replicates. Because DE detection entails testing multiple
hypotheses simultaneously ( potentially on the order of hun-
dreds of thousands of genes), special consideration needs to
be given to address any false positives within the putative
DE genes. Most DE methods therefore provide options for
multiple test corrections, including estimates of the false dis-
covery rate (expected proportion of false positives among
the obtained significant results), when assessing the signifi-
cance of DE between samples (143). Many environmental
metatranscriptomic analyses unfortunately have been hin-
dered by relatively small sample sizes and an overall paucity
of replication (both biological and technical replication).
This trend is slowly reversing as sequencing costs continue
to decline and metatranscriptomic methods become widely
available. For some studies, however, it may be preferable
to increase biological or technical replication at the expense
of sequencing depth. Alternatively, low-replicate data set
analyses may be a useful strategy for initial hypothesis gener-
ation, followed by reverse transcription qPCR analyses that
target genes of interest in highly replicated sampling
frameworks.
Small Noncoding RNA
A significant fraction of non-rRNA sequences in metatran-
scriptome data sets may not show homology to known pro-
teins or are homologous to uncharacterized or hypothetical
open reading frames. These sequences may represent novel
protein-coding genes or uncharacterized variants or isoforms
of known proteins, but also include a variety of microbial
sRNA molecules. Microbial sRNAs, many of which are
<250 nt, play important regulatory roles in diverse cellular
processes, ranging from housekeeping functions to virulence
(110, 123, 144–147). sRNAs in metatranscriptome data
sets can be identified by homology, for example, to sequences
in Rfam, an open-access database of noncoding RNA families
(148). The structure (secondary and tertiary) of noncoding
RNAs, including sRNAs, is often more evolutionarily con-
served than the primary RNA sequence, and has a strong
influence on the function of the RNA molecule (149–
151). Consequently, databases such as Rfam are queried
using algorithms that account for both primary sequence
and secondary structure conservation (the INFERNAL [152]
or CentroidHomfold-LAST [89] programs). Additionally, as
most microbial sRNAs occur within intergenic regions, the
mapping of transcriptome RNA fragments to intergenic
regions on reference genomes can help identify candidate reg-
ulatory molecules (55).
A comprehensive analysis of marine bacterioplankton
transcriptomes identified abundant sRNAs from known
sRNA families, such as riboswitches and signal recognition
particle RNAs, as well as a variety of putative sRNAs of
unknown function (55). However, sRNAs in other environ-
mental metatranscriptome data sets have been relatively
underexplored, potentially due to the challenge of defini-
tively assigning sRNA function and to the biased recovery
of longer RNA fragments by upstream RNA processing steps.
(As discussed already, certain RNA extraction or column
purification methods exclude molecules <200 nt; others per-
mit recovery of the full range of RNA molecule sizes, from kil-
obases to 10-mers [e.g., the mirVana miRNA isolation kit;
Ambion].) As metatranscriptome data accumulate and cura-
ted databases of noncoding RNAs continue to expand, we
anticipate an increased focus on sRNA recovery and analysis.
Patterns in this regulatory RNA pool may provide important
insight into how natural microbial communities respond to
environmental fluctuation.
Analysis Packages
A variety of public and commercial packages are available
for comprehensive analysis of meta-omic data sets (Table
3). These packages provide user-friendly pipelines that com-
bine many of the steps described here, from quality filter-
ing to statistical analysis, with a range of visualization
tools for data set comparison and presentation of hit counts
parsed according to function and taxonomy. Web-accessible
platforms such as MG-RAST support similarity searches on
external servers, thereby providing valuable resources for
users without access to local computing infrastructure. In
addition, MG-RAST serve as repositories of environmen-
tal metagenomic and metatranscriptomic data sets, many
of which are only searchable using through these platforms
(Table 3).
Although powerful, some meta-omic data processing pro-
grams may be restrictive for environmental samples whose
members or community metabolic functions are not well rep-
resented in the databases used for functional and taxonomic
classification. For example, in the BLAST-parsing program
MEGAN, which can be used to parse results of BLAST
searches against the NCBI-nr database, only genes with an
NCBI RefSeq identifier are included in the functional classi-
fication. Many reference genes in the NCBI-nr database,
however, do not have a RefSeq identifier. Indeed, a recent
analysis of microbial metagenomes from a marine oxygen
minimum zone indicated that the majority of reads with sig-
nificant BLASTX matches to genes in the NCBI-nr database
were excluded from a MEGAN-based classification (into
SEED categories; 90). Similarly, a BLASTX analysis of oxy-
gen minimum zone metatranscriptome sequences failed to
uncover some of the most highly expressed genes in the sam-
ples (encoding ammonia monooxygenase or Archaea; 102)
because homologs of these genes were not represented in
the KEGG Orthology database (KEGG may also used for
functional gene grouping in MG-RAST and HUMAnN;
Table 3). Identification of these sequences instead required
manual parsing of BLASTX results from a search against
the NCBI-nr database. This example illustrates the impor-
tance of manual data exploration and analysis using UNIX
and other programming languages to handle and extract
information from large data files.
The potential for classification schemes to exclude impor-
tant biological signals highlights the utility of using a variety
of databases for sequence annotation, as well as both auto-
mated and manual methods for exploring search results.
The information content of metatranscriptome data sets
will continue to grow as sequencing decreases in cost and
sequencing depth and sample number increase. Maximiz-
ing the extraction of biological information using metatran-
scriptomics will require a combination of the distributed use
of this technology across diverse environments and experi-
mental designs, as well as creative new approaches for
data analysis. The manual exploration of sequence data using
novel resources (e.g., task-specific scripts) therefore will
remain a vital component of meta-omic research, and an
important contributor to the refinement of standard analysis
pipelines.
I thank Ed DeLong for constructive input during the preparation of
this chapter. This work was made possible by generous support from the
Alfred P. Sloan Foundation and the National Science Foundation (Grant
1151698).
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-15
REFERENCES
1. Frias-Lopez J, Shi Y, Tyson GW, Coleman ML, Schuster
SC, Chisholm SW, Delong EF. 2008. Microbial commun-
ity gene expression in ocean surface waters. Proc Natl Acad
Sci USA 105:3805–3810.
2. Stewart FJ, Sharma AK, Bryant JA, Eppley JM,
DeLong EF. 2011. Community transcriptomics reveals
universal patterns of protein sequence conservation
in natural microbial communities. Genome Biol 12:
R26.
3. Moran MA, Satinsky B, Gifford SM, Luo H, Rivers A,
Chan LK, Meng J, Durham BP, Shen C, Varaljay VA,
Smith CB, Yager PL, Hopkinson BM. 2013. Sizing up
metatranscriptomics. ISME J 7:237–243.
4. Taniguchi Y, Choi PJ, Li GW, Chen H, Babu M, Hearn J,
Emili A, Xie XS. 2010. Quantifying E. coli proteome and
transcriptome with single-molecule sensitivity in single
cells. Science 329:533–538.
5. Steglich C, Lindell D, Futschik M, Rector T, Steen R,
Chisholm SW. 2010. Short RNA half-lives in the slow-
growing marine cyanobacterium Prochlorococcus. Genome
Biol 11:R54.
6. Stewart FJ. 2013. Preparation of microbial community
cDNA for metatranscriptomic analysis in marine plankton.
Meth Enzymol 531:187–218.
7. Barnett TC, Bugrysheva JV, Scott JR. 2007. Role of
mRNA stability in growth phase regulation of gene expres-
sion in the group A streptococcus. J Bacteriol 189:
1866–1873.
8. Kristoffersen SM, Haase C, Weil MR, Passalacqua KD,
Niazi F, Hutchison SK, Desany B, Kolsto AB, Tourasse
NJ, Read TD, Okstad OA. 2012. Global mRNA decay
analysis at single nucleotide resolution reveals segmental
and positional degradation patterns in a Gram-positive bac-
terium. Genome Biol 13:R30.
9. Deutscher MP. 2006. Degradation of RNA in bacteria:
comparison of mRNA and stable RNA. Nucleic Acids Res
34:659–666.
10. Deutscher MP. 2009. Maturation and degradation of ribo-
somal RNA in bacteria. Prog Mol Biol Transl Sci 85:
369–391.
11. Roume H, Muller EE, Cordes T, Renaut J, Hiller K,
Wilmes P. 2013. A biomolecular isolation framework for
eco-systems biology. ISME J 7:110–121.
12. Gifford SM, Sharma S, Rinta-Kanto JM, Moran MA.
2011. Quantitative analysis of a deeply sequenced marine
microbial metatranscriptome. ISME J 5:461–472.
13. Liu L, Li Y, Li S, Hu N, He Y, Pong R, Lin D, Lu L, Law
M. 2012. Comparison of next-generation sequencing sys-
tems. J Biomed Biotechnol 2012:251364.
14. Solonenko SA, Ignacio-Espinoza JC, Alberti A, Cruaud
C, Hallam S, Konstantinidis K, Tyson G, Wincker P,
Sullivan MB. 2013. Sequencing platform and library prep-
aration choices impact viral metagenomes. BMC Genomics
14:320.
15. Orsi WD, Edgcomb VP, Christman GD, Biddle JF. 2013.
Gene expression in the deep biosphere. Nature 499:
205–208.
16. Stewart FJ, Ottesen EA, DeLong EF. 2010. Development
and quantitative analyses of a universal rRNA-subtraction
protocol for microbial metatranscriptomics. ISME J 4:
896–907.
17. Urich T, Lanzen A, Qi J, Huson DH, Schleper C, Schus-
ter SC. 2008. Simultaneous assessment of soil microbial
community structure and function through analysis of the
meta-transcriptome. PLoS One 3:e2527.
18. Wommack KE, Bhavsar J, Ravel J. 2008. Metageno-
mics: read length matters. Appl Environ Microbiol 74:
1453–1463.
19. Schadt EE, Linderman MD, Sorenson J, Lee L,
Nolan GP. 2010. Computational solutions to large-
scale data management and analysis. Nat Rev Genet 11:
647–657.
20. Stein LD. 2010. The case for cloud computing in genome
informatics. Genome Biol 11:207.
21. Thakur RS, Bandopadhyay R, Chaudhary B, Chatterjee
S. 2012. Now and next-generation sequencing techniques:
future of sequence analysis using cloud computing. Front
Genet 3:280.
22. Qiu J, Ekanayake J, Gunarathne T, Choi JY, Bae SH, Li
H, Zhang B, Wu TL, Ruan Y, Ekanayake S, Hughes A,
Fox G. 2010. Hybrid cloud and cluster computing para-
digms for life science applications. BMC Bioinformatics
11:S3.
23. Tisdall JD. 2001. Beginning Perl for Bioinformatics. O’Reilly,
Beijing; Sebastopol, CA.
24. Bradnam K, Korf I. 2012. UNIX and Perl to the Rescue!: A
Field Guide for the Life Sciences (And Other Data-Rich Pur-
suits). Cambridge University Press, New York.
25. Goecks J, Nekrutenko A, Taylor J, Galaxy T. 2010. Gal-
axy: a comprehensive approach for supporting accessible,
reproducible, and transparent computational research in
the life sciences. Genome Biol 11:R86.
26. Meyer F, Paarmann D, D’Souza M, Olson R, Glass EM,
Kubal M, Paczian T, Rodriguez A, Stevens R, Wilke A,
Wilkening J, Edwards RA. 2008. The metagenomics
RAST server—a public resource for the automatic phyloge-
netic and functional analysis of metagenomes. BMC Bioin-
formatics 9:386.
27. Sun S, Chen J, Li W, Altintas I, Lin A, Peltier S, Stocks
K, Allen EE, Ellisman M, Grethe J, Wooley J. 2011. Com-
munity cyberinfrastructure for advanced microbial ecology
research and analysis: the CAMERA resource. Nucleic
Acids Res 39:D546–D551.
28. Markowitz VM, Chen IM, Palaniappan K, Chu K, Szeto
E, Grechkin Y, Ratner A, Jacob B, Huang J, Williams P,
Huntemann M, Anderson I, Mavromatis K, Ivanova NN,
Kyrpides NC. 2012. IMG: the Integrated Microbial
Genomes database and comparative analysis system.
Nucleic Acids Res 40:D115–D122.
29. De Wit P, Pespeni MH, Ladner JT, Barshis DJ, Seneca F,
Jaris H, Therkildsen NO, Morikawa M, Palumbi SR.
2012. The simple fool’s guide to population genomics via
RNA-Seq: an introduction to high-throughput sequencing
data analysis. Mol Ecol Resour 12:1058–1067.
30. Wolf JB. 2013. Principles of transcriptome analysis and
gene expression quantification: an RNA-seq tutorial. Mol
Ecol Resour 13:559–572.
31. Hoeijmakers WA, Bartfai R, Stunnenberg HG. 2013.
Transcriptome analysis using RNA-Seq. Meth Mol Biol
923:221–239.
32. Garber M, Grabherr MG, Guttman M, Trapnell C.
2011. Computational methods for transcriptome annota-
tion and quantification using RNA-seq. Nat Meth 8:
469–477.
33. Soneson C, Delorenzi M. 2013. A comparison of methods
for differential expression analysis of RNA-seq data. BMC
Bioinformatics 14:91.
34. Bullard JH, Purdom E, Hansen KD, Dudoit S. 2010.
Evaluation of statistical methods for normalization and dif-
ferential expression in mRNA-Seq experiments. BMC Bio-
informatics 11:94.
35. Dillies MA, Rau A, Aubert J, Hennequet-Antier C, Jean-
mougin M, Servant N, Keime C, Marot G, Castel D,
Estelle J, Guernec G, Jagla B, Jouneau L, Laloe D, Le
Gall C, Schaeffer B, Le Crom S, Guedj M, Jaffrezic F,
on behalf of The French StatOmique C. 2012. A compre-
hensive evaluation of normalization methods for Illumina
2.4.5-16 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
high-throughput RNA sequencing data analysis. Brief Bio-
inform 14:671–683.
36. Sun Z, Zhu Y. 2012. Systematic comparison of RNA-Seq
normalization methods using measurement error models.
Bioinformatics 28:2584–2591.
37. Feike J, Jurgens K, Hollibaugh JT, Kruger S, Jost G,
Labrenz M. 2012. Measuring unbiased metatranscriptom-
ics in suboxic waters of the central Baltic Sea using a new
in situ fixation system. ISME J 6:461–470.
38. Robidart J, Callister SJ, Song P, Nicora CD, Wheat CG,
Girguis PR. 2013. Characterizing microbial community
and geochemical dynamics at hydrothermal vents using
osmotically driven continuous fluid samplers. Environ Sci
Technol 47:4399–4407.
39. von Bergen M, Jehmlich N, Taubert M, Vogt C, Bastida
F, Herbst FA, Schmidt F, Richnow HH, Seifert J. 2013.
Insights from quantitative metaproteomics and protein-
stable isotope probing into microbial ecology. ISME J 7:
1877–1885.
40. Wang DZ, Xie ZX, Zhang SF. 2014. Marine metaproteo-
mics: current status and future directions. J Proteomics 97:
27–35.
41. Service RF. 2006. Gene sequencing. The race for the
$1000 genome. Science 311:1544–1546.
42. Wall DP, Kudtarkar P, Fusaro VA, Pivovarov R, Patil P,
Tonellato PJ. 2010. Cloud computing for comparative
genomics. BMC Bioinformatics 11:259.
43. Belasco J. 1993. mRNA degradation in prokaryotic cells: an
overview in control of messenger RNA stability, p 3–12. In
Belasco J, Brawerman G (eds), Control of Messenger RNA
Stability. Academic Press, San Diego, CA.
44. Scholin C, Doucette G, Jensen S, Roman B, Pargett D,
III RM, Preston C, Jones W, Feldman J, Everlove C,
Harris A, Alvarado N, Massion E, Birch J, Greenfield
D, Vrijenhoek R, Mikulski C, Jones K. 2009. Remote
detection of marine microbes, small invertebrates, harmful
algae, and biotoxins using the environmental sample pro-
cessor (ESP). Oceanography 22:158–167.
45. Ottesen EA, Marin RIII , Preston CM, Young CR, Ryan
JP, Scholin CA, DeLong EF. 2011. Metatranscriptomic
analysis of autonomously collected and preserved marine
bacterioplankton. ISME J 5:1881–1895.
46. Mutter GL, Zahrieh D, Liu C, Neuberg D, Finkel-
stein D, Baker HE, Warrington JA. 2004. Comparison
of frozen and RNALater solid tissue storage methods for
use in RNA expression microarrays. BMC Genomics 5:
88.
47. Su JM, Perlaky L, Li XN, Leung HC, Antalffy B, Arm-
strong D, Lau CC. 2004. Comparison of ethanol versus for-
malin fixation on preservation of histology and RNA in
laser capture microdissected brain tissues. Brain Pathol 14:
175–182.
48. Bailly J, Fraissinet-Tachet L, Verner MC, Debaud JC,
Lemaire M, Wesolowski-Louvel M, Marmeisse R. 2007.
Soil eukaryotic functional diversity, a metatranscriptomic
approach. ISME J 1:632–642.
49. Shrestha PM, Kube M, Reinhardt R, Liesack W. 2009.
Transcriptional activity of paddy soil bacterial commun-
ities. Environ Microbiol 11:960–970.
50. Carvalhais LC, Dennis PG, Tyson GW, Schenk PM.
2012. Application of metatranscriptomics to soil environ-
ments. J Microbiol Meth 91:246–251.
51. Simister RL, Schmitt S, Taylor MW. 2011. Evaluating
methods for the preservation and extraction of DNA and
RNA for analysis of microbial communities in marine
sponges. J Exp Mar Biol Ecol 397:38–43.
52. Moazed D. 2009. Small RNAs in transcriptional
gene silencing and genome defence. Nature 457:413–
420.
53. Sunkar R, Zhu JK. 2004. Novel and stress-regulated micro-
RNAs and other small RNAs from Arabidopsis. Plant Cell
16:2001–2019.
54. Wassarman KM, Repoila F, Rosenow C, Storz G, Gottes-
man S. 2001. Identification of novel small RNAs using
comparative genomics and microarrays. Genes Dev 15:
1637–1651.
55. Shi Y, Tyson GW, DeLong EF. 2009. Metatranscriptomics
reveals unique microbial small RNAs in the ocean’s water
column. Nature 459:266–269.
56. Damon C, Vallon L, Zimmermann S, Haider MZ,
Galeote V, Dequin S, Luis P, Fraissinet-Tachet L, Mar-
meisse R. 2011. A novel fungal family of oligopeptide
transporters identified by functional metatranscriptomics
of soil eukaryotes. ISME J 5:1871–1880.
57. Wang Z, Gerstein M, Snyder M. 2009. RNA-Seq: a revo-
lutionary tool for transcriptomics. Nat Rev Genet 10:
57–63.
58. Armour CD, Castle JC, Chen R, Babak T, Loerch P,
Jackson S, Shah JK, Dey J, Rohl CA, Johnson JM, Ray-
mond CK. 2009. Digital transcriptome profiling using
selective hexamer priming for cDNA synthesis. Nat Meth
6:647–649.
59. Peano C, Pietrelli A, Consolandi C, Rossi E, Petiti L,
Tagliabue L, De Bellis G, Landini P. 2013. An efficient
rRNA removal method for RNA sequencing in GC-rich
bacteria. Microb inform Exper 3:1.
60. Gilbert JA, Field D, Huang Y, Edwards R, Li W, Gilna P,
Joint I. 2008. Detection of large numbers of novel sequen-
ces in the metatranscriptomes of complex marine microbial
communities. PLoS One 3:e3042.
61. Kim KH, Bae JW. 2011. Amplification methods bias
metagenomic libraries of uncultured single-stranded and
double-stranded DNA viruses. Appl Environ Microbiol 77:
7663–7668.
62. Yilmaz S, Allgaier M, Hugenholtz P. 2010. Multiple dis-
placement amplification compromises quantitative analysis
of metagenomes. Nat Meth 7:943–944.
63. Chen Z, Duan X. 2011. Ribosomal RNA depletion for
massively parallel bacterial RNA-sequencing applications.
Meth Mol Biol 733:93–103.
64. He S, Wurtzel O, Singh K, Froula JL, Yilmaz S, Tringe
SG, Wang Z, Chen F, Lindquist EA, Sorek R, Hugen-
holtz P. 2010. Validation of two ribosomal RNA removal
methods for microbial metatranscriptomics. Nat Meth 7:
807–812.
65. Yi H, Cho YJ, Won S, Lee JE, Jin Yu H, Kim S, Schroth
GP, Luo S, Chun J. 2011. Duplex-specific nuclease effi-
ciently removes rRNA for prokaryotic RNA-seq. Nucleic
Acids Res 39:e140.
66. Zhulidov PA, Bogdanova EA, Shcheglov AS, Vagner LL,
Khaspekov GL, Kozhemyako VB, Matz MV, Meleshke-
vitch E, Moroz LL, Lukyanov SA, Shagin DA. 2004.
Simple cDNA normalization using kamchatka crab duplex-
specific nuclease. Nucleic Acids Res 32:e37.
67. Young BD, Anderson M. 1985. Quantitative analysis of
solution hybridization, p 47–71. In Hames BD, Higgins
SJ (eds), Nucleic Acids Hybridisation, a Practical Approach.
IRL Press, Oxford.
68. Giannoukos G, Ciulla DM, Huang K, Haas BJ, Izard J,
Levin JZ, Livny J, Earl AM, Gevers D, Ward DV, Nus-
baum C, Birren BW, Gnirke A. 2012. Efficient and robust
RNA-seq process for cultured bacteria and complex com-
munity transcriptomes. Genome Biol 13:R23.
69. Hewson I, Poretsky RS, Dyhrman ST, Zielinski B,
White AE, Tripp HJ, Montoya JP, Zehr JP. 2009. Micro-
bial community gene expression within colonies of the
diazotroph, Trichodesmium, from the southwest Pacific
Ocean. ISME J 3:1286–1300.
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-17
70. Spiess AN, Mueller N, Ivell R. 2003. Amplified RNA
degradation in T7-amplification methods results in biased
microarray hybridizations. BMC Genomics 4:44.
71. Nygaard V, Hovig E. 2006. Options available for profiling
small samples: a review of sample amplification technology
when combined with microarray profiling. Nucleic Acids
Res 34:996–1014.
72. Kurn N, Chen P, Heath JD, Kopf-Sill A, Stephens KM,
Wang S. 2005. Novel isothermal, linear nucleic acid ampli-
fication systems for highly multiplexed applications. Clin
Chem 51:1973–1981.
73. Clement-Ziza M, Gentien D, Lyonnet S, Thiery JP,
Besmond C, Decraene C. 2009. Evaluation of methods
for amplification of picogram amounts of total RNA for
whole genome expression profiling. BMC Genomics 10:
246.
74. Wu J, Gao W, Zhang W, Meldrum DR. 2011. Optimiza-
tion of whole-transcriptome amplification from low cell
density deep-sea microbial samples for metatranscriptomic
analysis. J Microbiol Meth 84:88–93.
75. Feldman AL, Costouros NG, Wang E, Qian M, Marincola
FM, Alexander HR, Libutti SK. 2002. Advantages of
mRNA amplification for microarray analysis. BioTechniques
33:906–912, 914.
76. Moreno-Paz M, Parro V. 2006. Amplification of low quan-
tity bacterial RNA for microarray studies: time-course anal-
ysis of Leptospirillum ferrooxidans under nitrogen-fixing
conditions. Environ Microbiol 8:1064–1073.
77. Rachman H, Lee JS, Angermann J, Kowall J, Kaufmann
SH. 2006. Reliable amplification method for bacterial
RNA. J Biotechnol 126:61–68.
78. Hansen KD, Brenner SE, Dudoit S. 2010. Biases in Illu-
mina transcriptome sequencing caused by random hexamer
priming. Nucleic Acids Res 38:e131.
79. Nam DK, Lee S, Zhou G, Cao X, Wang C, Clark T,
Chen J, Rowley JD, Wang SM. 2002. Oligo(dT) primer
generates a high frequency of truncated cDNAs through
internal poly(A) priming during reverse transcription.
Proc Natl Acad Sci USA 99:6152–6156.
80. Borodina T, Adjaye J, Sultan M. 2011. A strand-specific
library preparation protocol for RNA sequencing. Meth
Enzymol 500:79–98.
81. Levin JZ, Yassour M, Adiconis X, Nusbaum C, Thomp-
son DA, Friedman N, Gnirke A, Regev A. 2010. Com-
prehensive comparative analysis of strand-specific RNA
sequencing methods. Nat Meth 7:709–715.
82. Raghavan R, Sloan DB, Ochman H. 2012. Antisense
transcription is pervasive but rarely conserved in enteric
bacteria. mBio 3:e00156-12.
83. Parkhomchuk D, Borodina T, Amstislavskiy V,
Banaru M, Hallen L, Krobitsch S, Lehrach H,
Soldatov A. 2009. Transcriptome analysis by strand-
specific sequencing of complementary DNA. Nucleic Acids
Res 37:e123.
84. Mamanova L, Andrews RM, James KD, Sheridan EM,
Ellis PD, Langford CF, Ost TW, Collins JE, Turner DJ.
2010. FRT-seq: amplification-free, strand-specific tran-
scriptome sequencing. Nat Meth 7:130–132.
85. Margulies M, Egholm M, Altman WE, Attiya S, Bader JS,
Bemben LA, Berka J, Braverman MS, Chen YJ, Chen Z,
Dewell SB, Du L, Fierro JM, Gomes XV, Godwin BC,
He W, Helgesen S, Ho CH, Irzyk GP, Jando SC,
Alenquer ML, Jarvie TP, Jirage KB, Kim JB, Knight
JR, Lanza JR, Leamon JH, Lefkowitz SM, Lei M, Li J,
Lohman KL, Lu H, Makhijani VB, McDade KE,
McKenna MP, Myers EW, Nickerson E, Nobile JR, Plant
R, Puc BP, Ronan MT, Roth GT, Sarkis GJ, Simons JF,
Simpson JW, Srinivasan M, Tartaro KR, Tomasz A, Vogt
KA, Volkmer GA, et al. 2005. Genome sequencing in
microfabricated high-density picolitre reactors. Nature
437:376–380.
86. Quince C, Lanzen A, Curtis TP, Davenport RJ, Hall N,
Head IM, Read LF, Sloan WT. 2009. Accurate determina-
tion of microbial diversity from 454 pyrosequencing data.
Nat Meth 6:639–641.
87. Lohse M, Bolger AM, Nagel A, Fernie AR, Lunn JE, Stitt
M, Usadel B. 2012. RobiNA: a user-friendly, integrated
software solution for RNA-Seq-based transcriptomics.
Nucleic Acids Res 40:W622–W627.
88. Ewing B, Green P. 1998. Base-calling of automated
sequencer traces using phred. II. Error probabilities. Genome
Res 8:186–194.
89. Hamada M, Yamada K, Sato K, Frith MC, Asai K. 2011.
CentroidHomfold-LAST: accurate prediction of RNA sec-
ondary structure using automatically collected homologous
sequences. Nucleic Acids Res 39:W100–W106.
90. Ganesh S, Parris DJ, Delong EF, Stewart FJ. 2014. Meta-
genomic analysis of size-fractionated picoplankton in a
marine oxygen minimum zone. ISME J 8:187–211.
91. Fu L, Niu B, Zhu Z, Wu S, Li W. 2012. CD-HIT: accel-
erated for clustering the next-generation sequencing data.
Bioinformatics 28:3150–3152.
92. Schmieder R, Lim YW, Edwards R. 2012. Identification
and removal of ribosomal RNA sequences from metatran-
scriptomes. Bioinformatics 28:433–435.
93. Lee JH, Yi H, Chun J. 2011. rRNASelector: a computer
program for selecting ribosomal RNA encoding sequences
from metagenomic and metatranscriptomic shotgun libra-
ries. J Microbiol 49:689–691.
94. Kopylova E, Noe L, Touzet H. 2012. SortMeRNA: fast
and accurate filtering of ribosomal RNAs in metatranscrip-
tomic data. Bioinformatics 28:3211–3217.
95. Trapnell C, Williams BA, Pertea G, Mortazavi A,
Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter
L. 2010. Transcript assembly and quantification by
RNA-Seq reveals unannotated transcripts and isoform
switching during cell differentiation. Nat Biotechnol 28:
511–515.
96. Namiki T, Hachiya T, Tanaka H, Sakakibara Y. 2012.
MetaVelvet: an extension of Velvet assembler to de novo
metagenome assembly from short sequence reads. Nucleic
Acids Res 40:e155.
97. Robertson G, Schein J, Chiu R, Corbett R, Field M, Jack-
man SD, Mungall K, Lee S, Okada HM, Qian JQ, Griffith
M, Raymond A, Thiessen N, Cezard T, Butterfield YS,
Newsome R, Chan SK, She R, Varhol R, Kamoh B,
Prabhu AL, Tam A, Zhao Y, Moore RA, Hirst M, Marra
MA, Jones SJ, Hoodless PA, Birol I. 2010. De novo
assembly and analysis of RNA-seq data. Nat Meth 7:
909–912.
98. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K,
Bushman FD, Costello EK, Fierer N, Pena AG, Goodrich
JK, Gordon JI, Huttley GA, Kelley ST, Knights D, Koe-
nig JE, Ley RE, Lozupone CA, McDonald D, Muegge
BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ,
Walters WA, Widmann J, Yatsunenko T, Zaneveld J,
Knight R. 2010. QIIME allows analysis of high-throughput
community sequencing data. Nat Meth 7:335–336.
99. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann
M, Hollister EB, Lesniewski RA, Oakley BB, Parks DH,
Robinson CJ, Sahl JW, Stres B, Thallinger GG,
Van Horn DJ, Weber CF. 2009. Introducing mothur:
open-source, platform-independent, community-supported
software for describing and comparing microbial commun-
ities. Appl Environ Microbiol 75:7537–7541.
100. Gomez-Alvarez V, Teal TK, Schmidt TM. 2009. System-
atic artifacts in metagenomes from complex microbial com-
munities. ISME J 3:1314–1317.
2.4.5-18 ▪ MICROBIAL COMMUNITY ANALYSIS OF ENVIRONMENTAL SAMPLES
101. Stewart FJ, Dalsgaard T, Young CR, Thamdrup B,
Revsbech NP, Ulloa O, Canfield DE, Delong EF. 2012.
Experimental incubations elicit profound changes in com-
munity transcription in OMZ bacterioplankton. PLoS One
7:e37118.
102. Stewart FJ, Ulloa O, DeLong EF. 2012. Microbial meta-
transcriptomics in a permanent marine oxygen minimum
zone. Environ Microbiol 14:23–40.
103. Su X, Xu J, Ning K. 2012. Parallel-META: efficient meta-
genomic data analysis based on high-performance compu-
tation. BMC Syst Biol 6:S16.
104. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie
EL, Keller K, Huber T, Dalevi D, Hu P, Andersen GL.
2006. Greengenes, a chimera-checked 16S rRNA gene
database and workbench compatible with ARB. Appl Envi-
ron Microbiol 72:5069–5072.
105. Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun
Y, Brown CT, Porras-Alfaro A, Kuske CR, Tiedje JM.
2014. Ribosomal Database Project: data and tools for
high throughput rRNA analysis. Nucleic Acids Res 42:
D633–642.
106. Huson DH, Mitra S, Ruscheweyh HJ, Weber N, Schus-
ter SC. 2011. Integrative analysis of environmental sequen-
ces using MEGAN4. Genome Res 21:1552–1560.
107. Condon C (ed). 2009. Molecular Biology of RNA Process-
ing and Decay in Prokaryotes. Progress in Molecular Biology
and Translational Science series title. Elsevier, Amsterdam.
108. Nagarajan N, Pop M. 2013. Sequence assembly demysti-
fied. Nat Rev Genet 14:157–167.
109. Martin JA, Wang Z. 2011. Next-generation transcriptome
assembly. Nat Rev Genet 12:671–682.
110. Guell M, Yus E, Lluch-Senar M, Serrano L. 2011. Bacte-
rial transcriptomics: what is beyond the RNA horiz-ome?
Nat Rev Microbiol 9:658–669.
111. Schulz MH, Zerbino DR, Vingron M, Birney E. 2012.
Oases: robust de novo RNA-seq assembly across the
dynamic range of expression levels. Bioinformatics 28:
1086–1092.
112. Grabherr MG, Haas BJ, Yassour M, Levin JZ, Thompson
DA, Amit I, Adiconis X, Fan L, Raychowdhury R, Zeng
Q, Chen Z, Mauceli E, Hacohen N, Gnirke A, Rhind N,
di Palma F, Birren BW, Nusbaum C, Lindblad-Toh K,
Friedman N, Regev A. 2011. Full-length transcriptome
assembly from RNA-Seq data without a reference genome.
Nat Biotechnol 29:644–652.
113. Langmead B, Salzberg SL. 2012. Fast gapped-read align-
ment with Bowtie 2. Nat Meth 9:357–359.
114. Li H, Durbin R. 2009. Fast and accurate short read align-
ment with Burrows-Wheeler transform. Bioinformatics 25:
1754–1760.
115. Rumble SM, Lacroute P, Dalca AV, Fiume M, Sidow A,
Brudno M. 2009. SHRiMP: accurate mapping of short
color-space reads. PLoS Comput Biol 5:e1000386.
116. Lunter G, Goodson M. 2011. Stampy: a statistical algo-
rithm for sensitive and fast mapping of Illumina sequence
reads. Genome Res 21:936–939.
117. Hewson I, Poretsky RS, Beinart RA, White AE, Shi T,
Bench SR, Moisander PH, Paerl RW, Tripp HJ, Montoya
JP, Moran MA, Zehr JP. 2009. In situ transcriptomic anal-
ysis of the globally important keystone N2-fixing taxon
Crocosphaera watsonii. ISME J 3:618–631.
118. Poretsky RS, Hewson I, Sun S, Allen AE, Zehr JP,
Moran MA. 2009. Comparative day/night metatran-
scriptomic analysis of microbial communities in the
North Pacific subtropical gyre. Environ Microbiol 11:
1358–1375.
119. Besemer J, Lomsadze A, Borodovsky M. 2001. Gene-
MarkS: a self-training method for prediction of gene starts
in microbial genomes. Implications for finding sequence
motifs in regulatory regions. Nucleic Acids Res 29:
2607–2618.
120. Delcher AL, Harmon D, Kasif S, White O, Salzberg SL.
1999. Improved microbial gene identification with
GLIMMER. Nucleic Acids Res 27:4636–4641.
121. Hyatt D, Chen GL, Locascio PF, Land ML, Larimer FW,
Hauser LJ. 2010. Prodigal: prokaryotic gene recognition
and translation initiation site identification. BMC Bioin-
form 11:119.
122. Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M.
2012. KEGG for integration and interpretation of large-
scale molecular data sets. Nucleic Acids Res 40:
D109–D114.
123. Chambers JR, Sauer K. 2013. Small RNAs and their role
in biofilm formation. Trends Microbiol 21:39–49.
124. Overbeek R, Begley T, Butler RM, Choudhuri JV,
Chuang HY, Cohoon M, de Crecy-Lagard V, Diaz N,
Disz T, Edwards R, Fonstein M, Frank ED, Gerdes S,
Glass EM, Goesmann A, Hanson A, Iwata-Reuyl D, Jen-
sen R, Jamshidi N, Krause L, Kubal M, Larsen N, Linke
B, McHardy AC, Meyer F, Neuweger H, Olsen G, Olson
R, Osterman A, Portnoy V, Pusch GD, Rodionov DA,
Ruckert C, Steiner J, Stevens R, Thiele I, Vassieva O,
Ye Y, Zagnitko O, Vonstein V. 2005. The subsystems
approach to genome annotation and its use in the project
to annotate 1000 genomes. Nucleic Acids Res 33:5691–
5702.
125. Yu K, Zhang T. 2013. Construction of customized sub-
databases from NCBI-nr database for rapid annotation of
huge metagenomic datasets using a combined BLAST
and MEGAN approach. PLoS One 8:e59831.
126. Kanehisa M, Goto S. 2000. KEGG: Kyoto encyclopedia of
genes and genomes. Nucleic Acids Res 28:27–30.
127. Tatusov RL, Koonin EV, Lipman DJ. 1997. A
genomic perspective on protein families. Science 278:
631–637.
128. Chen C, Huang H, Wu CH. 2011. Protein bioin-
formatics databases and resources. Methods Mol Biol 694:
3–24.
129. Xiong X, Frank DN, Robertson CE, Hung SS, Markle J,
Canty AJ, McCoy KD, Macpherson AJ, Poussier P, Dan-
ska JS, Parkinson J. 2012. Generation and analysis of a
mouse intestinal metatranscriptome through Illumina
based RNA-sequencing. PLoS One 7:e36009.
130. David LA, Maurice CF, Carmody RN, Gootenberg DB,
Button JE, Wolfe BE, Ling AV, Devlin AS, Varma Y,
Fischbach MA, Biddinger SB, Dutton RJ, Turnbaugh
PJ. 2013. Diet rapidly and reproducibly alters the human
gut microbiome. Nature 505:559–563.
131. Powell S, Szklarczyk D, Trachana K, Roth A, Kuhn M,
Muller J, Arnold R, Rattei T, Letunic I, Doerks T, Jensen
LJ, von Mering C, Bork P. 2012. eggNOG v3 0: ortholo-
gous groups covering 1133 organisms at 41 different taxo-
nomic ranges. Nucleic Acids Res 40:D284–D289.
132. Abubucker S, Segata N, Goll J, Schubert AM, Izard J,
Cantarel BL, Rodriguez-Mueller B, Zucker J, Thiagara-
jan M, Henrissat B, White O, Kelley ST, Methe B,
Schloss PD, Gevers D, Mitreva M, Huttenhower C.
2012. Metabolic reconstruction for metagenomic data
and its application to the human microbiome. PLoS Com-
put Biol 8:e1002358.
133. Kent WJ. 2002. BLAT—the BLAST-like alignment tool.
Genome Res 12:656–664.
134. Ye Y, Choi JH, Tang H. 2011. RAPSearch: a fast protein
similarity search tool for short reads. BMC Bioinform 12:
159.
135. Kielbasa SM, Wan R, Sato K, Horton P, Frith MC. 2011.
Adaptive seeds tame genomic sequence comparison.
Genome Res 21:487–493.
 2.4.5. Generation and Analysis of Microbial Metatranscriptomes ▪ 2.4.5-19
136. Paulson JN, Stine OC, Bravo HC, Pop M. 2013. Differen-
tial abundance analysis for microbial marker-gene surveys.
Nat Meth 10:1200–1202.
137. Hardcastle TJ, Kelly KA. 2010. baySeq: empirical Baye-
sian methods for identifying differential expression in
sequence count data. BMC Bioinform 11:422.
138. Anders S, Huber W. 2010. Differential expression analysis
for sequence count data. Genome Biol 11:R106.
139. Robinson MD, McCarthy DJ, Smyth GK. 2010. edgeR:
a Bioconductor package for differential expression analysis
of digital gene expression data. Bioinformatics 26:139–140.
140. Tarazona S, Garcia-Alcalde F, Dopazo J, Ferrer A,
Conesa A. 2011. Differential expression in RNA-seq: a
matter of depth. Genome Res 21:2213–2223.
141. R Development Core Team. 2013. R: A Language and
Environment for Statistical Computing. Vienna, R. Develop-
ment Core Team.
142. Yendrek CR, Ainsworth EA, Thimmapuram J. 2012.
The bench scientist’s guide to statistical analysis of RNA-
Seq data. BMC Res Notes 5:506.
143. Pawitan Y, Michiels S, Koscielny S, Gusnanto A, Ploner
A. 2005. False discovery rate, sensitivity and sample size for
microarray studies. Bioinformatics 21:3017–3024.
144. Sorek R, Cossart P. 2010. Prokaryotic transcriptomics:
a new view on regulation, physiology and pathogenicity.
Nat Rev Genet 11:9–16.
145. Storz G, Vogel J, Wassarman KM. 2011. Regulation by
small RNAs in bacteria: expanding frontiers. Mol Cell
43:880–891.
146. Toledo-Arana A, Repoila F, Cossart P. 2007. Small non-
coding RNAs controlling pathogenesis. Curr Opin Micro-
biol 10:182–188.
147. Finnegan EJ, Matzke MA. 2003. The small RNA world.
J Cell Sci 116:4689–4693.
148. Burge SW, Daub J, Eberhardt R, Tate J, Barquist L,
Nawrocki EP, Eddy SR, Gardner PP, Bateman A. 2013.
Rfam 11 0: 10 years of RNA families. Nucleic Acids Res
41:D226–D232.
149. Breaker RR. 2012. Riboswitches and the RNA world.
Cold Spring Harb Perspect Biol 4:a003566.
150. Kozak M. 2005. Regulation of translation via mRNA
structure in prokaryotes and eukaryotes. Gene 361:13–37.
151. Warf MB, Berglund JA. 2010. Role of RNA structure in
regulating pre-mRNA splicing. Trends Biochem Sci 35:
169–178.
152. Nawrocki EP, Kolbe DL, Eddy SR. 2009. Infernal 1 0:
inference of RNA alignments. Bioinformatics 25:1335–
1337.
153. Rodrigue S, Materna AC, Timberlake SC, Blackburn
MC, Malmstrom RR, Alm EJ, Chisholm SW. 2010.
Unlocking short read sequencing for metagenomics.
PLOS ONE 5:e11840.
154. Magoc T, Salzberg SL. 2011. FLASH: fast length adjust-
ment of short reads to improve genome assemblies. Bioinfor-
matics 27:2957–2963.
155. Masella AP, Bartram AK, Truszkowski JM, Brown DG,
Neufeld JD. 2012. PANDAseq: paired-end assembler for
illumina sequences. BMC bioinformatics 13:31.
156. Zerbino DR, Birney E. 2008. Velvet: algorithms for de
novo short read assembly using de Bruijn graphs. Genome
research 18:821–829.
157. Guttman M, Garber M, Levin JZ, Donaghey J, Robinson
J, Adiconis X, Fan L, Koziol MJ, Gnirke A, Nusbaum C,
Rinn JL, Lander ES, Regev A. 2010. Ab initio reconstruc-
tion of cell type-specific transcriptomes in mouse reveals
the conserved multi-exonic structure of lincRNAs. Nature
biotechnology 28:503–510.
158. Peng Y, Leung HC, Yiu SM, Chin FY. 2012. IDBA-UD:
a de novo assembler for single-cell and metagenomic
sequencing data with highly uneven depth. Bioinformatics
28:1420–1428.
159. Laserson J, Jojic V, Koller D. 2011. Genovo: de
novo assembly for metagenomes. J Comput Biol 18:429–
443.
160. Boisvert S, Raymond F, Godzaridis E, Laviolette F, Cor-
beil J. 2012. Ray Meta: scalable de novo metagenome
assembly and profiling. Genome biology 13:R122.
161. Martin J, Bruno VM, Fang Z, Meng X, Blow M, Zhang T,
Sherlock G, Snyder M, Wang Z. 2010. Annotator: an
automated de novo transcriptome assembly pipeline from
stranded RNA-Seq reads. BMC genomics 11:663.
162. Pruitt KD, Tatusova T, Maglott DR. 2005. NCBI
Reference Sequence (RefSeq): a curated non-redundant
sequence database of genomes, transcripts and proteins.
Nucleic acids research 33:D501–504.
163. Tatusova T, Ciufo S, Fedorov B, O'Neill K, Tolstoy I.
2014. RefSeq microbial genomes database: new representa-
tion and annotation strategy. Nucleic acids research 42:
D553–559.
164. UniProt C. 2013. Update on activities at the Universal
Protein Resource (UniProt) in 2013. Nucleic acids research
41:D43–47.
165. Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J,
Boursnell C, Pang N, Forslund K, Ceric G, Clements J,
Heger A, Holm L, Sonnhammer EL, Eddy SR, Bateman
A, Finn RD. 2012. The Pfam protein families database.
Nucleic acids research 40:D290–301.
166. Haft DH, Selengut JD, White O. 2003. The TIGRFAMs
database of protein families. Nucleic acids research 31:371–
373.
 Introduction to Principles of Quality Assurance
KEVIN K. CONNELL
2.5.1
The ability to verify that an analytical system is in control
for environmental microbiological methods has improved
dramatically during the past decade. The availability of new
spiking materials during this period—combined with a funda-
mental shift in the approach to quality control for environ-
mental microbiological methods—has provided research
and production laboratories, study designers, regulatory agen-
cies, and data users with powerful new tools that increase
the usefulness of analytical data for decision making.
The chapters that follow present current best practices for
environmental microbiological analysis quality control
(QC), organized according to organism types, analytical proc-
ess, and analysis objectives. This introduction provides an
overview of the QC challenges for environmental micro-
biological analyses, the role that QC plays in the different
settings where analytical methods are used, and the advances
that new materials and new approaches have made in
increased data reliability and more effective method perform-
ance assessments.
CHALLENGES FOR QC OF ENVIRONMENTAL
MICROBIOLOGICAL ANALYSES
Quantitatively determining whether an analytical system is in
control for microbiology analyses in general—and environ-
mental to introduce list of reasons analyses, in particular—
historically has been challenging: because of difficulties in
assessing the true value of a spiked sample because of uneven
distribution of organisms in the suspensions used to spike
samples, because organism concentrations can change
between enumeration and spiking, and because of the varia-
bility of the procedure used to enumerate the spike. As a
result, although environmental chemical analyses tradition-
ally have been accompanied by quantitative QC samples to
demonstrate acceptable recovery, precision, and analytical
sensitivity, microbiological QC was typically limited to qual-
itative positive and negative controls, media checks, sterility
checks, and quantitative plate count comparisons between
analysts and duplicate analyses.
Although these generally qualitative measures help guard
against false positives, they are less helpful in guarding against
false negatives or results that are biased low. False negatives
may result from nondetects when the target organism is at
concentrations below the analytical sensitivity of the method
or is suppressed from detection due to matrix interferences
doi:10.1128/9781555818821.ch2.5.1
2.5.1-1
(e.g., the field sample actually contained 200 CFU/liter of
a target bacteria, but 0 were detected). Results that are biased
low occur when the concentration of a target organism is
underreported because only a subset of the organisms in the
field sample make it through the process to be detected
(e.g., the field sample actually contained 200 CFU/liter of a
target bacteria, but only 50 CFU/liter are detected).
Without quantitative QC samples and without meaning-
ful control limits, it is difficult to assess initial and ongoing
laboratory and method performance during routine use of
a method. Without quantitative QC, it also is challenging
for regulatory agencies or standards organizations to modify
and improve the methods without conducting large-scale,
side-by-side testing.
IMPROVEMENTS IN ROUTINE QC ANALYSES
When a method is used for routine monitoring or research,
routine QC analyses serve multiple purposes. “Initial Preci-
sion and Recovery and Ongoing Precision and Recovery”
analyses using spiked samples of an interference-free matrix
(such as reagent-grade water or phosphate-buffered saline)
provide data to the laboratory on whether the method is per-
forming acceptably under ideal conditions (independent
of the potential matrix interferences of the actual field sam-
ples) and whether the laboratory is performing the method
properly. When these analyses are quantitative, and the lab-
oratory is able to review percent recovery, rather than simply
whether a spiked QC sample is positive, the potential for
low or high biases by the method can be identified and
addressed to maintain data reliability for the field samples.
“Matrix spike” analyses using spiked samples of the actual
field matrix (such as drinking water or sewage sludge) provide
data to the laboratory—and data user—on whether the meth-
od’s performance is affected by constituents in the field sam-
ples being analyzed. If the method’s recovery in an individual
matrix spike sample, or its precision across multiple matrix
spike samples, falls outside of the method-specified control
limits, the concentrations or counts of the target organisms
in the accompanying field data may be biased low or affected
in other ways. This is particularly true if the method’s per-
formance on initial and/or ongoing QC analyses in the
interference-free matrix was acceptable.
These quantitative approaches have become prevalent
not only in classical environmental microbiological methods,
2.5.1-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
where the determinative technique entails counting colonies
or individual organisms on plates or slides but also in molec-
ular biological techniques. Although molecular methods
measure something less discrete than many classical methods,
the need for quantitative QC is the same. Moreover, because
the potential for environmental matrix interferences may
actually be greater for some molecular techniques than
some classical microbiological methods, the need for quanti-
tative QC to confirm that the target material is not inhibited
is critical.
The ability to perform these routine, quantitative QC
samples itself is a significant advancement in the quality
and reliability of environmental microbiological data. By
using quantitative QC data for environmental microbiologi-
cal analyses, laboratories and data users can now assess the
magnitude and, to some extent, nature, of the problems
that may occur. As a result, more efficient and effective steps
can be taken to correct the problem at the laboratory and
more meaningful interpretation of the field data during data
use and decision making.
QUANTITATIVE SPIKES
The current ability to quantitatively assess the performance
of many environmental microbiological methods through
QC samples is a result of two complementary initiatives: (a)
increased emphasis and refinement of quantitative assess-
ments through validation and implementation of USEPA’s
1600 series microbiological methods, and (b) advances in
the quality and availability of spiking procedures and materi-
als. Beginning with USEPA Method 1622 for measurement
of Cryptosporidium in drinking water source waters, a regula-
tory need for quantitation of the levels of these organisms
and the potential financial impact of these levels on new
treatment for drinking water utilities combined to necessitate
the generation and use of quantitative QC data to maximize
monitoring data reliability.
To meet this need, USEPA Method 1622 and an ever-
increasing array of other new methods have been developed
and validated with robust QC requirements for initial and
ongoing quantitative assessments of method and laboratory
performance. In many cases, these requirements are still
met using indirect estimates of the amount of a target organ-
ism spiked into a QC sample—multiple replicates of a spiking
suspension are quantified and the average of these replicates is
used to estimate the amount of the organism spiked into the
actual QC sample. Although a dramatic improvement over
the use of nonquantitative QC to assess method performance,
this technique does have limitations—the amount is only an
estimated, indirect prediction of what the actual spiked sam-
ple contains and the variability across the representative rep-
licates tends to be significant, with coefficients of variation of
up to 30%.
These problems were ameliorated with the use of flow
cytometry to prepare spiking materials for quantitative QC.
Flow cytometers are used to quantify and sort particles in a
fluid stream based on specified properties (size, shape, and,
if stained, fluorescence). Typically used in a clinical setting
for blood and other fluids, flow cytometry is now commonly
used to prepare spiking materials that are: (1) directly meas-
ured (the same organisms counted by the flow cytometer
are introduced into the QC sample), and (2) significantly
more precise, with coefficients of variation of as low as 2%.
Moreover, flow cytometry enables very low levels of organ-
isms—tens instead of tens of thousands—to be spiked into
QC samples. As a result, organisms in QC samples can be
more representative of actual field conditions where the
target organism occurs at low levels, enabling the analytical
system to be tested with realistic conditions. This reduces
the potential for false negatives that can result if the levels
in a QC sample are much higher than environmental levels
and the QC sample result is positive, but the method is not
actually detecting the lower levels in the environment.
CONSIDERATION OF ORGANISM TYPE AND
CONDITION, BASED ON QC DATA USE
Just as using spike levels for QC samples that are meaningful
with respect to the levels of the target organism in the envi-
ronment, so is the type and condition of the organism used to
spike. Ideally, organisms of the same strain (if the method is
species-specific) and same condition (live and not fixed or
stabilized for some protozoa, for example) should be used.
However, there is a balance between the ideal and the prac-
tical and cost-effective. In some settings, it is critically impor-
tant to minimize the differences between organisms used
for spiking and those expected in the environment, even if
this increases the cost of QC. In other settings, an imperfect
match may be adequate and more cost-effective.
A setting that exemplifies the first condition is the valida-
tion or evaluation of new environmental microbiological
method for subsequent routine use by other laboratories or
for a larger data collection effort. In these instances, approx-
imating the type and condition of target organism in the envi-
ronment is critically important to minimize the potential that
the QC organisms result in better or worse method perform-
ance than the environmental organisms. If this occurs, the
initial and ongoing QC data during subsequent use of the
method may inadvertently result in biased interpretation of
the field data. Examples include the use of one strain of a
bacteria species used for QC may grow more rapidly on selec-
tive media than the environmental strains or the use of
formalin- or heat-fixed protozoa that may be recovered
more effectively through sample processing than untreated
organisms in the environment. In both cases, false negatives
or results that are biased low could occur, despite the use of
quantitative QC.
In other settings, less stringency may be merited if the use
of the original organism type and condition is cost prohibitive
and prevents adequate frequency of QC samples to be ana-
lyzed. Data use is also a significant factor in establishing this
balance between fidelity to the nature of the organism in
the environment and the QC organism. If the data from rou-
tine environmental analyses will not be used for public health
decision making but is being used for range-finding or pre-
liminary assessments, the impact that potential differences
in method performance from organism type and condition
differences may not be as critical. Finally, if a laboratory dem-
onstrates through a one-time comparison study that no signif-
icant method performance differences are demonstrated
between the QC organism type and condition and the envi-
ronmental organisms, then the QC organisms should be used
for spiking if they are more cost-effective.
QC LIMITS
To be useful to a laboratory or data user, quantitative QC
results must be compared to meaningful criteria. Without
empirically derived limits for acceptable percent recovery
for matrix spike samples or acceptable relative standard devi-
ation for initial precision and recovery samples, these QC
samples may as well be qualitative. As a result, a critical aspect
 2.5.1. Introduction to Principles of Quality Assurance ▪ 2.5.1-3
of new method validation—particularly multilaboratory val-
idation for a method that is intended for widespread use—is
not only to assess method performance but to generate data
for development of quantitative QC acceptance criteria
that will establish realistic data quality expectations. Key ele-
ments in designing studies to meet this goal include the
following:
• Using multiple laboratories to assess method performance
in the hands of different analysts and under different
laboratory-specific conditions
• Using a referee laboratory or flow cytometry to prepare and
distribute spiking materials and minimize spike variability
as a confounding variable
• Analysis of multiple spiked samples for each matrix by all
laboratories to enable method recovery and method preci-
sion to be evaluated.
• Analysis of spiked reagent water or reference matrix sam-
ples to characterize method performance in the absence of
interfering materials
• Analysis of spiked real-world matrix samples to character-
ize method performance in the matrix in which the
method would be used to monitor the target pathogen
(s) or indicator(s)
• Analysis of blank samples to verify freedom from
contamination
Although additional analytical work and data analyses are
needed to generate empirically derived QC limits, this ele-
ment is critical to establishing meaningful criteria for evalua-
tion of subsequent, routine QC samples. QC limits derived
using this approach minimize the potential for laboratories
to fail to demonstrate initial acceptable capability to perform
the method because the criteria are unrealistically high, while
also minimizing the potential for field data to be used as reli-
able when the analytical system was out of control without
being flagged because the criteria are too low.
The chapters that follow present current best practices for
environmental microbiological analysis QC, based on organ-
ism types, analytical process, and analysis objectives.
Although the approach and details on the application of
appropriate QC differ across these topics, the goal is always
the same: perform the necessary checks on the environmental
microbiological methods to ensure that data of known and
reliable quality are generated.
 General Quality Control
ROBIN K. OSHIRO
2.5.2
GENERAL QUALITY CONTROL
A common assumption for conducting microbiological test-
ing is that the laboratory environment and its equipment
are properly operated, cleaned, calibrated, and maintained;
that the laboratory is a safe working environment; and that
the laboratory has a quality assurance (QA) plan that it main-
tains and adheres to.
In order to assure that these requirements are met, a
QA program needs to be in place. A QA program typically
requires 10–20% of an investigator’s time and generally
15% of total measurement time (1). The cost depends on
the extent of the specific QA requirements. Minimally, these
requirements should include careful record-keeping practi-
ces, field and laboratory control processes, and review to
ensure accuracy, completeness, and repeatability of the
work. Other QA requirements may be added because of
enforcement, litigation, or nuclear licensing activities, which
may require extensive documentation, procurement regula-
tion, and record keeping.
A goal of the QA program is to give management the
opportunity to provide input and take responsibility in the
planning, implementation, and assessment stages of the envi-
ronmental microbiology project. The organization’s policies
and management commitment regarding QA must be
explicit.
Organizations establish a QA program in a number of
ways, depending on management support, funding and organ-
izational structure, and the likelihood for ongoing QA
requirements from funding sources. Microbiology companies,
organizations, and laboratories may elect to establish a QA
program to ensure that work done is of the best quality and
so all components of data collection, data analyses, and veri-
fication can be verified. That is, that the data were collected
and analyzed according to established protocols that were
agreed on by staff, management, and a QA manager. Interna-
tional Standard Organization (ISO) Standards, the Environ-
mental Protection Agency (2) standards, or other standards as
determined by a specific funding agency may be required or
simply preferred as the laboratory standards.
All components of an organization, especially the QA staff
and their relationship to the management and technical staff,
are important to ensure coordinated efforts and logical flow of
information. For this reason, the organizational structure and
its members’ respective roles and responsibilities are defined
doi:10.1128/9781555818821.ch2.5.2
2.5.2-1
so they understand how they interact with others. Specifying
the differences between informal and formal communications
is important. Divisions of responsibility among the technical
staff, management, supervisory staff, and any contractors
should be clear. The responsibility for quality lies with all
personnel.
A standard requirement is for the QA staff to be independ-
ent of organizational cost and scheduling constraints. For this
reason, QA staff should not be supervised by the principal
investigator for the microbiological work. Additionally, QA
staff should not be supported by an account managed by the
same personnel being audited or assessed. Even in organiza-
tions of highest integrity, to do otherwise would constitute
the appearance of a conflict of interest. Additionally, the
QA manager should hold a position at the same or higher
level as the technical work managers. In a small organization,
independence might be difficult to accomplish, and an inde-
pendent consultant might have to be contracted periodically
to assess and document compliance with the QA program.
Qualification and training requirements must be identified
to ensure that work is performed safely, efficiently, and in
compliance with the QA requirements. Experience and edu-
cation must be verified to ensure that personnel can correctly
perform the work. Project personnel should be provided tech-
nical, project orientation, and QA training. Typically, such
training consists of some form of instruction followed by
documentation that the instruction has been completed,
and it must occur before personnel commence work. QA
training and the presentation of the QA program can be
accomplished through a classroom session or by reading
assignments followed by comprehension verification (e.g.,
written exams). However, training does not ensure that indi-
viduals will remember or adhere to the requirements. For this
reason, retention of the QA requirements is improved if they
are reviewed at times during the work with staggered and
refresher training.
Both technical procedures and data must be documented
to ensure that work is performed as planned and can be
repeated if necessary. For most environmental studies, this
is done via the scientific notebook, electronic records, or a
combination of these two methods. Generally, a process
that is performed the same way each time is a standard oper-
ating procedure (SOP). Deviations or experimental proce-
dures should be documented in a scientific notebook or an
electronic version of the same. Most guidelines for controlling
2.5.2-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
scientific notebooks have specific requirements (3) for the
manner in which entries are made and corrected, and to
ensure notebook security, so that they may be admissible as
legal documents. Scientific notebook use, including organiza-
tion, formatting, entry and attachment techniques, types of
reviews, and archival protocols, should be completed uni-
formly per a SOP. In addition, the laboratory should have a
document control system for handling documents, for exam-
ple, plans and procedures, and records to ensure that only the
most recent version is being used and that they accurately
reflect completed work (2).
Nonconformance with SOPs should be followed by
corrective action and documentation of the same. Noncon-
formance is defined as a deficiency in characteristic, docu-
mentation, or procedure that renders the quality of an item
or activity unacceptable or indeterminate (4). Nonconform-
ance is therefore a potentially serious deficiency, such as
measuring a sample with an instrument that does not meet
calibration requirements or performing a test outside the
SOP for it. Discovery of nonconformance is followed by iso-
lating the nonconforming item or condition so that it is not
used again until it is corrected, for example, labeling an
out-of-calibration instrument that cannot be used until it
meets calibration requirements.
Corrective action is defined as the measure(s) taken to cor-
rect conditions adverse to quality and, where necessary, pre-
vent recurrence (4). An example of a corrective action
would be to remeasure affected samples after the instrument
has met calibration requirements. Corrective actions are
documented with a technical justification that explains
why the corrective action to be implemented is the most log-
ical choice.
A QA record is a completed document that furnishes evi-
dence of the quality of items or activities (4) or study results.
Records are prepared, reviewed, and maintained to reflect the
processes resulting in the completed work. Examples of QA
records include plans and procedures used during the study
to direct work, raw and reduced data, calibration reports,
and chain-of-custody forms. Those specific documents and
data, and when they become QA records, are specified in pro-
cedures. Normally, documents and data become records when
personnel complete a segment of work and no longer need to
refer to the documents. Controls should be sufficient to
ensure that records are legible, accurate, complete, and secure.
Provisions should be made for storage, minimum retention
times, preservation, and disposition of records. This includes
provisions for proper storage of electronic data considering
the longevity of the various media. Controls must be in place
to ensure that records are traceable to samples, sites, data,
reports, and the study of origin.
Quality control (QC) is used to document system perform-
ance and reliability. Its use adds strength to a study by provid-
ing information to the data user of the conditions under
which the study was conducted; it can also show where there
are weaknesses in the study processes that must be overcome
to advance the investigation. Results of the QC determina-
tions must fall within certain limits of quality acceptance
for specific measurements. If QC measurements do not
meet acceptable limits, they must be flagged, and immediate
corrective action should be performed. QC results should be
presented as indicators of accuracy, precision, contamination,
and so on.
For QA/QC purposes, an assumption is made that the rel-
evant SOPs in the QA plan are followed and that the verify-
ing forms for the SOPs are properly and regularly completed
and kept as a record. The QA plan, SOPs, and forms should
have author, supervisor (e.g., laboratory supervisor), and QA
manager signatures and dates. In addition, these should be
reviewed at least on an annual basis and updated as needed.
Any new work in the laboratory should have related SOPs
and forms added to an amended QA plan.
It is important that the laboratory have a comprehensive
overall laboratory QA/QC plan because most methods do
not specify QA/QC requirements for the laboratory in which
the test will be conducted (i.e., its facilities and general work-
ing environment), its equipment, instruments, supplies, com-
mon reagents, or techniques (that are not specific to the test).
The SOPs and their accompanying monitoring and verifica-
tion forms should be a part of the QA plan.
The Occupational Safety and Health Administration
(OSHA), part of the Department of Labor, has requirements
for laboratory workers (5) in the United States. These regula-
tions are designed to keep laboratory workers safe from chem-
ical, biological, physical, and radioactive hazards, all of which
are applicable to the microbiological laboratory. OSHA has a
useful guidance (6) on laboratory safety that summarizes the
regulations. Some laboratories may not be required to follow
these regulations (e.g., public employees), but such laborato-
ries likely have state or local regulations that apply. Some of
the material from such regulations, as well as common sense
should and need to be included in the QA plan, especially
with regard to the following.
LABORATORY FACILITIES
Laboratories must be kept clean and safe. They must be
cleaned on a regular basis, and regular preventive measures
should be taken to keep the laboratory free of pests. Cleaning
should include regular and frequent trash removal, sweeping,
and mopping. In addition, less frequent but regular cleaning
(such as walls, windows) should be included. A list of the
cleaning supplies should be included in the form that records
how often this is done.
The QA plan should include a separate SOP for how to
decontaminate the laboratory to prevent contamination
and/or cross-contamination (since this is a special type of
cleaning). These procedures should be conducted prior to
when a sample arrives, during the analyses (e.g., see Quality
Control for Molecular Analyses), and when the sample is dis-
posed of. For samples that are considered biological or hazard-
ous waste, an SOP should be in place regarding proper
disposal.
The laboratory should also be comfortable and conducive
to work. It should be maintained at a comfortable temperature
with adequate lighting and air exchange (i.e., circulating
air). Consideration must be given to the kind and number
(typically 8–12/h) of air exchanges within the laboratory,
for example, whether the laboratory requires a no-return air
design (i.e., no air is returned back in the ventilation system
to be recirculated in the building).
Ideally, a laboratory should be designed in anticipation of
the kind of work that will be conducted. For example, a labo-
ratory working with pathogens should have negative air pres-
sure to help prevent the pathogens from escaping the lab. On
the other hand, a laboratory concerned with sterility, such as
a tissue culture laboratory, should have positive air pressure to
prevent contaminating air from entering. In smaller laborato-
ries, this may not be possible. Such laboratories should have
separate areas for reagent preparation, sample preparation,
and sample analysis to prevent cross-contamination among
these tasks; if this is not possible, then the laboratory must
have decontamination, cleaning, and QC measures to ensure

proper analyses. Although this is applicable to all microbio-
logical laboratories, it is especially acute for viral and molec-
ular testing.
Laboratories that analyze environmental samples should
be designed so that the sample passes through the laboratory
in a one-way path from sample arrival through analyses and
finally disposal. This is because the assumption is that the
sample is contaminated. If this is not possible, then the labo-
ratory should have sufficient decontamination protocols to
render multiple-use areas free from fomites, microbials, infec-
tious or potentially infectious material, DNA, or other con-
taminating substances.
LABORATORY PERSONNEL
Microbiologists conducting work in the laboratory need to be
trained not only in the technique being performed but also
with regard to the laboratory QA plan, relevant SOPs, and
the required monitoring forms. In addition, safety and hazard-
ous substance training (e.g., safe use of radioactive material,
chemical spill cleanup) should be completed prior to begin-
ning work in the lab. Supervisors should monitor analyst
training and adherence to laboratory protocols, identify
new training needs and the means to complete them.
LABORATORY EQUIPMENT
Each piece of equipment should have an SOP on proper oper-
ation, cleaning and maintenance. The instruction manuals
that came with the equipment should be kept in a designated
area. If calibration is required, that should also be included in
the SOP with a procedure if completed in the laboratory or an
indication that this task is performed under manufacturer
warranty or service contract. The purpose of an SOP, rather
than relying on the manufacturer instruction manual, is the
simplicity with which it should be written and the accompa-
nying form to exhibit compliance.
For equipment that is expected to maintain a specific tem-
perature (e.g., refrigerators, freezers, water baths, incubators),
daily monitoring should be done with the appropriate form
filled out. Examples of timetables and forms can be found
in various sources (7, 8).
For larger pieces of equipment that are connected to a ven-
tilation system (e.g., laminar flow or chemical hood) testing
and maintenance may be conducted by building facilities per-
sonnel unless a manufacturer warranty or service contract
applies. Both the Centers for Disease Control and Prevention
(9) and OSHA (5) have standards for this purpose.
TECHNIQUES/PROCESSES
Common techniques performed in the laboratory, for exam-
ple, autoclaving (either for decontamination or sterilization)
or media making, should have an SOP in the QA plan. The
SOPs should specify how to determine if these tasks have
been performed properly. For example, when autoclaving
for sterility, examination of sterility tape and the autoclave
log (to determine whether the correct pressure and time
were achieved and maintained), and the addition of aerobic
or anaerobic broth medium to bottles, tubes, or flasks, fol-
lowed by incubation and checking for growth should be com-
pleted. For media, culture with both a positive and negative
control should be conducted and appropriate records kept
for each batch of media. Other reagents, such as serological
or fluorescents should be verified and records kept of each
lot verification.
2.5.2. General Quality Control ▪ 2.5.2-3
SAMPLING PROCEDURES
If possible, validated or standard methods should be used to
collect samples. Make specific plans for collection, preserva-
tion, storage, record keeping, and transportation. Prior to col-
lection of samples, the following parameters need to be
determined: sample volume and type (grab or composite);
preservation techniques (e.g., use of formalin, azide), trans-
portation (e.g., overnight delivery), and storage (e.g., refriger-
ation) of samples; field equipment (e.g., pH meter, vacuum
pump); and forms and procedures to be used to record sample
collection, sampling conditions, and measurements. Holding
times, which are important in planning the collection of
microbiological samples, should be taken into consideration
when developing sampling plans to ensure that sample integ-
rity is maintained. The sampling plan should include a com-
bination of laboratory and field or travel blanks (travel blanks
carried or shipped with samples to detect contamination or
sample degradation during the trip), splits, duplicates, and
other QC sample types.
SAMPLE HANDLING
There should be a system in place for sample receipt, login,
and storage practices. All samples should be inspected for
containment integrity, temperature, and holding time status
on receipt. Any issues (e.g., shipment temperature, holding)
should be noted. There should be criteria established to allow
for the rejection of a sample. Chain-of-custody (COC) proce-
dures employ the use of signatures to record individuals
responsible for the sample throughout the collection and ana-
lytical process. The COC for each sample must be inspected
for accuracy against labels on and conditions of the sample
and sample containers. Any differences should be noted. In
addition documentation of the date, time, and persons col-
lecting and shipping samples should be checked. COC proce-
dures ensure sample protection by eliminating the possibility
of accidental or purposeful tampering and are necessary for
work related to litigation.
If a COC is not necessary, sample control to keep track of
sample identity, purpose, holding times, scheduling, analysis
types, and so on is helpful, especially when processing a large
number of samples, for more than a few sampling sites or sam-
pling teams, and for multiple studies.
ANALYTICAL PROCEDURES
SOPs should be developed for each analytical procedure
used in the laboratory. The SOPs should include method-
specific QC checks and any additional QC checks that are
required by the laboratory’s QA plan or certification body
(e.g., NELAC).
DEVIATIONS, RECORD KEEPING,
AND AUDITS
An SOP should be available to address nonconformance
events and deviations, whether procedural or process, planned
or unplanned. Both preventive and corrective actions should
be included. For prevention purposes, the SOP should be
updated based on experience within the laboratory. All such
deviations should be recorded and documented with regard
to what happened, why it happened, how it was corrected,
and how it will be prevented in future.
An SOP should be available on QA record keeping. It
should address how the records are stored, archived, and
2.5.2-4 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
retrieved; how long they are to be kept; and how to dispose of
them once they are no longer required.
For auditing purposes, many laboratories are required to
perform according to state, federal, or other guidelines, or
they may be certified by either these or other entities. Docu-
mentation may be required, and either planned or unplanned
audits may take place where staff will be required to demon-
strate knowledge and compliance with these guidelines.
The guidelines from such entities may require specifics that
are not necessarily within the QA plan (e.g., proof of U.S. cit-
izenship). Knowledge and adherence to all such specifics is
the responsibility of either the laboratory supervisor or the
QA manager, as specified in the auditing SOP.
SOPs and the QA plan should have signatures of the
author(s) and the approving officials (supervisor and QA
manager). This ensures both an agreement not only on the
content but an agreement to adhere to the content of these
documents.
REFERENCES
1. Cross-Smiecinski AJ, Stetzenbach LD. 1994. Quality
Planning for the Life Science Researcher. CRC Press, Boca
Raton, FL.
2. U.S. Environmental Protection Agency. 1992. Interim
draft: EPA requirements for quality management plans. U.S.
Environmental Protection Agency, Washington, DC.
3. Kanare HM. 1985. Writing the Laboratory Notebook. Ameri-
can Chemical Society, Washington, DC
4. ASME Nuclear Quality Assurance Committee. 1994.
Quality assurance requirements for nuclear facility applications,
vol. ASME NQA-1-1994. American Society of Mechanical
Engineers, New York, NY.
5. Code of Gederal Regulations. Part 1910. Occupational
Safety and Health Standards. Subpart Z. Toxic and Hazard-
ous Substances. Standard number 1910.1450. Occupational
exposure to hazardous chemicals in laboratories. http://www.
osha.gov/pls/oshaweb/owadisp.show_document?p_table=
standards&p_id=10106.
6. OSHA. 2011. Laboratory Safety Guidance, OSHA 3404-
11R. http://www.osha.gov/Publications/laboratory/OSHA3404
laboratory-safety-guidance.pdf.
7. Clesceri L, Greenberg AE, Eaton AD (eds). 2012. Standard
Methods for the Examination of Water and Wastewater, 22nd
ed. American Public Health Association, American Water
Works Association and Water Environment Federation,
Washington DC.
8. U.S. Environmental Protection Agency. 1978. Microbio-
logical Methods for Monitoring the Environment. EPA-
600/8-78-017 http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=
300014TD.txt.
9. CDC. 2009. Biosafety in Microbiological and Biomedical Labo-
ratories, 5th ed. HHS Publication No. (CDC) 21–112. U.S.
Department of Health and Human Services, Washington,
DC.
 Quality Control for Bacteriological Analyses
ELLEN BRAUN-HOWLAND
2.5.3
QUALITY CONTROL FOR
BACTERIOLOGICAL ANALYSES
Bacteriological analyses in environmental biology laborato-
ries differ from those performed in their clinical counterparts
in that quantitative results are commonly required. Fre-
quently, environmental samples are collected and analyzed
over long periods of time to determine, for example, temporal
and/or spatial effects of one or more environmental variables
on bacterial populations. Analysis of samples over extended
time periods points to the importance of stringent quality
control of experimental materials involved in the isolation
of environmentally derived organisms, so that recoveries are
consistent and valid comparisons can be made over time.
Quality control practices in many public health laboratories
performing bacteriological analyses for regulatory purposes
are guided by requirements set forth by the U.S. Environmen-
tal Protection Agency (1) or national accreditation organiza-
tions such as the NELAC Institute (2). Detailed guidelines for
essential quality control procedures applicable to analysis of
water samples can be found in Standard Methods for the Anal-
ysis of Water and Wastewater (3). The intent of this section
is to outline procedures designed to permit the production
of quantitatively consistent results in laboratories analyzing
bacteriological samples derived from the environment. For
detailed procedures, the reader is referred to the aforemen-
tioned normative references.
MEDIA COMPONENTS
To enhance consistency of results (both experimental and
quantitative recovery), commercially available media gener-
ally provides more consistency than media made from indi-
vidual ingredients. For this reason, they should be used
whenever possible. Purchased media are prepared according
to strict quality control guidelines and are pretested for recov-
ery of target organisms and inhibition of undesirable organ-
isms when applicable. Documentation of quality control
practices is available for each product. Although commercial
preparations generally yield more consistent results, it should
be noted that bacterial growth may be inconsistent between
manufacturers and even between different lots of media. Like-
wise, when media is prepared from individual ingredients,
consistency in recovery is best achieved when the same lot
numbers of ingredients are used. Quantitative recovery of
some organisms or experimental results can be influenced
doi:10.1128/9781555818821.ch2.5.3
2.5.3-1
by individual lot numbers of complex components, such as
yeast extract or tryptone (4, 5). All medium components
should be of high quality, preferably ASTM grade or better.
While some components of individual media, such as salts
or detergents, are stable for lengthier periods of time, the qual-
ity of dehydrated media is best maintained by storage in a
dark, cool, dry place under low humidity. Ideally, storage con-
ditions are monitored and recorded. Dehydrated media
should be discarded after the expiration date on the label or
six months after the container is opened. If this is not
economically feasible, at a minimum, performance must be
quantitatively verified to ensure comparable results over
time. Dehydrated media that have changed in appearance,
including caking or discoloring, should not be relied on for
the consistent recovery of organisms. Manufacturers’ instruc-
tions for individual medium components requiring more
stringent storage conditions, such as vitamins, antibiotics,
or fluorochrome-containing components, should always be
followed to ensure that the medium performs as desired.
The use of high-quality water in the preparation of bacter-
iological medium, particularly for use in the preparation of
sample diluents, is critical. Deionized or distilled water should
be evaluated for minimal contamination with metals and
other growth-inhibiting substances. Conductivity, pH, and
presence of growth-promoting compounds should be regu-
larly monitored. The production of growth-promoting com-
pounds can occur upon storage of distilled water as a result
of cryptic growth of contaminating bacteria or the result of
contaminating phototrophs if water is stored under lighted
conditions. To minimize the effect of heterotrophic contam-
ination, water used in media preparation should be quantita-
tively analyzed on a monthly basis. An increase in numbers
over time points to the need for decontamination of the stor-
age system. At a minimum, the system should be drained and
removable parts, such as Tygon tubing, should be replaced to
remove biofilms. Guidelines pertaining to quality of reagent
water can be found in Table 9020:II of Standard Method
for the Examination of Water and Wastewater (3).
MEDIA PREPARATION
Dehydrated media or individual medium components
should be weighed using a balance calibrated with weights
traceable to NIST or another national standard. Calibration
of the balance should be performed annually by a certified
2.5.3-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
company. To achieve traceability, the balance used for
medium preparation as well as the weights used for calibra-
tion should be recorded. Glassware should be type I borosi-
licate and if used for measuring volumes should be type I,
class A. Although documentation of measurement accuracy
is available for type A glassware, it is good practice to verify
a certain percentage of glassware for measurement accuracy
on receipt.
Although some organisms can survive and grow in dis-
tilled water, medium pH is critical to the successful isolation
of others, such as Legionella pneumophila. For this reason, and
to promote consistency of results over time, medium pH
should be verified after preparation of each batch; if using
agar plates, a surface pH meter probe should be used. If one
is not available, agar can be broken up in distilled water and
the pH of the resultant overlay can be measured. Medium
pH is typically measured at room temperature; sterilized agar
can be held in a tempering bath briefly (up to 3 h) until the
pH of molten agar is verified so that it can be adjusted, with
filter-sterilized reagents, as needed.
The pH meter should be calibrated before each use at two
pH values (two-point calibration) bracketing the desired pH
(e.g., 4 and 7 or 7 and 10). pH buffers traceable to NIST or
another national standard-setting organization should be
used and discarded after the expiration date on the label.
Ideally, aliquots of reference buffers are discarded after use,
and a fresh aliquot is poured for each measurement. In gen-
eral, probes for pH meters should be maintained in liquid hav-
ing an acidic pH and never stored in deionized water, as it may
result in diffusion of the ions from the internal KCl solution
into the ion-depleted water. The slope of a pH meter should
be checked and recorded monthly to ensure proper function-
ing of the pH probe. If the slope is less than 95% or greater
than 105%, the easiest correction is to replace the standard
buffers. If an acceptable slope is not achieved after replacing
the buffers, other corrective actions may include replacement
of the internal KCl solution or cleaning of the pH probe,
according to manufacturer’s instructions.
Some medium components are heat labile and must be
filter-sterilized and added after autoclaving, for example,
antibiotics, fluorochrome-containing components, and vita-
mins. These ingredients should be added to the medium after
tempering in a water bath to approximately 45°C. Record
lots of each component so that if organism recovery or appear-
ance changes, inconsistency in medium components can be
ruled out.
STERILIZATION
Media, glassware, and measuring devices that come into con-
tact with the sample are typically sterilized by autoclaving,
although flame sterilization of hockey sticks is permitted.
Autoclaves sterilize materials by exposing their contents
to saturated steam (121°C) under high pressure (15 lb/in2).
The time required to sterilize materials or solutions is volume-
and material-dependent, that is, more time is required to ster-
ilize 1 liter of solution than 100 ml. Bacteriological media
containing carbohydrates are typically sterilized 12–15 min
to avoid caramelization. Liquids are best autoclaved in con-
tainers that hold at least twice their volume to avoid boiling
over. When used, caps must be loosely tightened to avoid
explosion due to increased temperature and permit the
entrance of steam into the flask. For long-term storage, caps
should be closed finger-tight only after the solution has
cooled, or a vacuum will form in the flask/tube. For dry mate-
rials, which can be exhausted rapidly, the addition of a little
deionized water to loosely capped sample bottles will aid in
the sterilization process.
Autoclaves are typically maintained by professional staff
who have undergone specific training, although there are
some routine maintenance items that should be addressed.
Temperature must be verified annually using a traceable
maximum registering thermometer. Drainage screens must
be kept free of debris, and the door gasket should be kept
clean. Timers should be verified for accuracy against a NIST-
traceable timepiece on a quarterly basis. Sterilization is indi-
cated by autoclave tape, which contains a chemical indicator
that is visible after maintenance at 121°C, 15 lb/in2 for 15
min. Proper autoclave functioning is also ascertained through
the use of uninoculated medium controls, whereas steriliza-
tion of glassware or nontoxic solutions can be checked by
adding nonselective medium such as nutrient broth or brain
heart infusion directly to the flask or into the solution
and incubating a sufficient amount of time at a growth-
appropriate temperature. Commercially available biological
indicators can be used to demonstrate that an autoclave suc-
cessfully kills thermophilic bacterial spores at a frequency of
once a month. Chemical tapes (e.g., steam integrators) are
also commonly used in each batch of material autoclaved.
MEDIA STORAGE
Accreditation authorities provide guidelines for public health
laboratories performing bacteriological analysis on environ-
mental samples regarding storage of liquid and solid media
used for purposes of regulatory compliance (1–3). These indi-
cate that liquid media can be stored for up to 3 months refri-
gerated, if it is tightly capped. However, other sources indicate
that nonselective media can be stored in tightly capped tubes
for up to 6 months. Prepoured agar plates such as endo-type
medium have a holding time of 2 weeks if stored refrigerated
in the dark and in plastic bags or containers. Refrigerator and
freezer temperatures should be monitored, and the results
recorded, preferably twice a day. In general, common sense
should prevail: any change in the appearance of media, espe-
cially dehydration, indicates that it may not yield results con-
sistent with those obtained using freshly made medium.
Given the expense and time dedicated to sample collection,
together with the importance of repeatable quantifiable
experimental results, acceptable storage of media is highly
recommended.
INCUBATION CONDITIONS
Incubation temperature is critical to the growth rate, and
therefore quantitation, of bacteria. To verify the accuracy
of temperature measuring devices, mercury or alcohol-
containing thermometers should be calibrated at least annu-
ally against a NIST-traceable standard, while digital and
infrared thermometers should be calibrated quarterly. If there
are any temperature corrections, the measuring device should
be clearly marked with the correction factor and the corrected
temperature should be recorded. Incubator temperatures
should be monitored twice a day, once in the morning and
again in the afternoon, and the results recorded in a log book.
Thermometer gradations should reflect acceptable toler-
ances for use; for example, a thermometer having gradations
of 0.5°C would not be suitable for measuring incubation
conditions of 44.5 ± 0.2°C. Thermometers varying by a
degree or more are not reliable and should be removed from
use. To maintain temperature stability and avoid measure-
ment of transient fluctuations, thermometers used in freezers
 2.5.3. Quality Control for Bacteriological Analyses ▪ 2.5.3-3
are typically immersed in glycerol, whereas glycerol or water
solutions suffice for temperature measurement of refrigerators
or air incubators. Placement of thermometers in incubators
should reflect the areas of most common usage; in large,
walk-in type incubators or refrigerators, two or more ther-
mometers may be required. A circulating water bath will pro-
vide the most efficient heat transfer and achieve the desired
incubation temperature in the least amount of time. If an
air incubator is used, several hours may be required before
the medium reaches the desired temperature, particularly in
small units. Also, the addition of a large number of plates
or cold liquid media can change the internal temperature of
an incubator. For critical experiments, prewarming or cooling
of the medium in a water bath before placement in a dry incu-
bator will facilitate the required temperature change.
If several agar plates are to be incubated, they should be
inverted and stacked no more than four high to maintain con-
sistency in temperature between plates. They should be incu-
bated in plastic bags or tightly covered containers with a
moistened paper towel to minimize loss of liquid during incu-
bation and be situated in the incubator such that they do not
block air flow.
If organisms require aeration for growth or if experimen-
tal conditions require use of a rotator, quarterly calibration
against a NIST-traceable standard (e-clock) will ensure con-
sistency between experiments. Organisms requiring CO2
(reduced oxygen) should be grown in an incubator with a
gas gauge equipped with an alarm.
ANALYTICAL AND MEDIA CONTROLS
The importance of quantitative analysis of culture controls to
verify acceptable performance of bacteriological media can-
not be overemphasized. For this purpose, each batch of freshly
prepared medium should be inoculated with known numbers
of positive and negative control organisms, preferably before
use of the media. If this isn’t possible, then control cultures
should be analyzed when samples are first analyzed—with
the caveat that if medium quality is not acceptable, the exper-
imental results have no value. The organisms chosen for use as
controls must be appropriate for the assay. For example, ana-
lytical quality control of media used in water quality laborato-
ries typically makes use of E. coli, Enterobacter aerogenes or
other coliform, and a coliform negative organism, such as
Pseudomonas sp., to ascertain acceptable medium perform-
ance (see Table 9020:V, Standard Methods for the Examination
of Water and Wastewater). Typically, culture controls are per-
formed in at least duplicate, and quantitative results are com-
pared to the number of expected colonies. This can be
ascertained through ongoing quality control on media,
requiring statistical comparison of target organism recoveries
between batches (see use test; Standard Methods). For optimal
consistency, and if finances permit, quality control products
containing a known number of organisms are commercially
available. In addition to culture controls, a certain percentage
of tubes, flasks, or plates from each batch of medium should be
incubated in the absence of inoculation with bacterial
cultures to demonstrate sterility of the medium. If positive
or negative control cultures do not perform in an acceptable
manner, then the medium is not reliable for use in experi-
ments and should be discarded.
When using pour plates, the final plate from each bottle of
medium should remain uninoculated to demonstrate that the
medium was sterile prior to use. If the final poured plate shows
growth, then the results obtained using that bottle must be
discarded. Additional quality control includes the use of dilu-
tion blanks for membrane filtrations, which serve to demon-
strate the absence of contamination of filtration funnels,
filters, and diluents/rinse water. The introduction of extrane-
ous bacteria into an environmental sample or pure culture is
avoided by using aseptic technique.
METHOD-SPECIFIC QUALITY
CONTROL CHECKS
Some EPA 1600 series bacteriological methods (e.g., Meth-
ods 1600, 1603) include initial precision and recovery (IPR),
ongoing precision and recovery (OPR), and matrix spike sam-
ple analyses. For IPR and OPR analyses a reference matrix
(e.g., phosphate-buffered saline) is spiked with a known con-
centration of the target organism. IPR samples allow the ana-
lyst to demonstrate proficiency with the analytical technique,
while OPR analyses are conducted to demonstrate ongoing
proficiency. OPR samples may also serve as positive controls.
Matrix spike samples are conducted to determine the effect of
a particular matrix on recoveries. Matrix samples are spiked
with a known concentration of the target organism and ana-
lyzed according to the method.
Some methods and procedures may require that a per-
centage of positive and negative results (e.g., colonies, tubes)
be verified using an alternate procedure or biochemical
confirmation.
REFERENCES
1. U.S. Environmental Protection Agency. 2005. Manual
for the Certification of Laboratories Analyzing Drinking
Water, 5th ed, EPA-815-R-05-004. Office of Ground
Water and Drinking Water, Technical Support Center,
Cincinnati, OH.
2. National Environmental Laboratory Accreditation Con-
ference. 2003. 2003 NELAC Standard. EPA/600/R-04/003.
Washington, DC.
3. Clesceri L, Greenberg AE, Eaton AD (eds). 2012. Standard
methods for the examination of water and wastewater, 22nd
ed. American Public Health Association, American Water
Works Association and Water Environment Federation,
Washington, DC.
4. De Spiegeleer P, Sermon J, Lietaert A, Aertsen A, Michiels
CW. 2004. Source of tryptone in growth medium affects
oxidative stress resistance in Escherichia coli. J Appl Microbiol
97:124–133.
5. Nyberg R. 2000. Biological indicators and population veri-
fication. Inf Control Today January.
 Quality Control for Virological Analyses
RICHARD E. DANIELSON
2.5.4
QUALITY CONTROL FOR VIROLOGICAL CELL
CULTURE ANALYSES
Viruses are obligate intracellular parasites and therefore, to
cultivate viruses, they must be propagated in a host cell line
(1, 2). Certainly not all viruses that are the cause of human
disease are culturable in situ, but for those that are culturable
there are unique challenges for the laboratory technician in
propagating and maintaining the cell lines.
Since culturable viruses are perpetuated in living cell lines,
there are a wide variety of variables that can affect the outcome
of the virus–host interaction. Furthermore, such assays are
expensive not only in media, reagents, and disposables but
also in the time it takes a skilled analyst to propagate and pre-
pare the cell line, extract viruses from the sample, apply the
extract to the cell line, and diagnose the outcome of the assay.
Therefore, maintaining a strict quality assurance (QA) pro-
gram around these efforts is essential to maximize the perform-
ance of the cell line and limit the negative effect of variables
that can directly or indirectly affect test outcomes (2, 3).
Many activities associated with environmental virology
can be research-oriented. Therefore, while quality control
activities may have associated detailed procedures and guide-
lines, it may be necessary to build in flexibility to address iter-
ative modifications and refinements should they be necessary
and appropriate during the course of execution of a study.
TRAINING
It is difficult to find specific training courses in cell culture
maintenance as it is an expensive proposition and rarely prac-
ticed commercially. Typically, training in cell culture occurs
at the college or university environment. Individuals should
be trained and supervised by someone with proven expertise
(2, 4). Trainees (at any level of education) can be evaluated
based on their ability to prepare media and reagents, prepara-
tion of dilutions, direct observation of their technique in han-
dling the harvesting and distribution of cells, and the
outcome of assays (as appropriate).
Inclusive of training is the initial demonstration of capa-
bility (DOC) (2). This can be accomplished by devising a
program within the laboratory whereby trainees perform the
methods with quality control (QC) tests for cell line propaga-
tion and/or sample analysis. For example, the U.S. EPA rec-
ommends that at least seven proficiency samples be used to
determine the initial DOC (2). For someone who is training
doi:10.1128/9781555818821.ch2.5.4
2.5.4-1
in cell culture and assay, this can include a positive control
(laboratory fortified blank) and negative control (laboratory
reagent blank) along with five samples spiked with various
concentrations of the target virus. It would be best if the con-
centration of virus in these samples spanned the predicted
range of assay sensitivity.
Acceptance criteria can be based on the known concen-
tration of added spike and/or compared to the performance
of the method by others in the laboratory (3, 5). For someone
who is training in cell culture maintenance only, the labora-
tory can set a series of criteria for assessing successful perform-
ance of the procedure. These criteria may include correct
calculations in media and reagent preparation, the use of
aseptic technique, cell line manipulation technique, harvest-
ing and enumerating cell concentration, detection of con-
tamination of cell cultures, accuracy in documentation, and
the overall quality of outcomes of assays associated with the
prepared cell line.
It is equally important to maintain an ongoing DOC pro-
gram where the foregoing criteria can be assessed on a regular
basis (3, 5). Depending on the frequency of samples proc-
essed, this can range from monthly QC sets to batch-related
QC samples. Any analyst/technician who does not meet
the ongoing DOC must stop working on samples until
the cause of the failure has been identified and corrected
(2, 3, 5). Keep up-to-date records on all training, initial
DOC, and ongoing DOC on all personnel associated with
sample process and analysis. Data quality is addressed, in
part, by consistent performance of valid procedures docu-
mented in referenced analytical methods and laboratory pro-
cedures. The training and experience of project staff and
documentation of laboratory activities enhance data quality.
FACILITY DESIGN
There should be separate areas for the preparation of reagents,
the preparation of samples, and analysis (1, 2, 4). Virus extrac-
tion from filters and organic flocculation can be carried out in
a central sample processing area away from cell culture activ-
ities. For virus analysis by tissue culture, there should be dedi-
cated hoods for the preparation of cell lines and hoods for the
process of samples. In addition, to control unintended con-
tamination, there should be controlled movement of person-
nel into some of the rooms for some of the activities associated
with sample preparation and analysis (2, 4). Waste from cell
2.5.4-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
culture activities must be treated accordingly to local, state,
and/or federal regulations.
A reagent-grade water system is essential for maintaining
the purity of reagents and media (1, 6, 7). A definition of
what constitutes reagent-grade water can be found in sources
such as the U.S. Pharmacopeia, Standard Methods for the
Evaluation of Water and Wastewater, or the National Com-
mittee for Clinical Laboratory Standards. Water quality
should be checked by a certified analytical laboratory at least
annually (6, 7).
REAGENTS AND SUPPLIES
The goal is to have a clear association of all virological data
with specific lots of all materials employed in performing
analyses. Supplies and consumables are those items necessary
to support the sampling and analytical operation, including
but not limited to sample containers, filters, calibration solu-
tions, decontamination supplies, preservatives, and various
types of growth media (1–6). All supplies and consumables
ordered should be stored in appropriate conditions for the
item(s). Materials may have to meet specific requirements
as indicated by the appropriate manufacturer. For example,
only certified standard solutions within expiration time
should be used for pH calibration (6, 7). Additional quality
control items such as serologic pipettes, cell culture flasks,
and sample containers will include testing for sterility (2,
4, 6). Generally, a nonselective media (including antibiotic-
free cell culture media) can be used for this purpose. Maintain
records of all testing outcomes related to the various lot
number of material received. In-house-prepared media should
be routinely tested for sterility.
Each lot of culture media and supplements must be tested
to confirm that growth of the cell line and propagation of
viruses in the various assay formats are not negatively
affected (1, 2). Media supplements that can impact cell
growth and virus propagation are serum and/or semisynthetic
serum surrogates (1, 2). It may be necessary to order at least
two lots of serum per serum provider for evaluation. Test
each new lot of serum with all cell lines, as appropriate to
evaluate overall growth characteristics (e.g., development
of a monolayer, physical condition of the cells). It is just
as important to spike each lot of serum with the various virus
stock cultures to confirm that it is not inhibitory to virus
infection/replication and allows for the production of a suf-
ficient virus titer. It is important that this evaluation be con-
ducted well in advance of the exhaustion of the current lot
of media and supplements in use for comparison. This must
be planned carefully as evaluation of the effect on cell lines
can take two to three weeks. Testing for cell line sensitivity
to viruses and evaluation of the production of viruses can
also take two to three weeks. Overall this process may take
four to six weeks and consume laboratory resources. Once
the lot(s) of media and supplements has been identified,
order as much of the material as necessary to last as long
as possible under the proper storage conditions to delay
repeating this evaluation process.
EQUIPMENT
A list of equipment used in the laboratory should be generated
along with a maintenance schedule (2). Equipment used in
measuring (e.g., thermometers, micropipettors, balances)
need to be calibrated on a regular basis. While weights, hoods,
and microscopes can be calibrated and serviced annually,
micropipettors should be calibrated quarterly (6, 7). Self-
calibrate only if the laboratory is qualified and certified to
do so, otherwise use a certified third-party vendor (7).
SANITIZATION
In several steps of sample concentration, reusable equipment
such as filter housings, hoses, pumps, and pH probes will
require sanitizing. A 0.525% solution of sodium hypochlorite
can be prepared from germicidal household bleach (1:10 dilu-
tion). It is important that any application of a sanitizing sol-
ution must be followed by neutralization. For sodium
hypochlorite, prepare and apply a 1 M solution of sodium thi-
osulfate (1, 2, 8).
SAMPLE HANDLING PROCEDURES
There should be a system in place for sample receipt, log-in,
and storage practices (2, 7). All samples should be inspected
for containment integrity, temperature, and hold time status
on receipt. If there is a compromised sample container, or if
there are any deviations from prescribed shipment tempera-
ture and/or hold time, the client or source of the sample
should be notified immediately (7). There should be criteria
established to allow for the rejection of a sample. As was pre-
viously discussed, cell culture assays are very expensive, and if
the sample itself is not in acceptable condition, there is no
point in expending the effort in the process. Once samples
have been accepted into the laboratory, it is important that
they are stored separately from reagents and media (7).
ANALYTICAL PROCEDURES
The laboratory must have current copies of standard operating
procedures (SOPs) for all procedures available to personnel
(6, 7). It is essential that personnel follow SOPs to ensure
reproducibility and limit the variability inherent in microbio-
logical analyses. If using published methods, it is important to
write the procedure as they are performed in the laboratory to
account for the type of equipment, reagents, and media used.
SOPs should be read through and updated at least annually if
no other changes to the procedure have been made.
Everyday bench sheets are an essential mechanism for cap-
turing activity, information, and data collected. Bench sheets
must also be reviewed at least annually and the most current
version distributed to the staff.
QC CHECKS
A percent of the total analysis time should be dedicated to
quality control, including analyses of blanks and standards
(2, 3, 5–7). Analytical precision can be determined by multi-
ple analyses on a single sample (6). Accuracy can be calcu-
lated from internal standards, standard additions, or
external standards (6). Internal QC can include operator
training, recovery of known additions, evaluation of matrix
effects, analysis of standards and blanks, and routine equip-
ment checks.
QC requirements can include but are not limited to the
analysis of positive and negative controls, reagent blank
checks, sterility checks, spiked reference materials, and spiked
matrix samples (1, 2).
A QC sample set should accompany an analysis batch (2).
An analysis batch should be a group of samples prepared and
assayed with the same lots of media and reagents. Since some
of the reagents and media prepared for use may have only a
one- or two-week shelf-life, the QC sample set would be
applied to those samples processed during that time period.

The negative QC and equipment blank check is sterile
reagent-grade water processed through the entire protocol
and assay (1, 2). The positive QC sample is reagent-grade
water spiked with a known virus target. Only some viruses
can be enumerated, so in those cases use a concentration of
viruses that is likely to provide a positive outcome given the
expected recovery of the method. In those cases where there
are published expected performance criteria, the QC set shall
meet those standards. A positive result for the negative QC
sample indicates a failure to the batch of samples processed
associated with that QC set. Ongoing evaluation of each
media component should help identify the source(s) of the
contamination of a negative QC sample. A result outside
the acceptable limits in the positive control also constitutes
a batch failure. However, one practice is to use a rolling aver-
age of six QC sample batches to determine pass or fail (2).
There can be several reasons for inhibition; one can be resid-
ual disinfectant on the equipment used in the filtration con-
centration steps. It is absolutely vital that if a sanitizing
disinfectant is employed in preparing reusable equipment
that the proper neutralizer is applied (2, 8)
If all the QC measures are in place, then overall vari-
ability can be reduced, but the primary contribution of
variability with environmental sample analysis is the effect
of the matrix on process and assay. Therefore, matrix spikes
are vital when assessing recovery efficiency from samples (2).
As with the positive QC set control, a sufficient quantity of
target virus must be added to the first steps of the process. For
the preparation of relatively small sample sizes (e.g., 1 liter
for raw wastewater) the target can be added directly to the
sample container just prior to the organic flocculation step.
However, for other environmental sample sources, such as
secondary treated effluent, surface water, and groundwater,
larger volumes (up to 1,800 liters) must be filter concen-
trated. It may not be convenient to collect these large vol-
umes prior to filtration. Have the field personnel collect at
least an additional 10 liter sample to accompany the filter
(2, 8). The 10 liter portion can then be spiked and processed
through the filter. In some cases there may be methods with
published acceptable matrix spike criteria (EPA Method
1615). It is important to note and account for these results
in a final report.
Although rare, sometimes there are sources available for
proficiency evaluation studies provided by independent third
parties. Whenever possible, the laboratory should participate
in such studies as a check on overall laboratory performance.
Failure to meet any QC requirements requires that appro-
priate corrective actions be taken (2, 3, 5). The essential steps
in the corrective action process are as follows: investigate and
determine the cause of the problem, identify appropriate cor-
rective action, establish the effectiveness of and implement
the corrective action, verify that the corrective action has
eliminated the problem.
2.5.4. Quality Control for Virological Analyses ▪ 2.5.4-3
DATA REDUCTION, VERIFICATION,
AND REPORTING
Any conversions, transformations, or other calculations
required for reducing raw data collected by technicians or ana-
lysts will be noted in the relevant SOPs (2, 3, 5). Laboratory
personnel should be assigned primary responsibility for one or
more well-defined activities. Likewise, each individual will
also be responsible for any preliminary reduction of data
related to that activity.
A technical lead will be responsible for training staff as
appropriate in the statistical analysis needed for reporting pur-
poses. It is also anticipated that at least some statistical data
reduction is expected to be done by computer. For example,
most probable number indices that are the result of virus
assays can be determined by the U.S. EPA computer program,
Total Culturable Virus MPNV (2).
REFERENCES
1. U.S. Environmental Protection Agency. 1996. ICR Micro-
bial Laboratory Manual. EPA 600-R-95-178. Office of
Research and Development, Washington, DC.
2. U.S. Environmental Protection Agency. 2012. Method
1615. Measurement of Enterovirus and Norovirus Occur-
rence in Water by Culture and RT-qPCR. EPA 600-R-
10-181. Office of Research and Development, Washington,
DC.
3. U.S. Environmental Protection Agency. 1998. Guidance
for Quality Assurance Project Plans. EPA QA/G-5. EPA/
240/R-02/009. Office of Environmental Information, Wash-
ington, DC.
4. U.S. Environmental Protection Agency. 2004. Quality
Assurance/Quality Control Guidance for Laboratories Per-
forming PCR Analysis on Environmental Samples. EPA
815-B-04-001. Office of Water, Cincinnati, OH.
5. U.S. Environmental Protection Agency. 2001. Require-
ments for Quality Assurance Project Plans. EPA QA/R-5.
EPA 240-B-01-003. Office of Environmental Information,
Washington, DC.
6. Clesceri L, Greenberg AE, Eaton AD. (eds). Standard meth-
ods for the examination of water and wastewater, 22nd ed.
American Public Health Association, American Water
Works Association and Water Environment Federation,
Washington, DC.
7. U.S. Environmental Protection Agency. 2005. Manual for
the Certification of Laboratories Analyzing Drinking Water
Criteria and Procedures Quality Assurance. 5th ed. EPA
815-R-05-004. Office of Ground Water and Drinking
Water, Technical Support Center, Cincinnati, OH.
8. U.S. Environmental Protection Agency. 1995. Informa-
tion Collection Requirements Rule—Protozoa and Enteric
Viruses Sample Collection Procedures. EPA 814-B-
95-001. Office of Ground Water and Drinking Water, Wash-
ington, DC.
 Quality Control for USEPA Method 1623
Protozoan Analysis and PCR Analyses
GEORGE D. DI GIOVANNI AND GREGORY D. STURBAUM
2.5.5
QUALITY CONTROL FOR PROTOZOAN
ANALYSES
The following subsections of this chapter present quality
assurance (QA) procedures, including practical advice, for
laboratories performing U.S. EPA Method 1623 Crypto-
sporidium and Giardia protozoan analysis. As outlined in the
previous subsections, a comprehensive QA program is a
necessity for an environmental laboratory to accurately and
confidently report both presence/absence and quantifiable
results. A QA program outlines the components of a labora-
tory testing scheme that should be monitored, whether it be
daily, weekly, monthly and yearly, while the quality control
(QC) checks are put into place to ensure the discrete testing
method components used throughout the testing protocol
contribute minimal amount of error to the results. This out-
lined section is designed to give an overview of general labo-
ratory QC practices for microscopic detection of microbial
organisms.
EPA Method 1623 (and the variant Method 1623.1) is
used for the detection and enumeration of Cryptosporidium
oocysts and Giardia cysts in water (1, 2). Samples are filtered;
retained particles are then eluted from the filter by agitation
and washed with an aqueous buffered salt and detergent
solution. (Oo)cysts are further concentrated by centrifuga-
tion, and immunomagnetic separation (IMS) is used to
recover the organisms from the packed organic and inorganic
debris pellet. Dissociation of captured (oo)cysts from the
magnetic beads is achieved by adding acid and mixing by
vortexing. After dissociation, the IMS beads are retained
in the sample tube using a magnet and the supernatant con-
taining the (oo)cysts is transferred to a microscope slide. The
pH of the supernatant is neutralized and allowed to dry.
Microscopy is used to enumerate samples for (oo)cysts,
enabling the number per volume to be determined. To
view the (oo)cysts in a sample, they must be stained using
fluorescent-labeled antibodies. In addition, a second fluores-
cent stain, 40 , 60 -diamidino-2-phenylindole (DAPI), is
added to stain the nucleic acid present in the nuclei of
(oo)cysts. Samples are examined using epifluorescence
microscopy to visualize fluorescently stained (oo)cysts and
nuclei (if present). Confirmation of (oo)cysts requires visual-
izing internal structures (e.g., sporozoites, nuclei, median
bodies, and axonemes) by differential interference contrast
microscopy (DIC).
doi:10.1128/9781555818821.ch2.5.5
2.5.5-1
QC CHECKS
QA and QC parameters used to measure and determine
method performance, include ongoing precision recovery
(OPR) trials, method blanks (MBs), matrix spikes (MSs),
microscope interanalyst verification, and participation in a
proficiency testing program. Although the QC checks
described below are specifically for Method 1623/23.1, simi-
lar checks could be applied to other protozoa methods that
include microscopy.
• MB samples serve as negative controls and are processed
with reagent-grade water (i.e., distilled or deionized water)
at an equal volume at which the majority of the routine
samples are processed (e.g., 10 liters). MBs are control
checks to determine that equipment, materials, reagents,
and technique are free of contaminating (oo)cysts. At
least 1 MB sample is analyzed with each weekly batch of
field samples, and if more than 20 field samples are ana-
lyzed per week, then 1 MB per 20 samples is analyzed. If
(oo)cysts are detected in the MB sample, processing of
field samples is discontinued until the source of the con-
tamination is located and corrected.
• OPR samples serve as positive controls and are processed
with reagent grade water at an equal volume at which
the majority of the routine samples are processed (e.g.,
10 liters). OPRs are seeded with spiking suspensions
with precise numbers of Cryptosporidium oocysts and Giar-
dia cysts prepared using flow cytometry. OPRs are control
checks to determine that equipment, materials, reagents,
and technique are operating adequately to recover (oo)
cysts. At least 1 OPR sample is analyzed with each weekly
batch of field samples, and if more than 20 field samples
are analyzed per week, then 1 OPR per 20 samples (or por-
tion thereafter) is analyzed. As defined in Method 1623.1,
the acceptable recovery rate for Cryptosporidium oocysts is
33–100% and 22–100% for Giardia cysts. If (oo)cysts are
not detected or are not detected within the defined recov-
ery rates, processing of field samples is discontinued until
the reason for not recovering (oo)cysts is determined and
corrected. The percent recovery results of OPR samples
are to be plotted on control charts. Control charts will
help assist in identifying negative trends in percent recov-
ery and remedial action is to be taken immediately. For
example, if the spiking suspensions are suspected, they
2.5.5-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
can be validated by enumerating one vial from each batch
supplied. Enumeration is conducted by filtering the spik-
ing suspension (usually 1 ml) through a black carbonate
membrane followed by staining with fluorescein isothio-
cyanate–labeled monoclonal antibody and enumeration
by microscopy.
• MS samples measure method performance for each water
sample matrix by including organism spiking suspensions
as outlined for OPR samples. A duplicate field sample
of equal volume is spiked and processed alongside the
unspiked field sample. Both samples are examined and
total organism counts are recorded. The indigenous
organism count is subtracted from the MS sample organ-
ism count and the result expressed as the percent recovery
of seeded organisms.
• Analyst verification of microscopic enumeration and
characterization of (oo)cysts should be conducted monthly
to demonstrate accuracy and comparability of microscopy
skills. Each laboratory analyst enumerates the total num-
ber of (oo)cysts on the microscope slide, and counts
must be within ±10% of the median value. As a group,
analysts should characterize 10 oocysts and 10 cysts using
DAPI and DIC criteria and discuss if necessary to ensure
consistency. If an individual analyst is found to be outside
the required count, action is to be taken immediately to
identify needs for additional training.
• Laboratories conducting regular analyses of water samples
should participate in an external laboratory proficiency
testing (PT) program. PT programs can either evaluate
(a) the entire sample process protocol (filtration/concen-
tration/IMS/microscopy) by supplying filters previously
seeded with target organisms or flow cytometry spiking
solutions which the laboratory seeds a standardized sample
as it would with an OPR, or (b) supply prepared micro-
scope slides with an appropriate number of target organ-
isms for enumeration. If the entire sample processing
protocol is to be conducted, then the PT scheme organizer
should vary the sample matrix and number of spiked
organisms (including blanks) that will challenge the lab-
oratory’s ability to recover (oo)cysts. Participation in a PT
program not only demonstrates the ability of the labora-
tory to adequately detect the target organism(s), it helps
maintain and develop the skills of analysts.
QC FOR PCR ANALYSES
Since its advent in 1985, the PCR has become one of the
most widely used molecular techniques for conducting bio-
logical research and is increasingly being used for the detec-
tion and enumeration of microorganisms in environmental
samples. Capable of detecting as little as a single copy of tar-
geted genetic material, the detection limit of PCR methods is
equivalent to, and in many cases greater than, traditional
culture-based microbiological methods. In addition, PCR
methods are able to detect organisms that are difficult or can-
not be cultured in vitro. However, due to the exponential
amplification of target DNA sequences (i.e., PCR products
or amplicons), PCR is prone to false positive results if good
laboratory practices are not followed. In addition, environ-
mental samples often contain PCR inhibitors that can lead
to false negative results. Taken together, laboratories con-
ducting these technique demanding molecular protocols
need to establish and follow comprehensive QA/QC proce-
dures. While each organism type (e.g., protozoa, bacteria,
viruses) will require different protocols for lysis and recovery
of nucleic acids, the QC guidelines presented here generally
apply. Additional detailed QA programs and QC require-
ments are documented within the different EPA methods
and guidance documents, such as Quality Assurance/Quality
Control Guidance for Laboratories Performing PCR Analyses
on Environmental Samples (3), Method 1611: Enterococci in
Water by TaqMan Quantitative Polymerase Chain Reaction
(qPCR) Assay (4), and Method B: Bacteroidales in Water by
TaqMan Quantitative Polymerase Chain Reaction (qPCR)
Assay (5).
FACILITY DESIGN
As noted, due to the exponential number of amplicons
produced within a single PCR, cross-contamination with
amplicons from previous PCR assays is a significant source
of invalid PCR results. One way to mitigate this type of con-
tamination is to establish separate laboratory work areas, each
with dedicated equipment and supplies. At a minimum, three
physically separate rooms/areas are required of any laboratory
performing PCR. These include a reagent preparation room, a
sample preparation room, and an amplification and product
detection room. If additional space is available for subdivi-
sion, then the laboratory should appropriately be subdivided
into multiple areas.
In addition to separate work areas, unidirectional work-
flow is critical to reduce the risk of contamination. The
reagent preparation area is considered the “clean” area, while
the amplification and product area is considered “dirty,” and
workflow should always move from clean to dirty work areas.
For example, it is acceptable for an analyst to prepare PCR
reagents in the reagent preparation room, and then take the
PCR tubes into the sample preparation area for addition of
samples. In contrast, an analyst should not work in the sample
preparation area, especially not in the amplification and prod-
uct area, and then enter the reagent preparation room during
the same work day.
Equipment (e.g., micropipettors, vortex mixers), supplies
(e.g., pipette tips, microfuge tubes, tube racks), and reagents
(e.g., molecular grade water, lysis buffers, PCR reagents)
from the reagent and sample preparation areas should not
to be taken into the other area. Nothing from the ampli-
fication and product detection area is to be brought into
the sample preparation or reagent preparation areas, includ-
ing laboratory notebooks. It is recommended that the reagent
preparation and sample processing areas use UV light for
decontamination of surfaces. Dedicated laboratory jackets
and disposable gloves are to be worn in each of the designated
work areas and removed before leaving the work area.
The reagent preparation area is designated for the storage
of all PCR reagents and should be under positive airflow pres-
sure if possible. PCR master mix preparation and aliquoting
into PCR tubes should be performed in this room. Reagent
components should be aliquoted into “working” volumes to
reduce freeze/thaw cycles. The Taq polymerase enzyme and
other critical PCR reagents should be stored in manual defrost
–20°C freezers or within a freezer block to prevent degrada-
tion and inactivation over time.
The sample preparation room is for sample processing,
including target organism concentration and isolation as well
as nucleic acid extraction and purification. PCR negative and
positive control samples and method control samples are
also processed in this room. If possible, the room should be
under negative airflow pressure. Extracted nucleic acid is
added to PCR tubes with aliquoted PCR master mix and
immediately capped. Positive control organisms and template
 2.5.5. Quality Control for USEPA Method 1623 Protozoan Analysis ▪ 2.5.5-3
(e.g., plasmids or extracted nucleic acid) should be stored in
separate refrigerators and freezers than field or test samples.
Preferably, a dedicated biological safety cabinet with UV
decontamination would be used for processing and handling
of positive control material.
The amplification and product room houses PCR
machines for amplification and is where post-PCR manipula-
tion/analysis of PCR product occurs. If laboratory space is
available, it is recommended that separate areas/rooms are
established for the PCR machines and for post-PCR manipu-
lations. Unless the amplified product must be taken to an
outside facility (e.g., sent to a DNA sequencing laboratory),
it should not be removed from this room. If shipping or trans-
port is needed, the package containing the amplified product
should not be taken into the sample processing area. Labora-
tory counter tops should be routinely wiped with a 10%
bleach solution for decontamination, and if possible UV light
also used for decontamination of surfaces.
PCR POSITIVE AND NEGATIVE CONTROLS
PCR positive and negative controls are used to demonstrate
adequate performance of PCR methods. There are several
different types of PCR controls, and a more thorough descrip-
tion of each is provided in the U.S. EPA Quality Assurance/
Quality Control Guidance for Laboratories Performing PCR
Analyses on Environmental Samples (3). Positive controls
are frequently prepared by the addition of one of the follow-
ing: (a) whole organisms, (b) purified nucleic acid from the
organism of interest, or (c) cloned native or modified DNA
fragment of the target of interest. The concentration of the
positive control should be between 10 to 100 times that of
the predetermined PCR assay detection limit. This provides
a control which should amplify consistently while reducing
the risk of intralab contamination from handling excessive
quantities of positive control material. Great care should be
taken to avoid cross-contamination of samples with matrix
spikes and method positive controls.
There are four major types of positive controls:
• PCR positive control: verifies PCR reagent function and
proper thermal cycling. At least one replicate of this con-
trol should be run with every PCR batch.
• Inhibition positive control: determines if there is interfer-
ence from substances copurified from environmental sam-
ple matrices. Each matrix type should be examined for
inhibition. If the matrices (or resulting purified samples)
differ greatly in their physical, chemical, and biological
characteristics, then an inhibition control for each sample
may be needed.
• Method positive control: verifies that the entire analytical
process is functioning properly and the target microorgan-
ism can be detected. At least one replicate of this control
should be run with every batch of samples processed.
• MS: determines method recovery. Each matrix type should
be examined for recovery.
There are two major types of negative controls:
• PCR negative control: verifies that PCR reagents are
free from contaminating organisms or nucleic acids
which could cause false positive results. At least one
replicate of this control should be run with every PCR
batch. Analyzing three or more PCR negative controls
(and infrequently more) per batch may help identify low-
level contamination and is highly recommended. For
nested PCR protocols, it is essential that aliquots of the
first-round PCR negative controls be included as negative
controls for the nested PCR.
• Method negative control (laboratory blank): verifies that
sample integrity can be maintained throughout processing
and that gross cross-contamination has not occurred. At
least one replicate of this control should be run with every
batch of samples processed.
PREVENTING CONTAMINATION
As already mentioned, the physical separation of sample proc-
essing and work areas and the use of dedicated equipment and
consumables will greatly help reduce the risk of contamina-
tion. In addition, disposable gloves and dedicated lab coats
are the primary line of defense. From a cost-benefit stand-
point, it is better to frequently change gloves than to lose a
batch of costly samples and PCR reagents due to contamina-
tion. Disposable lab coats are affordable and readily available.
These can be clearly marked with their designated work area
and user’s name and can be discarded periodically or when-
ever deemed necessary.
The use of PCR product carryover prevention reagents
may also be used. Uracil (dUTP) can be incorporated into
the PCR cocktail in place of thymidine (dTTP). The result-
ing PCR products will contain uracil and are therefore differ-
ent from native DNA. The enzyme uracil DNA glycosylase is
also added to PCR cocktails and an enzymatic digestion step
performed prior to thermal cycling for amplification. This
step results in the digestion of contaminating uracil PCR
products and prevents false positives. It should be noted
that many DNA polymerases do not efficiently incorporate
uracil and PCR optimization may be required. One suggestion
is the use of a 4:6 dUTP:dTTP ratio as opposed to only dUTP.
This may help maintain PCR sensitivity while still providing
sufficient PCR product carryover prevention. Regardless, the
use of PCR product carryover prevention reagents or other
approaches is no substitute for proper technique and good
lab practices.
HOW TO ADDRESS QC FAILURES FOR
PCR ANALYSES
Often the first response after suffering false positives or
false negative results for PCR analyses is to discard all PCR
reagents. However, identifying which reagents were contami-
nated or were not functioning properly is a critical step in pre-
venting future failures. For example, was the molecular grade
water used to prepare the PCR master mix contaminated
because it was taken from the reagent room into the main lab-
oratory and contaminated? Was a particular reagent improp-
erly prepared and the analyst needs additional training?
Similarly, methodically reviewing sample processing proce-
dures and reagents is critical for understanding failures of
method in positive and negative controls.
TIPS FOR GOOD TECHNIQUE AND
LAB PRACTICES
Laboratory science is an art form and therefore comes more
naturally to some than others. However, each analyst can
ensure the quality of his or her work regardless of skill level
by using common sense and following recommendations
for good technique and lab practices. The following are
examples:
2.5.5-4 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
• Change gloves immediately after handling stocks of posi-
tive control organisms or positive control nucleic acids.
• Use aerosol-resistant filter barrier pipette tips to prevent
cross-contamination of samples and/or reagents.
• Never allow any part of a micropipette except the dispos-
able tip to extend past the mouth of a tube or bottle.
Transfer the sample or reagent to an appropriate size
tube or bottle. This is a common problem with the use
of microcentrifuge tubes and small volume (e.g., 10 µl)
pipette tips. This can be avoided by using extended-length
pipette tips.
• The use of microcentrifuge tube openers can prevent pain-
ful fingers but may also provide a false sense of security
with regard to sample integrity. If improperly used, tube
openers can be a source of cross-contamination due to
their contact with the underside of tube caps.
• PCR positive and negative controls are typically loaded
into the PCR tubes or plates after adding field or other
test samples. One suggestion is to load negative control
samples immediately after the PCR positive controls to
ensure that good technique has been used and cross-
contamination has not occurred.
• Sample cross-contamination and false positives can also
occur during gel electrophoresis due to spillover of
PCR product from an adjacent well. Rerunning a gel is
recommended when there are unexpected positive sam-
ples and other controls appeared normal, especially
when an adjacent sample is positive.
REFERENCES
1. U.S. Environmental Protection Agency. 2012. Method
1623.1: Cryptosporidium and Giardia in Water by Filtration/
IMS/FA. EPA 816-R-12-001. Office of Water (MS-140),
Cincinnati, OH.
2. U.S. Environmental Protection Agency. 2005. Method
1623: Cryptosporidium and Giardia in Water by Filtration/
IMS/FA. EPA 815-R-05-002. Office of Water (MS 4607),
Cincinnati, OH.
3. U.S. Environmental Protection Agency. 2004. Quality
Assurance/Quality Control Guidance for Laboratories Per-
forming PCR Analysis on Environmental Samples. EPA
815-B-04-001. Office of Water, Cincinnati, OH.
4. U.S. Environmental Protection Agency. 2012. Method
1611: Enterococci in Water by TaqMan® Quantitative Poly-
merase Chain Reaction (qPCR) Assay. Office of Water
(4303T), Washington, DC.
5. U.S. Environmental Protection Agency. 2010. Method B:
Bacteroidales in Water by TaqMan® Quantitative Polymerase
Chain Reaction (qPCR) Assay. Office of Water (4303T),
Washington, DC.
 The Role of Statistical Thinking in
Environmental Microbiology
J. VAUN MCARTHUR AND R. CARY TUCKFIELD
2.5.6
Unlike macrobiologists/ecologists who study relatively large
organisms, environmental microbiologists seldom see the
organisms they are examining. Instead, novel tools have
been developed that capture information on what the organ-
isms have been doing, harvest and characterize components
of their genetic inventory, catalog gene products, and many
other varied and exciting approaches. These methods are
capable of amassing huge amounts of data. This unprece-
dented ability to produce very large data sets creates several
interesting and difficult problems. Unfortunately many stud-
ies, regardless of the size of their spreadsheets, have insuffi-
cient data to allow meaningful conclusions to be made. In
addition with the availability of numerous desktop statistical
analysis packages, many designed for the types of data being
amassed, there is a false sense of power. Put the data into
the right format, input them into a program, and out come
reams of numbers with associated P-values. These programs
are incapable of determining whether the data have been
collected in a manner that does not violate the assumptions
of the analysis. It is the intent of the chapters in this section
to highlight some of the difficulties associated with many stud-
ies and provide direction on how to design studies that will
allow inferences and generalizations to be made. These chap-
ters are not a summary of statistical analytical techniques
but will focus on the problems of insufficient attention to stat-
istical thinking prior to the first data point being collected.
THE PROBLEM
Having reviewed countless submissions for a variety of jour-
nals, we have experienced far too many papers that have,
in essence, nothing to say. They have data and analyses,
P-values, and multidimensional plots, but they have nothing
to say. Indeed, because of the way the data have been col-
lected there is not much that can be said. However, pages
and pages of results and discussion follow with summaries
and conclusions. It is our contention that one need look
no further than the methods section to determine whether
it is needful to read the entire paper. However there is often
way too little information on which statistical tests are used
or not enough information on the types of data collected
recorded in the methods section. The primary danger, one
that should concern all editors and readers alike, is whether
the authors are justified in making conclusions that are
doi:10.1128/9781555818821.ch2.5.6
2.5.6-1
meaningful. Authors will attempt to make conclusions even
when they are not justified in doing so, often without know-
ing they are not justified in making such conclusions.
Consider the following scenario. A researcher goes to a
field location in February and collects one bottle of water
from an impacted site and one from a reference site. The
bottles are taken back to the lab, and three subsamples are
removed from each. These subsamples are then used to per-
form various molecular and chemical analyses. For example,
pH, PO4, NO3, ammonia, organic matter, and other physical
properties are analyzed. DNA is extracted from the sub-
samples and used for various molecular analyses, generating
multiple measurements for each subsample. The data are
then analyzed using various statistical software packages
and, wonder of wonders, “significant” differences are found
between the impacted and reference site for several variables.
The researchers, in an effort to be careful and corroborative,
decide to return to the field location and take another set
of samples in July. The same chemical and molecular analyses
are performed with the addition of a few new tests that
they feel might be explanatory. The results are tabulated
and statistical analyses performed with the inclusion of
tests to determine whether there is a time effect between Feb-
ruary and July. There appears to be a mild trend with a
P-value = 0.07. Undaunted, the researchers decide that if
they collected just a little more data they might be able to
verify this seeming pattern. So it is back to the field and col-
lect more samples. This time they collect sets of samples in
early March and late June and run the same battery of tests.
Statistical analyses are run, and what appears to be a trend
is found with significant differences between winter and
summer for a number of variables. What can the researchers
say from this study? Basically nothing!
Let us list the problems with this study (for the record, this
is an actual study that one of this chapters authors was asked
to review).
1. There was only a single sample collected on each date
from each location. Regardless of how many subsamples
might be taken from this sample there is only “one”
sample. Any means obtained from the subsamples are
not independent measures but are estimates of the true
value of the sample, not the field location. These sub-
samples provide information on how well the scientists
2.5.6-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY

could replicate their procedures. They do not reflect the

variance associated with each of the locations and thus
cannot be used to determine if differences actually exist.
2. Statistical analyses of the “single” sample shows a signifi-
cant difference. This finding (P-value) is an estimate of
the differences among the subsamples taken from the sin-
gle sample. Because the P-value was <0.05 the finding
means that the subsamples were different. It does not
mean that the locations were different, only that the var-
iability within a sample did not overlap with the other
samples. What can be said? Nothing since we do not
know if the samples are representative of the field loca-
tion or the season. Multiple samples would have needed
to be collected at each field site on each date. But that
isn’t even sufficient; the researchers would need to
show how the sample locations were chosen and whether
they adequately sampled to capture the variability within
FIGURE 1 Relationship between X and Y variables using narrow
the field location. If the samples were all taken in close
scale on the y axis. doi:10.1128/9781555818821.ch2.5.6.f1
proximity to each other, we might expect there to be
fewer differences than samples taken at increasing
distances. Again we can’t begin to answer that question
with the data collected. extreme when we think we see patterns in random data.
This can happen when we plot the data using inappropriate
3. Decision to sample again was not based on science but on scales or fail to include error bars with the data. For example
a desire to confirm a faulty observation. Why sample in consider Fig. 1. It looks like there is an increasing trend
July? Why not some other month? Time differences across the x-data. However this regression is not even close
can and do contribute huge sources of variation, but some to being significant (P = 0.59). Our eye sees a trend but there
justification should be given as to why July was chosen. really isn’t one there. One reason for this is that the scale of
There is nothing wrong with arbitrarily picking another the y-data is only from 0.98 to 1.08. If we change the scale
date, but careful thinking about physiology or biology to 0.9–1.1 (Fig. 2) the line appears to flatten and we would
could inform the decision. be less likely to think there is a trend—same data, different
4. The next sets of samples were collected in months close presentation, with a different likely interpretation. This prob-
to those sampled in the first year but not exactly. Is this a lem is exacerbated when there are only a few observations.
problem? We don’t know, and more important, we can- For example, tossing a coin 10 times may and most likely
not know from these data. Because only a single sample will not result in a 50:50 probability distribution for the
was collected from both locations on each date we have number of heads and tails. What if we get eight heads and
no estimate of the variance associated with the samples two tails? A highly unlikely event, of course, if we assume
in space or time. Attempts should have been made to that the coin is fair and unbiased. But without an underlying
collect as close as possible to the original dates so that knowledge of probability and statistics in this binomial proc-
the results are not compromised. ess, what would be our likely conclusion? We would tend
5. Additional analyses are performed on later samples than to rely strictly on the outcomes and presume that this coin
those performed initially. No temporal comparisons can would come up heads 80% of the time. Even if we suspect
be made on the new analyses. If the new data were that the probability distribution on the outcomes of heads
thought to have been important, it should have been and tails should be close to 50:50, our conclusion would be
measured on all samples from the beginning. that we expect more heads than tails in further coin tosses.

From this simple but real example, it is evident that the

researchers did not understand statistical thinking or the
ramifications of not thinking their “experiment” through.
The word experiment was put into quote marks because the
term can only be loosely used with regard to this study. There
was no experiment, and the observational data are incredibly
weak. Many other studies could have been used for illustra-
tion including ones using microchips, various DNA/RNA
techniques, or any other automated procedure that produces
large quantities of data or more specifically many observations
on a single sample. Having many measurements for a single
sample can be a good thing, or it can give a false sense of
meeting the requirements of a particular statistical analysis.
When there are many samples, each with many measure-
ments, the researcher is in a good place to be. Alternatively
having many measurements on just a few samples creates
unsupportable results. The next chapter in this section will
discuss this problem in much more detail.
Scientists are trained to make observations and to try FIGURE 2 Relationship between X and Y variables using broader
and see patterns in nature. However we can take this to the scale on the y axis. doi:10.1128/9781555818821.ch2.5.6.f2
 2.5.6. The Role of Statistical Thinking in Environmental Microbiology ▪ 2.5.6-3
But the empirical expectation of a 50:50 distribution is based
on large numbers of flips and if we were to toss the coin 100
or 1,000 times the expected distribution would become
more and more apparent. An a priori understanding of the sci-
ence of probability and statistics would leave us with no sur-
prise but a pleasant confirmation by what we observe. Data
analysis without some understanding on how the data were
generated can contribute very little to knowledge.
There are many reasons we don’t get a 50:50 distribution
when we toss a coin 10 times, including differences in flipping
speeds, wind, and height differences. There are many more
reasons environmental samples will be different from each
other, and thus it is absolutely necessary to collect enough
data to properly estimate the true distribution of whatever
we are measuring. It is through the mathematical methods
of statistics that we test to see if there are justifiable patterns
in the noise of nature.
Motulsky lists five steps to correctly apply statistical
analysis in an experiment (note this is not a list for observa-
tional studies—see below for discussion of these sorts of stud-
ies) (1):
1. Define a population you are interested in. This is an incred-
ibly important first step because it will define the breadth
of what you can say from the data. Population here refers
to the total possible outcomes of some variable of inter-
est. It does not necessarily refer to a biological popula-
tion. This step requires that we have some idea of what
we expect to be measuring. It implies that we know
the bounds of the variables. For example, if we wish to
know the growth rates of bacteria as a function of salinity,
then it is very important that we have some idea of how
much salinity is too much. Failure to have this under-
standing may result in an experiment where the salinity
there. Specifically Motulsky identified five problems that stat-
is too high and no results are generated—at least no
istical calculations are limited in overcoming (1):
results that are publishable.
2. Randomly select a sample of subjects to study. Subjects can
be bottles of water, soil cores, scrapings from teeth, indi-
vidual fish, lakes, streams, contaminated wetlands, and so
on. We seek to take samples that reflect the variability
found within our sample population. Thus taking a single
sample or measurement is insufficient to capture that var-
iability and more importantly can never be so. Often
environmental microbiologists seek to determine the
effect of some stress or impact. To determine an effect
they sample from two populations: one impacted the
other not. However, extreme care needs to be taken to
ensure that the subjects measured are reflective of the
populations. Random means exactly that. If the selec-
tion of subjects is biased either by the researcher or
by events, the results will have very limited value. Ran-
domization prevents variables that are not accounted
for from affecting our results in some consistent manner.
It makes these unaccounted variables act randomly.
Randomization reduces confounding and, if the sample
size is large enough, cancels out the effects of variables
not accounted for (2).
3. Randomly select half of the subjects to receive one treatment
and give the other half another treatment. Random means
that every subject has an equal chance of being assigned
to either group. It is very easy for experimental bias (see
below) to affect assigning subjects to groups. Care must
be taken especially when we are collecting data from
nature and are performing some sort of manipulative
study where we will need a control, randomization, and
replication. Seldom in nature are two sites totally equal
in all characteristics. If you are selecting locations ran-
domly in nature, you may get a control area that is
much larger or smaller than the location you seek to
treat. Any differences in the data may be from conditions
that exist prior to the experiment and not from the
experiment. Thus in field experiments it is important
to recognize the difference between controls and refer-
ence units. Reference units are examples of some ideal
condition (2). For example, if we are trying to determine
whether some anthropogenic input has affected some
microbial processes, we can compare the processes within
our study area and relate the observed patterns back to
the reference unit. As such reference units are not con-
trols because no treatment has been applied by the
researcher but rather something to compare “impacted”
site back to.
4. Measure a single variable in each subject. While this is the
simplest experimental design, often more than one vari-
able is screened at the same time. Increasing the number
of variables increases the need for larger sample sizes. In
addition, the more variables we include, the greater like-
lihood that there will be significant interactions among
the variable that complicate the interpretation of the
results.
5. From the data you have measured in the samples, use statisti-
cal techniques to make inferences about the likely distribution
of the variable to the population and about the effect of the
treatment. We discuss the importance of making infer-
ences later. However, the preceding four steps define
the level of inference that can be made from a specific
data set.
Statistics cannot make something out of data that isn’t
• The population you really care about is more diverse than
the population from which your data were sampled. If the
samples you collect are not representative of all possible
samples or of the true population, then they won’t capture
the level of diversity and variability found in the popu-
lation of interest and the results will not be representative
of what is really happening in nature. In many envi-
ronmental studies, single contaminated and uncontami-
nated sites are compared. In this case the subject is the
study site, and since there is only one of each we cannot
infer that our results apply to any other study sites regard-
less of whether they are contaminated with the same
pollutant.
• The subjects in the study were not randomly sampled from a
larger population. In many of the observational and exper-
imental studies we have reviewed in environmental micro-
biology the subjects (samples) are taken haphazardly
with little thought about their random selection. Samples
are often taken for convenience and not to capture the
variability of the measured objective. Several samples
might be taken in close proximity instead of being ran-
domly taken from all possible locations. Thus the study
is biased and the results are of little value in trying to
understand the phenomenon or process of interest. It
should be noted that completely random sampling is
not always the best approach. In some cases, based on
prior knowledge, we might know that there exist spa-
tially defined zones in such a case we may perform a
stratified random sampling scheme. Here we randomly
sample within each zone so as to capture the uniqueness
2.5.6-4 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
of the zone. In this instance, the zone-to-zone variation is
a useful partition when compared to the within zone
variation.
• The measured variable is a proxy for another variable you really
care about. For example, say we are interested in an organ-
ism that has only been identified via genetic techniques.
In a number of studies we have found that the density of
a culturable bacterium is correlated with results from
qPCR on the organism of interest. Rather than perform
many qPCR analyses, we screen the samples for the cul-
turable bacteria and use previous regression analyses to
estimate the density of the organism of interest. The prob-
lem is that all possible relationships between the two
organisms is not known and thus the proxy variable may
over- or underestimate the true density.
• The measurements may be made or recorded incorrectly
and assays may not always measure exactly the right thing.
Calibrating technicians is not a trivial concern. Failure
to calibrate instrumentation similarly can greatly affect
outcomes. In many studies more than one person makes
measurements or records data and differences among the
technicians may be as great as the effect you are trying
to measure.
• Scientific conclusions require looking at multiple outcomes, not
just one. The problems associated with multiple compari-
sons are discussed in the other chapters in this section.
A related problem that we have observed is when a set
of data is analyzed in many different ways until something
interesting pops out. Scientific honesty requires that
we state how we are going to analyze our data prior to col-
lecting the first data point. We should then analyze the
data in this way. Even if we don’t detect differences we
need to report this information and not continue to try
new analytical techniques until we find one that shows
something. In addition to these problems, researchers
often analyze using many techniques on their data until
some interesting pattern emerges. It has been said that if
you torture your data long enough, it will tell you anything
you want to hear (3).
Because of these limitations of statistical analyses, it is nec-
essary for researchers to exercise judgment in the interpreta-
tion of their results.
EXPERIMENTAL BIAS
Experimental bias creeps easily into experiments because
humans are subjective (4) even when they think they are
being objective. Sackett identified seven aspects of an experi-
ment where this bias can occur (4):
1. In reading up on the field—only certain papers are
read, only one side of an issue is read, only papers with
positive results are read, we are influenced by what seems
to be the hot topic of the moment at meetings or in
journals.
2. In specifying and selecting the study sample—Sackett (4)
identified 22 biases associated with specifying and select-
ing the study sample, many of which apply to clinical
studies. However, microbiologists can favor one strain
over another, they may be influenced by past studies, or
use the wrong sample size, and so on.
3. In executing the experimental maneuver (or exposure)—
Here we can be biased through contamination (known
or unknown), poor controls or improperly adminis-
tered controls, various compliance issues, and more.
This bias can also occur when different technicians
make the “same” measurements.
4. In measuring exposures and outcomes—Researchers can
bias this aspect through their expectations, instruments,
not accepting certain results as valid, substituting
another subject if one fails to show a certain result, and
many other practices.
5. In analyzing the data—As we will discuss, reporting signif-
icance levels is expected and misused. This is especially
true when we determine after an experiment the signi-
ficance level we choose to report. If “significant” results
are not obtained with one data analysis, researchers
continue trying other techniques until one of them
shows the results they were expecting. It may involve
using scales that show a statistical trend but have no bio-
logical relevance or more often have significance within
our set of subjects but not to a wider population.
6. In interpreting the analysis—Many times researchers do
multiple correlation analyses among every possible
variable to try to find something that might explain
their results, make new hypotheses up to fit the results
obtained after the fact , or subject the data to analyses
based on these new hypotheses until some significant
result is obtained.
7. In publishing the results—repeated submissions to several
journals following rejection until one of them agrees to
publish the results.
CAUSATION VERSUS CORRELATION
Understanding what factors may be influencing what we are
interested in discovering is essential and requires that we
understand the difference between correlation and causation.
Confidence in your summarized data increases if our findings
are consistent with reasonable mechanisms. Causation can be
inferred if we have eliminated other possibilities and shown
the phenomenon over a many different circumstances. It
is ensured if there is a plausible dependence between the
outcome we observe and the putative cause.
STATISTICAL SIGNIFICANCE AND INFERENCE
Readers of recent scientific literature are bombarded with a
continual stream of statements declaring that such and such
results are significant. Indeed, we have become so reliant on
statistical significance and P-values that submitted papers
can have great difficulty in getting published if they don’t
have their fair share. Unfortunately most researchers do not
understand what “significance” means or what a P-value is
really telling us. This is clearly evident when papers publish
P-values such as <0.000001. Such papers are implying asymp-
totic certainty (5).
Inference
Most researchers want to be able to make some inference
based on the results of their studies. There are two different
inference-making procedures (6). These two approaches are
“estimation” and “test of hypothesis.” With estimation a
researcher is asking whether the data collected are representa-
tive of the real population value. An estimate is the result of a
defined calculation, for example, a mean or standard devia-
tion. This procedure requires enough samples to be collected
such that the data obtained meets the requirements of the
central limit theorem. The value we are estimating is the
sample mean obtained from many samples, and we wish to
 2.5.6. The Role of Statistical Thinking in Environmental Microbiology ▪ 2.5.6-5
know whether this value fairly represents the true population
mean that we do not know. In essence we are trying to under-
stand something about a system or organism about which
little or nothing is currently known. For example, what is
the abundance of a specific microbe in sea ice, or what
is the GC content of some species/group? To reliably estimate
the value of interest, it is absolutely essential that we collect
enough samples or screen enough independent isolates.
The number of samples is critical; regardless of how time-
consuming or expensive the analysis is, without meeting
the requirements to make the estimate valid, our estimated
value is nearly meaningless.
Making inference through estimation is the goal of obser-
vational studies. In these studies, researchers cannot know the
true value of a parameter for some measure of the microbial
population, but they seek to estimate it. No experiment is
conducted. It is observational science and as such is still an
important contributor to knowledge. Data must be carefully
collected so that emergent patterns are reliable to the larger
scientific community. Thus careful statistical thinking must
be used prior to collecting observational samples. Haphazard
sampling, few samples, and limited subjects will all affect how
the data can be used. According to Tuckfield’s maxim (5),
data collected for some general and unspecified purpose can
answer very few specific questions. The scope of inference
for observational studies must be limited to situations that
are similar to the one being investigated (2).
The second procedure for making inference involves
testing a statistical hypothesis. A statistical hypothesis is not
necessarily the same as a biological hypothesis: hence the
problem. Researchers perform statistical analyses of their
data and then present P-values that are supposed to be dem-
onstrating that their results have supported their biological
hypothesis. Making inference using statistical hypotheses is
basically asking whether the value(s) we obtain from our
samples are equal to some specific value or to each other.
For example, having collected density estimates of bacteria
in natural undisturbed lakes of some region, we wish to
know whether some anthropogenic stress (heat, chemical,
physical, etc.) has altered this density. Based on our previous
sampling we know that the density is X and we wish to know
from a series of new samples whether the impact has altered
the density. To answer the question we have to state a statis-
tical hypothesis such as the mean density of the impacted
lakes is less than that of the un-impacted lakes. Our null
hypothesis then becomes that the density of the impacted
lakes does not differ from the un-impacted lake. In this exam-
ple the preexisting case is known. In many cases the null
hypothesis is what the scientist is trying to prove false (1).
A similar approach can be used even if we don’t know the
density of un-impacted lakes, but we expect it to be different
from the impacted lakes. In this instance the null hypothesis
is simply that the mean of un-impacted lakes is not different
from the mean of the impacted lakes. What is the scope of
inference of these studies? It depends on what the scientific
hypothesis was. Remember that the subjects in these studies
are lakes and that multiple sampling within a lake helps us
understand the “true” value for that lake. If we desire to use
our data to make inferences about all such impacted lakes,
it is essential that multiple lakes, as similar as possible (major
assumption), of both impacted and un-impacted be sampled
within the region. If this is done then we can infer that there
is or is not a difference between the factors we are measuring
in the lakes in this specific region.
Now consider what we could say if we had only sampled
one lake of each type. In this situation our scope of inference
is the two lakes. We could say nothing about any other lake
either impacted or not within the region.
Significance
Statistical significance has become the household word of
most scientific publications. Unfortunately it is little under-
stood, and far too often used to imply importance of the
work, and way too often the basis for acceptance or rejection
of a manuscript. For example a quick survey of recently pub-
lished studies had titles that included “biological signifi-
cance,” “ecological significance,” significance of sustainable
agriculture,” and “significance of size and nucleic acid con-
tent.” Whether intentionally applied or not, these titles sug-
gest that the results are significant or important. Statistics
becomes the tool for determining the value of a study regard-
less of the clear scientific thinking that went into the study!
Students and mentors alike are on a quest to find significant
results regardless of whether the results have any biological
meaning. Papers report ever decreasing P-values (5) and the
α ( predetermined level to accept or reject) is forgotten.
This level of α should not be arbitrary and should be based
on the thinking that went into the study prior to the first
data point being collected. How many papers qualify their
results by stating that they are nearly significant or some
variant of that statement? This practice is bolstered by a
psychological need to have significant results and that a
study isn’t worth reporting that does not meet that level.
Unfortunately journals usually don’t like papers that present
negative results or that don’t show significance. Rather
than throw out the whole study, authors come up with clever
ways of hinting that the data really do show something.
The real fact is that the data probably are valuable and the
authors have let a conclusion based on P-values prevent
them from determining what the data really demonstrate.
Many researchers fall for what Altman and Bland (7) call
the mistake of saying that lack of significance means lack of
difference, when in reality all it is saying is that there is insuf-
ficient evidence to conclude that there is a difference. Thus
they (7) title their paper “Absence of evidence is not evi-
dence of absence,” indicating that too few samples were col-
lected to provide the evidence. Again we repeat that statistics
cannot take the place of careful, determined, and clear scien-
tific thinking. Motulsky (1) suggests that instead of having
only two conditions (significant or not significant) we should
include a third category of inconclusive. This suggestion has
merit and allows the reader an opportunity to consider the
data on its own.
The P-values obtained from the statistical analysis can
only be interpreted based on the null hypothesis as defined
by the scientists. The P-value does not prove that the null
hypothesis is true because the null hypothesis is assumed to
be true. Rather, it is a measure of the strength of evidence
that the null hypothesis is likely true. In our example of com-
paring two or more groups, it is very unlikely that the groups
will have the same means. In such a case we have to decide if
the observed differences are meaningful. In some cases the dif-
ferences might be statistically significant but not biologically
significant. For example we might discover that the tempera-
ture optimum of some species of bacteria is 36.8°C rather than
37°C as previously reported. In such a case the new value does
little to further our biological understanding, regardless of
how significant it is statistically.
P-values need to be interpreted based on the context
of the experiment. Higgs (8) citing R. A. Fisher reported
“No scientific worker has a fixed level of significance at
which from year to year, and in all circumstances, he rejects
2.5.6-6 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
hypotheses; he rather gives his mind to each particular case in
the light of his evidence and his ideas.” The use of P < 0.05 is
arbitrary and causes numerous problems in interpreting the
real importance of a data set as editors, reviewers, and many
readers are more concerned with the outcomes relative to
this value than they are to what the data are saying. Fisher
goes on to say (as reported by 7), “In choosing the ground
upon which a general hypothesis should be rejected, personal
judgment may and should be properly exercised.” Higgs
suggests that at exactly the place where we are tempted into
using the word “significant” we should insert words that
demonstrate our personal judgment relative to what the
data are showing (8). This will require authors to carefully
think about what they mean in the place where they would
generally use the word “significant.” Of course if our P-values
>> 0.05 then personal judgment may just be hand waving. It
should be noted that Fisher considered a significant P-value
as providing justification to continue studying. Today when
researchers obtain nonsignificant P-values they continue
thinking that all they need is a larger sample size to finally
get significance.
Given that the meaning of the word “significance” has
changed considerably since Fisher first used it in statistical dis-
cussions, scientists, regulators, politicians, and the public
have fallen into a semantics trap. Today papers with signi-
ficant results are also considered significant, where in this
instance “significant” means important. That may or may
not be the case. When one hears a colleague suggest that
we read a significant article he or she has read recently,
what do they mean? Is it significant because it has statistically
significant results which may or may not be biologically
significant, or is it significant because the findings make a
major advance in scientific thinking or understanding?
The use of “significant” clearly has more than one intended
meaning.
Tuckfield (5) has argued that it is important to let the data
speak for themselves through the weight of evidence. With
any study where the data are continuous, either in time or
space, a first step should include summarizing the results vis-
ually. Many graphical approaches can be used to summarize
data, without statistical tests being imposed, that can offer
keen insights into patterns and processes. These presentations
are often more effective than complex statistical approaches
and they do not violate any assumptions. Furthermore, the
inherent variation associated with all data sets can be seen
and appreciated without the use of some summary metric.
However given the tendency of many researchers to find pat-
terns when they don’t exist, various summary statistics such as
box and whisker plots could be used to give considerable
insight into the data.
VARIATION
Because of the incredible improvements in analytical proc-
esses and equipment, much of the systematic variance in envi-
ronmental microbiological studies has been eliminated. One
can expect that identical samples will produce similar results.
If we truly reduce experimental bias as discussed, we may still
find differences even when none exist. What then are the
major sources of variation that environmental microbiologists
can expect? Let us consider the following:
1. Sampling bias
2. Lack of homogeneity among sub-samples
3. Differences among technicians
4. Confounding
Sampling Bias
When collecting samples, it is essential that care be taken to
select samples that represent the conditions you seek to char-
acterize. In most environmental microbiology studies the
actual size of a sample is many orders of magnitude larger
than the organisms of interest. Thus the variability within a
sample may be nearly as great as between samples. To capture
the variation among samples, it is necessary to take sufficient
samples that enable the variation to be estimated. As men-
tioned, one sample that is subsequently subsampled produces
results that are biased and not independent. To overcome this
problem, several small samples should be collected following
careful statistical thinking of the scientific question. It is
essential to remember that the scope of inference, or what
can be said from your data, is dependent on how the experi-
ment is designed and carried out. Most scientists wish to infer
more than their data will allow.
Lack of Homogeneity among Sub-samples
Subsampling is not always a terrible thing. As mentioned,
physical samples (e.g., volumes) taken in the field are often
too large to analyze as a single unit and thus a subsample
must be taken. Multiple subsamples from a single sample
are not acceptable except when we are interested in seeing
how well our analytical procedure is performing. If similar
results are obtained on subsamples from a single sample we
can have some assurance that the method is working and
repeatable. However, in general, a single subsample should
be taken from a sample. Subsampling should only be done
if one is sure that every possible subsample is similar to
another from the same sample. Lack of homogeneity among
subsamples will create increased levels of variability. In
most environmental studies there is variability between sam-
ples collected even a short distance from each other, and
within each sample there is variability among the subsamples.
Thus care must be taken to ensure that subsamples are repre-
sentative of the whole sample. To accomplish this we could
stratify our subsamples. For example we may take a 10 cm
core but are really interested in what is happening at a partic-
ular depth. In such a case we would take our subsamples from a
particular section of each core. If we are interested in what is
occurring within the entire 10 cm, it would be necessary to
homogenize the core prior to subsampling. Unfortunately
all of the tightly controlled relationships between species
and the abiotic conditions found with depth would be erased
through homogenizing.
In many microbiological procedures that seek to estimate
community composition, there is an implied assumption that
every subsample is equal. There are many technologies that
involve multiplates. The assumption in these technologies
is that the inoculum going into every cell is identical. This
is doubtful. Considering the environmental grain found
with respect to bacteria, no two samples will ever be exactly
similar. This is even undoubtedly true for samples taken in
close proximity to each other. Thus decisions on differences
in community composition need to be made based on means
of many such tests.
Differences among Technicians
Variability introduced because of differences among techni-
cians is common in laboratory and field studies. This source
of variability can be controlled in part by calibrating between
technicians. However many factors can affect the ability of
a single technician to obtain similar results on different
days. For example, you cannot control for differences in
 2.5.6. The Role of Statistical Thinking in Environmental Microbiology ▪ 2.5.6-7
technician reliability when they are working under stress
caused by domestic contentions the night before, illness,
financial concerns, and so on. When more than one techni-
cian is involved, the likelihood of increased variability is dif-
ficult to control, and these differences may produce results
that are statistically significant even when they are not.
Confounding
If a researcher does not have sufficient controls to eliminate
alternative explanations for an observed relationship between
dependent and independent variables, such a study is consid-
ered confounded. In most environmental studies relation-
ships between variables known or unknown can affect the
results of study through confounding. If a measured variable
is correlated with both the independent and dependent var-
iables, the results are not interpretable (9). The conclusions
made from such studies are considered spurious. An obvious
example involves seasons or time. For example, a researcher
sets out to collect a set of samples from along a lakeshore.
They wish to sample several long sections at various intervals
around the lake. To accomplish this task they randomly select
a starting location and then proceed to go around the lake,
always starting the next day’s sampling from the last point
of the previous day. The sampling effort takes more than a
month to finish. Unfortunately this sort of approach may
have confounded time with the samples. By the time the
researcher finished the sampling, any number of events natu-
ral or unnatural could have happened that would affect their
observed results. The first and last sample locations, which are
spatially close, may be different simply because of the time
that has passed since the first sampling. To prevent confound-
ing due to time or season, the researchers should have ran-
domly determined which sections were to be sampled.
Failure to include season in an analysis happens frequently.
Things both biotic and abiotic can change quickly between
sampling dates. Another example might be an experiment
where all replicates of a particular treatment are always placed
in an incubator or growth chamber in exactly the same loca-
tion. Subtle differences between conditions in the incubator
or growth chamber that are dependent on location would
confound the results.
MEASURES OF VARIABILITY
If careful statistical thinking is not put into practice, too often
results are obtained in which n is ambiguous or at least not
obvious. Consider a case where the researcher makes repeated
measurements on three sediment columns over time. For each
column there are six “samples.” Too often, from our experi-
ence, the researchers perform their analyses on all of the
data using an n = 18. However the true n for this example is
only 3. The multiple measurements for each column should
be averaged and the mean used to perform the statistics.
The only way n could equal 18 would be to have had 18
columns randomly assigned to the three treatments and
then the sampling order randomly assigned such that one of
each type of column was taken on each sampling date. If there
was a desire to determine differences on each sampling date
there would only be three samples taken for each date with
no replicates. Thus to make inference about different time
periods, additional replication would need to be designed
into the study. To do so we would need to have at least 36 col-
umns, where 2 columns of each treatment were randomly
sampled on each date. As can quickly be seen, as the complex-
ity of the experiment goes up, so does the sample size needed.
Analytical costs often limit the number of samples that can be
run. If such is the case the experiment would necessarily have
to be simplified so that sufficient samples were taken.
Having some appreciation for the need to show the varia-
bility of their measures researchers will often present error
bars. There are two different bars used that are very dif-
ferent in what they represent (10). While many studies state
in the figure legends what is being presented, most readers are
unaware of the differences and blithely continue read-
ing, assuming that the two measures are similar. Means
are graphed or tabulated with either a standard devia-
tion (±SD) or as a standard error of the mean (±SEM). Stand-
ard deviations areP calculated as the square root of the variance
(s 2) where s2 ¼ (Yi  Y) 2 =n  1 and Yi is one of n values

and Y is the mean of all values. The standard deviation is thus
based on the sum of squares of the differences between the
observed values and the mean and the number of samples
minus 1. This value is a quantifiable estimate of the variability
within a set of data. It has meaning relative to comparing
results from treatments or observations. If the samples are
good estimates of the natural variation, then the standard
deviation will not change much with increasing the sample
size. However, if the samples are not representative of the
true variability of the subjects, then increasing the sample
size can either increase of decrease the standard deviation.
Regardless changes in the standard deviation with increasing
or decreasing samples sizes will always be smaller than changes
in the standard error of the mean in response to increasing or
decreasing the sample size. Having larger samples helps you
quantify the variability more precisely.
In contrast the standard error of p the mean is calculated
from the SD as follows: SEM ¼ s= n where s is the SD.
This is a misguided estimate of variability among sampled
observations (11). Rather, it is an estimate of the precision
of the mean and can often be replaced with a confidence
interval, which provides more information and is easier to
interpret. An SEM is always smaller than the SD because
it is divided by the square root of n and if n is very large
the SEM can be quite small. Researchers often report SEM
because the values are small, don’t overlap, and give the
appearance of having data that are different. Reviewers
and editors need to ensure that SEMs are not reported
when in fact the researchers are purporting significant
results. Standard deviations allow comparisons to be made
between means.
Sometimes researchers wish to compare the variability or
scatter of data that have been collected in different metrics.
For example, a researcher might be interested in knowing
whether the variability in soil respiration rates is similar to
the variability in the concentration of mercury in soil pore
water. These two measures are made in different measurement
scales. One way to compare the relative variability is to calcu-
late the coefficient of variation (CV) on both types of sam-
ples. CV is calculated by dividing the SD by the mean.
Since the mean and the SD are in the same units, the CV
is a unitless measure. It is often expressed as a percent. Because
each CV has no units, it can be used to compare across meas-
ures that have been made in various ways.
Careful statistical thinking is often associated with com-
paring the variance within a set of samples to the variance
between samples. This is in fact is the basic idea known
as the analysis of variance (ANOVA). The variance as noted
above is the square of the SD. It can be partitioned such
that researchers can ask what portion of the variation is sys-
tematically explained by some factor or variable, compared
to the remainder, which is considered random variation.
However it is crucial that you the researcher understand
2.5.6-8 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
SD or the confidence interval for readers to grasp your analy-
ses. Many researchers feel that the purpose of statistics is to
determine whether their results are significant. As Moltusky
(1) points out, the whole point of statistics is to quantify
scientific evidence and uncertainty, not to prove biological
hypotheses. Determinations of whether a biological hypoth-
esis is supported or not comes from the clear scientific think-
ing of the investigators and their interpretation of the weight
of evidence.
POWER ANALYSES
There are a growing number of studies that have started
reporting the statistical power of their experiment, a trend
that is fundamentally flawed (12). Statistical power is the
probability of rejecting the null hypothesis when in fact it is
false. In other words, it’s the likelihood of avoiding a false neg-
ative conclusion. Unfortunately these estimates of statistical
power are being made a posteriori, that is, the researchers are
using the data they have collected to generate the estimate
rather than determining a priori the number of samples neces-
sary to detect a difference with a given level of variability. By
definition if their data did not detect a difference then they
didn’t have sufficient power. To present this probability
derived from their collected data and thus after the fact pro-
vides little new information. Power analysis can be an effec-
tive tool to help design an experiment. In this case the
statistic provides information on how large the study needs
to be to detect a difference or effect of a certain size given
the distributions of the populations of interest. This type of
analysis then gives the researcher information on how to
design their experiment. Many times such an analysis suggests
more samples than can reasonably be taken give economic
and time constraints. Thus simpler designs become necessary.
By “simpler” we mean fewer variables or combination of var-
iables and not fewer replicates.
SUMMARY
Environmental microbiology is thrusting ahead, creating a sea
of data. Some of these data are very good and provide clear
insights into normally intractable problems associated with
the complexities of the environment. However, there are
numerous examples of poorly designed studies that produce
poor data. Some of these studies are awash in observations
on a few samples. This feeling of having collected sufficient
data when in fact all they have done is measure a few samples
to the nth degree seems to be proliferating. We have presented
a brief summary of various problems environmental micro-
biologists may face as they seek to analyze their data. It has
been our experience that these problems can be addressed
with clear statistical and scientific thinking before the first
data point is collected.
The writing of this manuscript (JVM) was supported by the Depart-
ment of Energy under Award Number DE-FC09-07SR22506 to the
University of Georgia Research Foundation.
This report was prepared as an account of work sponsored by an agency
of the U.S. government. Neither the U.S. government nor any agency
thereof, nor any of their employees, makes any warranty, express or
implied, or assumes any legal liability or responsibility for the accuracy,
completeness, or usefulness of any information, apparatus, product, or
process disclosed, or represents that its use would not infrin ge privately
owned rights. Reference herein to any specific commercial product, proc-
ess, or service by trade name, trademark, manufacturer, or otherwise
does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the U.S. government or any agency thereof. The views and
opinions of authors expressed herein do not necessarily state or reflect
those of the U.S. government or any agency thereof.
REFERENCES
1. Motulsky H. 2010. Intuitive biostatistics: a nonmathema-
tical guide to statistical thinking. Oxford University Press,
New York.
2. Johnson DH. 2002. The importance of replication in wild-
life research. J Wildl Manage 66:919–932.
3. Coase RH. 1994. Essays on Economics and Economists. The
University of Chicago Press, Chicago, IL.
4. Sackett DL. 1979. Bias in analytic research. J Chron Dis 32
(1–2): 51–63. doi:10.1016/0021-9681(79)90012-2.
5. Tuckfield RC. 2004. Statistical thinking for students of
ecology. In, Odum EP, Barrett GW (eds.), Fundamentals of
ecology. 5th edn. Cengage Learning, Independence, KY.
6. Ott L. 1977. An Introduction to Statistical Methods and Data
Analysis. Duxbury Press, North Scituate, MA.
7. Altman DG, Bland JM. 1995. The absence of evidence is
not evidence of absence. Br Med J 311:485.
8. Higgs MD. 2013. Do we really need the S-word? Am Sci
101:6–9.
9. Wikipedia contributors, “Confounding,” Wikipedia, The
Free Encyclopedia, 19 April 2013, 01:17 UTC, http://en.
wikipedia.org/w/index.php?title=Confounding&oldid=
551064399, accessed 26 April 2013.
10. Altman DG, Bland JM. 2005. Standard deviations and
standard errors. Br Med J 331:903.
11. Curran-Everett D. 2008. Explorations in statistics: standard
deviations and standard errors. Adv Physiol Educ 32:
203–208.
12. Hoenig S, Heisey DM. 2001. The abuse of power: the per-
vasive fallacy of power calculations for data analysis. Am
Stat 55:1–6.
 Study Design
YILDIZ T. CHAMBERS AND ROBIN K. OSHIRO
2.5.7
STUDY DESIGN
When designing a study to develop or optimize a method,
evaluate method performance, validate a method, or test a
hypothesis, there are many factors to consider, for example,
matrix/matrices (e.g., water, soil), spiking, organism strain,
number of replicates, and quality control (QC) analyses.
The first step is to identify and clearly define the study objec-
tives. For example, studies to develop or optimize a method or
evaluate method performance may include assessment of sen-
sitivity, specificity, and limit of detection, whereas a study to
test a hypothesis would include specific analyses to test the
hypothesis. Once the objectives (e.g., interlaboratory valida-
tion [nationwide applicability], method comparability
[in-house use]) have been defined, the approach can be deter-
mined and the factors addressed. For example, the objectives
for an interlaboratory (round robin) validation study, at a
minimum, should include: (a) verify that the method works
in multiple laboratories among different users and (b) assess
the method performance in matrix of interest (e.g., recovery,
precision).
Ideally the study should be designed in such a manner that
the data generated will ensure that method performance char-
acteristics can be determined, including detection limits,
recovery, precision, accuracy, applicability, interferences,
and dynamic range, as appropriate (1).
The type of study or state of the method (e.g., recently
developed/optimized, single-laboratory validated) will dic-
tate which performance characteristics should be included
in the study. For example, method development or optimiza-
tion should generate data to characterize the method detec-
tion limit for the analyte (organism/strain) in the matrix of
interest, sensitivity, and specificity and evaluate interferences
due to the matrix of interest in addition to precision, recovery,
and accuracy. Method development studies may also evaluate
the dynamic range of the method (e.g., determination of the
lower and upper range) by analyses of multiple concentrations
of the analyte.
Method development or optimization studies may include
an evaluation of procedures to minimize interferences, in
addition to generating data to characterize precision, recov-
ery, and accuracy.
Single-laboratory validation studies may be used to opti-
mize methods and ensure that the method is appropriate for
the target analyte/matrix combination for which the method
doi:10.1128/9781555818821.ch2.5.7
2.5.7-1
will be used. Interlaboratory validation studies generally are
designed to generate data to characterize recovery, precision,
and accuracy across multiple laboratories and matrices (e.g.,
water samples collected from different geographical areas).
In general detection limits, interferences, and the dynamic
range of the method is characterized prior to interlaboratory
validation of a method.
REFERENCE MATRIX AND MATRICES
OF INTEREST
It is critical to identify an appropriate reference matrix
for each matrix type analyzed during the study. Ideally, a refer-
ence matrix should be representative of matrix of interest,
have consistent quality, should not contain the analyte of
interest, be widely available, and be inexpensive. Example
reference matrices include sterile reagent-grade water, sterile
buffer, and Milorganite (as a substitute for sewage sludge).
For method performance, reference matrix results are critical
for comparison purposes. Having the appropriate reference
matrix allows the differentiation between laboratory-specific
and matrix-specific causes of unexpected results.
Matrices from geographically diverse locations should
be used for interlaboratory validation studies to ensure that
the method can be used nationwide without matrix interfer-
ence or other issues that would result in unacceptable per-
formance of the test. Environmental Protection Agency
(EPA) studies generally include 10–12 locations for each
matrix (i.e., for fresh waters). For limited in-house use,
matrices should be representative of the matrices that will
be analyzed once the method is implemented. Method com-
parability testing may require as many as 10 geographically
dispersed matrices (2).
SPIKING APPROACH
To evaluate method performance, sample spiking is necessary
to determine recovery and precision. When determining the
appropriate spiking organism, multiple decisions need to be
made. For example, should a culture collection strain (e.g.,
ATCC, NCTC, well-characterized laboratory strain that
has had multiple passages in the laboratory) or wild-type
strain (recently isolated from the matrix of interest) be used
2.5.7-2 ▪ QA/QC IN ENVIRONMENTAL MICROBIOLOGY
to spike samples? Using a laboratory strain ensures that the
spike would be consistent but may not necessarily represent
environmental strains. Although a wild-type strain would
be more representative of what might be expected from the
matrix of interest, once the organism is passed/cultured in
the lab it is considered a laboratory culture. In some cases sew-
age could be used to spike samples. However, using sewage as
spike material does not guarantee that the target organism
levels or characteristics will be consistent among samples or
facilities.
If culture collection strains are used, they should be repre-
sentative of the organism of interest (e.g., similar sensitivities,
growth rate). During method development, multiple strains of
the target organism and near neighbors (similar biochemical
and structural characteristics) should be used to evaluate spe-
cificity and exclusivity to determine false negative and false
positive rates.
Once the organism/strain has been determined, the spik-
ing approach (spike type and level) needs to be determined.
Some common spike types used include commercially avail-
able (e.g., BioBall), referee-prepared, or laboratory-prepared
spikes. There are pros and cons associated with each spike
type. Commercially available spikes can reduce between-
laboratory variability because the spike is at a consistent
concentration and of known quality, is widely available,
and could be used in the future to demonstrate and main-
tain proficiency. However, commercially available spikes
can be expensive and may not be available at the required
concentration or specific strain or in the future. Referee-
prepared spikes limit between-laboratory variability because
all of the labs are spiking with the same material. However,
referee-prepared spikes may not be appropriate because spike
levels may be affected by shipping or may not always be avail-
able for labs to demonstrate and maintain proficiency in the
future. Laboratory-prepared spikes could be prepared at each
lab and could be used to demonstrate and maintain profi-
ciency at low cost and indefinitely. However, it is unlikely
that all of the laboratories are spiking samples at the same level
since the spikes are propagated and enumerated by each lab.
Once the spike type has been chosen, the spike level needs
to be determined. For interlaboratory validation studies, a sin-
gle level, representative of the expected target level (e.g., reg-
ulatory limit) and above the anticipated background levels in
the matrix of interest would be acceptable. However, it may be
necessary to use multiple spike levels to determine detection
limit to ensure the limit is appropriately characterized. For
presence/absence methods, the spike level needs to generate
a mix of positive and negative results (e.g., 40–90% positive).
Regardless of the method (quantitative versus qualitative),
spike levels should be above background levels in the matrices
of interest.
REPLICATES
The number of replicates analyzed during the study also needs
to be determined. The number of replicates analyzed per lab-
oratory during an interlaboratory validation study is depend-
ent on the number of laboratories and the variability of the
method. In general, a multilaboratory study should include
10–12 laboratories, to ensure six valid data sets; each labora-
tory would analyze a minimum of five replicates (one
unspiked and four spiked) per matrix (reference and matrix
of interest) (3). The number of replicates may be determined
based on statistical power analyses, if the estimate of method
variability is available.
QC ANALYSES
Study analyses should be accompanied by standard QC
checks, including any method-specific requirements (e.g.,
sterility checks, positive controls, blanks). If method perform-
ance criteria are available, each laboratory participating in the
study should conduct analyses to demonstrate proficiency
with the method. During a study the frequency and number
of method-specific QC checks may be increased to ensure
that data generated are of good quality. For example, matrix
spike samples may be required for every 20 samples, but may
be conducted with each batch of samples (<20) during the
study. Additional QC checks may also be required that are
study-specific. For instance, if a referee-prepared spike is being
used, analyses of trip controls may be included in the study
design. The trip controls would be analyzed by the referee lab-
oratory to determine if the integrity of the spikes was compro-
mised during shipping.
REFERENCES
1. U.S. Environmental Protection Agency. 1988. Availabil-
ity, Adequacy, and Comparability of Testing Procedures for
the Analysis of Pollutants Established Under Section 304
(h) of the Federal Water Pollution Control Act. EPA/600/
9-87/030. Cincinnati, OH.
2. U.S. Environmental Protection Agency. 2010. EPA Micro-
biological Alternate Test Procedure (ATP) Protocol for
Drinking Water, Ambient Water, Wastewater, and Sewage
Sludge Monitoring Methods. EPA-821-B-10-001. Office
of Water, Washington DC.
3. American Society for Testing and Materials. 1998. Annual
Book of ASTM Standards, vol. 11.01. Philadelphia, PA.
 Water Sampling and Processing Techniques for
Public Health–Related Microbes
VINCENT HILL
2.6.1
Choosing water sampling and processing techniques for
microbiological testing can be a challenge due to the multi-
tude of method options and relative lack of method stan-
dardization. While standardized and validated water testing
methods have been published by groups such as the U.S.
Environmental Protection Agency and in references like
Standard Methods for the Examination of Water and Wastewater
(1), these resources do not necessarily reflect the diversity of
method options available for the multitude of water sampling
applications that environmental scientists, engineers, and
public health officials may need to address. Identifying appro-
priate sample collection and processing techniques requires
an understanding of the target microbes, performance charac-
teristics of alternative methods, and potential water quality
impacts on sample processing and testing procedures.
Before evaluating the advantages and disadvantage of
alternative sampling and testing approaches, investigators
should first clearly establish the goals of their investigations
and associated data needs. These considerations include (a)
environmental sites and media of interest; (b) sampling fre-
quency to meet data needs and statistical testing require-
ments; (c) logistical constraints potentially affecting sample
collection, handling, transport, and storage; (d) analytical
targets of interest and relevance to study goals, including
physicochemical water quality parameters and microbial
analytes; and (e) data output criteria and requirements
(e.g., microbe viability, phenotype, genotype, concentration)
for monitoring, surveillance, epidemiological, modeling, or
other types of investigations. These considerations should
also be evaluated within the context of a sampling design
for data collection (2).
After establishing project goals and a scope of work for
sample collection, investigators should evaluate whether
data can or should be generated using an established “standard
method,” what detection limits are acceptable, and whether
the sample collection and processing procedures are appropri-
ate for the water type targeted and variable quality that may
be encountered. Regulated drinking water systems and recrea-
tional water venues will have established sampling locations,
sampling procedures, and analytical methods that investiga-
tors must use to generate data for regulatory purposes. For
example, in the United States the Environmental Protection
Agency (EPA) requires public drinking water systems to use
EPA-approved analytical methods when analyzing samples
to meet federal water quality monitoring requirements. These
doi:10.1128/9781555818821.ch2.6.1
2.6.1-1
methods are posted on the EPA website, water.epa.gov
(search for “Drinking Water Analytical Methods”). Simi-
larly, for coastal and Great Lakes recreational waters in the
United States, the EPA has promulgated ambient water qual-
ity criteria for enteric bacteria and approved methods for use
under the Clean Water Act (see water.epa.gov, search for
“Approved General-Purpose Methods”). When conducting
investigations of such systems and venues for nonregulated
parameters, it may still be prudent to collect samples using
established procedures to facilitate comparison with routine
water quality data and institutional response actions.
When it is of interest to characterize water quality by test-
ing for physicochemical parameters (e.g., turbidity, pH, disin-
fectant residual, conductivity, temperature) or microbial
indicators of fecal contamination (e.g., bacterial indicators
such as E. coli or enterococci on which federal and state water
quality standards are often based), it is often appropriate and
sufficient to perform discrete sampling of small volumes of
water (typically <500 ml). In fact, field testing (instead of
sample collection for laboratory analysis) is often preferable
for testing water for many physicochemical water quality
parameters, such as temperature, pH, dissolved oxygen, and
disinfectant residual, because the concentration or values of
these parameters can change rapidly after sample collection.
Small-volume sampling may be appropriate versus collecting
large-volume samples when an investigation includes
repeated sampling of a water system (3), given that a larger
number of small samples can be collected. However, when a
lower per sample detection limit is needed, large-volume sam-
ple collection is usually warranted, as either discrete “grab”
samples (up to a per container volume of ∼20 liters) or using
field-deployable sampling techniques. When multiple-grab
samples are collected from the same water body or sampling
site, these samples can be analyzed discretely or combined
to produce composite samples. Composite sampling can ena-
ble sensitive detection of target microbes or increase the
chance of detecting a target microbe whose spatial variability
in the water system may be significant.
Investigations focused on detecting and quantifying
human-infectious enteric pathogens in water often require
the collection of large-volume water samples (e.g., tens to
hundreds of liters) to achieve sufficiently low method detec-
tion limits. This is because enteric pathogens are allochtho-
nous in environmental water systems and typically present
at low concentrations due to factors such as advection,
2.6.1-2 ▪ SAMPLING METHODS
dispersion, and die-off (4). One caveat to this general rule
of thumb is for bacterial enteric pathogens, many of which
(e.g., pathogenic E. coli, Salmonella) can be readily and effi-
ciently cultured from small-volume water samples with mini-
mal processing, given proper culture conditions. Small
volume samples (e.g., ≤1 liter) may also be sufficient when
autochthonous pathogens (e.g., biofilm-associated pathogens
such as Legionella, pathogenic amoebas such as Naegleria fowl-
eri) are the focus of water quality investigations, if it is likely
that the pathogenic microorganism is an established member
of the microbial community (5, 6). However, in general,
large-volume water sampling techniques can be expected to
provide increased detection sensitivity for viruses, bacteria,
and parasites. Concomitantly, there are a multitude of options
available for collecting large-volume water samples, some of
which are designed specifically to capture one class of microbes
(e.g., virus-specific techniques) and others that are effective
at simultaneously capturing multiple classes of microbes.
While large-volume sampling methods can be powerful
tools for enabling sensitive detection of pathogenic microbes,
investigators should be aware of the potential effects that
organic and inorganic water constituents may have of the per-
formance of sampling methods and secondary sample process-
ing techniques, as well as potential effects on downstream
analytical methods. Secondary processing techniques, such
as filtration, sedimentation and precipitation, do not receive
the same level of experimental focus as primary sampling and
analytical techniques, but sample processing techniques are
critical for effectively concentrating water samples to obtain
the small volumes (often <10 ml) needed for analysis by
many common analytical techniques (e.g., PCR, microscopy,
tissue culture). This chapter describes techniques—and
considerations for their selection—that can be used to sample
and process a wide array of environmental water types
(including drinking water, groundwater, surface water, recrea-
tional water, and marine water) to enable testing for viruses,
bacteria, and parasites. Performance and application consid-
erations are addressed for each technique, including sample
volume capabilities, selectivity, microbial recovery efficiency,
robustness, ruggedness, cost, ease of operation, and whether
optimization and validation data are available for the appli-
cation of interest. When multiple methods are available for
the same application, comparisons are made to provide inves-
tigators with information to assist in choosing appropriate
methods for specific applications.
FIELD SAMPLING TECHNIQUES
Discrete Sampling
Small volumes (<500 ml) are often sufficient for bacterial
water quality indicators (e.g., E. coli, enterococci) because
drinking water regulations and ambient water quality criteria
are based on 100-ml sample volumes (in the United States).
Some commercial analytical methods incorporate 100 ml bot-
tles or sterile sampling bags for collecting field samples or
facilitating analysis.
When collecting a small-volume grab sample from a sur-
face water source for enteric microbial testing, the investigator
should collect the sample from within the top 15 cm (6
inches) below the water surface to collect a sample that is
more characteristic of the bulk water body and reflective of
potential human exposure (7, 8). However, there may be
applications when collecting a sample from the water surface
(or at greater depth) is more appropriate, if target microbes are
more likely to be present at those locations within the water
body (9). If collecting a grab sample from water that is flowing
(e.g., a river or stream), the sample bottle should be tilted up
with the opening facing upstream to reduce the potential con-
tamination of the sample by the sample collector. If the water
body is static (e.g., a lake or still area in a river), an artificial
current can be created by moving the sample bottle away
from the collector’s body in the direction that the bottle is
pointed. When the investigator must enter the water body
to collect the sample, care should be taken to minimize dis-
turbance of sediment.
For studies of household ( premise) plumbing and drinking
water distribution system quality, the tap (valve) should first
be opened for at least 2 min to allow stagnant water near
the tap to be flushed from the system, thereby enabling collec-
tion of a sample that is more representative of the water qual-
ity in the system (8). Similarly, when collecting a sample from
a well or borehole equipped with a hand or mechanical pump,
water should be pumped from the well for sufficient time to
purge—two to three well volumes of water to draw water
from the aquifer into the well for sampling. The purge volume
can be calculated using measurements of the well depth, cas-
ing diameter, and depth to water table using the following
equation:
V ¼ pr2 hðcf Þ
where
V = static volume of water in well (in U.S. gallons)
π = pi
r = inner radius of well casing (in feet)
h = length of water column (in feet) which is equal to the total
well depth minus depth to water
cf = conversion (gal/ft3) = 7.48 gal/ft3 (In this equation, 7.48
gal/ft3 is the necessary conversion factor.)
When these parameters are not known, the well or bore
hole can be pumped until field water quality parameters
(such as pH, specific conductivity, temperature) stabilize to
±10% as determined using continuous monitoring (10, 11).
To prevent contamination of a grab sample, the sample
collector should take care to not touch the opening of the
container with their hands or let the sample bottle touch
the spigot or associated piping/tubing.
Small-volume discrete sampling methods are effective
when ∼100 ml analyses are sufficient for a water quality inves-
tigation, but collection of slightly larger volume samples (e.g.,
1–2 liters) can also be effective for bacterial pathogen testing
and coliform testing (12). Containers are available for col-
lecting grab samples of up to 20 liters (e.g., Cubitainers or
equivalent). In general, the sample collection procedures
used for collecting small-volume grab samples can be used
for these larger grab samples, although it is likely easier and
safer to use a 1 liter sterile bottle and a sterile funnel to fill
large containers (e.g., 10–20 liters) that have a narrow mouth
opening. Small-volume grab samples are more appropriate for
enteric bacterial pathogen testing than for enteric viral or par-
asitic pathogen testing because bacteriological membrane fil-
ters can be added directly to culture media (typically broth,
but sometimes agar) used to culture the target enteric bacte-
rial pathogens (13, 14). Commonly used viral and parasitic
pathogen analytical methods (e.g., molecular, tissue culture,
microcopy) do not enable direct analysis or amplification of
pathogens from membrane filters. Direct analysis of viral indi-
cators of fecal contamination (e.g., coliphages) can be per-
formed using membrane filters when the water samples are
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-3
amended to facilitate virus adsorption to the filters, which
can then be placed on agar plates or in culture broth contain-
ing appropriate host bacteria (15). This general technique
of directly culturing viruses absorbed to membrane filters
has also been reported for enteroviruses (16). Relatively
small-volume samples can also be sufficient for investigating
pathogenic environmental bacteria, such as mycobacteria
and Legionella (17, 18) and free-living amoebae (19, 20).
For investigations of water systems that incorporate a filter,
such as swimming pools equipped with sand filters, ∼1 liter
grab samples of filter backwash can be effective for enteric
parasite testing (e.g., Cryptosporidium) (21) and possibly
also enteric bacteria (22), because the treatment system filters
concentrate (oo)cysts and similar sized particles.
One important factor that can affect the analysis of grab
samples in the general range of 500 ml to 2 liters is the poten-
tial effect of turbidity on sample processing. Bacteriological
membrane filters (0.2–0.45 µm) are subject to clogging, espe-
cially when processing lower quality water samples, such as
lake and river water. When clogging occurs, multiple mem-
brane filters can be added to the same volume of culture
broth, but there is a practical limit to this approach. More
often when clogging is anticipated, it may be prudent to use
a prefilter to remove large, nontarget particles or to employ
a centrifugation-based procedure to concentrate target
microbes in the sample by sedimentation. Prefilters (e.g., 10
µm pore size) are sometimes used by investigators when large-
volume samples are collected from relatively low-quality
water sources (e.g., surface water, wastewater) (23) or to sep-
arate microbes based on size difference (24). Sedimentation
is appropriate to concentrate bacteria and parasites, and is
generally applied to samples ≤1 liter for practical, time effi-
ciency reasons. Sedimentation is not effective for viruses
unless they are precipitated or otherwise attached to larger
particles (see section below on Virus Sedimentation by Pre-
cipitation and Flocculation).
Composite Sampling
Rivers, lakes, and other water bodies can exhibit high spatial
and temporal variability in microbial concentrations (7).
Composite sampling can be an effective way to account for
this variability. Composite sampling involves physically com-
bining and homogenizing environmental samples to form a
new sample (i.e., a composite sample) that is representative
of the larger population of samples. The chemical or biolog-
ical analyses of interest are then performed on the composite
sample. Composite sampling is often used in food testing
when samples from a produce field or production line are
composited prior to testing. Spatial and temporal variability
in rivers and lakes can be accounted for by collecting a series
of grab samples for compositing (25) or by periodically pump-
ing water through a filter that is used to capture microbes to
create a composite sample. Composite sampling has also
been used for microbiological studies of lakes to increase
the spatial coverage of the microbiological surveys across areas
for which data were not needed at each individual sample
location (6, 26). Because compositing physically averages
the individual samples, averaging the analytical results of a
few composites can produce an estimated mean that is as pre-
cise as one based on many more individual sample results (7).
Since fewer analyses are needed, composite sampling can sub-
stantially reduce study costs when analysis costs are high rel-
ative to the costs associated with the collection, handling,
and compositing of the samples.
Large-Volume Sampling
When water is contaminated with pathogens, even low
concentrations may be of concern as an infection risk for
exposed populations. Due to factors such as sampling delay,
microbial die-off, and dilution, it is often the case that human
pathogens and fecal microflora are present in environmental
waters at low concentrations that preclude the effective use
of grab sampling for water testing. Thus, it may be desirable
to use methods for concentrating large volumes of water
to enable detection and quantification of pathogens and
other microbes. Alternative large-volume sampling techni-
ques include physical separation techniques (e.g., microfilter
cartridges, continuous flow centrifugation, ultrafiltration)
and adsorption-elution techniques (used primarily for virus
recovery).
Physical Separation Techniques Specific to Parasites
Microfiltration is the most common technique for recovering
parasite cysts and oocysts from large-volume water samples
because (oo)cysts are relatively large compared to bacteria
and viruses (Fig. 1). The U.S. EPA has developed standar-
dized protocols (Method 1622 and 1623.1) for recovering
parasite (oo)cysts from drinking water and surface water sam-
ples using alternative microfilter cartridges (Pall Gelman
Envirochek HV and IDEXX Filta-Max filters) (27, 28).
Method 1623.1 can also be performed using continuous
flow centrifugation. While these methods were developed
for the recovery and detection of Cryptosporidium and Giardia
(oo)cysts, the filtration techniques have also been used to
recover Toxoplasma gondii (29, 30), Cyclospora cayatenensis
(31), and free-living amoebae (32) from water and wastewater
samples. The Envirochek filter has also been used to recover
Cryptosporidium and Giardia (oo)cysts from reclaimed water
(33, 34) and marine water (35). Microfilter cartridges
designed for parasite recovery have nominal pore sizes of
∼1 µm and thus are not designed for recovery of bacteria or
viruses by physical size exclusion. As with other large-volume
cartridge filtration techniques, microfilter cartridges can be
connected directly to a tap, if the piped system is under suffi-
cient pressure to force water through the filter, or connected
to a pump to collect the sample. According to Method
1623.1, samples can be filtered at 2 liters/min when using
Envirochek HV filters or 1–4 l/min when using Filta-Max
filters.
Continuous flow centrifugation (CFC) has been used pri-
marily as a laboratory technique for processing grab samples
for recovery of microbes from water samples—primarily to
recover parasites (36, 37), but the technique can also recover
bacteria (38). Recently, portable CFC instruments have been
developed that enable use of this technique to collect large-
volume water samples in the field (39). Because the techni-
que utilizes sedimentation to separate microbes (bacterial
sizes and larger) from water samples, the technique can theo-
retically be applied to a wider range of water or wastewater
samples than can comparative filtration techniques that are
more subject to clogging. Portable CFC instruments use a dis-
posable Latham bowl in which microbes are sedimented.
Because allowing sufficient residence time in the Latham
bowl is critical for effective microbial capture by sedimenta-
tion, the flow rate through the centrifuge must be balanced
with the centrifugal force applied (40). For recovery of proto-
zoan parasite (oo)cysts, water samples can be passed through
a portable CFC (Fig. 2) at flow rates of ∼1 liter/min when
the centrifuge is operated to apply 2,500 × g (40). If using
the portable CFC technique for recovery of smaller microbes,
2.6.1-4 ▪ SAMPLING METHODS

FIGURE 1 Range of filtration processes for collecting microbes and other materials by physical size exclusion.
doi: 10.1128/9781555818821.ch2.6.1.f1

FIGURE 2 Schematic of continuous flow centrifugation device (Courtesy of: Scientific Methods Inc).
doi: 10.1128/9781555818821.ch2.6.1.f2
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-5
such as microsporidia, a lower flow rate (e.g., 0.75 liter/min)
and higher centrifugal force (e.g., 3,000–4,000 × g) is needed
to effectively capture the microbes from a water sample
(39). Mean recoveries of Cryptosporidium and Giardia (oo)
cysts have been reported to be in the range of 47–71% for
10 liter source water samples and 35% for 1,000 liter tap water
samples (Cryptosporidium only) when processed by portable
CFC, including secondary concentration (39). After process-
ing the target water sample volume using the CFC procedure,
the centrifuge is operated for a few minutes to reduce the water
volume in the bowl to <250 ml. Then a surfactant-based sol-
ution (e.g., 5 ml of solution containing 1% sodium dodecyl
sulfate, 0.01% Tween 80, and 0.001% antifoam A) is injected
into the bowl. Using a wrist shaker, the bowl is vigorously
shaken for ∼20 min in several orientations to loosen sedi-
mented material from the Latham bowl surface. The bowl is
then inverted and the eluate drained into a 250 ml centrifuge
tube for further sample concentration (e.g., according to the
procedures described in Method 1623.1).
Ultrafiltration Techniques
Physical separation of microbes from water can also be per-
formed using ultrafiltration, an operational term for filters
having pore sizes defined in terms of molecular weight cut-off
in the range of 5,000 Da to 100,000 Da (∼4–50 nm) (Fig. 1).
Because of their small pore sizes, ultrafilters can be used to
recover viruses (41), bacteria (42), parasites (43), and even
large toxins (44) from water by size exclusion, making
FIGURE 3 Schematic of tangential flow hollow fiber ultrafiltration (Reprinted from ref. 48, with permission).
doi: 10.1128/9781555818821.ch2.6.1.f3
ultrafiltration (UF) a feasible technique for simultaneously
recovering multiple microbe types from the same water
sample (45, 46). UF has been effectively applied to a wide
range of water types, including drinking water, ground-
water, surface water, and marine water (29, 47–49). Ultrafilter
membranes relevant to water sampling are generally manu-
factured as flat sheets or hollow fibers, and are typically
composed of hydrophilic materials (e.g., polysulfone, cellu-
lose triacetate). Some tangential flow UF procedures have
been reported for water sampling using flat sheets held in
stackable cassettes (47, 50), but these cassettes are expensive
and need to be cleaned for reuse. More often, hollow-fiber
ultrafilters have been employed for large-volume water sam-
pling, using both the tangential flow UF approach since
the 1980s (49, 51) as well as dead-end UF more recently
(52–54). For this reason, only hollow fiber UF (HFUF) tech-
niques are described in detail in this chapter. Hollow
fiber ultrafilters produced for the hemodialysis industry have
become popular because their relative low cost makes it
feasible to dispose of them after a single use, thereby reduc-
ing labor time and cross-contamination concerns related to
filter reuse. Numerous hollow fiber ultrafilter sources are
available (52, 55), and these are generally interchangeable
because of design elements (e.g., standardized input and out-
put ports and side ports) that are common in the hemodialysis
industry.
Tangential (i.e., cross-flow) HFUF is performed by recir-
culating a water sample through the ultrafilter cartridge under
pressure using flexible tubing and a peristaltic pump (Fig. 3).
2.6.1-6 ▪ SAMPLING METHODS
Backpressure is created using a clamp to decrease the cross-
sectional tubing area through which the water is pumped.
This backpressure results in increased flux across the filter
membranes and higher filtrate rates as filtered water exits the
ultrafilter cartridge through one of the side ports. The more
the tubing is clamped, the greater the backpressure, and the
higher the filtrate rate. To enable effective tangential flow
through the ultrafilter cartridge, the flow rate is typically
balanced using the tubing clamp so that ≥50% of the flow
from the peristaltic pump exits the ultrafilter cartridge through
the end port (as cross-flow) versus through the side port (as
permeate/filtrate). If the tubing clamp is completely closed,
then all flow from the pump is forced through the filter fibers
and exits the side port as filtrate. This constitutes the dead-
end UF technique described later in this chapter. The
≥50% cross-flow rule of thumb is used to ensure that sufficient
water is moving across the membranes to enable the scouring
effect to reduce adherence of microbes and other particles to
ultrafilter fiber surfaces (thereby reducing potential fouling),
as suggested by various researchers (46, 50, 56, 57). Some
research has indicated that pretreatment of ultrafilters may
improve microbial recovery during tangential flow (TF)
HFUF by reducing microbial adhesion to ultrafilter fibers.
Such pretreatments have included use of calf serum (50, 58)
and chemical dispersants (45). Addition of chemical disper-
sants (e.g., sodium polyphosphate) (45, 59) prior to UF has
also been reported to significantly increase microbial recov-
eries for some applications, but use of such reagents may
not be appropriate when analyzing UF concentrates for cul-
ture of some bacteria (60). Desorption of microbes that do
adhere to filter fibers can be achieved through filter elution
or backflushing.
When assembling the set-up for TF HFUF, alternative
configurations may be used. In its simplest conformation, the
recirculation system is “open” and the water sample is recircu-
lated back into the sample container, from which the con-
centrated sample (retentate) is collected (Fig. 3) (45). For
applications in which potential exposure to aerosols during
UF is a biosafety concern, an intermediary bottle outfitted
with a vented cap can be included in the recirculation loop,
and the retentate collected in this “closed system” container
(59, 61). For additional biosafety protection, the procedure
can be performed in a biological safety cabinet (62). An auto-
mated TF HFUF device has also been developed using this
closed system design to facilitate performing TF HFUF in
the field (63). Using either an open or closed configuration,
the recirculation process is performed until the water sample
is concentrated to <500 ml. At this point, the retentate is
recovered and set aside. To improve microbial recoveries, elu-
tion or backflushing should be performed to recover microbes
from the ultrafilter that adsorbed to the filter fibers. Both elu-
tion (46, 58) and backflushing (45) have been shown to
significantly increase microbial recoveries for UF. While
no clear microbial recovery advantage has been identified
between these approaches, backflushing does enable both
chemical desorption and well as physical removal because
of the reverse flow through the filter fibers. After elution or
backflushing, the eluent (or backwash) is added to the reten-
tate prior to analysis or subsequent processing to further
concentrate the sample. After secondary processing, UF-
concentrated samples can be assayed using culture, micro-
scopy, and molecular techniques, but users should consider
that UF-concentrated samples can be inhibitory to certain
assay types, especially reverse transcription-PCR (RT-PCR)
(48, 49, 64). Nucleic acid extraction techniques and molec-
ular assay procedures designed to minimize and monitor
inhibition from water samples should be used when assaying
UF concentrates (49, 54, 65).
Dead-end ultrafiltration (DEUF) has been proposed as a
simpler, field-deployable alternative to TF HFUF methods
for recovering microbes from large-volume water samples
(52, 54). The DEUF technique differs from TF HFUF techni-
ques in that one of the ultrafilter end ports is plugged so that
the water sample cannot exit the ultrafilter through the end
cap (as occurs with TF HFUF methods) but instead must
pass perpendicularly through the hollow fiber membranes
(Fig. 4). Thus, particles and other water sample constituents
larger than the pores in the ultrafilter membranes are trapped
within the ultrafilter cartridge, as opposed to TF HFUF tech-
niques which result in the production of a concentrated reten-
tate sample (45). The DEUF method is simpler to perform
than tangential flow recirculation UF techniques because the
water sample passes through the ultrafilter cartridge only once
(as occurs with other single-pass filtration methods such as
U.S. EPA Method 1623.1 filtration for parasite [oo]cysts using
Envirochek filters [27] or virus adsorption-elution filtration
using NanoCeram filters [66]). When DEUF is completed,
the ultrafilter cartridge can be capped and the ultrafilter
shipped to a laboratory for recovery of microbes using either
backflushing or elution. TF HFUF is less easily performed
in the field, as it requires the water sample to be present in
a container (in which the water sample is recirculated), unless
an automated system is used such as has been developed by
the U.S. EPA (63). For large-volume water samples, multiple
containers must be filled for processing using manual TF
HFUF methods, whereas water sampled by the DEUF method
can be pumped directly from the water body through the
DEUF ultrafilter (Fig. 4a) or, if sampling a pressurized system,
the system pressure can be used to force water through the
DEUF ultrafilter without need for a pump (Fig. 4b). Flow rates
through an ultrafilter using DEUF can be 2–4 liters/min
and possibly higher. Once a water sample has been filtered
using DEUF, further sample processing can be performed as
described for TF HFUF methods.
Adsorption-Elution Techniques
Adsorption-elution techniques are usually applied specifi-
cally for virus recovery from water samples because most
viruses are negatively charged in environmental waters,
with isoelectric points below pH 7 (67, 68). While these
techniques can recover bacteria and parasites (60, 69, 70),
adsorption-elution filters and techniques are typically
designed for virus recovery because microfilters are available
for recovering bacteria and parasites based on size exclusion.
Virus adsorption-elution (VIRADEL) can be performed by
direct adsorption of virions to positively charged membranes
or by virion adsorption to negatively charged membranes
using multivalent cations (e.g., Mg2+, Al3+).
Positively charged microporous membranes have been
available since the late 1970s for recovering viruses from
water (71). Such filters are composed of media that has a rel-
atively higher positive charge than standard microfilters at
typical pH values for environmental water and drinking water
(72). As the water sample passes through the filter, the effi-
ciency of virion adsorption can be affected by various factors,
including water quality ( pH, ionic strength, organic content)
and virion surface characteristics. Virus adsorption to electro-
positive filters is generally strongest at pH ≤ 7.5 (72, 73),
although filters incorporating nano alumina (AlOOH) have
been reported to have similar adsorption effectiveness at
pH values between 6 and 9.5 (66). Thus, electropositive fil-
ters have an advantage over electronegative filters in that
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-7
FIGURE 4 Schematic of dead-end ultrafiltration for filtering water from (a) nonpressurized water bodies and (b) pressurized water systems.
doi: 10.1128/9781555818821.ch2.6.1.f4
pH adjustment (or other sample amendment) is often not
needed, which facilitates filtration of large-volume samples.
If pH adjustment is needed, the pH can be lowered by inject-
ing low-strength acid (e.g., 1 N HCl) with a high-pressure
peristaltic pump (73). Figure 5 illustrates the relatively simple
set-up as described in U.S. EPA Method 1615 for performing
the VIRADEL technique using a NanoCeram alumina nano-
fiber filter with optional injector and/or prefilter points
indicated (74). U.S. EPA Method 1615 was developed to
recover enteroviruses and human noroviruses from finished
drinking water, groundwater, surface water, and treated sew-
age effluent at water sample filtration rates of up to 10 liters/
min. The method does not address marine water sampling.
The high ionic strength of marine waters has been identified
as a significant issue for the VIRADEL technique using elec-
tropositive filters such as 1MDS and glass wool (75), and
these filters are generally not used for marine water sampling.
However, studies of alumina nanofiber electropositive filters
(e.g., NanoCeram, ViroCap) indicate that such electroposi-
tive filters can be effective for recovering viruses from marine
water (76, 77). For applications in which water samples may
contain high organic content (e.g., surface water), studies
indicate that soluble organics (e.g., humic acid) can interfere
with virus adsorption to electropositive filters and result in
low recovery efficiencies (78, 79).
While charge-modified glass and cellulose medium filters
(e.g., Zeta Plus 1MDS) filters have been the standard for
decades for recovering viruses from fresh water and drinking
water—including being designated by the U.S. EPA for
determining total culturable viruses in drinking water under
the Information Collection Rule (80)—cost-effective alter-
native electropositive filters are becoming more popular.
The U.S. EPA has investigated the NanoCeram alumina
nanofiber filter as a potential alternative to 1MDS filters for
2.6.1-8 ▪ SAMPLING METHODS
FIGURE 5 Schematic of filtration set-up using an electropositive cartridge for recovery of viruses from water (Reprinted from ref 74).
doi: 10.1128/9781555818821.ch2.6.1.f5
ICR-like water quality investigations (66, 81). The glass wool
filtration technique was first reported in 1993 (82) and is
commonly used in European countries for freshwater viro-
logical quality studies, including a recent evaluation of fresh-
water recreational water quality (83). To perform the
technique, glass wool (which is essentially fiberglass, although
an oiled sodocalcic glass wool product is typically used) is
manually packed into a column to a specified density (73).
The column is then connected to a pump or pressurized water
source to collect a filtered water sample. Depending on col-
umn size and glass wool packing density, flow rates of 0.2 to
0.5 liter/min have been reported (73, 84). As with other
VIRADEL methods, viruses adsorbed to glass wool filters
must be eluted from the filters, typically using an organic sol-
ution (such as beef extract) at high pH (e.g., 9.5). In addition
to use in European countries for freshwater sampling, the glass
wool technique has been used in the United States to inves-
tigate groundwater quality (73, 85). However, the major
factor limiting widespread uptake of the glass wool technique
is that premade glass wool cartridges are not commercially
available, so the filters must be manually produced by packing
stainless steel or plastic holders with glass wool material
(73, 83).
As with electropositive filters, electronegative filters have
been used for decades for recovering viruses from water (86).
The effect of salts in promoting enteric virus attachment to
membrane filters was described as early as 1967 by Wallis
and Melnick (87), who also noted that proteinaceous media
and surfactants can be effective for eluting viruses from filters.
Through the 1980s and 1990s the electronegative VIRADEL
technique was applied to freshwater samples using sample
acidification to ∼pH 3.5 with or without the addition of mul-
tivalent cations (in MgCl2 or AlCl3 salts) to create a “salt
bridge” between virus particles and membrane filters (75,
79, 88). When the electronegative VIRADEL technique is
applied to seawater, the ionic strength is high enough such
that viruses effectively adsorb to nitrocellulose membrane fil-
ters without need for added salts, and the method can be per-
formed either with sample acidification (83, 89) or without
sample acidification (90).
After filtering the water sample with an electronegative
filter, viruses may be eluted using a wide array of proteina-
ceous eluents, including beef extract (1–3%, pH 8.0–9.5)
(91, 92), skimmed milk ( pH 9.5) (93), or amino acids (e.g.,
0.5 M glycine) (86). However, some proteinaceous eluents,
such as beef extract, have been associated with inhibition of
molecular assays (90, 94). An alternative approach reported
by the research group of Haramoto and Katayama for electro-
negative VIRADEL and sample preparation for RT-PCR is to
use an acid rinse (H2SO4) to remove multivalent cations
and environmental inhibitors from the membrane filter,
followed by virus desorption using NaOH (95). To perform
their “cation-coated filter method,” 5 ml of 250 mM AlCl3
is passed through a 0.45 µm, 90 mm electronegative filter to
coat the filter in multivalent cations. Then the water sample
is passed through the filter. When using MgCl2, a 2.5 M
MgCl2 solution is added 1:100 to a water sample (e.g., 5 ml
added to a 500 ml sample) to obtain a final concentration
of 25 mM, then the sample is passed through a electronega-
tive membrane filter. For both approaches, the membrane fil-
ter is rinsed with 200 ml of 0.5 mM H2SO4 ( pH 3.0), and
then viruses are eluted using 10 ml of 1.0 mM NaOH ( pH
10.8). The filtrate is captured in a tube containing 50 µl of
100 mM H2SO4 ( pH 1.0) and 100 µl of 100 × Tris-EDTA
buffer ( pH 8.0) for neutralization. After this point, further
sample concentration can be accomplished using a micro-
concentrator (e.g., Centriprep YM-50; Millipore) (96). This
method has been effective for enabling detection of enteric
viruses in seawater, drinking water, surface water, and sewage
(97, 98).
Quality Control for Water Sample Collection
Preservation, Handling, and Transport
A chain-of-custody form should be used to record sample col-
lection information (e.g., date, time, site description, GPS
coordinates), appropriate physicochemical field data (e.g.,
turbidity, temperature, pH, conductivity, disinfectant resid-
ual, dissolved oxygen), sample handling issues ( preservation
[temperature, dechlorination], holding times), and names of
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-9
personnel handling, receiving, and relinquishing samples.
Sample preservation for microbial water quality investiga-
tions is important to preserve sample integrity by stabilizing
microbes and minimizing the potential for their degradation
during sample handling, transport, and storage. Best practice
requirements and conditions for preserving water samples for
microbial testing depend primarily on the detection assays
planned for testing the samples.
Quality control considerations that should be addressed
prior to collecting water samples for microbial testing include
temperature control, chemical preservation, holding times,
container types, and the need for blanks and duplicates
(99). Samples for microbiological testing should always be
collected in sterile containers fitted with screw caps or other-
wise sealed in a way that prevents sample leakage. Cartridge
filters used for large-volume water sample collection should
be placed in sealed containers that prevent leakage from the
sample into the storage/shipping container and which pre-
vent sample contamination during storage and/or transport.
Double bagging may be prudent to provide secondary con-
tainment for a filter sample. Water and filter samples
collected for microbiological testing should be iced or refri-
gerated (but never frozen) to maintain samples at a tempera-
ture of 2–8°C, especially when bacteriological or other
culture testing is planned. However, there may be some spe-
cial applications, such as testing for pathogenic free-living
amoebae, mycobacteria, Vibrio cholerae, and Legionella, for
which samples should not be chilled (100, 101). For bacter-
iological testing of water samples, the U.S. EPA recommends
that testing should be initiated within 6 h of sample collec-
tion (102); drinking water must be assayed within 30 h after
sample collection under the Total Coliform Rule. Water sam-
ples collected for nonculture analyses (e.g., microscopy or
molecular testing) should still be stored and transported
under chilled conditions (2–8°C) and shipped to the analyt-
ical laboratory within 24 h as a best practice (27). When
residual chlorine or monochloramine is detected or suspected
in a water sample, the sample should be dechlorinated during
or immediately after collection using sodium thiosulfate
(or equivalent dechlorinating reagent) to avoid potential
false-negative culture results (103) or molecular testing results
(104). Sodium thiosulfate added to a final concentration of
10 mg/liter (i.e., 1% w/v) is sufficient for dechlorinating
tap water and swimming pool water samples containing typi-
cal chlorine levels (≤4 mg/liter as Cl2). Sodium thiosulfate
can be added to empty sample bottles in powder form before
samples are collected or added as a liquid (e.g., from a 10%
solution) after sample collection.
When planning an environmental sampling project it
is important to consider what quality control measures should
be (or are required to be) taken. This is especially important
when data from environmental sampling may be used for reg-
ulatory or legal purposes. Quality control measures may
include collecting duplicate samples, including sample blanks
and/or spiked samples during sample shipment, and monitor-
ing the temperature in shipping containers (e.g., using tem-
perature loggers) (105).
LABORATORY SAMPLE PROCESSING
TECHNIQUES
Direct Sedimentation by Centrifugation
Centrifugation is an often used sample processing technique
in environmental microbiology laboratories and can be used
to concentrate microbes from samples on the order of 1 liter
down to a few milliliters. Centrifugation procedures have an
advantage over filtration procedures in that they are not
physically constrained by water quality (as filters are when
they clog), but they are practically limited by the volume of
sample that can be accommodated by standard centrifuges
and rotor types. For example, whereas 1 liter of a relatively
clean water sample may be processed through one membrane
filter, this same volume may require multiple centrifuge bot-
tles. Consequently, centrifuging many large-volume samples
can consume a substantial amount of analyst time because
the centrifuging process can become a bottleneck in the pro-
cedure. When preparing to centrifuge water samples to con-
centrate microbial particles, it is important to understand the
specifications of the centrifuge and rotor being used, as well as
the sedimentation characteristics of the target microbes (e.g.,
size, sedimentation coefficient) and water sample (e.g., vis-
cosity, temperature). Most environmental microbiology labo-
ratories use centrifuges equipped with a swinging bucket rotor
to concentrate microbes in the tip of centrifuge tubes. Com-
mercial centrifuge tubes for water sample processing are typi-
cally conical and available in many sizes (e.g., 15, 50, 200,
and 500 ml). To give support to the tubes during centrifuga-
tion, adapters having the appropriate size and shape are
inserted into the centrifuge buckets. It is important to under-
stand that centrifugation conditions specified by one labora-
tory for an application may not be as effective in another
laboratory that uses a different centrifuge and rotor. First, it
is important to understand that relative centrifugal force
(RCF, reported as ×g) is the primary physical factor affecting
particle sedimentation rates. Different rotors may need to be
operated at different rotor speeds (reported as rpm) to achieve
the same RCF. The RCF for a rotor can be determined using a
nomogram (en.wikipedia.org/wiki/Nomogram), given the
radius of rotation (in millimeters) and rpm. The radius of rota-
tion will differ from the top of the centrifuge tube to the bot-
tom; the midpoint of the tube is typically used to calculate
RCF using the common usage formula:
RCF ¼ 11:17rðrpm=1,000Þ2
where
r = radius to tube midpoint (in cm).
For any particle, the time needed to completely sediment
the particle can be determined using the equation:
t ¼ ððD  LÞ=ðD þ LÞÞ ðN=ðd2 ðg  pÞv2 ÞÞ
where
t = time in min,
D = radial distance in cm for rmax,
L = radial distance to meniscus,
N = viscosity of the water sample,
p = density of the particle to sediment,
d = diameter of the particle in centimeters, and
v = rotational velocity in rpm.
However, for many applications these inputs are not
known or the data are not readily available. In these cases,
a rule of thumb approach can be used based on the scientific
literature. For sedimentation of parasite (oo)cysts, 15 min of
centrifugation at 1,500 × g has been reported for sedimenting
Cryptosporidium and Giardia (oo)cysts in 250 ml conical tubes
(27, 106). For sedimenting bacterial cells, 30 min of centrifu-
gation at 4,000 × g has been reported for E. coli cells in 150 ml
2.6.1-10 ▪ SAMPLING METHODS
samples (42), while 30 min at 15,000 × g has been reported
for sedimenting bacteria in 250 ml sample volumes (107).
Centrifugation at an RCF of >8,000 × g has been suggested
as a rule of thumb for sedimenting bacteria in complex sample
matrices (108).
When a centrifugation procedure is reported with a spe-
cific time and RCF (e.g., 15 min at 4,000 × g), the maximum
RCF for each rotor can be used to convert the procedure from
one rotor to another:
t1 ¼ ðt2  RCF2 Þ=RCF1
where t1 is the run time in procedure 1 and t2 is the run time in
procedure 2; RCF1 is the RCF in procedure 1 and RCF2 is the
RCF in procedure 2.
However, if k factors are available for each rotor, then the
conversion is simply:
t1 ¼ ðk1  t2 Þ=k2
In addition to using sufficient time and RCF to effectively
sediment target microbes in water samples, centrifugation
procedures should also consider the recovery of microbes
from centrifuge tubes. Microbes have a tendency to adsorb
to surfaces, and this tendency is likely exacerbated when con-
centrating microbes in centrifuge tubes at high g forces. To
minimize adsorption, a nonionic surfactant (e.g., Tween 80)
can be added to water samples at a final concentration of
0.1–1% prior to centrifugation (106). When centrifuging fil-
ter eluents, these typically already contain surfactant or
organic compounds that reduce microbial adsorption to cen-
trifuge tubes (27, 59). When centrifuging water samples for
bacterial recovery and culture, it is important to avoid using
sample additives that may inhibit culture. When removing
supernatant, gentle aspiration is typically the best approach
to minimize disturbance of the pellet and loss of microbes.
There is a trade-off between minimizing the resuspended pel-
let volume and the risk of losing microbes in the removed
supernatant. There are no general rules to suggest for indicat-
ing how close to a pellet the supernatant can be drawn down
to, as various factors (e.g., centrifuge tube size, aspiration
instrument/suction, pellet stability) are application-specific.
Training in aspiration and pellet resuspension techniques
should be completed under the guidance of an environmental
microbiologist with expertise in these procedures.
Centrifugation can also be used to concentrate viruses in
water samples, but a high-speed centrifuge is required because
of the small size of virus particles. For direct (unassisted) sed-
imentation of viruses, use of an ultracentrifuge is needed to
achieve g forces on the order of 100,000 × g (109–111). Ultra-
centrifugation has generally been applied to wastewater
samples or to water sample processing as a secondary concen-
tration technique after primary concentration using a filtra-
tion method (112). Because ultracentrifuges are relatively
expensive, require specialized training and maintenance, and
are limited to processing relatively small volumes of water
(e.g., ≤30 ml), they are not commonly used to concentrate
viruses in environmental samples. Many water sample filtra-
tion techniques result in eluent and water concentrate vol-
umes on the order of hundreds of milliliters, which are too
large to practically process using ultracentrifugation. Ultra-
centrifugation has been effectively used to recover viruses from
small volume samples, especially sewage samples. Girones
et al. reported effective recovery of adenoviruses in 30 ml sew-
age samples after performing ultracentrifugation at 48,400 × g
for 3 h 45 min at 4°C (113). Sylvain et al. used ultracentri-
fugation to recover bacteriophages and poliovirus-1 from
water sample concentrates produce using TF UF. The TF
UF procedure resulted in water concentrates having volumes
of ∼180 ml. The entire sample concentrate volume was fur-
ther concentrated using ultracentrifugation at 158,000 × g
for 1 h 30 min) (114).
Virus Sedimentation by Precipitation
and Flocculation
Viruses can also be concentrated using more commonly avail-
able centrifuges, including floor model centrifuges capable of
reaching RCF values of ∼12,000 × g and benchtop models
capable of reaching RCF values of ∼4,000 × g, when virus
sedimentation is aided by precipitation or flocculation tech-
niques. Such techniques include use of polyethylene glycol
(PEG) precipitation, precipitation using metallic salts (e.g.,
aluminum hydroxide, ferric chloride), and organic floccula-
tion (using reagents such as celite or skimmed milk). Precip-
itation techniques are typically performed on volumes ≤1
liter, but some researchers have reported using organic floccu-
lation for volumes as large as 20 liters. PEG precipitation has
been a standard technique for precipitating viruses for deca-
des (115–117). For water samples, PEG precipitation has
been typically performed using PEG 8000 at concentrations
of 7% to 12% w/v and sodium chloride (0.2–0.9 M) at pH
7.2–7.4 (59, 73, 118, 119). For relatively low organic water
samples (such as drinking water and groundwater) that are
not concentrated by a method in which an organic eluent
(such as beef extract) is used, some researchers have added
1% bovine serum albumin to aid in floc development during
the PEG precipitation process (41, 59). After PEG precipita-
tion reagents are added and dissolved, samples are incubated
at 4°C for as few as 2 h (59, 118) or overnight (73, 119, 120).
After incubating PEG-precipitated samples, 100–200 ml vol-
umes are centrifuged at relatively high RCF values, typically
between 7,000 and 12,000 × g for 30 min (59, 118, 119),
but possibly as low as 4,200 × g for 45 min (73).
Precipitation using metallic salts is based on traditional
drinking water treatment methods that use chemicals such
as alum and ferric chloride to remove viruses and other par-
ticles from drinking water by coagulation and flocculation
(121). Ferric chloride, aluminum chloride, and aluminum
hydroxide have been frequently reported for recovery of
viruses from water. The aluminum hydroxide precipitation
procedure is described in detail in Standard Methods for
the Examination of Water and Wastewater (Method 9510
D) (1). When using aluminum chloride, it is added to the
sample (say, a 1 liter volume) to achieve a concentration of
0.9 N AlCl3 at pH 6 (1, 122). The sample is then stirred
slowly for 15 min at room temperature to allow flocs to form.
Precipitated viruses are then sedimented by centrifugation at
3,500 × g for 20 min. For virus precipitation using ferric
chloride, ferric chloride can be added to a concentration of
250–405 mg/liter (26, 121). According to procedures out-
lined by the U.S. EPA and others, ferric chloride precipita-
tion of beef extract eluents should be performed at pH 3.5,
with sedimentation of the floc by centrifugation at 2,500 × g
for 15 min (26, 123, 124).
Filter eluents containing beef extract can also be precipi-
tated using PEG (described previously) or organic floccula-
tion. Organic flocculation was first used for virus recovery
in the 1970s (125) and then became a standard procedure
for the U.S. EPA (80) and in Standard Methods for the
Examination of Water and Wastewater (Method 9510 C)
(1). The organic flocculation technique was developed to
concentrate viruses in beef extract eluents used to recover
viruses from municipal sludges and positively charged filters.
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-11
When applied to beef extract eluates, the pH of the eluates is
lowered to 3.5. At this pH value, organics in the eluate precip-
itate and form flocs, which can then be separated from the
sample by centrifugation. In the 1980s, the effectiveness of
the organic flocculation procedure was found to be affected
by the composition/source of the beef extract used in the pro-
cedure (92, 126). To perform the classical organic floccula-
tion procedure, the pH of beef extract eluent is lowered to
3.5 (e.g., using 1 N HCl) and the sample is stirred at room
temperature for 30 min to form flocs (127, 128). Flocs are
sedimented by relatively low-speed centrifugation (e.g.,
2,500 × g for 15 min or 3,000 × g for 10 min), and the pel-
leted material is suspended in phosphate buffer (e.g., 0.15
M NasHPO4, pH 7.5). While the classical procedure became
standardized for secondary concentration of viruses from
water samples, the virus recovery efficiency of organic floccu-
lation has been reported to be highly variable in numerous
studies, and the low pH used in the method has been
associated with viral inactivation (128, 129). However, the
procedure is still included in Standard Methods for the
Examination of Water and Wastewater and has recently
been incorporated into a standardized U.S. EPA method for
recovery of enterovirus and norovirus in water (74).
To improve the effectiveness of the organic flocculation
technique, various materials have been investigated as addi-
tives to improve virus recovery. The modification that has
been most influential is the use of celite filter aid, a diatoma-
ceous silica product (130). Since its reporting in 1986, the
celite procedure is now performed when nonflocculating
beef extract (e.g., BBL beef extract or equivalent) is used dur-
ing primary concentration of the viruses from a water sample.
In the current method, celite is added to 100 ml beef extract
eluent samples at 0.1 g/100 ml. When the celite is dispersed
evenly, the pH is adjusted to 4.0 by slowly adding 1 N HCl
(131). After stirring for 10 min, celite particles can be col-
lected by centrifuging at 2,500 × g for 15 min (130) or by
passing the sample through a 42 mm, 2 µm prefilter (e.g.,
AP-20 filter from Millipore) using vacuum filtration (131).
In recent reports, the filtration approach has been favored
(66, 131). After filtration, the collected celite particles are
then washed using 5 ml of sodium phosphate or PBS at pH
9.0 to elute viruses from the celite. McMinn et al. found
that PBS was more effective than sodium phosphate for virus
recovery and avoiding PCR inhibition (131). These research-
ers also reported that calcinated-powder celites having fine to
medium particle size were associated with the highest recov-
ery of adenoviruses. These researchers reported average
adenovirus recoveries using the modified celite method that
were comparable to the organic flocculation method. In
general, for secondary concentration of viruses in beef extract
eluents no consistent differences in recovery efficiency or
method detection limits have been determined in studies
comparing PEG precipitation, organic flocculation, or the
celite method (118, 131).
Organic flocculation has also been applied to large-
volume grab samples (e.g., ≤10 liters) to recover viruses using
skimmed milk. When applied to 10 liter volumes of marine
water, the skimmed milk flocculation procedure was reported
to recover an average of 42–52% of seeded adenovirus (132).
When applied to 5 and 10 liter river water samples, the tech-
nique was reported to recover ∼50% of seeded viruses (133).
To perform the skimmed milk organic flocculation techni-
que, the water sample is acidified (e.g., using 1 N HCl) to
pH 3.5 and the conductivity adjusted (if needed) to ≥
1,500 µS/cm using artificial sea salts (134). To the water sam-
ple, preflocculated 1% skimmed milk is added at a ratio of
1:100 and the sample is stirred for 8–10 h at room tempera-
ture. Then the stirring is stopped and the flocs allowed to sedi-
ment by gravity for 8–10 h. After the sedimentation period,
the supernatant is removed using a peristaltic pump and the
remaining sediment (approximately 500 ml) is transferred
to a centrifuge tube. The flocs are then concentrated by cen-
trifugation at 8,000 × g for 30 min at 4°C. The pelleted flocs
are resuspended using phosphate buffer to produce final con-
centrates of ∼10 ml for viral analysis. While the skimmed
milk technique has been reported to be effective for recover-
ing viruses in various types of water the method is limited,
as with other precipitation and flocculation techniques,
application of the method is limited to grab samples, with a
practical limit of 10–20 liters per sample.
Centrifugal Filtration Techniques
Centrifugal microconcentrators use ultrafilter membranes
and relatively low-speed centrifugation to filter liquids. The
largest volume centrifugal filters that are commercially avail-
able are able to process 70 ml of liquid, so the technique is
limited in practical terms to processing 100–200 ml of liquid
per filter unit. Such centrifugal filter units may be available in
a range of pore sizes, from 3 kDa to 100 kDa, and are relatively
expensive compared to alternative secondary concentrations
procedures such as chemical precipitation and organic floccu-
lation. While little comparative data are available, Hill et al.
reported that 30 kDa Centricon Plus-70 centrifugal filter
units (Millipore) were as effective as PEG precipitation for
recovering GI and GII noroviruses in 100 liter groundwater
samples that were concentrated using TF HFUF (41). Hill
et al. also reported using 30 kDa Centricon Plus-70 units to
concentrate viruses in 100 liters tap water concentrates pro-
duced by TF UF (48). In this study, the researchers followed
the manufacturer’s instructions, reporting recoveries of 61–
120% for MS2 and phi X174 bacteriophages. In another
study where TF UF was used to concentrate 100 liters ground-
water and 100 liters surface water samples, 30 or 100 kDa Cen-
tricon Plus-70 centrifugal filters were used to process 70 ml
volumes of UF concentrates (135). When processing ground
water samples, the centrifugal filter units were used to directly
concentrate viruses in the UF concentrates. However, when
processing surface water samples, the researchers preclarified
the UF concentrates prior to using the microconcentrators
to minimize possible clogging of the filters. After centrifuging
the samples at 5,000 × g for 5 min at 4°C, the supernatants
were processed using the centrifugal filter units. A similar pro-
cedure was also utilized to process ground water, surface water,
and finished drinking water samples in a field station in
Ghana (136).
SUMMARY
Investigators must consider many factors when planning and
conducting water quality studies. To meet study goals and
data quality objectives, sample collection plans should be
developed to describe the sample types, volumes, locations,
collection procedures, and transport considerations that are
appropriate to facilitate effective testing for target microbes.
When collecting samples for pathogen testing, large-volume
samples are often needed, especially for viruses and parasites
for which culture techniques are not available or are other-
wise technically challenging. Numerous techniques have
been reported for large-volume sample collection. Many of
these are specific for recovering certain classes of pathogens,
such as viruses or parasites, but others (such as UF) can be
effective for simultaneous capture of diverse microbes.
2.6.1-12 ▪ SAMPLING METHODS
However, whenever collecting large volume water samples,
analysts should have a plan for minimizing the effects that
water sample constituents can have on the performance and
efficiency of sample processing techniques and downstream
analytical methods. Whether avoiding filter clogging, filter
adsorption inefficiency, or loss during centrifugation, it is crit-
ically important for sampling and sample processing techni-
ques to incorporate media, reagents, or procedures to reduce
microbial losses and inhibitors of analytical procedures. There
is no single gold standard sampling method that is effective or
warranted for any water type, so it is incumbent on investiga-
tors to be educated about the environment they are studying,
the microbes they are targeting, and the advantages and dis-
advantages of alternative sampling and sample processing
methods.
REFERENCES
1. American Public Health Association AWWA, Water
Environment Federation. 2005. Standard Methods for the
Examination of Water and Wastewater, 21st ed. American
Public Health Association, New York.
2. U.S. EPA. 2002. Guidance on Choosing a Sampling
Design for Environmental Data Collection for Use in
Developing a Quality Assurance Project Plan. EPA QA/
G-5S, Office of Environmental Information, EPA/240/
R-02/005.
3. Haas CN. 1993. Microbial sampling—is it better to sam-
ple many times or use large samples. Wat Sci Technol 27:
19–25.
4. Corapcioglu M, Haridas A. 1984. Transport and fate of
microorganisms in porous media: a theoretical investiga-
tion. J Hydrol 72:149–169.
5. Marrie TJ, Haldane D, Bezanson G, Peppard R. 1992.
Each water outlet is a unique ecological niche for Legionella
pneumophila. Epi Infect 108:261–270.
6. Kemble SK, Lynfield R, DeVries AS, Drehner DM, Pom-
putius WF, Beach MJ, Visvesvara GS, da Silva AJ, Hill
VR, Yoder JS, Xiao LH, Smith KE, Danila R. 2012. Fatal
Naegleria fowleri infection acquired in Minnesota: possible
expanded range of a deadly thermophilic organism. Clin
Infect Dis 54:805–809.
7. U.S. EPA. 2010. Sampling and Consideration of Variabil-
ity (Temporal and Spatial) for Monitoring of Recreational
Waters. EPA-823-R-10-005, Office of Water.
8. U.S. EPA. 1978. Microbiological Methods for Monitoring
the Environment—Water and Wastes. EPA-600/8-78-017,
Environmental Monitoring and Support Laboratory.
9. Hsu BM, Huang CC, Chen JS, Chen NH, Huang JT.
2011. Comparison of potentially pathogenic free-living
amoeba hosts by Legionella spp. in substrate-associated bio-
films and floating biofilms from spring environments Wat
Res 45:5171–5183.
10. Nicholson A. 2007. Groundwatech Chemistry. In Ground-
water and wells, 3rd ed. Johnson Screens, New Brighton,
MN.
11. U.S. EPA. 1985. Practical Guide for Ground-Water
Sampling. EPA/600/2-85/104, Office of Research and
Development.
12. Hargy TM, Rosen J, LeChevallier M, Friedman M,
Clancy JL. 2010. A high-volume sampling method for
total coliform and E. coli. J Am Wat Works Assoc 102:
79–86.
13. Kemp R, Leatherbarrow AJH, Williams NJ, Hart CA,
Clough HE, Turner J, Wright EJ, French NP. 2005. Prev-
alence and genetic diversity of Campylobacter spp. in envi-
ronmental water samples from a 100-square-kilometer
predominantly dairy farming area. Appl Environ Microbiol
71:1876–1882.
14. Higgins JA, Belt KT, Karns JS, Russell-Anelli J, Shelton
DR. 2005. Tir- and stx-positive Escherichia coli in stream
waters in a metropolitan area. Appl Environ Microbiol 71:
2511–2519.
15. Sobsey MD, Schwab KJ, Handzel TR. 1990. A simple
membrane filter method to concentrate and enumerate
male-specific RNA coliphages. J Am Wat Works Assoc
82:52–59.
16. Papageorgiou GT, Moce-Llivina L, Christodoulou CG,
Lucena F, Akkelidou D, Ioannou E, Jofre J. 2000.
A simple methodological approach for counting and iden-
tifying culturable viruses adsorbed to cellulose nitrate mem-
brane filters. Appl Environ Microbiol 66:194–198.
17. Falkinham JO, Iseman MD, de Haas P, van Soolingen D.
2008. Mycobacterium avium in a shower linked to pulmo-
nary disease. J Wat Health 6:209–213.
18. Leoni E, De Luca G, Legnani PP, Sacchetti R, Stampi S,
Zanetti F. 2005. Legionella waterline colonization: Detec-
tion of Legionella species in domestic, hotel and hospital
hot water systems. J Appl Microbiol 98:373–379.
19. John DT, Howard MJ. 1995. Seasonal distribution of
pathogenic free-living amebae in Oklahoma waters. Parasi-
tol Res 81:193–201.
20. Yoder JS, Straif-Bourgeois S, Roy SL, Moore TA, Visves-
vara GS, Ratard RC, Hill VR, Wilson JD, Linscott AJ,
Crager R, Kozak NA, Sriram R, Narayanan J, Mull B, Kah-
ler AM, Schneeberger C, da Silva AJ, Poudel M, Baum-
gartner KL, Xiao LH, Beach MJ. 2012. Primary amebic
meningoencephalitis deaths associated with sinus irrigation
using contaminated tap water. Clin Infect Dis 55:E79E85.
21. Shields JM, Gleim ER, Beach MJ. 2008. Prevalence of
Cryptosporidium spp. and Giardia intestinalis in swimming
pools, Atlanta, Georgia. Emerg Infect Dis 14:948–950.
22. Hutcheson C, Cira R, Gaines SL, Jones KR, Howard W,
Hornsby D, Rustin C, Hlavsa MC, Murphy JL, Jothiku-
mar N, Miller CD, Cantrell B, Hill VR, Beach MJ. 2013.
Microbes in pool filter backwash as evidence of the need for
improved swimmer hygiene—metro-Atlanta, Georgia,
2012. Morbid Mortal Week Rep 62:385–388.
23. U.S. EPA. April 2001. USEPA Manual of Methods for
Virology: Chapter 14 (Concentration and Processing of
Waterborne Viruses by Positive Charge 1MDS Cartridge
Filters and Organic Flocculation). EPA/600/4-84/013.
24. Wommack KE, Hill RT, Colwell RR. 1995. A simple
method for the concentration of viruses from natural-water
samples. J Microbiol Meth 22:57–67.
25. Bertke EE. 2007. Composite analysis for Escherichia coli at
coastal beaches. J Great Lakes Res 33:335–341.
26. Xagoraraki I, Kuo DHW, Wong K, Wong M, Rose JB.
2007. Occurrence of human adenoviruses at two recrea-
tional beaches of the great lakes. Appl Environ Microbiol
73:7874–7881.
27. U.S. EPA. 2005. Method 1622: Cryptosporidium in Water
by Filtration/IMS/FA. EPA 815-R-05-001.
28. U.S. EPA. 2012. Method 1623.1: Cryptosporidium and
Giardia in Water by Filtration/IMS/FA. EPA 816-R-
12-001.
29. Shapiro K, Mazet JAK, Schriewer A, Wuertz S, Fritz H,
Miller WA, Largier J, Conrad PA. 2010. Detection of Tox-
oplasma gondii oocysts and surrogate microspheres in water
using ultrafiltration and capsule filtration. Water Res 44:
893–903.
30. Villena I, Aubert D, Gomis P, Ferte H, Inglard JC,
Denis-Bisiaux H, Dondon J-M, Pisano E, Ortis N, Pinon
J-M. 2004. Evaluation of a strategy for Toxoplasma gondii
oocyst detection in water. Appl Environ Microbiol 70:
4035–4039.
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-13
31. Sturbaum GD, Ortega YR, Gilman RH, Sterling CR,
Cabrera L, Klein DA. 1998. Detection of Cyclospora
cayetanensis in wastewater. Appl Environ Microbiol 64:
2284–2286.
32. Marciano-Cabral F, Jamerson M, Kaneshiro ES. 2010.
Free-living amoebae, Legionella and Mycobacterium in tap
water supplied by a municipal drinking water utility in
the USA. J Water Health 8:71–82.
33. Huffman DE, Gennaccaro AL, Berg TL, Batzer G,
Widmer G. 2006. Detection of infectious parasites in
reclaimed water. Water Environ Res 78:2297–2302.
34. Quintero-Betancourt W, Gennaccaro AL, Scott TM, Rose
JB. 2003. Assessment of methods for detection of infectious
Cryptosporidium oocysts and Giardia cysts in reclaimed
effluents. Appl Environ Microbiol 69:5380–5388.
35. Magana-Ordorica D, Mena K, Valdez-Torres JB, Soto-
Beltran M, Leon-Felix J, Chaidez C. 2010. Relationships
between the occurrence of Giardia and Cryptosporidium
and physicochemical properties of marine waters of the
Pacific Coast of Mexico. J Water Health 8:797–802.
36. Higgins JA, Trout JM, Fayer R, Shelton D, Jenkins MC.
2003. Recovery and detection of Cryptosporidium parvum
oocysts from water samples using continuous flow centrifu-
gation. Water Res 37:3551–3560.
37. Swales C, Wright S. 2000. Evaluation of a continuous flow
centrifuge for recovery of Cryptosporidium oocysts from large
volume water samples. Water Res 34:1962–1966.
38. Borchardt MA, Spencer SK. 2002. Concentration of
Cryptosporidium, microsporidia and other water-borne
pathogens by continuous separation channel centrifuga-
tion. J Appl Microbiol 92:649–656.
39. Zuckerman U, Tzipori S. 2006. Portable continuous flow
centrifugation and method 1623 for monitoring of water-
borne protozoa from large volumes of various water matri-
ces. J Appl Microbiol 100:1220–1227.
40. Zuckerman U, Armon R, Tzipori S, Gold D. 1999. Eval-
uation of a portable differential continuous flow centrifuge
for concentration of Cryptosporidium oocysts and Giardia
cysts from water. J Appl Microbiol 86:955–961.
41. Hill VR, Mull B, Jothikumar N, Ferdinand K, Vinje J.
2010. Detection of GI and GII noroviruses in ground water
using ultrafiltration and TaqMan Real-time RT-PCR. Food
Environ Virol 2:218–224.
42. Mull B, Hill VR. 2009. Recovery and detection of
Escherichia coli O157:H7 in surface water, using ultrafiltra-
tion and Real-Time PCR. Appl Environ Microbiol 75:
3593–3597.
43. Kimble G, Amburgey J, Hill V. 2012. Comparison of
hollow-fiber ultrafilters with pleated capsule filters for sur-
face and tap water samples using U.S. EPA Method 1623.
J Environ Eng 138:899–901.
44. Raphael BH, Lautenschlager M, Kahler A, Pai S, Parks
BA, Kalb SR, Maslanka SE, Shah S, Magnuson M, Hill
VR. 2012. Ultrafiltration improves ELISA and Endopep
MS analysis of botulinum neurotoxin type A in drinking
water. J Microbiol Meth 90:267–272.
45. Hill VR, Polaczyk AL, Hahn D, Narayanan J, Cromeans
TL, Roberts JM, Amburgey JE. 2005. Development of a
rapid method for simultaneous recovery of diverse microbes
in drinking water by ultrafiltration with sodium polyphos-
phate and surfactants. Appl Environ Microbiol 71:
6878–6884.
46. Morales-Morales HA, Vidal G, Olszewski J, Rock CM,
Dasgupta D, Oshima KH, Smith GB. 2003. Optimization
of a reusable hollow-fiber ultrafilter for simultaneous con-
centration of enteric bacteria, protozoa, and viruses from
water. Appl Environ Microbiol 69:4098–4102.
47. Olszewski J, Winona L, Oshima KH. 2005. Comparison
of 2 ultrafiltration systems for the concentration of seeded
viruses from environmental waters. Can J Microbiol 51:
295–303.
48. Hill VR, Kahler AM, Jothikumar N, Johnson TB, Hahn
D, Cromeans TL. 2007. Multistate evaluation of an
ultrafiltration-based procedure for simultaneous recovery
of enteric microbes in 100-liter tap water samples. Appl
Environ Microbiol 73:4218–4225.
49. Gibson KE, Schwab KJ. 2011. Tangential-flow ultrafiltra-
tion with Integrated inhibition detection for recovery of
surrogates and human pathogens from large-volume source
water and finished drinking water. Appl Environ Microbiol
77:385–391.
50. Winona LJ, Ommani AW, Olszewski J, Nuzzo JB, Osh-
ima KH. 2001. Efficient and predictable recovery of viruses
from water by small scale ultrafiltration systems. Can
J Microbiol 47:1033–1041.
51. Dziewulski DM, Belfort G. 1983. Virus concentration
from water using high-rate tangential-flow hollow fiber
ultrafiltration. Water Sci Technol 15:75–89.
52. Smith CM, Hill VR. 2009. Dead-end hollow-fiber ultrafil-
tration for recovery of diverse microbes from water. Appl
Environ Microbiol 75:5284–5289.
53. Leskinen SD, Harwood VJ, Lim DV. 2009. Rapid dead-
end ultrafiltration concentration and biosensor detection
of enterococci from beach waters of Southern California.
J Water Health 7:674–684.
54. Mull B, Hill VR. 2012. Recovery of diverse microbes in
high turbidity surface water samples using dead-end ultrafil-
tration. J Microbiol Meth 91:429–433.
55. Holowecky PM, James RR, Lorch DP, Straka SE, Lind-
quist HDA. 2009. Evaluation of ultrafiltration cartridges
for a water sampling apparatus. J Appl Microbiol 106:
738–747.
56. Kfir R, Hilner C, Dupreez M, Bateman B. 1995.
Studies evaluating the applicability of utilizing the same
concentration techniques for the detection of protozoan
parasites and viruses in water. Water Sci Technol 31:
417–423.
57. Simmons OD, Sobsey MD, Heaney CD, Schaefer FW,
Francy DS. 2001. Concentration and detection of Crypto-
sporidium oocysts in surface water samples by method 1622
using ultrafiltration and capsule filtration. Appl Environ
Microbiol 67:1123–1127.
58. Liu PB, Hill VR, Hahn D, Johnson TB, Pan Y,
Jothikumar N, Moe CL. 2012. Hollow-fiber ultrafiltra-
tion for simultaneous recovery of viruses, bacteria and
parasites from reclaimed water. J Microbiol Meth 88:
155–161.
59. Polaczyk AL, Narayanan J, Cromeans TL, Hahn D,
Roberts JM, Amburgey JE, Hill VR. 2008. Ultrafiltration-
based techniques for rapid and simultaneous concentration
of multiple microbe classes from 100-L tap water samples.
J Microbiol Meth 73:92–99.
60. Polaczyk AL, Roberts JM, Hill VR. 2007. Evaluation
of 1MDS electropositive microfilters for simultaneous
recovery of multiple microbe classes from tap water.
J Microbiol Meth 68:260–266.
61. Lindquist HDA, Harris S, Lucas S, Hartzel M, Riner D,
Rochele P, DeLeon R. 2007. Using ultrafiltration to con-
centrate and detect Bacillus anthracis, Bacillus atrophaeus
subspecies globigii, and Cryptosporidium parvum in 100-liter
water samples. J Microbiol Methods 70:484–492.
62. Rhodes ER, Hamilton DW, See MJ, Wymer L. 2011.
Evaluation of hollow-fiber ultrafiltration primary concen-
tration of pathogens and secondary concentration of
viruses from water. J Virol Meth 176:38–45.
63. CDC, U.S. EPA. 2011. Comparison of Ultrafiltration
Techniques for Recovering Biothreat Agents in Water.
EPA 600/R-11/103, Office of Research and Development.
2.6.1-14 ▪ SAMPLING METHODS
64. Rajal VB, McSwain BS, Thompson DE, Leutenegger
CM, Kildare BJ, Wuertz S. 2007. Validation of hollow
fiber ultrafiltration and real-time PCR using bacteriophage
PP7 as surrogate for the quantification of viruses from water
samples. Water Res 41:1411–1422.
65. Hill VR, Jothikumar N, Vinje J, Cromeans TL. 2010.
Sample Preparation Methods for Molecular Techniques for
Drinking Water, Project 3108. Water Research Foundation.
Denver, CO.
66. Karim MR, Rhodes ER, Brinkman N, Wymer L, Fout
GS. 2009. New electropositive filter for concentrating
enteroviruses and noroviruses from large volumes of water.
Appl Environ Microbiol 75:2393–2399.
67. Michen B, Graule T. 2010. Isoelectric points of viruses.
J Appl Microbiol 109:388–397.
68. Gerba CP. 1984. Applied and theoretical aspects of virus
adsorption to surfaces. Adv Appl Microbiol 30:133–168.
69. Haramoto E, Katayama H, Asami M, Akiba M. 2012.
Development of a novel method for simultaneous concen-
tration of viruses and protozoa from a single water sample.
J Virol Meth 182:62–69.
70. Brassard J, Guevremont E, Gagne MJ, Lamoureux L.
2011. Simultaneous recovery of bacteria and viruses from
contaminated water and spinach by a filtration method.
Int J Food Microbiol 144:565–568.
71. Sobsey MD, Jones BL. 1979. Concentration of poliovirus
from tap water using positively charged microporous filters.
Appl Environ Microbiol 37:588–595.
72. Sobsey MD, Glass JS. 1980. Poliovirus concentration from
tap water with electropositive adsorbent filters. Appl Envi-
ron Microbiol 40:201–210.
73. Lambertini E, Spencer SK, Bertz PD, Loge FJ, Kieke BA,
Borchardt MA. 2008. Concentration of enteroviruses,
adenoviruses, and noroviruses from drinking water by
use of glass wool filters. Appl Environ Microbiol 74:
2990–2996.
74. U.S. EPA. 2012. Method 1615: Measurement of Enterovi-
rus and Norovirus Occurrence in Water by Culture and
RT-qPCR. EPA 600/R-10/181, Version 1.1. Office of
Research and Development.
75. Lukasik J, Scott TM, Andryshak D, Farrah SR. 2000.
Influence of salts on virus adsorption to microporous filters.
Appl Environ Microbiol 66:2914–2920.
76. Gibbons CD, Rodriguez RA, Tallon L, Sobsey MD. 2010.
Evaluation of positively charged alumina nanofibre car-
tridge filters for the primary concentration of noroviruses,
adenoviruses and male-specific coliphages from seawater.
J Appl Microbiol 109:635–641.
77. Bennett HB, O’Dell HD, Norton G, Shin G, Hsu FC,
Meschke JS. 2010. Evaluation of a novel electropositive fil-
ter for the concentration of viruses from diverse water
matrices. Water Sci Technol 61:317–322.
78. Guttman-Bass N, Catalano-Sherman J. 1986. Humic acid
interference with virus recovery by electropositive micro-
porous filters. Appl Environ Microbiol 52:556–561.
79. Sobsey MD, Glass JS. 1984. Influence of water quality on
enteric virus concentration by microporous filter methods.
Appl Environ Microbiol 47:956–960.
80. U.S. EPA. 1996. ICR Microbial Laboratory Manual. EPA/
600/R-95/178, Office of Research and Development.
81. Cashdollar JL, Brinkman NE, Griffin SM, McMinn
BR, Rhodes ER, Varughese EA, Grimm AC, Parshioni-
kar SU, Wymer L, Fout GS. 2013. Development and eval-
uation of EPA method 1615 for detection of enterovirus
and norovirus in water. Appl Environ Microbiol 79:
215–223.
82. Vilagines P, Sarrette B, Husson G, Vilagines R. 1993.
Glass wool for virus concentration at ambient water PH
level. Water Sci Technol 27:299–306.
83. Wyn-Jones AP, Carducci A, Cook N, D’Agostino M,
Divizia M, Fleischer J, Gantzer C, Gawler A, Girones
R, Höller, de Roda Husman AM, Kay D, Kozyra I,
López-Pila J, Muscillo M, Nascimento MSJ, Papageor-
iou G, Rutjes S, Sellwood J, Szewzyk R, Wyer M.
2011. Surveillance of adenoviruses and noroviruses in
European recreational waters. Water Res 45:1025–1038.
84. Deboosere N, Horm SV, Pinon A, Gachet J, Coldefy C,
Buchy P, Vialette M. 2011. Development and validation
of a concentration method for the detection of influenza
A viruses from large volumes of surface water. Appl Environ
Microbiol 77:3802–3808.
85. Bradbury KR, Borchardt MA, Gotkowitz M, Spencer
SK, Zhu J, Hunt RJ. 2013. Source and transport of human
enteric viruses in deep municipal water supply wells. Envi-
ron Sci Technol 47:4096–4103.
86. Sobsey MD, Glass JS, Jacobs RR, Rutala WA. 1980.
Modifications of the tentative standard method for
improved virus recovery efficiency. J Am Water Works Assoc
72:350–355.
87. Wallis C, Melnick JL. 1967. Concentration of enterovi-
ruses on membrane filters. J Virol 1:472–477.
88. Rose JB, Singh SN, Gerba CP, Kelley LM. 1984.
Comparison of microporous filters for concentration of
viruses from wastewater. Appl Environ Microbiol 47:
989–992.
89. Pallin R, WynJones AP, Place BM, Lightfoot NF. 1997.
The detection of enteroviruses in large volume concen-
trates of recreational waters by the polymerase chain reac-
tion. J Virol Meth 67:57–67.
90. Katayama H, Shimasaki A, Ohgaki S. 2002. Development
of a virus concentration method and its application to
detection of enterovirus and Norwalk virus from coastal
seawater. Appl Environ Microbiol 68:1033–1039.
91. Watt PM, Johnson DC, Gerba CP. 2002. Improved
method for concentration of Giardia, Cryptosporidium,
and poliovirus from water. J Environ Sci Health A Tox Haz-
ard Subst Environ Eng 37:321–330.
92. Hurst CJ, Dahling DR, Safferman RS, Goyke T. 1984.
Comparison of commercial beef extracts and similar mate-
rials for recovering viruses from environmental samples.
Canadian J Microbiol 30:1253–1263.
93. Wyn-Jones AP, Pallin R, Dedoussis C, Shore J, Sellwood
J. 2000. The detection of small round-structured viruses in
water and environmental materials. J Virol Meth 87:
99–107.
94. Schwab KJ, Deleon R, Sobsey MD. 1995. Concentration
and purification of beef extract mock eluates from water
samples for the detection of enteroviruses, hepatitis A virus,
and Norwalk virus by reverse transcription-PCR. Appl
Environ Microbiol 61:531–537.
95. Haramoto E, Katayama H, Ohgaki S. 2004. Detection of
noroviruses in tap water in Japan by means of a new method
for concentrating enteric viruses in large volumes of fresh-
water. Appl Environ Microbiol 70:2154–2160.
96. Kitajima M, Oka T, Haramoto E, Katayama H, Takeda N,
Katayama K, Ohgaki S. 2010. Detection and genetic anal-
ysis of human sapoviruses in river water in Japan. Appl Envi-
ron Microbiol 76:2461–2467.
97. Victoria M, Rigotto C, Moresco V, Correa AD, Kolesni-
kovas C, Leite JPG, Miagostovich MP, Barardi CRM.
2010. Assessment of norovirus contamination in environ-
mental samples from Florianopolis City, Southern Brazil.
J Appl Microbiol 109:231–238.
98. Haramoto E, Katayama H, Utagawa E, Ohgaki S. 2009.
Recovery of human norovirus from water by virus concen-
tration methods. J Virol Meth 160:206–209.
99. U.S. EPA. 2007. Validation and Peer Review of U.S. Envi-
ronmental Protection Agency Sampling Methods for
 2.6.1. Water Sampling and Processing Techniques for Public Health–Related Microbes ▪ 2.6.1-15
Chemical and Radiochemical Parameters. FEM Document
Number 2007–02.
100. Stockman LJ, Wright CJ, Visvesvara GS, Fields BS,
Beach MJ. 2011. Prevalence of Acanthamoeba spp. and
other free-living amoebae in household water, Ohio,
USA—1990–1992. Parasitol Res 108:621–627.
101. Falkinham JO. 2011. Nontuberculous mycobacteria from
household plumbing of patients with nontuberculous
mycobacteria disease. Emerg Infect Dis 17:419–424.
102. U.S. EPA. 2000. Improved Enumeration Methods for the
Recreational Water Quality Indicators: Enterococci and
Escherichia coli. EPA/821/R-97/004. Office of Science and
Technology.
103. Wiedenmann A, Langhammer W, Botzenhart K. 2001. A
case report of false negative Legionella test results in a chlori-
nated public hot water distribution system due to the lack of
sodium thiosulfate in sampling bottles. Int J Hyg Environ
Health 204:245–249.
104. Sen K, Schable NA, Lye DJ. 2007. Development of an
internal control for evaluation and standardization of a
quantitative PCR assay for detection of Helicobacter
pylori in drinking water. Appl Environ Microbiol 73:
7380–7387.
105. U.S. EPA. 2002. Guidance for Quality Assurance Project
Plans. EPA QA/G-5, EPA/240/R-02/009, Office of Envi-
ronmental Information.
106. McCuin RM, Clancy JL. 2005. Methods for the recovery,
isolation and detection of Cryptosporidium oocysts in waste-
waters. J Microbiol Meth 63:73–88.
107. Bracken CL, Hendricks CW, Harding AK. 2006. Appa-
rent bias in river water inoculum following centrifugation.
J Microbiol Meth 67:304–309.
108. Stevens KA, Jaykus LA. 2004. Bacterial separation and
concentration from complex sample matrices: a review.
Crit Rev Microbiol 30:7–24.
109. Prata C, Ribeiro A, Cunha A, Gomes NCM, Almeida A.
2012. Ultracentrifugation as a direct method to concen-
trate viruses in environmental waters: virus-like particle
enumeration as a new approach to determine the efficiency
of recovery. J Environ Monit 14:64–70.
110. Steyer A, Torkar KG, Gutierrez-Aguirre I, Poljsak-
Prijatelj M. 2011. High prevalence of enteric viruses in
untreated individual drinking water sources and sur-
face water in Slovenia. Int J Hyg Environ Health 214:
392–398.
111. Cliver DO, Yeatman J. 1965. Ultracentrifugation in the
concentration and detection of enteroviruses. Appl Micro-
biol 13:387–392.
112. Skraber S, Gantzer C, Helmi K, Hoffmann L, Cauchie
HM. 2009. Simultaneous concentration of enteric viruses
and protozoan parasites: a protocol based on tangential
flow filtration and adapted to large volumes of surface
and drinking waters. Food and Environ Virol 1:66–76.
113. Girones R, Allard A, Wadell G, Jofre J. 1993. Application
of PCR to the detection of adenoviruses in polluted waters.
Water Sci Technol 27:235–241.
114. Sylvain S, Gantzer C, Karim H, Lucien H, Henry-Michel
C. 2009. Simultaneous concentration of enteric viruses and
protozoan parasites: a protocol based on tangential flow fil-
tration and adapted to large volumes of surface and drink-
ing waters. Food and Environ Virol 1:66–76.
115. Philipson L, Albertsson PA, Frick G. 1960. The purifica-
tion and concentration of viruses by aqueous polymer phase
systems. Virology 11:553–571.
116. Albertsson PA, Frick G. 1960. Partition of virus particles
in a liquid two-phase system. Biochim Biophys Acta 37:
230–237.
117. Yamamoto KR, Alberts BM. 1970. Rapid bacteriophage
sedimentation in the presence of polyethylene glycol and
its application to large-scale virus purification. Virology
40:734–744.
118. Enriquez CE, Gerba CP. 1995. Concentration of enteric
adenovirus 40 from tap, sea, and waste water. Water Res
29:2554–2560.
119. Schwab KJ, DeLeon R, Sobsey MD. 1996. Immunoaffin-
ity concentration and purification of waterborne enteric
viruses for detection by reverse transcriptase PCR. Appl
Environ Microbiol 62:2086–2094.
120. Rodriguez RA, Thie L, Gibbons CD, Sobsey MD. 2012.
Reducing the effects of environmental inhibition in
quantitative real-time PCR detection of adenovirus and
norovirus in recreational seawaters. J Virol Meth 181:
43–50.
121. Lal SM, Lund E. 1975. Recovery of virus by chemical pre-
cipitation followed by elution. Prog Water Technol 7:
687–693.
122. Fernandez-Molina MC, Alvarez A, Espigares M. 2004.
Presence of hepatitis A virus in water and its relationship
with indicators of fecal contamination. Water Air Soil Pollut
159:197–208.
123. Payment P, Fortin S, Trudel M. 1984. Ferric chloride floc-
culation for nonflocculating beef extract preparations. Appl
Environ Microbiol 47:591–592.
124. U.S. EPA. 1995. Virus monitoring protocol for the
Information Collection Requirements Rule. EPA/814-B-
95-002, Office of Ground Water and Drinking Water.
125. Katzenelson E, Fattal BHT. 1976. Organic flocculation:
an efficient second step concentration method for the
detection of viruses in tap water. Appl Environ Microbiol
32:638–639.
126. Safferman RS, Rohr ME, Goyke T. 1988. Assessment of
recovery efficiency of beef extract reagents for concentrat-
ing viruses from municipal wastewater sludge solids by the
organic flocculation procedure. Appl Environ Microbiol
54:309–316.
127. Guttmanbass N, Armon R. 1983. Concentration of sim-
ian rotavirus SA-11 from tap water by membrane filtration
and organic flocculation. Appl Environ Microbiol 45:
850–855.
128. Ikner LA, Gerba CP, Bright KR. 2012. Concentration
and recovery of viruses from water: a comprehensive review.
Food Environ Virol 4:41–67.
129. Hill VR, Polaczyk AL, Kahler AM, Cromeans TL, Hahn
D, Amburgey JE. 2009. Comparison of hollow-fiber ultra-
filtration to the USEPAVIRADEL technique and USEPA
method 1623. J Environ Qual 38:822–825.
130. Dahling DR, Wright BA. 1986. Recovery of viruses
from water by a modified flocculation procedure for
second-step concentration. Appl Environ Microbiol 51:
1326–1331.
131. McMinn BR, Cashdollar JL, Grimm AC, Fout GS.
2012. Evaluation of the celite secondary concentration
procedure and an alternate elution buffer for the recovery
of enteric adenoviruses 40 and 41. J Virol Meth 179:
423–428.
132. Calgua B, Mengewein A, Grunert A, Bofill-Mas S,
Clemente-Casares P, Hundesa A, Wyn-Jones AP,
López-Pila J, Girones R. 2008. Development and applica-
tion of a one-step low cost procedure to concentrate viruses
from seawater samples. J Virol Meth 153:79–83.
133. Calgua B, Fumian T, Rusinol M, Rodriguez-Manzano J,
Mbayed VA, Bofill-Mas S, Miagostovich M, Girones R.
2013. Detection and quantification of classic and emerging
viruses by skimmed-milk flocculation and PCR in river
water from two geographical areas. Water Res 47:
2797–2810.
134. Bofill-Mas S, Hundesa A, Calgua B, Rusinol M, de Motes
CM, Girones R. 2011. Cost-effective method for microbial
2.6.1-16 ▪ SAMPLING METHODS
source tracking using specific human and animal viruses.
J Visual Exper 3(58):e2820.
135. Gibson KE, Schwab KJ. 2011. Detection of bacte-
rial indicators and human and bovine enteric viruses
in surface water and groundwater sources potentially
impacted by animal and human wastes in lower
Yakima Valley, Washington. Appl Environ Microbiol 77:
355–362.
136. Gibson KE, Opryszko MC, Schissler JT, Guo YY,
Schwab KJ. 2011. Evaluation of human enteric viruses in
surface water and drinking water resources in southern
Ghana. Am J Trop Med Hyg 84:20–29.
 Surface Sampling
LAURA J. ROSE, JUDITH NOBLE-WANG, AND MATTHEW J. ARDUINO
2.6.2
Microbial surface sampling techniques have been in use for
various purposes for the past century. Publications as early
as 1931 document techniques for sampling food contact
surfaces (1–7). Current methods stem from this early work
in the food industry, but have been adapted to meet specific
needs for numerous applications. In addition, some industries
have established standards and/ or guidance specific for their
needs (e.g., microbiological evaluation of clean rooms or
controlled pharmaceutical environments).
Guidelines and standards now exist for monitoring
contamination in the food manufacturing industry (8, 9).
Food preparation surfaces are sampled as part of a hazard
analysis critical control point approach to identify and mon-
itor the remediation effectiveness of environmental critical
control points (8) within a production process. Results of
sampling inform how effective control measures are at pre-
venting contamination of product with pathogens. Several
methods can be used, such as swabs, sponges, or contact
plates.
As the aeronautical industry expanded into space explora-
tion, concern about the contamination of other planets with
Earth organisms and the detection of extraterrestrial life led to
the formation of a group that developed methods and stand-
ard procedures for examining space hardware (10). Spacecraft
were monitored before flight and after returning from a mis-
sion (11–13). The international committee on space research
(COSPAR) developed a planetary protection policy that
requires member nations to provide reports on spacecraft
that includes “the estimated biological burden at launch,
the methods used to obtain the estimate (e.g. assay techniques
applied to spacecraft or a proxy), and the statistical uncer-
tainty in the estimate” (14).
For pharmaceutical manufacturing, the United States
Pharmacopoeial Convention provides standards for evalua-
tion of clean rooms and controlled environments (15) and
compounding facilities (16). Both standards, USP 1116 and
USP 797, specify using contact plates and swabs to monitor
cleaning and disinfection of surfaces, but if validated, other
methods may be used as well (17).
Though routine surface sampling is not recommended
for health care settings, it may be a tool to support an inves-
tigation of a disease outbreak if environmental reservoirs or
fomites are implicated (18–21). One example is the finding
in 1999 in an intensive care unit of a correlation between pos-
itive surface sample isolates of Acinetobacter baumannii and
doi:10.1128/9781555818821.ch2.6.2
2.6.2-1
the number of patients with A. baumannii colonization and/
or infection (22).
The response to a biothreat incident may also require sur-
face sampling to characterize the extent of the contamina-
tion, determine who may have been exposed, and evaluate
the efficacy of remediation or clean-up efforts (23–25), as
evidenced by the response to the 2001 Bacillus anthracis
mail contamination incident (25, 26). The Centers for
Disease Control and Prevention invested in evaluating sam-
pling methods for the recovery of B. anthracis spores from
smooth, nonporous surfaces. The sampling devices evaluated
included swabs, wipes, and sampling sponges. Swab and
sponge processing and culturing methods for the recovery of
B. anthracis spores from nonporous surfaces (27, 28) and other
devices and protocols were evaluated for this purpose as well
(29–32).
Surface sampling has been used in research settings to
determine potential environmental reservoirs of pathogens
(33–36). However, standard methods and strategies may
not exist for all situations, and the investigator should care-
fully consider several factors before engaging in environ-
mental surface sampling.
CONSIDERATIONS
No sampling method is capable of recovering 100% of the
organisms present. The recovery efficiency of a sampling
method is influenced by many factors, such as the interactions
among the organism, surface, and sampling materials; the
presence of organic matter and other organisms; the shipping
and/or storage conditions (time, temperature, transport
media); and the assays by which the organisms are detected
and/or quantified (Box 1). Table 1 shows some organisms
and percent recoveries reported from surfaces with various
sampling tools and under the conditions specified in the
references. Each step, from sampling to enumeration, can
have an impact on the resulting data, and since decisions
about the bioburden of an area in question are based on a rel-
atively small number of representative samples, established
sampling procedures must be applied uniformly and consis-
tently (37).
Sampling Plan
With these considerations in mind, a sound plan should be in
place before beginning to sample. The objectives of a
2.6.2-2 ▪ SAMPLING METHODS

BOX 1 Variables Affecting Sampling, Transportation and Storage, Detection, and Analysis

• Ability of the agent to survive or produce active toxins on surfaces for extended periods of time in various environmental
conditions
○ Spores of B. anthracis, Clostridium spp., and fungal spores can persist for years on surfaces such as paper or nonporous

surfaces (110)
○ Some vegetative bacteria (e.g., Staphylococcus aureus) can survive for weeks or months depending on the presence of
organic material, available moisture, or a suitable carrier (111–113)
○ If viability of the microorganism is unlikely, then sampling for detection by alternate means (i.e., PCR, DNA
sequencing) should be considered or used in conjunction with culture
• Interference from particulates, other organisms, growth inhibitors, or assay inhibitors
○ Particulates such as dust, heavy metals, or fibers may inhibit growth, interfere with molecular assays, or make recog-
nition of characteristic colony morphology difficult in culture (114)
○ Large numbers of background bacteria present can compete with the target organism in culture, or interfere with
molecular assays by providing an excess of nontarget DNA (115)
○ Clay and organic substances (e.g., humic acid in soil) can inhibit growth in culture and interfere with molecular assays

(23, 116).
• Adherence properties of microorganisms
○ Vegetative bacteria are more likely than bacterial spores to be affected by the chemical and/or physical properties of
surfaces (117)
○ Adherence and sampling efficiency can be influenced by ambient temperature and humidity, and choice of sampling

device and moistening fluid

• Size of surface being sampled
○ Recovery efficiency may be lost with increased surface area sampled

• Transport and storage conditions

○ When transporting to the laboratory for analysis, viability and/or stability of the sample may be compromised if con-

ditions are not optimal for the given organism (103, 104)
○ Choice of transport media may be important for quantitative analysis to ensure organisms aren’t inhibited or growth
enhanced
• Time, temperature since sampling
• Laboratory assay used for detection
○ Optimum elution methods may vary with target organism
○ Materials and wetting compounds may not be compatible with assay reagents
○ Growth media, diluents, or neutralizers may be inhibitory to the organism being detected

sampling plan must be well defined to provide usable, defend-
able, applicable, and scientifically meaningful data for an
investigation or study. Essential components of a sound sam-
pling plan include (a) properly trained personnel, (b) a quali-
fied laboratory, (c) laboratory-approved sample media and
supplies, (d) appropriate safety policies and procedures, (e)
thorough record keeping and documentation, and (f ) quality
assurance (QA) quality control procedures (38), and (g) a
clear understanding of how the results will be interpreted
and applied (20). Two basic elements must be present in
any sampling plan to ensure the reliability of the results: the
use of sterile equipment, and the use of the aseptic technique.
Use of prepackaged sterile supplies (within expiration date)
and equipment are the easiest way to ensure that sampling
is done with sterile materials. Aseptic technique ensures
that no additional contamination is introduced into the sam-
ple during the collection process and that a sample represent-
ing the actual flora on the surface has been collected (37). To
ensure aseptic technique during sampling, the concept of the
“clean person/contaminated person” is suggested. In practical
terms, this becomes a two-person sampling team, with one
individual handling the sampling device (the sampler) and
the other individual handling all other supplies (support
person) (37). The support person uses aseptic technique to
open packages and make sampling devices and equipment
available to the sampler. The sampler is the only person who
touches the sample media, collects the sample, and places the
device into an open specimen container that is held by the
support person. Once the sample is placed into the specimen
container, the support person decontaminates the outer sur-
faces of the container and prepares it for storage or shipment
to the laboratory. When sampling for potential pathogens,
persons involved with microbiologic sampling should use per-
sonal protective equipment appropriate to the risk and task
(e.g., gloves, lab coat, gown, respirator) (38, 39).
Sampling Sites
When deciding on the locations to be sampled, consider
whether the locations are truly representative and whether
the numbers of samples to be taken will provide the confi-
dence needed for the investigation. Several approaches to
designing a sampling plan can be found in the U.S. EPA
 2.6.2. Surface Sampling ▪ 2.6.2-3

TABLE 1 Examples of recovery efficiencies; percent of organisms recovered, relative to known inoculum placed on surfaces, counted as
cultivable colonies (CFU), infective virus (PFU), or cell equivalents of DNA copies

Surface Inoculum (CFU or Recovery efficiency

Organism Surface Sampling tool Reference
area PFU) (%)
B. anthracis spores Steel Swab 26 cm2 49 25.7 27
B. anthracis spores Steel Cellulose sponge-stick 645 cm2 26 32.4 28
B. atrophaeus spores Steel Vacuum filter sock 100 cm2 102–103 32.1 31
Bacillus atrophaeus Vectrana Foam swab 25 cm2 104 1.4 118
fabric
Bacillus atrophaeus Vectran fabric Nylon flocked swab 25 cm2 104 5.9 118
Bacillus atrophaeus Vectran fabric Foam spatula 400 cm2 1,551 13.2 118
Bacillus atrophaeus Steel Nylon flocked swab 25 cm2 100 45.2 119
Bacillus atrophaeus Vectran fabric Cellulose Sponge-stick 400 cm2 1,551 0.5 118
Escherichia coli Steel Dacron swab 25 cm2 104 0.06 120
Yersinia pestis CO92 Steel Dacron swab 4 in2 106 94b 121
Yersinia pestis A1122 Steel Dacron swab 4 in2 106 <1b 121
Staphylococcus aureus Steel Dip slide/selective 20 102–105 51.3 51
(MRSA) media
Staphylococcus aureus Steel Swab/pour plate 825 102–105 0.12 51
(MRSA)
MS2 bacteriophage Steel Polyester swab 25 mm2 5012 39 61
a
Rough fabric used as spacecraft airbag material.
b
Cell equivalents of DNA.

document QA/G-5s, “Guidance on Choosing a Sampling
Design for Environmental Data Collection” (40). Two com-
monly used approaches are targeted sampling (sometimes
called judgmental sampling), and statistical sampling, some-
times called probabilistic sampling (38, 41). The targeted
sampling approach consists of using good judgment to
sample where the contamination or organism is most likely
to be, based on the available evidence (forensic, epidemio-
logical, historical, etc.). The efficacy of this approach is
only as good as the evidence available. The statistical or prob-
abilistic approach, simply put, involves dividing up the area
into a grid and sampling a portion the spaces of the grid
in a random manner. There are many types of statistical
sampling approaches, each varying in how chance is used to
select sample points and the methods used to draw conclu-
sions from the data. Some examples include simple random
sampling, stratified sampling, and cluster sampling (40, 42,
43). It may also be possible to use a Bayesian model to com-
bine targeted sampling and statistical sampling, as described
by Sego et al. (44). The probabilistic approaches can provide
better statistical confidence in some applications, but if a
large area is to be covered, the number of samples needed
for this approach may be prohibitive.
The goal of the sampling can be to obtain qualitative
results or quantitative results and in some cases a combination
of both. Qualitative results would indicate if a given organism,
or any organism, is detected on the surface. Quantitative
(or semi-quantitative) results would indicate the estimated
number of organism detected on the surface and could be
reported as a number per given area. Sampling for quantita-
tive results requires an attention to the amount of surface
area sampled, since results will be reported as the concentra-
tion of the agent in a given area sampled. Guidelines for areas
and surface types for each sampling device can be seen in
Table 2. To reduce variability, there must be uniformity of
sample collection materials or devices, as well as uniform per-
sonnel techniques. Laboratory processing for quantitative
results involves more time and effort, as well as more supplies
and reagents; communication with the laboratory early in
the planning stages is therefore crucial. Qualitative sampling
still requires an understanding of the method so that the limit
of detection is understood.
Special Circumstances
When undertaking environmental sampling to evaluate the
effectiveness of a large-space decontamination process, bio-
logical indicators (e.g., commercially available spore strips)
should always be included in the sampling design (23). Envi-
ronmental sampling of surfaces does not stand alone on its
own merit when assessing the efficacy of decontamination.
Rather, there must be evidence that the decontamination
process was successful (i.e., inactivation of biological indica-
tors placed throughout the treated space).
The presence of residual disinfectants should be con-
sidered when sampling surfaces have been cleaned and disin-
fected. This is most important when using such disinfectants
such as sodium hypochlorite, quaternary ammonium com-
pounds, hydrogen peroxide, and phenolics. In such cases,
inclusion of specific neutralizers in the rinse fluids, moisten-
ing fluids, or sampling media is important to prevent carrying
the residual disinfectant activity into the assay or culture
media (Table 3). Such carryover may result in toxic effects
and reduced numbers or erroneous assay results (e.g., false-
negative results). The sampling team should consult with
laboratory personnel when making decisions on the appro-
priate neutralizers to be incorporated in the sampling device
during sample collection. The choice of neutralizer should
be based on the type of disinfectant used for deconta-
mination, compatibility with the assay, and knowledge of
the neutralizer’s toxicity to the target microbe(s) (21, 23,
2.6.2-4 ▪ SAMPLING METHODS
TABLE 2 Comparison of surface sampling devices or sample types
Sample type/examplesa
Swab
Foam: Individually
wrapped, scored shaft, # 25–
1607 1PF SC Puritan
Medical Products
(Guilford, ME)
Rayon: swab in 2.5 ml rinse
solution #908C Copan
Diagnostics (Murrieta, CA)
Wipe
Rayon: Versalon Gauze
200 × 200 Sponge, #8042
Covidien (Mansfield, MA)
Cellulose: Sponge folded
over a handle #SSL10NB
Sponge-Stick, 3M (St Paul,
MN)
Foam: BisKit #BIS-40001A
Quicksilver Analytics
(Abingdon, MD)
Agar contact
Prepared RODAC plates
(#21232 D/E neutralizing
agar, BD Diagnostic
Systems (Franklin Lakes,
NJ)
Dip slide: Hygicult TPC
#68010 Orion Diagnostica
(Espoo, Finland)
PetriFilm: aerobic count
plates #6400 3M™ (St
Paul, MN)
Vacuum sock
Dust Collection Sock,
X-Cell 200 Midwest
Filtration (Cincinnati,
OH)
Microvacuum Cassette
Micro-vac cassette: with
0.45 µm MCE filter #222–
9543, SKC (Eighty Four,
PA)
Micro-Vac carpet/Dust
Sampling Cassette:
#7345CC, Zefon (Ocala,
FL)
Wet Vacuum
M-Vac, MSI (Bluffdale,
UT)
Typical application
Moisten with sterile lab-determined
wetting solution, swab area of
specified size.
Surface type: Smooth, nonporous, or
crevices and corners. Often collected
with other sample media. Extent of
contamination and critical item
screening.
Area: 100 cm2
Moisten with sterile lab-determine
wetting solution and wipe area of
specified size
Surface type: Smooth, nonporous
Area: typically 600–700 cm2
Screening small nonporous surfaces,
sample collected by firmly pressing
the contact plate or other agar surface
onto the surface to be sampled. Do
not move or slide the plate.
Surface type: smooth, flat or gently
curved nonporous
Area: surface area of agar
Sample collected with filter sock
inserted into a vacuum hose,
connected to a HEPA filtered
vacuum pump.
Surface type: porous or nonporous,
smooth or rough
Area: typically 1 m2, but dependent
on target organism and amount of
dust present.
Sample collected with cassette attached Versatile for multiple surface
to a vacuum pump, typically set at
10–30 liters/min, though optimum
vacuum rate may vary with target
organism, dust load and surface type.
Surface type: porous or nonporous,
smooth or rough
Area: typically 100 cm2, but
dependent on target organism, type
of membrane in cassette, amount of
dust, and type of surface sampled
Surface rinse solution is sprayed on
surface and simultaneously
vacuumed up into collection bottle
Surface type: nonporous, smooth or
rough
Area: up to 10 ft2
Advantages
Good for smaller areas or Sampling area limitations
devices, corners, and crevices
of equipment or shelves
Versatile for molecular
targets and multiple
organisms; bacteria, fungi,
viruses, biological toxins
Versatile for molecular targets Sampling area limitations
and multiple organisms;
bacteria, fungi, viruses,
biological toxins
Simple, no processing needed
Versatile for multiple surface
types and larger sample areas
Best for spore-forming
bacteria, fungal spores and
fragments, biological toxins
types and larger sample areas
Best for spore-forming
bacteria, fungal spores and
fragments, biological toxins
Direct collection, no need to
elute organisms from device
Less potential for desiccation
of vegetative cells than other
vacuum methods
Limitations
Gauze wipes or sponge without
handles: fresh sterile gloves
needed for each sample
Sampling area limited to surface
area of agar
Can only detect viable
organisms capable of growth on
media
If large volume of dust collected
along with sample, detection of
target organisms may be difficult
Vacuum pump may need power
source
Vacuum pump can create airflow
that may disturb other nearby
sample locations
Not ideal for viable vegetative
cells or viruses, due to potential
for desiccation.
If large volume of dust collected
along with sample, detection of
target organisms may be difficult
Vacuum pump may need power
source
Not ideal for viable vegetative
cells or viruses, due to potential
for desiccation
Disinfection of device between
samples may be problematic
Addition of moisture to surface
not appropriate for all
applications
(Continued on next page)
 2.6.2. Surface Sampling ▪ 2.6.2-5

TABLE 2 (Continued )
Sample type/examplesa Typical application Advantages Limitations
Forensic Vacuum Sample is collected with Forensic Versatile for multiple surface High-power HEPA vacuum may
Filter Forensic Vacuum vacuum filter using HEPA vacuum types and larger sample areas compromise integrity of filter,
Filter, #FF-1, 3M (St Paul, pump. Best for spore-forming resulting in loss of sample.
MN) Surface type: porous or nonporous, bacteria, fungal spores and Vacuum rate should remain
smooth or rough. fragments, biological toxins under 600 liters/min to avoid
Area: Typically 1 m2, but dependent filter collapse
on target organism and amount of Not ideal for viable vegetative
dust present cells or viruses, due to potential
for desiccation
Bulk sample A portion of contaminated or suspect Versatile for multiple surface Need dedicated sterile cutting tool
Cut section of carpet, surface or substance on the surface is types, molecular targets and and forceps or spatula for each
HVAC filter, wallboard, aseptically packaged and sent to lab multiple organisms; bacteria, sample
or fabric for direct extraction and analysis fungi, viruses, biological No standard method for
Loose powder Surface type: porous or nonporous, toxins processing, interpretation of
smooth or rough results difficult
Area: Determined by processing
method, and what can be handled in
laboratory
a
The use of trade names and commercial sources is for identification purposes only and does not imply endorsement by CDC or the U.S. Public Health Service,
Department of Health and Human Services.
Material in this table is compiled from references (12, 13, 20, 23, 61, 86).
Key: cm = centimeter; ft2 = feet square; HEPA = high efficiency particulate air; HVAC = heating, ventilation; and air conditioning; in2 = inches square

45). Incorporating appropriate control specimens in the measured as a number of samples testing positive relative to
sampling can be helpful to monitor the efficacy of and possi- the total number of samples taken. The LOD, important
ble toxic, bacteriostatic, or fungistatic properties of the neu- whether a qualitative or quantitative result is desired, has
tralizers (21). Additionally, the sampling device materials been defined as “the lowest concentration of analyte, target,
themselves should be checked for potential antimicrobial or signature that can be consistently detected in a specified
substances (46). sample” (47). For statistical considerations, some have gone
further and defined the LOD90 or LOD95 as the lowest
Interpretation concentration detected at least 90% or 95% of the time,
Although the interpretation of a positive sample may be respectively (48). The efficiency of the method as a whole,
self-evident, an environmental surface sample that results which includes the collection, transport, elution, and analysis
in a “nondetect” does not represent a surface free of the target. assays, must be taken into account when determining
The sensitivity and limit of detection (LOD) of the sampling the LOD and when interpreting negative samples (20). The
and detection methods must be considered. The sensitivity of analytical assay plays an important part in determining the
a method is defined as the ability to detect the target when LOD, especially if only a small portion of the eluate is ana-
present and can vary depending on concentration. A meth- lyzed, as is typical when using molecular methods (such as
od’s sensitivity can be estimated for various concentrations PCR) for detection. Since no sampling method is 100%
or contamination levels in controlled laboratory studies, efficient, negative results can only be reported as below the
LOD for the method being used. When a confidence in
the quantity of organisms or target present on a surface is
required, the limit of quantitation is needed, defined as
TABLE 3 Neutralizing agentsa the lowest amount of analyte in a sample (or on a surface)
that can be determined with acceptable precision and accu-
Disinfectant Neutralizer or neutralizing media racy under the stated conditions (49). This may include
Sodium hypochlorite, Sodium thiosulfate, Dey Engley considering the statistical error inherent in low numbers,
chlorine dioxide, iodine (D/E) broth or agar (Becton such as the convention of using only between 25 and 250
Dickinson, Sparks MD) CFUs (50).
Formaldehyde, Glycine, D/E broth or agar
glutaraldehyde METHODS FOR SURFACE SAMPLING
Hydrogen peroxide Catalase The methods to collect surface samples have been organized
Phenolics Tween 80, D/E broth or agar in this section based on whether they will be used to collect
Quaternary ammonium Lecithin + Lubrol W, Letheen broth from porous or nonporous surfaces. Each sampling method,
compounds or agar (Becton Dickinson), or D/E its typical application, advantages, and limitations can be
broth or agar seen in Table 2.
Vaporized hydrogen None needed—end products H2O
Nonporous Surfaces
peroxide and O2
Nonporous surfaces are smooth, impermeable surfaces such as
a
Adapted from (45) ceramics, vinyl, stainless steel, metals, painted and coated
2.6.2-6 ▪ SAMPLING METHODS

wood surfaces, and plastics. Effective sampling of nonporous

surfaces is typically done with premoistened swabs, sponges,
or wipes or direct contact methods (7, 21, 51–53). Swabs,
sponges, and wipes also require an effective elution method
to remove the organism from the device. Elution typically
involves submerging the device in an elution liquid and
agitating by shaking, vortexing, stomaching, or sonicating.
The most effective moistening fluids and elution fluids may
vary with target organism or molecule and surface type, but
can include various buffers or general purpose broth media
(Table 4). The addition of surfactants such as Pluronic or
Tween 80 to the moistening fluid changes cell surface hydro-
phobicity (54) and interaction energies between microbes
and the sampling materials (swabs, wipes, vacuum filters,
etc.) and processing equipment (test tubes, pipettes, etc.)
(55). This can enhance the ability to disperse organisms
that are hydrophobic, such as B. anthracis (53) and Clostridium FIGURE 1 Examples of swabs used for surface sampling, from top:
difficile spores, and reduce adherence of organism to sampling polyester, cotton, foam, large foam tip with shaft attached to cap for
and processing equipment. transport tube. doi:10.1128/9781555818821.ch2.6.2.f1
Direct contact methods do not require elution from a
device; rather, the organisms are collected by contact of the
device directly with the surface. Examples of direct contact
methods include agar contact or replicate organism detection hard-to-reach or unusually shaped areas, such as corners
and counting plates (RODAC, Beckton Dickinson, Franklin and crevices, and areas inaccessible to larger devices. Many
Lakes, NJ), Petrifilm device (3M Medical-surgical division, swab types are available, with absorbent materials made of
St. Paul, MN), membrane filter (7), or adhesive tape (tape-lift calcium alginate, rayon, cotton, polyester, or foam (Fig. 1).
method, 56). The efficiency of each material may vary for different target
organism, since adherence to the environmental surface
Swabs and adherence to the swab material are determined by the
physical and chemical interactions of the organism with the
Swabs are typically used to sample small areas 26 cm2 to 100 surface and the swab material (9, 55). If molecular analysis
cm2 (8, 21, 23, 27, 57), though some studies were conducted will be performed directly from the swab eluate, potential
using 1, 5, and 6 cm2 (58–60). Swabs are convenient for assay inhibitors from the swab materials must be considered.
The use of calcium alginate swabs, which dissolve and release
collected microbes on submersion in sodium hexametaphos-
phate, may be appropriate for some applications. However,
TABLE 4 Examples of eluents and diluents for environmental the use of calcium alginate/hexametaphosphate has been
surface samplinga reported to be inhibitory of some species (2, 5) and for this
reason is not widely used.
Solutions Concentration in water Premoistened swabs have been shown to perform better
Ringer ¼ strength than dry swabs at removing spores and other organisms from
Peptone water 0.1%–1.0% steel (9, 53) plastic, wood, and cotton cloth (60) and viruses
from PVC and stainless steel (61). A number of wetting
Buffered peptone water 0.067 M phosphate, 0.43%
agents can be used, such as sterile water, buffered water, or
NaCl, 0.1% peptone
saline (Table 4). Wetting agents containing 0.2–0.02% sur-
Phosphate-buffered saline 0.02 M phosphate, 0.9% factant such as Polysorbate (Tween 80 or Tween 20) or polox-
NaCl amers (Pluronic) can enhance recovery of organisms from a
Sodium chloride (NaCl) 0.25%–0.9% surface. Premoistening swabs with a solution containing a
Calgon ringerb ¼ strength neutralizer is advised when sampling surfaces were previously
Thiosulfate ringerc ¼ strength cleaned with a disinfectant (Table 3). Before choosing a
premoistening liquid, the laboratory should be consulted to
Minimal essential media (MEM)d Full strength
ensure compatibility with the assay methods that will be
Viral collecting brothe Full strength used to analyze the sample in the laboratory. If a quantita-
Tryptic soy broth (TSB) Full strength tive assessment is being conducted, ensure that the premois-
Brain-heart infusion broth (BHI) Full strength tening fluid does not contain nutrients that would promote
supplemented with 0.5% beef growth during transport to the laboratory. Pre-moistened
extract water swabs are moved across the sampling area in a systematic man-
a
ner, typically back and forth in one direction, then again in a
A surfactant (e.g., polysorbate [i.e., Tween 80]) may be added to eluents direction 90° from the first, then diagonally (37). Upon com-
and diluents. A concentration ranging from 0.01% to 0.1% is generally used,
depending on the specific application. Foaming may occur during use. pleting the sampling, the swab should be aseptically placed
b
Used for dissolution of calcium alginate swabs. into a tube of saline, buffer, or appropriate transport medium
c
d
Used for neutralization of residual chlorine. and maintained at the appropriate temperature during
Has been used to sample for viruses (36). shipment to the lab. Though transport tubes with semisolid
e
Brain-heart infusion broth with 1% bovine serum and antibiotics; has
been used to sample for viruses (122). transport media are commercially available and typically
Unless otherwise noted, material in this table is compiled from (21), (46), used for clinical specimens, if quantitation of the sample is
and (36). a goal of sampling, this type of transport system is not

recommended because cells are lost to the semisolid agar
when removing the swabs at the laboratory.
Standardizing the sampling technique is difficult because
much of the technique is dependent on the sampler, for
example, the degree of pressure applied, the amount of time
spent sampling the area, the degree of swab rotation, and so
on. This variability between individuals can be minimized
with an effective training program.
Wipes and Sponges
Wipes and sponges can be used to sample slightly larger non-
porous surface areas than swabs, simply because of the larger
size of the wipe or sponge, but the area should still be limited
and/or defined for each investigation. Typical areas sampled
are between 625 cm2 (28) and 1 m2 (19), with some studies
conducted with other surface areas (100.4 cm2, 57). Kirshner
and Puleo (62) found that fewer organisms were recovered
from 1.49 m2 than from 0.74 m2. They suggest that because
the wipe (clean room cloths) dries out, the efficiency is
reduced with the larger surface area. Sponges and wipes can
be made from rayon, polyester, or cotton fabric or gauze,
cellulose, or polyurethane materials (Fig. 2). Herzog et al.
(63) also found that the recovery efficiency was reduced when
sampling bacteriophage P22 from 1,000 cm2 vs. 100 cm2 sur-
faces using moistened fabric wipes. The recovery was reduced
by 25%, 40%, and 50% for the larger acrylic, steel, and lam-
inate surfaces, respectively. Some wipe materials are available
premoistened, sterile, and attached to a handle. Alterna-
tively, the wipe or sampling sponge can be aseptically pre-
moistened immediately before sampling. As with swabs,
premoistening generally improves the collection and recovery
efficiency of the wipe and sponge (60), but the premoistening
fluid should be chosen to be compatable with the detection
assay, and not enhance or inhibit growth during transport
to the laboratory. There are times, however, that adding
moisture to the surface would not be appropriate, such as in
a food processing area in which added moisture would encour-
age growth of bacteria (9). Typically a wipe or sponge is pre-
moistened and then handled aseptically while moving it
systematically across the area to be sampled in a horizontal,
then vertical, then diagonal motion, folding or turning the
wipe with each new direction so as to expose a new surface
FIGURE 2 Wipe materials, clockwise from top left: premoistened
cellulose sponge, gauze wipe, Sponge stick.
doi:10.1128/9781555818821.ch2.6.2.f2
2.6.2. Surface Sampling ▪ 2.6.2-7
of the wipe. The wipe or sponge is then placed into a sterile
specimen container (e.g., tube, cup, jar, or specimen bag)
for transport to the laboratory. If the target organism is sensi-
tive to degradation during transport, a transport liquid or
medium and ice packs should be considered. Elution of the
organisms from the sponge or wipe can be accomplished
by shaking or stomaching while submerged in a sterile buffer
or broth (Table 4).
Commercially available polyester electrostatic cloths
have also been evaluated and found effective at recovery of
Staphylococcus aureus (64) and household dust for endotoxin
testing (65). Another type of sponge device evaluated for
sampling Bacillus atrophaeus spores over a 1 m2 area is the
Biological Sample Kit or BisKit (QuickSilver Analytics, Bel-
camp, MD) (29). The device consists of a foam sponge held in
a plastic casing and attached to the lid of the casing. A bottle
of buffer is attached to the container to allow dispensing
directly onto the sponge to hydrate the sponge and express
the microbes from the sponge while still in the casing.
Contact Agar Plate Sampling
The contact agar plate imprint technique consists of pressing
an agar plate onto a solid suface to directly pick up organisms
onto the agar. The agar plate (20 cm2–26 cm2) is convex on
the bottom and overfilled on the top so that the agar surface
extends slightly beyond the side walls of the plate, enabling
contact with the surface being sampled. Because the sampler
simply presses the plate to the surface and lifts (rather than
wipes), the surface area being sampled is limited to the
surface area of the agar plate, and this works best on relatively
clean surfaces (66, 67), as the target organisms can be masked
by growth of other organisms. The agar plate can then be
incubated under the appropriate conditions and inspected
for growth of the target organisms. This method has been
evaluated for recovery of a variety of nosocomial pathogens,
biothreat agents, and foodborne pathogens (68–70). Quanti-
tative data from contact agar plates may not be as accurate as
data obtained from swabs or wipes, since contact agar plates
will also pick up microcolonies and aggregates of organisms
that can mask a target organism on incubation and growth,
whereas when samples are collected with swabs or wipes the
aggregates and microcolonies can be disbursed during pro-
cessing (2). Some studies have found contact agar plates to
be superior to swabs and wipes when presented as the percent
of sites positive for a given organism (71–73). These studies
were conducted in a hospital setting by sampling adjacent
sites that were presumed to be contaminated equally. In
another controlled study, the contact plate method was found
more sensitive when CFUs were compared to CFUs from a
swab sample streaked across an agar plate (74). However, a
study of food preparation surfaces concluded the use of swab
sampling to be more accurate in predicting the surface bacte-
rial populations than the contact plate method (75). In these
comparison studies, the recovery efficiency of the swab
method was shown to vary with species and strain (73, 74).
Neutralizing agents can be incorporated into the agar sam-
pling media for sampling after cleaning and disinfection
of a surface, but choosing the neutralizer specific to the
disinfectant and at minimal concentration is advisable to
avoid inhibitory effects on the target organism (76). The con-
tact plate method is less labor intensive than other methods,
is able to detect a range of organism with the use of a variety
of selective media, and requires only the use of an incubator
after sampling. This method can be a useful tool to esti-
mate the contamination of generally clean surface, though
2.6.2-8 ▪ SAMPLING METHODS
preliminary work should be undertaken to ensure it is effec-
tive and efficient for a given target organism or application.
Dip Slides and Petrifilm
Two additional approaches that are similar to contact agar
plate sampling include the use of dip slides and Petrifilm
(3M) for sampling. Petrifilm and dip slide (e.g., Enviro-
check, VWR, West Chester, PA; Solar-Cult, Solar Biologi-
cals, Ogdenburg, NY, Hygicult, Orion Diagnostica, Espoo,
Finland) sampling are commonly used in the food industry
to monitor surfaces for contamination and occasionally in
clinical settings.
Dip slides were developed for the detection of microbes in
liquids, but have been adopted for use in sampling surfaces.
Dip slides are hinged plastic paddles covered with a culture
medium on both sides. The paddle is enclosed in a plastic
tube and attached to the tube cap. The paddle can be held
aseptically by the cap and pressed onto a surface like a contact
agar plate, or it can be inoculated by streaking a swab across
the agar on the paddle. The paddle is then placed back into
its sterile tube for incubation. The paddles are easy to use
but allow for a limited surface area to be sampled, since
each side of the paddle is approximately 12 cm2. A multilab
validation study compared dip slide sampling to contact plate
and swab sampling and found little difference in percent
recovery and precision between the methods (77).
Petrifilm plates are intended for use as a cost-effective,
convenient replacement for conventional spread plating
methods. Some investigators began using them for sampling
as well (78, 79). They consist of a thin layer of dehydrated
media loaded into a paper and clear film holder. The media
is rehydrated with 1 ml buffered water, and allowed to sit
for 15 min. At this point the Petrifilm plate can be used as
a contact agar plate, by pressing gently and evenly onto the
surface to be sampled. Upon removal from the surface, the
clear plastic film is carefully rolled over the agar surface to
seal in the moisture, then incubated for the time and temper-
ature appropriate for the target organism (78).
Tape Lift
Tape lift samples of dust are typically used to visually provide a
semi-quantitative analysis of the genera of fungal spores
present on a surface (80, 81). This method involves imprint-
ing the sticky side of celluloid tape to a contaminated surface
and lifting to pick up adhered fungal spores. An experienced
mycologist can stain and examine the tape microscopically to
identify the genus from the morphology of the fungal struc-
tures and/or spores. This method can provide an estimate of
mold contamination but cannot provide information on via-
bility of the organisms (82, 67). Bacterial sampling can also be
conducted with an adhesive sheet and the adhesive sheet
stained for fluorescent microscopy or the DNA extracted
and analyzed by PCR or nucleotide sequencing (56, 83).
Immersion or Containment Rinsing
Small items such as medical devices may be sampled by
directly immersing in an elution fluid and shaking or sonicat-
ing to dislodge the microbes (21). Similarly, a device or piece
of equipment may be filled with fluid, brushed with a sterile
brush (84) or closed to contain the fluid, then agitated to dis-
lodge microbes, and the fluid decanted into a sterile vessel for
analysis (21).
Porous Surfaces
Porous surfaces are typically surfaces that are permeable to
water, air, or other fluids and include carpets, draperies,
upholstery, and ceiling tiles. Concrete and asphalt may be
included as porous surfaces because, although they may not
be totally permeable, the surfaces are irregular and contain
pores. Porous surfaces present additional challenges to sam-
ple. Though swabs and wipes can be used on porous surfaces,
the efficiency of these devices is diminished because of the
characteristics of the surface (30, 60). Persistent bacteria,
bacterial spores, and fungal spores deposited onto porous or
nonporous surfaces can be recovered with various vacuum
devices (23, 30, 60, 85).
Vacuum Filter Sock
Vacuum filter socks such as the X-Cell 100 (Midwest Filtra-
tion, Cincinnati, OH, Fig. 3), were developed for collection
of dust and allergens (86), but were used for collection of
B. anthracis spores during the 2001 anthrax mail contamina-
tion incident (24, 25). The filter socks can be used to collect
dust and bulk material, which is then sent for analysis. The
filter socks are not HEPA-quality filters, but are used with
vacuums containing internal HEPA filters (such as the Atrix
Express HEPA vacuum, Burnsville, MN) that help reduce
the reaerosolizing of dust and organisms downstream from
the collection. Because the filter sock is not a HEPA collec-
tion device, the collection efficiency of bacteria may be
low due to initial loss of sample passing through the sock.
One study demonstrated a 3–6% recovery of a low level
(1.8–270 CFU/100 cm2) of B. anthracis spores from a stainless
steel surface using the filter socks (30). Another study
reported a mean 29% recovery from stainless steel when the
inoculum was higher (2–5 log10 CFU/cm2) (31).
Calfee et al. (87, 88) compared filter socks to other vac-
uum devices for collection of B. atrophaeus spores and found
the recoveries varied with the speed of the sampling and
the materials inoculated (carpet, concrete, and upholstery,
HVAC filter). Prezant et al. (89) found that though the socks
collected more dust than the cassette filters, the cassette filters
yielded higher CFUs per gram of dust, suggesting that the
socks collect larger particles but allow the smaller organisms
to pass through. One advantage of the filter sock method is
that a larger area can be sampled than with a swab or wipe;
but if a large amount of dust is also collected, culture and
detection of the target organism may be more difficult. To
minimize the presence of background organisms and inhibi-
tors, the sample area should be limited.
FIGURE 3 Left to right: vacuum sock loaded into cardboard
holder with sampling nozzle, transport cup, and vacuum sock before
loading into holder. doi:10.1128/9781555818821.ch2.6.2.f3

The vacuum filter socks are not sterile as purchased. For
ease of handling, they are available as a kit preassembled in
a disposable cardboard holder with a nozzle (Fig. 3). With
this filter sock kit, the sampler simply inserts the device
into hose of the portable HEPA vacuum and moves the nozzle
across the designated sample area in a systematic manner
to cover the entire surface. After vacuuming, the nozzle is
removed; the sock filter is aseptically removed from the card-
board holder, the opening of the sock is secured with a clip
or zip tie, and the filter is inserted into a sterile specimen
cup or sealable plastic bag for transport to the laboratory
(90). Gloves should be changed between samples, and the
vacuum hose should be disinfected or changed between
samples.
Microvacuum
The microvacuum cassette or closed-face filter cassette was
developed to monitor total particulates in air, and later
used to monitor the breathing zone of industrial workers
(91). Surface sampling for organisms can be conducted in a
similar manner as described in the ASTM standard method
D7144, which was developed for surface dust sampling
intended for subsequent metals detection (92). The method
uses 25 or 37 mm diameter filters made of polytetraflorene
(PTFE) or mixed cellulose ester (MCE) pre-loaded into a
sealed plastic holder or cassette such as seen in Fig. 4. Calfee
et al (87, 88) found 0.8 µm pore size MCE superior to 0.3 µm
pore size PTFE 37 mm cassettes for recovery of B. atrophaeus
spores from upholstery, concrete, and carpet, and equivalent
to vacuum filter socks when sampling B. atrophaeus spores
from HVAC filters. No standard vacuum rate has been estab-
lished for surface sampling with these devices, but a vacuum
pump of sufficient power is needed to collect the target organ-
ism without causing breakthrough of the sample or damage
to the collection filter. A vacuum pump meant for personal
airspace sampling (1–4 liters/min) may not be powerful
enough for surface sampling, but 20 liters/min worked well
for bacterial spores (87, 88). Prezant et al. (61) suggest a
standard 28.3 liters/min pump for collection of dust contain-
ing fungal spores. Air sampling pumps that compensate for
lower flow on increased loading will adjust the power to main-
tain the vacuum rate. This can be detrimental for surface
sampling, in that the pump will shut off completely if the
loading is too great to maintain the designated flow rate.
The ASTM method D7144 also specifies use of a sampling
nozzle for the cassettes made out of Tygon tubing with one
FIGURE 4 37 mm cassette with filter inside, sampling nozzle and
vacuum hose adapter attached.
doi:10.1128/9781555818821.ch2.6.2.f4
2.6.2. Surface Sampling ▪ 2.6.2-9
end cut at a 45° angle. This type of nozzle was used by Ashley
et al. (93) to collect surface dust, and by Calfee et al. (87, 88)
to collect B. atrophaeus spores.
The collection and recovery efficiency of these devices
are dependent on airflow (air currents or wind at the site),
pump flow rate, relative humidity, dust presence, and filter
type (91, 94, 95) as well as the electrostatic charge of the
environmental surface (96). Wang et al. (96) reported on
the factors affecting the efficiency of vacuum sampling of
dust from carpets and notes that the collection efficiency
was optimum between 60% and 85% relative humidity and
that more dust was recovered from low loop carpets than
from shag carpets. No standard surface area for sampling has
been established for these devices, but areas reported in the
literature range from 0.1 to 3 m2 (88, 89).
Other Vacuum Devices
The Microbial-Vac (M-Vac, Microbial-Vac Systems, Sandy,
UT) is a hand-held wet-vacuum sampler that may be useful
in some situations. The M-Vac sprays sterile solution onto
the surface being sampled to penetrate pores and dislodge
microbes, while simultaneously vacuuming up the microbes,
DNA, and particles into a collection bottle. One study
reported 2–10 times higher recovery efficiencies with the
M-Vac when compared with swabs and wipes for the recovery
of B. subtilis spores and Escherichia coli from various surfaces
(85). Another study found the M-Vac sensitivity ( percent
of samples target detected) to be statistically lower than
that of swabs and sponges when sampling Listeria monocyto-
genes from four surface materials (97). Advantages of this
method are the ease of use and the reduction or elimination
of sample preparation. The device is rinsed with hot water
between samples to disinfect the sample pathway, though
this may not be sufficient for some sample types. As was men-
tioned with premoistened swab and wipe sampling, consider
the consequences of adding liquid to the surface being
sampled, as amplification of remaining organisms may occur.
Household portable bag-type vacuum cleaners have been
used to assess house dust for fungal contamination by using
the vacuum bag as a collection device. Pinard and others
(98) recommended sampling from surfaces other than the
floor to minimize sampling tracked-in fungal spores and to
vacuum 1–2 m2 of each surface for 5 min. These samples
can be maintained at 4°C for up to 6 days before culturing.
Hung et al. (99) found that vacuums with bags made of
high-retention materials such as DuPont Hysurf provided
higher yields of allergen when compared with other standard
commercial vacuum bags. When analyzing the dust for
fungi, some analysts sieve the dust to remove large particles
before sprinkling directly on agar culture media or mixing
with a buffer and culturing (100).
Bulk Sampling
Bulk sampling or excision sampling is used by removing a
section of the surface for analysis. This is a destructive method
and therefore not widely used. Examples of bulk samples
include aseptically cutting sections of a carpet, upholstery,
clothing, animal carcass, or a paper food container. The sec-
tion is typically placed in a sterile container for shipment
to the laboratory for analysis. An elution buffer or liquid is
added and the sample shaken, vortexed, or placed in a paddle
blender such as a Stomacher (Seward Laboratory Systems,
Davie, FL) to dislodge the organisms for culture or molecular
detection. Communication with the lab is critical before
sending this type of sample, because few clinical or public
health labs are equipped to analyze them. Sampling in this
2.6.2-10 ▪ SAMPLING METHODS
manner is not endorsed by any agency because the results
from this type of sampling are difficult to interpret.
Bulk sampling can also refer to collection of a substance on
a surface, such as powders, dust, or other particulates. A stand-
ard method for collection of bulk powders suspected of being
a biothreat agent was developed in 2010 (101). The method
involves the use of a dry swab to push the powder onto a lami-
nated card, which is then placed into a wide-mouth specimen
container. The remaining residual powder is then collected
with a sterile moistened swab.
FIELD AND LAB BLANKS
To ensure no cross-contamination has occurred while sam-
pling, field blanks should be collected at the same time as
collecting the specimen. Collecting a field blank involves
removing the sampling device from its sterile packaging
(while in the same environment as the actual samples are col-
lected), then placing it in the appropriate sterile transport
container (without touching any surfaces). This will inform
the sampler of the potential for cross-contamination of the
samples due to mishandling or environmental conditions.
Generally, one field blank for every 10 surface samples is rec-
ommended (37). Good laboratory practice also includes
inserting laboratory blank samples into the workflow. Labo-
ratory blanks are unused, previously unopened sample mate-
rials (swabs, wipes, vacuum filters, etc.) that are processed
and analyzed alongside the surface sampled and field blank
devices.
PACKAGING AND TRANSPORT
Timely analysis of the sample is important for accurate results,
so the means and conditions of transport should be part of the
sampling plan. Samples can undergo extreme temperature
fluctuations during shipping due to seasonal temperatures,
high altitude transport conditions, or nonoptimal packaging.
Optimal temperatures, transport media, and hold times can
vary by organism type, genus, or even species, but packaging
with sufficient cold packs to maintain temperatures close to
5°C is generally recommended (9, 37, 102–104).
If samples contain pathogenic organisms or toxic sub-
stances, they should be packaged and shipped according
to appropriate International Air Transport Association
(IATA) Dangerous Goods Regulations (105) and Code of
Federal Regulations 49 CFR Parts 171–180 (106). Regula-
tions on transporting infectious substances through the U.S.
Postal Service are codified in publication 52—Hazardous,
Restricted, and Perishable Mail (107)—and Section 135 of
the International Mail Manual (108). Shipping samples con-
taining select agents require compliance with the federal
select agent regulations (109).
CONCLUDING REMARKS
Though microbiologic sampling of the environment appears
at first glance to be a simple task, each of the sampling
strategies and methods discussed in this chapter are complex,
and the success of the sampling endeavor depends on metic-
ulous attention to details in sampling design and aseptic tech-
nique. Environmental sampling may never be extremely
precise and accurate because of all the variables that come
into play, such as the organisms’ interactions with the surface
materials, dust, organic matter, and other organisms, as well as
environmental conditions such as temperature, humidity,
and air movement. Minimizing the variables in collection,
transport, and analysis will help researchers obtain better
information from surface samples. A good sampling strategy
appropriate for the purpose of the investigation, a consistent
approach, understanding the limitations of the different
sampling methods, and knowledge of the target organism
are the most important factors in obtaining the best pos-
sible information from a microbiologic environmental sam-
pling event.
REFERENCES
1. Angelotti R, Foter MJ, Busch KA, Lewis KH. 1958.
A comparative evaluation of methods for determining
the bacterial contaminaiton of surfaces. Food Res 23:
175–185.
2. Baldock JD. 1974. Microbiological monitoring of the food
plant: methods to assess bacterial contamination on surfa-
ces. J Milk Food Technol 37:361–368.
3. Tredinnick JE, Tucker J. 1951. The use of calcium algi-
nate wool for swabbing dairy equipment. J Appl Microbiol
14:85–88.
4. Tiedeman WD, Fuchs AW, Gunderson NO, Hucker GJ,
Mallman WL. 1948. A proposed method for control of food
utensil sanitation. Am J Pub Health Nat Health 38:68–70.
5. Patterson JT. 1971. Microbiological assessment of surfaces.
J Food Technol 6:63–72.
6. Hammer BW, Olson HC. 1931. Bacteriology of Butter III. A
Method for Studying the Contamination from Churns. Iowa
Agricultural Experiment Station Bulletin, Ames, IA.
7. Craythorn JM, Barbour AG, Matsen JM, Britt MR, Gar-
ibaldi RA. 1980. Membrane filter contact technique for
bacteriological sampling of moist surfaces. J Clin Microbiol
12:250–255.
8. Evancho G, Sveum W, Moberg L, Frank J. 2001. Micro-
biological monitoring of the food processing environment,
p. 25–35. In Downes FP, Ito K (eds), Compendium of Meth-
ods for the Microbiological Examination of Foods, 4th ed.
American Public Health Association, Washington, DC.
9. Holah JT, Staff Campden & Chorleywood Food
Research Association. 1999. Effective Microbiological Sam-
pling of Food Processing Environments. Campden & Chorley-
wood Food Research Association, Gloucestershire, UK.
10. National Aeronotics and Space Administration. 2010.
Handbook for the Microbiological Examination of Space
Hardware, NASA-HDBK-6022. NASA, Washington, DC.
11. Favero MS. 1971. Microbiologic assay of space hardware.
Environ Biol Med 1:27–36.
12. Puleo JR, Oxborrow GS, Fields ND, Herring CM, Smith
LS. 1973. Microbiological profiles of four Apollo space-
craft. Appl Microbiol 26:838–845.
13. Puleo JR, Fields ND, Bergstrom SL, Oxborrow GS, Sta-
bekis PD, Koukol R. 1977. Microbiological profiles of the
Viking spacecraft. Appl Environ Microbiol 33:379–384.
14. Committee on Space Research. 2002. COSPAR Planetary
Protection Policy, October 2002 ed. COSPAR/IAU Work-
shop on Planetery Protection, Paris, France.
15. Anonymous. 2009. General Chapter <1116>, Microbio-
logical evaluation of clean rooms and other controlled
environments. In United States Pharmacopeial Conven-
tion (ed), United States Pharmacopeia, vol. 32. United States
Pharmocopeia, Rockville, MD.
16. Anonymous. 2009. General Chapter <797>, Pharmaceut-
ical compounding: sterile preparations, p. 318–349. In
United States Pharmacopeial Convention (ed), United States
Pharmocopeia, vol. 32. United States Pharmocopeia,
Rockville, MD.
17. Sutton S. 2007. Microbial surface monitoring, p. 73–92. In
Dixon AM (ed), Environmental Monitoring for Cleanrooms

and Controlled Environments. Informa Healthcare USA,
New York.
18. Eickhoff TC. 1974. Microbiologic sampling of the hospital
environment. Health Lab Sci 11:73–75.
19. Gilchrist MJR. 1992. Microbiological culturing of envi-
ronmental and medical-device surfaces, p. 11.10.14. In
Isenberg HD (ed), Clinical Microbiology Procedures Hand-
book. American Society for Microbiology Press, Washing-
ton, DC
20. Sehulster LM, Rose LJ, Noble-Wang J. 2012. Microbio-
logic sampling of the environment in healthcare facilities,
p. 1059–1075. In Mayhall GC (ed), Hospital Epidemiology
and Infection Control. Lippincott Williams & Wilkins,
Philadelphia, PA.
21. Bond WW, Sehulster LM. 2004. Microbiological cultur-
ing of environmental and medical-device surfaces,
p 13.10.11–13.10.12. In Isenberg HD, Miller JM, Bell M
(eds), Clinical Microbiology Procedures Handbook. American
Society for Microbiology Press, Washington, DC
22. Denton M, Wilcox MH, Parnell P, Green D, Keer V,
Hawkey PM, Evans I, Murphy P. 2004. Role of environ-
mental cleaning in controlling an outbreak of Acinetobacter
baumannii on a neurosurgical intensive care unit. J Hosp
Infect 56:106–110.
23. Busher A, Noble-Wang J, Rose L. 2008. Surface sampling,
p. 95–131. In Emanuel P, Roos JW, Niyogi K, (eds), Sam-
pling for Biological Agents in the Environment. American
Society for Microbiology Press, Washington, DC.
24. Edmonds JM. 2009. Efficient methods for large-area sur-
face sampling of sites contaminated with pathogenic micro-
organisms and other hazardous agents: current state, needs,
and perspectives. Appl Microbiol Biotechnol 84:811–816.
25. Sanderson WT, Hein MJ, Taylor L, Curwin BD, Kinnes
GM, Seitz TA, Popovic T, Holmes HT, Kellum ME,
McAllister SK, Whaley DN, Tupin EA, Walker T, Freed
JA, Small DS, Klusaritz B, Bridges JH. 2002. Surface sam-
pling methods for Bacillus anthracis spore contamination.
Emerg Infect Dis 8:1145–1151.
26. Teshale EH, Painter J, Burr GA, Mead S, Wright SV,
SCseh LF, Zabrocki R, Collins R, Kelly KAH, JL, Swer-
dlow DL, Team AR. 2002. Environmental sampling for
spores of Bacillus anthracis. Emerg Infect Dis 8:1083–1087.
27. Hodges LR, Rose LJ, O’Connell H, Arduino MJ. 2010.
National validation study of a swab protocol for the recov-
ery of Bacillus anthracis spores from surfaces. J Microbiol Meth
81:141–146.
28. Rose LJ, Hodges L, O’Connell H, Noble-Wang J.
2011. National validation study of a cellulose sponge
wipe-processing method for use after sampling Bacillus
anthracis spores from surfaces. Appl Environ Microbiol 77:
8355–8359.
29. Buttner MP, Cruz P, Stetzenbach LD, Klima-Comba AK,
Stevens VL, Emanuel PA. 2004. Evaluation of the Biolog-
ical Sampling Kit (BiSKit) for large-area surface sampling.
Appl Environ Microbiol 70:7040–7045.
30. Estill CF, Baron PA, Beard JK, Hein MJ, Larsen LD,
Rose L, Schaefer FW III, Noble-Wang J, Hodges L,
Lindquist HD, Deye GJ, Arduino MJ. 2009. Recovery
efficiency and limit of detection of aerosolized Bacillus
anthracis Sterne from environmental surface samples. Appl
Environ Microbiol 75:4297–4306.
31. Brown GS, Betty RG, Brockmann JE, Lucero DA, Souza
CA, Walsh KS, Boucher RM, Tezak MS, Wilson MC.
2007. Evaluation of vacuum filter sock surface sample
collection method for Bacillus spores from porous and
non-porous surfaces. J Environ Monit 9:666–671.
32. Brown GS, Betty RG, Brockmann JE, Lucero DA, Souza
CA, Walsh KS, Boucher RM, Tezak M, Wilson MC,
Rudolph T. 2007. Evaluation of a wipe surface sample
2.6.2. Surface Sampling ▪ 2.6.2-11
method for collection of Bacillus spores from nonporous
surfaces. Appl Environ Microbiol 73:706–710.
33. Getchell-White SI, Donowitz LG, Groschel DH. 1989.
The inanimate environment of an intensive care unit as a
potential source of nosocomial bacteria: evidence for long
survival of Acinetobacter calcoaceticus. Infect Ctrl Hosp Epi-
demiol 10:402–407.
34. Weber DO, Gooch JJ, Wood WR, Britt EM, Kraft RO.
1976. Influence of operating room surface contamination
on surgical wounds: a prospective study. Arch Surg 111:
484–488.
35. Scherer K, Mäde D, Ellerbroek L, Schulenburg J, Johne
R, Klein G. 2009. Applicaton of a swab sampling method
for the detection of Norovirus and Rotavirus on artificially
contaminated food and environmental surfaces. Food Envi-
ron Virol 1:42–49.
36. Keswick BH, Pickering LK, DuPont HL, Woodward
WE. 1983. Survival and detection of rotaviruses on envi-
ronmental surfaces in day care centers. Appl Environ Micro-
biol 46:813–816.
37. Centers for Disease Control and Prevention. 2012. Sur-
face sampling procedures for Bacillus anthracis spores from
smooth, non-porous surfaces. http://www.cdc.gov/niosh/
topics/emres/surface-sampling-bacillus-anthracis.html.
38. National Response Team. 2003. National response team
technical assistance for anthrax response: interim final
draft. Updated July 2005. http://yosemite.epa.gov/sab/sab
product.nsf/17A8EBA9CAA030468525758D006D7A1B/
$File/HSAC+Anthrax+TAD+2005+4-21-09+Meeting.pdf.
39. Chosewood LC, Wilson DE, (eds). 2009. Biosafety in
Microbiological and Biomedical Laboratories, 5th ed, HHS
publication no. (CDC) 21–1112. http://www.cdc.gov/bio
safety/publications/bmbl5/.
40. U.S. Environmental Protection Agency. 2002. Guidance
on Choosing a Sampling Design for Environmental Data Col-
lection. EPA QA/G-5S. U.S. Environmental Protection
Agency, Washington, DC.
41. Gillen M. 2008. The role of sampling in the phases of a
biologial event: fact and fiction in an airport scenario,
p. 73. In Emanuel P, Roos JW, Niyogi K (eds), Sampling
for Biological Agents in the Environment. American Society
for Microbiology Press, Washington, DC.
42. Government Accountability Office. 1992. Using Statis-
tical Sampling. GAO/PEMD-10.1.6. U.S. Government
Accountability Office, Washington DC.
43. Thompson SK. 2012. Sampling. Wiley, Hoboken, NJ.
44. Sego LH, Shulman S, Anderson KK, Wilson JE, Pulsi-
pher BA, Sieber WK. 2010. Acceptance sampling Using
Judgmental and Randomly Selected Samples. PNNL-19315.
Pacific Northwest National Laboratory, Richland, WA.
45. Russell AD. 2001. Principles of antimicrobial activity
and resistance, p. 31–56. In Block SS (ed), Disinfection,
Sterilization, and Preservation, 5th ed. Lippincott, Williams
& Wilkins, Philadelphia, PA.
46. American National Standard. 2011. Sterilization of health
care products - Microbiological methods-Part 1:Determination
of the population of microorganisms on product. ANSI/
AAMI/ISO 11737-1:2006/(R)2011. Association for the
Advancement of Medical Instrumentation, Arlington,
VA.
47. Budowle B, Schutzer SE, Morse SA, Martinez K,
Chakraborty R, Marrone BL, Messenger SL, Murch
RS, Jackson PJ, Williamson P, Harmon R, Velsko SP.
2008. Criteria for validation of methods in microbial
forensics. Appl Environ Microbiol 74:5599–5607.
48. Krauter PA, Piepel GF, Boucher R, Tezak M, Amidan
BG, Einfeld W. 2012. False-negative rate and recovery effi-
ciency performance of a validated sponge wipe sampling
method. Appl Environ Microbiol 78:846–854.
2.6.2-12 ▪ SAMPLING METHODS
49. Anonymous. 1980. General chapters: validation of
compendial methods, p. 549. In U.S. Pharmacopeial Con-
vention (ed), United States Pharmacopeia, vol. 31, U.S.
Pharmacopeial Convention, Rockville, MD. http://www.
pharmacopeia.cn/v29240/usp29nf24s0_c1225.html.
50. Tomasiewicz DM. 1980. Most suitable number of colonies
on plates for counting. J Food Prot 43:282–286.
51. Obee P, Griffith CJ, Cooper RA, Bennion NE. 2007. An
evaluation of different methods for the recovery of
methicillin-resistant Staphylococcus aureus from environ-
mental surfaces. J Hosp Infect 65:35–41.
52. Poletti L, Pasquarella C, Pitzurra M, Savino A. 1999.
Comparative efficiency of nitrocellulose membranes versus
RODAC plates in microbial sampling on surfaces. J Hosp
Infect 41:195–201.
53. Rose L, Jensen B, Peterson A, Banerjee SN, Srduino MJ.
2004. Swab materials and Bacillus anthracis spore reco-
very from nonporous surfaces. Emerg Infect Dis 10:
1023–1029.
54. Brown DG, Jaffè PR. 2006. Effects of nonionic surfactant
on the cell surface hydrophobility and apparent Hamaker
constant of a Sphigomonas sp. Environ Sci Techol 40:
195–201.
55. Da Silva SM, Filliben JJ, Morrow JB. 2011. Parameters
affecting spore recovery from wipes used in biological sur-
face sampling. Appl Environ Microbiol 77:2374–2380.
56. Ichijo T, Hieda H, Ishihara R, Yamaguchi N, Nasu M.
2013. Bacterial monitoring with adhesive sheet in the
international space station—“Kibo”, the Japanese experi-
ment module. Microbes Environ 28:264–268.
57. Valentine NB, Butcher MG, Su Y-F, Jarman KH,
Matzke M, Webb-Robertson B-J, Panisko EA, Seiders
BAB, Wahl KL. 2008. Evaluation of sampling tools for
environmental sampling of bacterial endospores from
porous and nonporous surfaces. J Appl Microbiol 105:
1107–1113.
58. Buttner MP, Cruz-Perez P, Stetzenbach LD. 2001.
Enhance detection of surface-associated bacteria in indoor
environments by quantitative PCR. Appl Environ Microbiol
67:2564–2570.
59. Brown GS, Betty RG, Brockmann JE, Lucero DA, Souza
CA, Walsh KS, Boucher RM, Tezak MS, Wilson MC,
Rudolf T, Lindquist HAD, Martinez KF. 2007. Evalua-
tion of rayon swab surface sample collection method for
Bacillus spores from nonporous surfaces. J Appl Microbiol
103:1074–1080.
60. Frawley DA, Samaan MN, Bull RL, Robertson JM,
Mateczun AJ, Turnbull PCB. 2008. Recovery efficiencies
of anthrax spores and ricin from nonporous or nonabsorb-
ent and porous or absorbent surfaces by a variety of sam-
pling methods. J Forensic Sci 53:1102–1107.
61. Julian TR, Tamayo FJ, Leckie JO, Boehm AB. 2011.
Comparison of surface sampling methods for virus recovery
from fomites. Appl Environ Microbiol 77:6918–6925.
62. Kirschner LE, Puleo JR. 1979. Wipe-rinse technique for
quantitating microbial contamination on large surfaces.
Appl Environ Microbiol 38:466–470.
63. Herzog AB, Pandey AK, Reyes-Gastelum D, Gerba CP,
Rose JB, Hashsham SA. 2012. Evaluation of sample recov-
ery efficiency of bacteriophage P22 on fomites. Appl Envi-
ron Microbiol 78:7915–7922.
64. Lutz JK, Crawford J, Hoet AE, Wilkins JR III, Lee J.
2013. Comparative performance of contact plates, electro-
static wipes, swabs and a novel sampling device for the
detection of Staphylococcus aureus on environmental surfa-
ces. J Appl Microbiol 115:171–178.
65. Thorne PS, Metwali N, Avol E, McConnell RS. 2005.
Surface sampling for endotoxin assessment using electro-
static wiping cloths. Ann Occup Hyg 49: 401–406.
66. Hall LB, Hartnett MJ. 1964. Measurement of the bacterial
contamination on surfaces in hospitals. Pub Health Rpts 79:
1021–1024.
67. Favero MS, McDade JJ, Robertsen JA, Hoffman RK,
Edwards RW. 1968. Microbiological sampling of surfaces.
J Appl Bacteriol 31:336–343.
68. Niskanen A, Pohja MS. 1977. Comparative studies on the
sampling and investigation of microbial contamination of
surfaces by the contact plate and swab methods. J Appl Bac-
teriol 42:53–63.
69. Dull PM, Wilson KE, Kournikakis B, Whitne EAS, Bou-
let CA, Ho JYW, Ogston J, Spence MR, Mckenzie MM,
Phelan MA, Popovic T, Ashord D. 2002. Bacillus anthracis
aerosolization associated with a contaminated mail sorting
machine. Emerg Infect Dis 10:1044–1047.
70. Gomez G, Arino A, Carraminana JJ, Rota C, Yanguela J.
2012. Comparison of sampling procedures for recovery of
Listeria monocytogenes from stainless steel food contact
surfaces. J Food Prot 75:1077–1082.
71. Clabots CR, Bettin KM, Peterson LR, Gerding DN.
1991. Evaluation of cycloserine-cefoxitin-fructose agar
and cycloserine-cefoxitin-fructose broth for recovery of
Clostridium difficile from environmental sites. J Clin Micro-
biol 29:2633–2635.
72. Lemmen SW, Hafner H, Zolldann D, Amedick G,
Lutticken R. 2001. Comparison of two sampling
methods for the detection of gram-positive and gram-
negative bacteria in the environment: moistened swabs
versus Rodac plates. Int J Hyg Environ Health 203:
245–248.
73. Hacek DM, Trick WE, Collins SM, Noskin GA,
Peterson LR. 2000. Comparison of the Rodac imprint
method to selective enrichment broth for recovery of
vancomycin-resistant enterococci and drug-resistant Enter-
obacteriaceae from environmental surfaces. J Clin Microbiol
38:4646–4648.
74. Buggy BP, Wilson KH, Fekety R. 1983. Comparison of
methods for recovery of Clostridium difficile from an envi-
ronmental surface. J Clin Microbiol 18:348–352.
75. Holah JT. 1992. Cleaning and disinfection, p. 319–341. In
Dennis C, Stringer MF (eds), Chilled Foods: A Comprehen-
sive Guide. Ellis Horwood, Chichester, UK.
76. Schiemann DA, Chatfield DE. 1976. Evaluation of neu-
tralizers in RODAC media for microbial recovery from dis-
infected floors. J Environ Health 38:401–404.
77. Salo S, Line A, Alanko T, Sjoberg A, Wirtanen G. 2000.
Validation of the microbiological methods Hygicult dip-
slide, contact plate, and swabbing in surface hygiene con-
trol: a Nordic collaborative study. J AOAC Internat 83:
1357–1365.
78. McGoldrick KF, Fox TL, McAllister JS. 1986. Evaluation
of a dry medium for detecting contamination on surfaces.
Food Technol 40:77–80.
79. Mikolay A, Huggett S, Tikana L, Grass G, Braun J, Nies
DH. 2010. Survival of bacteria on metallic copper
surfaces in a hospital trial. Appl Microbiol Biotechnol 87:
1875–1879.
80. Horner WE, Barnes C, Codina R, Levetin E. 2007.
Guide for interpreting reports from inspections/investiga-
tions of indoor mold. J Allergy Clini Immunol 121:592–597.
81. Portnoy JM, Barnes CS, Kennedy K. 2004. Sampling for
indoor fungi. J Allergy Clin Immunol 113:189–198.
82. Prezant B, Weekes DM, Miller DJ. 2008. Recognition,
Evaluation, and Control of Indoor Mold. American Industrial
Hygiene Association, Fairfax VA.
83. Yamaguchi N, Ishidoshiro A, Yoshida Y, Saika T, Senda
S, Nasu M. 2003. Development of an adhesive sheet for
direct counting of bacteria on solid surfaces. J Microbiol
Meth 53:405–410.

84. Deva AK, Vickery K, Zou J, West RH, Selby W, Benn
RAV, Harris JP, Cossart YE. 1998. Detection of persistent
vegetative bacteria and amplified viral nucleic acid from
in-use testing of gastrointestinal endoscopes. J Hosp Infect
39:149–157.
85. Bradley BJ, Saddler FS, Pyle BH, Broadaway SC,
McFeters GA. 2002. High Efficiency Pathogen Collection
and Rapid Detecton of Surface Microbes. SAE Technical Paper
2002-01-2450. Society for Automotive Engineers, Warren-
dale, PA.
86. Macher JM. 2001. Review of methods to collect settled
dust and isolate culturable microorganisms. Indoor Air 11:
99–110.
87. Calfee MW, Rose LJ, Morse S, Mattorano D, Clayton M,
Touati A, Griffin-Gatchalian N, Slone C, McSweeney N.
2013. Comparative evaluation of vacuum-based surface
sampling methods for collection of Bacillus spores.
J Microbiol Meth 95:389–396.
88. Calfee MW, Rose LJ, Tufts J, Morse S, Mattorano D,
Clayton M, Touati A, Griffin-Gatchalian N, Slone C,
McSweeney N. 2014. Evaluation of sampling methods
for Bacillus spore-contaminated HVAC filters. J Microbiol
Meth 96:1–5.
89. Prezant B. 2005. Comparison of PVC cassettes versus
vacuum bags for collection and enumeration of culturable
fungi in settled dust, p. 218–225. In Johanning E (ed),
Bioaerosols, Fungi, Bacteria, Mycotoxins and Human
Health: Pathophysiology, Clinical Effects, Exposue Assess-
ment, Prevention and Control Indoor Environments
and Work. Fungal Research Group Foundation, Albany,
NY.
90. Centers for Disease Control and Prevention. 2012.
Guidance on Packaging and Shipping Vacuum Socks
Used for the Collection of Bacillus anthracis Samples.
http://www.cdc.gov/niosh/topics/anthrax/pdfs/socksguid
ance.pdf.
91. Aizenberg V, Grinshpun SA, Willeke K, Smith J, Baron
PA. 2000. Performance characteristics of the button per-
sonal inhalable aerosol sampler. Am Indust Hyg Assoc 61:
398–404.
92. ASTM. 2005. Standard Practices for Collection of Surface
Dust by Microvacuuming Sampling for Subsequent Metals
Determination, D71440. ASTM International, West Con-
shohocken, PA.
93. Ashley K, Applegate GT, Wise TJ, Fernback JE, Gold-
camp MJ. 2007. Evaluation of a standardized micro-
vacuum sampling method for collection of surface dust.
J Occ Environ Hyg 3:215–223.
94. Burton NC, Grinshpun SA, Reponen T. 2007. Physical
collection efficiency of filter materials for bacteria and
viruses. Ann Occup Hyg 51:143–151.
95. Burton NC, Adhikari A, Grinshpun SA, Hornung R,
Reponen T. 2005. The effect of filter material on bioaero-
sol collection of Bacillus subtilis spores used as a Bacillus
anthracis simulant. J Environ Monit 7:475–480.
96. Wang E, Rhoades GG, Wainman T, Lioy PJ. 1995.
Effects of environmental and carpet variables on vacuum
sampler collection efficency. Appl Occ Environ Hyg 10:
111–119.
97. Nyachuba DG, Donnelly CW. 2007. Comparison
of 3M™ Petrifilm™ Environmental Listeria Plates
against standard enrichment methods for the detection
of Listeria monocytogenes of epidemiological signifi-
cance from environmental surfaces. J Food Sci 72:
M346–M354.
98. Pinard MF, Widden P, Debellis T, Deblois C, Mainville
C. 2005. House dust: an efficient and affordable tool to
assess microbial contamination in homes, p. 191–198. In
Johanning E (ed), Bioaerosols, Fungi, Bacteria, Mycotoxins
2.6.2. Surface Sampling ▪ 2.6.2-13
and Human Health: Patho-physiology, Clinical Effects, Expo-
sure Assessment, Prevention and Control in Indoor Environ-
ments and Work. Fungal Research Group Foundation,
Albany, NY.
99. Hung LL, Miller JD, Dillon HK. 2005. Field Guide for the
Determination of Biological Contaminants in Environmental
Samples, 2nd ed. American Industrial Hygiene Association,
Fairfax, VA.
100. Horner WE, Worthan AG, Morey PR. 2004. Air-
and dustborne mycoflora in houses free of water damage
and fungal growth. Appl Environ Microbiol 70:6394–
6400.
101. ASTM. 2010. Standard Practice for Bulk Sample Collection
and Swab Sample Collection of Visible Powders Suspected
of Being Biological Agents from Nonporous Surfaces.
E2458-10. ASTM International, West Conshohocken,
PA.
102. Wilson ML. 1996. General principals of specimen collec-
tion and transport. Clin Infect Dis 22:766–777.
103. Gilbert SE, Rose LJ, Howard M, Bradley MD, Shah S,
Silvstri E, SchaeferFW III, Noble-Wang J. 2014. Evalua-
tion of swabs and transport media for the recovery of Yersi-
nia pestis. J Microbiol Meth 96:35–41.
104. Perry KA, O’Connell HA, Rose LJ, Noble-Wang JA,
Arduino MJ. 2013. Storage effects on sample integrity of
environmental surface sampling specimens with Bacillus
anthracis spores. Biosafety S1:002.
105. International Air Transport Association (IATA) Dan-
gerous Goods Regulations. http://www.iata.org/publica
tions/dgr/Pages/index.aspx.
106. U.S. Government Printing Office. Code of Federal Regula-
tion, 49 CFR Parts 171–180. Government Printing Office,
Washington, DC. http://www.ecfr.gov.
107. U.S. Postal Service. 2014. Publication 52—Hazardous,
Restricted, and Perishable Mail. http://pe.usps.com/text/
pub52/welcome.htm.
108. U.S. Postal Service. 2014. Mailing Standards of the United
States Postal Service—International Mail Manual. PSN
7610-02-000-9904. http://pe.usps.gov/text/imm/immc1_014.
htm.
109. Federal Select Agent Program. 2014. Federal select agent
guidance on the shipment and receipt of packages with
select agents and toxins. http://www.selectagents.gov/
guidance-shipreceipt.html.
110. Mitscherlich E, Marth EH. 1984. Microbial Survival in the
Environment: Bacteria and Rickettsiae Important in Human
and Animal Health. Springer, New York.
111. Smith SM, Eng RHK, Padberg FT. 1996. Survival of
nosocomial pathogenic bacteria at ambient temperature.
J Med 27:293–302.
112. Sinclair R, Boone SA, Greenberg D, Keim P, Gerba CP.
2008. Persistence of category A select agents in the envi-
ronment. Appl Environ Microbiol 4:555–563.
113. Calfee MW, Wendling M. 2012. The effects of environ-
mental conditions on persistence and inactivation of
Brucella suis on building material surfaces. Lett Appl Micro-
biol 54:504–510.
114. Rutjes SA, van den Berg HH, Lodder WJ, de Roda
Husman AM. 2006. Real-time detection of Noroviruses
in surface water by use of a broadly reactive nucleic acid
sequence-based amplification assay. Appl Environ Microbiol
72:5349–5358.
115. Bej AK, Majububani MH. 1994. Application of the poly-
merase chain reaction in environmental microbiology. In
Griffin AGH (ed), PCR Technology: Current Innovations.
Academic Press, San Diego, CA.
116. Wilson IG. 1997. Inhibition and facilitation of
nucleic acid amplification. Appl Environ Microbiol 63:
3741–3751.
2.6.2-14 ▪ SAMPLING METHODS
117. Downey AS, Da Silva SM, Olson ND, Filliben JJ, Mor-
row JB. 2012. Impact of processing method on recovery
of bacteria from wipes used in biological surface sampling.
Appl Environ Microbiol 78:5872–5881.
118. Probst A, Facius R, Wirth R, Wolf M, Moissl-Eichinger
C. 2011. Recovery of Bacillus spore contaminants from
rough surfaces: a challenge to space mission cleanliness
control. Appl Environ Microbiol 77:1628–1637.
119. Probst A, Facius R, Wirth R, Moissl-Eichinger C. 2010.
Validation of a nylon-flocked-swab protocol for efficient
recovery of bacterial spores from smooth and rough surfaces.
Appl Environ Microbiol 76:5148–5158.
120. Moore G, Griffith C. 2007. Problems associated with tradi-
tional hygiene swabbing: the need for in-house standardiza-
tion. J Appl Microbiol 103:1090–1103.
121. Hong-Geller E, Valdez YE, Shou Y, Yoshida TM, Mar-
rone BL, Dunbar JM. 2010. Evaluation of Bacillus anthracis
and Yersinia pestis sample collection from nonporous surfa-
ces by quantitative real-time PCR. Lett Appl Microbiol 50:
431–437.
122. Winther B, McCue K, Ashe K, Rubino J, Hendley
JO. 2011. Rhinovirus contaminiation of surfaces in homes
of adults with natural colds: transfer of virus to fingertips
during normal daily activities. J Med Virol 83:906–909.
 Soil Sampling for Microbial Analyses
JOHN BROOKS
2.6.3
Beneath our feet lies a microbial “mega-metropolis,” with the
number of microorganisms in 1 g of soil outnumbering
the current human population. These bacteria, viruses, and
small eukaryotes control the biological, biogeochemical,
and nutrient fate associated with all organisms on the soil sur-
face. Soil is a complex mosaic of weathered inorganic rock,
organic leaf litter, and biological life; understanding the fate
of the soil microbiological constituents is only a single frame
of the story, but for our intents and purposes, it is the most
important. To work with soils from a microbiological per-
spective, the basic underlying features of soil need to be under-
stood, particularly the soil class, organic matter, and moisture
levels, to name a few. The interaction(s) between each of
these factors and the microbiological population defines the
basis of environmental microbiology. Like any metropolis,
heterogeneity is the way of the soil micro world; between
bacterial numbers reaching 1012 g−1, unicellular eukaryotes,
fungi, small animals, and finally viruses, which outnumber
all others combined, soil is one of the most qualitatively and
quantitatively complex ecosystems. Capturing the microbial
diversity of soil has only now been realized; the advent of
molecular techniques has guaranteed a wide-open culture-
independent exploration of microbes in and around us. Tradi-
tionally, microbial diversity was assessed using plate culture,
carbohydrate consumption, or other physiological measure-
ments. It is well known that only a fraction of the microbial
diversity has been captured using these limited methods (1).
All soil sampling, or any sampling for that matter, needs
to take into account the research objectives and what
they aim to achieve and the obstacles in the way to achieve
these objectives. Recent research funding calls have stressed
the importance of multidisciplinary teams, which address
complex problems from multiple angles. Understanding
soil microbial functional diversity or pathogen survival,
for example, can only be accomplished via (a) careful soil
sampling, (b) understanding the importance of soil field
variations, and (c) influence of wildlife, groundwater, and
treatments, to name a few. Certain questions need to be
answered prior to embarking on any field study, including
(a) what type of field; (b) what are the field nuances and char-
acteristics; (c) what type of experimental design addresses
a and b; (d) what type, where in the field, and how many
samples should be collected; and (e) what type of assay(s)
will be performed, and what will be done with the data?
doi:10.1128/9781555818821.ch2.6.3
2.6.3-1
This chapter aims to introduce the reader to soil microbial
investigations, specifically soil sampling scheme, tools and
equipment, processing for culture and molecular analyses,
unique sampling situations, and finally a case study presenting
a typical soil sampling situation. A straightforward discussion
regarding potential problems, caveats, and lessons learned
will be used to facilitate this conversation.
SOIL SAMPLING SCHEME
Sampling scheme (Fig. 1) may be the single most important
aspect to a well-designed soil project. Deciding the frequency,
number of samples, type, depth, and locations are all critical
steps prior to embarking on a project; failure to properly
address one of these may invalidate all data or at least fail
to assess treatment effects. Spatial variability must be taken
into account as minute (in centimeters) physical or chemical
differences may contribute to large microbial variability,
potentially obscuring treatment effects (2, 3). Similarly, sam-
pling variability, for example, the variability introduced via
sampling method, personnel, or training, will also obscure
treatment effects (4). Proper sample strategy must take into
account the most simplistic objective of the study. If the sam-
ple strategy cannot adequately address this, then it needs to
be reevaluated. Does the field have any obvious spatial varia-
bility? How many treatments will be imposed? Will samples
be collected to a specific depth? Should a statistical design
(e.g., block) be implemented to account for these variations?
Often, sample strategy requires presampling. Presampling
ideally addresses the previously posed questions. Case in
point: a field site may be level and flat for two-thirds of the
field and suitable for experimental design implementation.
The other third may be sloped and prone to excessive water
retention. Obviously, this may affect physical sample collec-
tion and analysis, in addition to microbial activity, resulting
in random effects that may be misconstrued as experimen-
tal treatment effects. A simple random sample collection
(Fig. 1b) could be used to estimate the amount of uncertainty
or variation in a field (5). Simple random sampling removes
bias by allowing all samples to be collected from randomized
locations in the field, plot, or replicate; that is, every possible
sample location has an equal chance of inclusion in the sam-
ple population and thereby placing equal amounts of bias on
each sample, even if that bias in unknown (5). If obvious
2.6.3-2 ▪ SAMPLING METHODS
(a)
(b)
FIGURE 1 (a) Common experimental designs, including (1) complete randomized; (2) complete randomized block; and (3) split-plot
randomized block designs. (b) Typical sampling regimes which can be applied in each of the experimental designs in part (a).
doi:10.1128/9781555818821.ch2.6.3.f1
physical differences exist (as in the foregoing example), a sys-
tematic or stratified random sample set could be collected
which would include the confounding factor (5). Since
these approaches would be considered for a preexperimental
investigation, it would not be expected that statistical guid-
ance would be necessary, though it would be helpful. In addi-
tion, literature and data freely available online is a good place
to begin data mining to help avoid some of these potential
issues. The U.S. Department of Agriculture Natural Resour-
ces Conservation Service (http://websoilsurvey.nrcs.usda.
gov/app/HomePage.htm) and National Agricultural Statis-
tics Service (http://www.nass.usda.gov/Statistics_by_State/)
provide detailed soil maps and physicochemical characteris-
tics, and land use statistics, respectively, for nearly the entire
continental United States. These data are invaluable to the
planning stages of any soil sampling study.
Once presampling is finished, it’s time to decide on an
experimental design. This would be ideally designed around
the most direct approach that addresses study objectives,
while providing strong statistical power to properly ensure
experimental robustness and avoid bias. While this chapter
is not designed to discuss statistics, it is an integral aspect to
experimental design. For further information regarding statis-
tics, the reader is referred to Preparation of Soil Sampling Proto-
cols: Sampling Techniques (available at http://www.epa.gov/
oust/cat/mason.pdf ) and (5).
Typically experimental plots are laid out in a fully random
or complete random block experimental design, though other
designs are available given the constraints of a specific field or
site (6). Other designs consist of Latin square and incomplete
block (6). The use of two or more treatments and potential
site bias lends itself to complete random block design, while
more homogeneous fields may be better served by a fully ran-
dom design (6). If by chance the bias in question is known, an
experimental design can be prepared that takes the bias into
consideration. By using complete random block design (6),
the bias can be blocked by accounting for it in the experimen-
tal layout. For instance, if a portion of a field with an obvious
physical bias exists, such as a slope that is 3% greater than the
rest of the study site, the experimenter could then block for
this bias. In this case, assuming that one-fourth of the study
site is affected by the slope; the experiment could be designed
to use this section as a block, while the other three blocks
would be located throughout the remainder of the field.

Each block would then have one replicate from each treat-
ment for a total of four replicates, with treatments randomly
placed throughout the block, hence complete (one block
contains all treatments) randomized (treatments randomly
placed) block design.
In some instances, complete random design cannot be
implemented. In these situations, an incomplete balanced
block design would be used (7). For example, all treatments
may not fit into a block or the random variability of a field
(observed in presampling) precludes the use of large complete
blocks, therefore the incomplete block design should be
implemented (7). Nested designs can be used when treat-
ments are replicated in different years, locations, or using dif-
ferent brands of the same treatment, for instance; in these
cases, year, location, and treatment brand would be nested
during data analyses.
The number of samples needed to describe the replicate or
population needs to be considered. Additionally, subsample
collection in either a random or systematic approach, should
also be considered (Fig. 1a). In most situations, a simple ran-
dom approach can be used within a block replicate, with the
number of samples per replicate directly correlated to plot
size. Ideally, a large number of samples would be collected
from each replicate plot, thus allowing the calculation of
the “true” population mean; however, rarely does this occur
in practice, as analyzing each individual sample increases
analysis costs considerably. A good rule of thumb is appro-
ximately 10–15 randomly collected subsamples, to be com-
posited (e.g., combine and thoroughly mix subsamples to
create one representative sample) to characterize an inter-
mediate sized plot (≤1 ha). The numbers scale up or down
from that point. Composite sampling is commonly used to
reduce subsample numbers and intrareplicate variability (4).
Sample collection schemes for flat (e.g., forage) and row
(e.g., cotton or food crop) fields will differ substantially. For
instance, sample collection in row crop environments can
be complex as row crops have very distinct sampling regions
due to the soil treatment, rows, furrows, and/or the crop’s
influence on the rhizosphere biology (8, 9). Soil cores col-
lected near the base of the plant will yield a far different
microbial population than cores collected in the furrow. Fer-
tilizer is often applied near the base of the plant, essentially to
apply the fertilizer near the root zone, and thus creates a soil
gradient in the horizontal as well as the vertical planes. Addi-
tionally, new methods for manure application (e.g., poultry
litter, swine effluent) can yield distinct areas in the field
where organic matter is highly localized (4). For instance,
poultry litter and swine effluent can now be applied or
injected in distinct bands or “trenches” created by coulters
or discs. Likewise, in the case of Class B biosolids, a soil incor-
poration process is required immediately following land appli-
cation, which creates a localized biosolids zone below the soil
surface (10). Projects involving these practices need to be
treated differently, particularly based on study objectives. If,
for instance, a particular project’s goal was to ascertain the
survival of Salmonella in swine effluent applied to row crop
lands, then one would want to collect soil cores where the
liquid effluent was applied (e.g., in the furrow) and not at
the top of the row. However, the same project may be inter-
ested in obtaining soil nutrient levels, thus cores collected
near the applied manure may not yield the most accurate
soil nutrient levels. Figure 2 shows a soil collection design
that may address these issues (4). In this case, sample cores
collected at the top of the triangle are lined up atop the man-
ure band or trench with the sides of the triangles flanking at
45° angles. The cores collected atop the band would be far
2.6.3. Soil Sampling for Microbial Analyses ▪ 2.6.3-3
more concentrated than those collected at the farthest point
from the band. Thus, if the study goal is to ascertain pathogen
survival, the top of the band would be of interest, while a
composite sample (of all cores) would yield the most realistic
nutrient level for the row. The key to compositing would be to
take equal amounts of soil from each core, so that each core is
equally represented in the composite.
Forage sites are typically more simple and homogeneous,
which allows sample collection to proceed in a more straight-
forward fashion (e.g., no rows, furrows). That being said, no
field is uniform when considering topology, hydrology, and
vegetation. Study treatments will also affect sample collec-
tion, specifically, when the application involves broadcast
treatments (e.g., solid manure or pesticide). For instance,
the application of solid fertilizers (e.g., manure or biosolids)
may appear uniform when considered on a relatively macro-
scale (e.g., 100 tons ha−1). The researcher can consider the
manure to be dispersed evenly on a hectare scale, but on a
relatively meso- and microscale, the application may not be
homogeneous. A clod of manure may be present in one
core and not present in a core 15 cm away. In this case, strict
adherence to sample collection design (e.g., random sampling
in a block design) is necessary. For instance, if the purpose of
the study was to measure the impact of manure application on
Salmonella soil levels, then a random sample approach will
most likely be carried out. That is, every sample core must
have an equal chance of being collected and represented,
regardless of whether a large solid piece of manure is present.
If, for instance, the purpose of the study was to assess Salmo-
nella die-off, then soil cores with manure clods would be tar-
geted. In the former scenario, a simple randomized sample
collection would be followed, while the latter could be
considered a “judgment” sampling approach (i.e., an inherent
bias is known).
Sampling time(s) will vary based on needs. In some cases,
samples may need to be collected immediately with follow-up
samples collected every day for a week, then slowly spaced to a
longer time period (e.g., sample collection at 1, 3, 7, 14, 21,
30, and 60 days) (6). This may be necessary when measuring
pathogen survival in soil, in which case data points need to be
closely spaced at the beginning of the project, with larger
intervals occurring after one month (9, 11). Long-term
microbial ecology, on the other hand, may require sample
intervals at 1, 6, 12, 24, and 36 months and longer following
initial experimental treatment (10). Since immediate micro-
bial inactivation or die-off data are not needed, ecological
studies are interested in overall long-term effect of the treat-
ment, so in this case, immediate impacts may not be that
necessary. Moreover, repeated sample collections in close
proximity in time and space may require special attention as
plot integrity may be influenced by a large number of soil
cores. A repeated collection ANOVA analysis or similar
may be needed to account for variable-time interactions (12).
Soil sampling depth is dependent on whether the purpose
is to collect samples from the surface, rhizosphere, or vadose.
In general, many microbial soil sampling approaches
comprise collection at the soil surface or rhizosphere. Soil
surface collection typically comprises from 0–30 cm below
the surface, given the type of soil and available equipment.
Rhizosphere samples are collected near the rhizome or root
surface and require a more specialized approach to collection
(13, 14). In either case, soil sample depth is imperative since
knowledge of a given contaminant travel distance or incor-
poration of a soil amendment will determine the sample
depth, when considering the two-dimensional sampling
plane. For instance, the depth below the surface to which a
2.6.3-4 ▪ SAMPLING METHODS

FIGURE 2 (a) Banded poultry litter (dark spot in soil profile and core) as an example of localized, concentrated nutrients and microbial
activity. In this example the treatment (e.g., litter) is applied in a band which concentrates all litterborne bacteria and nutrients into a small
section of the applied row. (b) A sample collection template used to identify the litter band and locations flanking the band; thus ensuring a
reproducible and representative sample. Sample cores would be collected from atop the band (top through the middle of the triangle) and flank-
ing the band; in this case one band soil core would represent 1/13 of the sample assuming all flanking cores were collected (six on each side).
Alternatively, if the study is associated with pathogen die-off in the band, soil cores could be collected exclusively from the middle of the tri-
angle. Photos courtesy of Dr. Haile Tewolde. doi:10.1128/9781555818821.ch2.6.3.f2

soil treatment is incorporated may only be 10 cm; therefore,
the researcher may be interested in microbial activity in the
shallow portion of the profile (e.g., top 10 cm). Vadose and
deep-vadose sampling (15, 16) are far more difficult, involv-
ing drilling and specialized tractor-mounted probes. Further-
more, minimal microbial activity and fewer microorganisms
warrant the use of sterile sampling approaches and specialized
culture and molecular methods. With the advent of high-
throughput DNA sequencing, this largely unexplored subsur-
face is likely to receive more attention.
SOIL SAMPLING TOOLS AND EQUIPMENT
There are many tools available to collect soil surface and deep
core samples; the basic premise is aseptically (or as much as
fieldwork allows) collect a sample with minimal disturbance
to the surrounding soil and microbial population. The type
of soil core sampler can vary from simple soil probes to slurry
samplers and soil core drills. At its most simple form, surface
soil can be collected using a handheld trowel. In this case, the
user is only interested in sampling the top few centimeters of
the surface, rather than a distinct soil depth (e.g., 0–15 cm).
Sample trowels may be preferable for initial presampling
events, in which case a defined soil core with vertical stratifi-
cation is not necessary.
Soil probes or augers are typically used to pull soil sample
cores (Fig. 3). Often, the soil probe is the tool of choice;
probes are typically made from stainless steel, which promotes
durability, a sharp edge, and efficient field disinfection using
70% ethanol. Probes are easily used in most soil types, given
that the soil is soft enough to be collected. Probes can be
conveniently scored and marked with vertical lines from 0,
5, 10, 15 . . . 30 cm, which is useful when it’s necessary to par-
tition soils into distinct cores. For instance, a 0–30 cm core
could be sectioned into 0–5, 5–15, and 15–30 cm sections.
Unfortunately, deep core samples cannot be collected with
a simple probe; for that, a more specialized piece of equipment
is necessary.
Deep soil core samples can be collected using soil drills
and mounted probes, which are capable of collecting
soil cores from a few meters (Fig. 3) to hundreds of meters
below the surface. Deep core samples are costly and difficult
to acquire sterilely. Drill-rigs are large pieces of equipment,
requiring tractor mounts, specialty drill bits, cooling fluids,
and scraping teeth. Many drill-rigs are designed to collect
soil cores by using a sampling barrel near the end of a drill
 2.6.3. Soil Sampling for Microbial Analyses ▪ 2.6.3-5

FIGURE 3 Soil cores collected with a (a) core sampler mounted to a tractor capable of reaching approximately 1 m, and a (b) standard soil
probe capable of reaching approximately 30–45 cm. Photos courtesy of Drs. Karamat Sistani and Ardeshir Adeli.
doi:10.1128/9781555818821.ch2.6.3.f3

bit, which is then pushed through the soil. Pushing a drill
bit hundreds of meters through the soil surface generates
substantial heat; therefore, the use of cooling air or fluids
is used to reduce heat and remove drill cuttings. Either
way, contamination of the outer surface can be monitored
through the use of tracer dyes, which dictate the level of
contaminated soil through the core profile. Any contami-
nated soil would then be removed using scraping teeth. For
more information regarding deep core sampling, the reader
is encouraged to read (17).
Generally speaking, all soil samples are collected and then
immediately stored in a cooler packed with ice packs or wet
2.6.3-6 ▪ SAMPLING METHODS
ice (carefully avoiding potential cross-contamination) prior
to transport back to the laboratory (18). Samples are typically
placed in thin, sterile, plastic resealable bags, directly from soil
probe to bag, with minimal manipulation. In addition,
plastic-capped tubes can be used, but may be more difficult
to manipulate in the field. When using a soil probe, the use
of a chisel or soil knife (which are easily disinfected) are
used to remove or slice the soil core (e.g., horizontal slices
at 5, 10, and 15 cm) prior to placement in the bag (19). Sam-
ples should be placed in a cooler with ice or ice packs, partic-
ularly if the samples were collected from a cool environment.
Samples collected for pathogen analysis need to be collected
and stored at cooler temperatures (e.g., 4°C) to facilitate
transport to the laboratory with minimal loss in viability.
Likewise, when microbial physiology must be measured, it is
critical to maintain soil samples close to ambient conditions,
particularly if that measurement cannot be made in the field.
In other instances, the soil vertical profile of the core needs to
be maintained; in this case, transport must facilitate move-
ment of either individual pieces of the soil profile or the entire
section.
SOIL SAMPLE PROCESSING: BACTERIA,
VIRUSES, AND PARASITES
A number of microbial parameters can be measured in soil
samples, assuming proper processing is adhered to, including
pathogens, microbial ecology and activity, nucleic acid, and
protein expression.
Sample storage is just as important as sample collection in
many instances; even a few days of improper storage can limit
the story a sample can tell. With busy projects and busier labs,
sample storage is sometimes unavoidable. Conventional wis-
dom dictates that samples in which culture or physiological
methodologies will be employed need immediate processing,
while samples that involve DNA can be frozen at −20°C or
lower (20). Samples can be stored in plastic bags, but it should
be noted that samples can become anaerobic or excessively
dry through prolonged storage (21). In addition, if a soil sam-
ple is to be analyzed for labile chemical constituents, such as
nitrogen, then care should be taken to avoid volatile losses.
Storage at 4°C is a temporary measure and can be maintained
for 2–3 weeks; this time frame is “adequate” for microbial
analysis, but longer time frames should be avoided or held
at −20°C and below (22). It should go without mention,
but subfreezing storage reduces some cultivated vegetative-
bacteria, while some parasitic cysts, fungal spores, and viruses,
may persist in a frozen state (23). Storage during short time
frames and low temperatures reduce changes to the microbial
population; however, certain microbial measurements (e.g.,
oxygen-labile organisms, pathogens, physiology) will be
irreparably altered during short- and long-term storage (24).
Samples can be homogenized by passing slightly dried soil
through a 2 or 4 mm mesh screen to remove stones, roots, or
other leaf litter commonly found in surface and subsurface
soils. A caveat associated with this type of pretreatment is
that soils need to be dried below most typical field moisture
contents (<30% moisture content). Drying may harm any
culture-based pathogen or any other sensitive population
assay (25). DNA-based assays may not be negatively affected
by drying (16).
All assays (at least culture) require sample elution in a buf-
fer, typically consisting of sterile water, PBS ( phosphate-
buffered saline), saline, peptone water, or some other type
of aqueous growth medium or diluent. The basic idea is to
(a) suspend the easily dislodged soil microbial population
into a liquid matrix to facilitate assay; (b) remove the soil
surface-bound microbial fraction and suspend into the liquid
fraction; and (c) break open soil peds and release the micro-
bial population. Once again, as with other sections in this
chapter, the choice is dependent on the end use of the sam-
ples and the project. For instance, sterile water, PBS, or saline
would be used as simple diluents, while a growth medium
(e.g., tryptic soy broth, peptone water) could be used to
enrich a bacterial population or subpopulation. Soil is nor-
mally added at a rate of 10 g per 95 ml diluent (19). This
essentially provides an initial 10-fold dilution based on the
moist soil (10 g) in the suspension. This initial dilution can
then be homogenized, either via stomacher, horizontal
shaker, or hand shaker. Stomaching systems are advantageous
in that they provide a single-use sterile bag system in which
soil and diluent are mixed. The system is then sealed and
placed into a stomaching system that gently “massages” the
soil/diluent solution to dislodge most bacteria or eukaryotic
cells from soil particles. This enables the solution to be serially
diluted for assays, be it via culture or molecular methods. A
horizontal shaker or hand-wrist shaker essentially accom-
plishes the same effect. Once again, this is user choice, and
no one system is better than the other, though horizontal
and hand-wrist shakers are capable of housing more samples,
sometimes with less dislodgement of soil-bound cells. Stom-
aching systems are often limited to a single sample, though
some stomachers can handle more than one sample simulta-
neously. Additional approaches may include bead-beating in
concert with horizontal shaking; in this case, the sample will
be loaded with sterile glass beads as well as diluent, and then
shaken horizontally, with the beads removing any soil-bound
microorganisms. Sonication uses energy waves to dislodge
cells from soil; this technique has been applied effectively,
but care must be taken not to apply too much pressure as cells
can readily degrade (26).
Study of the soil rhizosphere is more involved (13, 14).
Essentially, the rhizosphere, containing intact roots, root
hairs, and adhering soil, is first removed from the surrounding
soil. The adhering soil is then lightly removed by dispersing it
in sterile washing diluent. This washing diluent is then used
in further assays.
Following initial sample elution, most soil samples will
require serial dilution prior to assay. Dilution is typically
done with the same diluent as the original elution solution.
Sterility is vital to ensure that the multiple handling steps
don’t introduce contaminating microorganisms or nucleic
acid to the solution. Serial, 10-fold dilutions are carried out
sequentially, until the dilution extinction is reached.
Calculation of CFU (colony-forming units), MPN (most
probable number), or GU (genomic units) g−1 soil is typically
done on a dry gram basis. This is a standard way of presenting
soil data for most disciplines; the purpose is to present data in a
manner which is not subject to change (e.g., soil moisture can
always change, but dry mass does not). To properly calculate
on a dry gram basis, one must calculate the gravimetric mois-
ture content of the soil using equation 1, where m = moist soil
mass and d = dry soil mass. First, 10 g of field moist soil is
measured (m), which is then dried at 104°C for 24 h to obtain
a final dry mass (d ).
Gravimetric Soil Moisture Content
 
md
%umc ¼  100 (1)
d
Presenting soil microbial values based on soil dry gram is
advantageous, because it enables the use of a standardized

value (since the dry gram is already dry and will not change
further) which can be compared with other values, regardless
of the amount of moisture in a sample. Soil moisture is typi-
cally presented as percent soil moisture. In practice, gravimet-
ric soil moisture is useful for most situations; however, in some
scenarios soil is too moist (e.g., more than 50% water), which
prevents the calculation of soil moisture using equation 1. In
this case, it may be advantageous to dry the soil slightly before
analyses, if permissible, or calculate soil dry matter, as is com-
monly applied for calculations of waste-matter solid content.
The main difference here is that soil dry content is a function
of wet mass (e.g., dry mass divided by wet mass), whereas
gravimetric soil moisture is a function of dry mass.
Soil Microbial Assays
All assays will essentially reflect either the viable cultivated
microbial fraction or the noncultivated microbial fraction.
Cultivated (e.g., culturable) assays encompass simple dilution
plating approaches or physiological measurements. nonculti-
vated assays include molecular approaches using either PCR,
quantitative PCR (qPCR), or library-based ecological assays.
Culturable approaches are the most traditional, although
one can argue that molecular approaches have certainly
supplanted them, at least in the field of microbial ecology.
Dilution spread-plating approaches are the most basic. The
selection of the initial assay type and media are paramount
for a successful project; the choice between a spread plate or
enrichment approach is directed by study objectives. The for-
mer providing for a cost-effective quantitative solution with
more high-throughput capabilities, while the latter provides
for detection of low numbers in either a presence/absence
format or quantitative most probable number approach.
The choice may come down to labor, time, cost, and the
anticipated microbial level in the soil. Colony counts, via
spread plate approaches, are simple and cost-effective, even
in large numbers. The fact that the entire process, from sam-
ple to CFU, can take as little as 24 h enables a fast turn-around
compared to many other culture-based techniques. Spread-
plating is conducted with a number of media types; bacterial
targets range from generic heterotrophic bacteria to specific
bacterial groups or pathogens. Since the focus is on soil and
soil microbial communities, the choice of agar must facilitate
slower-growing bacteria not necessarily adapted to a nutrient-
rich agar. Most soil microorganisms are living in a limited
nutrient environment; thus, applying approaches used for
clinical isolates (e.g., organisms adapted to oligotrophic envi-
ronments) may not be suitable. Low nutrient and generic het-
erotrophic culture media are typically used when measuring
soil bacteria. Some typical media choices are tryptic soy
agar (TSA), plate count agar (PCA), Reasoners 2 agar
(R2A), diluted nutrient agar, soil extract agar, and VL55
agar. It is important to note that nutrient-rich media such
as TSA may be too rich for many nutrient-starved bacteria;
therefore, it may be prudent to attempt a half or less strength
media (27). While one media type can be used, it is often
more advantageous to attempt to culture using a wide selec-
tion of culture media; for instance, the use of nutrient-rich
TSA and R2A could both be used to capture a wide selection
of bacteria. When attempting to quantify the “general” pop-
ulation, a series of 10-fold serial dilutions (e.g., low to high
or 10−1 to 10−6) will need to be plated, particularly from an
unknown soil type or quality. The initial soil suspension is
considered the first 10-fold dilution with subsequent dilutions
following in serial fashion. Soils collected in the summer,
particularly when moisture content is low (e.g., lack of precip-
itation), will most likely need fewer dilutions, particularly
2.6.3. Soil Sampling for Microbial Analyses ▪ 2.6.3-7
when bacteria are present at 104 CFU g−1. Conversely,
“healthy” soils with high organic matter and adequate mois-
ture may number as high as 108 CFU g−1; thus requiring
more dilutions. Sample location may determine this as well.
The deeper the sample, the less concentrated it will be;
likewise, the samples collected nearest to fertilizer, rhizo-
sphere, or high moisture or organic matter will tend to be
more concentrated.
In addition to generic media, assays can be made selective
for biogeochemical functional groups or bacteria which can
grow on specific organic contaminants (28). Minimal media
and other broth and agar formulations can be amended with
many types of organic contaminants, such as 2,4 chloroben-
zoate (28) and trichloroethylene (29) or heavy metals (30),
which can be used to isolate microbial populations capable
of tolerating and/or degrading such contaminants. Minimal-
media formulations are designed to selectively isolate bacte-
ria, which can grow on the sole carbon source included in
the media. Likewise, minimal media amended with heavy
metals in conjunction with an organic substrate would facil-
itate the selection and isolation of heavy metal–resistant
bacteria. In this case, the minimal media would contain a sin-
gle or mix of carbon sources, macronutrients, and the heavy
metal. These minimal media preparations can be poured as
agar plates or broth solution. When quantification is not
necessary, a 10 g soil aliquot can be inoculated into the broth
solution. In these cases, the soil would be incubated in an
effort to enrich the microbial population; efforts such as
this would be used to isolate unique resistant bacteria, identify
potential for degradation pathways, or study microbial com-
munity interactions (28).
Pathogen and indicator assays are wide ranging, but in this
chapter the focus will be on zoonotic/human fecal pathogen
and indicator detection. Samples collected for these situa-
tions must be collected with caution, as many, if not all, of
these microorganisms will be physiologically stressed. Con-
cern for pathogen survival must be taken into account and
often leads the researcher to immediately store samples in a
cooler and kept on ice until they are brought to the laboratory.
Bacterial fecal pathogens or indicators can be cultured
using serial-plating techniques. Hektoen enteric, MacCon-
key, sorbitol-MacConkey, mTEC, manitol salt, mEnterococ-
cus, mCP agars, and a host of others can be used. One point to
consider when using serial plating is that pathogens will be
one to six orders of magnitude less than background bacteria.
Spread plating at low dilutions (e.g., 10−1) may facilitate
pathogen detection at low levels, but it will also provide for
high background bacterial levels; in these cases, the selective
and differential nature of the media would most likely be
overcome. Addition of selective antibiotics can also aid in
reducing background bacteria, but can also bias samples.
Some pathogens (e.g., fecal pathogens) will be present at
such low levels that plating techniques are not feasible. To
properly detect these pathogens, one must enrich and select
for the pathogen while not harming the pathogen. This is a
highly stressful “tightrope” to walk, one which is far from
resolved. Pathogen levels will often be below 500 CFU g−1
soil and in a difficult-to-culture state; therefore, in these situa-
tions, it is often beneficial to default to an MPN enrichment
approach. The basic premise behind an MPN approach is to
begin enrichment of a sample followed by stepwise increases
in selective enrichments, until an endpoint is reached
whereby the MPN can be calculated. This approach enables
enrichment of difficult to culture bacteria, while also enrich-
ing bacterial numbers to a level that can be statistically enum-
erated. Most probable number approaches are time, labor, and
2.6.3-8 ▪ SAMPLING METHODS
consumable intensive approaches, but are used to great effect
when pathogen levels are relatively low (19). Many have
noted the culture biases introduced with MPN approaches
(31). In some cases the background soil microbial levels are
too inhibitory; false negatives are often reported, particularly
from assays targeting fastidious bacteria such as Campylobacter
spp. (32). The ability to assay relatively large amounts of soil
is a benefit of MPNs. In a standard nine-tube MPN assay, 1 g
can be accommodated in the first set of three tubes, followed
by serial 10-fold dilutions in the next two sets of tubes. A large
volume MPN can include soil aliquots up to 10 g in the first
three tubes of a nine-tube assay. Conversely, a standard
plating approach may only assay, at most, 0.1 g of soil per
plate, but more realistically 0.01 g soil per plate. However,
it must be stated that cultivation of bacteria on spread plates
provides for a tangible colony forming unit (CFU g−1),
whereas MPN provides a statistically derived MPN g−1 value.
A single colony can be isolated with further purification or
selective steps, but the value will still be considered an
MPN. The amount of time spent transferring samples from
tube to tube is a caveat associated with MPNs. Because
MPNs are based on the enrichment of the sample, initially
in a nonselective medium, the transfer of the sample from
one enrichment tube to the next facilitates bacterial selec-
tion. Because soil often harbors background bacterial levels
at 108 CFU g−1, the common MPN approach will most likely
involve selective enrichment. Most MPNs can take as long
as 4 days from sample collection to confirmed isolate; thus,
this type of assay can limit sample throughput. Keep in
mind, a simple 9-tube MPN with three transfer steps will
use 27 tubes, followed by isolation and isolate confirmation.
MPNs combined with qPCR (33) can reduce the number
of individual tubes; alternatively, some methods have been
modified to utilize small sample volumes, often working
with 96-well or 48-well microtiter plates (34). Alternatively,
a simple enrichment assay can be substituted, which reduces
consumable materials; however, only plus/minus data are
collected (19).
Viruses and parasites can be quantified and cultivated
from soil samples. Animal virus assays often involve MPN
approaches, with cell culture substituting for the steps in
the previous bacterial MPN assays (35). In this case, the
soil must be extracted for viruses using beef extract or similar
extraction approaches (36). The extract would then be
assayed via MPN. Likewise, coliphage can be quantified
and isolated using these similar approaches, although coli-
phage assays can be conducted using a semi-solid double
agar overlay approach (37). Parasites such as Cryptosporidium
spp. can be enumerated using MPN approaches as well (38).
SOIL SAMPLES FOR MOLECULAR ANALYSES
Soil samples intended for DNA quantification are treated dif-
ferently than soil meant for immediate culture. In the former,
the soil can be frozen at −20 to −80°C. In this case the imme-
diacy of the cultivation assay can be ignored, thus allowing for
more long-term storage of the samples. Nucleic acid–degrad-
ing enzymes are inhibited at temperatures below −20°C, and
thus samples can be stored for a number of months up to years.
Of course, it is prudent to immediately process samples, if
possible, to avoid any unwanted changes to the microbial
population due to storage. Samples should be stored in inert
plastic bags or containers and should be clearly marked
with all pertinent information to facilitate reference to the
samples should the researcher not handle the sample for
long time periods. Soil aliquots for nucleic acid extraction
are often small (<1 g), thus a small aliquot can be sto-
red for long-term storage. Studies where RNA is targeted
(e.g., transcriptomics) should be handled with extreme
care, as RNA is well known for its labile nature. In these
cases, RNA should be extracted immediately. There are
many RNA extraction procedures, as well as standardized
kits which can be applied quickly and efficiently to extract
RNA. Similar considerations should be given to studies
of protein expression (e.g., proteomics). An alternative to
immediate processing is immediate flash freezing using
liquid nitrogen (39). Storage considerations should be well
thought out prior to sample collection. A poorly carried out
storage plan could result in an imbalanced treatment of
sample sets, whereby some samples may be immediately proc-
essed and others not. This imbalance could result in an
unknown statistical bias.
A number of methods exist to extract and isolate microbial
DNA from soil samples; however, the “right” method signifi-
cantly impacts data interpretation (40). Typically, commer-
cial DNA extraction kits are used to isolate various forms of
microbial DNA. The extraction kit used is highly dependent
on the soil type, moisture, organic matter, anticipated humic
inhibitory substances, soil pH, and anticipated microbial
target levels. Microbial DNA kits have been designed to
work with soils containing high organic matter (e.g., fecal
matter), high moisture (e.g., slurry/sediment), and low micro-
bial content (e.g., extreme sites such as deserts). Though this
chapter will not delve into the specifics of each kit or type, the
basic idea is as follows: (a) dislodging of soil microbes from soil
particles, (b) extraction of cellular nucleic acid, (c) purifi-
cation of nucleic acid, and (d) nucleic acid collection. One
source of discrepancy between kits is in the extraction of
cellular nucleic acid. Some kits utilize a “bead-beating”
approach, which is a physical means of removing the cellular
wall to extract DNA, in addition to further dislodging bound
bacteria from the surface and interior of soil particles. Other
kits rely on chemical and heat lysis approaches. It is best to
obtain samples from a few competing kits to test representa-
tive samples. The DNA can then be compared and qualified
on a basis of A260/280, DNA quantity, and finally PCR
amplification potential. Often, kits perform more efficiently
with specific sample matrices (41).
Some kits can extract up to 10 g of soil, which allows the
researcher to work with a more representative soil aliquot,
rather than the typical 0.25 g. In all cases this increases assay
detection sensitivity. DNA extraction using traditional
approaches such as phenol/chloroform, cetyltrimethylammo-
nium bromide, freeze-thawing, and nitrogen are still used.
These techniques yield a highly purified microbial DNA
extract, with few inhibitors. It is important to note that these
traditional techniques often yield DNA, which is less biased
than kit-extracted DNA; the major caveats associated with
these approaches are the time, labor, low throughput, and
use of harsh chemicals. It may be advantageous to combine
some of these approaches to facilitate difficult to isolate
DNA or to improve efficiency (41). In addition, some
microbes do not lyse easily, particularly microbes from harsh
soil environments such as contaminant sites, deserts, and
extreme sites.
The use of molecular techniques is quickly becoming the
gold standard in environmental microbiology, particularly
in fields such as microbial ecology. These methods have
supplanted the cultivation-based approaches, which as most
are aware, are biased and may only represent the very rare
cultivated members of the microbiome (1). That being said,
every research project has its specific needs, and cultivation

is still very necessary, particularly when conducting whole
genome-based characterization, or when phenotypic infor-
mation is needed, such as in antibiotic-resistance based
characterizations.
As in cultivation-based approaches, there are two main
types of molecular assays: qualitative and quantitative. Qual-
itative assays are based on either presence/absence PCR or in
DNA sequence identification, such as in library-based analy-
ses. In most cases, the ability to amplify DNA extracted from
soil samples is the major limiting point. Successful amplifica-
tion is dependent on proper soil DNA extraction and removal
of key inhibitors such as heavy metals, divalent metal cations,
competing organic humic substances, and other unidentified
PCR inhibitors (42). The majority of DNA extraction kits
have been designed to remove most of these inhibitors,
though they are far from perfect. For instance, soils originating
from low-pH sites, such as surface-mined sites, or from sites
receiving regular biosolids land applications may need oppos-
ing DNA extraction methodologies. Pretreating samples prior
to DNA extraction may be necessary; in this case, pH adjust-
ments, washing soil particles, or passage through gel agarose
or capillary tubes are common approaches.
PCR has allowed the identification or quantification of
previously ignored microbial populations. In soil, PCR is
not without its own biases. For instance, PCR will selectively
amplify certain members of the microbial population due to
ease of cell lysis and/or removal from soil particles. In addi-
tion, high G-C (guanine-cytosine) DNA may require more
effort to achieve ample amplification; proper primer design
will reduce this bias, though.
Most molecular-based studies revolve around some form of
PCR; however, presence/absence PCR can only provide so
much data (i.e., the gene of interest is either there or not).
Quantitative PCR (qPCR) can be considered as a substitute
for quantification data, traditionally obtained via spread-
plating or MPN analyses. qPCR enables the researcher to
quantify genes of interest via slight modifications to tradi-
tional PCR; in many instances, this has become the new
standard for quantification analyses. qPCR, like all PCRs, is
subject to biases associated with DNA from nonviable organ-
isms. Pre-PCR modifications have enabled researchers to
amplify DNA from the viable fraction (43) but require another
manipulation of template DNA prior to amplification.
One way to view PCR-based assays, is that it gives the
researcher a genotypic or “genetic potential” view of the
microbial population, allowing the user to estimate what
the microbial population is capable of achieving, assum-
ing all genes are expressed phenotypically. Fecal patho-
gens in soil can be quantified using qPCR, but the presence
of DNA may not equate to the presence of a viable
pathogen. In this case, either a modified qPCR approach
(e.g., propidium monoazide) (43) or enrichment-based
MPN-qPCR (33) may be more suitable. Quantitative PCR
assays have been developed, including functional (e.g.,
phosphatase—enzyme activity), biogeochemical functio-
nal (e.g., nosZ—nitrous oxide reductase), fecal indicator
(e.g., uidA—E. coli), antibiotic resistance (e.g., tetM—
tetracycline resistance), and total bacteria or eukaryotes
(e.g., 16S and 18S rRNA, respectively).
As with qPCR, microbial ecological assays such as clone
libraries, metagenomics, transcriptomics, proteomics, and
microarrays enable the researcher to obtain a wide range of
microbial ecological information. The “omics” era has now
arrived, and these methods allow for large amounts of infor-
mation (>106 DNA sequences) to be collected at a low cost
per sequence (44). These studies allow for soil microbial
2.6.3. Soil Sampling for Microbial Analyses ▪ 2.6.3-9
ecologists to truly capture the diversity of a given soil. While
this chapter won’t delve into too many details, sampling soil
for “omics” analyses would still follow the same guidelines
outlined here. At this point, the limiting factor may be
an overwhelming amount of data generated per sample.
This area of microbiology is still burgeoning and expanding
rapidly.
SPECIAL SITUATIONS
Multidisciplinary projects create unique circumstances in
which soil samples are cocollected, -analyzed, and -correlated
with chemical, physical, microbial, and molecular informa-
tion. Certain situations may arise, which require rethinking
of typical sample collection. Figure 2 shows a cotton field
in which poultry litter was banded by applying litter approx-
imately 10 cm below the soil surface (4). In this situation,
simply collecting a random soil core would not suffice. Repre-
sentative soil cores require a sampling, which needs to take
into account the banded litter as well as the surrounding
soil, all equally represented (see Fig. 2). One approach would
be to collect soil cores, equally weighted, from atop the litter
as well as the flanking soil. The cores could then be homogen-
ized and a proper representative sample collected. One needs
to be aware of the depth at which the localized manure or
fertilizer was delivered; in most instances, the presence or
location of the band of fertilizer will be difficult to determine
from the surface. If the researcher is interested in collecting
only soil impacted by the fertilizer, then it would be more pru-
dent to collect a soil profile which passes through the band,
only collecting the profile which includes the band, as well
as soil above and below it.
Drainage ditches, and equally sloped soils yield surprising
variability. Samples collected at the base of the ditch contain
large amounts of organic matter which in turn leads to more
microbial diversity. In these cases, the researcher must con-
sider samples collected at the base of the drainage ditch as
well as on either sloped side. In all likelihood, samples would
yield varying degrees of microbial diversity. In addition, if
sample collection coincides with a rain event, the base of
the ditch will most likely be very moist. If samples are slurry
in consistency, the researcher may need to adjust sample col-
lection from a simple probe to one modified for slurry or
sludge collection (Fig. 4).
Often times, the sample site may have localized physio-
chemical influences, such as acid mine drainage or high metal
content, both as a result of anthropogenic activities (45).
These sites may have locations where soil pH has been highly
affected by these anthropogenic activities. In this case, a
gridded soil survey may be necessary to characterize the loca-
tion; alternatively, a simple stratified sample approach may be
enough, if lines of demarcation (e.g., low and high pH) are
well known. In addition, the presence of low pH or heavy
metal contamination, which is common at coal mining sites,
may inhibit DNA extraction and PCR. In cases where the
mine soils have been recently reclaimed, a low microbial
diversity could be expected. Culture analysis should not
have much of a problem with these low levels, as sample ali-
quot tends to be greater than molecular methods; however,
DNA analysis may require larger amounts of soil to properly
characterize the soil or increase sensitivity.
EXAMPLE
The following is a typical scenario one could expect
to encounter when conducting a soil sample collection;
2.6.3-10 ▪ SAMPLING METHODS
FIGURE 4 Soil core samples collected from a drainage ditch in
the Mississippi Delta. The soil was collected shortly following
large rain events; hence, collecting intact soil cores using a soil probe
is not feasible. Soil core liners, in this case, can be pushed through
the soil (inset) and capped at the either end. Alternatively sludge
samplers could be used if the soil is too moist to maintain profile
integrity. Photos courtesy of Dr. Beth Baker.
doi:10.1128/9781555818821.ch2.6.3.f4
additionally, the scenario is followed by one approach
to sampling said field, but it should be known that many
approaches could be used here. Assume a field divided
into 30 subplots (100 m2) receives three different types of
fertilizer, delivered via surface and subsurface land applica-
tion, with four replicates. Six subplots were used as controls,
with no fertilizer applied. The objective is to obtain soil cores
to deduce the effect of the various fertilizers on the micro-
bial quality and quantity using both culture and molecular-
based approaches. The surface-applied fertilizer was not
incorporated or disked in, while the subsurface application
was applied approximately 10 cm below the soil surface in
a trench 15 cm to the east of the crop row. Since this is a
time-based ecological and pathogen study, samples are to be
collected at time 0, one-fourth, one-half, 1, 6, 12, 18, and
36 months following land application. An appropriate
number of sample bags would be prelabeled the day before,
in addition to packing: (a) soil core probes and chisels, (b)
gloves, (c) cooler, (d) ice packs, (e) camera, (f ) measuring
tape, and (g) notebooks/pens/markers. One approach would
involve collecting soil cores using a 3 cm diameter soil probe
at each plot. Cores would be collected from two locations,
excluding the outer 1 m2 edge of the plot. The cores would
be collected from 0–15 cm below the soil surface, with divi-
sion of the core into 0–7.5 and 7.5–15 cm subcores; 10 indi-
vidual cores would be composited to comprise one sample
per plot. Since the idea is to compare surface to subsurface
application, the same number of cores will need to be col-
lected from the subsurface-applied plots. The exact location
of the subsurface trench is known; therefore, cores can be
pulled from directly above the trench as well as surrounding
the trench. The key here is to equally represent the localized
delivery of fertilizer as well as the surrounding soil. For
instance, if the row is 60 cm in width and the core is 3 cm,
the individual cores could be spaced every 9 cm to allow for
a total of five cores per row beginning at the crop base, which
would be repeated at another location in the plot. Samples
would be stored in an appropriately sized cooler with ice
packs and kept from direct sunlight.
Upon returning to the laboratory, a subsample would be
separated for molecular analyses (to be frozen) while culture
analyses would begin immediately. Following sample homog-
enization, a 10 g aliquot would be added to 95 ml sterile PBS
and stomached for 90 s. Proper serial dilutions would be made
and aliquots plated to half-strength R2A to quantify total
aerobic heterotrophic bacteria. Since pathogens are of inter-
est in this study, samples will also be plated to mTEC and
Hektoen enteric agar. Molecular analyses would consist of
extracting DNA from 1 g of soil per sample and amplifying
the 16S rRNA gene followed by sequence library creation
and phylogenetic analyses. Overall, this approach can be
modified and expanded in a number of ways, including the
individual sample core collection, sample depth, analyses,
and statistics.
REFERENCES
1. Shade A, Hogan CS, Klimowicz AK, Linske M, McManus
PS, Handelsman J. 2012. Culturing captures members of the
soil rare biosphere. Environ Microbiol 14:2247–2252.
2. Loescher H, Ayres E, Duffy P, Luo H, Brunke M. 2014.
Spatial variation in soil properties among North American
ecosystems and guidelines for sampling designs. PLoS One
9:1–31.
3. Goovaerts P. 1998. Geostatistical tools for characterizing the
spatial variability of microbiological and physic-chemical
soil properties. Biol Fertil Soils 27:315–334.
4. Tewolde H, Way TR, Pote DH, Adeli A, Brooks JP, Shan-
kle MW. 2013. Method of soil sampling following subsurface
banding of solid manures. Agron J 105:519–526.
5. Webster R. 2001. Statistics to support soil research and their
presentation. Euro J Soil Sci 52:331–340.
6. USEPA. 1992. Monitoring small-scale field test of microor-
ganisms U.S. EPA 700 R-92-008. Washington, DC.
7. Lopez MV, Arrue JL. 1995. Efficiency of an incomplete
block design based on geostatistics for tillage experiments.
Soil Sci Am J 59:1104–1111.
8. Stine SW, Song I, Choi CY, Gerba CP. 2005. Effect of rel-
ative humidity on preharvest survival of bacterial and viral
pathogens on the surface of cantaloupe, lettuce, and bell
peppers. J Food Prot 68:1352–1358.
9. Byrne-Bailey KG, Gaze WH, Kay P, Boxall ABA, Hawkey
PM, Wellington EMH. 2009. Prevalence of sulfonamide
resistance genes in bacterial isolates from manured agricul-
tural soils and pig slurry in the United Kingdom. Antimicrob
Agents Ch 53:696–702.
10. Zerzghi H, Brooks JP, Gerba CP, Pepper IL. 2010. Influ-
ence of long-term land application of Class B biosolids on
soil bacterial diversity. J Appl Microbiol 109:698–706.
11. Hutchison ML, Walters LD, Moore T, Thomas DJI, Avery
SM. 2005. Fate of pathogens present in livestock wastes
spread onto fescue plots. Appl Environ Micro 71:691–696.
12. Piepho HP, Buchse A, Richter C. 2004. A mixed modeling
approach for randomized experiments with repeated meas-
ures. J Agron Crop Sci 190:230–247.
13. Nichols TD, Wolf DC, Rogers HB, Beyrouty CA,
Reynolds CM. 1997. Rhizosphere microbial populations in
contaminated soils. Wat Air Soil Poll 95:165–178.
14. Yang CH, Crowley DE. 2000. Rhizosphere microbial com-
munity structure in relation to root location and plant iron
nutritional status. Appl Environ Microbiol 66:345–351.
15. Frederickson JK, Li SW, Brockman FJ, Haldeman DL,
Amy PS, Balkwill DL. 1995. Time-dependent changes in
viable numbers and activities of aerobic heterotrophic bacte-
ria in subsurface samples. J Microbiol Meth 21:253–265.
16. Frostegard A, Courtois S, Ramisse V, Clerc S, Bernillon
D, Le Gall F, Jeannin P, Nesme X, Simonet P. 1999.

Quantification of bias related to the extaction of DNA
directly from soils. Appl Environ Microbiol 65:5409–5420.
17. Phelps TJ, Fliermans CB, Garland TR, Pfiffner SM,
White DC. 1989. Methods for recovery of deep terrestrial
subsurface sediments for microbiological studies. J Microbiol
Meth 9:267–279.
18. USEPA. 1999. Field Sampling Guidance Document
#1215: Sediment Sampling. (http://www.epa.gov/region6/
qa/qadevtools/mod5_sops/soil_sampling/r9soilsample_gui.pdf).
Accessed 18 March 2014.
19. McLaughlin MR, Brooks JP, Adeli A, Read JJ. 2010. Com-
parison of selected nutrients and bacteria from common con-
tiguous soils inside and outside swine lagoon effluent spray
fields after long-term use. J Environ Qual 39:1829–1840.
20. Stenberg B, Johansson M, Pell M, Sjodahl-Svensson K,
Stenstrom J, Torstensson L. 1998. Microbial biomass and
activities in soil as affected by frozen and cold storage. Soil
Biol Biochem 30:393–402.
21. Stotzky G, Goos RD, Timonin MI. 1962. Microbial
changes occurring in soil as a result of storage. Plant Soil
16:1–18.
22. Maryniuk S, Oron J. 2008. Survival of Rhizobia in two soils
as influenced by storage conditions. Polish J Microbiol 57:
257–260.
23. Walker VK, Palmer GR, Voordouw G. 2006. Freeze-thaw
tolerance and clues to the winter survival of a soil commun-
ity. Appl Environ Microbiol 72:1784–1792.
24. Dadenko EV, Kazeev KS, Kolesnikov SI, Val’kov VF.
2009. Changes in the enzymatic activity of soil samples
upon their storage. Eurasian Soil Sci 42:1380–1385.
25. Gerba CP, Smith JE. 2005. Sources of pathogenic microor-
ganisms and their fate during land application of wastes.
J Environ Qual 34:42–48.
26. Janssen PH, Yates PS, Grinton BE, Taylor PM, Sait M.
2002. Improved culturability of soil bacteria and isolation
in pure culture of novel members of the divisions Acidobac-
teria, Actinobacteria, Proteobacteria, and Verrucomicrobia.
Appl Environ Microbiol 68:2391–2396.
27. Azevedo NF, Braganca SF, Simoes LC, Cerqueira L,
Almeida C, Keevil CW, Viera MJ. 2012. Proposal for a
method to estimate nutrient shock effects in bacteria. BMC
Res Notes 5:1–6.
28. Gentry TJ, Wang G, Rensing C, Pepper IL. 2004.
Chlorobenzoate-degrading bacteria in similar pristine soils
exhibit different community structures and population
dynamics in response to anthropogenic 2-, 3-, and 4-
chlorobenzoate levels. Microb Ecol 48:90–102.
29. Mars AE, Prins GT, Wietzes P, de Koning W, Janssen DB.
1998. Effect of trichloroethylene on the competitive behav-
ior of toluene-degrading bacteria. Appl Environ Microbiol 65:
208–215.
30. Hassen A, Saidi N, Cherif M, Boudabous A. 1998. Resist-
ance of environmental bacteria to heavy metals. Bio Technol
64:7–15.
31. Musgrove MT, Berrang ME, Byrd JA, Stern NJ, Cox NA.
2001. Detection of Campylobacter spp. in ceca and crops with
and without enrichment. Poult Sci 80:825–828.
2.6.3. Soil Sampling for Microbial Analyses ▪ 2.6.3-11
32. Williams LK, Sait LC, Cogan TA, Jorgensen F, Grongo-
no-Thomas R, Humphrey TJ. 2012. Enrichment culture
can bias the isolation of Campylobacter subtypes. Epidemiol
Infect 140:1227–1235.
33. McLaughlin MR, Brooks JP, Adeli A. 2009.
Characterization of selected nutrients and bacteria from
anaerobic swine manure lagoons on sow, nursery, and
finisher farms in the mid-south USA. J Environ Qual
38:2422–2430.
34. Oh SW, Kang DH. 2005. Rapid enumeration of Enterobacter
sakazakii in reconstituted milk formula by fluorogenic
most-probable number assay using 96-well microtiter plate.
J Rapid Meth Aut Mic 13:318–328.
35. Reynolds KA, Gerba CP, Abbaszadegan M, Pepper IL.
2001. ICC/PCR detection of enteroviruses and hepatitis
A virus in environmental samples. Can J Microbiol 47:
153–157.
36. Straub TM, Pepper IL, Gerba CP. 1995. Comparison of
PCR and cell culture for detection of enteroviruses in
sludge-amended field soils and determination of their trans-
port. Appl Environ Microbiol 61:2066–2068.
37. Straub TM, Pepper IL, Gerba CP. 1992. Persistence of
viruses in desert soils amended with anaerobically digested
sewage sludge. Appl Environ Microbiol 58:636–641.
38. Carey CM, Lee H, Trevors JT. 2006. Comparison of most
probable number-PCR and most probable number-foci
detection method for quantifying infectious Cryptosporidium
parvum oocysts in environmental samples. J Microbiol Meth
67:363–372.
39. Weber CF, Balasch MM, Gossage Z, Porras-Alfaro A,
Kuske CR. 2012. Soil fungal cellobiohydrolase I gene
(cbhI) composition and expression in a loblolly pine planta-
tion under conditions of elevated atmospheric CO2 and
nitrogen fertilization. Appl Environ Microbiol 78:3950–
3957.
40. Martin-Laurent F, Philippot L, Chaussod R, Germon JC,
Soulas G, Catroux G. 2001. DNA extraction from soils:
old bias for new microbial diversity analysis methods. Appl
Environ Microbiol 67:2354–2359.
41. Fitzpatrick KA, Kersh GJ, Massung RF. 2010. Prac-
tical method for extraction of PCR-quality DNA from
environmental soil samples. Appl Environ Microbiol 76:
4571–4573.
42. Watson RJ, Blackwell B. 2000. Purification and
characterization of a common soil component which inhib-
its the polymerase chain reaction. Can J Microbiol 46:
633–642.
43. Nocker A, Sossa-Fernandez P, Burr MD, Camper AK.
2007. Use of propidium monoazide for live/dead distinc-
tion in microbial ecology. Appl Environ Microbiol 73:5111–
5117.
44. Simon C, Daniel R. 2011. Metagenomic analyses: past
and future trends. Appl Environ Microbiol 77:1153–1161.
45. Pepper IL, Zerzghi HG, Bengson SA, Iker BC, Banerjee
MJ, Brooks JP. 2012. Bacterial populations within copper
mine tailings: long-term effects of amendment with class a
biosolids. J Appl Microbiol 113:569–577.
 Microbiological Sampling of Wastewater and Biosolids
NICOLETTE A. ZHOU, ERIC C. THOMPSON, AND JOHN SCOTT MESCHKE
2.6.4
INTRODUCTION
Overview of Wastewater Treatment
Wastewater is water impacted by anthropogenic sources.
Municipal wastewater, commonly referred to as sewage
(derived from “to the sea”), is a mix of domestic, commercial,
and industrial wastes. Domestic waste streams are composed of
feces, urine, food scraps, flush water, and other items flushed
or washed into sewer conveyances. Agricultural wastewater
may contain fecal wastes and can be an issue for receiving
waters. They may vary in strength and composition from
municipal wastewater. As a consequence of its source com-
position, the major contaminants in wastewater include bio-
degradable organic matter, volatile organic compounds,
xenobiotics, metals, suspended solids, nutrients, and micro-
bial pathogens and parasites (1). The objective of wastewater
treatment is to remove or significantly reduce contaminants
in these classes to a level at which they cannot adversely affect
human health or degrade water quality (2).
Wastewater treatment varies significantly by location.
Minimally, wastewater may be merely conveyed offsite and
then disposed of in a receiving water body without treatment.
In some less developed regions, wastewater receives little to
no treatment prior to discharge in a receiving water body
(e.g., pit latrines draining to a river, hanging toilets) (3).
On the opposite end of the spectrum, wastewater in other
locations is highly treated and reused for nonpotable and pot-
able uses (2).
Wastewater treatment can be divided into decentralized
and centralized treatment. “Decentralized treatment” refers
to treatment at or very near the site of generation, and the
types of treatment range from modern septic systems to
more primitive stabilization ponds or lagoons (2).
Centralized wastewater treatment involves conveyance of
the wastewater from the site of generation to a centralized
treatment plant. These treatment plants are designed to proc-
ess large volumes of wastewater from multiple sources on a
continual basis. A combination of physical, chemical, and
biological technologies and practices are used that systemati-
cally treat wastewater and sewage sludge. Modern wastewater
treatment typically involves pretreatment to remove large sol-
ids and grit, followed by primary settling of organic solids and
oil and grease removal. Secondary treatment refers to biolog-
ical stabilization of wastewater to remove biochemical oxygen
demand (BOD) to prevent oxygen depletion in receiving
doi:10.1128/9781555818821.ch2.6.4
2.6.4-1
waters. In some areas of the developed world, wastewater is
commonly disinfected to reduce the microbiological load,
including pathogenic organisms, in the effluent prior to dis-
charge (2).
This chapter focuses primarily on sampling of centralized
wastewater or sewage. However, many of the sampling aspects
described herein will be directly translatable to other types of
wastewater systems.
Matrix Description and Characterization
As wastewater moves through the treatment process, its char-
acter changes. Untreated wastewater is commonly referred to
a “raw wastewater” or influent. Water that has passed through
primary or secondary treatment is referred to as “primary efflu-
ent” or “secondary effluent,” respectively. After disinfection,
the wastewater is considered “disinfected effluent” or final
effluent. In some advanced treatment plants, advanced sec-
ondary, tertiary, or quaternary treatment processes are used
to further remove nutrients, pathogens, and some other con-
taminants. The effluent from these processes are similarly
named based on type of treatment, or may be collectively
referred to as “advanced treated effluent” or “reuse water.”
Raw Wastewater
Influent wastewater contains a high quantity of suspended
solids and microorganisms, including pathogens. As a result,
it is considered reflective of the prevalence of pathogen car-
riage within the source population and is best suited for envi-
ronmental surveillance of pathogen circulation within a
population (4, 5). However, these wastewaters may be heter-
ogeneous and are preferably sampled at locations with turbu-
lent flow to ensure good mixing. Influent can be sampled at
the headworks of the treatment plant or upstream at points
of convergence in the conveyance system. Sampling sites in
the conveyance system may be impacted by diurnal flow var-
iability more than are sampling sites in the plant (4).
Primary Effluent
During primary treatment, readily settleable solids and float-
ing materials are removed, reducing the suspended solids.
The concentration of solids produced during primary treat-
ment ( primary solids) ranges from 2% to 12%, depending
on the strength of the wastewater and if waste-activated sludge
from secondary treatment is returned to the primary sedimen-
tation tanks (2). The microbiological populations are also
2.6.4-2 ▪ SAMPLING METHODS
reduced, with log10 removals of 0.15 to 0.8 for indicator bac-
teria (6–8), 0 to 0.5 for viruses (6–8), and 0.62 for pathogenic
protozoa (8).
Secondary Effluent
During secondary treatment, biodegradable constituents
are stabilized and nutrients are removed or transformed in
a microbiologically mediated process. Following stabilization,
suspended and nonsettleable colloidal solids are captured
by flocs and settled (2). Effluent concentrations of microor-
ganisms and viruses are typically reduced by 1–2 log10 during
secondary treatment (2). However, removed organisms are
enriched in the settled solids (9). Solids from secondary treat-
ment are typically combined with primary solids and treated
as sludge.
Disinfected Effluent
Disinfection inactivates viruses, bacteria, and eukaryotic
microbes with varying effectiveness depending on the dis-
infectant. Typical disinfectants for wastewater include chlor-
ine and ultraviolet (UV) irradiation (2). Chlorine dioxide
and ozone are also used, but are less common in the United
States (2). Chlorine and chlorine dioxide remove viruses
well, though they are not effective for some chlorine-resistant
microorganisms, such as the protozoa Cryptosporidium oocysts
and Giardia cysts (2, 10). Viruses are also well removed by
ozone and moderately removed by UV and pasteurization (2).
Protozoa are well removed by UV and pasteurization and
moderately removed by ozone (2). Bacteria are well removed
by chlorine, chlorine dioxide, ozone, and pasteurization and
are moderately removed by UV (2). When sampling chlori-
nated effluent, it is essential to neutralize any residual disin-
fectant or samples may not be representative (11). Final
effluent samples should be collected at the most representa-
tive site prior to effluent discharge.
Tertiary Effluent and Reuse Water
Tertiary treatment is used in some locations to improve the
removal of residual suspended solids, and it is completed after
secondary treatment prior to disinfection (2). Sampling of
these types of effluent have very different considerations rel-
ative to other types of wastewater because of the extremely
low level of contaminants present. As such, they are not cov-
ered in detail in this chapter.
Sludge
Sewage sludge below 7% solids content behaves as a moder-
ately viscous liquid (12). It is heterogeneous and will settle
on standing. Settling may be overcome by adequate mixing.
Liquid sludge should be collected using the procedures dis-
cussed for other wastewater samples and is best sampled at
process points where it is well mixed or flow is turbulent.
For sampling of locations that are not well mixed, cross-
sectional sampling and compositing of subsamples may be
necessary (11).
Thickened sludge, with a higher than 7% solids content,
is resistant to flow, which causes nonuniform solids distri-
bution in tanks and lagoons. Sampling is best performed using
a thief sampler or weighted bottle sampler and using a cross-
sectional sampling approach to determine distribution of
organisms (12).
Dried or dewatered sludges can be sampled from conveyor
belts or sludge piles using a spoon, scoop, shovel, or auger
(12). Due to heterogeneity in piles, compositing of grab sam-
ples is necessary.
Biosolids
Sewage sludge produced by wastewater treatment processes to
meet the land application standards of the 40 CFR Part 503
regulation is commonly referred to as “Biosolids” (13, 14).
Biosolids are classified as either Class A or Class B depending
on the pathogen content or the processes used to further
reduce biological pathogens (Salmonella spp., enteric viruses,
and viable helminth ova) (13, 15). Processed biosolids with
no free liquids are best sampled when they are being trans-
ferred between processes. Grab samples should be collected
across the entire width of the conveyor belt and combined
into a composite sample. For biosolids that are stored in piles,
surface samples may not reflect the microbiological quality
of the pile (16–18). A cross-sectional sampling of these piles
should be performed to determine its heterogeneity.
SAMPLING SCHEMES
Purpose and Rationale for Sampling
There are a number of reasons that wastewater may need to
be sampled, including monitoring of treatment performance,
regulatory compliance in support of public health protection,
and conducting of research. These types of research may range
from understanding or optimizing treatment processes to
investigating the microbial ecology (19, 20). In addition,
sampling of raw wastewater can be used for environmental
surveillance of infectious agents (4, 5, 21) (and potentially
antimicrobial resistance genes [22, 23]) that are circulating
in a population.
There are a variety of sampling schemes suitable for micro-
biological or virological analysis of wastewater. The type of
scheme used will depend on the rationale for sampling. For
regulatory monitoring, components of the sampling plan
may be well outlined in the regulation. Plans for research or
process monitoring may allow for more flexibility. Any defen-
sible sampling plan should be developed based on well-
defined data quality objectives (24).
Quality Assurance Project Plan (QAPP)
A QAPP is a project-specific document that outlines how
environmental data will be collected, measured, and docu-
mented to ensure that the data meet project requirements.
From a sampling perspective, a QAPP specifically outlines
the rationale for the project or monitoring program and speci-
fies the details of the sampling procedures and requirements,
including a description of the procedures for sample col-
lection, documentation, handling, tracking, and custody. A
QAPP further describes in comprehensive detail the neces-
sary quality assurance (QA), quality control (QC), and other
technical activities that must be implemented to ensure
that the results of the work performed will satisfy the stated
performance criteria. The U.S. Environmental Protection
Agency (EPA), Department of Defense, and the Department
of Energy have published a joint program manual, the Uni-
form Federal Policy for Quality Assurance Project Plans (25).
Use of a QAPP is mandated for all projects under EPA author-
ity (25), though it is good practice for all projects to ensure
data quality.
Controls
A competent sampling plan will incorporate appropriate
controls to ensure method performance and determine bias,
precision, and risk of contamination. The number and types
of controls vary from method to method. Most standard
microbiological methods for analysis of wastewater involve
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-3
the use of recovery, sterility, and precision controls. Recovery
controls may include initial precision recovery samples to
demonstrate initial proficiency with a method, ongoing preci-
sion recovery samples to demonstrate consistent method
performance, and matrix spike samples to ensure method per-
formance in each matrix (26). Some methods also require
matrix spike duplicates to check the precision of a method
for a matrix (27). Good practice also includes the use of field
blanks, rinsate or dilution blanks, and positive and negative
assay controls. Field blanks (samples of sterile water trans-
ferred to sampling container in the field) should be collected
once each day samples are collected (24, 26). Research study
methods have more flexibility than standard methods in the
type and use of controls, but the same considerations to ensure
QC should be made. Sampling designs should include
adequate replication to determine the natural variability of
the samples. Good sampling design will include replication
at each level, from sample collection to assay.
Chain of Custody
The proper use of formal chain-of-custody procedures is an
important component of a good sampling plan, particularly
for any sampling performed for regulatory or legal reasons
(11). Properly designed and executed chain-of-custody forms
will ensure sample integrity from collection to data reporting.
Components of a well-implemented chain-of-custody plan
include the use of (a) labels to clearly identify samples; (b)
sample seals to detect unauthorized tampering; (c) a sampling
log book; (d) chain-of-custody forms to accompany samples
to laboratory; (e) prompt delivery or shipping of samples to
laboratory for analysis within specified transport times; (f )
inspection and formal acceptance (or rejection if conditions
are not met) of sample for analysis; and (h) analysis and for-
mal reporting of results.
At a minimum, information on the sample label should
include a unique sample number, sample type, identifier of
the collector, date and time of collection, place of collection,
and type of sample preservative or use of reducing agents (e.g.,
sodium thiosulfate and EDTA). Concurrently, all informa-
tion relevant to sample collection should be recorded in the
field sample log book and on the chain-of-custody form.
The chain-of-custody form should also contain signatures of
persons involved in the chain of possession and include dates
and times of possession. A separate sample analysis request
sheet should accompany samples to the laboratory to ensure
use of proper analytic methods.
Log Books
Field and laboratory sample log books should be maintained
to record information necessary to accurately reconstruct
the sample collection and analysis processes. A field log
book should record information related to sample collection
in the field, including purpose of sampling; location of sam-
pling point; where in treatment chain; name and address of
field contact; producer of material being sampled and address;
type of sample; and method, date, and time of preservation.
Any additional relevant metadata for the sample should
also be recorded (e.g., temperature, chlorine residual). Any
deviations from sample collection standard operating proce-
dures, handling, or receipt procedures should be recorded.
A separate sample log should be used to record information
on samples received in the laboratory, including details of
sample identification and origin, the necessary chain-of-
custody information, analyses performed, and final results.
Inadequately identified samples should not be accepted for
analysis (11).
Sampling Plans
Four factors must be established for every microbiological
sampling program: (a) number and frequency of samples to
be collected, (b) methods of sampling, (c) microbiological
parameters to be analyzed, and (d) sampling locations (28).
For regulatory sampling, these factors are typically defined
by the pertinent discharge permits. For nonregulatory sam-
pling, sampling parameters should be governed by statistical
methods selected for data analysis, sound scientific judgment,
and logistical considerations. Any sampling plan established
should adequately support the objectives of the intended
sampling.
Generally, the sampling frequency and number of samples
should consider the flow pattern into and through the waste-
water treatment plant (11). If the flow pattern is relatively
consistent across the day, independent samples collected
throughout the day may be equally representative. If the
flow peaks at certain times of the day, samples should be dis-
tributed to adequately represent the flow. For environmental
surveillance of disease transmission in a population, the
objective of the sampling may differ from regulatory sampling.
A judgmental sampling plan, which involves selecting sam-
pling locations subjectively based on knowledge of the site,
may be more appropriate, and sample collection targeting
the early peak in flow may be best (24).
Sample Types
Most microbiological analysis of wastewater is performed on
simple grab samples (11). Grab samples are single, discrete,
or individual samples collected at a specific time and location.
They provide information about microbiological concentra-
tions only at the moment from which they were collected.
The volume of the grab sample should be sufficient to carry
out all microbial analyses required. Typical volumes will
vary based on microorganism types, standard operating proce-
dures, and number of analytic replicates.
Grab samples for microbiological analyses are typically
collected manually. Automated systems for collection of sam-
ples from wastewater exist, but they are not commonly used
due to the limited holding times required for most microbio-
logical assays, cost, and concerns about cross-contamination
(29, 30). In addition, such equipment can allow solids sepa-
ration during sampling due to pumping rates.
Composite samples are routinely used for other (chemical
or physical) wastewater analyses (11). They are essentially
multiple discrete grab samples collected sequentially (over
time) and/or spatially that are combined in a manner propor-
tional to the flow. The benefits of composite samples include
a reduced analytic load, capture of natural variability in sam-
ples, and an expectation that the resulting composite will be
more representative of heterogeneous matrices. Compositing
of samples in the field can be difficult to do properly, so indi-
vidual samples are commonly sent to the laboratory and com-
posited by the technician prior to analysis.
Composite samples are not generally used for microbio-
logical analyses of wastewater, although they are more com-
monly used for collection of sludge and biosolid samples.
The main limitation of composite sampling is that individ-
ual samples are typically collected over a length of time
incompatible with hold times for many microbiological assays
(hours to a day). Composite durations for bacteriological and
viral analysis have a max time of an hour due to microbiolog-
ical changes in the first aliquot (12). However, composit-
ing can be useful for microbiological parameters with
longer acceptable hold times (e.g., helminth ova) however.
2.6.4-4 ▪ SAMPLING METHODS
Composite duration for helminth ova is 24 h as long as no
chemical or thermal stress (ammonia hydroxides, oxidants)
is expected (12).
Sample Collection
Samples for microbiological examination of wastewater or
biosolids should be collected in nonreactive borosilicate glass
or plastic bottles or in presterilized plastic bags appropriate for
microbiological use (28, 30). Plastic bags are especially useful
for free-flowing solids and thickened sewage sludges (12).
Containers should have tight-fitting lids to prevent leaking
during transit. Glass bottles that are cracked, chipped, or
etched should not be used for sample collection as constitu-
ents could stick to uneven surfaces. Similarly, crazed plastic
bottles or bottles that leak should be discarded (11). Where
legal action may be involved, the use of tamper-evident clo-
sures should be considered (11, 31).
Sterile sampling containers should be used for micro-
biological analyses. Disposable, single-use vessels are com-
mercially available (32). Multiple-use containers may also
be used, but must be cleaned and sterilized between uses.
This requires maintenance and monitoring of an adequate
sterilization program. Sampling bottles/containers should be
steam sterilized in an autoclave for at least 15 min at 121°C
and 15 PSI or baked for at least 2 h at 170°C (11, 18, 28,
30). If bottles or containers are sterilized at the laboratory,
one bottle per batch should be tested for sterility. If commer-
cial products are purchased for sample collection, one con-
tainer from each lot should be tested for sterility. If sterility
tests fail, resterilize the entire batch or lot. Containers must
also be free of any detergent residue that would adversely
affect the target microorganisms.
Sampling equipment used for microbiological sampling of
biosolid or sludge samples (e.g., coring devices, coliwasas,
pitchers, conduits, shovels, trowels) should be of dedicated
use and thoroughly cleaned and sterilized between uses
(12). Suitable sampling equipment materials generally
include glass, stainless steel, and plastic. Tools that are not
compatible with autoclaving may be disinfected by soaking
in a bleach solution (32). However, if bleach is used, equip-
ment must be rinsed thoroughly to prevent residual bleach
affecting the microbial population in the next sample.
Sampling Procedures
Safety Precautions
Written sampling protocols and a biological/chemical
hygiene plan that refers to worker safety should be included
in each QAPP. When collecting, samples should be consid-
ered as potentially infectious, requiring proper training in
biological safety. Personnel collecting biosolids or effluent
samples should wear proper personal protective equipment,
such as eye protection, disposable gloves, laboratory jackets,
and a mask (if aerosols are of immediate concern). Exterior
of sample bottles should be rinsed on site with 10% bleach
and placed in secondary containment for transport to the
laboratory.
Wastewater
Samples collected should be representative of the overall
microbiological content of the media sampled. Samples
should be collected from points in the treatment chain where
the flow is well mixed. Keep the sterilized sampling container
completely closed until the sample is to be collected. When
removing the lid, avoid contact with the inside of the lid.
Do not touch or handle the neck of the container. Grab
samples are collected by filling the sample containers by hold-
ing them just beneath the surface of the waste stream or under
the flow at an outfall pipe. The mouth of the container should
be faced into the current while keeping the hands, sampler,
and any other equipment downstream to minimize the chance
of contamination. Where it is impractical or unsafe to sam-
ple by hand, a sampling rod or bucket and rope should be
used. Consistent sampling procedures should be maintained
between samples. When the sample is collected, leave ample
air space in the bottle (at least 2.5 cm) to facilitate mixing by
shaking before examination (11). Laboratories should reject
sample bottles that are overfilled and request resampling or,
alternatively, add overfilled samples to a larger sterile sample
bottle in the laboratory to ensure adequate mixing.
Biosolids
Although the 503 Rule does not specifically state how to col-
lect biosolids samples for laboratory analysis, EPA Methods
1680, 1681, and 1682 provide guidance for biosolids product
sample collection at treatment plants (16–18). Biosolids sam-
pling frequency is determined by the annual metric tons of
biosolids produced and subsequently used for land applica-
tions or surface disposal (13). Frequency ranges from once a
year for facilities using or disposing of relatively small amounts
of biosolids, to once a month for facilities handling larger
amounts. Grab samples should be collected for pathogens
but may be composited at the laboratory. No fixed number
of individual samples are required (except for Class B patho-
gens). Enough material must be collected for the sample to be
representative of the media sampled.
Biosolids samples should be collected for analysis after the
final treatment process is complete, prior to shipping the
product from the treatment plant. The sampling location
may vary depending on how each plant is designed, but gen-
erally samples may be collected from digesters, conveyor belts,
pipes, bins, compost piles, and truck beds. For digester biosol-
ids, a large container should be placed under the sampling
pipe and the pipe flushed of stagnant biosolids. Once suffi-
ciently flushed, a large prelabeled sterile container should
be used to collect the product. Enough head space should
be left in the container before sealing to accommodate gas
production by the sample. For solids collected from a con-
veyer belt, a sterile sampling utensil (flat-blade trowel or
scoop) should be used to collect and transfer the material
into a sterile container. The container should be sealed and
transported to the laboratory for analysis. For drier biosolids
in compost piles, bins, or truck beds, a representative sample
should be collected by first removing the top overburden
material (approximately 6 inches), and then collecting ran-
dom samples with a sterile flat shovel from multiple locations.
Samples should be transferred into sterilized containers. If
compositing at the sample site, dispense the collected mate-
rial on a presterilized surface and thoroughly mix the sample
aggregate. Reduce the sample size by quartering and combine
a representative portion of the sample into a single prelabeled
sterile container. Close and seal the container before storage
or shipment.
Sample size reduction in the field can be difficult for
microbial samples because of the potential for microbial con-
tamination. As a result, the laboratory may be better equipped
to perform compositing and subsampling than samplers in the
field. Liquid samples can be mixed in the sample bottles by
shaking the bottles as long as the head space allows for
adequate mixing. Transfer of multiple smaller samples into
a larger bottle with adequate head space for mixing can be
performed to facilitate compositing and sample reduction.
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-5
Sample size reduction for thickened sewage sludges is more
difficult and may require hand mixing of a composite of sub-
samples and then subsampling a large number of small grabs
from the composited sample to form the smaller sample for
analysis (13). Dry solids samples can be adequately mixed
by shaking as long as sufficient head space exists in the sample
container. For microbial reduction testing, additives such as
wood chips should be removed from the sample by sieving
before size reduction or sample preparation.
Sample Preservation and Handling
Components of the waste stream may impact the stability of
the sample for microbiological analyses. For sampling of efflu-
ents disinfected with chemical oxidants, the addition of a
reducing agent to sampling containers is necessary to neutral-
ize any residual disinfectant activity. Sodium thiosulfate
(Na2S2O3) is an appropriate neutralizing agent for residual
halogens and prevents continued inactivation during sample
transit (11). For sampling chlorinated wastewater effluents,
sufficient Na2S2O3 should be added to give a concentration
of 100 mg/liter in the sample (11). Section 9060 of Standard
Methods (22nd ed.) provides instructions for the prepara-
tion of the sodium thiosulfate solution. When samples are
expected to contain high concentrations of metals, a chelat-
ing agent (e.g., EDTA) must be added to the sample con-
tainer before sterilization or mixed with Na2S2O3 before
adding to the bottle. EDTA should be added to the sample
bottles so that the final concentration with the sample is
372 mg/liter EDTA (33).
After sample collection, samples should be transport on
ice and stored under refrigeration until processed to minimize
biological changes in sample composition (28). Ensure that
sample bottles will not be immersed in water while in transit
(28). For local transport, samples may be transported in a
cooler on ice or cold packs. If shipping of the samples to the
laboratory is required, sample containers should be tightly
sealed, packed upright, cushioned, insulated, and chilled on
cold packs to keep the sample at approximately <10°C with-
out freezing the sample (11, 27). Follow IATA regulations for
packaging and shipping hazardous goods (www.IATA.org).
Ensure appropriate permits are in place (CDC, USDA) for
shipping and receiving.
Sample Hold Times
Microbiological analysis of wastewater and biosolids should
be initiated as soon as possible after collection of samples to
avoid changes in the microbial population. Samples that are
not processed within the specified time and under the pro-
per conditions can yield erroneous results, especially with
the less stable microorganisms (i.e., bacteria). As a result,
the allowable holding times for regulatory sampling of bacte-
ria tend to be very short (e.g., 6 h plus a 2 h processing time)
(34). In contrast, holding times for viruses, protozoa, and hel-
minths can be considerably longer (e.g., holding times for
Cryptosporidium and Giardia may be 96 h before initiation
of concentration) (27). Samples for analysis of bacteria and
helminth ova should be held at <10°C but not allowed to
freeze until processed. Samples for enteric virus analysis can
be held at <0°C.
Wastewater
To minimize microbial community changes, perform sample
analysis as soon as possible after sample collection. Transit
time should be less than 6 h (11). Sample temperature should
be verified and recorded on receipt through the use of either a
control water sample bottle or an infrared thermometer.
Samples should be processed within 24 h (11), though 6 h
or less is preferable (32). For organisms with well-described
standard methods, maximum hold times are typically speci-
fied in the method, for example, fecal coliforms (6 h) (30).
Biosolids
To meet Class A pathogen and indicator requirements, bio-
solids samples typically must be analyzed for Salmonella or
fecal coliforms within 8 h of collection (6 h max time for
transport and 2 h for sample processing at the laboratory)
(34). For Class B requirements, the typical holding time for
fecal coliforms testing is 8 h (6 h max time for transport
and 2 h for sample processing at the laboratory) (16, 17).
However, for fecal coliform analysis of Class A composted,
Class B anaerobically digested, or Class B aerobically digested
biosolids, the holding time may be extended to 24 h when
using EPA Method 1680 or 1681 for these types of sewage
sludge (16, 17).
Although best practice is still to process samples as soon as
possible after collection, biosolids samples being tested for
Class A requirements may be held at <0°C (best practice:
−70°C or colder) for up to 2 weeks for enteric virus analysis,
and samples can be store at 4°C for 1 month for helminth ova
analysis (15). Samples for helminth ova should not be frozen,
or helminth numbers may be reduced due to freeze thaw dam-
age (15).
Quantitative Approaches
The analysis of a microbiological or virological analyte in
wastewater may be scored either on a categorical ( positive
or negative) or discrete (e.g., enumerated colonies, (oo)cysts,
or plaques) basis. In some cases, the nonquantitative catego-
rical results can be expressed as a statistical estimation of con-
centration or titer using a most probable number approach.
The unit volume by which results are typically reported
may depend on the regulatory practice, volume sampled,
and custom. Bacteria are commonly reported on a per 100
ml basis, whereas protozoa or viruses may be reported per liter
or greater volume (2, 27). For biosolids, samples may be
reported based on per mass basis rather than per volume. Fecal
coliforms are reported on a MPN (or CFU) per gram of total
solids basis, while Salmonella are reported based on MPN per
4 g of total solids basis. Enteric viruses (PFU) and helminths
(ova) are also expressed on per 4 g of total solids basis for bio-
solids (13).
REGULATORY SAMPLING LIMITS
Relevant Regulation
The legal authority for regulatory sampling of wastewater is
the Federal Water Pollution Control Act (33 U.S.C.
§1251), otherwise known as the Clean Water Act (CWA).
The National Pollutant Discharge Elimination System
(NPDES) is a permit-based program under the CWA that
mandates compliance monitoring. Under the NPDES pro-
gram, wastewater treatment plants are required to obtain
permits before effluent can be discharged into receiving
waters or biosolids can be used for land applications or applied
to surface areas for disposal. The CWA established the
authority of the EPA to grant NPDES permits, but most of
the implementation is performed by state agencies under
authorization (35). The EPA handles implementation for
states that are not authorized and on tribal lands.
2.6.4-6 ▪ SAMPLING METHODS
Wastewater
Microbiological quality of effluent discharges from mun-
icipal wastewater treatment plants are regulated based on
fecal coliform load. The term coliform is nontaxonomical,
but generally applies to certain genera within the family
Enterobacteriaceae. Fecal coliforms are operationally defined
as Gram-negative, non–spore-forming, rod-shaped bacteria
that ferment lactose with the production of gas at 44.5°C
within 24 h (28). Fecal coliforms are widely used as an indi-
cator of fecal pollution because they are present in high num-
bers in fecal waste and are safer and easier to quantify than
pathogens directly.
Under the CWA, municipal wastewater treatment plants
are a category of discharger for which technology-based efflu-
ent limits have been promulgated. The microbiological per-
formance of wastewater treatment is measured by
monitoring of fecal coliform levels in final effluent. The tech-
nology based standard typically placed in the NPDES permits
for wastewater treatment plants is 200 organisms/100 ml as a
30-day geometric mean, and 400 organisms/100 ml as a 7-day
geometric mean. However, in cases where receiving waters
into which a treatment plant discharges are impaired for fecal
coliforms, more stringent discharge standards may be
enforced in discharge permits.
Biosolids
Under the Part 503 regulation, biosolids are classified as Class
A or B based on their microbiological quality (15). To meet
Class A regulatory requirements, the treatment plant process
must reduce pathogenic microorganisms levels in the final
biosolids product to below detectable levels to protect public
health and the environment (Table 1), and fecal coliforms
are used to demonstrate biosolids safety and process equiva-
lency. Before Class B biosolids can be applied for beneficial
use or surface disposal, seven samples of treated biosolids
must be collected and analyzed to verify that the fecal coli-
form geometric mean meets the regulatory compliance limit
(Table 1) (36).
Pathogenic microorganisms monitored in Class A biosol-
ids include Salmonella spp., enteric viruses, and viable hel-
minth ova. Salmonella spp. are so widespread that sufficient
numbers typically exist in untreated sewage sludge allowing
it to be used as an efficient indicator of biosolids treatment
(13). Enteric viruses are used as an indicator of biosolids treat-
ment effectiveness for two primary reasons: their routine pres-
ence in untreated sewage sludge and inherent resistance to
TABLE 1 Standards for Class A and B biosolids (Part 503
Pathogen Density Limits, U.S. EPA, 2003)
Indicator or Pathogen Density Limits (dry wt. basis)
Class A
Salmonella <3 MPN/4 g TSa or
Fecal coliforms <1,000 MPN/g TS and
Enteric viruses <1 PFU/4 g TS and
Viable helminth ova <1 ova/4 g TS
Class B
Fecal coliforms <2,000,000 MPN or CFU/g
(geometric mean)
a
TS: total solids (biosolids).
Values are in most probable number (MPN), colony-forming units
(CFU), or plaque-forming units (PFU).
environmental stresses and treatment, including heat and
low pH. Historically, helminth ova were commonly found
in relatively large numbers in sewage sludge in the United
States (37, 38). They demonstrate high resistance to treat-
ment conditions, making them useful for evaluating treat-
ment effectiveness.
The number of samples that must be analyzed for compli-
ance with Class A microbiological parameters is not specified,
but it is strongly recommended that multiple samples per sam-
pling event be analyzed for biosolids (13). The number of
samples taken must be sufficient to adequately represent qual-
ity of the biosolids being sampled. The final results for Class A
microbial analysis must be reported for each individual sample
(not averaged) (13).
If bacteria are not fully inactivated in a sample, concerns
over the potential for regrowth of the bacteria exist (13).
As a result, fecal coliform and Salmonella spp. determinations
must be made sufficiently close to the time that biosolids will
actually be used or disposed of to be indicative of whether the
potential for regrowth has been controlled. In contrast,
enteric viruses and viable helminth ova do not regain viabil-
ity once destroyed.
METHODS FOR SAMPLING OF BACTERIA
Bacterial Targets
Sampling for detection of bacteria in wastewater and biosolids
involves sample collection followed by concentration and/or
enrichment. Target bacteria include indicator bacteria, bac-
terial pathogens, process bacteria, and other specialized
bacteria.
Indicator bacteria include fecal coliforms and several
other groups of bacteria that have been suggested as indicators
of fecal pollution and have been measured in wastewater.
These bacteria include sulfate-reducing clostridia, fecal strep-
tococci, and enterococci. In addition, some targets have been
suggested as source-specific indicators and have been
included in microbial source tracking studies (e.g., Bacteroides
spp. and E. coli) (39, 40).
Many pathogenic bacteria are transmitted by the fecal-oral
pathway and their presence can be detected in wastewater.
Target bacterial pathogens include Salmonella spp., Shigella
spp., E. coli, Campylobacter spp., Vibrio cholerae, and Yersinia
enterocolitica. Others, like Leptospira, are shed in urine and
may also be present in wastewater. Still others are ubiquitous
in environmental waters (e.g., Aeromonas spp.).
While monitoring of bacterial pathogens is not required in
wastewater or biosolids (other than Salmonella), it can be
helpful to determine the source of outbreaks or track the cir-
culation of a pathogen within a population. Discharge of
wastewater to recreational waters can lead to outbreaks of
some waterborne diseases. Salmonella spp. is currently the sin-
gle most common cause of bacterial food poisoning in the
United States (41). The bacteria causes nonfatal gastroenter-
itis and can be shed in high concentrations in feces. Shigella
has a low infectious dose and causes approximately 500,000
cases of diarrhea in the United States each year (42). Out-
breaks of Shigella have been associated with fecal contamina-
tion of drinking water and are most common among children
(11). Pathogenic E. coli has a low infectious dose and
can cause diarrhea (11, 43). There have been several major
waterborne outbreaks from E. coli in water supplies and recrea-
tional waters impacted by wastewater (11). Campylobacter also
causes diarrhea, and people may be infected from water con-
taminated by animal feces and outbreaks in community water
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-7
supplies (11). Cholera, an acute diarrheal disease, is caused by
the bacteria Vibrio cholerae typically in waste-contaminated
waters (11). Yersinia enterocolitica can cause intestinal infection
and may be caused by consuming contaminated water (11).
It has been detected in secondary effluents and in untreated
and treated drinking water (11). Leptospira can cause leptospi-
rosis when individuals are exposed to contaminated waters
such as recreational water or floodwaters (11). Though Aero-
monas spp. can cause intestinal distress, they are naturally
found in waters worldwide. They have been detected in treated
drinking water, wastewater, and sludge (11).
Treatment plants may also take process control samples in
efforts to optimize performance or as internal QC and QA
checks on the effectiveness of treatment process. Organisms
that might be targeted include ammonia-oxidizing bacteria,
nitrite-oxidizing bacteria, filamentous bacteria, and other
specialized bacteria like trace organic contaminant (TOrC)-
degrading bacteria. Nitrifying bacteria include the ammonia-
oxidizing bacteria Nitrosospira and Nitrosomonas europea
as well as the nitrite-oxidizing bacteria Nitrobacter spp. and
Nitrospira. These bacteria have been monitored in wastewater
to determine factors effecting nitrification for a better under-
standing of how to improve this treatment process (20, 44,
45). Sampling for specialized bacteria like TOrC-degrading
bacteria can lead to identification of these bacteria and a bet-
ter understanding of TOrC removal during wastewater treat-
ment (46–48). This may lead to improve TOrC removal
through process modifications or additions.
General Sampling Considerations
Bacteriological samples have short holding times and should
always be transported and stored between 0°C and 10°C.
Insulated containers should be used to ensure proper mainte-
nance of storage temperature. Upon arrival in the laboratory,
samples should be refrigerated and analyzed as soon as possi-
ble, preferably within 2 h of collection. Samples should be
brought to room temperature before analysis. If sample hold-
ing times are not mandated, stability of sample could be
empirically determined for each matrix and target. Maximum
hold times for some bacteria have been indicated, including
total coliforms (6 h) (30), fecal coliforms (6 h) (30), fecal
streptococci (6 h) (30), enterococci (6 h) (26), and Salmo-
nella (6 h) (34).
Samples volumes for bacterial analysis have generally not
been specified for wastewater sampling, but sample volumes
required for bacterial detection generally range from 250 ml
to 10 liters. Samples collected should not be less than 100 ml
(11). The sample volumes required depend greatly on the
(a) treatment stage sampled, (b) the bacteria sampling for,
and (c) the goal of sampling.
Treatment stage sampled: Typical sample volume is inver-
sely correlated to level of treatment. The less treated the
wastewater is the less that is required for sampling. For raw
sewage 250 ml to 1 liter should be sufficient. Activated sludge
samples of 500 ml to 2 liters are typically collected (46, 49–
53). When sampling primary or secondary effluent, 1 to 10
liters should be sufficient. Also, if the wastewater is very
dilute, perhaps due to rain events or just consistently weak
wastewater, a minimum of 2 liters is required (54). For plants
that have advanced treatment, more volume may be required
for the final effluent (10–20 liters).
Target bacteria: Sample volumes have been specified for
detection of a few pathogens, including Salmonella (500 ml
to 20 liters), Shigella (100 ml to 1 liter), Campylobacter jejuni
(1–10 liters), Leptospira (100 ml to 1 liter), Legionella (1 liter),
and Aeromonas (200 ml to 1 liter) (11).
Sampling goal: The volume to be sampled is also dependent
on goal of the sampling (identification, detection, isolation,
or enumeration). Volumes stated above are for detection, iso-
lation, or enumeration using techniques like membrane filtra-
tion, multiple tube fermentation, or other culturing
techniques. If the end goal is identification by molecular
techniques, smaller volumes may be acceptable. When using
quantitative PCR (qPCR) for quantification of ammonia and
nitrite-oxidizing bacteria, DNA is typically extracted from 2
ml of activated sludge and diluted before running qPCR due
to the very high concentrations of genomic DNA in the sam-
ples as well as other PCR inhibitors present in wastewater
(55). Molecular techniques may still require “enrichment”
of the bacteria of interest. For example, when applying termi-
nal restriction fragment length polymorphism (T-RFLP) to
activated sludge samples (DNA extracted from 0.5 ml acti-
vated sludge) to assess the nitrifying community, nested
PCR was necessary to amplify the target genes prior to
T-RFLP (44). Fluorescent in situ hybridization (FISH) is a
powerful tool for visualizing the location of nitrifying bacteria
(45), filamentous bacteria that can cause bulking issues at the
wastewater treatment plant (56–58), and other bacteria of
interest (59, 60) in activated sludge flocs. When preparing
activated sludge samples for FISH, very small samples are
required (<1 ml), and dilution may be necessary. However,
detection limits may be higher than with other techniques,
and for use with wastewater effluents concentration and/or
enrichment may be required (60–62).
Additional processing may be required for very turbid sam-
ples prior to concentration or enrichment for the bacteria of
interest. Homogenization by blending of raw sewage, primary
effluent, or activated sludge samples may be necessary for rep-
resentative subsamples (28). Dilutions may also be required
prior to enumeration, isolation, or identification of bacteria
from turbid wastewater (28). Dilutions should not be held
for greater than 30 min prior to processing or else the bacteria
may become stressed, which can result in false negatives (11).
If dilutions are prepared, then a dilution blank should be proc-
essed to demonstrate sterility. If buffer is used to rinse the sam-
ple from sides of filtration apparatus, a rinsate control should
also be performed.
Concentration Methods
After sample collection, a known volume of wastewater is typ-
ically concentrated by filtration or centrifugation, allowing
for enumeration of the bacteria in later steps. Filters that
can be used for wastewater samples include diatomaceous
earth (11), glass fiber filters (11), or membrane filtration (typ-
ically cellulose ester, cellulose acetate, or cellulose nitrate fil-
ters) (63). Diatomaceous earth filtration involves placing an
absorbent pad on a filter device, covering it with diatoma-
ceous earth, filtering the samples, and placing the diatoma-
ceous earth “plug” in enrichment media. Disadvantages of
diatomaceous earth filters include that they clog easily with
turbid water and some bacteria can pass through them if the
filter is not formed properly (28). Glass fiber filters should
be used when filtering very large volumes (tested with samples
up to 70 liters) (64), and after filtration the filter will be
placed in enrichment media (28). If samples are highly turbid,
a series of filters can be used (28). A 0.45 µm pore size mem-
brane filter can be used for wastewater effluents. Filters used
for detection of bacteria by membrane filtration should
meet the criteria laid out in Standard Methods 9020B section
4i (11). Namely, the filters should have a diameter suitable for
the solids load and volume of the sample, a 0.45 µm mean
pore diameter, and they should retain bacterial cells (11).
2.6.4-8 ▪ SAMPLING METHODS
Filters for bacterial concentration should be sterile prior to fil-
tration. Sterility checks of filters and culture/enrichment
media should be performed.
Filtration for concentration has been demonstrated for a
variety of bacteria. Membrane filtration can be used for con-
centration and then detection of total coliforms in wastewater
(11), E. coli in wastewater (65), and Yersinia enterocolitica in
wastewater effluents (11). Salmonella can be concentrated
by membrane filtration (11) as well as diatomaceous earth
or glass fiber filters (11,28). For Campylobacter, wastewater
can be concentrated by filtering with a stainless steel filter
device that has filters ranging from 3 µm to 0.45 µm (11).
Mycobacterium in wastewater samples should be concentrated
by filtration (47 mm diameter, 0.45 µm mean pore diameter,
black filter, 100–500 ml depending on turbidity) (11).
A combination of centrifugation and filtration is also use-
ful for concentrating target bacteria. To concentrate Shigella,
centrifugation (1,520 × g, 15 min) of 200–250 ml is recom-
mended (11). Leptospira can be concentrated for by centrifu-
gation and a series of filters. Leptospira pass through 0.22 µm
filters, allowing them to be separated from other bacteria in
wastewater (11).
Concentration for some pathogens can also be accom-
plished using gauze pads (66, 67) by placing the pads in the
wastewater for 24 h to 1 week for collection. This technique
has been used to sample for V. cholerae (66), Salmonella (28,
67), and Aeromonas (68). However, it is not quantitative
and does not reflect changes in bacterial concentrations
over time (28).
Separation and Enrichment Methods
Primary enrichment of wastewater samples using broth media
can allow for easier isolation and/or identification of the bac-
teria of interest. Wastewater, concentrated wastewater, filter
“plugs,” or gauze pads can be added directly to the enrichment
broth. Selective and/or differential broth media can be used
with a variety of indicator bacteria and pathogens. Selective
media allows for target bacteria to grow but inhibits the
growth of other organisms. Differential media allow for differ-
entiation between closely related bacteria through the use
of chemicals or dyes. Some examples include enrichment of
fecal coliforms in phenol red lactose broth (63); total coli-
forms in lauryl tryptose broth (28); E. coli in lauryl tryptose
broth (11); fecal streptococci in purple broth base with carbo-
hydrates (28); Salmonella in selenite-based (with or without
dulcitol) (28), tetrathionate-based (with or without brilliant
green) (28), or malachite green-magnesium–based media (use
of two or more enrichment media are recommended) (11);
Shigella in selenite F broth or GN broth (11); Campylobacter
in Campylobacter broth, Campy-thio broth, Gifu anaerobe-
modified semisolid medium, or Preston medium (69);
V. cholerae in alkaline peptone broth (11); and Aeromonas
in a 1% alkaline peptone water followed by presence-absence
tests (11, 68).
Selective broth enrichments are also used for isolation of
TOrC-degrading bacteria. Enrichment of these bacteria is
achieved by inoculating activated sludge into nutrient-rich
media containing the target TOrC (70) and by enrichment
of activated sludge in modified mineral media containing
only the TOrC as the carbon source with repeated transfers
into fresh media after TOrC degradation occurs (46, 52). Iso-
lation of the bacteria is then accomplished by plating the
enrichments on nutrient-rich agar containing the TOrC.
The immunomagnetic separation/adenosine triphosphate
(IMS/ATP) method has recently been applied to wastewater
samples for detection of E. coli, Enterococcus spp., and
Bacteriodes fragilis (71–73). This technique involves separat-
ing the bacteria by immunomagnetic separation and then
quantifying the bacteria based on the bioluminescence from
ATP release from the cells (73).
METHODS FOR SAMPLING VIRUSES
Viral Targets
Viruses transmitted via the fecal-oral route are commonly
excreted in relatively large numbers with feces and are present
in domestic sewage. However, viral pathogens are not normal
flora in the intestinal tract; they are excreted only by infected
individuals. In addition, dilution, natural inactivation, and
wastewater treatment further reduce viral numbers. As a
result, pathogenic viruses are typically present in wastewater
at levels well below the levels in which they are shed and
may need significant concentration for adequate detection.
Human enteric viruses commonly found in wastewater
include the genera Norovirus, Rotavirus, Enterovirus, Hepatovi-
rus, Hepevirus, Mamastrovirus, and Mastadenovirus. All
together, there are more than 200 human enteric viruses
that may be found in wastewater (11). Some viruses (e.g.,
Norovirus) may demonstrate pronounced seasonality. How-
ever, recent studies using modern molecular techniques dem-
onstrated that domestic sewage contains many types of viruses
year-round. Enteric viruses are used as an indicator of biosol-
ids treatment effectiveness for two primary reasons: their rou-
tine presence in untreated sewage sludge and inherent
resistance to environmental stresses and treatment, including
heat and low pH.
Bacteriophages are viruses of bacteria. In wastewater, these
viruses far outnumber human enteric viruses. Coliphages are a
particular group of bacteriophage that infect E. coli and are
frequently used as fecal indicators. Monitoring of coliphages
in wastewater effluents could be an important tool as direct
potable reuse becomes increasingly more common. Detection
of male-specific and somatic coliphages in waters can be
accomplished by a two-step enrichment method (74) or by
the single agar layer procedure (75), although for water reuse,
testing may require significant concentration. Viruses of food
plants are also commonly present in wastewater, with recent
studies suggesting that they may be a very sensitive surrogate
indicator for human enteric viruses. Studies have demon-
strated that pepper mild mottle virus is extremely widespread
and abundant in municipal wastewater in diverse geographic
regions (76, 77).
General Sampling Considerations
Routine examination of water and wastewater for enteric
viruses is not currently recommended because of the com-
plexity of methods and the necessity of costly equipment
and facilities for safe processing and analysis. However, detec-
tion of viruses may be prudent or essential in special circum-
stances (e.g., wastewater reclamation, disease outbreaks, or
special research studies).
A number of methods have been described for the recov-
ery and detection of viruses from wastewater. However, cur-
rent methods are not well standardized, and even the
best-described methods continue to be modified. As a result,
any study initiated should include a preliminary evaluation of
virus recovery efficiency for the matrix being sampled by the
chosen method.
Holding times for samples that will be analyzed for viruses
are typically longer than for bacterial targets, and prompt
chilling of samples is appropriate for viruses as well as bacteria.
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-9
If sample concentrates cannot be processed and assayed
immediately after collection, they can be stored at room tem-
perature for up to 2 h or at refrigerator temperatures (4–10°C)
for up to 24 h with minimal virus losses (13). If samples are to
be stored for more than 24 h, they should be frozen at −70°C
or colder, rather than −10 to −20°C, because some enteric
viruses may become extensively inactivated at −10 to −20°
C. If possible, samples should be concentrated, eluted, and
extracted prior to freezing. This will increase their stability
by reducing bacterial and fungal contamination (11).
Recovery of viruses from wastewater is typically performed
on grab samples. Depending on the rationale for sampling,
a volume of less than 1 liter and possibly as small as a few
milliliters may suffice for recovering viruses from raw or pri-
mary treated sewage. However, larger volumes of sample are
typically collected and concentrated than are assayed for
viruses. The effective volume assayed may not be entirely
clear from published reports, which may not distinguish sam-
ple volume from assayed volume. For example, if a method
concentrates 2 liters of primary effluent to 100 ml, but then
only assays 2 ml of the concentrate for virus, then the effec-
tive sample volume would be 40 ml. Maximizing the effective
volume analyzed will increase the sensitivity of the method,
provided losses in performance do not increase with scale.
Concentration Methods
Due to their small size, size exclusion filtration has not been as
routinely used for virus recovery from wastewater samples.
Rather concentration methods have relied on virus adsorption
elution (VIRADEL), ultracentrifugation, or phase extraction.
Sample volume will be dictated in part by the concentration
methods selected. For filtration methods, either positive- or
negative-charged filtration media may be used for virus recov-
ery. Viruses tend to be negatively charged at natural pHs thus
will be attracted to positively charged media (78). Negatively
charged media can be pretreated with multivalent cations to
collect viruses (78). The appropriate filter diameter will
depend partly on water quality and the expected virus concen-
tration in the waste stream. Single-stage microporous filter
adsorption-elution methods have been used to recover viruses
from 100 ml raw sewage on 47 mm diameter flat disc filters and
from 3.8 to 4.6 liters secondary and tertiary sewage effluent on
90 or 142 mm diameter filters (11). A method’s efficacy may
vary widely depending on water quality. Recovery of viruses
from more thoroughly treated samples or from poorly mixed
samples may require significantly higher sample volumes,
potentially necessitating concentration at the site of collec-
tion and the use of alternative filter configurations (e.g., car-
tridge filters, tangential flow ultrafiltration). Cartridge filters
can accommodate sample volumes up to tens of liters of waste-
water (79). Hollow fiber ultrafilters are a type of tangential
flow ultrafiltration and can also process tens of liters, though
by size exclusion rather than VIRADEL (80).
Viruses can be directly concentrated from small volumes
of wastewater using a polyethylene glycol (PEG)/dextran
phase separation method that has been extensively used for
environmental surveillance of poliovirus (4). This method
involves creating a phase separation for a 500 ml sample by
addition of a PEG solution and a dextran solution in a sepa-
ration funnel, then recovering of the virus from the lower
hydrophilic phase and interphase.
Viruses in eluate volumes too large to be conveniently and
economically assayed directly in cell cultures (e.g., those
obtained from processing large volumes of water through car-
tridge or large disc filters) can be concentrated further (recon-
centrated) via several methods. Viruses in proteinaceous or
organic buffer eluates from all types of water can be reconcen-
trated by either “organic flocculation,” 1,2 aluminum hydrox-
ide adsorption precipitation (11). Organic flocculation, now
used widely, involves precipitating viruses by acidifying elu-
ates to pH 3.5, recovering precipitate via centrifugation,
and then resuspending it in a small volume of alkaline buffer
(13). Methods include precipitation with PEG/NaCl, and
skim milk precipitation methods have also been described
(81, 82). The major limitations of precipitation methods
are losses in virus recovery and inconsistency between virus
types and matrices. Additionally, viruses in nonproteinaceous
eluates (e.g., glycine-NaOH) can be reconcentrated via
adsorption to and elution from small microporous filters.
The eluate is adjusted to pH and ionic conditions for opti-
mum virus adsorption and filtered through a secondary
adsorbent, and then adsorbed viruses are eluted with a small
volume of eluent. Another method for reconcentration of pri-
mary filter eluates is centrifugal ultrafiltration (83).
Many virus concentration methods have achieved
adequate virus recoveries for water or wastewater samples in
studies with known seeded quantities of a specific enteric
viruses seeded into a matrix. However, few methods have
been broadly tested on a panel of different viruses across vary-
ing matrices. Although method effectiveness in field trials is
difficult to evaluate, some virus concentration methods have
successfully recovered naturally occurring enteric viruses, but
the performance recovery has not been consistently reported
throughout the literature.
Elution/Extraction Techniques
A number of eluants have been evaluated for use with the
above-described filtration methods. The most common
eluant used with positively charged filters is either a beef
extract–based eluant or a beef extract in combination with
amino acids (78). A soybean-based eluant (Optima RE)
has also been described with comparable recoveries (11).
Current negatively charged filter methods use a NaOH-based
elution following an H2SO4 acid rinse (84).
Separation and Purification Techniques
Viral sample concentrates, especially from wastewater, are
commonly contaminated with bacteria and fungi, which
can overgrow cell cultures and interfere with virus detection
and assay. In highly treated wastewaters, contamination is
controlled by the addition of antibiotics (e.g., penicillin-
streptomycin or gentamycin-kanamycin) and antifungal
agents (e.g., amphotericin B or nystatin), which are added
immediately after the sample is concentrated (11). If
penicillin-streptomycin or gentamycin-kanamycin are inad-
equate, use one or more additional antibiotics (e.g., aureomy-
cin, neomycin, or polymyxin B). To maximize the antibiotic
effects, samples should be incubated for 1–3 h at 25–37°C
after adding antibiotics. Freezing at −70°C after incubation
with antibiotics will also enhance bacterial destruction. Sam-
ples should be stored frozen at −70°C until assayed for viruses.
Organic extraction of samples with chloroform (CHCl3) or
Vertrel will further reduce bacterial and fungal contamina-
tion (11).
Viral metagenomics (virome studies) allows for detection
of viruses that are not cultivatable (85). Application of viral
metagenomics to wastewater samples have identified patho-
genic viruses in wastewater and provided new insights into
the abundances of these viruses (86–88). A good method
for purification of virions from concentrated samples is iso-
pycnic (density gradient centrifugation) using a cesium chlor-
ide gradient that separates viruses from bacteria and other
2.6.4-10 ▪ SAMPLING METHODS
debris based on density (89). This is a classical virological
technique but can be a bit clumsy to perform. Use of a DNAse
and/or RNAse treatment to eliminate free nucleic acid is
another important step in virome studies. After concentra-
tion and purification of the wastewater samples, RNA and
DNA can be extracted using readily available extraction
kits followed by amplification prior to sequencing (86, 87).
METHODS FOR SAMPLING OF
EUKARYOTIC MICROBES
Eukaryotic Targets
Sampling of wastewater or biosolids for detection of eukary-
otic microbes typically involves sample collection, followed
by concentration and separation prior to microscopic or
immunological detection. Target eukaryotic microbes
include pathogenic protozoa and helminth ova.
Pathogenic protozoa are widely distributed and form cyst,
oocyst, or spore stages that allow them to survive in the
environment and treatment processes. They are relatively
resistant to inactivation by chemical disinfectants and have
caused numerous waterborne outbreaks. From an epidemio-
logical standpoint, Cryptosporidium spp. and Giardia duo-
denalis (syn. G. intestinalis, G. lamblia) are by far the most
important (90). Not surprisingly they are the only two proto-
zoan targets that have standardized methods available for
their analysis. Except for Cryptosporidium spp. and Giardia
spp., there are no standardized or approved methods for
detecting pathogenic protozoa in water. However, various
approaches have been used to detect some of these protozoa.
Procedures typically use 1–5 m porosity filters or hollow
fiber ultrafilters, centrifugation for primary or secondary con-
centration, and then microscopy (with or without special
staining procedures) and/or PCR. Other protozoan parasites
associated transmitted by the fecal-oral route, and thus may
be expected in wastewater, include Entamoeba histolytica,
Cyclospora cayetanensis, Cystoisospora belli, Blastocystis hominis,
and Balantidium coli (11). Microsporidia (Encephalitozoon spp.
and Enterocytozoon bieneusi) are emerging pathogens of
immunocompromised individuals (91). They are more closely
related to fungi than to protozoa, but have been isolated from
wastewater. Acanthamoeba spp. and Naegleria fowleri are free-
living amoeba, but have been isolated from sewage (92).
Intestinal helminth infections have been steadily decreas-
ing in the United States. At one time, helminth ova were
commonly found in relatively large numbers in sewage sludge
in the United States (37, 38). Due to cleaner water supplies
and better personal hygiene practices, levels of helminth
ova in sewage sludge have decreased. Consequently, it is ques-
tionable if measurable levels of helminth ova exist in
untreated sludges, thus the absence of helminth ova in treated
sludge can no longer serve as an acceptable indicator of
adequate reduction by biosolids treatment. However, hel-
minth ova are still useful in evaluating treatment effectiveness
in process equivalency evaluations due to their high resist-
ance to treatment conditions (93).
General Sampling Considerations
Holding times for samples analyzed for protozoa and helminth
ova are longer than for bacterial targets though maintaining
the proper temperature is necessary (<10°C, but above freez-
ing). For example, samples for detection of Giardia and Cryp-
tosporidium using EPA Method 1693 are typically collected in
10-liter LDPE cubitainers and shipped to the analytic labora-
tory on ice. Cryptosporidium oocysts or Giardia cysts present in
samples can degrade, potentially biasing analytical results. If
the samples are frozen or the temperature is greater than 20°
C on receipt, the samples should be rejected. Samples should
be stored <10°C (but not frozen) prior to the initiation of
concentration (maximum of 96 h after collection) (27, 94).
Also, although viable helminth ova are relatively stable in
samples below 35°C (95°F) when chemicals such as lime,
chlorine, or ammonia are not used in the treatment process
(13), it is recommended to store them at 4°C (39.2°F).
This allows helminth ova to be held for up to 1 month before
analysis. Freezing should be avoided because it may result in
loss of viability for helminth ova (13).
Good practice dictates that sample processing should
be completed as soon as possible by the laboratory. If exten-
sion of holding times/temperatures is unavoidable, spiking
experiments should be conducted to determine their effects
on recovery. Specified holding times must be met when
using the method for regulatory compliance samples. While
sample processing should be completed the day the sample
is received whenever possible, it is not practical in many
instances. If necessary, sample processing may be halted after
filtration, application of the purified sample onto the slide,
or staining. For example, the standard method for Crypto-
sporidium and Giardia allows up to 72 h from application
of the purified sample to the slide to staining, and up to
7 days are acceptable between sample staining and examina-
tion (27).
Samples for microbiological analysis of eukaryotic targets
are typically collected as bulk samples and shipped to the
laboratory for processing through the entire method, with
typical volumes of 1–10 liters. However, the sample volumes
required depend greatly on the (a) treatment stage sampled
and (b) detection method to be used. In general, the less
treated the wastewater, the smaller the volume required.
This is particularly the case when sampling for helminth
ova. Many techniques have been published for the recovery
of helminth ova from wastewater; some techniques use larger
wastewater volumes when the waters have low total sus-
pended solids (TSS) (40 liters) than when the waters have
high TSS (1 liter) (95), and another recommends sampling
1 liter of raw or partially treated wastewater or sampling 10 lit-
ers of wastewater effluent (96). For sludge or biosolids, sam-
pling a minimum of 1-liter samples are typically collected,
with 300 g (estimated dry weight) of dried or thickened solids
or 1 liter of liquid samples (equivalent to ∼50 g of dry solids)
typically being processed for analysis (13). The detection
method used will also affect the sample volume required.
For example when sampling for detection of Giardia cysts by
microscopic examination, concentration of 1–100 liters of
raw and treated wastewater is required for effective detection,
depending on level of treatment (97). In contrast, detection
by qPCR may be accomplished from concentration of 1-liter
raw wastewater samples (98). It is important to remember that
frequently only a fraction of the concentrated sample is
assayed, thus the effective sample volume may be consider-
ably lower than the sample volume.
Also, the EPA method for detection of helminth ova
by microscopic examination requires 5 liters of wastewater
regardless of treatment stage (95), whereas detection by
real-time PCR has been demonstrated with concentration
of 1 liter raw wastewater and secondary effluent (99). Sam-
pling procedures for detection of helminth ova depends on
the treatment stage of the wastewater. It is important to
take 24-h composite samples or representative grab samples
of raw wastewater because there may large variations in hel-
minth ova concentrations throughout the day. As there are
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-11
less diurnal variations in helminth ova concentrations in
wastewater effluent, it is not necessary to take 24-h composite
samples (96).
Concentration and Elution Methods
Concentration of samples containing pathogenic protozoa
(Giardia cysts and Cryptosporidium oocysts) should be concen-
trated by filtration and centrifugation (94, 100). This can be
done using the Envirochek filter for 10-liter samples, or the
Envirochek HV filter or Filta-Max foam filter for 50-liter sam-
ples (27, 94). When using the Filta-Max foam filter protocol,
the sample can be further concentrated by centrifugation of
the eluent or by using the Filta-Max concentrator, which
filters the eluent. Samples containing Cryptosporidium can
also be concentrated using a portable continuous-flow centri-
fugation (94). Ultrafiltration (through 50 to 80 kDa molecu-
lar weight cut-off hollow fiber filters) has also been
demonstrated to efficiently recover Cryptosporidium oocysts
from large volumes of environmental water samples (80).
Sample for detection of helminth ova may be concen-
trated by filtration, centrifugation, and/or sedimentation.
Hookworm ova have been concentrated from secondary efflu-
ent by filtration on 90 mm polycarbonate filters with an 8 µm
pore size and from raw wastewater by centrifugation (99). The
World Health Organization (WHO) recommends concen-
trating the wastewater by settling it, siphoning off the super-
natant, centrifuging the sediments, and again removing the
supernatant. The solids were then processed to separate the
helminth ova (96).
EPA Method 1623 is widely considered the gold standard
for detection of Cryptosporidium and Giardia in water, and
adaptation of Method 1623 to raw sewage, primary clarified
effluent, and treated wastewater using Envirochek HV filters
detected Cryptosporidium in 67–83% of samples of influent,
secondary effluent, and final disinfected effluent from waste-
water reclamation facilities (1–400-liter volumes) (11). The
method has been formally adapted for performance in disin-
fected wastewaters and combined sewer overflow waters as
EPA Method 1693 (27). Due to the widely varying composi-
tion of matrices in raw sewage, secondary effluent, and tertiary
effluent, different methods are necessary to detect Crypto-
sporidium and Giardia in different types of wastewater. A pre-
liminary evaluation of sample volume/packed pellet ratios
can be rapidly obtained by centrifuging 250 ml of bulk sam-
ple and evaluating the packed pellet volume (27). If the sam-
ple appears to be filterable proceed to Section 12.2 of the US
EPA Method 1693 and follow the filtration option. If the
sample is not filterable, follow the direct centrifugation
option. The target sample volume for examination by the fil-
tration option is 10 liters or a maximum packed pellet volume
of 2 ml. The target sample volume for examination by the
direct centrifugation option is 1 liter or a maximum packed
pellet volume of 2 ml.
Separation Methods
The current approach for separation of Giardia cysts and Cryp-
tosporidium oocysts from other concentrated debris involves
purification by immunomagnetic separation (94). Well-
defined commercial immunoreagents are available for Crypto-
sporidium and Giardia but are not typically available for most
other protozoan targets (101, 102). Cryptosporidium oocysts
and Giardia cysts can also be separated from wastewater sam-
ples by isopycnic centrifugation. This method involves mix-
ing the sample with a density gradient and centrifuging
until separated according their buoyancy. Density gradients
used include Percoll solution (103, 104), sucrose (105), and
potassium citrate (105). This method can be applied to other
protozoan targets.
Separation of helminth ova has been completed by centri-
fugation, sedimentation, and/or flotation using MgSO4 or
ZnSO4. Hookworm ova in secondary effluent were separated
by centrifugation at low speeds and in raw wastewater were
separated by the MgSO4 flotation method (99). The WHO
recommended method suggests separating helminth ova
from wastewater by a combination of flotation (ZnSO4) and
biphasic sedimentation (95, 96).
Nucleic Acid Purification
Extraction of DNA from hardy (oo)cysts, spores, or ova can be
challenging. However, successful methods have been pub-
lished for many eukaryotic microbes. For example, DNA
extraction for detection of Giardia has been accomplished
using concentrated samples, which were then purified by
freeze/thawing, addition of proteinase K, sonication, addition
of Chelex 100 and PVP 360, and applied to DNeasy columns
(98). DNA from concentrated and purified hookworm ova in
secondary effluent and raw wastewater samples was extracted
using the MO Bio Power Soil DNA kit with several freeze/
thaw steps (99).
REFERENCES
1. Bitton G. 1999. Wastewater Microbiology, 2nd ed. Wiley-
Liss, New York, NY
2. Tchobanoglous G, Stensel HD, Tsuchihashi R, Burton
F, Abu-Orf M, Bowden G, Pfrang W. 2014. Waste-
water Engineering: Treatment and Resource Recovery, 5th
ed. McGraw-Hill Education, New York, NY.
3. Morella E, Foster V, Banerjee S. 2008. Climbing the Lad-
der: The State of Sanitation in Sub-Saharan Africa. World
Bank and the Water and Sanitation Program.
4. WHO. 2015. Guidelines on environmental surveillance
for detection of polioviruses. http://www.polioeradication.
org/Portals/0/Document/Resources/GPLN_publications/
GPLN_GuidelinesES_April2015.pdf last accessed 4/4/
2016.
5. Nakamura T, Hamasaki M, Yoshitomi H, Ishibashi T,
Yoshiyama C, Maeda E, Sera N, Yoshida H. 2015.
Environmental surveillance of poliovirus in sewage water
around the introduction period for inactivated polio vac-
cine in Japan. Appl Environ Microbiol 81:1859–1864.
6. Lucena F, Duran AE, Morón A, Calderón E, Campos C,
Gantzer C, Skraber S, Jofre J. 2004. Reduction of bacterial
indicators and bacteriophages infecting faecal bacteria in
primary and secondary wastewater treatments. J Appl Micro-
biol 97:1069–1076.
7. Zhang K, Farahbakhsh K. 2007. Removal of native coli-
phages and coliform bacteria from municipal wastewater
by various wastewater treatment processes: implications to
water reuse. Water Res 41:2816–2824.
8. Payment P, Plante R, Cejka P. 2001. Removal of indicator
bacteria, human enteric viruses, Giardia cysts, and Crypto-
sporidium oocysts at a large wastewater primary treatment
facility. Can J Microbiol 47:188–193.
9. Berg G. 1973. Removal of viruses from sewage, effluents,
and waters. Bull World Health Organ 49:451–460.
10. Zhou P, Di Giovanni GD, Meschke JS, Dodd MC. 2014.
Enhanced inactivation of Cryptosporidium parvum oocysts
during solar photolysis of free available chlorine. Environ
Sci Technol Lett 1:453–458.
11. American Public Health Association. 2012. Standard
Methods for the Examination of Water and Wastewater,
22nd ed. American Public Health Association, Washing-
ton, DC.
2.6.4-12 ▪ SAMPLING METHODS
12. Florida Department of Environmental Protection. 2014.
DEP-SOP-001/01 FS 2400 Wastewater Sampling.
13. U.S. EPA. 2003. Environmental Regulations and Tech-
nology Control of Pathogens and Vector Attraction in Sew-
age Sludge. EPA-625/R-92-013. EPA, Washington, DC.
14. National Research Council of the National Academies.
2002. Biosolids applied to land: advancing standards and prac-
tices. National Academies Press, Washington, DC [March
14, 2016]. Available from: http://www.nap.edu/catalog/
10426.
15. U.S. EPA. 1994. A Plain English Guide to the EPA Part 503
Biosolids Rule EPA/832-R-93/003. EPA 832/R-93/003.
Office of Wastewater Management. EPA, Washington, DC.
16. U.S. EPA. 2010. Method 1680: Fecal Coliforms in Sewage
Sludge (Biosolids) by Multiple-Tube Fermentation using
Lauryl Tryptose Broth (LTB) and EC Medium. EPA 821-
R-04-026. EPA Office of Water. EPA, Washington, DC.
17. U.S. EPA. 2006. Method 1681: Fecal Coliforms in Sewage
Sludge (Biosolids) by Multiple-Tube Fermentation using
A-1 Medium. EPA-821-R-06-013. EPA Office of Water.
EPA, Washington, DC.
18. U.S. EPA. 2006. Method 1682: Salmonella in Sewage
Sludge (Biosolids) by Modified Semisolid Rappaport-
Vassiliadis (MSRV) Medium. EPA-821-R-04-028. EPA
Office of Water. EPA, Washington, DC.
19. Rodriguez E, Garcia-Encina PA, Stams AJM, Maphosa F,
Sousa DZ. 2015. Meta-omics approaches to understand
and improve wastewater treatment systems. Rev Environ
Sci Bio-Technol 14:385–406.
20. Aoi Y, Miyoshi T, Okamoto T, Tsuneda S, Hirata A,
Kitayama A, Nagamune T. 2000. Microbial ecology of
nitrifying bacteria in wastewater treatment process exam-
ined by fluorescence in situ hybridization. J Biosci Bioeng
90:234–240.
21. Kokkinos P, Ziros P, Meri D, Filippidou S, Kolla S,
Galanis A, Vantarakis A. 2011. Environmental surveil-
lance: an additional/alternative approach for virological
surveillance in Greece? Int J Environ Res Public Health 8:
1914–1922.
22. Auerbach EA, Seyfried EE, McMahon KD. 2007. Tetra-
cycline resistance genes in activated sludge wastewater
treatment plants. Water Res 41:1143–1151.
23. Rizzo L, Manaia C, Merlin C, Schwartz T, Dagot C, Ploy
MC, Michael I, Fatta-Kassinos D. 2013. Urban wastewater
treatment plants as hotspots for antibiotic resistant bacteria
and genes spread into the environment: a review. Sci Total
Environ 447:345–360.
24. Zhang C. 2007. Fundamentals of Environmental Sampling
and Analysis. John Wiley & Sons, Hoboken, NJ.
25. U.S. EPA. 2005. Uniform Federal Policy for Quality Assur-
ance Project Plans: Evaluating, Assessing, and Document-
ing Environmental Data Collection and Use Programs—
Part 1: UFP-QAPP Manual. EPA-505-B-04-900A. EPA,
Washington, DC.
26. U.S. EPA. 2006. Method 1600: Enterococci in water by
membrane filtration using membrane-enterococcus indoxyl-
beta-D-glucoside agar (mEI). EPA 821-R-02-022. EPA,
Washington, DC.
27. U.S. EPA. 2014. Method 1693: Cryptosporidium and Giardia
in disinfected wastewater by concentration/IMS/IFA EPA
821-R-14-013. EPAOffice of Water. EPA, Washington, DC.
28. U.S. EPA. 1978. Microbiological methods for monitoring
the environment. EPA-600/8-78-017. EPA, Washington,
DC.
29. Roser D, Skinner J, LeMaitre C, Marshall L, Baldwin J,
Billington K, Kotz S, Clarkson K, Ashbolt N. 2002.
Automated event sampling for microbiological and related
analytes in remote sites: a comprehensive system. Water Sci
Technol Water Supply 2:123–130.
30. U.S. EPA. 1982. Handbook for sampling and sample pres-
ervation of water and wastewater. EPA-600/4-82-029.
EPA, Washington, DC.
31. New England Interstate Water Pollution Control Com-
mission. 2006. The Wastewater Treatment Plant Operators
Guide to Biosolids Sampling Plans. http://www.neiwpcc.
org/neiwpcc_docs/biosampleguide/biosampleguide_web.pdf
last accessed 4/4/2016.
32. Duncan D, Harvey F, Walker M, Australian Water Qual-
ity Centre. 2007. Regulatory monitoring and testing: water
and wastewater sampling. EPA. http://http://www.epa.sa.
gov.au/files/8494_guide_wws.pdf last accessed 4/4/2016.
33. American Public Health Association. 1998. Standard
Methods for the Examination of Water and Wastewater, 20th
ed. American Public Health Association, Washington, DC
34. U.S. EPA. 1993. Methods for microbiological analyses of
sewage sludges. EPA, Washington, DC.
35. U.S. EPA Office of Water. 2016. NPDES State Program
Information [Internet]. [March 15, 2016]. Available from
https://www.epa.gov/npdes/npdes-state-program-information
36. U.S. EPA. 2016. e-CFR: Title 40: Protection of Environ-
ment PART 503—STANDARDS FOR THE USE OR
DISPOSAL OF SEWAGE SLUDGE [Internet]. Electronic
Code of Federal Regulations. Available from: http://www.
ecfr.gov/cgi-bin/text-idx?SID=7db152db3d2c1f1cb7951811881
c83c8&mc=true&node=sp40.30.503.d&rgn=div6.
37. Yanko WA. 1988. Occurrence of Pathogens in Distribu-
tion and Marketing Municipal Sludges EPA/600/S1-87/
014. EPA, Washington, DC.
38. Jimenez-Cisneros BE. 2007. Helminth ova control in
wastewater and sludge for agricultural reuse. In Grabow
WOK, Water and Health. UNESCO, Eolss, Oxford, UK,
429–450.
39. Srinivasan S, Aslan A, Xagoraraki I, Alocilja E, Rose JB.
2011. Escherichia coli, enterococci, and Bacteroides thetaio-
taomicron qPCR signals through wastewater and septage
treatment. Water Res 45:2561–2572.
40. Mayer RE, Bofill-Mas S, Egle L, Reischer GH, Schade
M, Fernandez-Cassi X, Fuchs W, Mach RL, Lindner G,
Kirschner A, Gaisbauer M, Piringer H, Blaschke AP,
Girones R, Zessner M, Sommer R, Farnleitner AH.
2016. Occurrence of human-associated Bacteroidetes
genetic source tracking markers in raw and treated waste-
water of municipal and domestic origin and comparison
to standard and alternative indicators of faecal pollution.
Water Res 190:265–276.
41. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV,
Widdowson M-A, Roy SL, Jones JL, Griffin PM. 2011.
Foodborne illness acquired in the united states—major
pathogens. Emerg Infect Dis 17:7–15.
42. CDC. 2015. Shigella—Shigellosis [Internet]. [March 15,
2016]. Available from: http://www.cdc.gov/shigella/general-
information.html
43. CDC. 2015. E. coli [Internet]. [March 14, 2016]. Available
from: http://www.cdc.gov/ecoli/general/index.html
44. Siripong S, Rittmann BE. 2007. Diversity study of nitrify-
ing bacteria in full-scale municipal wastewater treatment
plants. Water Res 41:1110–1120.
45. Wagner M, Rath G, Koops HP, Flood J, Amann R. 1996.
In situ analysis of nitrifying bacteria in sewage treatment
plants. Water Sci Technol 34:237–244.
46. Zhou NA, Lutovsky AC, Andaker GL, Gough HL,
Ferguson JF. 2013. Cultivation and characterization of
bacterial isolates capable of degrading pharmaceutical
and personal care products for improved removal in acti-
vated sludge wastewater treatment. Biodegradation 24:
813–827.
47. Zhou NA, Kjeldal H, Gough HL, Nielsen JL. 2015.
Identification of putative genes involved in bisphenol a
 2.6.4. Microbiological Sampling of Wastewater and Biosolids ▪ 2.6.4-13
degradation using differential protein abundance analysis
of Sphingobium sp. BiD32. Environ Sci Technol 49:12232–
12241.
48. Kjeldal H, Zhou NA, Wissenbach DK, von Bergen M,
Gough HL, Nielsen JL. 2016. Genomic, proteomic, and
metabolite characterization of gemfibrozil-degrading organ-
ism Bacillus sp. GeD10. Environ Sci Technol 50:744–755.
49. Forster S, Snape JR, Lappin-Scott HM, Porter J. 2002.
Simultaneous fluorescent gram staining and activity assess-
ment of activated sludge bacteria. Appl Environ Microbiol
68:4772–4779.
50. Rosenberg Goldstein RE, Micallef SA, Gibbs SG,
George A, Claye E, Sapkota A, Joseph SW, Sapkota
AR. 2014. Detection of vancomycin-resistant entero-
cocci (VRE) at four U.S. wastewater treatment plants that
provide effluent for reuse. Sci Total Environ 466–467:
404–411.
51. Zhao D, Huang R, Zeng J, Yu Z, Liu P, Cheng S, Wu QL.
2014. Pyrosequencing analysis of bacterial community and
assembly in activated sludge samples from different geo-
graphic regions in China. Appl Microbiol Biotechnol 98:
9119–9128.
52. Miller TR, Colquhoun DR, Halden RU. 2010. Identifica-
tion of wastewater bacteria involved in the degradation of
triclocarban and its non-chlorinated congener. J Hazard
Mater 183:766–772.
53. Ajonina C, Buzie C, Rubiandini RH, Otterpohl R. 2015.
Microbial pathogens in wastewater treatment plants
(WWTP) in Hamburg. J Toxicol Environ Health-Part
-Curr Issues 78:381–387.
54. U.S. EPA. 1990. Methods for the investigation and pre-
vention of waterborne disease outbreaks. EPA/600/1-90/
005a. EPA, Washington, DC
55. Harms G, Layton AC, Dionisi HM, Gregory IR, Garrett
VM, Hawkins SA, Robinson KG, Sayler GS. 2003. Real-
time PCR quantification of nitrifying bacteria in a munic-
ipal wastewater treatment plant. Environ Sci Technol 37:
343–351.
56. van der Waarde JJ, Geurkink B, Henssen M, Heijnen G.
1998. Detection of filamentous and nitrifying bacteria in
activated sludge with 16S rRNA probes. Water Sci Technol
37:475–479.
57. Manz W, Amann R, Ludwig W, Wagner M, Schleifer
K-H. 1992. Phylogenetic oligodeoxynucleotide probes
for the major subclasses of proteobacteria: problems and sol-
utions. Syst Appl Microbiol 15:593–600.
58. Daims H, Brühl A, Amann R, Schleifer KH, Wagner M.
1999. The domain-specific probe EUB338 is insufficient
for the detection of all Bacteria: development and evalua-
tion of a more comprehensive probe set. Syst Appl Microbiol
22:434–444.
59. Loge FJ, Emerick RW, Thompson DE, Nelson DC,
Darby JL. 1999. Development of a fluorescent 16S
rRNA oligonucleotide probe specific to the family Entero-
bacteriaceae. Water Environ Res 71:75–83.
60. Moreno Y, Botella S, Alonso JL, Ferrus MA, Hernandez
M, Hernandez J. 2003. Specific detection of Arcobacter and
Campylobacter strains in water and sewage by PCR and flu-
orescent in situ hybridization. Appl Environ Microbiol 69:
1181–1186.
61. Konuma S, Satoh H, Mino T, Matsuo T. 2001. Compar-
ison of enumeration methods for ammonia-oxidizing bacte-
ria. Water Sci Technol 43:107–114.
62. Moreno Y, Ferrús MA, Alonso JL, Jiménez A, Hernández
J. 2003. Use of fluorescent in situ hybridization to evidence
the presence of Helicobacter pylori in water. Water Res 37:
2251–2256.
63. Lin SD. 1978. Development and evaluation of a 2 step
membrane filter method for fecal coliform recovery in
chlorinated sewage effluents. Illinois State Water Survey
Report of Investigation 87, pp. 1–14.
64. Levin MA, Fischer JR, Cabelli VJ. 1974. Quantitative
large-volume sampling technique. Appl Microbiol 28:515–
517.
65. U.S. EPA. 2009. Method 1603: Escherichia coli (E. coli) in
water by membrane filtration using modified membrane-
thermotolerant Escherichia coli agar (modified mTEC).
EPA 821-R-06-011. EPA, Washington, DC.
66. Barrett TJ, Blake PA, Morris GK, Puhr ND, Bradford
HB, Wells JG. 1980. Use of Moore swabs for isolating
Vibrio cholerae from sewage. J Clin Microbiol 11:385–388.
67. Moore B. 1948. The detection of paratyphoid carriers in
towns by means of sewage examination. Mon Bull Minist
Health Public Health Lab Serv 7:241–248.
68. Moyer NP. 1996. Isolation and enumeration of aeromonads.
In Austin B, Altwegg M, Gosling PJ, Joseph SW (eds.),
The Genus Aeromonas John Wiley & Sons, Chichester,
UK p. 39.
69. Goossens H, Butzler JP. 1992. Isolation and Identification of
Campylobacter spp. Campylobacter jejuni: Current Status
and Future Trends. American Society for Microbiology,
Washington, DC
70. Jiang L, Yang J, Chen J. 2010. Isolation and characteristics
of 17 beta-estradiol-degrading Bacillus spp. strains from acti-
vated sludge. Biodegradation 21:729–736.
71. Agidi S, Vedachalam S, Mancl K, Lee J. 2013. Effective-
ness of onsite wastewater reuse system in reducing bacterial
contaminants measured with human-specific IMS/ATP
and qPCR. J Environ Manage 115:167–174.
72. Muller EE, Grabow WOK, Ehlers MM. 2003. Immuno-
magnetic separation of Escherichia coli O157: H7 from envi-
ronmental and wastewater in South Africa. Water SA 29:
427–432.
73. Bushon RN, Likirdopulos CA, Brady AMG. 2009. Com-
parison of immunomagnedic separation/adenosine triphos-
phate rapid method to traditional culture-based method for
E. coli and enterococci enumeration in wastewater. Water
Res 43:4940–4946.
74. U.S. EPA. 2001. Method 1601: Male-specific (F+) and
Somatic Coliphage in Water by Two-step Enrichment Pro-
cedure. EPA 821-R-01-030. EPA Office of Water. EPA,
Washington, DC.
75. U.S. EPA. 2000. Method 1602: Male-specific (F+) and
Somatic Coliphage in Water by Single Agar Layer (SAL)
Procedure. EPA 821-R-01-029. EPA Office of Water.
EPA, Washington, DC.
76. Rosario K, Symonds EM, Sinigalliano C, Stewart J, Breit-
bart M. 2009. Pepper mild mottle virus as an indicator of
fecal pollution. Appl Environ Microbiol 75:7261–7267.
77. Kitajima M, Iker BC, Pepper IL, Gerba CP. 2014. Rela-
tive abundance and treatment reduction of viruses during
wastewater treatment processes—identification of poten-
tial viral indicators. Sci Total Environ 1488–489:290–6.
78. Ikner LA, Gerba CP, Bright KR. 2012. Concentration
and recovery of viruses from water: a comprehensive review.
Food Environ Virol 4:41–67.
79. Fagnant CS, Beck NK, Yang M-F, Barnes KS, Boyle DS,
Meschke JS. 2014. Development of a novel bag-mediated
filtration system for environmental recovery of poliovirus.
J Water Health 12:747–754.
80. Liu P, Hill VR, Hahn D, Johnson TB, Pan Y, Jothikumar
N, Moe CL. 2012. Hollow-fiber ultrafiltration for simulta-
neous recovery of viruses, bacteria and parasites from
reclaimed water. J Microbiol Meth 88:155–161.
81. Calgua B, Rodriguez-Manzano J, Hundesa A, Suñen E,
Calvo M, Bofill-Mas S, Girones R. 2013. New methods
for the concentration of viruses from urban sewage using
quantitative PCR. J Virol Meth 187:215–221.
2.6.4-14 ▪ SAMPLING METHODS
82. Kiulia NM, Mans J, Mwenda JM, Taylor MB. 2014. Nor-
ovirus GII.17 predominates in selected surface water sour-
ces in Kenya. Food Environ Virol 6:221–231.
83. Soto-Beltran M, Ikner LA, Bright KR. 2013. Effective-
ness of poliovirus concentration and recovery from treated
wastewater by two electropositive filter methods. Food
Environ Virol 5:91–96.
84. Haramoto E, Katayama H. 2013. Application of acidic elu-
tion to virus concentration using electropositive filters.
Food Environ Virol 5:77–80.
85. Mokili JL, Rohwer F, Dutilh BE. 2012. Metagenomics
and future perspectives in virus discovery. Curr Opin Virol
2:63–77.
86. Aw TG, Howe A, Rose JB. 2014. Metagenomic
approaches for direct and cell culture evaluation of the viro-
logical quality of wastewater. J Virol Meth 210:15–21.
87. Bibby K, Peccia J. 2013. Identification of viral pathogen
diversity in sewage sludge by metagenome analysis. Environ
Sci Technol 47:1945–51.
88. Tamaki H, Zhang R, Angly FE, Nakamura S, Hong P-Y,
Yasunaga T, Kamagata Y, Liu W-T. 2012. Metagenomic
analysis of DNA viruses in a wastewater treatment plant
in tropical climate. Environ Microbiol 14:441–452.
89. Casas V, Rohwer F. 2007. Phage metagenomics. Methods
Enzymol 421:259–268.
90. Marshall MM, Naumovitz D, Ortega Y, Sterling CR.
1997. Waterborne protozoan pathogens. Clin Microbiol
Rev 10:67–85.
91. Izquierdo F, Castro Hermida JA, Fenoy S, Mezo M,
González-Warleta M, del Aguila C. 2011. Detection of
microsporidia in drinking water, wastewater and recrea-
tional rivers. Water Res 45:4837–4843.
92. García A, Goñi P, Clavel A, Lobez S, Fernandez MT,
Ormad MP. 2011. Potentially pathogenic free-living amoe-
bae (FLA) isolated in Spanish wastewater treatment plants.
Environ Microbiol Rep 3:622–626.
93. U.S. EPA. 2015. Basic Information: Pathogen Equivalency
Committee [Internet]. [March 16, 2016]. Available from:
https://www.epa.gov/biosolids/basic-information-pathogen-
equivalency-committee#fecal.
94. U.S. EPA. 2005. Method 1623: Cryptosporidium and
Giardia in Water by Filtration/IMS/FA. EPA, Washington,
DC.
95. Maya C, Jimenez B, Schwartzbrod J. 2006. Comparison of
techniques for the detection of helminth ova in drinking
water and wastewater. Water Environ Res Res Publ Water
Environ Fed 78:118–124.
96. Ayres RM, Mara DD. 1996. Analysis of Wastewater for
Use in Agriculture—A Laboratory Manual of Parasitologi-
cal and Bacteriological Techniques. World Health Organ-
ization, Geneva, Switzerland.
97. Nasser AM, Vaizel-Ohayon D, Aharoni A, Revhun M.
2012. Prevalence and fate of Giardia cysts in wastewater
treatment plants. J Appl Microbiol 113:477–484.
98. Guy RA, Payment P, Krull UJ, Horgen PA. 2003. Real-
time PCR for quantification of Giardia and Cryptosporidium
in environmental water samples and sewage. Appl Environ
Microbiol 69:5178–5185.
99. Gyawali P, Sidhu JPS, Ahmed W, Jagals P, Toze S. 2015.
Rapid concentration and sensitive detection of hookworm
ova from wastewater matrices using a real-time PCR
method. Exp Parasitol 159:5–12.
100. Giangaspero A, Cirillo R, Lacasella V, Lonigro A, Mar-
angi M, Cavallo P, Berrilli F, Di Cave D, Brandonisio
O. 2009. Giardia and Cryptosporidium in inflowing water
and harvested shellfish in a lagoon in Southern Italy. Para-
sitol Int 58:12–17.
101. Hsu B-M, Huang C. 2007. IMS method performance anal-
yses for Giardia in water under differing conditions. Environ
Monit Assess 131:129–34.
102. Bukhari Z, McCuin RM, Fricker CR, Clancy JL. 1998.
Immunomagnetic separation of Cryptosporidium parvum
from source water samples of various turbidities. Appl Envi-
ron Microbiol 64:4495–4499.
103. Nieminski EC, Schaefer FW, Ongerth JE. 1995. Compar-
ison of two methods for detection of Giardia cysts and Cryp-
tosporidium oocysts in water. Appl Environ Microbiol 61:
1714–1719.
104. Arrowood MJ, Sterling CR. 1987. Isolation of Crypto-
sporidium oocysts and sporozoites using discontinuous
sucrose and isopycnic Percoll gradients. J Parasitol 73:
314–319.
105. Rose JB, Landeen LK, Riley KR, Gerba CP. 1989. Eval-
uation of immunofluorescence techniques for detection of
Cryptosporidium oocysts and Giardia cysts from environ-
mental samples. Appl Environ Microbiol 55:3189–3196.
Environmental Public Health
Microbiology
VOLUME EDITOR: MARYLYNN V. YATES
SECTION EDITORS: GARY A. TORANZOS, MARK P. BUTTNER, ED TOPP, VALERIE J. HARWOOD,
AND MARYLYNN V. YATES

3.1 WATER 3.2.3 Analysis of Bioaerosol Samples / 3.2.3-1

PATRICIA CRUZ AND MARK P. BUTTNER
3.1.1 Current and Developing Methods for the 3.2.4 Fate and Transport of Microorganisms in
Detection of Microbial Indicators in Air / 3.2.4-1
Environmental Freshwaters and Drinking GARY S. BROWN AND ALAN JEFF MOHR
Waters / 3.1.1-1 3.2.5 Airborne Fungi and Mycotoxins / 3.2.5-1
TASHA M. SANTIAGO-RODRIGUEZ, DE-WEI LI, ECKARDT JOHANNING, AND
JULIE KINZELMAN, AND GARY A. TORANZOS CHIN S. YANG
3.1.2 Best Practices for Cyanobacterial Harmful 3.2.6 Airborne Bacteria, Archaea, and
Algal Bloom Monitoring / 3.1.2-1 Endotoxin / 3.2.6-1
TIMOTHY G. OTTEN AND HANS W. PAERL PETER S. THORNE, CAROLINE DUCHAINE,
3.1.3 Assessing the Efficiency of Wastewater AND PASCALE BLAIS LECOURS
Treatment / 3.1.3-1 3.2.7 Airborne Viruses / 3.2.7-1
GRACIELA RAMÍREZ TORO AND HARVEY SYED A. SATTAR, NITIN BHARDWAJ, AND
MINNIGH M. KHALID IJAZ
3.1.4 Epidemiologic Aspects of Waterborne 3.2.8 Aerobiology of Agricultural
Infectious Disease / 3.1.4-1 Pathogens / 3.2.8-1
SAMUEL DOREVITCH ESTELLE LEVETIN
3.1.5 Waterborne Enteric Viruses: Diversity, 3.2.9 Legionellae and Legionnaires’
Distribution, and Detection / 3.1.5-1 Disease / 3.2.9-1
MORTEZA ABBASZADEGAN AND ABSAR ALUM CLARESSA E. LUCAS AND
3.1.6 Detection of Protozoa in Surface and Finished BARRY S. FIELDS
Waters \ 3.1.6-1
ABSAR ALUM, ERIC N. VILLEGAS, SCOTT
P. KEELY, KELLY R. BRIGHT, LAURA 3.3 SOIL
Y. SIFUENTES, AND MORTEZA ABBASZADEGAN 3.3.1 Pathogenic Viruses and Protozoa
3.1.7 Drinking Water Microbiology / 3.1.7-1 Transmitted by Soil / 3.3.1-1
MARYLYNN V. YATES PASCAL DELAQUIS, JULIE BRASSARD, AND
ALVIN GAJADHAR
3.3.2 Natural Soil Reservoirs for Human
3.2 AIR Pathogenic and Fecal Indicator
3.2.1 Introduction to Aerobiology / 3.2.1-1 Bacteria / 3.3.2-1
PAULA KRAUTER AND LINDA MARIA LAURA BOSCHIROLI, JOSEPH
D. STETZENBACH FALKINHAM, SABINE FAVRE-BONTÉ,
3.2.2 Sampling for Airborne SYLVIE NAZARET, PASCAL PIVETEAU,
Microorganisms / 3.2.2-1 MICHAEL SADOWSKY, MURULEE
SERGEY A. GRINSHPUN, MARK P. BUTTNER, BYAPPANAHALLI, PASCAL DELAQUIS, AND
GEDIMINAS MAINELIS, AND KLAUS WILLEKE ALAIN HARTMANN
3.4 MICROBIAL SOURCE TRACKING
3.4.1 The Evolving Science of Microbial Source
Tracking / 3.4.1-1
VALERIE J. HARWOOD, CHARLES HAGEDORN,
AND MICHAEL SADOWSKY
3.4.2 Validation of Microbial Source Tracking
Markers and Detection Protocols:
Considerations for Effective
Interpretation / 3.4.2-1
ASJA KORAJKIC, DON STOECKEL, AND JOHN
F. GRIFFITH
3.4.3 Overview of Microbial Source Tracking
Methods Targeting Human Fecal Pollution
Sources / 3.4.3-1
ORIN C. SHANKS, HYATT GREEN, ASJA
KORAJKIC, AND KATHARINE G. FIELD
3.4.4 Methods of Targeting Animal Sources of Fecal
Pollution in Water / 3.4.4-1
ANICET R. BLANCH, ELISENDA BALLESTÉ,
JENNIFER WEIDHAAS, JORGE SANTO
DOMINGO, AND HODON RYU
3.4.5 Microbial Source Tracking: Field Study
Planning and Implementation / 3.4.5-1
JULIE KINZELMAN AND WARISH AHMED
3.4.6 Fecal Indicator Organism Modeling and
Microbial Source Tracking in Environmental
Waters / 3.4.6-1
MEREDITH B. NEVERS, MURULEEDHARA
N. BYAPPANAHALLI, MANTHA
S. PHANIKUMAR, AND RICHARD L. WHITMAN
3.5 MICROBIAL RISK ASSESSMENT
3.5.1 Risk Assessment Framework / 3.5.1-1
MARYLYNN V. YATES
3.5.2 Exposure Assessment / 3.5.2-1
SUSAN R. PETTERSON AND NICHOLAS
J. ASHBOLT
3.5.3 Dose-Response Modeling and Use: Challenges
and Uncertainties in Environmental
Exposure / 3.5.3-1
MARK H. WEIR
 Current and Developing Methods for the Detection
of Microbial Indicators in Environmental
Freshwaters and Drinking Waters
TASHA M. SANTIAGO-RODRIGUEZ, JULIE KINZELMAN, AND
GARY A. TORANZOS
3.1.1
Freshwaters are vital for recreational and drinking purposes;
however, they may become contaminated, resulting in
gastrointestinal and respiratory illnesses and eye and skin
infections, many caused by enteric pathogens (1). While
human enteric pathogens can directly affect humans, ill-
nesses resulting from the exposure to zoonotic enteric patho-
gens can also occur. Enteric pathogens can be detected by
culture methods, but these can be relatively expensive, not
all enteric pathogens are culturable, and their detection
requires highly trained personnel. For these reasons, the
microbiological quality of recreational and drinking waters
has been routinely assessed using microbial indicators for
the past 100 years (2).
The use of microbial indicators offers several advantages
compared to the detection of enteric pathogens: (a) detec-
tion methods are relatively simple, (b) results are obtained
within a reasonable amount of time (although this is cur-
rently under debate), (c) analyses do not required highly
trained personnel, and (d) indicators are often present in
higher numbers than pathogens and thus concentration
methods are not required (3). As discussed shortly, no cur-
rently used microbial indicator possesses all the characteris-
tics required for an ideal indicator of fecal pollution.
Additionally, those microbial indicators intended to identify
the source of the fecal contamination (microbial source
tracking) are a subject of debate as many do not necessarily
accomplish this with 100% reliability (2). For this reason,
several researchers have proposed the use of a “toolbox” of
methods, in which the drawbacks of one indicator should
be covered by another (4).
This chapter focuses on microbial indicators and the
methods used to assess the microbiological quality of sur-
face and drinking waters. Several methods may apply to
both aquatic environments, and others may be specific
to either surface (fresh or marine) or drinking waters.
Although this chapter discusses the use of enteric microor-
ganisms to infer fecal pollution, the use of microbial
source tracking (MST) tools will be discussed as well.
The MST field has increased in the recent years due to
the importance of identifying possible sources of fecal con-
tamination in an effort to focus remediation, avoid eco-
nomic losses, and reduce or eliminate possible risks to
public health.
doi:10.1128/9781555818821.ch3.1.1
3.1.1-1
CURRENT MICROBIAL INDICATORS
OF FECAL CONTAMINATION
Total and Thermotolerant Coliforms
The presence of total coliforms in freshwaters and drinking
waters has long been used to infer fecal pollution. These bac-
teria are present in the intestinal tract of animals, can ferment
lactose and produce gas within 24–48 h at 35°C, and include
the genera Klebsiella, Enterobacter, Citrobacter, and Escherichia.
In freshwaters, the presence of total coliforms may indicate
fecal pollution, although their presence has also been associ-
ated with nonfecal sources of contamination (e.g., runoff ). In
drinking waters, the presence of total coliforms (and other
indicator bacteria) may indicate failures of treatment and dis-
tribution systems. While total coliforms are used as indicators
of fecal pollution, some species can also cause gastrointestinal
illnesses. For this reason, the U.S. Total Coliform Rule was
established in 1989 to ensure that coliforms are not detected
in drinking waters (5). Because of their ubiquitous nature in
the environment, total coliforms are not recommended as
an indicator of fecal pollution in freshwaters used for recrea-
tional purposes (6).
Thermotolerant coliforms are part of the intestinal
microbiota of warm-blooded animals, and fecal discharges
contain large numbers of bacteria belonging to this group.
Thermotolerant coliforms are distinguished from total coli-
forms by their ability to grow at 45°C (7, 8). Of the thermo-
tolerant coliforms, the genera Escherichia and Klebsiella are
currently used as an index of fecal contamination in fresh-
waters and drinking waters. However, thermotolerant coli-
forms may also originate from nonfecal sources, such as soil
(9), can replicate in subtropical and tropical aquatic envi-
ronments, and have been found to be part of their environ-
mental microbiota (10, 11). In addition, thermotolerant
coliforms have been detected in the absence of enteric patho-
gens and vice versa, and thus there is a lack of correlation
between the number of these indicator bacteria and enteric
pathogens (8, 12).
Escherichia coli as Indicator
Among the coliforms, E. coli has received greater attention
due to its apparent specificity to fecal material (13). Advan-
tages and disadvantages of using E. coli as indicators of fecal
3.1.1-2 ▪ WATER
pollution include those discussed for the total and thermoto-
lerant coliforms. Several methods are currently being used to
detect and quantify E. coli in freshwaters and drinking waters
and are discussed below.
Enzyme-Based Methods for the Detection of E. coli
Enzyme-based methods are generally accepted for the detec-
tion of E. coli in both freshwaters and drinking waters
due to their ease of use and affordability. Among the enzyme-
based methods, Colilert is the most popular, simultaneously
detecting total coliforms and E. coli. The activity of β-galac-
tosidase and β-glucuronidase (enzymes that are characteristic
of total coliforms and E. coli, respectively) cleave ONPG
(o-nitrophenyl-β-D-galactopyranoside) and MUG (4-meth-
ylumbelliferyl-β-D-glucuronide), producing characteristic
color reactions. Results are obtained within 24 h after incuba-
tion at 35°C. Similar methods, such as Colilert-18 and
Colisure, are currently available, have been simultaneously
tested and compared, and in the case of Colilert-18 can
reduce the incubation period from 24 to 18 h. Other
enzymatic methods, such as E*Colite, uses X-Gal (5-bromo-
4-chloro-3-indolyl-β-D-galactopyranoside) and MUG as sub-
strates. Although many variants are available, one of the
disadvantages of these methods is the time it takes to
obtain results (18–48 h), and the potential need for confir-
mation (14).
Filtration Methods for the Detection of E. coli
m-Coli Blue 24, a liquid-based media approved by the U.S.
EPA, can be classified as both an enzyme-based and a
filtration method. Samples are filtered and membranes are
placed into the m-Coli Blue liquid media. Antibiotics in
the media inhibit the growth of background bacteria, the
dye TTC (2,3,5-triphenyltetrazolium chloride) turns total
coliform colonies red, and the action of β-glucuronidase
results in blue E. coli colonies. One drawback of this method
is that it has shown a failure rate of approximately 23% to
detect target microorganisms, and false positive results are
obtained when water samples are spiked with Aeromonas
spp. (14).
Supplemented agar-based media are among the most
common filtration techniques. One example is membrane
thermotolerant E. coli (m-TEC) agar, which relies on the
action of sodium lauryl sulfate and sodium deoxycholate as
selective agents against Gram-positive bacteria and bromoc-
resol purple and bromophenol red as pH indicators. Posterior
confirmation is needed when using this medium, in which
filters are saturated with a urea substrate and held at room
temperature for 15–20 min. Typical colonies exhibit a yellow
or yellow-brown color (15). A modification of this method
(Method 1603) was approved by the U.S. EPA, which does
not require a confirmatory phase. The modified m-TEC
medium contains 5-bromo-6-chloro-3-indolyl-β-D-glucuro-
nide, that when catabolized to glucuronic acid produces a
typical red or magenta-colored colony (16).
Most-Probable Number (MPN) Method
The MPN method can be used to detect both coliforms and
E. coli. For freshwaters, a series of tubes should be inoculated
with 10-fold dilutions of the sample, and three to five tube
series are inoculated for each dilution. One major disadvant-
age of this method is the time it takes to obtain results: 48–72
h from inoculation to confirmation. This may present an
increased risk to public health as immediate actions cannot
be taken. Although E. coli is generally nonpathogenic, cer-
tain strains harboring genes encoding for Shiga toxins and
other virulence factors may represent a risk to public health;
thus, its rapid detection in freshwaters used for recreation and
consumption is vital. However, MPN is still the method of
choice when dealing with highly turbid samples (17).
qPCR Methods for the Detection of E. coli
Molecular methods, such as quantitative PCR (qPCR) are
relatively new. They are not yet broadly applied due to the
complexity of the analysis, potential for inhibition (or under-
estimation due to assay failure), availability of capable testing
facilities and cost. However, qPCR allows for rapid detection
and direct quantification of target organism (E. coli) DNA
without the need to wait for microbial growth to occur.
Thus, results can be generated in as little as 2–6 h after receipt
by the laboratory. There are two more common assays in use
for the quantification of E. coli by qPCR: (a) a proprietary
qPCR (Scorpion) assay developed by Noble et al. targeting
the uidA genes and utilizing lyophilized beads (Cepheid
OmniMix) or spray-coated cuvettes (BioGx SmartBeads)
containing all reagents and (b) a TaqMan assay developed
by the U.S. EPA (18, 19). qPCR assays for E. coli, at present,
have been primarily employed in the assessment of recrea-
tional surface waters, where beach action decision agreement
with enzyme-based assays (Colilert-18) approached 94%
(18). To demonstrate equivalent public health protection,
comparative analyses between culture and qPCR methods
may be required if the intent is to use the results generated
from qPCR assays for regulatory purposes. The city of Racine,
Wisconsin (USA) has used qPCR assays for E. coli to success-
fully regulate it on bathing beaches since 2012.
Enterococci as Indicators
Enterococci are Gram-positive bacteria that include members
of the genus Enterococcus, are able to grow at a wide temper-
ature range (10–45°C), alkaline pH (9.6), and high salinities
(6.5% NaCl). Enterococci have been accepted as indicators
of fecal pollution because of their presence in the feces of
warm-blooded animals and their prevalence in aquatic envi-
ronments similar to that of many bacterial pathogens (20).
Current regulatory limits for enterococci in recreational fresh-
waters are 33 CFU/100 ml. One of the major drawbacks of
enterococci as indicators of fecal pollution is that nonfecal
sources (e.g., runoff, sediments, and algae) also contribute
to the detection of these bacteria in temperate and tropical
freshwaters (11). In addition, enterococci have been detected
in pristine tropical waters as well as when pathogens are
absent (8, 21). Although several enterococcal species can
be used to infer fecal contamination, they may not accurately
assign source attribution. Another disadvantage of entero-
cocci as indicators of fecal contamination is that their detec-
tion often requires confirmation methods (22). Although
enterococci may be used to indicate freshwater quality, their
primary use is in marine waters.
Enzyme-Based Method for the Detection of
Enterococci
IDEXX Enterolert uses MUG that when catabolized produces
a blue fluorescence color, which can be viewed in UV light.
Results are obtained within 24 h, and no additional confir-
mation test is required (23).
Filtration Methods for the Detection of Enterococci
Culture media currently used for the detection of enterococci
in freshwaters and drinking waters rely on the inhibition
of Gram-negative bacteria using azide dextrose. For instance,
 3.1.1. Detection of Microbial Indicators in Environmental Freshwaters and Drinking Waters ▪ 3.1.1-3
m-Enterococcus (mE) agar possesses azide dextrose and TTC
as an indicator. This media, however, may ignore the ability
of other Gram-positive bacteria, such as staphylococci, to
reduce TTC (G. A. Toranzos, unpublished data). A modifica-
tion of this method has been approved by the U.S. EPA
(Method 1600). In Method 1600, enterococci are detected
by adding indoxyl β-D- glucoside (whose cleavage by
β-glucosidase results in a characteristic blue halo), nalidixic
acid (for the inhibition of background bacteria), and TTC
to mE agar base. Colonies with a blue halo are considered
to be enterococci. Percents of false positives and negatives
are usually relatively low (approximately 6%) (22), but as
with the coliform and E. coli methods, the time it takes to
obtain results (24 h) may not truly demonstrate the microbial
quality of waters at the time of the contamination. Thus, rapid
methods are being developed and tested, such as the qPCR
method, and the bacteriophage method, described below.
qPCR Methods for the Detection of Enterococci
The U.S. EPA included a qPCR assay for enterococci in a
suite of assays used to quantify microbial indicators of water
quality as part of the NEAAR epidemiological studies (24).
Health effect data indicated a positive relationship between
incidence in illness and increased concentrations of entero-
cocci as quantified by qPCR. The U.S. EPA assay (Method
1611), utilizes a TaqMan assay, similar to the E. coli assay as
previously described (25). Results are available within 4 h
of sample receipt by the laboratory. A modification of this
method, utilizing an environmental master mix reagent helps
reduce inhibition (Method 1609) (26). As with the qPCR
assay for E. coli, site-specific applicability should be ensured
prior to implementation. However, demonstration of equiva-
lent public health protection is unnecessary because of this
assay’s inclusion in epidemiological studies.
Bacteriophages as Indicators
Bacteriophages are present in the intestinal tract of warm-
blooded animals. Phages are viruses that infect bacteria and
are very specific in terms of the host they infect as well as
the source (i.e., those isolated from the feces of humans
have not been isolated from the feces of other animals)
(27). Because of this specificity, phages possess the potential
for use as indicators of fecal pollution. Other reasons for con-
sidering phages as indicators of fecal contamination include
the need to find models of human enteric viruses and a reli-
able method to assess the virological quality of waters. The
similar morphology, structure, and behavior of bacterio-
phages to that of many human enteric viruses suggest that
they should be more reliable indicators of the virological qual-
ity of water sources than indicator bacteria (28–30). One
major drawback of the bacteriophage method is the host
strains used for their detection are not necessarily the same
type strain and may vary depending on the laboratory.
Thus, for valid comparisons between sample types and geo-
graphical regions, the same bacterial strain should always be
used. However, the bacteriophage method is still the only
one showing viability in short periods of time.
Among the bacteriophages proposed as indicators of fecal
contamination, specifically that coming from a human
source, are those infecting Bacteroides fragilis (31–35).
B. fragilis phages are absent when enteric viruses are absent,
do not replicate under environmental conditions, seem to
be more resistant to various water treatments than human
enteric viruses, and have been proposed as a tool for viral
waterborne disease control (36). However, they have only
been detected in certain geographical regions of the United
States and Europe. In addition, given that the conditions
needed to grow B. fragilis are anaerobic, the detection of their
bacteriophages could be relatively difficult (37).
Other bacteriophage methods include the coliphage
method, which has been approved by the U.S. EPA (38).
Coliphages, or viruses that infect E. coli, are currently used
as an indicator of fecal contamination in groundwaters; their
presence in other water types also indicates fecal contamina-
tion. Coliphages are classified as somatic or F (male)-specific
(39–41). Somatic coliphages infect bacteria by attaching to
specific surface receptors. In contrast, F (male)-specific coli-
phages only infect “male” bacteria by recognizing receptors
in the pili. F (male)-specific coliphages possess genomes of
+ RNA or + DNA and belong to the Leviviridae and Inoviridae
families, respectively. F + RNA coliphages have been well
characterized as markers of fecal pollution and as a tool for
MST purposes, but less is known about F + DNA coliphages
(41, 42). The use of coliphages as a tool for MST purposes
may have several disadvantages. It has been suggested that
serotypes II and III of F + RNA coliphages could be used to
infer human fecal contamination (43, 44) (Table 1); how-
ever, other studies have found that the same serotypes can
also be found in animal feces (44). Other drawbacks of coli-
phages as viral indicators of fecal pollution is the fact that
their survival and inactivation rates under various environ-
mental conditions are, in most cases, not comparable to those
exhibited by human enteric viruses (41, 45).
Recently, phages that infect a specific type strain of
E. faecalis (enterophages) have been proposed as alternate
indicators of human fecal contamination in recreational
waters. The enterophage method is a relatively simple culture
technique, in which a water sample is mixed with an equal
amount of liquefied media, the bacterial host, calcium chlor-
ide (to facilitate the adsorption of the phages to the bacterial
host), and sodium azide (to inhibit the growth of background
bacteria). Results are obtained within 4–6 h after incubation.
Enterophages have been detected in domestic sewage and
environmental waters with point sources of human fecal pol-
lution. In addition, enterophages have exhibited survival and
inactivation rates similar to those of enteric viruses under sim-
ilar conditions (temperature, chlorine, and presence of envi-
ronmental microbiota) (46–48). Enterophages have been
mainly tested in tropical regions, and thus their prevalence
needs to be determined in other geographical areas. Disad-
vantages of enterophages may include their relative low num-
bers in sewage and environmental water samples. An
indicator of fecal pollution should be present in higher num-
bers than enteric pathogens, which usually exhibit very low
concentrations.
Other Indicators
Aerobic and facultative anaerobic bacteria have been tested
as indicators of fecal pollution in wastewater treatment
plants and include bacteria from the genera Acinetobacter,
TABLE 1 Host specificity of different groups of F + RNA
coliphages (modified from (36))
Group Host
I Nonhuman animals
II Humans and occasionally pigs
III Exclusively human
IV Nonhuman origin with rare human associations
3.1.1-4 ▪ WATER

Aeromonas, Alcaligenes, Citrobacter, Enterobacter, Flavobacte-
rium, Klebsiella, Moraxella, Mycobacteria, Proteus, Pseudomo-
nas, and Serratia. Among this group, Pseudomonas and
Aeromonas have received greater attention. Certain Pseudo-
monas spp. have been associated with infections associated
with the exposure to fecally contaminated surface waters
and ingestion of contaminated drinking water. Aeromonas
has been found associated with infection of open wounds
and problems related to the ingestion of waters contaminated
with fecal material (8). The method of choice for detection of
aerobic and facultative anaerobic bacteria is known as the
heterotrophic plate count (HPC), which detects viable bac-
teria by their ability to use organic compounds. Regulations
state that aerobic bacteria numbers in drinking water should
not exceed 500 organisms/ml, but concentrations may vary
according to conditions such as chlorine and temperature.
The HPC method cannot be used to identify the source of
the contamination (49, 50).
results for H2S are accompanied by the presence of enteric
pathogens. Similarly, negative results are accompanied by
the absence of enteric pathogens. A quantitative approach
of the method has been tested in drinking waters, in which
genera and species identification by terminal restriction frag-
ment length polymorphisms has been performed. Most of the
bacterial genera and species identified by this method corre-
spond to those having a fecal origin, thus representing a con-
cern to public health (54).
Legionella
Legionella spp., including L. pneumophila, can cause lung
infections as a result of inhalation of aerosols containing these
bacteria. Other Legionella spp. rarely cause infections, for
example L. anisa, which has been detected in hospital water
systems. However, its presence in water may indicate the pres-
ence of pathogenic Legionella spp. (55).

Clostridium perfringens IMPORTANCE OF MICROBIAL INDICATORS

C. perfringens may be considered a pathogen due to the pro-
duction of virulence factors such as hemolysins, collagenases,
and lipases. Their ability to form endospores makes them
resistant to removal and chlorination, and their presence
in wastewaters makes them an attractive indicator of the per-
formance of wastewater treatment systems. Media for the
detection of Clostridium include mCP, Tryptose Sulphite
Cycloserine (TSC), and CP ChromoSelect agar. One major
drawback of these media is the detection of other Clostridium
spp. (51, 52).
H2S Producers
Hydrogen sulfide (H2S) producers have been used to evaluate
the microbial quality of drinking waters. Also known as the
“paper strip method,” it is a simple and affordable field test
used to determine the presence of enteric bacteria able to
reduce organic sulfur and produce H2S (53). One disadvant-
age of the method is that sulfate-reducing bacteria from non-
fecal sources, even those isolated from pristine sources, can
produce H2S. However, studies focusing on this method for
the detection of enteric pathogens have found that positive
TO PUBLIC HEALTH
Regulations Pertaining to Microbial Indicators in
Freshwaters and Drinking Waters
Threshold exceedance values for currently used microbial
indicators have been determined based on health effects. Epi-
demiological studies determine the health effects associated
with the presence of microbial indicators of fecal pollution.
These studies aim to determine symptoms associated with gas-
trointestinal and respiratory illness and ear, eye, and skin
infections after contact with (e.g., swimming) or ingesting
fecally contaminated waters, as well as the correlation with
different levels of microbial indicators. Reports often involve
threshold values of indicator bacteria, symptoms associated
with the mentioned illnesses, and how variations in the
severity of the health effects correlate with the extent of the
fecal contamination (56). Regulations are based on symptoms
associated with the exposure and ingestion of fecally contami-
nated waters and are summarized in Table 2 (57–60).
Positive correlations are often observed between gastro-
intestinal illness and increased densities of enterococci,
thermotolerant coliforms, and E. coli. Notably, most of the

TABLE 2 Bacteriological drinking water and recreational freshwater standards or guidelines

Maximum no. of indicator organisms permitted per 100 ml of water type:
Total coliformsa Thermotolerant coliforms Turbidity
Guidelines established by: Enterococci (recreational)
c c
Drinking Recreational Drinking Recreational (NTUb)
World Health Organization 1–10 0 <1–5
Canada <10 0 200d 35e <1–5
European Economic 0 <10,000f 0–4
Community
United States 0 200g 0 <2,000f 1 (monthly)
a
In systems analyzing <40 samples per month, the maximum contaminant level specifies that no more than one sample per month may be total coliform
positive. In systems analyzing >40 samples per month, the maximum contaminant level specifies that no more than 5% of the monthly samples may be total
coliform positive (57).
b
NTU, nephelometric turbidity units.
c
Recreational refers to primary contact (swimming) waters.
d
Geometric mean of at least five samples (58) when experience has shown that greater than 90% of the thermotolerant coliforms are E. coli.
e
Geometric mean of at least five samples taken during a period not to exceed 30 days (59).
f
Compulsory limit. If exceeded in more than 20% of samples with at least 14 days of sampling, then bathing is prohibited (60).
g
This is a U.S. EPA criterion. Since no uniform national standards exists, it may vary from state to state.
Note: Although concentrations of indicators are still being used, new approaches are being implemented in which water quality may be determined based on a
range of concentrations and population densities in the watershed. The latter approach may be more useful as a tool for the protection of public health.
 3.1.1. Detection of Microbial Indicators in Environmental Freshwaters and Drinking Waters ▪ 3.1.1-5
thresholds are lower than current water quality guidelines
(61). Other studies have reported a correlation between
enterococci and skin illness (62), but these variations may
be due to differences in the experimental designs as well as
the indicators used, participants, and water body type.
Correlations of Indicators with Enteric Pathogens
Ideally the characterization of indicators of fecal pollution
should include correlation analyses with the prevalence of
enteric pathogens. Studies of this type are limited since not
many laboratories possess the facilities to detect enteric
pathogens. The presence of specific microbial indicators
may be positively correlated with the presence of bacterial
pathogens, as in the case of total coliforms and C. perfringens
with Salmonella, but in other cases, coliforms do not correlate
with enteric pathogens (63, 64). Possible reasons for this
include differences in the prevalence and survival times
and/or the ability of coliforms to become part of the environ-
mental microbiota of waters (3).
In terms of the enterococci, their prevalence has been
noted by some researchers to be positively correlated with
human enteric viruses (65). Other studies, however, have
found no correlation between enterococci, as detected by
molecular or culture methods, and pathogenic bacteria
(e.g. Bacteroides spp. with Campylobacter spp., pathogenic
bacteria that causes gastroenteritis) (66). Differences in the
correlation analyses between indicators and pathogens may
be due to the water types tested, as aquatic ecosystems may
be differently impacted by ecological factors (e.g., salinity,
dilution, transport, turbidity, and rainfall). Other influential
variables may include differences in the type and numbers
of indicators and pathogens (as the latter are often detected
in lower concentrations), pathogen source, sample size, and
statistical methods. Statistical methods, particularly, may
have a great influence on the results as these depend on the
sample size. Recent studies have suggested that discrepancies
between correlations between indicators and pathogens may
be due to the insufficient data for assessing these correlations
(67). It has also been suggested that microbial indicators
of fecal pollution may not necessarily be used to infer the
presence of pathogens; rather, that these are all statistical
probabilities of their co-occurrence (68).
Correlation of Indicators with Environmental Factors
Temperature can affect the detection of microbial indicators
in freshwaters. Temperate and subtropical regions are charac-
terized by marked changes in the prevalence of microbial
indicators, dependent on ambient seasonal temperature fluc-
tuations. The survival of microbial indicators is also influ-
enced by water temperature as longer survival times have
been reported at lower temperatures. Low temperatures in
the Northern Hemisphere (January–April and September–
December) are associated with high numbers of microbial
indicators, whereas higher temperatures (May–August) often
exhibit the opposite effect (69). This may also imply that
solar radiation influences the prevalence and survival of
enteric microorganisms in these surface waters, as solar radia-
tion is higher during the summer months (70). On the other
hand, in tropical regions where water temperatures are rela-
tively constant throughout the year, the numbers of microbial
indicators often exceed current standards. This has been
attributed to the ability of these bacteria to replicate in trop-
ical environments (71). Thus, careful consideration must be
taken when comparing microbial indicators across different
geographical regions.
Another important environmental factor that influences
the detection of microbial indicators is rainfall. One of the
main reasons for monitoring rainfall and the microbial quality
of surface waters is because of the possible resuspension of
pathogens from sediments to surface waters. The persistence
of enteric microorganisms tends to be greater in sediments
than surface waters; thus there is a greater health risk after pre-
cipitation events (72). U.S. outbreaks of waterborne diseases
have been preceded by rainfall (73). Yet the effect of rainfall
on the detection of enteric microorganisms may also depend
on the time of the fecal contamination event. Organisms that
persist longer in the environment may have a greater oppor-
tunity to attach to sediment particles and possibly be resus-
pended after rainfall events (47).
MST TOOLS
Library-Independent Methods
Different sets of tools can be applied when determining fecal
pollution sources and the use of redundant, or multiple, fecal
source tracking (FST) tools can often increase confidence in
the appropriate assignation of the source of pollution (74,
75). The simplest and most frequent employed FST tool is
the quantification of fecal indicator bacteria (FIB), such as
E. coli and enterococci, by defined substrate or agar-based
methods. While they are the most commonly employed,
C. perfringens has been examined as a secondary indicator
in some subtropical and tropical recreational waters (76)
(Watson Okubo, Hawaii Department of Health, personal
communication). Examining the association between fluctu-
ations in FIB concentrations in response to a predetermined,
site-specific set of meteorological, hydrological, or other envi-
ronmental triggers should be the first step in the development
of a robust study design which ultimately brings to bear, and
targets, chemical source tracking or MST tools (77). Caution
should be urged, however, when examining multiple primary
or secondary indicators as the concentrations of indicator
microbes do not necessarily correlate with one another, espe-
cially when they are measured by different methods (78).
With the advent of more rapid testing methods (such as
quantitative real-time PCR and qPCR) the time from sample
collection to result has been decreased from multiple days to
as little as a few hours (18, 25). Revised recreational water
quality criteria, published by the U.S. EPA in 2012, include
guideline values for enterococci by both culture-based assay
and qPCR (25). The implementation of rapid analytical
methods for the quantification of FIB is also under consider-
ation within the European Union (79). Real-time qPCR
assays for FIB target specific gene sequences, such as the
23S rRNA gene in Enterococcus spp. or uidA in E. coli (18,
25). Unique gene sequences in bacterial species allow for tar-
get specificity. The rapidity with which FIB concentrations
can be determined will provide an additional level of public
health protection as well as help in tracking pollution events
back to their original source. Once a likely source or sources
have been identified via FIB assessments, confirmation can
occur through the application of an ever growing toolbox of
FST tracking methods (80). Libraries of FIBs compiled from
local or regional host sources (sewage, avian, domesticated
and wild animals) have been successfully utilized as part of
a holistic FST tracking program in recreational and other sur-
face waters (81, 82).
However, the success of library-dependent methods is
defined by the size, representativeness, and geographic specif-
icity of the library (83). It may also require an extensive
3.1.1-6 ▪ WATER
amount of time to accumulate a sufficient source apportion-
ment database if previous efforts have not taken place.
Because of the limitations imposed by library-dependent
methodologies, other source attribution methods have been
explored.
Library-Independent Methods
As the name implies, library-independent methods do not
require the development of a host-specific library. Rather,
these methods rely on the detection and/or quantification
of a genetic marker or set of markers associated with source
specific pathogens or pathogen indicators. A genetic marker
is a specific location of a gene or segment of DNA that is
unique to the bacterial target. Library-independent methods
can be either presence/absence or quantitative in nature.
These can include phage typing (F+ and somatic coliphage),
gene-specific PCR (E. coli toxins), host-specific PCR (Bacter-
oides, Bifidobacteria, Enterococcus spp., enterovirus, and
adenovirus), and total community analysis (16S rRNA)
(84–88). Host-specific F+ RNA coliphage techniques have
been developed as a rapid and more cost-effective means of
identifying fecal pollution sources when monitoring the
microbiological quality of drinking, recreational, shellfish
harvesting, and other surface waters (84, 89). A study by
Lee et al. successfully employed this technique to determine
predominant fecal sources impacting ground and surface
water in South Korea (90). Somatic coliphages have also
been used in conjunction with C. perfringens to determine
the efficacy of virus removal in drinking water (91). Bacterio-
phages of common FIB and host-specific source tracking
markers (Bacteroides) have also shown promise as MST
tools (92).
F+ RNA and somatic coliphage techniques possess the
ability to distinguish human from animal sources, are easy
to perform, and do not require a reference library (38). How-
ever, environmental conditions (low numbers) and loss
of phage infectivity among subgroups after release into the
environment may lessen the success of this method for
source attribution (44). Some researchers have suggested
that reverse transcription PCR (RT-PCR) may provide a
more accurate determination of the in situ concentration of
F+-specific RNA coliphages due to its ability to detect low
copy numbers; facilitating more effective remediation strat-
egies for impacted environments (93, 94). RT-PCR is also
a technique commonly employed in the identification of
RNA viruses (95).
Host-specific Bacteroides primers, targeting the 16S rRNA
genes, have been developed that are capable of distinguishing
between human, ruminant, horse, and pig fecal pollution (96,
97). The use of host-specific Bacteroides markers has proven
useful in determining the origin of fecal pollution from a vari-
ety of aquatic environments (98, 99). Bacterial gene-specific
markers, such as the nifH marker of Methanobrevibacter smithii,
have been used in concert with Bacteroides HF183 (host-
specific) or viral markers such as adenoviruses or polyomavi-
ruses for accurate and sensitive detection of fecal pollution in
environmental waters samples in Australia (100). E. coli
toxin genes have also been used to characterize potential
human health risk from exposure to beach sands at Great
Lakes coastal recreational waters (101). Advantages of
gene- or host-specific PCR include the ability to isolate the
target from a complex (gut or external) environment without
prior culturing, sensitivity, and rapidity. However, these
methods only focus on a single gene or molecular target
from a single bacterial group that can be present in relatively
low numbers within the host or environment; regional
variations in specificity have also been observed (98).
Although total community analysis (a culture-independent
library MST method) is generally cumbersome, time-con-
suming, and expensive, it has been successfully used in the
construction and analysis of smaller clone libraries (<50
sequences) from environmental samples. It also can be used
to verify the specificity of primers (such as 16S rRNA Bacter-
oides specific primers) used in PCR assays or to verify the pres-
ence of host-specific bacteria environmental samples. Recent
advances in gene- or host-specific marker methods have
allowed for quantification (qPCR) and/or decreased the sam-
ple processing time (RT-PCR) from a half day or longer to a
few hours while still retaining the discriminatory ability, a dis-
tinct advantage over other slower FST methods (102, 103).
While rapidity is advantageous, a comprehensive under-
standing of the distinct differences between culture-based
assays and molecular methods is critical for successful imple-
mentation. Although culture-based assays (agar or defined
substrate) measure the ability of a microorganism to be culti-
vated and exhibit characteristic growth on selective media,
PCR assays (qualitative or quantitative) account for total cel-
lular DNA, which may be derived from live cells, dead cells,
or cells in a viable but not culturable state. Therefore, a
one-to-one relationship is unlikely to exist with the difference
in units of measure sometimes being several orders of magni-
tude different due to the number of target sequences per cell.
For example, the use of UV disinfection can also complicate
the detection of host-specific markers using methods such as
PCR, as the microorganisms are killed yet the DNA may per-
sist (104).
PCR methods in general, whether gene-specific, host-
specific or real time, can also be subject to inhibition or
underestimation. Inhibition of the qPCR assay can result
as a consequence of certain environmental constituents in
the sample matrix, for instance, humic acid (105). Reaction
inhibition results in assay failure, through either an underes-
timation of the target or complete lack of target detection
(106). The U.S. EPA has defined inhibition as a ≥3.0 cycle
threshold difference between the calibrator specimen process-
ing control (SPC) and the sample SPC (a known quantity of
nontarget DNA added to the assay) (25). Underestimation
of target numbers may also result from competition for the
substrate in environmental samples containing high amounts
of nontarget DNA (107). Despite inherent drawbacks associ-
ated with rapid molecular methods, their importance to FST
and the enumeration of FIBs from aqueous environments to
enhance the protection of public health cannot be refuted.
Current water quality guidelines are being revised to include
rapid molecular methods for the enumeration of FIB from
marine and fresh recreational waters, and it is likely that
they will be more broadly implemented to other water-based
applications in the foreseeable future given the good correla-
tion seen between these methods and culture-based assays
(108, 109).
CONCLUDING REMARKS
Bacterial indicators have successfully been used to protect
public health from fecally contaminated fresh surface waters
and drinking waters for the past 100 years, and variations
on the theme will be in use for decades to come. There are
a variety of new detection methods, and the advent of molec-
ular methods is facilitating the rapid detection of microbial
indicators. Although detection methods change, the target
indicator microorganisms do not. Therefore, their utility
may need to be revisited because very little is known about
 3.1.1. Detection of Microbial Indicators in Environmental Freshwaters and Drinking Waters ▪ 3.1.1-7
their ecology. It has been shown that FIB can survive for long
periods of time or even become part of the environmental
microbiota, thus their presence has to be placed within this
context. Although new indicators have been proposed,
namely, different types of bacterial viruses, they still have to
be tested and the water industry, as well as government agen-
cies responsible for the guidelines, do not seem to be com-
pletely convinced of the need for new indicators. A new
indicator has to be at least as good as the ones presently
used, and in fact better, especially indicators of enteric
viruses.
The best way to protect public health from enteric bacte-
rial pathogens and other diseases is to thoroughly know the
watershed and be ready to react to any possible changes in
water quality with rapidity. Regulatory agencies around the
world are now using a holistic approach that goes beyond
the detection of microbial indicators. For example, the hazard
assessment critical control points approach is much more real-
istic than focusing solely on the quantification of fecal indica-
tors. Although the latter is still part of the overall approach,
and this chapter has presented several methods for their
detection, contamination events and risk to public health
may be site-specific up to a point. Thus regulators, as well
as stakeholders, should have this in mind when designing
the best approach for the protection of public health in the
context of waterborne enteric pathogens.
Real-time methods, coupled with environmental assess-
ments, including the monitoring of rain events, extreme
weather, or other events for correlation the presence of fecal
contamination in waters, will enhance public health protec-
tion. In fact, public health will be better protected by a holis-
tic approach to water quality. This will be achieved through
the use of many different assessment tools, including bacte-
rial, chemical, and meteorological methods that will yield
information on the complete watershed. This approach may
allow stakeholders to be proactive rather than reacting to
the presence of indicators, as in the past.
REFERENCES
1. Dufour AP. 1984. Bacterial indicators of recreational water
quality. Can J Public Health 75:49–56.
2. Harwood VJ, Levine AD, Scott TM, Chivukula V, Luka-
sik J, Farrah SR, Rose JB. 2005. Validity of the indicator
organism paradigm for pathogen reduction in reclaimed
water and public health protection. Appl Environ Microbiol
71:3163–3170.
3. Savichtcheva O, Okabe S. 2006. Alternative indicators of
fecal pollution: relations with pathogens and conventional
indicators, current methodologies for direct pathogen mon-
itoring and future application perspectives. Water Res 40:
2463–2476.
4. Scott TM, Rose JB, Jenkins TM, Farrah SR, Lukasik J.
2002. Microbial source tracking: current methodology and
future directions. Appl Environ Microbiol 68:5796–5803.
5. Lechevallier MW. 2014. Conducting self-assessments
under the revised Total Coliform Rule. J Am Water Works
Assoc 106:90–102.
6. de Brauwere A, Ouattara NK, Servais P. 2014. Modeling
fecal indicator bacteria concentrations in natural surface
waters: a review. Crit Rev Environ Sci Technol 44:
2380–2453.
7. Golomidova A, Kulikov E, Isaeva A, Manykin A, Letarov
A. 2007. The diversity of coliphages and coliforms in horse
feces reveals a complex pattern of ecological interactions.
Appl Environ Microbiol 73:5975–5981.
8. Toranzos G, McFeters G, Borrego J, Savill M. 2007.
Detection of microorganisms in environmental fresh-
waters and drinking waters. In Manual of Environmen-
tal Microbiology, 3rd ed., ASM Press, Washington, DC,
pp. 249–264.
9. Beauchamp CJ, Simao-Beaunoir AM, Beaulieu C,
Chalifour FP. 2006. Confirmation of E. coli among other
thermotolerant coliform bacteria in paper mill effluents,
wood chips screening rejects and paper sludges. Water Res
40:2452–2462.
10. Rivera SC, Hazen TC, Toranzos GA. 1988. Isolation of
fecal coliforms from pristine sites in a tropical rain forest.
Appl Environ Microbiol 54:513–517.
11. Whitman RL, Shively DA, Pawlik H, Nevers MB, Byap-
panahalli MN. 2003. Occurrence of Escherichia coli and
enterococci in Cladophora (Chlorophyta) in nearshore water
and beach sand of Lake Michigan. Appl Environ Microbiol
69:4714–4719.
12. Fong TT, Lipp EK. 2005. Enteric viruses of humans and
animals in aquatic environments: health risks, detection,
and potential water quality assessment tools. Microbiol
Mol Biol Rev 69:357–371.
13. Edberg SC, Rice EW, Karlin RJ, Allen MJ. 2000. Escher-
ichia coli: the best biological drinking water indicator
for public health protection. Symp Ser Soc Appl
Microbiol:106S–116S.
14. Olstadt J, Schauer JJ, Standridge J, Kluender S. 2007. A
comparison of ten USEPA approved total coliform/E. coli
tests. J Water Health 5:267–282.
15. Dufour AP, Strickland ER, Cabelli VJ. 1981. Membrane
filter method for enumerating Escherichia coli. Appl Environ
Microbiol 41:1152–1158.
16. U.S. EPA. 2002. Method 1603: Escherichia coli (E. coli) in
water by membrane filtration using modified membrane-
thermotolerant Escherichia coli agar (modified mTEC). U.
S. Environmental Protection Agency, Office of Water,
Washington, DC.
17. Olson BH. 1978. Enchanced accuracy of coliform testing
in seawater by a modification of the most-probable-number
method. Appl Environ Microbiol 36:438–444.
18. Noble RT, Blackwood AD, Griffith JF, McGee CD,
Weisberg SB. 2010. Comparison of rapid quantitative
PCR-based and conventional culture-based methods for
enumeration of Enterococcus spp. and Escherichia coli in rec-
reational waters. Appl Environ Microbiol 76:7437–7443.
19. Chern EC, Siefring S, Paar J, Doolittle M, Haugland RA.
2011. Comparison of quantitative PCR assays for Escheri-
chia coli targeting ribosomal RNA and single copy genes.
Lett Appl Microbiol 52:298–306.
20. Wheeler AL, Hartel PG, Godfrey DG, Hill JL, Segars
WI. 2002. Potential of Enterococcus faecalis as a human
fecal indicator for microbial source tracking. J Environ
Qual 31:1286–1293.
21. Bonilla N, Santiago T, Marcos P, Urdaneta M, Domingo
JS, Toranzos GA. 2010. Enterophages, a group of phages
infecting Enterococcus faecalis, and their potential as alter-
nate indicators of human faecal contamination. Water Sci
Technol 61:293–300.
22. U.S. EPA. 2002. Method 1600: Enterococci in water by mem-
brane filtration using membrane-Enterococcus indoxyl-B-
d-glucoside agar (mEI). U.S. Environmental Protection
Agency, Washington, DC.
23. Chen C, Doherty K, Gu H, Dichter G, Naqui A. 1995.
Enterolert—A Rapid method for the detection of Entero-
coccus spp. Idexx Laboratories, Westbrook, ME.
24. U.S. EPA. 2012. Recreational water quality criteria. U.S.
Environmental Protection Agency, Washinton, DC.
25. U.S. EPA. 2012. Method 1611: Enterococci in water by
Taqman quantitative polymerase chain reaction (qPCR) assay.
3.1.1-8 ▪ WATER
U.S. Environmental Protection AgencyOffice of Water,
Washington, DC.
26. U.S. EPA. 2013. Acceptability of the EPA qPCR test at your
beach. U.S. Environmental Protection Agency, Office of
Water, Washington, DC.
27. Ogilvie LA, Caplin J, Dedi C, Diston D, Cheek E, Bowler
L, Taylor H, Ebdon J, Jones BV. 2012. Comparative
(meta)genomic analysis and ecological profiling of human
gut-specific bacteriophage phiB124–14. PLoS One 7:
e35053.
28. Yahya M, Hmaied F, Jebri S, Jofre J, Hamdi M. 2015.
Bacteriophages as indicators of human and animal faecal
contamination in raw and treated wastewaters from Tuni-
sia. J App Microbiol 118:1217–1225.
29. Havelaar AH, Nieuwstad TJ. 1985. Bacteriophages and
fecal bacteria as indicators of chlorination efficiency of
biologically treated waste-water. J Water Poll Cont Fed 57:
1084–1088.
30. Armon R, Kott Y. 1996. Bacteriophages as indicators of
pollution. Crit R EnvironSci Techn 26:299–335.
31. Tartera C, Lucena F, Jofre J. 1989. Human origin of Bac-
teroides fragilis bacteriophages present in the environment.
Appl Environ Microbiol 55:2696–2701.
32. Tartera C, Jofre J. 1987. Bacteriophages active against Bac-
teroides fragilis in sewage-polluted waters. Appl Environ
Microbiol 53:1632–1637.
33. Lucena F, Muniesa M, Puig A, Araujo R, Jofre J. 1995.
Simple concentration method for bacteriophages of Bacter-
oides fragilis in drinking water. J Virol Meth 54:121–130.
34. Jofre J, Olle E, Ribas F, Vidal A, Lucena F. 1995. Potential
usefulness of bacteriophages that infect Bacteroides fragilis
as model organisms for monitoring virus removal in drink-
ing water treatment plants. Appl Environ Microbiol 61:
3227–3231.
35. Gantzer C, Henny J, Schwartzbrod L. 2002. Bacteroides
fragilis and Escherichia coli bacteriophages in human faeces.
Int J Hyg Environ Health 205:325–328.
36. Ebdon JE, Sellwood J, Shore J, Taylor HD. 2012. Phages
of Bacteroides (GB-124): a novel tool for viral waterborne
disease control? Environ Sci Technol 46:1163–1169.
37. Payan A, Ebdon J, Taylor H, Gantzer C, Ottoson J, Papa-
georgiou GT, Blanch AR, Lucena F, Jofre J, Muniesa M.
2005. Method for isolation of Bacteroides bacteriophage
host strains suitable for tracking sources of fecal pollution
in water. Appl Environ Microbiol 71:5659–5662.
38. Salter RS, Durbin GW. 2012. Modified USEPA Method
1601 to Indicate Viral Contamination of Groundwater
(PDF). J Am Water Works Assoc 104:E480–E488.
39. Havelaar AH, Hogeboom WM. 1984. A method for the
enumeration of male-specific bacteriophages in sewage. J
Appl Bacteriol 56:439–447.
40. Hernández-Delgado E, Sierra M, Toranzos G. 1991. Col-
iphages as alternate indicators of fecal contamination in
tropical waters. Environ ToxicolWater Qual 6:131–143.
41. Gantzer C, Maul A, Audic JM, Schwartzbrod L. 1998.
Detection of infectious enteroviruses, enterovirus genomes,
somatic coliphages, and Bacteroides fragilis phages in
treated wastewater. Appl Environ Microbiol 64:4307–4312.
42. Long SC, El-Khoury SS, Oudejans SJ, Sobsey MD, Vinjé
J. 2005. Assessment of sources and diversity of male-spe-
cific coliphages for source tracking. Environ Engineer Sci
22:367–377.
43. Havelaar AH, Furuse K, Hogeboom WM. 1986. Bacterio-
phages and indicator bacteria in human and animal faeces.
J Appl Bacteriol 60:255–262.
44. Cole D, Long SC, Sobsey MD. 2003. Evaluation of F +
RNA and DNA coliphages as source-specific indicators
of fecal contamination in surface waters. Appl Environ
Microbiol 69:6507–6514.
45. Allwood PB, Malik YS, Maherchandani S, Vought K,
Johnson LA, Braymen C, Hedberg CW, Goyal SM.
2004. Occurrence of Escherichia coli, noroviruses, and F-
specific coliphages in fresh market-ready produce. J Food
Prot 67:2387–2390.
46. Santiago-Rodriguez TM, Davila C, Gonzalez J, Bonilla
N, Marcos P, Urdaneta M, Cadete M, Monteiro S, Santos
R, Domingo JS, Toranzos GA. 2010. Characterization of
Enterococcus faecalis-infecting phages (enterophages) as
markers of human fecal pollution in recreational waters.
Water Res 44:4716–4725.
47. Santiago-Rodriguez TM, Tremblay RL, Toledo-
Hernandez C, Gonzalez-Nieves JE, Ryu H, Santo Domi-
ngo JW, Toranzos GA. 2012. Microbial quality of tropical
inland waters and effects of rainfall events. Appl Environ
Microbiol 78:5160–5169.
48. Santiago-Rodriguez TM, Marcos P, Monteiro S, Urda-
neta M, Santos R, Toranzos GA. 2013. Evaluation of
Enterococcus-infecting phages as indices of fecal pollution.
J Water Health 11:51–63.
49. LeChevallier MW, McFeters GA. 1985. Interactions
between heterotrophic plate count bacteria and coliform
organisms. Appl Environ Microbiol 49:1338–1341.
50. Bartram J, Cotruvo J, Exner M, Fricker C, Glasmacher A.
2004. Heterotrophic plate count measurement in drinking
water safety management: report of an Expert Meeting
Geneva, 24–25 April 2002. Int J Food Microbiol 92:
241–247.
51. Bisson JW, Cabelli VJ. 1980. Clostridium perfringens as a
water pollution indicator. J Water Pollut Control Fed 52:
241–248.
52. Sorensen DL, Eberl SG, Dicksa RA. 1989. Clostridium
perfringens as a point source indicator in non-point polluted
streams. Water Res 23:191–197.
53. Manja KS, Maurya MS, Rao KM. 1982. A simple field test
for the detection of faecal pollution in drinking water. Bull
World Health Organ 60:797–801.
54. McMahan L, Grunden AM, Devine AA, Sobsey MD.
2012. Evaluation of a quantitative H2S MPN test for fecal
microbes analysis of water using biochemical and molecular
identification. Water Res 46:1693–1704.
55. van der Mee-Marquet N, Domelier AS, Arnault L,
Bloc D, Laudat P, Hartemann P, Quentin R. 2006.
Legionella anisa, a possible indicator of water conta-
mination by Legionella pneumophila. J Clin Microbiol 44:
56–59.
56. Pruss A. 1998. Review of epidemiological studies on
health effects from exposure to recreational water. Int J Epi-
demiol 27:1–9.
57. Bucklin KE, Mcfeters GA, Amirtharajah A. 1991. Pene-
tration of coliforms through municipal drinking-water fil-
ters. Water Res 25:1013–1017.
58. Canada HaW. 1992. Guidelines for Canadian drinking
water quality. Canadian Goverment Publishing Centre,
Ottawa, Canada.
59. Canada HaW. 1993. Guidelines for Canadian drinking
water quality. Canadian Government Publishing Centre,
Ottawa, Canada.
60. Rheinheimer G, Walker N. 1992. Aquatic Microbiology.
Wiley, New York.
61. Mujeriego R, Bravo J, Feliu M. 1982. Recreation in coas-
tal waters: public health implications. Vièmes J Etud
Poll:585–594.
62. Fleisher JM, Fleming LE, Solo-Gabriele HM, Kish JK,
Sinigalliano CD, Plano L, Elmir SM, Wang JD, Withum
K, Shibata T. 2010. The BEACHES Study: health effects
and exposures from non-point source microbial contami-
nants in subtropical recreational marine waters. Int J Epide-
miol 39:1291–1298.
 3.1.1. Detection of Microbial Indicators in Environmental Freshwaters and Drinking Waters ▪ 3.1.1-9
63. Morinigo MA, Córnax R, Muñoz MA, Romero P, Bor-
rego JJ. 1990. Relationships between Salmonella spp and
indicator microorganisms in polluted natural waters. Water
Res 24:117–120.
64. Polo F, Figueras MJ, Inza I, Sala J, Fleisher JM,
Guarro J. 1999. Prevalence of Salmonella serotypes in
environmental waters and their relationships with indica-
tor organisms. Antonie Van Leeuwenhoek 75:285–292.
65. Ottoson J, Stenstrom TA. 2003. Faecal contamination of
greywater and associated microbial risks. Water Res 37:
645–655.
66. Hellein KN, Battie C, Tauchman E, Lund D, Oyarzabal
OA, Lepo JE. 2011. Culture-based indicators of fecal con-
tamination and molecular microbial indicators rarely corre-
late with Campylobacter spp. in recreational waters. J Water
Health 9:695–707.
67. Wu J, Long SC, Das D, Dorner SM. 2011. Are microbial
indicators and pathogens correlated? A statistical analysis of
40 years of research. J Water Health 9:265–278.
68. Payment P, Locas A. 2011. Pathogens in water: value and
limits of correlation with microbial indicators. Ground
Water 49:4–11.
69. Lipp EK, Kurz R, Vincent R, Rodriguez-Palacios C, Far-
rah SR, Rose JB. 2001. The effects of seasonal variability
and weather on microbial fecal pollution and enteric
pathogens in a subtropical estuary. Estuaries 24:266–276.
70. Sinton L, Hall C, Braithwaite R. 2007. Sunlight inactiva-
tion of Campylobacter jejuni and Salmonella enterica, com-
pared with Escherichia coli, in seawater and river water.
J Water Health 5:357–365.
71. Bonilla TD, Nowosielski K, Cuvelier M, Hartz A, Green
M, Esiobu N, McCorquodale DS, Fleisher JM, Rogerson
A. 2007. Prevalence and distribution of fecal indicator
organisms in South Florida beach sand and preliminary
assessment of health effects associated with beach sand
exposure. Mar Pollut Bull 54:1472–1482.
72. Goyal SM, Gerba CP, Melnick JL. 1977. Occurrence and
distribution of bacterial indicators and pathogens in canal
communities along the Texas coast. Appl Environ Microbiol
34:139–149.
73. Shehane SD, Harwood VJ, Whitlock JE, Rose JB. 2005.
The influence of rainfall on the incidence of microbial fae-
cal indicators and the dominant sources of faecal pollution
in a Florida river. J Appl Microbiol 98:1127–1136.
74. Gourmelon M, Caprais MP, Mieszkin S, Marti R, Wery
N, Jarde E, Derrien M, Jadas-Hecart A, Communal PY,
Jaffrezic A, Pourcher AM. 2010. Development of micro-
bial and chemical MST tools to identify the origin of the
faecal pollution in bathing and shellfish harvesting waters
in France. Water Res 44:4812–4824.
75. Edge TA, Hill S. 2005. Occurrence of antibiotic resistance
in Escherichia coli from surface waters and fecal pollution
sources near Hamilton, Ontario. Can J Microbiol 51:
501–505.
76. Wade TJ, Sams E, Brenner KP, Haugland R, Chern E,
Beach M, Wymer L, Rankin CC, Love D, Li Q, Noble
R, Dufour AP. 2010. Rapidly measured indicators of rec-
reational water quality and swimming-associated illness at
marine beaches: a prospective cohort study. Environ Health
9:66.
77. Boehm AB, Fuhrman JA, Mrse RD, Grant SB. 2003.
Tiered approach for identification of a human fecal pollu-
tion source at a recreational beach: case study at Avalon
Bay, Catalina Island, California. Environ Sci Technol 37:
673–680.
78. Shibata T, Solo-Gabriele HM, Sinigalliano CD, Gidley
ML, Plano LR, Fleisher JM, Wang JD, Elmir SM, He
G, Wright ME, Abdelzaher AM, Ortega C, Wanless D,
Garza AC, Kish J, Scott T, Hollenbeck J, Backer LC,
Fleming LE. 2010. Evaluation of conventional and alter-
native monitoring methods for a recreational marine beach
with nonpoint source of fecal contamination. Environ Sci
Technol 44:8175–8181.
79. Oliver DM, Hanley ND, van Niekerk M, Kay D, Heath-
waite AL, Rabinovici SJM, Kinzelman JL, Fleming LE,
Porter J, Shaikh S, Fish R, Chilton S, Hewitt J, Connolly
E, Cummins A, Glenk K, McPhail C, McRory E, McVit-
tie A, Giles A, Roberts S, Simpson K, Tinch D, Thairs T,
Avery LM, Vinten AJA, Watts BD, Quilliam RS.
2015. Molecular tools for bathing water assessment in
Europe: Balancing social science research with a rapidly
developing environmental science evidence-base. Ambio:
1–11.
80. Giebel R, Worden C, Rust SM, Kleinheinz GT, Robbins
M, Sandrin TR. 2010. Microbial fingerprinting using
matrix-assisted laser desorption ionization time-of-flight
mass spectrometry (MALDI-TOF MS) applications and
challenges. Adv Appl Microbiol 71:149–184.
81. Kinzelman J, McLellan SL, Daniels AD, Cashin S, Singh
A, Gradus S, Bagley R. 2004. Non-point source pollution:
determination of replication versus persistence of Escheri-
chia coli in surface water and sediments with correlation of
levels to readily measurable environmental parameters.
J Water Health 2:103–114.
82. Olivas Y, Faulkner BR. 2008. Fecal source tracking by
antibiotic resistance analysis on a watershed exhibiting
low resistance. Environ Monit Assess 139:15–25.
83. Lu Z, Lapen D, Scott A, Dang A, Topp E. 2005. Identify-
ing host sources of fecal pollution: diversity of Escherichia
coli in confined dairy and swine production systems. Appl
Environ Microbiol 71:5992–5998.
84. Stewart-Pullaro J, Daugomah JW, Chestnut DE, Graves
DA, Sobsey MD, Scott GI. 2006. F + RNA coliphage typ-
ing for microbial source tracking in surface waters. J Appl
Microbiol 101:1015–1026.
85. Okabe S, Okayama N, Savichtcheva O, Ito T. 2007.
Quantification of host-specific Bacteroides-Prevotella 16S
rRNA genetic markers for assessment of fecal pollution in
freshwater. Appl Microbiol Biotechnol 74:890–901.
86. Oberste MS, Maher K, Williams AJ, Dybdahl-Sissoko N,
Brown BA, Gookin MS, Penaranda S, Mishrik N, Uddin
M, Pallansch MA. 2006. Species-specific RT-PCR ampli-
fication of human enteroviruses: a tool for rapid species
identification of uncharacterized enteroviruses. J Gen Virol
87:119–128.
87. Jiang S, Noble R, Chu W. 2001. Human adenoviruses and
coliphages in urban runoff-impacted coastal waters of
Southern California. Appl Environ Microbiol 67:179–184.
88. Zoetendal EG, Akkermans AD, De Vos WM. 1998. Tem-
perature gradient gel electrophoresis analysis of 16S rRNA
from human fecal samples reveals stable and host-specific
communities of active bacteria. Appl Environ Microbiol
64:3854–3859.
89. Love DC, Sobsey MD. 2007. Simple and rapid F+ coli-
phage culture, latex agglutination, and typing assay to
detect and source track fecal contamination. Appl Environ
Microbiol 73:4110–4118.
90. Lee JE, Lee H, Cho YH, Hur HG, Ko G. 2011. F+ RNA
coliphage-based microbial source tracking in water
resources of South Korea. Sci Total Environ 412–413:
127–131.
91. Payment P, Franco E. 1993. Clostridium perfringens and
somatic coliphages as indicators of the efficiency of drink-
ing water treatment for viruses and protozoan cysts. Appl
Environ Microbiol 59:2418–2424.
92. Gomez-Donate M, Payan A, Cortes I, Blanch AR,
Lucena F, Jofre J, Muniesa M. 2011. Isolation of bacterio-
phage host strains of Bacteroides species suitable for tracking
3.1.1-10 ▪ WATER
sources of animal faecal pollution in water. Environ Micro-
biol 13:1622–1631.
93. Kirs M, Smith DC. 2007. Multiplex quantitative real-time
reverse transcriptase PCR for F+-specific RNA coliphages:
a method for use in microbial source tracking. Appl Environ
Microbiol 73:808–814.
94. Smith DC. 2006. Microbial source tracking using F-specific
coliphages and quantitative PCR. Doctoral dissertation,
University of Rhode Island.
95. Griffin DW, Gibson CJ III, Lipp EK, Riley K, Paul JH
III, Rose JB. 1999. Detection of viral pathogens by reverse
transcriptase PCR and of microbial indicators by standard
methods in the canals of the Florida Keys. Appl Environ
Microbiol 65:4118–4125.
96. Bernhard AE, Field KG. 2000. A PCR assay to discrimi-
nate human and ruminant feces on the basis of host differ-
ences in Bacteroides-Prevotella genes encoding 16S rRNA.
Appl Environ Microbiol 66:4571–4574.
97. Harmsen HJ, Wildeboer-Veloo AC, Raangs GC,
Wagendorp AA, Klijn N, Bindels JG, Welling GW.
2000. Analysis of intestinal flora development in breast-fed
and formula-fed infants by using molecular identification
and detection methods. J Pediatr Gastroenterol Nutr 30:
61–67.
98. Gawler AH, Beecher JE, Brandao J, Carroll NM, Falcao
L, Gourmelon M, Masterson B, Nunes B, Porter J, Rince
A, Rodrigues R, Thorp M, Walters JM, Meijer WG.
2007. Validation of host-specific Bacteriodales 16S rRNA
genes as markers to determine the origin of faecal pollution
in Atlantic Rim countries of the European Union. Water
Res 41:3780–3784.
99. Lee DY, Weir SC, Lee H, Trevors JT. 2010. Quantitative
identification of fecal water pollution sources by TaqMan
real-time PCR assays using Bacteroidales 16S rRNA genetic
markers. Appl Microbiol Biotechnol 88:1373–1383.
100. Ahmed W, Sidhu JP, Toze S. 2012. Evaluation of the nifH
gene marker of Methanobrevibacter smithii for the detection
of sewage pollution in environmental waters in Southeast
Queensland, Australia. Environ Sci Technol 46:543–550.
101. Bauer L, Alm E. 2012. Escherichia coli toxin and attach-
ment genes in sand at Great Lakes recreational beaches.
J Great Lakes Res 38:129–133.
102. Dick LK, Field KG. 2004. Rapid estimation of numbers of
fecal Bacteroidetes by use of a quantitative PCR assay for 16S
rRNA genes. Appl Environ Microbiol 70:5695–5697.
103. Green HC, Dick LK, Gilpin B, Samadpour M, Field KG.
2012. Genetic markers for rapid PCR-based identification
of gull, Canada goose, duck, and chicken fecal contamina-
tion in water. Appl Environ Microbiol 78:503–510.
104. Bae S, Wuertz S. 2009. Rapid decay of host-specific fecal
Bacteroidales cells in seawater as measured by quantita-
tive PCR with propidium monoazide. Water Res 43:
4850–4859.
105. Opel KL, Chung D, McCord BR. 2010. A study of PCR
inhibition mechanisms using real time PCR. J Forensic Sci
55:25–33.
106. Shanks OC, Sivaganesan M, Peed L, Kelty CA, Black-
wood AD, Greene MR, Noble RT, Bushon RN, Stelzer
EA, Kinzelman J, Anan’eva T, Sinigalliano C, Wanless
D, Griffith J, Cao Y, Weisberg S, Harwood VJ, Staley
C, Oshima KH, Varma M, Haugland RA. 2012. Interla-
boratory comparison of real-time PCR protocols for quan-
tification of general fecal indicator bacteria. Environ Sci
Technol 46:945–953.
107. Hata A, Katayama H, Kitajima M, Visvanathan C, Nol C,
Furumai H. 2011. Validation of internal controls for
extraction and amplification of nucleic acids from enteric
viruses in water samples. Appl Environ Microbiol 77:
4336–4343.
108. Kinzelman JL, Bushon RN, Dorevitch S, Noble RT.
2011. Comparative evaluation of molecular and culture
methods for fecal indicator bacteria for use in inland recrea-
tional waters. International Water Association. WERF
Report PATH7R09.
109. Srinivasan S, Aslan A, Xagoraraki I, Alocilja E, Rose JB.
2011. Escherichia coli, enterococci, and Bacteroides
thetaiotaomicron qPCR signals through wastewater and
septage treatment. Water Res 45:2561–2572.
 Best Practices for Cyanobacterial Harmful Algal
Bloom Monitoring
TIMOTHY G. OTTEN AND HANS W. PAERL
3.1.2
The phylum Cyanobacteria is comprised of some of the
Earth’s oldest microorganisms, estimated to have first
appeared approximately 2.7 billion years ago (1). This group
is highly diverse in both appearance and function, although
all are primary producing (autotrophic) bacteria that obtain
their energy via oxygenic photosynthesis; they also fix inor-
−
ganic carbon (CO2/CO2– 3 /HCO3 ) into complex organic car-
bon that fuels the base of the microbial food web. Their
oxygenic photosynthetic activity early in the Earth’s history
is credited with paving the way for higher life forms to evolve
in what is commonly referred to as the Great Oxidation Event
(2). In spite of these beneficial attributes, cyanobacteria may
represent a significant threat to aquatic ecosystems and
human health. Under the right conditions, such as warm,
stratified, stagnant, and nutrient-rich waters, cyanobacterial
cells can accumulate into dense, paint-like scums that float
near the surface (3) (Fig. 1). When these bloom assemblages
die off, they represent a large sink of organic carbon that is
subject to microbial remineralization that may lead to large,
localized areas that are devoid of dissolved oxygen (hypoxia),
a process that often culminates in fish kills.
In addition to their ecological impacts, most cyanobacte-
rial genomes contain a variety of nonribosomal peptide
synthetases and polyketide synthases that are involved in
the production of myriad secondary metabolites, many of
which exhibit toxin-like characteristics (4). This chapter
focuses on the four major recognized classes of cyanotoxins
(anatoxin-a, cylindrospermopsin, microcystin/nodularin, and
saxitoxin); however, the reader should be aware that other
classes of compounds exist, such as aeruginosin, anabaeno-
peptin, cyanobactin, cyanopeptolin, and microviridin, and
these and other presently uncharacterized peptides may war-
rant inclusion into future algal bloom management plans as
we learn more about them (5). Cyanobacterial harmful algal
blooms (CyanoHABs) are considered a public health threat
for animals or humans that use impaired surface waters for
drinking, irrigation, or recreation. Pets and livestock are dis-
proportionately affected by algal toxins because they tend
to drink directly from contaminated waters. Every year there
are dozens of animal deaths related to acute intoxication from
cyanobacteria-impacted waters. Historically, veterinarians
have only been able to provide supportive therapy (e.g., fluids
and nutrients) for sickened animals. However, there has been
at least one clinical case that identified a potential therapy for
minimizing liver damage from acute microcystin intoxication
doi:10.1128/9781555818821.ch3.1.2
3.1.2-1
by administering cholestyramine, which efficiently binds
with microcystin within the gastrointestinal lumen, thereby
preventing enterohepatic recirculation and liver damage (6).
APPROACHES TO CYANOHAB MONITORING
Although cyanotoxins are widely acknowledged as a public
health risk, there are no formal regulations in place in the
United States that mandate CyanoHAB monitoring or man-
agement of impaired waters. The lack of broad federal over-
sight—such as under the umbrella of the Safe Drinking
Water or Clean Water Acts—has led some individual states
to craft their own CyanoHAB management programs,
although less than half (n = 21) of U.S. states had programs
in place as of 2014. One immediate problem with these state
initiatives is that they lack regulatory enforcement. Instead,
participation in CyanoHAB monitoring is noncompulsory,
which has led to widespread underreporting of bloom events.
Of the U.S. states that have chosen to enact management
plans, most have adopted frameworks similar to those pro-
posed by the World Health Organization’s guidelines for
managing toxic cyanobacteria in water (7) (Fig. 2). Note
that the actual recreational guidance values presented here
have been updated from the original version to better reflect
our current understanding of cyanotoxins and their public
health risks.
A beneficial aspect of an adaptive management approach
is that it provides water quality managers with tiered options
best suited to their needs and financial constraints. Under this
framework, the least cumbersome option is to simply post a
lake advisory if algal accumulations become visible within
the water column or at the surface. This course of action
may be appropriate if the lake is infrequently used or if finan-
cial constraints preclude additional analysis. However, most
water managers will strive to minimize unnecessary economic
losses associated with lake closings (e.g., reduced tourism),
especially if cyanobacterial abundances or toxin concentra-
tions are below the defined action level. One approach to
CyanoHAB monitoring is to simply assay for the entire suite
of cyanotoxins from each sample collected, although this may
be cost prohibitive depending on the quantity and frequency
of samples collected. However, because not all phytoplank-
ton blooms will be composed of cyanobacteria—eukaryotic
phytoplankton such as diatoms and green algae often bloom
in the spring and winter in temperate lakes—it makes sense to
3.1.2-2 ▪ WATER

FIGURE 1 Photographs of cyanobacterial blooms from around the world: (a) Microcystis bloom in Lake Taihu, China; (b) Microcystis bloom
in Lake Erie, USA; (c) Lyngbya bloom in Lake Atitlán, Guatamala; (d) Anabaena, Microcystis, and Nodularia in the Baltic Sea–Gulf of Finland;
(e) Aphanizomenon bloom in Lake Dianchi, China; (f and g) Microcystis blooms in Lake Taihu, China; (h) unidentified cyanobacterial bloom in
Taivallahti Bay (Baltic Sea), Finland; (i) Microcystis bloom in the Neuse River, NC (USA); ( j) unidentified cyanobacterial bloom in St. John’s
River, FL (USA); (k) Nodularia bloom in Gulf of Finland (Baltic Sea); (l) Trichodesmium bloom in Sanibel Inlet (Gulf of Mexico) FL (USA).
Adapted from (3). doi: 10.1128/9781555818821.ch3.1.2.f1

first determine if CyanoHAB taxa are present. This can be
done by viewing/counting the sample under a microscope
or by using genetic assays such as PCR or quantitative PCR
(qPCR) to identify/quantify total or potentially toxigenic
cyanobacteria. When potentially toxigenic cyanobacterial
genera are present at concentrations greater than 40,000
cells/ml or 5,000 toxin gene equivalents/ml, then these sam-
ples are candidates for toxin screening since they are likely to
contain measurable amounts of cyanotoxin. The benefits of
this approach include: (a) cost savings by not performing
expensive cyanotoxin analyses when these compounds are
unlikely to be present, (b) the use of two independent assays
 3.1.2. Best Practices for Cyanobacterial Harmful Algal Bloom Monitoring ▪ 3.1.2-3

FIGURE 2 Recommended guidelines for estimating public health risks from recreational exposure to cyanobacteria-impacted waters. Once
an advisory is issued, samples should be collected weekly until 2 weeks have passed and both the cell concentrations or qPCR gene equivalents
and toxin concentrations are below the action level. Potentially toxigenic cyanobacteria and their relevant toxins are listed in Table 1.
doi:10.1128/9781555818821.ch3.1.2.f2

(microscopy or qPCR + toxin screening) provides confirma-
tion of one another and reduces the likelihood of false posi-
tives, and (c) when paired specifically with qPCR, sample
throughput can be increased, allowing for rapid risk assess-
ment of large quantities of samples which will improve risk
characterization efforts.
Table 1 provides an overview of cyanobacterial genera that
have been demonstrated to be potential toxin producers in at
least some ecosystems and provides physical and ecological
characteristics useful in their identification. Before going
any further, it is fundamentally important to convey that
no single species or genus is definitively toxic or nontoxic;
instead, the genes involved in toxin biosynthesis are widely
distributed across most cyanobacterial taxa. Therefore toxin
production is essentially a strain-level trait that varies over
time, space, and environmental condition. Table 1 provides
a list of the most commonly encountered freshwater cyano-
bacterial taxa that may also produce toxins. It remains unclear
if some other commonly observed genera such as Gloeotrichia
or Woronochinia are capable of toxin production, however,
these cells should be included in the total cyanobacteria
counts nonetheless. While toxin gene estimates by qPCR
tend to strongly correlate with toxin concentrations and rep-
resent a more sensitive early warning method for detection of
potential toxin producers than cell counting, it does not
definitively convey if toxins are being synthesized. Therefore,
toxin analyses remain the “gold standard” for enacting and
removing public health advisories, whereas qPCR and micro-
scopy are highly useful for informing which samples to per-
form toxin screening on and for advancing our ecological
and physiological understanding of a system. If the toxin con-
centrations are below the action level, then no lake advisory
needs to be issued—even if the population exceeds the cell
abundance or qPCR gene equivalent thresholds (see Fig.
2). In the event that an alert threshold is passed, a lake advi-
sory should be posted and samples should be analyzed weekly
until 2 weeks have passed where the cell concentrations or
qPCR gene equivalents and toxin concentrations are below
the action limits. Note that both cells/genes and their rele-
vant toxins should be below the action limits before lifting
an advisory because cell lysis could produce a scenario where
few cyanobacterial cells are present in the water column, even
though dissolved toxin levels may remain high due to their
long half-lives, which range from hours to weeks (8–11).
It is important to note that individual states do not always
choose to follow the exact risk approximations set forth by the
WHO guidelines; this is in part because the original guide-
lines failed to estimate toxin quotas for anything except
microcystin. Of which, the WHO guidelines assumed a worst-
case scenario of 0.2 pg microcystin-LR equivalent per cell—
which equates to 4 µg/liter per 20,000 cells (low risk) or 20
µg/liter per 100,000 cells (high risk) if a bloom was composed
of 100% toxic cells. Microcystin risks are interpreted differ-
ently from state to state; for example, Oregon uses 10 µg/liter
3.1.2-4 ▪ WATER

TABLE 1 Most commonly observed toxin-producing cyanobacterial genera in North American lakes and their physical and ecological
characteristics

Genus Potential toxins Characteristics

Anabaena (Dolichospermum) anatoxin-a, homoanatoxin-a, anatoxin-a(S), cylindrospermopsin, microcystin, saxitoxin F, C, H, A, B/P
Aphanizomenon anatoxin-a, homoanatoxin-a, cylindrospermopsin, microcystin, saxitoxin F, C, H, A, P
Cylindrospermopsis anatoxin-a, homoanatoxin-a, cylindrospermopsin, microcystin, saxitoxin F, H, A, P
Lyngbya anatoxin-a, homoanatoxin-a, lyngbyatoxin, saxitoxin F, S, B/P
Microcystis microcystin U, C, P
Nostoc microcystin F, H, A, O, P
Nodularia nodularin F, H, A, B/P
Oscillatoria anatoxin-a, homoanatoxin-a, cylindrospermopsin, microcystin, saxitoxin F, P
Planktothrix anatoxin-a, homoanatoxin-a, microcystin, saxitoxin F, P
A = akinetes, B = benthic, C = colonial, F = filamentous, H = heterocystous, O = hormogonia, P = pelagic, S = sheathed, U = unicellular.

microcystin for recreational waters as its action threshold,
whereas California only permits 0.8 µg/liter and Washington
allows 6 µg/liter. Even larger variation exists among these
states when comparing action thresholds for other cyanotox-
ins; these values range from 20 to 90 µg/liter for anatoxin-a; 4
to 6 µg/liter for cylindrospermopsin; and 10 to 75 µg/liter for
saxitoxin. Regarding finished drinking water, guideline val-
ues have been proposed that are meant to be protective
over a lifetime of exposure; they are as follows: 6 µg/liter
anatoxin-a, 2 µg/liter homoanatoxin-a, 1 µg/liter anatoxin-a
(S), 1 µg/liter cylindrospermopsin, 1 µg/liter microcystin-LR,
1 µg/liter nodularin, and 3 µg/liter saxitoxin (12). Note that
all of these guidance values may be subject to change as we
learn more about these compounds and the diverse secondary
metabolites produced by cyanobacteria.
SAMPLING CONSIDERATIONS
Several characteristics exhibited by cyanobacteria make accu-
rate surveillance of these populations challenging. As cell
concentrations grow, the size of colony assemblages in turn
increases. Larger colonies have higher buoyancy owing to
internal gas vesicles that enable them to maintain a position
near the air–water interface (13). However, their position
within the water column is not fixed. During daylight hours,
most cyanobacterial cells will accumulate near the water’s sur-
face; however, following a period of active photosynthesis the
cells accumulate carbohydrates, which increases their density
until they sink, resulting in a characteristic vertical migration
pattern at night. The distance that cells travel is dependent on
colony size and shape, but a general estimate would be on the
order of up to 5 m depth overnight for many taxa. This verti-
cal migration pattern allows the cells to access dissolved
nutrients that are often only present in limited quantities
near the water’s surface but often abundant below the ther-
mocline. Throughout the night, the cellular density decreases
due to respiration and carbohydrate utilization; until eventu-
ally their buoyancy potential exceeds their weight and they
float back to the surface (14, 15). This pattern has important
implications for developing a sampling strategy. If samples are
collected early in the morning, or late in the evening, a large
quantity of cells may still be located below the surface,
whereas during peak daylight most will be within the top
0.5 m on a calm day. As such, it is recommended that samples
should consistently be collected at approximately the same
time of day to facilitate meaningful comparisons over time.
The first step in designing a CyanoHAB monitoring plan
is to determine the desired outcomes. For example, public
health sampling approaches are aimed at estimating the max-
imum possible risks associated with direct contact with a con-
centrated scum. As such, these samples are collected along
shorelines where algal accumulations are greatest due to phys-
ical factors such as wave or wind forcing. Analyses of these
samples may reveal extraordinarily high cell and toxin con-
centrations that are not necessarily representative of the
entire water body’s condition. Instead, these values are used
to determine public health risks and are applied to the frame-
work outlined in Fig. 2, which indicates when a lake advisory
should be posted. However, if the goal is to better understand
cyanobacterial bloom dynamics, then environmental samples
should be collected away from the shore to minimize physical
accumulation effects. It is not uncommon to collect one sam-
ple near the surface (0.5 m depth) and one that is
depth-integrated (e.g., 0–8 m) by using a van Dorn sampler
or simply a rubber hose with a weighted end or a PVC pipe
that can be lowered vertically through the water column
and plugged on one end to create a vacuum seal. The tube
can then be pulled up and dispensed into a bucket or other
collection vessel. These types of environmental samples
will better represent actual conditions and are particularly
valuable over longer periods of time to determine temporal
and spatial patterns, cell growth rates, and responses to chang-
ing environmental conditions. This approach can also be
used to help establish nutrient management plans, such as
total maximum daily loads of nitrogen and phosphorus. If
the samples are meant to characterize possible health risks
in drinking water sources, they should come from the raw
water collected from the reservoir intake pipe prior to disin-
fection, once after flocculation/sedimentation, and then
once again from the finished drinking water.
SAMPLE COLLECTION AND STORAGE
The first task is to identify where sampling will take place.
Some knowledge about the water body’s general usage is
required to determine where exposures are most likely to
occur. Commonly sampled areas for recreational monitoring
include boat launches, swimming beaches, or alongside
docks. Drinking water samples should be collected near the
influent intake pipe; one at the depth of the pipe and another
that is depth-integrated at the same location. In general, a
sample volume of 0.5–1 liter will suffice for most downstream
applications. The type of collection vessel is not of critical
importance, but amber glass is the most inert and least prone
to adsorb extracellular toxins, although breakage concerns
during transport may ultimately favor the use of plastic
 3.1.2. Best Practices for Cyanobacterial Harmful Algal Bloom Monitoring ▪ 3.1.2-5
bottles—in which case, HDPE type 2 plastic is preferred. If
polyethylene bottles are used for collection, they should first
be triple rinsed with lake water due to the potential for toxin
sorption. Individuals should also take caution when collect-
ing samples to not come into direct contact with an algal
scum since many are considered to contain dermal toxins
that cause skin irritation and rashes; at a minimum, nitrile/
latex gloves should be worn to reduce skin exposure risk. If
direct contact does occur, the affected area should be washed
thoroughly with soap and water. Subsurface samples can be
collected by placing the unopened container fully beneath
the surface of the water, then removing the cap and filling
completely. Samples should be stored immediately in the
dark on wet ice, but never allowed to freeze, which will result
in cell lysis and toxin release as well as making microscopic
identification much more difficult. Once samples are returned
to the laboratory, the contents of the bottle should be thor-
oughly mixed prior to aliquoting it into subsamples for later
analyses. Ideally a churn splitter will be used to ensure even
mixing; however, simply inverting the bottle several times,
then quickly pouring out the desired volume will suffice. In
general a portion of each sample (e.g., 100 ml) should be pre-
served using Lugol’s iodine solution at a final concentration of
1% vol/vol. For visual reference, this mixture should appear
“whiskey” colored. These samples can then be stored for
several months at room temperature if kept out of direct sun-
light and can be used for microscopic identification and
enumeration.
If toxin analysis is desired, samples should be stored
according to the analytical method that will be used. For
enzyme-linked immunosorbent assays (ELISA), 1 ml aliquots
of the sample can be stored in microcentrifuge tubes at −20°C
or −80°C for up to 1 year. This type of sample can only be
used to quantify total toxin (extracellular + intracellular),
because upon thawing significant cell lysis will occur. Quan-
tification of total toxin is the most common approach for
public health monitoring. However, there may be times the
investigator is interested in quantifying only the intracellular
or extracellular fraction of the toxin pool. In this case the
samples need to be passed through a filter to collect the cells
(intracellular toxin fraction) or the filtrate (extracellular
toxin fraction). Filtering the sample has the additional bene-
fit of concentrating the cells that may be required if cyanobac-
terial abundance is low or if toxin concentrations are below a
given assay’s limit of detection. The cells that are collected on
the filters can then have their toxins extracted at a later date if
desired. There are many filter materials to choose from (e.g.,
glass microfiber or polyethersulfone membrane) although any
with pore size diameters less than 2 µm should capture all
potential toxin-producing cyanobacterial cells. This
approach may not capture pico-planktonic cyanobacteria
(<2 µm diameter), although this group is not presently
known to represent a toxin risk. If desired, membrane filters
with pore sizes of 0.2 µm are available and can be used to cap-
ture all cyanobacteria within a sample.
Samples should also be filtered if genomic analyses are
desired, since cell-enclosed DNA, RNA, or protein can be
extracted from these filters. For all purposes, it is important
to record the volume of sample passed through the filters to
relate the observed values back to the original environmental
concentration. To pass the sample through the filter, a vac-
uum manifold is most commonly employed, although hand
pumps may be more accessible if processing samples in the
field. Incorporation of a relief value is recommended during
filtration to reduce the likelihood of lysing cells due to high
suction pressure, which will result in loss of their cellular
contents (e.g., nucleic acids and toxins). Following filtration,
the filters can be rolled into a cone using clean forceps and
placed into microcentrifuge tubes or folded in half and
wrapped in aluminum foil. Storage at −80°C is recom-
mended, although the samples should be stable at −20°C
for up to 1 year. Best practice QA/QC is to use clean and auto-
claved filter rigs for each sample. If an autoclave is unavail-
able, the filter rigs should be thoroughly cleaned in a dilute
bleach solution (5% vol/vol) and rinsed in dH2O for each dis-
crete sample filtered. Additionally, a negative control consist-
ing of sterile dH2O processed similarly as the samples should
be processed each time environmental samples are filtered.
Note that if nucleic acid extraction is immediately desired, fil-
tration can be replaced by a high speed centrifugation step
(∼5,000 × g for 5 min) to concentrate the cells. However,
due to the buoyancy exerted by gas vesicles within many of
these organisms, it will be difficult to achieve a satisfactory
amount of pelleting without first exposing the sample briefly
(1–2 s) to a pressure of ∼150 psi, which should burst most cya-
nobacterial gas vesicles without causing cell lysis (16).
MICROSCOPIC IDENTIFICATION AND
ENUMERATION
The most frequently used method for cyanobacterial identifi-
cation and enumeration is the Utermöhl sedimentation tech-
nique (17). This approach requires an inverted microscope
capable of at least 100 × magnification and a counting cham-
ber with a settling cylinder attached to it. Depending on the
cylinder size, 2–10 ml of Lugol’s preserved sample is dispensed
into the cylinder and allowed to settle; typically sedimenta-
tion will take approximately 4 h per cm of height of the sam-
ple. Before dispensing the preserved sample into the settling
chamber, it should be gently inverted to evenly mix the phy-
toplankton. After settling, the top contents of the chamber
can be carefully discarded using a serological pipette, and
then the phytoplankton can be enumerated by counting
the total number of organisms (if sample has low abundance)
or by counting a subsection of the field and then multiplying
by the appropriate correction factor. Since natural assemb-
lages of cyanobacteria tend to aggregate into colonies, accu-
rate quantification is difficult. There are two approaches
that can be utilized to improve counting accuracy; one
involves breaking up the colonies and counting the individ-
ual cells and the other simply estimates cell abundance based
on colony biovolumes.
The first approach is to disintegrate the mucilaginous
sheath surrounding the colonies so that single cells or fila-
ments can be counted. There is a variety of ways to do this:
hydrolysis in 0.1 M NaOH, incubation at 80°C for 15 min
followed by 30 s vortex mixing and/or sonication at 20 kHz
for 30 s can all be helpful in disaggregating cyanobacterial
assemblages (18). However, any method that is employed
should be checked to ensure it results in little to no cell lysis.
It will take some experience, but to improve counting accu-
racy the number of phytoplankton cells added to the counting
chamber should not be overconcentrated; which is easily
done with bloom or public health samples. Approximately
100 phytoplankton units per counting area is ideal to mini-
mize counting error. If the sample contains a very high con-
centration of cells it should be diluted 1:10 or greater with
1% Lugols vol/vol in distilled H2O prior to settling and count-
ing. When counting filamentous cyanobacteria, the number
of cells comprising an individual filament (excluding special-
ized cells such as akinetes and heterocysts) can be counted for
the first 30 filaments observed, then averaged to produce an
3.1.2-6 ▪ WATER
estimate of mean cells per filament. This will expedite the
counting process going forward, since only the filaments
will need to be counted. These values can be easily converted
into cells/ml by multiplying by the estimated factor. Counting
error across replicates can be calculated by the equation: error
(±%) = 100 × sq.root(2/n); where n is the total number of
units (replicates, grid squares, etc.) that were counted (19).
For public health monitoring there should never be greater
than ±30% error across replicates.
The other counting method commonly utilized is to meas-
ure biovolumes of cyanobacterial genera within a sample. To
do this, the specific taxa endemic to the water body of interest
needs to first be characterized based on its colonial shape and
individual cell dimensions. Thus, a stage micrometer is
required to make accurate measurements of length. For exam-
ple, the most common colony geometries observed in cyano-
bacteria and their volumetric dimensions are: sphere (V = 4/
3 × π × r 3), prolate/ovaloid sphere (V = π/6 × d 2 × h) and cyl-
inder (V = π × r 2 × h), although other shapes may be observed
(20). Note that these same equations are also used to deter-
mine individual cell or filament volumes. Biovolumes will
need to be calculated for every new set of samples due to
cell and colony size variations that change for many reasons,
including cell physiological status, local environment, and
biotic interactions (e.g., grazing pressure). Combined, these
two measurements of colony volume and individual cell vol-
ume can be used to come up with an estimate of cells per fluid
volume. As an example, a spherical Microcystis sp. colony is
observed to measure 100 µm across its diameter, thus the vol-
ume (V ) = 4/3 × π × 503 = 523,598.78 µm3. If the average
cell diameter of the individual Microcystis sp. cells is measured
to be 4 µm, then the biovolume of one cell is (V = 4/3 × π ×
23 = 33.51 µm3). The number of cells making up the colony
can then be estimated by dividing colony biovolume by cell
biovolume (523,598.78/33.51 = 15,625.15 cells). This proc-
ess is repeated for all of the colonies, and then each genus is
summed up and divided by the total volume added to the
counting chamber to come up with an estimate of cells/ml
for each.
TOXIN QUANTIFICATION
Likely the most widespread method for cyanotoxin analysis
used today is ELISA. These assays rely on antibodies binding
with specific antigens present in a sample. The presence and
concentration of the antigen (i.e., toxin) is then determined
by adding an enzymatic substrate that produces a colorimetric
change. The only specialized equipment required for this
method is a spectrophotometer plate reader that measures
the magnitude of the colorimetric change in each well. These
values are then related to a standard curve generated from the
absorbance values for a range of standards of known concen-
tration that come with the kits. These kits cost approximately
$450 per 96-well assay plate, although individual strips of 8
wells can be removed to get multiple runs out of a single
kit. However, standard curves will need to be generated for
each analysis and will reduce the number of total wells avail-
able for sample quantification. There are now commercially
available ELISA kits for the quantification of all four major
classes of freshwater cyanotoxins: anatoxin-a, cylindrosper-
mopsin, microcystin-LR (monoclonal) and total microcys-
tin/nodularin ( polyclonal, all congeners), and saxitoxin.
Liquid samples (lake water + cyanobacterial cells) are the
most commonly analyzed substrate for total toxin—which
includes the extracellular and intracellular fractions, although
cells can first be concentrated by filtration or freeze-drying if
desired. Various cell lysis techniques have been successfully
employed to access the intracellular toxin fraction, including
three repetitive freeze-thaw cycles, autoclaving, sonication,
bead beating, lyophilization and resuspension in dH2O or
MeOH, boiling and chemical extraction using compounds
such as 1% TritonX-100 and 75% methanol, or proprietary
kits (e.g., Abraxis QuikLyse) (7, 21, 22). Cell lysis should
be performed in glass scintillation vials to minimize toxin
sorption to container walls. The average time required to
carry out this type of analysis is 3–4 h, which includes the
toxin extraction step. Limits of detection for these assays
are approximately 10 µg/liter for anatoxin-a, 0.04 µg/liter
for cylindrospermopsin, 0.1 µg/liter for microcystin/nodu-
larin, and 0.015 µg/liter for saxitoxin. One drawback of these
assays is that if multiple toxins may be present—such as with
Anabaena—then the operator would need to conduct multi-
ple different ELISA analyses, whereas the next two methods
discussed allow multiple toxins to be quantified sequentially
from the same sample run.
High-performance liquid chromatography (HPLC) was
one of the first methods developed for identification and
quantification of cyanotoxins, and it is still commonly used.
Under this approach, extracted (hydrophobic) toxins are dis-
solved within a solvent (e.g., acetonitrile or MeOH) and then
pumped through reverse-phase C18 analytical columns to
separate individual constituents. The time it takes for an
organic compound to move through the column depends
on its physical interaction with the sorbent being pumped
through the column (23). Analysis relies on comparisons of
the retention time and the spectra generated by each com-
pound exposed to either UV or photodiode arrays—the latter
being more common—then subsequently relating them to
known analytical standards (24). As such, this procedure
can be used to detect any dissolved organic compound in
which a standard is available. HPLC has been used to identify
and quantify all major classes of cyanotoxins and can also
determine specific toxin variants (e.g., MC-LR, MC-YR,
MC-LA). Additionally, dilute samples can be concentrated
onto reverse-phase C18 and carbon cartridges and then
eluted with a 90% MeOH-H2O solution for more accurate
quantification. The limitation of this approach is that the
only toxin congeners that can be quantified are those that
have a reference standard. Since some cyanotoxins contain
dozens of different congeners, HPLC may underestimate the
total toxin concentration of a sample.
Liquid chromatography–mass spectrometry (LC-MS)
analysis of cyanotoxins has become more commonplace in
recent years. This approach provides the most powerful reso-
lution (highest specificity) of cyanobacterial toxin congeners
and has the lowest minimum detection limits (highest
sensitivity). HPLC minimum detection limits range from
approximately 46 ng/liter for anatoxin-a, 100 ng/liter for
cylindrospermopsin, and 2–10 µg/liter for microcystins, mak-
ing it generally more sensitive than ELISA for detecting these
analytes (25). The most significant drawback is that this
equipment is very expensive—in the range of $150,000 to
$300,000—and it requires extensive specialized training to
accurately operate. Mass spectrometric techniques are contin-
ually being refined and improved with much of the customiza-
tion pertaining to the ionization and ion analysis steps.
However, on the most basic level, liquid chromatography col-
umns are first used to separate individual organic compounds
just as in HPLC. Then the mass spectometer converts the
analytes into charged (ionized) states; it is these ions and
ion fragments that are then analyzed on the basis of their
mass to charge ratio (m/z). This approach also relies on the
 3.1.2. Best Practices for Cyanobacterial Harmful Algal Bloom Monitoring ▪ 3.1.2-7
incorporation of analytical standards, therefore analysis is gen-
erally restricted to a subset of toxin congeners due to logistic
limitations. This drawback is less pronounced for some classes
of cyanotoxins, such as anatoxin-a and cylindrospermopsin,
where additional variants are rare (26). However, there are
over 90 different microcystin congeners identified to date
and at least 27 variants of saxitoxin (27, 28). In this case, only
the most common forms can be feasibly quantified, whereas
polyclonal ELISA kits have the benefit of detecting all toxin
congeners and may therefore more accurately convey the total
public health risk. For a detailed protocol, see the recently
released standard method for measuring microcystins in fin-
ished drinking water using LC-MS/MS issued by the U.S.
EPA (EPA Method 544, Document#EPA/600/R-14/474).
While ELISA is the most accessible toxin quantification
method in terms of cost and specialized equipment required,
it suffers from its inability to identify specific toxin congeners
or very low concentrations of toxin. However, these concerns
are primarily of interest from a research perspective and may
not directly influence management responses. Note that
there is at least one “dipstick”-type product on the market
that is akin to a litmus paper test for measuring microcystin
concentrations. This product may be useful for a quick indi-
cation of the presence/absence of microcystin in a sample,
but it lacks the sensitivity to base public health management
decisions based solely on these values. As such, a positive
sample using this approach should be followed up with a
more quantitative method to accurately characterize the
risk. One final observation worthy of note: it has recently
been discovered that microcystin can bind with a variety of
intracellular proteins during periods of oxidative stress, and
that this bound microcystin–protein complex is often unde-
tected by the assays outlined here (29). The extent that we
may collectively be underestimating toxin risks, at least in
the case of microcystin and nodularin, because of this phe-
nomenon is poorly understood, but at least one study has
identified that these toxin–protein complexes are signifi-
cantly underestimated by ELISA and LC-MS (30). The bio-
activity of these protein-bound microcystin complexes is
unknown. In light of this finding, it is possible that the micro-
cystin guidelines values that were based on environmental
and culture-based observations may have significantly under-
estimated cellular quotas for these compounds. Figure 3 dis-
plays microcystin concentrations relative to cyanobacterial
abundances (Microcystis sp. + Planktothrix sp.) based on data
from the U.S. Environmental Protection Agency’s 2007
National Lakes Assessment of 1,028 North American Lakes.
The diagonal line transecting the data is based on the maxi-
mum expected toxin concentrations relative to cyanobacte-
rial abundance taken from the World Health Organization’s
guidelines for managing toxic cyanobacteria in water (7).
The large number of exceedances (data points to the left of
the diagonal line) suggests that this estimation of cell quota
maxima is too low or that there are other prolific producers
of microcystin in U.S. waters in addition to Microcystis and
Planktothrix. A review of the scientific literature identified
several observations of microcystin cell quotas exceeding
those estimated by the WHO guidelines, with one being as
much as 3.5-fold higher (0.726 pg/cell) (31). Unfortunately,
cell quota data are lacking for the other cyanotoxins. There-
fore, the maximum expected toxin loads relative to cell abun-
dance for these compounds is presently unresolved. By using
qPCR assays that target specific toxin genes, the percentage of
toxigenic cells relative to total cells can be discerned (32). It
is possible that when toxin concentrations exceed the maxi-
mum expected based on cell concentration it is simply
because the population is dominated by a toxic strain. As
such, water quality managers are encouraged to determine this
type of information for their systems and to share this data
broadly to facilitate state, regional, and national standards.
NUCLEIC ACID ANALYSIS
There are a variety of methods for extracting cyanobacterial
DNA, most of which require the use of a microcentrifuge to
separate high molecular weight proteins and cellular debris
from the DNA. The first step is to break open or lyse the cells
to gain access to the nucleic acids within. However, cyano-
bacteria have durable peptidoglycan cell walls and often pro-
duce profuse quantities of polysaccharide-rich mucilage that
makes complete cell lysis a challenge (33). A combination
of physical (e.g., boiling, freeze-thawing, bead beating) and
chemical (e.g., lysozyme, proteinase k, sodium dodecyl
sulfate) disruption techniques are typically employed to
ensure total lysis. The homogenate is then spun for 60 s at
13,200 × g to pellet high molecular weight compounds,
whereas the nucleic acids remain in suspension. The superna-
tant is then carefully collected without disturbing the pellet
and is further purified by either adsorbing it onto a silica col-
umn obtained from a commercially available DNA extraction
kit or by chemical precipitation. For the latter, phenol/
chloroform extraction followed by an ethanol precipitation
step is by far the most commonly used method for obtaining
high-quality DNA from a variety of cell types and substrates.
To do this—while working in a sterile fume hood—one vol-
ume of phenol:chloroform:isoamyl alcohol (25:24:1) is added
to supernatant containing the DNA and is vigorously mixed
or vortexed for 20 s. The microcentrifuge tube is then spun at
room temperature for 5 min at 16,000 × g, which produces
two distinct phases within the tube. The upper aqueous phase
is collected while being careful to not carry over any of the
phenol-containing bottom layer. The DNA is then precipi-
tated by adding 0.5 volumes of 7.5 M NH4OAc and 2.5 vol-
umes of 100% EtOH. For improved recovery an inert carrier
protein, such as glycogen (20 µg), can be added to the mix-
ture prior to precipitation, although this is optional. The
tube is then incubated overnight at −20°C or for at least 1
h at −80°C. The sample is then centrifuged for 30 min at
16,000 × g at 4°C. At this point an opaque pellet will usually
be visible at the bottom of the microcentrifuge tube, but this
depends on the starting concentration of cells. The superna-
tant can now be carefully removed and discarded; the pellet is
then washed in 150 µl of fresh 70% ethanol (EtOH) in dH2O,
then centrifuged for 5 min at 16,000 × g at 4°C. The EtOH is
removed and the pellet is washed once again as before and
then centrifuged. Finally the EtOH is removed and any resid-
ual EtOH is allowed to volatilize off of the pellet by placing
the open tube on ice with a Kimwipe covering it for 2–5
min. It is important to remove the majority of the EtOH since
it can interfere with downstream enzymatic reactions such as
PCR or ligation, but it is also important to not overdry the
pellet, which can cause problems dissolving the DNA back
into solution. The DNA pellet is resuspended in 100 µl ster-
ile, nuclease-free dH2O or a pH (7.5–8.0) stable buffer such as
1 × Tris EDTA or 1 × Tris-HCl. The DNA is now ready for
downstream analysis or can be stored indefinitely at −80°C or
for at least 1 year at −20°C. Note that it is best practice to per-
form a DNA extraction negative in parallel with any sample
extractions to rule out the possibility of cross-contamination
in any downstream analysis.
Conventional PCR is the original variation of the
method, and it is still routinely used to qualitatively establish
3.1.2-8 ▪ WATER

FIGURE 3 Microcystin concentrations observed in 2007 EPA National Lakes Assessment for sites containing the two most prolific micro-
cystin producing genera; Microcystis sp. and Planktothrix sp. and each lake’s relative trophic status. The trend line represents maximum expected
microcystin concentrations (0.2 pg/cell) relative to cyanobacterial cell concentration based on World Health Organization guidelines (7), hor-
izontal lines denote recreational and drinking water maximum contaminant levels (MCL) recommended by WHO. Note that as lake trophic
status increases, cyanobacterial biomass, microcystins, and the percent likelihood of exceeding the MCL also increase.
doi:10.1128/9781555818821.ch3.1.2.f3

the presence or absence of cyanobacterial genes. This
approach has utility as an early warning sentinel for the pres-
ence of potentially toxic cyanobacteria, especially since
detection sensitivity is enhanced by the standard practice of
concentrating cells from environmental samples onto filters
prior to DNA extraction. While this approach may signal
the presence of cyanobacteria, it does not reliably indicate
the quantity of cyanobacteria present. For this, real-time
qPCR is often used because it can quantify the number of spe-
cific gene copies present in a sample. This process requires a
reference standard in which to relate the values observed
from an environmental sample. There are several methods
for producing DNA standards for qPCR. The most rudimen-
tary is to count the number of cells in a culture under the
microscope, extract the DNA from this known quantity of
cells, and then relate the cycle threshold values obtained in
the qPCR assays to those from unknown environmental sam-
ples. Limitations of this approach stem from cell counting
inaccuracies which will lead to misinterpretation of environ-
mental estimates, the possibility that gene copy numbers may
vary from one strain or time point to the next, and that reac-
tions may suffer reduced amplification efficiency if the refer-
ence strain’s hybridization sites contain any nucleotide
mismatches relative to the primers. These problems can be
overcome by the use of plasmid standards or commercially
available synthetic gene constructs (e.g., gBlocks). For the
former, a PCR assay is conducted and the amplification prod-
uct is separated by electrophoresis on an agarose gel, then the
DNA band is extracted, purified, and ligated into a plasmid
vector. This vector is then cloned into E. coli and grown over-
night, which results in an exponential increase in the number
of plasmid copies carrying the DNA insert. Next, a mini-prep
kit is used to harvest the plasmid DNA from the cells. The
circular plasmids are then linearized by restriction enzyme
digestion, purified, and quantified to estimate copy number.
This approach will produce DNA fragments of the gene target
of interest that will have perfect homology with the primers.
As a result, PCR efficiency will be improved and the assay
limits of detection, sensitivity, and specificity can be deter-
mined. The biggest drawback to this approach lies in accu-
rately quantifying the amount of plasmid DNA template
that is generated. UV spectrophotometers are most com-
monly used to measure DNA concentrations; they do so by
determining the absorbance ratios of the samples exposed
to light at 260 nm and 280 nm wavelengths (A260/A280),
where values between 1.7 and 2.0 indicate suitably pure
DNA for downstream analysis. A limitation of these devices
is that they historically required a large volume of DNA (e.g.,
400 µl–1 ml), although there are now spectrophotometers
such as the UV NanoDrop that can quantify DNA/RNA
from as little as 1 µl of sample. An additional technique using
a plate reader spectrophotometer is to incorporate fluorescent
dyes that bind double-stranded DNA and allow it to be meas-
ured based on its fluorescence, which is more sensitive than
measuring its absorbance. Finally, one of the more sophisti-
cated approaches for DNA quantification is the use of micro-
fluidic chips run on specialized equipment such as the Agilent
BioAnalyzer. This method only requires 1 µl of DNA and can
also quantify DNA fragment lengths within a sample.
There is a new technique rapidly gaining acceptance—
known as digital PCR—for nucleic acid detection and quan-
tification. These devices operate by dividing a sample into
thousands of independent qPCR reactions, some of which
will contain DNA template whereas others will not. After
 3.1.2. Best Practices for Cyanobacterial Harmful Algal Bloom Monitoring ▪ 3.1.2-9
30–40 cycles of PCR, each subreaction is quantified and the
data are fitted to a Poisson distribution that results in a highly
sensitive, absolute count of the target molecules present in
the sample. As such, this device could be used to generate
and accurately quantify DNA templates to be used for
qPCR standards in the types of reactions described above,
although its greatest strength is that it can quantify unknown
templates without relying on external standards, making it
more approachable for users with limited molecular biology
experience.
Currently, PCR-based approaches for cyanobacteria mon-
itoring are almost exclusively performed within an academic
setting. However, it is only a matter of time before water qual-
ity laboratories begin incorporating molecular diagnostics
into their monitoring strategies, especially since in many
regions these methods are already in place for monitoring
enteric bacteria and other pathogens. There are no public
health guidelines in place for interpreting cyanobacterial risks
based on toxin gene abundances. Although these data better
describe the potential for toxin production than cell counts,
since the population may be composed of a mixture of toxic
and nontoxic strains, the amount of toxin that is ultimately
produced depends on how these genes are being regulated
and not just how many are present. Even so, these genes
seem to be constitutively expressed—at least at a basal level,
with up to several-fold variations in RNA transcript content
possible depending on environmental conditions and cell
physiologic status. What this means is that if toxin genes
are detected in a sample there is a high probability that toxins
will be present as well. Molecular diagnostics provide several
other advantages, including information on cyanobacterial
bloom dynamics such as distribution patterns of toxic strains
and population succession rates; most important, with limits
of detection of 10 genes or fewer, toxigenic cells can be
detected earlier than by microscopy or toxin analysis.
A reasonable way to include qPCR analyses into a moni-
toring plan is illustrated in Fig. 2. Although the gold standard
remains toxin testing, qPCR has the potential to supplant
microscope counting due to the reasons listed above. To do
this, qPCR assays should focus on gene targets that are highly
conserved within the organisms of interest and gene targets
should also ideally only be present as a single copy within
the chromosome. The most widely used target for bacterial
identification continues to be 16S ribosomal RNA gene; it
represents the most robust sequence database for taxonomic
identification and can be used for PCR primer design. There
are general primer sets that are meant to amplify the 16S
rRNA gene from all cyanobacteria (Table 2). These ampli-
cons can then be sequenced and a BLAST analysis can be
used to identify the genera present. The problems with this
approach include the relatively long turnaround time for
sequencing and that the high sequence conservation of this
locus means that resolution beyond the genus level is gener-
ally not possible. Amplification of the 16S–23S rRNA inter-
genic spacer region (ITS) can be used to provide additional
resolution since it includes a noncoding region that is more
prone to mutation, which enables species- or strain-level
comparisons. The benefit of these primers is that the ampli-
con size can serve as a rough guide of the cyanobacteria
present—even without sequencing, since each genus will
produce slightly different sized bands when viewed on an
agarose gel (34). Note that contrary to their original appli-
cation, these primers can be used without the inclusion of a
GC clamp and do not require denaturing gradient gel electro-
phoresis to visualize 16S–23S rRNA ITS differences (35).
However, in many cyanobacterial genomes the ribosomal
operons are present in multiple copies (36, 37). Therefore,
it makes sense to use this gene for conventional PCR to deter-
mine the presence of cyanobacteria, since multiple gene cop-
ies within a single cell will enhance detection sensitivity.
However, variable copy numbers make absolute quantifica-
tion by qPCR unreliable since there is no way to know how
the results relate to actual cell abundances. Instead, one of
the most common operons targeted for total or group-specific
cyanobacterial enumeration are the photopigment genes c-
phycocyanin A and B (cpcAB) (38). Table 2 contains com-
monly used primers for cyanobacterial analysis and their
related information. As a general rule of thumb, SYBR green
I qPCR chemistry is used to quantify amplicons ranging from
approximately 250–400 bp in size, whereas Taqman chemis-
try is more target specific since it includes a separate fluoro-
phore probe and is used on smaller amplicons ranging from
approximately 60 to 200 bp.
For absolute quantification of target genes using qPCR,
two additional concerns need to be addressed. The first is
that the DNA extraction efficiency must be calculated to
properly interpret the results. This can be done by comparing
the cycle threshold (Cq) values from a qPCR assay using a
known concentration of cells that had the DNA extracted
by boiling for 15 min (assuming 100% cell lysis) followed
by a phenol-chloroform purification as outlined above versus
the same concentration of cells extracted from a filter by
whichever method chosen (e.g., commercially available
DNA extraction kit). This correction factor is important for
determining the assay’s actual limit of detection. For instance,
if the DNA extraction efficiency is only 50% from a filter and
the minimum number of detectable gene copies is 10, then
the minimum limit of detection is 20 cells. The other concern
is when there is too much organic matter present in a
sample—which may occur during a dense cyanobacterial
bloom. Under these circumstances it is possible for the
qPCR reactions to be inhibited owing to various compounds
such as salts, heavy metals, humic and tannic acids, or a suite
of polysaccharides, proteins, and organic compounds (39).
An inhibited reaction may still yield a positive signal,
although the intensity is reduced. Therefore all qPCR sam-
ples should at least initially be screened for inhibitors, espe-
cially when applying the assays to a new ecosystem, or if the
samples contain high cell concentrations to determine the
level of variation introduced by PCR inhibition. The easiest
way to test for inhibition is to spike a known amount of exog-
enous DNA or an armored-RNA standard (e.g., for reverse
transcriptase qPCR assays) into an environmental sample
and compare the Cq values from the qPCR assay for the
spiked sample versus the control (40, 41). If the Cq values
are within one unit of each other, then PCR inhibition can
be ruled out. However, if there is evidence of inhibition,
then the samples should be diluted 1:10 in nuclease-free
dH2O and the assay is repeated. For very dense cyanobacte-
rial blooms, or samples collected in water bodies containing
high humic loads, heavy metals, or other chemicals, it may
be necessary to dilute 1:100 or more to attenuate out the
inhibiting substances. Since excessive dilution will result
in decreased assay sensitivity, in these cases it may be nec-
essary to perform an additional clean-up step on the DNA
to remove inhibitors.
Finally, since some genera of cyanobacteria, such as Ana-
baena sp., have been shown to be possible producers of all
major classes of cyanotoxins, the use of a multiplex qPCR
approach may be beneficial if these types of genera are
present. For this approach, SYBR green I chemistry cannot
be used because it will not distinguish between the different
3.1.2-10 ▪ WATER

TABLE 2 Commonly used PCR/qPCR primers for CyanoHAB detection and quantification and their optimal thermal cycler settings (any
fluorophore/quencher that is compatible can be used with these assays)

TA Size
Target (gene) Primers and probes (50 –30 ) Reference
(°C) (bp)
Forward: AGCCACACTGGGACTGAGACA
All cyanobacteria (16S) Probe: [FAM]-CCTACGGGAGGCAGCAGTGGG-[BHQ] 60 80 42
Reverse: TCGCCCATTGCGGAAA
Forward: G(T/C)CACGCCCGAAGTC(G/A)TTAC 34
All cyanobacteria Probe: Amplicon too large for Taqman, must use SYBR green 60 variablea
(16S–23S rRNA) Reverse: CCTCTGTGTGCCTAGGTATC
Forward: GGCTGCTTGTTTACGCGACA 46
All cyanobacteria Probe: Amplicon too large for Taqman, must use SYBR green 55 685 38
(cpcBA) Reverse: CCAGTACCACCAGCAACTAA
Forward: TCTGGTATTCAGTCCCCTCTAT
Anatoxin-a (anaC) Probe: Amplicon too large for Taqman, must use SYBR green 58 366 47
Reverse: CCCAATAGCCTGTCATCAA
Forward: GTCTGCCCACGTGATGTTATGAT
Cylindrospermopsin Probe: [FAM]-CCTTTGGGAACGAAATTCTCGAAGCAACT-[BHQ] 60 71 42
(cyrA) Reverse: CGTGACCGCCGTGACA
Forward: AATAAATCATAATTTAGAACSGGVGATTTAGG
Microcystin (mcyE/ndaF) Probe: [FAM]-AATCAAGTTAAGGTVAATGGYTATCG-[BHQ] 60 128 42
Reverse: AATAAATCATAACGRBTVADTTGRTATTCAATTTCT
Forward: GGAGTGGATTTCAACACCAGAA
Saxitoxin (sxtA) Probe: [FAM]-TGCCGATTTAGAAGAAAGTATCCTCTCAG-[BHQ] 60 148 42
Reverse: GTTTCCCAGACTCGTTTCAGG
a
16S–23S rRNA ITS amplicon sizes (bp) for various genera: Anabaena (500, 700); Anabaenopsis (450, 725); Aphanizomenon (475, 700); Cylindrospermopsis
(425, 600); Gloeotrichia (475, 775); Leptolyngbya (500); Lyngbya (675); Microcystis (550); Nodularia (575, 875); Nostoc (500, 775).

amplicons generated during the PCR, since it makes all
double-stranded DNA fluoresce. Therefore target-based
chemistries, such as Taqman or Scorpion probes, must be uti-
lized. Each gene probe is labeled with a different fluorophore
that fluoresces at a different wavelength. The thermal cycler
has multiple channels for detecting each dye’s unique wave-
length and can measure the fluorescence intensity of each
probe, allowing quantification of multiple gene targets simul-
taneously from a single sample. This approach is more compli-
cated than just assaying for single genes, but has the benefit of
reduced cost and time relative to running four separate assays.
This multiplex approach has been successfully used to quan-
tify potentially toxigenic cyanobacteria (42).
RNA analysis is not covered in detail in this chapter,
although we comment on sample collection, storage, and
extraction considerations. Since RNAs are transient mole-
cules, it is important that they are stored properly to protect
against degradation. The problem with just freezing a water
sample is that upon thawing, the cells will lyse and RNases
within the sample can rapidly degrade the RNA. Therefore,
it is recommended that samples are filtered immediately in
the field using a vacuum manifold or a hand pump, then
placed in nuclease-free microcentrifuge tubes and either flash
frozen in liquid nitrogen or stored in a mono-phasic solution
of phenol and guanidine isothiocyanate, such as Trizol (Invi-
trogen), then placed on dry ice. Once back at the laboratory,
they can be moved to −80°C for long-term storage. A third
option that is convenient when processing samples in the
field is to place the filters in a preservative such as RNAlater
(Qiagen) and store them overnight at 4°C. This preservative
will permeate the cells and stabilize the RNA without causing
cell lysis. After incubation at 4°C for 24 h, the RNA is stable
for up to 1 week at room temperature or can be archived
indefinitely at −20°C or −80°C. For RNA extraction, these
samples can be centrifuged at 4°C for 5 min, then the RNA-
later is carefully pipetted off of the filter. An extraction buffer
(800 µl), such as Trizol, is then immediately pipetted onto
the filtered sample, followed by vigorous vortexing and/or
bead beating for 3–5 min, then left for 5 min at room temper-
ature. The next step is to transfer as much of the liquid as pos-
sible to another nuclease-free microcentrifuge tube and add
120 µl chloroform per 600 µl of recovered homogenate. For
optimal recovery, 20 µg of glycogen can optionally be added
at this time to help improve RNA precipitation. The sample is
then mixed by inverting for 20 s, then allowed to sit for 2 min
at room temperature. Finally, the sample is centrifuged for 15
min at 11,000 rpm and at 4°C. The mixture should separate
into three distinct phases: a lower (red) chloroform phase
containing proteins, lipids, and filter debris; an opaque/white
interphase containing mostly DNA; and a clear, upper aque-
ous phase containing most of the RNA. Collect the upper
aqueous phase into a new microcentrifuge tube, taking care
to not include any of the interphase, then centrifuge for 5
min at 11,000 rpm at 4°C. Collect the upper aqueous phase
once again and then add 400 µl isopropanol and incubate
for 10 min at room temperature, then centrifuge at 4°C for
15 min at 12,000 rpm. The RNA will precipitate and form
a gel-like pellet. Decant the supernatant and add 250 µl of
fresh 75% EtOH (in DEPC-H2O), vortex gently to resuspend
the pellet, then centrifuge for 5 min at 8,500 rpm at 4°C.
Carefully remove the supernatant and air dry the pellet on
wet ice with a Kimwipe covering the open centrifuge tube.
Dissolve the pellet with DEPC-H2O such that the final con-
centration is approximately 1–10 µg/ml of RNA. Note that
the volume of water to add will take some experimentation
since it depends on the size of the RNA pellet. Leave the
 3.1.2. Best Practices for Cyanobacterial Harmful Algal Bloom Monitoring ▪ 3.1.2-11
sample on ice while occasionally mixing it over a 10 min
period, then centrifuge for 5 min at 8,500 rpm and 4°C to pel-
let any remaining insoluble material. Collect the supernatant
and measure the A260/A280 ratio to determine purity. The
RNA is ready for use at this point or can be stored indefinitely
at −80°C. Note that for reverse-transcriptase qPCR, it is rec-
ommended that a DNase I treatment step is conducted to
ensure total removal of any contaminating DNAs prior to
cDNA synthesis.
OTHER CONSIDERATIONS
Not all cyanobacteria exhibit a pelagic lifestyle; many form
benthic mats in shallow areas that receive adequate light
penetration through the water column. In faster moving
systems—especially rivers—benthic cyanobacteria abundan-
ces may exceed those of pelagic cells. Many genera of benthic
cyanobacteria are recognized as taste and odor producers, in
addition to their toxigenic potential. The obvious difficulty
in making accurate visual assessments of cyanobacterial risks
means that these organisms are often left unaccounted for
until an adverse event occurs. Early warning signs may
include the production of earthy or musty taste and odors in
finished drinking water (e.g., geosmin or 2-methylisoborneol).
Often these benthic strains go unnoticed until there is an
acute intoxication event involving a pet or livestock that
drank from a ditch or a bedrock puddle found along rocky
shorelines. Thus, water quality managers should be alert to
the possibility of these organisms—especially if there are
ever periods of water quality deterioration without any appa-
rent causative agent (i.e., visual bloom). In these cases,
benthic samplers such as an Eckman dredge or even a PVC
pipe can be used to scrape or core into the sediment at approx-
imately 1 m intervals through the water column and down to
either twice the Secchi depth—an approximation of 1% pho-
tosynthetic active radiation availability—or 5 m if a Secchi
disk is unavailable. Since benthic cyanobacteria will only
grow as deeply as the light permits, the appropriate sampling
depth will depend on light intensity and water clarity. Finally,
if the benthic substrate can be removed, such as small rocks or
cinder blocks, then the algae can be scraped off and collected
in a sample jar and processed in a manner analogous to any
other water column sample.
While neurotoxin production has long been associated
with certain genera of marine phytoplankton, including the
dinoflagellates Gymnodinium sp., Alexandrium sp., and Pyrodi-
nium sp., and some salt-tolerant cyanobacteria such as Lyng-
bya sp., Anabaena sp., and Aphanizomenon sp. (43), the
hepatotoxins microcysin and cylindrospermopsin have classi-
cally been viewed as freshwater toxins. This perspective has
changed following recent reports of marine animal deaths
due to acute microcystin intoxication (44). Additional inves-
tigations along coastal areas exposed to freshwater inputs have
documented the widespread occurrence of microcystins in
this environment (45). Due to the stable nature of these com-
pounds and their ability to bioaccumulate in shellfish, it is
clear that physical transfer along the freshwater–marine con-
tinuum should not be overlooked when developing manage-
ment strategies.
CONCLUDING REMARKS
Cyanobacteria are pervasive throughout many North Amer-
ican aquatic ecosystems. Human activities are exacerbating
the impacts of these organisms. The lack of any central regu-
latory approach greatly hampers our monitoring, mitigation,
and educational outreach efforts. County and state public
health and environmental quality departments will need to
spearhead these efforts, and ideally, they will adopt a standard
set of procedures for sample collection, analysis, and response.
A central data repository could serve as a bridge between
managers, policy makers, and academic laboratories that are
all independently interested in issues pertaining to cyanobac-
terial blooms. Such a repository could also greatly inform our
collective knowledge of cyanobacterial ecology and toxicity
by comparing a range of impaired and unimpaired aquatic
ecosystems. This would guide public health outreach and
lake remediation efforts by targeting the highest risk systems,
as well as possibly revealing novel management options by
investigating non-impacted lakes that are otherwise expected
to contain cyanobacterial blooms based on their physico-
chemical characteristics. Large data sets such as the EPA
National Lakes Assessment have demonstrated that we
have historically been underestimating microcystin produc-
tion in cyanobacteria and that the guidelines should be
adjusted to account for this discrepancy. A similar approach
is warranted for all other major classes of cyanotoxins to better
frame their production in an environmental context, which
in turn will help determine if our management guidelines
adequately protect public health or if they are in need of
revision.
REFERENCES
1. Schirrmeister BE, de Vos JM, Antonelli A, Bagheri HC.
2013. Evolution of multicellularity coincided with increased
diversification of cyanobacteria and the great oxidation
event. Proc Natl Acad Sci 110:1791–1796.
2. Lyons TW, Reinhard CT. 2011. Earth science: sea change
for the rise of oxygen. Nature 478:194–195.
3. Paerl HW, Otten TG. 2013. Harmful cyanobacterial
blooms: causes, consequences, and controls. Microbial Ecol
65:995–1010.
4. Shih PM, Wu D, Latifi A, Axen SD, Fewer DP, Talla E,
Calteau A, Cai F, Tandeau de Marsac N, Rippka R, Herd-
man M, Sivonen K, Coursin T, Laurent T, Goodwin L,
Nolan M, Davenport KW, Han CS, Rubin EM, Eisen JA,
Woyke T, Gugger M, Kerfeld CA. 2013. Improving the cov-
erage of the cyanobacterial phylum using diversity-driven
genome sequencing. Proc Natl Acad Sci 110:1053–1058.
5. Humbert J-F, Barbe V, Latifi A, Gugger M, Calteau A,
Coursin T, Lajus A, Castelli V, Oztas S, Samson G, Longin
C, Medigue C, Tandeau de Marsac N. 2013. A tribute to dis-
order in the genome of the bloom-forming freshwater cyano-
bacterium Microcystis aeruginosa. PLoS One 8(8):e70747.
6. Rankin KA, Alroy KA, Kudela RM, Oates SC, Murray
MJ, Miller MA. 2013. Treatment of cyanobacterial (Micro-
cystin) toxicosis using oral cholestyramine: case report of a
dog from Montana. Toxins 5:1051–1063.
7. Chorus I, Bartram J. 1999. Toxic cyanobacteria in water: a
guide to their public health consequences, monitoring and
management, In Chorus I, Bartram J (eds), Toxic Cyanobac-
teria in Water. E & FN Spon on Behalf of the World Health
Organisation, London.
8. Chiswell RK, Shaw GR, Eaglesham G, Smith MJ, Norris
RL, Seawright AA, Moore MR. 1999. Stability of cylin-
drospermopsin, the toxin from the cyanobacterium, Cylin-
drospermopsis raciborskii: effect of pH, temperature, and
sunlight on decomposition. Environ Toxicol 14:155–161.
9. Edwards C, Graham D, Fowler N, Lawton LA. 2008. Bio-
degradation of microcystins and nodularin in freshwaters.
Chemosphere 73:1315–1321.
3.1.2-12 ▪ WATER
10. Kayal N, Newcombe G, Ho L. 2008. Investigating the fate
of saxitoxins in biologically active water treatment plant fil-
ters. Environ Toxicol 23:751–755.
11. Kaminski A, Bober B, Lechowski Z, Bialczyk J. 2013.
Determination of anatoxin-a stability under certain abiotic
factors. Harmful Algae 28:83–87.
12. Burch MD. 2008. Effective doses, guidelines & regulations.
Proceedings of the Interagency, International Symposium
on Cyanobacterial Harmful Algal Blooms (ISOC_HAB),
p. 831–853. In Hudnell HK (ed), State of the science and
research needs, Springer, New York.
13. Reynolds CS, Oliver RL, Walsby AE. 1987. Cyanobacterial
dominance: the role of buoyancy regulation in dynamic lake
environments. N Z J Mar Freshwater Res 21:379–390.
14. Kromkamp JC, Mur LR. 1984. Buoyant density changes in
the cyanobacterium Microcystis aeruginosa due to changes in
the cellular carbohydrate content. FEMS Microbiol Lett 25:
105–109.
15. Visser PM, Passarge J, Mur LR. 1999. Modelling vertical
migration of the cyanobacterium Microcystis. Hydrobiologia
349:99–109.
16. Walsby AE. 1994. Gas vesicles. Microbiol Rev 58:94–144.
17. Utermöhl H. 1958. Zur Vervollkommnung der quantitative
Phytoplankton-Methodik Mitt int Verein theor angew. Lim-
nol 5:567–596.
18. Box JD. 1981. Enumeration of cell concentrations in suspen-
sions of colonial freshwater microalgae, with particular refer-
ence to Microcystis aeruginosa. Brit Phycol J 16:153–164.
19. Laslett GM, Clarke RM, Jones GJ. 1997. Estimating the
precision of filamentous blue-green algae cell concentration
from a single sample. Environmetrics 8:313–339.
20. Hillebrand H, Dürselen CD, Kirschtel D, Pollingher
U, Zohary T. 1999. Biovolume calculation for pelagic and
benthic microalgae. J Phycol 35:403–424.
21. Loftin KA, Meyer MT, Rubio F, Kamp L, Humpries E,
Whereat E. 2008. Comparison of Two Cell Lysis Procedures
for Recovery of Microcystins in Water Samples from Silver
Lake in Dover, Delaware, with Microcystin Producing Cyano-
bacterial Accumulations. U.S. Geological Survey, Reston, VA.
22. Kim IS, Nguyen GH, Kim S, Lee J, Yu H-W. 2009. Evalua-
tion of methods for cyanobacterial cell lysis and toxin
(microcystin-LR) extraction using chromatographic and mass
spectrophotometric analyses. Environ Eng Res 14:250–254.
23. Purdie EL, Young FM, Menzel D, Codd GA. 2009. A
method for acetonitrile-free microcystin analysis and purifi-
cation by high-performance liquid chromatography, using
methanol as mobile phase. Toxicon 54:887–890.
24. Meriluoto J, Codd GA (eds). 2005. Toxic: Cyanobacterial
Monitoring and Cyanotoxin Analysis. Abo Akademi Uni-
versity Press, Abo, Finland.
25. Yen HK, Lin TF, Liao PC. 2011. Simultaneous detection of
nine cyanotoxins in drinking water using dual solid-phase
extraction and liquid chromatography-mass spectrometery.
Toxicon 58:209–218.
26. Stüken A, Jakobsen KS. 2010. The cylindrospermopsin
gene cluster of Aphanizomenon sp. strain 10E6: organization
and recombination. Microbiology 156:2438–2451.
27. Ho L, Sawade E, Newcombe G. 2012. Biological treatment
options for cyanobacteria metabolite removal—a review.
Water Res 45:1536–1548.
28. Puddick J, Prinsep MR, Wood SA, Miles CO, Rise F, Cary
SC, Hamilton DP, Wilkins L. 2013. Structural characteriza-
tion of new microcystins containing tryptophan and oxi-
dized tryptophan residues. Mar Drugs 11:3025–3045.
29. Zilliges Y, Kehr J-C, Meissner S, Ishida K, Mikkat S,
Hagemann M, Kaplan A, Börner T, Dittmann E. 2011.
The cyanobacterial hepatotoxin microcystin binds to pro-
teins and increases the fitness of microcystis under oxidative
stress conditions. PLoS One 6(3):e17615.
30. Meissner S, Fastner J, Dittmann E. 2013. Microcystin pro-
duction revisited: conjugate formation makes a major contri-
bution. Environ Microbiol 15:1810–1820.
31. Lyck S. 2004. Simultaneous changes in cell quotas of micro-
cystin, chlorophyll a, protein and carbohydrate during differ-
ent growth phases of a batch culture experiment with
Microcystis aeruginosa. J Plankton Res 26:727–736.
32. Otten TG, Xu H, Qin B, Zhu G, Paerl HW. 2012. Spatio-
temporal patterns and ecophysiology of toxigenic Microcystis
blooms in Lake Taihu, China: implications for water quality
management. Environ Sci Technol 46:3480–3488.
33. Tillett D, Neilan BA. 2000. Xanthogenate nucleic acid iso-
lation from cultured and environmental cyanobacteria.
J Phycol 36:251–258.
34. Janse I, Meima M, Kardinaal WEA, Zwart G. 2003.
High-resolution differentiation of cyanobacteria by using
rRNA-internal transcribed spacer denaturing gradient gel
electrophoresis. Appl Environ Microbiol 69:6634–6643.
35. Otten TG, Paerl HW. 2011. Phylogenetic inference of col-
ony isolates comprising seasonal Microcystis blooms in lake
Taihu, China. Microbial Ecol 62:907–918.
36. Klappenbach JA, Saxman PR, Cole JR, Schmidt TM.
2001. rrndb: the ribosomal RNA operon copy number data-
base. Nucleic Acids Res 29:181–184.
37. Schirrmeister BE, Dalquen DA, Anisimova M, Bagheri
HC. 2012. Gene copy number variation and its significance
in cyanobacterial phylogeny. BMC Microbiol 12:177.
38. Neilan BA, Jacobs D, Goodman AE. 1995. Genetic diver-
sity and phylogeny of toxic cyanobacteria determined by
DNA polymorphisms within the phycocyanin locus. Appl
Environ Microbiol 61:3875–3883.
39. Schrader C, Schielke A, Ellerbroek L, Johne R. 2012. PCR
inhibitors—occurrence, properties and removal. J Appl
Microbiol 113:1014–1026.
40. Haughland RA, Siefring SC, Wymer LJ, Brenner KP,
Dufour AP. 2005. Comparison of Enterococcus measure-
ments in freshwater at two recreational beaches by quantita-
tive polymerase chain reaction and membrane filter culture
analysis. Water Res 39:559–568.
41. Rueckert A, Cary SC. 2009. Use of an armored RNA stand-
ard to measure microcystin synthetase E gene expression in
toxic Microcystis sp. by reverse-transcription QPCR. Limnol
Oceanogr Meth 7:509–520.
42. Al-Tebrineh J, Pearson LA, Yasar SA, Neilan BA. 2011. A
multiplex qPCR targeting hepato- and neurotoxigenic cya-
nobacteria of global significance. Harmful Algae 15:19–25.
43. Hackett JD, Wisecaver JH, Brosnahan ML, Kulis DM,
Anderson DM, Bhattacharya D, Plumley FG, Erdner
DL. 2013. Evolution of saxitoxin synthesis in cyanobacteria
and dinoflagellates. Mol Biol Evol 30:70–78.
44. Miller MA, Kudela RM, Mekebri A, Crane D, Oates SC,
Tinker MT, Staedler M, Miller WA, Toy-Choutka S,
Dominik C, Hardin D, Langlois G, Murray M, Ward K,
Jessup DA. 2010. Evidence for a novel marine harmful algal
bloom: cyanotoxin (microcystin) transfer from land to sea
otters. PLoS One 5(9):e12576.
45. Vareli K, Jaeger W, Touka A, Frillingos S, Briasoulis E,
Sainis I. 2013. Hepatotoxic seafood poisoning (HSP) due
to microcystins: a threat from the ocean? Mar Drugs 11:
2751–2768.
46. Iteman I, Rippka R, Tandeau de Marsac N, Herdman M.
2000. Comparison of conserved structural and regulatory
domains within divergent 16S rRNA-23S rRNA spacer
sequences of cyanobacteria. Microbiology 146:1275–1286.
47. Rantala-Ylinen A, Kana S, Wang H, Rouhiainen L, Wahls-
ten M, Rizzi E, Berg K, Gugger M, Sivonen K. 2011. Ana-
toxin-a Synthetase Gene Cluster of the Cyanobacterium
Anabaena sp. Strain 37 and Molecular Methods to Detect
Potential Producers. Appl Environ Microbiol 77:7271–7278.
 Assessing the Efficiency of Wastewater Treatment
GRACIELA RAMÍREZ TORO AND HARVEY MINNIGH
3.1.3
Wastewater treatment (WWT) is applied to a waste stream to
prevent or minimize deleterious effects to receiving waters
(RW). This chapter provides an overview of how we mea-
sure the effectiveness of WWT, in other words, how well
the processes meet the goal of providing a final effluent
(FE) that does not degrade the RW. Not all wastewater treat-
ment plants (WWTPs) discharge to water bodies—there are
plants that do not discharge to any receptor, often called
“zero-discharge” plants—though most WWTPs do discharge
and we use the analogue of “receiving water” since the goal of
effective treatment is the same no matter the end destina-
tion of the effluent. By the same token, most WWTPs also
produce “solid” effluents (biosolids), in the sense of more
than 3% or 4% solids and we also consider these.
We provide a superficial overview of treatment to give a
context to methods or techniques of assessment. We do not
attempt to describe any particular process in detail, only the
measurements that will affect the FE. There are numerous
sources that describe the many processes available and some
of the best known include references (1–4). Of particular
interest to smaller and poorer communities is the recent
“return” of WWT ponds. The U.S. Environmental Pro-
tection Agency (EPA) recently published a design and
operation manual (5) that includes design, operation, and
maintenance changes and additional processes and treat-
ments to provide removals not earlier considered in ponds.
Removals of organic and inorganic constituents in influents
occur in the same way in WWT ponds as in more complex
constructed systems, though generally with fewer opportu-
nities to vary treatment and these will be slower of effect.
Most of the approximately 16,000 WWTPs in the country
in 1996 (∼86%, ∼13,000) (6, 7) are described as providing
treatment at a secondary or greater than secondary level,
which suggests that about the same number provide some
form of biological treatment. The latter treatment involves
or depends on microbial activity and the most common
forms are suspended or attached growth systems. The most
advanced systems are intended to better remove nutrients
(N and P) and generally include additional oxidation, some-
times in alternating aerobic and anoxic chambers. These
usually occur subsequent to conventional AS aeration units
and sometimes as side streams or as a final stage for either
plant mixed liquor (ML) or effluent.
Figure 1 shows a conventional activated sludge plant.
Incoming wastes are treated aerobically and most biological
doi:10.1128/9781555818821.ch3.1.3
3.1.3-1
treatment systems are aerobic. Pollutants are removed by a
mixed culture where an incoming waste stream passes through
a screen and the heavy, mostly inert solids are removed by
plain sedimentation. The settled waste stream flows into a
contact basin, where the stream is mixed with return acti-
vated sludge (RAS; underflow from secondary clarifiers)
and aerated; the combined flow is now called mixed liquor
(ML). After aeration the ML flows into secondary clarifiers,
a portion of the settled material is returned to the aeration sys-
tem, and a portion is sent to sludge concentration, possibly
further treatment, and discharged, usually to landfills or
ocean discharges or used for land application.
This is generally the schema for even advanced oxygen
processes (AOPs) where there may be additional stages or
side streams whereby aerobic processes, possibly alternat-
ing with anaerobic or anoxic stages, and additional sludge
removal might be applied to process streams before being
discharged as FE.
A typical attached-growth plant would historically be
what was known as a trickling filter, while a more modern ver-
sion might be a membrane system—for example, a rotating
biological contactor where media is attached to membranes
and moved through the waste stream or where membranes
are suspended and the waste stream moves through them,
known as a membrane biological reactor (MBR). The effec-
tive difference between suspended growth and fixed growth
systems is the physical state of the culture providing the
treatment, fixed or suspended. Other than the physical state
of the culture the processes are generally the same—primary
settling, aerobic biological treatment, and secondary settling.
A schematic of a trickling filter plant is shown at Fig. 2. The
settled effluent from the primary clarifier is spread over the
media in the filter bed by the arms that typically rotate to
cover the surface of the media in the bed.
In converting these waste streams into effluents, we usually
see at least one liquid and one solids effluent. Although both
are forms of the FE, we, in common with most regulatory
and operating bodies, think of these as two different effluents,
the plant effluent (liquid) and the plant-produced sludge or
biosolids. Both will contain chemical and biological pollu-
tants removed from the sewage or their relicts less the portion
lost to the atmosphere during treatment. We discuss the meas-
ures of efficiency of treatment for both these forms of FE.
The pollutants of interest have increased dramatically over
the years; first due to better means of analyses and increases in
3.1.3-2 ▪ WATER

FIGURE 1 Typical suspended growth (activated sludge) plant. doi:10.1128/9781555818821.ch3.1.3.f1

detection sensitivity allowing us to identify constituents
earlier only supposed or presumed to be present and, second,
due to the proliferation of products in waste streams in part
because of increased use of those products by populations
served by treatment plants.
The EPA recently published a report on the performance
of WWTP utilizing secondary treatment and treating ≥10
million gallons per day (MGD)(8) (Table 1). This report
examined the secondary treatment standards requiring efflu-
ent discharge requirements of 40 CFR 133 (at a minimum)
for most general municipal discharges to meet 30-day average
effluent concentration limits of 30 mg/liter for biochemical
oxygen demand 5-day test (BOD5) and total suspended solids
(TSS). Alternative 30-day average limits for carbonaceous

FIGURE 2 Typical attached growth (trickling filter [biofilter]) plan. doi:10.1128/9781555818821.ch3.1.3.f2


TABLE 1 Report on secondary treatment performancea
Category
N Results TSS limit = 30
Monthly median for plants
95th percentile
N Results TSS limit ≤ 30
Monthly median for plants
95th percentile
N Results BOD5 limit = 30
Monthly median for plants
95th percentile
N Results BOD5 limit ≤ 30
Monthly median for plants
95th percentile
N Results cBOD5 limit = 25
Monthly median for plants
95th percentile
N Results cBOD5 limit ≤ 25
Monthly median for plants
95th percentile
a
All POTWs in this report have <6 combined sewer outfalls and treat 10
MGD or more.
Source: U.S. EPA, Report on the Performance of Secondary Treatment Tech-
nology (8).
BOD5 (cBOD5) of 25 mg/liter also apply to many facilities
nationally in lieu of BOD5 limits.
These data suggest that Publicly-owned Treatment Works
(POTWs) generally comply with effluent standards for TSS,
BOD5 and cBOD5. Carbonaceous BOD5 is merely the BOD5
measured without nitrogenous BOD, that is, the BOD5 anal-
ysis conducted with a nitrification inhibitor. There have been
a number of performance indicator systems proposed for the
satisfactory operation, reliability, and durability of WWTPs.
Some are fairly elaborate and designed for long-term assess-
ment of satisfactory design, treatment, safety and personnel
management. (9, 10)
OBJECTIVES OF TREATMENT
It may be useful to think of WWT from a regulatory stand-
point—removing pollutants from the WW and preventing
damage to the water quality (WQ) in the RW are its primary
objectives. These are accomplished in a number of ways, but
most are included in regulations and the permits that govern
dischargers. There are different measures of their efficiency.
Some pollutants are conservative—they are merely moved
among the various discharges. For example, for phosphorus,
the amount of P in the influent will equal the amount of P
in the FE plus the amount in the biosolids. Removal or trans-
formation can occur in many processes but always results in a
constituent that is entrained in the aqueous or solid effluent
or, to a much lesser extent, may be released to the atmosphere
as a gas or vapor. From a plant operator’s point of view, how-
ever, the paradigm used to solve problems resembles that used
by an auto mechanic where the symptoms of system failure,
informed by experience and authority, are used to inform
changes to correct failures or improve performance, testing
and repeating until the device or problem is fixed. The level
of sophistication in these attempts in WWT varies with the
discharge requirements of the plant owner and, probably
3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-3
more important, the resources of the owner or the area where
the plant is or is to be located. For assessment of WWTPs
mg/liter some measures remain much as they have been for decades,
688 whereas for many of the more intractable assessments—
8
pathogens in FE, need for and efficacy of advanced treatment
regimes, and ultimate fate of biosolids—new techniques and
20 methods are becoming available.
854 Table 2 lists general types of pollutants, their effects on
7.5 RWs, and general measurement methods applied to estimate
19.0 their occurrence, density, or concentration. Table 2 does not
363
list the newer, metagenomic methods, covered in depth else-
where in this manual.
9.2 Pathogens and indicators—Historically, in WWT, indi-
24.0 cators are used because direct measurement of the indicated
451 species or constituent is impossible, too expensive, or too dif-
9.1 ficult for routine use. Indicators do have the benefit of almost
23.0
100 years of experience in use as measures of disinfection
effectiveness. Pathogens, of course, are a real public health
452 problem if present in RWs (11).
5.2 Oxygen demand or depletion—The discharge of any oxygen
15.0 depleting organisms, aggregates of organisms, or oxidizable
688 material into RW will rapidly have negative effects on those
4.0
waters since desirable organisms in the RW are aerobic and
oxygen saturation of water is low at physiological tempera-
13.0 tures, ≤25°C. Typically, assuming a WWTP is performing
adequately in this regard, damaging discharges will be the
result of plant upsets or failures. Measures of the efficacy of
WWT based on oxygen depletion have a history at least as
long as the use of microbial indicators, and this takes into
account the impossibility of classifying or accounting for
the densities of all organisms that might be involved in
oxygen depletion and all biodegradable organics in the FE.
The BOD test is an aggregate method. There are a number
of BOD tests in use, differentiating carbonaceous and nitro-
genous oxygen utilization and using various times to comple-
tion but the commonest by far is the 5-day test (BOD5).
There exist, in addition, the chemical oxygen demand test
(COD) and the total organic carbon (TOC) analysis. Both
have their place and analysts have used empiric relationships
between BOD5 and COD results in the same waste stream to
shorten analytical times for process control. In sum, COD
includes organics in forms not easily biodegraded and tends
to be higher than BOD. An excellent description of BOD
methodology and discussion is given by Pipes and Zmuda in
the last edition of this manual (12).
N and P—The concentrations of N and P in FE are critical
to the biosphere in RWs. Microgram per liter concentrations
in RWs can cause excess growth or blooms of undesirable
species. Although it is feasible to measure amounts of N
and P in FE, this illustrates a caveat for assessments of the
effects of FE on RWs—the effects of any constituent in any
FE will depend in large part on the quality and condition of
RWs. For this reason the algal growth assay provides reliable
information on specific FE (and the species of N and P
therein) and its effect on a particular RW.
Toxic constituents and pharmaceutical and personal care prod-
ucts (PPCPs)—Any FE may have any number of constituents
that are toxic, but their expression is the result of many vari-
ables besides occurrence or concentration. These will include
the susceptibility of sensitive organisms, interactions, and
synergies between FE and RW and other physical and chem-
ical attributes of the RW. While it is nominally possible to
identify and quantify all toxic constituents in a specific FE
(since we may verify and quantify the existence of the constit-
uent or precursors in the influent), it is neither practical nor
sensible to do so. There exist a number of biological toxicity
3.1.3-4 ▪ WATER

TABLE 2 Pollutants and measurements of efficiency of WWT

Class of pollutant Effect on RW or subsequent use of RW Measurement technique
Microbial pathogens RW becomes unsuitable for human Microbiological
consumption or recreation In disinfection (below, and other chapters)
Indicator organisms May signal possible presence of fecal Microbiological
pollution or pathogenic organisms In disinfection (below, and other chapters)
Oxygen-demanding, biodegradable Deplete available O2 in RW Microbiological assay
organic compounds (BOD5 test)
Chemical (COD test)
Or TOC test
Classic nutrients (N and P) Stimulate growth in RW, principally algae Chemical
and aquatic plants of concern Microbiological—algal growth test
Toxic constituents Kills or interferes with reproduction or Microbiological toxicity tests
Including pharmaceuticals and growth of aquatic life Biological toxicity assays
personal care products (PPCP) Specific chemical tests
Settleable solids Physical—creation of sludge banks Physical
Biological—smothering benthos leading
to anaerobic benthic conditions
Fats, oils, and grease Surface films interfering with surface Chemical or physical extraction
phenomena Not covered

assays that may be applied to FEs and RWs and collectively
called whole effluent tests (WETs). Though these are not
normally part of NPDES permits (National Pollution Dis-
charge Elimination System), each such permit must have a
narrative that references the state water quality standards for
RW. There are a number of plants that have such requirement
in their permits and there are states that require dischargers,
usually large dischargers, to sample for either specific chemi-
cals or perform WETs.
POLLUTANT REMOVAL AND REDUCTION
Specific pollutants are generally conservative, while measures
of the efficiency of treatment are often ratios of a parameter
in the influent and effluent to a process or the WWTP as a
whole. We mentioned phosphorus as a conservative pollutant
where:
Ctotal ¼ Ci þ Ce
 
Ci  Ce
%Efficiency ¼ 100 
Ci
Waste Load Reduction ¼ ðCi  Ce ÞðQÞð f Þ
Where Ci and Ce are the concentration of a pollutant in the
influent and effluent, respectively, Q is WWTP flow rate, and
f is a unit conversion factor. For example, when Q is in MGD
and the concentration is in mg/liters, then waste load reduc-
tion can be in g/day or lb/day. Then, the percent efficiency
and waste load to an RW from a plant with 20 mg/liters of
P in the influent and 1.3 mg/liters in the effluent and a daily
flow of 1 MGD would be:
ð20 mg=l  1:3 mg=lÞ
%Efficiency ¼ 100  ¼ 93:5%
20 mg=l
1:3mg kg
Waste Load ¼ 1 MGD  ¼ 4:921
L day
Conversely, parameters or measures of process efficacy,
though often called “removals,” are not. For example, the
BOD5 of a waste stream will (hopefully) be reduced through
treatment, but BOD is not a constituent but a parameter—a
property of the organisms that consume the components
of the complex organic waste—and it is more proper to refer
to these changes as reductions. Compounding this problem is
the fact that most regulatory agencies include FE BOD5 in
permits and that WWTPs are designed and operated in part
on BOD5 measurement. In design, especially, experience
with processes and about a century of empirical studies have
resulted in designer and operator trust in generalized relation-
ships of temperature, time, dissolved oxygen and volatile
suspended solids and treatment units. Volatile suspended
solids are a gravimetric measure of organic material in the
solids in WW.
WWT AND SLUDGE TREATMENT
WWTPs use two partially serialized processes, one for WW
and one for sludge (or, more currently, biosolids). In micro-
bial treatment of WW, sludge is either settleable material,
for example, most TSS in the plant influent are organic and
macroscopic materials—paper, food particles, hair—and
most are removed in primary clarification. “Biosolids” are
more properly the accretions of microbes that grow in the
processes, digesting and transforming the waste into cells
and metabolites. Neither the material nor the terms are exclu-
sive; we use “biosolids” for both to follow current usage but
with the caveat concerning nonexclusivity.
Many of the pollutants of concern are removed in the
processes and most likely will be associated with the biosolids
fraction and the latter are normally subjected to separate treat-
ment intended to change the character of the biosolids,
reducing the volume and nature of the solids with no refer-
ence to WQ. These treatments are generally called digestion
and may be aerobic or anaerobic. The reasons for further bio-
solids treatment are the biodegradation of some of the solids as
well as changing the solids’ characteristics so that (a) they are
oxidized very slowly after discharge (stabilization), and (b)
they thicken and dewater more easily. Sometimes stabiliza-
tion is accomplished chemically by altering the pH of the
sludge to >10 often with lime (lime stabilization). Typically
the aqueous phase from digestion is recycled to the head of
 3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-5
the plant so it may be diluted and treated again. In short,
the focus of biosolids treatment is on the characteristics
of the biosolids, not WQ.
Disposal of biosolids to land are regulated under 40 CFR
Part 503, Standards for the Use or Disposal of Sewage Sludge.
As to the effects and hazards of biosolids, Eisenberg and col-
leagues (13) develop and present a risk-based method to
assess risk to human health from exposure to pathogens in
biosolids. They utilized historical data from a number of treat-
ment plants and modeled single and double digestion in mes-
ophilic anaerobic units with and without lime treatment.
They used Class A biosolids (then, for viruses, <1 plaque-
forming-unit [PFU]/g, currently <1 PFU/4 g biosolids [dry
weight]) and used sample densities of 99% below the Class
A level. They found that their model, adapted for aerosol,
groundwater, and direct exposure as well as for occupational
and residential populations, provided estimates of occurrence
of rotavirus that could be verified in environmental samples.
Of particular importance is the possibility of using the models
for the typical environmental scenario where most samples
contain no detectable pathogens (“nondetects”).
PROCESS CONTROL
Oxygen Utilization
Most large WWTPs use aerobic processes. In these processes
oxygen is consumed by assimilation of organic material into
cell mass, endogenous respiration, predation, nitrification,
and algal growth. There is an assumed relationship between
oxygen utilization and organic compound removal that has
not yet been thoroughly modeled. Overall, BOD is the most
common measure of the efficacy of aerobic WWT processes.
Most WWT is performed in aerobic systems because:
• It provides rapid treatment.
• It suffers fewer intractable operating upsets.
• BOD loading is one of the design factors for WWTPs.
• BOD is usually one of the measures of whether FE WQ
meets permit standards for regulatory authorities.
• BOD has a very long history and experience in measuring
the efficacy of treatment and to control the process(es).
• BOD is a suitable nonspecific measure of a very complex
microbiological system for transforming organic C into
biosolids and estimating its operational health.
• Most anaerobic treatments are designed for the removal
of specific pollutants or for constituents or waste streams
that are difficult to remove aerobically or where anaerobic
treatment provides some other benefit.
It is precisely the generality of BOD measurement that pro-
vides its utility in assessing the state of the system. Although
many different aerobic processes are available, BOD can serve
them all, as a measurement that is a useful, relatively easy way
to assess performance of individual processes and the WWTP
as a whole. Methodological problems with the analysis are
well known and can be accounted for and corrected. Results
may be compared among plants, processes, and periods. Over
time, a municipal WWTP (MWWTP) treating the “same”
WW (insofar as this is possible—changes in user habits and
discharges only vary slowly over time and other major factors
like temperature and rainfall will tend to be cyclical) will be
able to adduce a biosolids per mass unit of BOD5 and use this
in managing operations.
There are a great many BOD tests (instantaneous, 7-, 10-,
and 20-day), but the commonest measure used is the 5-day
test. This reflects the most reliable measure of C removal since
carbonaceous BOD is generally finished in that period and
nitrogen-transforming (nitrogenous) BOD occurs thereafter.
The addition of nitrification inhibitors essentially ensures
that the BOD5 in carbonaceous only and has come to be
called carbonaceous BOD5 (CBOD5).
Although for specific wastes BOD may be used to estimate
the amount of biosolids produced, it does not measure the
removal of anything. BOD is a measure of oxygen utilization
by the heterotrophic organisms (or the oxygen required by
organic material) in the treatment stream; in the absence
or inactivity of the culture oxygen is not consumed, in the
absence of organic material the culture will still consume oxy-
gen. It may be thought of as a measure of the oxygen demand
of the organic material in the waste stream that can be assimi-
lated by the microbiological culture, but it is not the same as
the organic material and the contribution of any particular
organic component to oxygen demand is essentially impos-
sible to judge because of the complexity of the WW itself,
both as influent and as activated sludge.
BOD is a bioassay and may be seen as an antediluvian
ancestor of metagenome analyses—sort of bucket molecular
microbiology; just as molecular methods focus on functional
gene arrays or operational taxonomic units and not specific
organisms, BOD ignores the constituent members of the
aggregation that provides the effect and uses a secondary
measure of the effect itself. Now, with metagenomic methods
we can begin to understand the nature and characteristics of
groups of organisms with specific characteristics or responsi-
ble for specific transformations. See (14–17).
Total Suspended Solids
While one of the typical permit limitations and a required
measurement in FE, TSS in effluents are typically different
from those in influents. TSS in the FE of even a marginally
effective plant is almost entirely microbial flocs. It is generally
not a useful measure of biological treatment, although
• TSS measurements in process streams (RAS, ML volatile
suspended solids [MLSS]) are often used to control treat-
ment in suspended growth systems.
• TSS measurements may also be used as a quick check on
biological processes since BOD5 takes 5 days for results
and few plants are equipped for respirometer (instantane-
ous oxygen demand measures) analyses. These methods
require a period during which empirical relationships
for the specific plant and waste stream are developed.
• Some operators use what is often called an oxygen uptake
rate (OUR) where diluted MLSSs are saturated with air
and oxygen utilization is measured over a short time
(<1 h) and standardized against the MLVSS. This can
be compared with “normal” OURs for the type of plant
and the “normal” OUR (developed over time for the proc-
ess) for the plant, and a significant change may suggest
the entry of a toxicant or a change in the waste stream
or the process microbiome.
• Others may use TSS measures to estimate BOD of proc-
esses, between treatment units or in the FE (18).
The characteristics of the TSS are also very important;
the occurrence of filamentous or bulking sludge ( poor set-
tling characteristics) can affect all downstream processes,
including disinfection (19), and may signal a surfeit of partic-
ulate material in biological treatment flocs, adding to the
hydrolysis needed and making the flocs less amenable to
separation (20).
3.1.3-6 ▪ WATER
PATHOGENS AND INDICATOR
MICROORGANISMS
The authors of this section in the third edition (12) noted a
number of emerging issues that remain cogent. They noted
“(i) the need to identify alternate indicators of pathogens in
WW, FE, RW, and biosolids; (ii) the need for better methods
to detect pathogens themselves in WW, FE, RW, and biosol-
ids; (iii) the challenge antibiotic-resistant bacteria represent
to the WWTP industry; (iv) the need to develop bioassays
to study the removal of endocrine disrupting and other
organic contaminants from WW; and (v) research needs asso-
ciated with the beneficial use of WW sludge or biosolids.” All
of these have been addressed in the intervening period, and
much use has been made of metagenomic and metabolomic
methods to identify the existence and abundance of specific
groups of organisms in WWTP communities.
Assessing Disinfection and the Occurrence
of Pathogens
The absence of pathogens from the FE is an important
measure of the effectiveness of WW treatment but is rarely
determined. In any event, the efficiency of treatment is not
calculated; it is not necessary to determine the total number
or percentage of the pathogens inactivated, nor is it germane.
Treatment success is evidenced by the absence of pathogens
in the FE and in the RW. There are, of course, the problems
of changes on the target organisms due to disinfectants and
disinfection (21) and the primal concern of monitoring oper-
ation of disinfections units (22). Some advances have been
made in developing probes for virulence (23) and in using
molecular methods for identifying causes of waterborne dis-
ease outbreaks (24) or isolating frank pathogens from water
matrices (25). There is also some concern about protection
of pathogenic organisms by higher organisms in environs
downstream of the FE (26). Excellent reviews of pathogen
molecular identification techniques are found in (27) and
for biosensors in (28), both in (29). A sampling of studies
follows.
Enumerating Pathogens
Bibby, Viau, and Peccia (30) studied PCR amplicons from
WWT biosolids treated by mesophilic and thermophilic
phased anaerobic digestion, composting, and soils to which
biosolids were applied. They found similarities in biosolids
from the same treatment scheme but biosolids from different
stabilization methods were different. Based on Jukes-Cantor
distance (0.03), all samples contained pathogens and most
belonged to the genera Clostridia and Mycobacteria. They
suggest that the technique might help screen samples for
subsequent analysis and ensure that pathogens are not left
out of risk assessment.
Carbon nanotubes and Field Effect Transistors (FETs)
were used by Villamizar et al. (31) to identify Salmonella infan-
tis. Antibodies are adsorbed onto the nanotubes and in this
study detected 100 CFU/ml in 1 h. Streptococcus pyogenes
and Shigella sonnei did not interfere at densities of 500 CFU/
ml. Similar devices could be used with suitable antibodies
for other bacteria or viruses.
The occurrence and risk of pathogens in sediments of
streams below combined sewer outfalls (combined collection
systems for sanitary and storm water flows) and WWTP
have also been studied (32), and some studies offer an addi-
tional insight into particle association of pathogens (33). A
caution on the use of indicators (and pathogens, when meth-
ods become usable and available) is provided by Gin and
Goh (34), who tested viable but not culturable (VBNC)
enterococci in cultures of 107 CFU/ml, stressed with dif-
ferent levels of light and salinity. At low light intensities
(<20 W/m2 or 200 mW/cm2), for all salinities, VBNC cells
can be expected to last for extended periods. A qPCR techni-
que for human adenovirus and polyomavirus was tested for
co-occurrence with norovirus in WWTP influents, effluents,
and RWs and was considered apt as a conservative indicator
of contamination with FE and for norovirus, with the reserva-
tions normal for indicators (35).
Alcalde, Folch, and Tapias (36) monitored removals of
indicators (bacteria, viruses, and spores of sulfite-reducing
Clostridia as an analog for protozoa) at a conventional acti-
vated sludge (CAS) plant with advanced treatment for water
reclamation. They found similar removals for E. coli and fecal
enterococci in both the secondary effluent and in the filtered,
chlorinated, reclaimed water effluent. Neither of the bacterial
indicators was sufficient in itself for indication of viruses or
protozoa for the reclaimed effluent (filtered, chlorinated)
nor for viruses in the secondary treatment effluent.
Chern and colleagues (37) compared qPCR for markers
for E. coli, Enterococcus, Clostridium spp., Bacteroides, total
Bacteroidales, as well as the human-associated Bacteroides
markers, HF183 and HumM2. Culturable enterococci,
E. coli, Bacteroides, and C. perfringens were also determined.
Both undisinfected and disinfected effluents were collected
from plants using chlorination or UV for disinfection of
effluents for four seasons. All plants and both forms of disin-
fection showed significant reduction of culturable indicators.
There were no changes in the densities of genetic markers
in effluents disinfected by either chlorine or UV and they
found that the use of such densities provides no useful infor-
mation as to efficacy of disinfection.
Cheng et al. (38) studied the frequency of occurrence
of noroviruses, Enterococcus faecalis and Enterococcus
faecium in influent, final effluent, and biosolids from four
secondary wastewater treatment plants with undisinfected
effluents in northwestern Ireland. They found reduced
frequencies in effluents and in most biosolids. Norovirus
genotype 2 occurred in all seasons in most effluents and
biosolids. Only E. faecalis was found in biosolids in winter
in a single plant. This study underscores the occurrence
of indicator and possible pathogen removals in AS systems
and the need for care in the discharge of effluents including
biosolids.
Simmons and Xagoraraki analyzed 30 wastewater and
6 biosolids samples from five WWTPs in Michigan (39).
They sampled influent, pre- and post-disinfection, and bio-
solids from each plant using qPCR and cell culture for viruses
and documented releases into the environment.
METHODS OF DISINFECTION
Disinfection in WWTP is still almost exclusively effected by
chlorination. Currently, permit requirements and routine
monitoring continue to require culture methods and a few
indicator organisms, specifically fecal (thermotolerant) coli-
forms and E. coli or enterococci for FE and enterococci for
RW in most places. Although operators may use alternative
indicators and molecular techniques to assess operations,
rules continues to require culture methods for compliance
reporting. Relatively liberal agencies or policies allow substi-
tution of other methods after significant demonstrations of
equivalence with approved culture methods. However, there
are a number of alternative methods in use and a sampling is
discussed next.
 3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-7
Solar Radiation
Generally, although reactions are too slow for application to
WWTPs, the effectiveness should be noted since it may be
applied to specific small plants or for low usages of effluents.
For example, solar disinfection can be used to reduce protozoa
in water (40). Bichai, Polo-López, and Fernández Ibañez (41)
showed that solar disinfection could be effective against
E. coli at densities of 104. In their study FEs were used for irri-
gation. They also found that compound parabolic collector
reactors were superior to standard SODIS polyethylene
terephthalate bottles.
UV Radiation
The only readily available method to assess UV operation is
measurement of radiation flux in the receiving stream. A
number of alternative assessments involving measurement
of inactivation or indicators have been offered. In a WEF
2005 report (42), Blum evaluated two culture-independent
techniques, specifically to evaluate the efficacy of disinfection
in inactivating fecal coliforms and relate differences to the
physiology of the organisms. In addition he investigated the
effect of those differences in the recovery of those organisms
from disinfection. He found that two proteins (used in a pro-
tein profile analysis), one each associated with organisms
in the growth and resting phase were associated with sensitiv-
ity to disinfection; cells with more of the growth-phase pro-
tein were associated with cells from relatively oxygen-rich
areas or treatments (he used effluent from a trickling filter)
and were easier to inactivate, whereas organisms from a low-
oxygen environment (lower areas of clarifiers) were more
difficult and were more capable of being resuscitated. He
also notes the equivalence of disinfection by chlorination
or by heating. This work was undertaken assuming organisms
in the vegetative state.
Blum intentionally dealt with vegetative organism only,
and, unsurprising to most operators, UV inactivation of efflu-
ents from AS depends in large part on the nature of the flocs.
Azimi, Allen, and Farnood (43) suggested a model for the
effect of AS biofloc densities on the efficiency of UV disinfec-
tion and showed a relationship between the size and densities
of flocs and UV efficiency. Managing the biological flocs in
WWTP could, as Blum hoped, improve the effectiveness of
UV disinfection in WWTPs and physical disruption of flocs
may be required for best results. In the same vein, Aziri and
other researchers posited and refined a mechanism (43, 44)
for floc formation and its effects on UV disinfection. In
(44) they evaluated the additive effect of polyphosphate on
inactivation by UV of organisms associated with flocs. Others
(45–47) have noted the effects of polyphosphates, both nat-
urally occurring and additive, on disinfection. Following the
effect of floc formation and WWTP process control, Azimi
and others evaluated the flocs from a CAS system and a
nitrogen-removal process in series on disinfection and organ-
ics removal (48). In addition, see the section on ozone for
effects of floc on both ozonation and chlorination.
In a study of norovirus using qPCR and F-specific RNA
bacteriophage GA using both qPCR (total concentration)
and a plaque assay (infectious concentration) of a WWTP
using UV for disinfection researchers concluded that qPCR
in isolation underestimates the reduction of infectious virus
during wastewater treatment (49).
In a study of UV disinfection of WWTP effluent for
tetracycline-resistant heterotrophs and E. coli (50) found
that resistant heterotrophs and E. coli were inactivated.
Although E. coli did not undergo dark repair, resistant
heterotrophs did, so, while UV is useful for disinfection, O3
or Cl might provide better disinfection results over time.
This possibly points to the major reservation or concern
around UV disinfection: the presence of a multitude of
pathogens (i.e., bacteria, viruses, worms, protozoa) and the
potential health risk of the possible photo and dark repair
of reversibly damaged bacteria. However, the combination
with electrolytically generated oxidants may add another
layer of protection and, in preventing or reducing biofilm
formation on the quartz sleeves of lamps, reduce one of the
major concerns in UV disinfection of WW (51). Another
possible concern arose as a result of the formation of nitro-
sodimethylamine (a potent carcinogen) when WW with nat-
urally occurring organic material or nitrates was subjected to
UV disinfection (52).
Dong Li et al. showed that spores from aerobic spore
forming (ASF) organisms did not associate with particles
(essentially AS flocs) in FE to provide shielding from UV
inactivation. They compared homogenized and unhomogen-
ized portions and both cultured (Bacillus subtilis, ATCC 3633)
and indigenous aerobic spores. There was no difference in
effectiveness between homogenized and nonhomogenized
portions and both achieved 1.5 log reductions at 24 mJ/cm2
and 1.7 log at 40 mJ/cm2. They suggest that ASF organisms
might serve as an indicator to assess the efficacy of UV
disinfection for Giardia and Cryptosporidium in FE (53).
Hedrick and colleagues showed the effectiveness of UV doses
against actinospore or triactinomyxon stages of Myxobolus
cerebralis (54).
Ozone
One of the important problems with ozonation is estimat-
ing O3 dosage and hydroxyl (OH−) formation in wastewater.
Some sort of indicator system is one option and a second,
differential UV254 absorbance (ΔUV254) and total fluores-
cence (ΔTF) were studied. Researchers developed empirical
correlations for these two radiation measures and inactivation
of three microbial surrogates, E. coli, MS2, and Bacillus subtilis
spores, in nine lab systems with wastewater and eight waste-
waters at pilot and plant scales (55). They note that online
ΔUV254 and ΔTF analyzers are readily available—which
would allow use of the empirical relationships to adjust
treatment—and that the same method could be extended
to other, advanced oxidation systems like UV-peroxide. In
the disinfection of flocs ozone was found to be less effective
in penetrating flocs than chlorine and concurrent or sequen-
tial use might be useful (56).
Miscellaneous Alternatives
Metallic ions (Fe3+ and Al3+), used in conjunction with per-
acetic acid (CH3CO3H) has proven to be an effective agent
against indicators and some pathogens in treating tertiary
effluents (57). “Tertiary” refers to WWTPs with processes
beyond the secondary, nutrient removal, for example. Perace-
tic acid is less effective against spores, viruses, and protozoa
than other agents (58).
Rodriguez-Chueca et al. experimented with RF-induced
heating of entrained Fe species in assaying the efficacy of
Fenton-like processes with and without hydrogen peroxide
in inactivating E. coli and Enterococci spp. in simulated FEs.
Fenton processes refer to processes with hydrogen peroxide
(H2O2) at pH values lower than 4. They were hopeful, report-
ing inactivation of 3–4 logs10 for E. coli but noted the need to
experiment in real-world plants (59).
Bandala et al. (60) discussed disinfection and organics
removal using Fe(VI) as FeSO42− showing impressive results
3.1.3-8 ▪ WATER
in concurrent removal of over 40% of organic material
as TOC and 99.9% of total and thermotolerant coli-
forms at 5–7 mg/liters of FeSO4−2 in 5 min for the bacteria
and 10 min for filtered samples for organics removal.
High-frequency ultrasound has also been studied and
shows promise for bacteria and yeasts, though yeasts were
less affected. Inactivation was thought due to the formation
of free radicals and HO2 generated by cavitation (61). Nano-
filtration and reverse osmosis (RO) have also been studied as
have coagulation, powdered activated carbon and sand, or
dual-media filtration with fair success (62). All these proc-
esses need relatively clear FEs or systems that are readily
cleared from occlusion.
NUTRIENT REMOVAL
Almost all techniques for enhanced removal require addi-
tional oxidation, often in conjunction with anoxic stages.
Most assessment of removals may be made by direct measure-
ment of N or P. However, the effect of the FE on the RW is
best assessed by the algal growth test due to the many, poorly
understood factors controlling toxicity in RW.
Advanced WWT may not be uniformly efficient. In an
excellent review of antibiotic removals from urban WWTPs
Michael et al. (63) noted the general superiority of advanced
oxidation but note the difficulty in comparing removals even
from plants with nominally the same treatments due to the
lack of detailed understanding of the processes and condi-
tions in each. They also note that careful monitoring and
operation are required.
TESTING FOR EFFLUENT TOXICITY:
WHOLE EFFLUENT TOXICITY
This is, of course, the primal measure of WWT—the effect
the effluent might have on the RW. The full-scale bio-
assays tend to be very complex and quite time and labor inten-
sive. The standard reference is (64) and, again, the reader is
referred to Pipes and Zmuda for an excellent survey (12).
Are bioassays and useful in categorizing intoxicants but extre-
mely time consuming and difficult. There are saline and
freshwater assays and organisms including algae, plants, fish,
bacteria, and insects. An example is the red algal sexual
reproduction toxicity test. Parallel stock cultures of Champia
parvula must be maintained for testing in carbon-stripped
culture water. Nutrients are added regularly and water is main-
tained at 23°C ± 1°C and salinity at 30 pph ± 2 pph, aerated
weekly while standing mass is reduced ∼50%. Analysts
must change alternate culture’s media weekly and new cul-
tures could be started at this time with 1 cm branch tips.
Nominally, there should be enough material for test after 3
weeks. Under test the photoperiod is 16 h light, 8 h dark at
500 ft-candles (75 µE/m2/s). After inoculating test culture
Erlenmeyer flasks (1 liter glass acid-stripped and sterile
with 800 ml culture medium). The algae are inoculated five
female and one male branch tips into each of five sets of a
dilution of FE and RW and incubated for at least 2 days.
Then growth of new branches is verified and they are counted
and related to the concentration of the FE. There are com-
mercial tests (e.g., Environmental Bio-detection Products
Algaltoxkit f, which uses the standard Selenastrum capricornu-
tum now called Pseudokirchneriella subcapitata) that include
the test strain and can be completed in about 5–7 days with
no need to maintain cultures.
There are also a number of bacterial assays, including and
similar to Ames’s tests. For example, one test uses an
engineered strain of Salmonella typhimurium and is equivalent
to ISO 13829 for interference with the production of beta-
galactosidase.
Another newer possible assay is based on recent work and
suggests that FEs might actually reduce diversity in RWs.
Drury et al. (65) sampled two streams, one with industrial
effluents as well as a large WWTP (245 MGD) and one
with a small WWTP (5 MGD); neither effluent was dis-
infected. They sampled the streams above and below the
outfalls of the plants. Downstream the rivers had higher
inorganic nutrients but much reduced sediment bacterial
communities. While the two streams were significantly differ-
ent above the discharges, they were indistinguishable below.
This might suggest that the effluents were toxic to benthic
organisms, although this was not tested—an example of a pos-
sible indicator of effluent toxicity. Finally, the EPA has pro-
duced software using QSAR (quantitative structure-activity
relationship) techniques to estimate toxicity for any com-
pound or molecule (66).
EMERGING ISSUES
PPCPs and Antibiotic Resistance in Wastewater
An excellent review article is (67), in which new molecular
and microscope methods are detailed. There is a new standard
method for PPCPs, Section 6801, which assays for 13 PPCPs:
trimethoprim, acetaminophen, caffeine, primidone, sulfame-
thoxazole, fluoxetine, carbamazepine, naproxen, diclofenac,
ibuprofen, bisphenol A, gemfibrozil, and triclosan, which
were among those suggested in a WERF research project
(68). The WERF products were selected as indicators for
this class of compounds, and their removal efficiencies were
assayed in standard and advanced WWT processes. With
the typical caveats concerning indicators, this is a great first
step in producing a method to allow quantitation of occur-
rence of these compounds in WWTP flows. Indicators for
PPCPs were evaluated, and the authors proposed a suite of
indicators that could be used to estimate the efficacy of partic-
ular treatments in removing PPCP from WWT FE and bio-
solids. The project identified 25 candidate compounds
based on frequency of occurrence and concentration and exis-
tence of methodology for analysis that were removed by bio-
transformation or sorption, although most could be removed
by both. Toxicological importance was not a primary crite-
rion. Of these, 22 indicators were chosen: acetaminophen,
atenolol, benzophenone, bisphenol A, caffeine, carbamaze-
pine, cimetidine, DEET, diphenhydramine, fluoxetine, gem-
fibrozil, ibuprofen, iopromide, meprobamate, naproxen,
sucralose, sulfamethoxazole, TCEP, TCPP, triclocarban, tri-
closan, and trimethoprim. The indicators were evaluated
at a number of conventional treatment plants and plants
with various forms of sludge treatment. Mass balance analyses
for the indicators generally was good, and factors affecting
removals were those identified by others, sludge retention
time (the time that AS was retained in the aerobic system),
hydraulic retention time, temperature, and pH. Biotransf-
ormation and sorption were developed or verified in the
study. The study suggested a readily available fate model
(ASTreat—unfortunately no longer available, but the EPA
still offers ECOtox) to aggregate and help in assessing the
removals at specific plants. WERF generally uses the acronym
TOrC—total organic compounds.
This fits nicely with an earlier project, Development of
Diagnostic Tools for Trace Organic Compounds and Multiple
Stressors (69), which is part of a WERF undertaking to
 3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-9
characterize and identify risks from TOrCs. In this study, a
number RW bioassays are suggested, and the authors suggest
that the reaction paradigm they propose be used retrospec-
tively, that is, that following identification of morphological
or other gross changes in a stream’s biota the factors identified
in the report be used to assess the likelihood that the changes
noted might be the result of TOrCs. Both projects helped cre-
ate a WERF database on TOrCs at (traceorganicsecotool.
werf.org). Users “may store, query, and search TOrC data,
as well as biological and aquatic life habitat information for
sites in the U.S.”
A survey of EPA sludge samples from 2001 showed the
occurrence and frequency of PPCPs was not too different
from 2006–2007 samples. The researchers used five mega-
composite samples, and this suggests that a metagenomic
method may be acceptable for survey and monitoring (70).
A survey of hospital and psychiatric center effluents (71)
for 100 pharmaceuticals with concentrations based on usage
data provided by the hospitals. There were differences in the
two types of hospitals. The authors point out that of the four
pharmaceuticals likely to be found most often in these efflu-
ents, only one (diclofenac) has experimental fate and ecotox-
icity data. They suggest that pretreatment of hospital wastes
might be beneficial and suggest sorption rather than urine
separation.
Researchers looking at AS used in ammonia removal
(nitrification reactors) found 75 amoA sequences from five
WWTPs operating with low dissolved oxygen levels and
long retention times. In one of the four major phylogenetic
clusters amoA sequences were “virtually identical” to those
from all five activated sludge samples (from Oregon, Wiscon-
sin, Pennsylvania, and New Jersey) (72). Also, in a study of
risk assessments of land application of biosolids (73), the
authors suggest that their findings be used to prioritize further
study and that effective pathogen removal or inactivation for
biosolids is the most effective method to reduce exposure and
infectious risk rather than the common increased distance or
offset of application from inhabitants. Some other methods
for removals were reported on. For example, the removal of
ofloxacin by solar-driven Fenton’s technique catalyzed by
TiO2 (63).
An accurate assessment of the pollutant load in WW, FE,
RW, and biosolids is still not currently possible for all pollu-
tants. However, a sobering thought is that “the largest part
of the Earth’s microbial biomass is stored in cold environ-
ments, which represent almost untapped reservoirs of novel
species, processes, and genes. In this study, the first metage-
nomic survey of the metabolic potential and phylogenetic
diversity of a microbial assemblage present in glacial ice is
presented. DNA was isolated from glacial ice of the North-
ern Schneeferner, Germany. Pyrosequencing of this DNA
yielded 1,076,539 reads (239.7 Mbp)” (74).
Both land application and landfilling of sludge can pro-
vide reservoirs of protozoal pathogens, although disinfection
can be effective (75, 76). The microbial community in AS in
south India was studied by metagenomics. The major domain
were prokaryotes, mostly bacterial from nine phyla, and most
genes for the metabolism of aromatics led to an experiment
with salicylate-induced sludge that improved COD reduction
(77). One study examined the occurrence and loss of
antibiotic-resistance genes (ARGs) and bacteria (ARBs) in
the effluents and biosolids of different treatment plants
(membrane biological reactor, AS, oxidative ditch and rotary
biological contactors) and sludge treatments (dewatering,
gravity thickening, anaerobic digestion, and lime stabi-
lization). Influents, pre- and postdisinfected effluents, and
biosolids were analyzed for tetracycline-resistant genes, sulfo-
namide resistance gene, and tetracycline- and sulfona-
mide-resistant bacteria. MBR effluents were lower in these
than in conventional treatment, and disinfection by UV
and chlorination did not significantly reduce ARGs or
ARBs. Advanced biosolids treatment—anaerobic digestion
and lime stabilization—were more effective in reduction of
ARBs and most ARGs (78).
Exciting advances in eliciting metagenomes from specific
environments have and hopefully will continue to help in
the assessment of the very complex WWTP processes in FE
and in RW (79–83). Validation and QA/QC for these meth-
ods is becoming available, for example, ALE in (84). The
software is available at http://www.alescore.org. However,
these methods are still extremely difficult even with improved
methods of characterizing metagenomes. Even in potable
water they are extremely difficult; Pinto et al. note “The bac-
terial community in all samples was taxonomically diverse
and consisted of a total of 4369 operational taxonomic units
(OTUs) at a 97% similarity cutoff.” OTUs varied with loca-
tion in the distribution system and season (in Ann Arbor,
MI) and were related to the microbiome in the raw water (85).
In a 2010 report publicizing the worldwide crisis in water,
the United Nations referred to “sick water” (86). Following
the analogy, WWT can be seen as the analog of medical
care and WWTPs as the environmental equivalent of hospi-
tals. Not unlike the medical trade, the methods and techni-
ques of WWT have burgeoned in means of treatment and
in monitoring processes. The future will include increased
use of new methods utilizing metagenomics and meta-
bolomics, which offer a level of gross understanding of the
extremely complex matrix that is wastewater and wastewater
under treatment. As we come to understand what the interac-
tions between these metagenomes and specific pollutants,
in particular as we identify functional genome units, we will
approach nearly real-time measures of process removal effi-
ciencies. Citing a comment by Rossello-Morá and colleagues
using fluorescent in situ hybridization–microautoradiography
studying substrate uptake in halophiles that bacteria behave
differently in laboratory settings than in the environment.
Dazzo, Schmid, and Hartmann note that this is “a lesson
repeated often in microbial ecology” (87, p. 726). This is par-
ticularly apt for WWTP since halophile populations and
environments are likely to be much simpler than WW.
Advanced Treatments for PPCPs and
Antibiotic-Resistant Organisms
An interesting whole-plant assessment for PPCP removals
use life cycle assessment tools illustrated in (88). Igos
and colleagues (89) offer a good summary for nonspecific
PPCPs comparing general removals for centralized CAS
(CCAS), upgraded CCAS, centralized conventional (CC)
with O3, powdered activated carbon (PAC), and UV and
then distributed treatments at the sources (generally assumed
to be hospitals). In this methodology only removals (reported
or measured for a specific site) and environmental cost
factors—electrical use, H2O2, for example, and removals for
the specific technique are compared. Removals reported for
10 pharmaceuticals were used along with real plant data
for CCAS and the distributed WWTPs. Of the methods
studied, UV was the least advantageous while upgrading a
CCAS was the most. This may be a good approach to value-
less assessment utilizing the indicator suite noted in the
WERF 2012 study. A similar study looked at different types
of tertiary treatment for P removal including high-rate
3.1.3-10 ▪ WATER
sedimentation, microsieve, dual media filtration (all with UV
disinfection), and polymer ultrafiltration or ceramic microfil-
tration membranes. While membrane removals were best,
gravity techniques provided very good treatment at much
reduced environmental cost (47).
A study of estrogenic activity using cell-based reporter
gene assay by the Water Reclamation District of Greater
Chicago, one of the most complex in the United States,
found that WWTPs were only one of the major contributors
and that streams in highly urbanized areas above outfalls had
estrogenic activities near or above FEs (78).
The need for careful study of site or waste stream–specific
advanced treatment for PPCP removals is pointed to very
effectively by a recent study (90). The authors studied ozona-
tion and PAC with and without subsequent sand filtration on
16 TOrCs. Ozonation resulted in better removal of the spe-
cific organic contaminants (0.7 g O3/g DOC and 18 min)
than PAC (20 mg/liter for 60 min) and better genotoxic
effects. But Ames assays showed dose-dependent mutagenic-
ity for O3 and toxicity in trout assays; sand filtration reduced
this effect, but not to the level of untreated effluent. The same
was seen in effluents from four separate plants suggesting
that these results might be expected in any FE. A study
from Croatia (91) pointed up the removals in aqueous and
solids phases for modest removals in that country. Fourteen
pharmaceuticals (sulfadiazine, sulfathiazole, sulfapyridine,
sulfamethazine, sulfamethoxazole, n-acetylsulfamethoxazole,
trimethoprin, norfloxacin, ciprofloxacin, enrofloxacin, azi-
thromycin, dehydroerythromycin, clarithromycin, roxithro-
mycin) were studied in sorbed and liquid phases in influent
and effluents, and most removals were physicochemical (in
solids), from 2% to 70%. The highest biological removal
was about 55% for clarithromycin.
A similar study suggested that the ability of cultures grown
aerobically on anthracite provided removals of up to 95%
for organic micropollutants spiked at 2 ng/liter in tertiary
effluent and that these removals correlated with those on
granular activated carbon (GAC) and this might serve as
an indicator of bioavailability of these compounds during bio-
filtration on anthracite (92).
Endocrine-Disrupting Compounds (EDCs)
Removal of these compounds by “chlorination, ozonation”
(93), and UV radiation (94) have all been reported as effec-
tive in the removal of EDCs in wastewater effluents. However,
these treatment processes are also likely to produce unknown
disinfection by-products, some of which might exhibit a
degree of estrogenicity. Removal of dissolved estrogenic con-
taminants by adsorption onto solid sorbents represents an
attractive treatment alternative that might prove to be impor-
tant adjuvant in the control of these types of contaminants.
Adsorption, in which pollutants can be removed without pro-
ducing unwanted by-products, has unique advantages over
traditional wastewater treatment processes (95) and rein-
forced by Liu and Carr (96).
Limpiyakorn et al. (97) reviewed studies including those
that have used molecular methods to show the presence of
archaeal ammonia mono-oxygenase (amoA) genes in
WWTPs. They provide a protocol and illustrations of how
to use MAR-FISH (microautoradiography) in combination
with fluorescence in situ hybridization (FISH) to obtain
information about the ecophysiology of probe-defined single
cells in mixed microbial communities. They give an overview
of how to use MAR-FISH in various ecosystems and include
sampling, incubation with radiotracers, the MAR procedure
in combination with FISH, and other staining techniques.
They note that “microscopy and stable isotope probing of
nucleic acids are needed to better understand the activity,
function, abundance, and diversity of amoA-encoding arch-
aea in wastewater treatment plants.”
SUMMARY
Many of the hoped-for methods for assaying whole-biome
activities and new methods of classification of metagenomic
material have or are coming to fruition. Nowhere will these
advances be as welcome as in the design, operation, and
management of WWT and WWTPs. Traditional and tried
bioassays like BOD5 will continue to be important. In fact,
a European review of design for future implementation of
their wastewater standards continues to be based on BOD5.
However, as newer techniques become familiar and more
evidence is adduced to support their implementation as con-
trol strategies, we will come closer to understanding how—
and why—these systems work and how we can monitor and
improve them without the hit-or-miss strategies of the
past; while based on seemingly sensible approximations of
operative relations and causalities, the “seem” has often inter-
vened. As in all things, improved understanding allows
improved operation.
The authors are indebted to and highly recommend the notably clear
and cogent explanation of BOD and WET by past authors of this section,
Wesley O. Pipes and James T. Zmuda.
REFERENCES
1. Metcalf & Eddy. 2003. Wastewater Engineering: Treatment
and Reuse, 4th ed. McGraw-Hill, New York.
2. Spellman F. 2013. Handbook of Water and Wastewater Treat-
ment Plant Operations, 3rd ed. CRC Press, New York.
3. Hannah T, Wilner PE, Hubbell S, Stephanie L, Spalding
PE, Lynne E. 2012. Wastewater Treatment Plant Design
Handbook. WEF, Alexandria, VA.
4. ASCE and WPCF. 1977. Wastewater Treatment Plant Design.
ASCE and WPCF, Washington, DC.
5. U.S. EPA. 2011. Principles of Design and Operations of
Wastewater Treatment Pond Systems for Plant Operators,
Engineers and Managers. Division LRaPC, Laboratory
NRMR (eds). U.S. EPA, Cincinnati, OH.
6. Census. 2000. Summary File 1. US Census Bureau, Wash-
ington, DC [cited 2003]. Available from: http://factfinder.
census.gov/home/en/datanotes/expsflu.htm.
7. Goldstein R, Smith W. 2002. U.S. Electricity Consumption
for Water Supply & Treatment—The Next Half Century.
Topical Report 1006787. Electric Power Research Institute,
Palo Alto, CA.
8. U.S. EPA. 2013. Report on the Performance of Secon-
dary Treatment Technology. EPA-821-R-13-001, Contract
No.: EPA-821-R-13-001. U.S. EPA, Washington, DC.
9. Quadros S, Rosa MJ, Alegre H, Silva C. 2010. A perfor-
mance indicators system for urban wastewater treatment
plants. Water Sci Technol 62:2398–2407. doi: 10.2166/
wst.2010.526.
10. Silva C, Quadros S, Ramalho P, Alegre H, Rosa MJ. 2014.
Translating removal efficiencies into operational per-
formance indices of wastewater treatment plants. Water Res
57:202–214. doi: http://dx.doi.org/10.1016/j.watres.2014.
03.025.
11. Arnone RD, Walling JP. 2007. Waterborne pathogens in
urban watersheds. J Water Health 5:149–162.
12. Pipes WO, Zmuda JT. 2007. Assessing the efficiency of
wastewater treatment, pp 334–354. In Hurst CJ, Crawford
RL, Garland JL, Lipson DA, Mills AL, Stetzenbach LD
 3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-11
(eds), Manual of Environmental Microbiology, 3rd ed. ASM,
Washington, DC.
13. Eisenberg JNS, Moore K, Soller JA, Eisenberg D JMC Jr.
2008. Microbial risk assessment framework for exposure to
amended sludge projects. Environl Health Perspect 116:
727–733. doi: 10.1289/ehp.10994.
14. Bibby K, Viau E, Peccia J. 2011. Viral metagenome analysis
to guide human pathogen monitoring in environmental
samples. Lett Appl Microbiol 52:386–392. doi:
10.1111/j.1472-765X.2011.03014.x.
15. Basner A, Antranikian G. 2014. Isolation and biochemical
characterization of a glucose dehydrogenase from a hay
infusion metagenome. PLoS One 9:e85844. doi:
10.1371/journal.pone.0085844.
16. Batista-Garcia RA, Sanchez-Reyes A, Millan-Pacheco C,
Gonzalez-Zuniga VM, Juarez S, Folch-Mallol JL, Pastor
N. 2014. A novel TctA citrate transporter from an acti-
vated sludge metagenome: structural and mechanistic pre-
dictions for the TTT family. Proteins 82:1756–1764. doi:
10.1002/prot.24529.
17. Abe T, Kanaya S, Uehara H, Ikemura T. 2009. A
novel bioinformatics strategy for function prediction of
poorly-characterized protein genes obtained from meta-
genome analyses. DNA Res 16:287–297. doi: 10.1093/
dnares/dsp018.
18. Kumar KS, Kumar PS, Babu DMJR. 2010. Performance
evaluation of waste water treatment plant. Int J Engineer
Sci Technol 2:7785–7796.
19. Flores-Alsina X, Comas J, Rodriguez-Roda I, Gernaey
KV, Rosen C. 2009. Including the effects of filamentous
bulking sludge during the simulation of wastewater treat-
ment plants using a risk assessment model. Water Res 43:
4527–4538. doi: http://dx.doi.org/10.1016/j.watres.2009.
07.033.
20. Wang B-B, Peng D-C, Hou Y-P, Li H-J, Pei L-Y, Yu L-F.
2014. The important implications of particulate substrate
in determining the physicochemical characteristics of
extracellular polymeric substances (EPS) in activated sludge.
Water Res 58:1–8. doi: http://dx.doi.org/10.1016/j.watres.
2014.03.060.
21. Blatchley ER, Gong WL, Alleman JE, Rose JB, Huffman
DE, Otaki M, Lisle JT. 2007. Effects of wastewater disinfec-
tion on waterborne bacteria and viruses. Water Environ Res
79:81–92.
22. Gaki E, Banou S, Ntigkakis D, Andreadakis A,
Borboudaki K, Drakopoulou S, Manios T. 2007. Quali-
tative monitoring of tertiary treated wastewater reuse
extensive distribution system: total coliforms number
and residual chlorine concentration. J Environ Sci Health
A Tox Hazard Subst Environ Eng 42:601–611. doi:
10.1080/10934520701244367.
23. Hayes SL, Rodgers MR, Lye DJ, Stelma GN, McKinstry
CA, Malard JM, Vesper SJ. 2007. Evaluating virulence of
waterborne and clinical Aeromonas isolates using gene
expression and mortality in neonatal mice followed by
assessing cell culture’s ability to predict virulence based on
transcriptional response. J Appl Microbiol 103:811–820.
doi: 10.1111/j.1365-2672.2007.03318.x.
24. Hewitt J, Bell D, Simmons GC, Rivera-Aban M, Wolf S,
Greening GE. 2007. Gastroenteritis outbreak caused
by waterborne norovirus at a New Zealand ski resort.
Appl Environ Microbiol 73:7853–7857. doi: 10.1128/
AEM.00718-07.
25. Hoffman RM, Wolk DM, Spencer SK, Borchardt MA.
2007. Development of a method for the detection of water-
borne microsporidia. J Microbiol Meth 70:312–318. doi:
10.1016/j.mimet.2007.05.005.
26. Bichai F, Payment P, Barbeau B. 2008. Protection of
waterborne pathogens by higher organisms in drinking
water: a review. Can J Microbiol 54:509–524. doi:
10.1139/w08-039.
27. Cupples AM, Xagoraraki I, Rose JB. 2010. New molecular
methods for detection of waterborne pathogens, pp 57–94.
In Mitchell R, Gu J-D (eds), Environmental Microbiology,
2nd ed. John Wiley & Sons, Hoboken, NJ.
28. Ripp S, DiClaudio ML, Sayler GS. 2010. Biosensors as
environmental monitors, pp 213–233. In Mitchell R, Gu
J-D (eds), Environmental Microbiology, 2nd ed. John Wiley
& Sons, Hoboken, NJ.
29. Mitchell R, Gu JD. 2010. Environmental Microbiology, 2nd
ed. John Wiley & Sons, Hoboken, NJ.
30. Bibby K, Viau E, Peccia J. 2010. Pyrosequencing of the 16S
rRNA gene to reveal bacterial pathogen diversity in biosol-
ids. Water Res 44:4252–4260.
31. Villamizar RA, Maroto A, Rius FX, Inza I, Figueras MJ.
2008. Fast detection of Salmonella infantis with carbon nano-
tube field effect transistors. Biosens Bioelectron 24:279–283.
doi: 10.1016/j.bios.2008.03.046.
32. Donovan EP, Staskal DF, Unice KM, Roberts JD, Haws
LC, Finley BL, Harris MA. 2008. Risk of gastrointestinal
disease associated with exposure to pathogens in the sedi-
ments of the lower passaic river. Appl Environ Microbiol 74:
1004–1018. doi: 10.1128/aem.01203-07.
33. Droppo IG, Liss SN, Williams D, Nelson T, Jaskot C,
Trapp B. 2009. Dynamic existence of waterborne patho-
gens within river sediment compartments. Implications for
water quality regulatory affairs. Environ Sci Technol 43:
1737–1743.
34. Gin KY-H, Goh SG. 2013. Modeling the effect of light and
salinity on viable but non-culturable (VBNC) Enterococcus.
Water Res 47:3315–3328. doi: http://dx.doi.org/10.1016/j.
watres.2013.03.021.
35. Hewitt J, Greening GE, Leonard M, Lewis GD. 2013. Eval-
uation of human adenovirus and human polyomavirus as
indicators of human sewage contamination in the aquatic
environment. Water Res 47:6750–6761. doi: http://dx.doi.
org/10.1016/j.watres.2013.09.001.
36. Alcalde L, Folch M, Tapias JC. 2012. Removal and rela-
tionships of microbial indicators in a water treatment and
reclamation facility. J Water Health 10:549–559. doi:
10.2166/wh.2012.213.
37. Chern EC, Brenner K, Wymer L, Haugland RA. 2013.
Influence of wastewater disinfection on densities of cultura-
ble fecal indicator bacteria and genetic markers. J Water
Health 12:410–417. doi: 10.2166/wh.2013.179.
38. Cheng H-WA, Lucy FE, Broaders MA, Mastitsky SE,
Chen C-H, Murray A. 2012. Municipal wastewater treat-
ment plants as pathogen removal systems and as a contami-
nation source of noroviruses and Enterococcus faecalis.
J Water Health 10:380–389. doi: 10.2166/wh.2012.138.
39. Simmons FJ, Xagoraraki I. 2011. Release of infectious
human enteric viruses by full-scale wastewater utilities.
Water Res 45:3590–3598. doi: http://dx.doi.org/10.1016/j.
watres.2011.04.001.
40. Gaafar MR. 2007. Effect of solar disinfection on viability
of intestinal protozoa in drinking water. J Egypt Soc Parasitol
37:65–86.
41. Bichai F, Polo-López MI, Fernández Ibañez P. 2012. Solar
disinfection of wastewater to reduce contamination of let-
tuce crops by Escherichia coli in reclaimed water irrigation.
Water Res 46:6040–6050. doi: http://dx.doi.org/10.1016/j.
watres.2012.08.024.
42. Blum P. 2005. Understanding Viability of Pathogens during Dis-
infection. IWA Publishing, Alexandria, VA.
43. Azimi Y, Allen DG, Farnood RR. 2012. Kinetics of
UV inactivation of wastewater bioflocs. Water Res 46:
3827–3836. doi: http://dx.doi.org/10.1016/j.watres.2012.
04.019.
3.1.3-12 ▪ WATER
44. Azimi Y, Allen DG, Farnood RR. 2014. Enhancing dis-
infection by advanced oxidation under UV irradiation in
polyphosphate-containing wastewater flocs. Water Res 54:
179–187. doi: http://dx.doi.org/10.1016/j.watres.2014.01.011.
45. Carvalho G, Lemos PC, Oehmen A, Reis MAM. 2007.
Denitrifying phosphorus removal: linking the process per-
formance with the microbial community structure. Water
Res 41:4383–4396. doi: http://dx.doi.org/10.1016/j.watres.
2007.06.065.
46. Lanham AB, Oehmen A, Saunders AM, Carvalho G,
Nielsen PH, Reis MAM. 2013. Metabolic versatility
in full-scale wastewater treatment plants performing
enhanced biological phosphorus removal. Water Res 47:
7032–7041. doi: http://dx.doi.org/10.1016/j.watres.2013.
08.042.
47. Remy C, Miehe U, Lesjean B, Bartholomäus C. 2014.
Comparing environmental impacts of tertiary wastewater
treatment technologies for advanced phosphorus removal
and disinfection with life cycle assessment. Water Sci Technol
69:1742–1750. doi: 10.2166/wst.2014.087.
48. Azimi Y, Chen X, Allen DG, Pileggi V, Seto P, Droppo IG,
Farnood RR. 2013. UV disinfection of wastewater flocs: the
effect of secondary treatment conditions. Water Sci Technol
67:2719–2723. doi: 10.2166/wst.2013.148.
49. Flannery J, Keaveney S, Rajko-Nenow P, O’Flaherty V,
Doré W. 2013. Norovirus and FRNA bacteriophage deter-
mined by RT-qPCR and infectious FRNA bacteriophage
in wastewater and oysters. Water Res 47:5222–5231. doi:
http://dx.doi.org/10.1016/j.watres.2013.06.008.
50. Childress H, Sullivan B, Kaur J, Karthikeyan R. 2013.
Effects of ultraviolet light disinfection on tetracycline resist-
ant bacteria in wastewater effluents. J Water Health 12:
404–409. doi: 10.2166/wh.2013.257.
51. Haaken D, Dittmar T, Schmalz V, Worch E. 2014. Disin-
fection of biologically treated wastewater and prevention of
biofouling by UV/electrolysis hybrid technology: influence
factors and limits for domestic wastewater reuse. Water Res
52:20–28. doi: http://dx.doi.org/10.1016/j.watres.2013.12.
029.
52. Sakai H, Takamatsu T, Oguma K, Murakami M, Kosaka
K, Asami M, Takizawa S. 2014. Effects of natural
organic matter and nitrate on the behavior of nitrosodime-
thylamine during ultraviolet irradiation and chlorami-
nation. J Water Supply Res Technol 63:260–267. doi:
10.2166/aqua.2014.021.
53. Li D, Craik SA, Smith DW, Belosevic M. 2009. The
assessment of particle association and UV disinfection of
wastewater using indigenous spore-forming bacteria. Water
Res 43:481–489. doi: http://dx.doi.org/10.1016/j.watres.
2008.10.025.
54. Hedrick RP, Petri B, McDowell TS, Mukkatira K, Sealey
LJ. 2007. Evaluation of a range of doses of ultraviolet
irradiation to inactivate waterborne actinospore stages of
Myxobolus cerebralis. Dis Aquat Organ 74:113–118.
55. Gerrity D, Gamage S, Jones D, Korshin GV, Lee Y,
Pisarenko A, Trenholm RA, von Gunten U, Wert EC,
Snyder SA. 2012. Development of surrogate correlation
models to predict trace organic contaminant oxidation and
microbial inactivation during ozonation. Water Res 46:
6257–6272. doi: http://dx.doi.org/10.1016/j.watres.2012.
08.037.
56. Dietrich JP, Loge FJ, Ginn TR, Başağaoğlu H. 2007. Inac-
tivation of particle-associated microorganisms in wastewater
disinfection: modeling of ozone and chlorine reactive diffu-
sive transport in polydispersed suspensions. Water Res 41:
2189–2201. doi: http://dx.doi.org/10.1016/j.watres.2007.
01.038.
57. Pradhan SK, Kauppinen A, Martikainen K, Pitkänen T,
Kusnetsov J, Miettinen IT, Pessi M, Poutiainen H,
Heinonen-Tanski H. 2013. Microbial reduction in
wastewater treatment using Fe3+ and Al3+ coagulants and
PAA disinfectant. J Water Health 11:581–589. doi:
10.2166/wh.2013.241.
58. U.S. EPA. 2013. Emerging Technologies for Wastewater
Treatment and In-Plant Wet Weather Management. U.S.
EPA Office of Wastewater Management, 832-R-12-011.
59. Rodríguez-Chueca J, Mediano A, Ormad MP, Mosteo R,
Ovelleiro JL. 2014. Disinfection of wastewater effluents
with the Fenton-like process induced by electromagnetic
fields. Water Res 60:250–258. doi: http://dx.doi.org/10.
1016/j.watres.2014.04.040.
60. Bandala ER, Miranda J, Beltran M, Vaca M, López R,
Torres LG. 2009. Wastewater disinfection and organic mat-
ter removal using ferrate (VI) oxidation. J Water Health 7:
507–513. doi: 10.2166/wh.2009.003.
61. Gao S, Hemar Y, Ashokkumar M, Paturel S, Lewis GD.
2014. Inactivation of bacteria and yeast using highfrequency
ultrasound treatment. Water Res 60:93–104.
62. de Lima Isaac R, Urbano dos Santos L, Tosetto MS, Franco
RMB, Guimarães JR. 2014. Urban water reuse: microbial
pathogens control by direct filtration and ultraviolet dis-
infection. Journal of Water and Health 12:465–473. doi:
10.2166/wh.2014.221.
63. Michael I, Rizzo L, McArdell CS, Manaia CM, Merlin C,
Schwartz T, Dagot C, Fatta-Kassinos D. 2013. Urban
wastewater treatment plants as hotspots for the release of
antibiotics in the environment: a review. Water Res
47:957–995. doi: http://dx.doi.org/10.1016/j.watres.2012.
11.027.
64. U.S. EPA. 2002. Methods for Measuring the Acute Toxicity
of Effluents and Receiving Waters to Freshwater and Marine
Organisms. EPA-821-R-02-012. U.S. EPA Office of Water,
Washington, DC.
65. Drury B, Rosi-Marshall E, Kelly JJ. 2013. Wastewater
treatment effluent reduces the abundance and diversity
of benthic bacterial communities in urban and sub-
urban rivers. Appl Environ Microbiol 79:1897–1905. doi:
10.1128/aem.03527-12.
66. User’s Guide for T.E.S.T. (version 4.1) (Toxicity Estima-
tion Software Tool). 2012. U.S. EPA, Washington, DC.
67. Gedalanga P, Kotay SM, Sales CM, Butler CS, Goel R,
Mahendra S. 2013. Novel applications of molecular biolog-
ical and microscopic tools in environmental engineering.
Water Environ Res 85:917–950.
68. Salveson A, Rauch-William T, Drury D, Dickenson E,
Drewes JE, Snyder S. 2012. Trace Organic Compound Indi-
cator Removal during Conventional Wastewater Treatment.
WERF & IWA, Alexandria, VA.
69. Diamond J, Thornton K, Bartell S, Munkittrick K, Kidd
K, Kapo K, Latimer H, Gilliam J, Gerritsen J. 2010.
Development of Diagnostic Tools for Trace Organic Compounds
and Multiple Stressors. WERF & IWA, Alexandria, VA,
CEC5R08b.
70. McClellan K, Halden RU. 2010. Pharmaceuticals and
personal care products in archived U.S. biosolids from the
2001 EPA national sewage sludge survey. Water Res 44:
658–668. doi: http://dx.doi.org/10.1016/j.watres.2009.12.032.
71. Escher BI, Baumgartner R, Koller M, Treyer K, Lienert J,
McArdell CS. 2011. Environmental toxicology and risk
assessment of pharmaceuticals from hospital wastewater.
Water Res 45:75–92. doi: http://dx.doi.org/10.1016/j.
watres.2010.08.019.
72. Park H-D, Wells GF, Bae H, Criddle CS, Francis CA.
2006. Occurrence of ammonia-oxidizing archaea in waste-
water treatment plant bioreactors. Appl Environ Microbiol
72:5643–5647. doi: 10.1128/aem.00402-06.
73. Viau E, Bibby K, Paez-Rubio T, Peccia J. 2011. Toward a
consensus view on the infectious risks associated with land
 3.1.3. Assessing the Efficiency of Wastewater Treatment ▪ 3.1.3-13
application of sewage sludge. Environ Sci Technol 45:
5459–5469. doi: 10.1021/es200566f.
74. Simon C, Wiezer A, Strittmatter AW, Daniel R. 2009. Phy-
logenetic diversity and metabolic potential revealed in a gla-
cier ice metagenome. Appl Environ Microbiol 75:7519–7526.
doi: 10.1128/AEM.00946-09.
75. Graczyk TK, Kacprzak M, Neczaj E, Tamang L, Graczyk
H, Lucy FE, Girouard AS. 2007. Occurrence of Cryptospori-
dium and Giardia in sewage sludge and solid waste landfill
leachate and quantitative comparative analysis of sani-
tization treatments on pathogen inactivation. Environ Res
106:27–33.
76. Graczyk TK, Kacprzak M, Neczaj E, Tamang L, Graczyk
H, Lucy FE, Girouard AS. 2007. Human-virulent micro-
sporidian spores in solid waste landfill leachate and sewage
sludge, and effects of sanitization treatments on their inacti-
vation. Parasitol Res 101:569–575.
77. More RP, Mitra S, Raju SC, Kapley A, Purohit
HJ. 2014. Mining and assessment of catabolic pathways
in the metagenome of a common effluent treatment plant
to induce the degradative capacity of biomass. Bioresour
Technol 153:137–146. doi: 10.1016/j.biortech.2013.
11.065.
78. Munir M, Wong K, Xagoraraki I. 2011. Release of anti-
biotic resistant bacteria and genes in the effluent and
biosolids of five wastewater utilities in Michigan. Water Res
45:681–693. doi: http://dx.doi.org/10.1016/j.watres.2010.
08.033.
79. Saha S, Hunter WB, Reese J, Morgan JK, Marutani-Hert
M, Huang H, Lindeberg M. 2012. Survey of endosymbionts
in the Diaphorina citri metagenome and assembly of a
Wolbachia wDi draft genome. PLoS One 7:e50067. doi:
10.1371/journal.pone.0050067.
80. Whiteley AS, Jenkins S, Waite I, Kresoje N, Payne H,
Mullan B, Allcock R, O’Donnell A. 2012. Microbial 16S
rRNA Ion Tag and community metagenome sequencing
using the Ion Torrent (PGM) platform. J Microbiol Meth
91:80–88. doi: 10.1016/j.mimet.2012.07.008.
81. Wommack KE, Bhavsar J, Polson SW, Chen J, Dumas M,
Srinivasiah S, Furman M, Jamindar S, Nasko DJ. 2012.
VIROME: a standard operating procedure for analysis of viral
metagenome sequences. Stand Genom Sci 6:427–439. doi:
10.4056/sigs.2945050.
82. Youssef N, Steidley BL, Elshahed MS. 2012. Novel
high-rank phylogenetic lineages within a sulfur spring
(Zodletone Spring, Oklahoma), revealed using a combined
pyrosequencing-sanger approach. Appl Environ Microbiol
78:2677–2688. doi: 10.1128/aem.00002-12.
83. Brum JR, Culley AI, Steward GF. 2013. Assembly
of a marine viral metagenome after physical fractiona-
tion. PLoS One 8:e60604. doi: 10.1371/journal.pone.
0060604.
84. Clark SC, Egan R, Frazier PI, Wang Z. 2013. ALE: a
generic assembly likelihood evaluation framework for assess-
ing the accuracy of genome and metagenome assemblies.
Bioinformatics 29:435–443. doi: 10.1093/bioinformatics/
bts723.
85. Pinto AJ, Schroeder J, Lunn M, Sloan W, Raskin L.
2014. Spatial-temporal survey and occupancy-abundance
modeling to predict bacterial community dynamics in the
drinking water microbiome. mBio 5:e01135-14. doi:
10.1128/mBio.01135-14.
86. Corcoran E, Nellemann C, Baker E, Bos R, Osborn D,
Savelli H (eds). 2010. Sick Water? The Central Role of Waste-
water Management in Sustainable Development. A Rapid
Response Assessment. United Nations Environment Pro-
gramme. UNEP/UN-HABITAT, Arendal, Norway.
87. Dazzo FB, Schmid M, Hartmann A. 2007. Immunofluores-
cence microscopy and fluorescence in situ hybridization
combined with CMEIAS and other image analysis tools
for soil- and plant-associated microbial autecology, pp
712–733. In:Hurst CL, (ed), Manual of Environmental Micro-
biology, 3rd ed. ASM, Washington, DC.
88. U.S. EPA NRMRL. 2006. Life Cycle Assessment: Prin-
ciples and Practice. EPA/600/R-06/060. NRMRL, U.S.
EPA, Cincinnati, OH,
89. Elorri I, Benetto E, Venditti S, Köhler C, Cornelissen A.
2013. Comparative and integrative environmental assess-
ment of advanced wastewater treatment processes based on
an average removal of pharmaceuticals. Water Sci Technol
67:387–394. doi: 10.2166/wst.2012.581.
90. Magdeburg A, Stalter D, Schlüsener M, Ternes T,
Oehlmann J. 2014. Evaluating the efficiency of advanced
wastewater treatment: target analysis of organic contami-
nants and (geno-)toxicity assessment tell a different story.
Water Res 50:35–47.
91. Senta I, Terzic S, Ahel M. 2013. Occurrence and fate of dis-
solved and particulate antimicrobials in municipal waste-
water treatment. Water Res 47:705–714. doi: http://dx.doi.
org/10.1016/j.watres.2012.10.041.
92. Rattier M, Reungoat J, Keller J, Gernjak W. 2014.
Removal of micropollutants during tertiary wastewater treat-
ment by biofiltration: role of nitrifiers and removal mecha-
nisms. Water Res 54:89–99. doi: http://dx.doi.org/10.1016/
j.watres.2014.01.030.
93. Liu RX, Wilding A, Hibberd A, Zhou JL. 2005. Partition
of endocrine-disrupting chemicals between colloids and dis-
solved phase as determined by cross-flow ultrafiltration.
Environ Sci Technol 39:2753–2761.
94. Gori R, Lubello C, Ferrini F, Nicese F. 2004. Reclaimed
municipal wastewater as source of water and nutrients for
plant nurseries. Water Sci Technol 50:69–75.
95. Li J, Werth CJ. 2001. Evaluating Competitive Sorption
Mechanisms of Volatile Organic Compounds in Soils and
Sediments Using Polymers and Zeolites. Environmental Sci-
ence and Technology 35:568-74. doi: 10.1021/es0.
96. Liu JC, Carr SA. 2013. Removal of estrogenic com-
pounds from aqueous solutions using zeolites. Water Envi-
ron Res 85:2157–2163. doi: http://dx.doi.org/10.2175/
106143013X13736496909356.
97. Limpiyakorn T, Furhacker M, Haberl R, Chodanon T,
Srithep P, Sonthiphand P. 2013. amoA-encoding archaea
in wastewater treatment plants: a review. Appl Microbiol
Biotechnol 97:1425–1439.
 Epidemiologic Aspects of Waterborne Infectious Disease
SAMUEL DOREVITCH
3.1.4
More than 150 years have passed since John Snow identified
the critical roles of wastewater and drinking water in causing
the London cholera epidemic of 1854. Since then, despite
continued progress in epidemiology, microbiology, civil
engineering, public health practice, and medical services,
waterborne disease persists, even in nations with advanced
public health systems. Waterborne infectious diseases do
not randomly impact individuals within populations. The
distribution and determinants of disease—in other words,
the epidemiology of disease—must be understood to devise
public health action that will effectively prevent waterborne
disease transmission (as Snow did when he removed the
Board Street pump handle).
The “epidemiologic triad” of infectious diseases remains
a useful framework for understanding the distribution and
determinants of disease caused by these agents. Elements of
the triad are the infectious agent, the susceptible host, and
an environment in which the agent can be transmitted to the
host. Agents of waterborne illness include a wide range of
bacteria, viruses, protozoa, and toxins elaborated by microbes.
Surface waters (lakes, ponds, rivers, and oceans) support the
persistence of many infectious agents. Biofilms found in water
distribution system, including in-premise plumbing, can serve
as a reservoir for many recognized pathogens. While only
humans serve as hosts for some waterborne infectious agents
(such as hepatitis A), agents of zoonotic disease (such as
Cryptosporidium spp.) have reservoirs in other animals.
The ingestion, inhalation, and direct dermal contact
with waterborne microbes or microbial toxins can result in
a variety of clinical syndromes, of which gastroenteritis is
the most frequently reported and best studied. Symptoms
of gastroenteritis include nausea, vomiting, diarrhea, abdomi-
nal cramps, and in some cases, fever. Inhalation of water-
borne microbes can cause lower respiratory tract infections,
such as bronchitis and pneumonia, which produce symptoms
such as cough, shortness of breath, chest discomfort, and
fever. Conjunctivitis can be caused by ocular exposures,
and otitis externa (“swimmer’s ear”) can be caused by expo-
sures to the skin of the ear canal. Constitutional, rather
than organ-specific symptoms (fever, chills, headache, and
muscle ache) are common in leptospirosis. Although dermal
exposures typically cause local symptoms (skin rash, irrita-
tion, and/or itching), exposures to Leptospira spp. or Vibrio
vulnificus in surface waters at areas of broken skin (scrapes
or cuts) can allow these pathogens to gain entry to deeper
doi:10.1128/9781555818821.ch3.1.4
3.1.4-1
skin structures and cause systemic infections. Gastroenteritis
acquired through water recreation is relatively common,
and it is typically mild and self-limited. Some rare waterborne
infections are life-threatening. V. vulnificus can cause massive
soft tissue destruction, in some cases requiring amputation of
effected extremities. Naegleria fowleri, a waterborne amoeba
found in warm freshwaters, can cause primary amoebic ence-
phalitis, an infection of the central nervous system that is
nearly uniformly fatal. Waterborne E. coli O157:H7 infec-
tions can cause hemolytic uremic syndrome, an acute kidney
disease than may require dialysis.
This chapter describes the epidemiology of waterborne
diseases from the perspective of societies in which safe
drinking water and effective wastewater treatment are widely
available, with much of the epidemiologic information com-
ing from the United States and Canada. The research and dis-
ease surveillance methods that have generated our present
understanding of waterborne disease are described, along
with limitations of those methods. Areas of active research
regarding emerging public health concerns in recent years
are emphasized.
EPIDEMIOLOGIC METHODS USED IN
UNDERSTANDING AND PREVENTING
WATERBORNE INFECTIOUS DISEASES
Outbreak Investigations
Much of what is known about waterborne disease comes from
reports of disease outbreaks. The sequence of events that
results in the identification of a waterborne disease outbreak
(WBDO) begins with the notification of health departments
of a suspected outbreak by individuals who develop illness,
their health care providers, or clinical laboratory personnel
if the microbe identified is on a list of “notifiable” diseases.
If such notification occurs, local public health authorities
must decide whether an investigation is warranted given their
limited resources. If preliminary information suggests that
an outbreak has occurred, particularly if it involves severe dis-
ease or numerous individuals, an investigation may follow.
The gastrointestinal symptoms experienced by a cluster of
people are not specific to WBDOs and could be caused by
other sources (food) and pathways ( person-to-person spread).
Thus, outbreak investigation is often the only way to deter-
mine when an outbreak is waterborne, identify the causative
3.1.4-2 ▪ WATER
agent(s), and describe the circumstances that caused the out-
break. The complexity of the investigation depends on the
availability of resources, the number of people impacted,
and the severity of illness (whether emergency department
visits, hospitalizations, or deaths occurred). In such studies
people who did not develop symptoms (controls) are inter-
viewed in addition to those who did develop symptoms (cases)
to identify exposures that were more common among the cases
than the controls.
Outbreaks that are thought to be waterborne should be
reported to national authorities. In the United States, local
and state health department investigations are then reported
to the U.S. Centers for Disease Control and Prevention
(CDC), Waterborne Outbreak Disease Surveillance System
(WBDOSS), which published summaries of outbreak inves-
tigation findings ( pathogens, illness type, factors that contrib-
uted to the outbreak, temporal trends, etc.) approximately
every two years (1–5). WBDOSS considers two or more cases
of a similar illness that are linked temporally to a common
water exposure to be a waterborne outbreak. In Canada,
researchers have described and analyzed WBDOs (6, 7) while
very large outbreaks have been evaluated and reported by
local and provincial authorities (8). FoodNet Canada is a rel-
atively new system that includes comprehensive surveillance
for foodborne and waterborne infections at one sentinel loca-
tion in each of three provinces: Alberta, British Columbia,
and Ontario (9).
WBDOs that are captured by surveillance systems provide
an important but incomplete picture of all WBDOs and cer-
tainly of waterborne disease overall. A key limitation is that
the United States relies on passive surveillance, rather than
active surveillance. Public health agencies do not “look for”
outbreaks; rather, individuals who suspect that an outbreak
has occurred can notify local public health officials. As a
result, the recognized WBDOs represent a small but unknown
fraction of all WBDOs that occur. Outbreaks that are recog-
nized may be atypical in the severity of cases and the degree
to which obvious clustering of cases has occurred. For exam-
ple, multiple cases of gastroenteritis among individuals from a
small community (who know one another and seek care from
a small number of medical providers) who used a local swim-
ming pool would be relatively likely to be recognized as an
outbreak. A similar number of cases with a comparable
severity illness in an outbreak that centered around an urban
beach would be less likely to be recognized if those who
became ill were from many neighborhoods and see different
health care providers. Another reason that outbreaks of gas-
troenteritis are difficult to recognize is that in the United
States about 20% of adults with diarrhea seek medical atten-
tion and of those about 20% provide a stool sample for anal-
ysis (10). In Canada, for every one positive stool culture
reported to public health authorities, an estimated 300 cases
of gastrointestinal illness occur in the community (11).
Despite these limitations, WBDO surveillance systems pro-
vide important information about the microbes identified
and the relevant exposure pathways. Outbreaks investigations
generate data needed to improve surface water protections,
the management of swimming pools and water parks, the
treatment of drinking water and wastewater, and to address
deficiencies in water distribution systems.
Other U.S. Surveillance Systems for
Waterborne Disease
Infections caused by the highly virulent waterborne Vibrio
spp. bacteria are tracked in CDC’s COVIS, the Cholera and
Other Vibrio Illness Surveillance system. Noncholera Vibrio
species infections had been reported voluntarily by several
coastal states beginning in 1989 and since 2007, when infec-
tions caused by noncholera Vibrio species have been nation-
ally reportable, surveillance has become more complete.
Another distinct waterborne disease surveillance system is
HABISS, the CDC’s Harmful Algal Bloom-Related Illness
Surveillance System. This system compiles information vol-
untarily submitted by state or local health departments about
human and animal cases of illness thought to be due to harm-
ful algal blooms (including cyanobacterial blooms). Among
the data elements requested are demographics, clinical
descriptions, environmental conditions, and any data regard-
ing toxin measurement.
Cohort Studies
Epidemiologic studies conducted in the United States over
the past decade that collectively enrolled tens of thousands
of participants have described waterborne illness in thousands
of individuals (12–18). However, outbreak reporting systems
did not identify any outbreaks at the times and locations the
studies were conducted. Thus, the term “sporadic cases” of
waterborne disease has been used to describe those that are
not part of recognized outbreaks. For that reason, outbreak
surveillance data do not address the frequent occurrence of
these relatively mild cases of illness, which may be caused
by different agents and might require different strategies for
prevention than do outbreaks. A very informative but time-
intensive and costly method of such research is the prospec-
tive cohort study. This involves identifying people free of
symptoms at baseline (before they swim) and monitoring
them over time to describe the frequency with which illness
occurs. This approach has been applied to characterize the
short-term health risks of water recreation and tap water con-
sumption. In an ideal cohort study of water recreation, a dem-
ographically identical group of people that has no exposure to
the recreational water of interest is monitored during the same
time period as a group of people who are exposed to that
water. Cases of illness are defined based on the reporting of
symptoms. However, whether a case of illness is due to water
exposure in a given study participant who swam at a beach,
cannot be determined. This is because the symptoms of gas-
trointestinal, respiratory, or skin conditions acquired as a
result of water recreation are similar to those of gastrointesti-
nal, respiratory, or skin conditions acquired through other
pathways (such as through food or via person-to-person trans-
mission). For that reason, the risk of illness observed in the
water-exposed study participant group is compared with risk
of illness in the unexposed group. The difference in risk
between the two groups produces the attributable risk, such
as the number of people per 1,000 water recreators who
develop illness as a result of water recreation.
A challenge in interpreting findings of prospective
cohort studies is that the two groups (exposed and unexposed)
are unlikely to be identical in important ways, such the distri-
bution of chronic health problems between those who choose
to go into the water for recreation and those who choose not
to (the unexposed group) (14, 15, 19, 20). Statistical methods
are used to adjust for potential “confounders”—variables such
as health status or age that are associated with the outcome
and differ between the exposed and unexposed study groups.
To characterize the observed attributable risks of water
recreation (often in the range of 10–50 cases per thousand),
several thousand participants must be enrolled. To determine
whether the risks differ in subgroups (for example, children
compared with adults), thousands of individuals may need
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-3
to be enrolled in each subgroup if the risk difference is subtle.
Thus, while prospective cohort studies can generate impor-
tant information, they are resource intensive and can take
years (several bathing seasons) to complete.
Randomized Controlled Trials (RCTs)
In RCTs individuals are randomly assigned to one of two (or
more) exposure groups. In studies of water recreation, the
conditions are remaining at the beach but staying out of the
water (unexposed) and going into the water and performing
head submersion (exposed) (21–23). In studies that evaluate
the role of drinking water in the occurrence of gastroenteritis,
the exposure groups have included families that receive
enhanced drinking water treatment in the home versus those
who do not but whose standard supplied drinking water meets
water quality requirements (18, 24–29). As a result of ran-
domization, the study groups (exposed, unexposed) should
be nearly identical in terms of the distribution of age and
the presence of underlying health conditions. For this reason,
the statistical adjustment methods required in cohort studies
may not be needed and for that reason RCTs require much
smaller numbers of participants than do cohort studies.
An important limitation of the RCT approach used
to study water recreation is that the exposure to which
some participants are randomized is head immersion several
times in a specific time interval (typically 10 or 15 minutes).
Although this helps standardize the exposures of individuals,
the exposure may be quite different than those experienced
during swimming. If head immersion results in significantly
less exposure than swimming, such studies could underesti-
mate the health effects of swimming. Both the cohort and
RCT studies of water recreation are subject to two additional
limitations. First, the water-exposed individuals are (of
course) not blinded to their study group (water exposure ver-
sus not). Those who have water exposure and are concerned
about the health effects of such exposure may be more likely
to report symptoms compared with those who may not be
aware of water quality hazards, a phenomena referred to a
risk perception bias (30, 31). This could result in an overesti-
mation of health risk. Second, even if relationships between
water quality and health risk truly exist, they may not be
detectable if the range of water measures is narrow at the loca-
tions and times that the research is conducted.
Health Care Utilization Data Analysis
In recent years, data provided to state public health agencies
by hospitals has been used to evaluate whether gastroenteritis
may be related to precipitation, which can flush pathogens
found on the ground into surface waters by sheet flow, storm
sewers, or combined (septic and storm) sewer systems. Many
of these studies have utilized a case-crossover design, in which
the rate of illness during a specific time period (generally
between 1 and 10 days) following precipitation is compared
to the rate of illness in the same location during a comparable
time period that had not been preceded by precipitation. The
hospitalization or emergency department visit and/or stool
culture data are obtained from a hospital system, state health
department, or federal data system (such as Medicare). Unlike
the cohort studies or RCTs, which define health outcomes
based on symptom reporting by study participants, hospital
discharge data include diagnoses made by health care pro-
viders who have spoken with and examined patients. Investi-
gators have begun to use this approach to analyze rainfall
as a predictor of infections caused by specific pathogens,
such as Clostridium difficile (32), rather than nonspecific
gastroenteritis.
Strengths of this method include the ability to readily
obtain information at relatively little cost. Large health data
sets can then be linked to weather, land use, CSO outfall loca-
tion (33), water treatment and sociodemographic data (34).
Some analyses have focused on specific populations, such as
children (35) or older adults enrolled in the U.S. Medicare
program (36). One limitation of using public health data
sets, rather than data collected for research purposed in a
cohort study, is that exposure data of individuals (did the
individual engage in water recreation? did he or she drink
bottled water rather than tap water?) are not available.
THE EPIDEMIOLOGY OF RECREATIONAL
WATERBORNE ILLNESS
Recreational Waterborne Disease: Outbreaks
Information regarding recreational waterborne disease out-
breaks reported to the CDC in 2007–2008 (3) and 2009–
2010 (5) is summarized in Table 1. Of the 227 outbreaks,
171 (75.0%) were linked to water recreation, as were
13,976 of 18,940 (73.8%) cases of illness in the waterborne
outbreaks. The vast majority of outbreaks were of gastrointes-
tinal illness, with much smaller numbers of respiratory,
skin, and other symptom categories. Recreation in treated
water venues, such as water parks and swimming pools,
accounts for the largest number of all reported recreational
outbreaks (136/171 = 79.5%) and ill individuals (13,204/
13,976 = 94.5%). Nearly all recreational outbreaks occurred
in April through September. Recognized outbreaks were rel-
atively small in size, with fewer than 10 people known to be
impacted in most. The protozoan parasite Cryptosporidium
spp. accounts for an overwhelming majority of the outbreaks
and cases linked to treated recreational water venues. Even
after excluding a massive outbreak of cryptosporidiosis
(5,697 cases) that centered around swimming pools, Crypto-
sporidium spp. accounted for the majority of outbreak cases.
Recognized outbreaks linked to recreation at untreated
waters (lakes, rivers, and beaches) are caused by a wider vari-
ety of etiologic agents. It is thought that at treated waters
(swimming pools, water parks, spas), a wide variety of patho-
gens are brought into the water by people—particularly chil-
dren who are not fully toilet trained—but that treatment
processes ( primarily chlorination) effectively renders bacteria
and viruses noninfectious, but not Cryptosporidium spp.
oocysts. The much smaller number of recognized outbreaks
at untreated water may reflect differences in risk, but it may
also reflect differences in the likelihood of identifying out-
breaks linked to beaches versus water parks and swimming
pools. The outbreaks at untreated waters generally occurred
at small inland lakes, rather than marine or Great Lakes
beaches, which may reflect the greater dilution of pathogens
in coastal waters than small bodies of water (37). Confirmed
leptospirosis outbreaks were not reported in 2007–2010. Out-
breaks linked to recreation on rivers in Florida (38), Illinois
(39), and Western Europe (40–42).
Among 958 cases of waterborne illness reported in the
2009–2010 period, 61 (6.4%) were hospitalized and none
were known to have died from their illness. This does
not include case of waterborne illness caused by the proto-
zoan N. fowleri or Vibrio spp. bacteria, which have consi-
derably higher case-fatality rates. During the 2007–2008
reporting period, eight fatal cases of primary amoebic menin-
goencephalitis were caused by N. fowleri after recreation
in warm untreated freshwater. All of these fatal cases occurred
among people under the age of 22 years. During 2007–2008,
3.1.4-4 ▪ WATER

TABLE 1 Etiologic agents in waterborne disease outbreaks reported by the CDC, 2007–2010
Recreational water Drinking water
Untreated
Etiology Treated waters Ground Surface Total
waters
Outbr. Cases Outbr. Cases Outbr. Cases Outbr. Cases Outbreak Cases
Campylobacter spp. 1 6 0 0 6 877 0 0 7 883
E. coli O157:H7 4 20 3 56 2 39 0 0 9 115
Legionella spp. 0 0 14 130 7 24 20 100 41 254
Plesiomonas shigelloides 1 2 0 0 0 0 0 0 1 2
Providencia 0 0 0 0 1 55 0 0 1 55
Pseudomonas spp. 0 19 214 0 0 0 0 19 214
Salmonella spp. 0 0 0 0 2 1,305a 0 0 2 1,305
Shigella spp. 3 72 3 32 0 0 0 0 6 104
Bacteria, total 9 100 39 432 18 2,300 20 100 86 2,932
Schistosomes 6 311 0 0 0 0 0 0 6 311
Cryptosporidium spp. 5 33 83 12,549b 1 35 0 0 89 12,617
Cryp. hominis + Giardia intestinalis 0 0 2 26 0 0 0 0 2 26
Cyclospora cayetanensis 0 0 0 0 0 0 1 82 1 82
Giardia spp. 0 0 3 19 4 95 0 0 7 114
Parasite, total 9 344 88 12,594 5 130 1 82 103 13,150
Hepatitis A 0 0 0 0 0 0 0 0 0 0
Norovirus 4 134 4 147 5 312 0 0 13 593
Virus, total 4 134 4 147 5 312 0 0 13 593
Multiple pathogens 2 99 0 0 3 246 1 41 6 386
Cyanobacterial toxin 11 61 0 0 0 0 0 0 11 61
Unidentified 0 34 5 31 2 90 1 1,663c 8 1,818
Mult/cyano/unID total 13 194 5 31 5 336 2 1,704 25 2,265
Total 35 772 136 13,204 33 3,078 23 1,886 227 18,940
Excludes chemical agents such a swimming pool chemicals, but includes toxins associated with algal blooms.
“Multiple pathogens” refers to outbreaks with more than one category of microbe (e.g., virus and bacteria) that were identified as etiologic agents.
a
A single outbreak of Salmonella typhimurium linked to an untreated community water system accounted for an estimated 1,300 of these cases.
b
A single outbreak linked to a swimming pool accounted for 5,697 of these cases.
c
A single outbreak linked to a community water system with treatment deficiencies accounted for 1,663 cases of gastroenteritis. Water in the treatment plant
tested positive for fecal coliforms and E. coli, but the agent responsible for symptoms was not identified. In 2009–2010 two drinking water outbreaks (12 cases) were
in systems that used both groundwater and surface water. For the purposes of this table they were categorized as ground water outbreaks.

vibriosis resulted in a considerably higher rate of hospitaliza-
tion and death than for outbreaks overall (31.4% and 3.8% of
236 cases, respectively). Nearly all vibriosis patients were
exposed to marine or brackish water on the Atlantic and
Gulf of Mexico coasts.
Recreational Waterborne Illness: Sporadic
Epidemiologic studies provide the best available estimates of
the observed risk of sporadic illness following water recrea-
tion. Cohort studies conducted more than 30 years ago by
the U.S. Environmental Protection Agency (EPA) found
that the risk of gastrointestinal illness among swimmers could
be predicted by concentrations of enterococci at marine
beaches (43) and by either E. coli or enterococci at freshwater
beaches (44). This work formed the basis of the 1986 Ambi-
ent Water Quality Criteria (45), which set values of indicator
bacteria to limit the risk of illness to a goal of 8 cases/1,000
swimmers in fresh water and 19 cases/1,000 swimmers in
marine waters. In the years following the publication of the
1986 criteria, it became clear that water quality at a given
location changes rapidly at beaches. As a result of water qual-
ity changes during the 18–24-h interval between water
sampling and obtaining bacterial culture results, erroneous
beach management decisions were made (leaving a beach
open for swimming when it should have been closed). To
address this limitation, the EPA conducted a new set of epi-
demiologic studies, the National Epidemiological and Envi-
ronmental Assessment of Recreational Water (NEEAR),
which evaluated the quantitative PCR method to support
same-day beach management decisions (such as issuing a
swim advisory or swim ban).
NEEAR data shed light on the frequency with which
sporadic cases of gastrointestinal illness occurs among swim-
mers at beaches that are potentially impacted by wastewater
discharge. The unadjusted rate of gastrointestinal illness
was 23 cases/1,000 higher among swimmers than among non-
swimmers at Great Lakes beaches. At marine beaches, the
difference was 49 cases of gastrointestinal illness per 1,000
people on days that enterococci concentrations exceeded
EPA criteria of 35 colony forming units/100 ml. On days
the enterococci concentrations met EPA criteria, the differ-
ence between swimmers and nonswimmers in the rate of gas-
trointestinal illness was 18 cases per 1,000 (12). Another
cohort study, the Chicago Health, Environmental Exposure,
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-5
and Recreation Study (CHEERS) found that following
boating, fishing, rowing, and paddling at Chicago-area inland
waters and Lake Michigan, some of which were heavily
polluted by nondisinfected wastewater, the risk of gastrointes-
tinal illness attributable to water recreation was approxi-
mately 15 cases per 1,000 people (14). Thus, these large
studies of swimming and incidental contact recreation found
that about 1–5% of people who engage in water recreation at
beaches, rivers, and lakes, appear to develop gastrointestinal
illness attributable to water recreation, with the higher end
of that range observed among swimmers in marine waters
that exceed EPA enterococci criteria values. The sporadic
recreational waterborne illness identified in cohort studies is
substantially less severe illness identified in outbreaks. The
acute gastrointestinal illness observed in a cohort study of
incidental contact recreation on highly polluted waters gen-
erally did not result in a visit to a physician and led to
hospitalization in less than 1% of cases (14), compared to
the 6.4% noted in outbreak-associated cases of recreational
waterborne illness.
Indicator Bacteria and Risk of Recreational
Waterborne Illness
Because of the wide variety of pathogens associated with
waterborne disease, the complexity of water sampling
procedures for protozoa and viruses, and the challenges of
viral culture, epidemiologist studies have utilized measures
of waterborne fecal indicator bacteria (FIB) rather than
pathogens as predictors of illness among study participants.
While epidemiologic studies consistently demonstrate that
swimmers are at increased risk for gastrointestinal illness
(compared to nonswimmers), research findings are much
less consistent as to whether concentrations of FIB in beach
water are predictive of the risk of illness. At Great Lakes
and marine beaches impacted by treated wastewater, concen-
trations of the indicator bacteria enterococci, measured by
either cultivation or qPCR, are predictive of the occurrence
of acute gastrointestinal illness (13, 20). In a large study of
swimming at marine beaches not impacted by wastewater
discharge, no relationship was observed between indicator
bacteria and the risk of gastrointestinal illness (15). At a
California beach that was intermittently impacted by flow
from inland waters, gastrointestinal illness was associated
with FIB (enterococci by culture or qPCR, fecal coliforms,
total coliforms) when water flowed from inland locations
to the beach area (berm open); when the berm was closed, ill-
ness was not associated with indicator bacteria concentrations
(16). Neither indicator bacteria, coliphages, Giardia spp.,
Cryptosporidum spp., nor turbidity were predictive of GI ill-
ness in a recent study of incidental contact water recreation
on inland waters, including some there were heavily impacted
by nondisinfected wastewater discharge (31). Interest has
grown in evaluating the use of fecal indicator viruses to assess
the risk posed by water recreation. Coliphages, bacterio-
phages that infect E. coli, have been evaluated as predictors
of health risk among water recreators enrolled cohort studies.
Although some studies have found coliphages to be predictive
of illness (46), other have not (47). Among 6,350 partici-
pants in NEEAR study sites at three U.S. marine beaches
that were potentially impacted by wastewater discharges,
the gastrointestinal illness rate was significantly higher among
swimmers than nonswimmers (8.4% versus 4.9%) on days
that coliphages were detectable using a rapid (5-h) test
method; illness rates among swimmers and non-swimmers
were comparable on days that F+ coliphages were
undetectable (12). Coliphage measurement is currently
under consideration by the EPA as method of evaluating fecal
contamination of surface waters (48). Together findings of
studies of illness in relation to measures of bacterial and viral
indicator microbes suggests that characteristics of the recrea-
tional water site, the extent of water exposure, the source of
indicator microbes, how the microbes are measured, and
other factors impact the value of the indicators as predictors
of illness risk.
Rapid Molecular Approaches and Public Health
Monitoring of Beach Water
A major shift in beach monitoring methods for public notifi-
cation purposes is under way. In the NEEAR study, the qPCR
measures of enterococci in water samples from freshwater
beaches were better predictors of illness than were the enter-
ococci measures obtained from cultivation (13). The method
requires greater technical expertise and more sophisticated
laboratory facilities than does cultivation. Furthermore, it is
not yet known (and may never be) whether the implementa-
tion of qPCR monitoring program reduces rates of sporadic GI
illness among beachgoers. Nevertheless, beach managers are
beginning to implement qPCR monitoring to provide timely
information to the public.
Demographic, Medical, and Exposure Risk Factors for
Sporadic Waterborne Gastroenteritis
In addition to measures of water quality, other variables pre-
dict whether individuals will develop illness following water
recreation. An increased susceptibility to sporadic recrea-
tional waterborne gastrointestinal illness among children
was suggested in studies of swimming (13, 15), but not in a
study of incidental contact recreation (14). Gender does
not appear to be a risk factor for illness. The risk of developing
gastrointestinal symptoms among swimmers was found to
increase with increasing water exposure, defined as no expo-
sure, any exposure, face exposure, and swallowing water
(15). Similar findings were noted in a study of incidental con-
tact water recreation (14, 31). Frequent user of a location for
canoeing and other incidental contact activities appears to
have a significantly lower risk of illness than do less frequent
(or new) users (31, 46). People who have a chronic gastroin-
testinal condition were found to be at significantly greater risk
of acute gastrointestinal illness follow incidental contact
recreation (14).
Pathogens Responsible for Sporadic Waterborne
Gastroenteritis
Very little is known about pathogens responsible for sporadic
cases of recreational waterborne gastrointestinal illness (in
contrast to outbreaks, in which pathogens responsible for
illness are typically identified). One pilot study conducted
in the United Kingdom tested stool samples from participants
before and one and three weeks after beach visits. Samples
were provided by participants with and without beach water
exposure, regardless of whether or not they developed gastro-
intestinal symptoms. Based on the analysis of samples from
260 people, the presence of pathogens in stool samples was
unrelated to beach water exposure (49). A larger cohort study
recently analyzed stool samples from 740 people who devel-
oped gastrointestinal symptoms, of whom 510 had recent
exposure to recreational water and 230 did not. No pathogen
or category of pathogen microbe (virus, bacteria, protozoa)
was more frequent among those with exposure to recreational
water than among those without such exposure (50). Thus
3.1.4-6 ▪ WATER
observational studies have yet to identify microbes that
account for sporadic recreational waterborne gastrointestinal
illness.
Recreational Waterborne Disease: Outcomes
Other Than Gastrointestinal Illness
Numerous cohort studies have evaluated skin rash (or skin
symptoms) as a consequence of water recreation and in
some studies, the risk of rash in relation to measures of indica-
tor bacteria. A systematic review and meta-analysis of such
studies concluded that skin symptoms are more frequent
among swimmers in both marine and freshwater compared
to non-simmers, and are more frequent when FIB concen-
trations are high (51). The risk of eye symptoms was found
to be elevated among people who engaged in limited contact
recreation on the heavily polluted Chicago River system (14
cases per 1,000 uses of the river relative attributable to water
recreation) but not among users of cleaner waters (14). Respi-
ratory symptoms have been linked to water exposure among
those who perform head immersion in marine water (23)
and in some studies of marine swimmers (12) but not others
(15). Because the pathogens identified in outbreaks of respi-
ratory infection have been described extensively in the con-
text of nonrecreational water exposure (drinking water,
water used in air conditioning systems), this topic is discussed
under “Waterborne Respiratory Infection.”
INFECTIOUS DISEASES TRANSMITTED VIA
DRINKING WATER
Although the drinking water supply in advanced economies
is generally quite safe, outbreaks of infectious gastroenteritis
do occur and can cause relatively severe illness in large num-
bers of people. In addition, respiratory disease outbreaks and
sporadic cases of pneumonia have been linked to drinking
water systems.
Gastrointestinal Illness Associated with
Drinking Water: Outbreaks
Of the 57 reported U.S. drinking water–associated outbreaks
in in 2007–2010 33 (58%) were linked to groundwater
systems (wells) and accounted for 3,078 (61.2%) of drinking
water cases (see Table 1). Drinking water outbreaks in the
United States demonstrate little seasonality (unlike the
observed peak of recreational water outbreaks in summer).
Bacteria accounted for the majority of ground water and
surface water outbreaks, though viruses and protozoa caused
significant numbers of groundwater-associated outbreaks.
Outbreaks due to drinking water in Canada between 1974
and 2001 showed a generally similar distribution of pathogens
to that described in Table 1, though protozoa accounted for a
somewhat higher proportion of outbreaks in Canada (6). An
important contribution of the CDC surveillance system
summaries of drinking water WBDOs is the description of
deficiencies in elements of drinking water systems that con-
tributed to outbreaks. In 2007–2010 numerous drinking
water–associated outbreaks and cases were attributed to the
contamination of water in distribution systems, in some cases
because of cross-connections between drinking water and
nonpotable water pipes. Other outbreaks were due to conta-
minated source water in systems that did not include water
treatment (such as chlorination).
Three gastrointestinal illness outbreaks are described to
illustrate several types of pathogens, exposure pathways, and
system deficiencies.
Outbreak 1: A massive outbreak of cryptosporidiosis
occurred in Milwaukee, Wisconsin, in 1993, causing approx-
imately 54 deaths and 403,000 cases of gastrointestinal illness
(52) in a metropolitan area of approximately 1.6 million
people. Multiple factors are through to have contributed to
this outbreak, including heavy antecedent rainfall, runoff
from agricultural areas into Lake Michigan (source of Milwau-
kee’s drinking water) that contained Cryptosporidium spp.
oocysts, improper operation of water treatment equipment,
and inadequate response to dramatically elevated turbidity
of finished drinking water. One lesson learned from this
epidemic was that public drinking water systems could not
depend on the conventional processes of coagulation fol-
lowed by the filtration used in Milwaukee. Subsequent
regulatory changes by the EPA were phased in to provide
additional protections for surface waters and ground water
under the direct influence of surface water to prevent another
large-scale drinking water outbreak caused by Cryptosporidium
spp. (53–55).
Outbreak 2: In May 2000, an outbreak of bloody diarrhea
was noted in the Canadian town of Walkertown, Ontario.
A total of 2,300 people were thought to have developed gas-
trointestinal illness, 65 were hospitalized, and 6 died (8).
Over the ensuring weeks, the outbreak was found to involve
both E. coli O157:H7 (167 confirmed cases) and Campylo-
bacter spp. (116 confirmed cases). Twenty-seven people
developed the hemolytic-uremic syndrome, a serious compli-
cation of E. coli O157:H7 that can result in kidney failure.
The outbreak was thought to have been caused by conta-
mination of a groundwater well by runoff of fecal matter
from livestock farms after heavy rains. In addition to the acute
health consequences of waterborne infection, in the years
since the outbreak, long-term complications have been noted
to be more frequent among individuals who had acute
diarrhea at the time of the initial outbreak than Walkerton
residents who did not. These chronic health conditions
include irritable bowel syndrome (56), hypertension, and
renal impairment (57). In addition to highlighting chronic
sequellae of a waterborne disease outbreak, this outbreak
highlighted the potential for contamination of groundwater
during wet weather. The EPA Groundwater Rule of 2006
was promulgated in part to reduce the likelihood of a
Walkertown-type outbreak in the United States (58).
Outbreak 3: Three weeks after a Wisconsin restaurant
opened, 229 staff and patrons developed gastroenteritis, 6 of
whom were hospitalized (59). Of the 18 stool samples ana-
lyzed, 4 were positive for Norovirus genotype II, one for Cam-
pylobacter spp., and one for Salmonella enterica. A new,
recently inspected 85 m deep private drinking water well was
in operation, as was a septic system with a leach field 188 m
away from the well. After an extensive hydrologic investiga-
tion, it was determined that fractures in the dolomite forma-
tion in which the systems were located had allowed fairly
rapid transit (6–15 days) of tracer dye from the restaurant toi-
let and septic leach field to the drinking water well. This out-
break demonstrated that septic systems that meet all
regulatory requirements may not be adequate for public
health protection in geological formations that allow rela-
tively rapid movement of wastewater toward the drinking
water sources.
Gastrointestinal Illness Associated with
Drinking Water: Sporadic Cases
Researchers have tried to determine whether drinking water
distributed through public systems that meets regulatory
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-7
requirements cause sporadic cases of gastrointestinal illness.
The RCT approach has been used to answer that question,
with households randomized to receive standard tap water
versus water that receives additional treatment (reverse osmo-
sis, ozonation, ultraviolet radiation, and/or filtration). Five
studies conducted in the United States, Canada, and Aus-
tralia suggested that the risk of illness is lower among those
whose drinking water receives additional purification (18).
A large, rigorous study of drinking water and gastrointestinal
illness was conducted in the area of Davenport, Iowa. The
methods included randomization, blinding (individuals did
not know whether they were receiving water with enhanced
treatment) and a crossover phase (those who had initially
received an active enhanced treatment device later received
a sham device that delivered tap water, and those who began
with a sham device finished the study with an active device).
Among 1,296 participants who lived in 456 households
and were followed for one year, there was no reduction in
risk associated with home purification (29). This suggests
that if sporadic cases of gastrointestinal illness associated
with tap water that meets regulatory standards occurred,
they were too rare to be detected. Among older adults in
Sonoma County, California, sporadic cases of gastrointestinal
illness do appear to occur as a result of consumption of tap
water that meets regulatory requirements. This is based on a
study of adults ages 55 years and older, in which enhanced
filtration resulted in a reduction in gastrointestinal illness
occurrence. The group of older adults who used a sham water
treatment device had an observed incidence rate of 1.98 cases
of gastrointestinal illness per year, and this was reduced by
approximately 12% when an active enhanced water treat-
ment device was in use (though participants unaware of
whether they had an active device) (27).
Sporadic cases of viral gastroenteritis appear to occur
among users of groundwater. WAHTER (Water and Health
Trial for Enteric Risk), a set of recent environmental and
epidemiologic studies of community drinking water wells
(25, 60, 61) found that gastrointestinal illness was associated
with concentrations of norovirus and enteroviruses (but not
adenovirus) measured by qPCR in tap water samples (24).
Children appeared to be particularly susceptible to gastro-
intestinal illness during periods of high concentrations
of norovirus. WAHTER study data suggest that virus contam-
ination of groundwater wells and associated distribution sys-
tems accounts for 0–11% and 0.1–5% of sporadic cases of
acute gastrointestinal illness, respectively. Water distribution
systems were estimated to account for about 2–6 cases of gas-
trointestinal illness per 100 people a year (25). To put this in
context, approximately 65 cases of gastroenteritis per 100
people a year are thought to occur in the United States based
on data from a surveillance system for foodborne illness (62).
WATERBORNE RESPIRATORY INFECTIONS
An outbreak of pneumonia occurred among people who
spent time at a hotel in Philadelphia, Pennsylvania, where
the 1976 American Legion convention was held. A total of
182 people developed an acute respiratory illness, of whom
147 (81%) were hospitalized and 29 (16%) died (63).
An epidemiologic, clinical microbiology, and environmental
microbiology investigation identify the causative agent
and pathways of transmission. Silver staining of lung tissue
from patients with “Legionnaire’s disease” was cultured and
identified as a non–acid fast Gram-negative bacillus (64),
subsequently named Legionella pneumophila. Since then
L. pneumophila has been identified as the cause of many
outbreaks of respiratory disease, with water ( particularly
warm water) serving as a reservoir for the microbe. As sum-
marized in Table 1, Legionella accounted for more waterborne
outbreaks than any other bacteria, the majority of which were
linked to treated recreational water, such as hot tubs or spas.
In addition to the outbreak data presented in the table, seven
legionellosis outbreaks occurred in 2009–2010 linked to
“water not intended for drinking” or “water of unknown
intent.” Those outbreaks included 99 cases, of which 49
were hospitalized and 6 died. This fatality rate of 6 per 99 cases
is dramatically higher than that observed in outbreaks of gas-
trointestinal illness. Four of the seven outbreaks, which
resulted in six deaths, occurred in hospital/long-term care/
assisted living facilities. The outbreaks in two of these health
care–associated settings were linked to ornamental fountains.
Many prior outbreaks occurred in the context of health care
settings, with some linked to hospital air cooling systems
(65) and a decorative fountain in a hospital lobby (66).
Like the 1976 breaks of legionellosis in Philadelphia, many
have occurred outside of health care settings. The largest out-
break (64 cases, 17 hospitalizations) of the 2007–2010 period
occurred in a military facility and was linked to a cooling/air
conditioning systems. Hot tubs (67, 68) have been the setting
of numerous outbreaks. Community outbreaks have been
linked to aerosolized water from cooling towers several kilo-
meters away (69).
Lung infections by Mycobacterium spp. other than
M. tuberculosis are referred to as nontuberculous myco-
bacterial (NTM) infections, the majority of which are caused
by M. avium. M. avium pneumonia infections can be life-
threatening and are difficult to treat successfully. The annual
prevalence of NTM pneumonia has been increasing in recent
years and is estimated to be about 1.6–6.6 per 100,000 in
several areas of the United States (70); among those over
the age of 65, the prevalence is estimated to be much higher
at 47 cases per 100,000 (71).
L. pneumophila, NTM, Pseudomonas spp., and several
other microbes have been categorized as “opportunistic prem-
ise plumbing pathogens.” These phylogenetically diverse
bacteria share several characteristics that allow them to persist
and grow in pipes leading to homes and hotels (water distri-
bution systems), and the plumbing within those facilities.
These shared characteristics include chlorine resistance,
which can be several hundred times greater than that of
E. coli (72). Not only do some of these bacteria tolerate chlor-
ine, their tolerance appears to increase following chlorine
exposure. Another common characteristic of these respiratory
pathogens is that they persist and in many cases grow within
amoebae present in the biofilms of water distribution systems
and premise plumbing, in complex endosymbiotic relation-
ships (73, 74). Additionally, tolerance of low-oxygen and
low-carbon environments allows some of these pathogens,
such as M. avium, to increase in number with increasing dis-
tance from water treatment plants. Recent work in Florida
and Virginia demonstrates that nucleic acids from Legionella
spp., and less frequently M. avium, are commonly found in
home drinking water (75). Acanthamoebae were detected
in 4% of biofilm samples (but not in water samples) (75).
Several case reports have described exact genetic matches
(based on DNA sequencing) between M. avium present in
home (typically bathroom) plumbing of individuals with
M. avium pneumonia and isolates from the lungs of those
patients (76–78).
In addition to pulmonary infections, limited epidemio-
logic data point to biofilms as sources of pulmonary and eye
infections. Epidemiologic studies have also suggested that
3.1.4-8 ▪ WATER
keratitis, a vision-threatening infection of the cornea, can be
caused by Acanthamoeba spp. among contact lens users.
While these protozoa have contaminated contact lens solu-
tions, they are also commonly found in water distribution sys-
tem biofilms. It has been suggested that water distribution
system biofilms may contribute to the occurrence of Acantha-
moeba keratitis (79).
MICROBIAL TOXINS AND WATERBORNE
ILLNESS
In addition to causing illness by multiplying within hosts
(infection), waterborne microbes can cause illness as a result
of toxin production. The prokaryotes cyanobacteria and
algae, a diverse group of eukaryotes that includes diatoms
and dinoflagellates, are the most frequently recognized toxi-
genic microbes responsible for waterborne illness in humans.
Direct exposure pathway in include dermal contact, inges-
tion, or inhalation of toxin-containing surface water. Indirect
pathways include the consumption of finfish or shellfish that
have ingested toxin or the toxin-producing aquatic microbes.
Cyanobacteria and algae are widely present in surface
waters, and under certain conditions, they exhibit exuberant
growth and “bloom.” In freshwaters those blooms are visible
as a surface “scum.” In freshwater lakes, reservoirs, and other
impounded waters, conditions that promote blooms appear to
include sunlight, relatively high temperature, and nutrient
loading, particularly dissolved phosphorous (80). Species of
the Microcystis, Anabaena, Cylindrospermopsis, and other cya-
nobacterial genera may produce toxins in harmful algal
blooms (HABs), though the determinants of toxin produc-
tion in blooms are not well understood. Many cyanobacterial
species can produce multiple toxins, and the same toxin can
be produced by different cyanobacterial genera. Monitoring
of cyanobacterial harmful algal bloom monitoring is pre-
sented in detail in Chapter 3.1.2.
The toxins produced by cyanobacteria include cyclic
peptides, alkaloids, and lipopolysaccharides. Toxicity testing
in rodents demonstrated that several cyclic peptides toxins
produced by cyanobacteria are, on a microgram per kilogram
body weight basis, among the most lethal substances known.
Among the best studied cyanobacterial toxins are microcys-
tins, saxitoxins, and anatoxins. Acute effects of cyanotoxin
exposure in 11 bloom events that occurred in freshwater dur-
ing 2009–2010 were summarized by in a report from that
CDC’s Harmful Algal Bloom-Related Illness Surveillance
System (HABISS) (81). Skin and gastrointestinal symp-
toms may be similar to those observed in WBDOs caused
by infections. Unlike WBDOs, HAB-associated outbreaks
included constitutional (fever, fatigue, and malaise), respira-
tory, and neurological symptoms. Neurological symptoms
associated with HABS include dizziness, headache, and con-
fusion. Within a given outbreak, symptoms differed among
individuals, and many had symptoms referable to multiple
organ systems. No deaths were reported among the 61 people
who developed illness associated with HAB exposure, but a
majority of those for whom information was available visited
a health care professional (emergency department or clini-
cian’s office).
In a review of 2007–2011 HABISS data, 80 reports of
waterborne human illness (which excludes cases related to
fish consumption) included information regarding cyano-
bacteria, algae, or toxins (82). The most frequently identified
cyanobacteria, in decreasing order of detection were Microcys-
tis spp., followed by Anabaena; the most frequently detected
toxins were anatoxins, followed by microcystins. Recognized
algal blooms have become common in the United States
in recent decades, in particularly in Ohio (83). In August
2014, a large algal bloom resulted in measurable concen-
trations of microcystins in eastern Lake Erie, near Toledo,
Ohio. Approximately 400,000 people in three counties in
Ohio and one in the adjacent state of Michigan were advised
by not to drink tap water for two days.
Beyond case reports and outbreak investigations, epide-
miologic studies of HAB exposure have been limited. A
prospective cohort study was conducted at rivers and inland
lakes in Australia and Florida, in which water samples were
analyzed for cyanobacteria and selected cyanotoxins (84).
People who were in water containing relatively high concen-
trations of cyanobacteria were twice as likely to report respira-
tory symptoms as people who were in waters with relatively
low concentrations.
A disastrous outbreak of cyanobacterial intoxication
occurred in Brazil in which water used for hemodialysis had
become contaminated with microcystin (85). Of the 130
patients treated at an outpatient hemodialysis center, 116
abruptly developed visual disturbance, nausea, vomiting,
and other symptoms. Over the subsequent days, 50 died,
the majority due to acute liver failure. An extensive investi-
gation with input from environmental microbiologists, epi-
demiologists, nephrologists, and pathologists was able to
reconstruct the sequence of events that caused outbreak. As
a consequence of a drought, water had been transported by
truck from a reservoir to the dialysis center, where the water
was treated on site by sand filtration, carbon adsorption, dei-
onization, and micropore filtration. Microcystins were found
in reservoir water, water obtained from the delivery truck,
the water-holding tank and the carbon and ion resins used
at dialysis center, but they were not found at another dialysis
center in the same state that continued to receive treated
municipal water during the drought. Microcystins were also
found in liver tissue of 17 of victims.
Indirect Exposure to Waterborne Toxins
through Food
Among 458 cases of suspected and confirmed human illness
following exposure to cyanobacteria or algae, the most fre-
quent category of illness which accounted for ciguatera fish
poisoning in more than half the cases (82). A dinoflagellate
toxin, ciguatoxin, produced by Gambierdiscus toxicus, can
cause a constellation of neurologic and gastrointestinal
symptoms that can develop after one eats fish containing
high concentrations of the toxin. The fish are typically reef
fish from subtropical or tropical waters that are high on the
food chain, such as barracuda, black grouper, blackfin snap-
per, and others (http://www.cdc.gov/nceh/ciguatera/#about).
The toxin is not degraded significantly by cooking the
fish. Neurologic symptoms such as tingling sensations of the
face or extremities, dizziness, confusion, and perceiving cold
as heat, can persist for months or years following ingestion
of contaminated fish. The marine dinoflagellate Karenia
brevis can produce brevitoxin during periods of prolific
growth, which are referred to as “red tides” because of the
brownish-red coloration at the location of the bloom. Red
tides due to K. brevis occur in waters of the Gulf of Mexico
and can result in the death of marine animals and can directly
impact human health. An outbreak of respiratory symptoms
was reported in Florida among dredging workers who likely
had heavy exposures to brevitoxin (86).
In addition to acute health effects of toxins generated by
cyanobacteria and algae, the risk of cancer years after exposure
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-9
is a potential concern. The International Agency for
Research on Cancer (IARC) systematically reviewed the lit-
erature of and concluded that microcystin-LR appears to
be a tumor promoter, with evidence that the microcystin-LR
modulates oncogene activity. Several epidemiologic studies,
nearly all in China, suggest that drinking water from ponds
or ditches, locations where microcystins may be present,
increases the risk of colon and hepatocellular (liver) cancer.
Many of the epidemiologic study had significant limitations,
such inadequate microcystin-LR data and lack of adjustment
for exposure to known liver carcinogens, such as aflatoxin
and hepatitis B infection. Based on the available evidence,
which was considered inadequate to make a more definitive
classification, IARC concluded that microcystin-LR is a pos-
sible human carcinogen (Group 2b).
EMERGING CONCERNS AND METHODS
Climate Change and Waterborne Infectious Diseases
Waterborne disease outbreaks are expected to become more
frequent as the climate of much of the Northern Hemisphere
becomes warmer and wetter. A common feature of many
recent outbreaks is their association with heavy precipitation
and/or flooding (87). Because leptospira persist in freshwater
environment (88) and cause outbreaks of disease following
flooding (89–92), an increased burden of leptospirosis can
be expected. In recent decades the frequency of heavy rain
events in spring has increased, as have mean summer temper-
atures in the Midwest and Northeast portions of the United
States. These trends are expected to continue and perhaps
accelerate in coming decades (93). Both of these factors are
thought to promote the occurrence of HABs, as warmer
lake temperatures appear to promote bloom growth and
spring rains promote runoff of dissolved phosphates applied
to fields in fertilizers (80, 94).
In addition to outbreaks, occurrence of sporadic cases of
gastrointestinal illness has been studied in relation to the
occurrence of any rainfall (35), heavy rainfall (33), flooding
(95), the streamflow in a river (96), or other events, such as
beach closures (36), with most studies showing some asso-
ciation. In the United States (97) and Canada (98) scholars
and government agencies are promoting the development of
preparedness and adaptation strategies to reduce the public
health consequences climate change, including waterborne
gastrointestinal illness.
Emerging Epidemiologic Methods
Hospital discharge or emergency department data sets are
based on clinical diagnoses coded with the International
Classification of Disease system and are only available to
researchers months after a clinical encounter occurs. New
approaches to studying associations between water quality
(or precipitation) have utilized data that could be made avail-
able in near real time. For example, a recent study of drinking
water turbidity and gastrointestinal illness in New York City
made use of emergency department syndromic surveillance
data (99). In the United States, main two syndromic surveil-
lances systems are used by health departments (BioSense
and ESSENSE), which track in near real time the frequency
with which people present to emergency departments with
general categories of symptoms (in other words, syndromes)
rather than final diagnoses. With the widespread implemen-
tation of electronic medical records and the availability of
such data to public health departments, this approach holds
potential to conduct analyses of short-term variability in
water drinking water quality and in the occurrence of the
gastrointestinal syndrome. It also holds potential to identify
outbreaks of disease in real time (100) and to identify out-
breaks that otherwise might have gone undetected.
Pharmacy sales data of over-the-counter medication to
treat diarrhea or nausea been associated with gastrointestinal
syndromes (101) or diarrhea caused by norovirus (102).
Tracking Google searches for information about cold/flu
medication or flu symptoms has been shown to be a good pre-
dictor of public health data about influenza incidence (103).
It is possible that the application of analogous information
for gastrointestinal symptoms and medications could be useful
in identifying and understanding outbreaks of gastrointestinal
illness related to environmental exposures.
Improved Clinical Diagnostics
Methods to identify pathogens in water samples have impro-
ved tremendously in the past decade, and there is interest
in testing water samples for those pathogens, rather than
(or in addition to) testing for FIB. Unfortunately, the epide-
miologic studies that have sought to characterize patho-
gens responsible for sporadic illness have not succeeded
in that effort (49, 50). Even in the context of WBDOs,
pathogens frequently are not identified using routine clinical
testing. Promising methods for improving clinical and/or
public health diagnostic methods are under development.
Presently, clinicians must specify a particular pathogen to
be identified in a clinical test, rather than taking a “shotgun
approach” and ordering many bacterial, viral, and para-
site analyses. Multiplexed PCR methods should lead to
improved consistency in the number and range of relevant
pathogen included in routine testing. It also may lead to
improved assay sensitivity in pathogen detection in clinical
samples (104).
A limitation of PCR testing is that it begins with a decision
of which microbes one wants to detect, and then uses appro-
priate primers and probes. As a result, pathogens that are not
suspected (or not yet recognized as pathogens) would escape
detection. A metagenomic approach does not require an
a priori specification of microbes of interest. Metagenomic
analyses in this context entails blindly sequencing nucleic
acids found in stool samples of symptomatic individuals,
and then comparing sequences to those found in GenBank.
Metagenomic methods might improve sensitivity of detec-
tion and also lead to the identification of previously unrecog-
nized pathogens. Using metagenomics, sequences of eight
RNA viruses and one potentially pathogenic parasite were
identified in 31 stool samples obtained from 26 outbreaks of
gastrointestinal illness (not specifically waterborne) in New
Zealand in which routine clinical laboratory testing did
not identify pathogens (105). One of the viruses identified,
human parechovirus 3, had not previously been identified
in gastroenteritis outbreaks in Australasia, demonstrating
the value of a metagenomics approach.
A very different approach to identifying recent potentially
waterborne infections involves testing saliva samples for
specific antibodies, such as norovirus or Cryptosporidium
spp. (106, 107). Saliva samples were collected repeatedly
from individuals to identify a significant increase in the con-
centration of anti-norovirus antibody following norovirus
infection. This method may in the future prove useful in
cohort, clinical, and outbreak investigations, particularly if
multiplexing allows the detection of an increasing concen-
tration of any one of a panel of antibodies. Salivary anti-
body testing is appealing, in that it does not require stool or
blood testing.
3.1.4-10 ▪ WATER
SUMMARY
Sporadic cases of disease and recognized disease outbreaks
have been linked to treated recreational water (such as in
swimming pools and water parks) and untreated surface
waters (beaches, rivers, lakes). The recreational water out-
breaks most commonly involve acute gastrointestinal symp-
toms Additionally, contaminated drinking water and water
not intended for drinking (cooling systems of hotels and hos-
pitals) have caused disease outbreaks that results in symptoms
of the gastrointestinal or the respiratory tract. Pathogens
responsible for disease outbreaks include bacteria, viruses,
protozoa, and toxins elaborated by cyanobacteria. Pathogens
responsible for sporadic disease, which epidemiologic studies
suggest does occur, are not well characterized. Outbreaks of
disease, which have resulted in significant morbidity and
mortality, should be preventable through water/wastewater
treatment, maintenance of treatment and distribution sys-
tems, monitoring of water quality, and effective communica-
tion of water quality problems to the public. Future research
will improve our understanding of pathogens responsible for
illness and associations between climate and waterborne dis-
ease, and may lead to a better estimate of the burden of water-
borne disease.
REFERENCES
1. CDC. 2006. Surveillance for waterborne disease and out-
breaks associated with recreational water—United States,
2003–2004. MMWR Surveill Summ 55:1–30.
2. CDC. 2008. Surveillance for waterborne disease and out-
breaks associated with recreational water use and other
aquatic facility-associated health events—United States,
2005–2006. MMWR Surveill Summ 57:1–29.
3. Hlavsa MC, Roberts VA, Anderson AR, Hill VR, Kahler
AM, Orr M, Garrison LEH, L.A., Newton A, Hilborn
ED, Wade TJ, Beach MJ, Yoder JS. 2011. Surveillance
for waterborne disease outbreaks and other health events
associated with recreational water—United States, 2007–
2008. MMWR Surveill Summ 60:1–32.
4. CDC. 2013. Surveillance for waterborne disease outbreaks
associated with drinking water and other nonrecreational
water—United States, 2009–2010. MMWR Morb Mortal
Wkly Rep 62:714–720.
5. Hlavsa MC, Roberts VA, Kahler AM, Hilborn ED, Wade
TJ, Backer LC, Yoder JS, CDC. 2014. Recreational water-
associated disease outbreaks—United States, 2009–2010.
MMWR Morb Mortal Wkly Rep 63:6–10.
6. Schuster CJ, Ellis AG, Robertson WJ, Charron DF,
Aramini JJ, Marshall BJ, Medeiros DT. 2005. Infectious
disease outbreaks related to drinking water in Canada,
1974–2001. Can J Public Health 96:254–258.
7. Moffat H, Struck S. 2011. Water-borne Disease Outbreaks in
Canadian Small Drinking Water Systems. National Collabo-
rating Centres for Public Health, Vancouver, BC.
8. Bruce-Grey-Owen Sound Health Unit. 2000. The
Investigative Report of the Walkerton Outbreak of Waterborne
Gastroenteritis May–June 2000. Owen Sound, ON: Bruce-
Grey-Owen Sound Health Unit.
9. Public Health Agency of Canada. FoodNet Canada.
http://www.phac-aspc.gc.ca/foodnetcanada/index-eng.php.
Accessed June 24, 2015.
10. Scallan E, Jones TF, Cronquist A, Thomas S, Frenzen P,
Hoefer D, Medus C, Angulo FJ. 2006. Factors associated
with seeking medical care and submitting a stool sample
in estimating the burden of foodborne illness. Foodborne
Pathog Dis 3:432–438.
11. Flint JA, Dore K, Majowicz SE, Edge VL, Sockett P.
2004. From stool to statistics: reporting of acute gastrointes-
tinal illnesses in Canada. Can J Public Health 95:309–313.
12. Wade TJ, Sams E, Brenner KP, Haugland R, Chern E,
Beach M, Wymer L, Rankin CC, Love D, Li Q, Noble
R, Dufour AP. 2010. Rapidly measured indicators of rec-
reational water quality and swimming-associated illness at
marine beaches: a prospective cohort study. Environ Health
9:66.
13. Wade TJ, Calderon RL, Brenner KP, Sams E, Beach M,
Haugland R, Wymer L, Dufour AP. 2008. High sensitiv-
ity of children to swimming-associated gastrointestinal
illness: results using a rapid assay of recreational water qual-
ity. Epidemiology 19:375–383.
14. Dorevitch S, Pratap P, Wroblewski M, Hryhorczuk DO,
Li H, Liu LC, Scheff PA. 2012. Health risks of limited-
contact water recreation. Environ Health Perspect 120:
192–197.
15. Colford JM, Jr., Wade TJ, Schiff KC, Wright CC, Grif-
fith JF, Sandhu SK, Burns S, Sobsey M, Lovelace G,
Weisberg SB. 2007. Water quality indicators and the risk
of illness at beaches with nonpoint sources of fecal contam-
ination. Epidemiology 18:27–35.
16. Colford JM, Jr., Schiff KC, Griffith JF, Yau V, Arnold
BF, Wright CC, Gruber JS, Wade TJ, Burns S, Hayes
J, McGee C, Gold M, Cao Y, Noble RT, Haugland R,
Weisberg SB. 2012. Using rapid indicators for Enterococcus
to assess the risk of illness after exposure to urban runoff
contaminated marine water. Water Res 46:2176–2186.
17. Colford JM, Hilton JF, Wright CC, Arnold BF, Saha S,
Wade TJ, Scott J, Eisenberg JNS. 2010. Apparent benefit
of water filters may be an artifact of study design response.
American Journal of Public Health 100:1558–1559.
18. Colford JM, Jr., Roy S, Beach MJ, Hightower A, Shaw
SE, Wade TJ. 2006. A review of household drinking water
intervention trials and an approach to the estimation of
endemic waterborne gastroenteritis in the United States.
J Water Health 4(Suppl 2):71–88.
19. Arnold BF, Schiff KC, Griffith JF, Gruber JS, Yau V,
Wright CC, Wade TJ, Burns S, Hayes JM, McGee C,
Gold M, Cao YP, Weisberg SB, Colford JM. 2013.
Swimmer illness associated with marine water exposure
and water quality indicators impact of widely used assump-
tions. Epidemiology 24:845–853.
20. Wade TJ, Calderon RL, Sams E, Beach M, Brenner KP,
Williams AH, Dufour AP. 2006. Rapidly measured indi-
cators of recreational water quality are predictive of
swimming-associated gastrointestinal illness. Environ
Health Perspect 114:24–28.
21. Kay D, Fleisher JM, Salmon RL, Jones F, Wyer MD,
Godfree AF, Zelenauch-Jacquotte Z, Shore R. 1994. Pre-
dicting likelihood of gastroenteritis from sea bathing:
results from randomised exposure. Lancet 344:905–909.
22. Wiedenmann A, Kruger P, Dietz K, Lopez-Pila JM,
Szewzyk R, Botzenhart K. 2006. A randomized controlled
trial assessing infectious disease risks from bathing in fresh
recreational waters in relation to the concentration of
Escherichia coli, intestinal enterococci, Clostridium perfrin-
gens, and somatic coliphages. Environ Health Perspect
114:228–236.
23. Fleisher JM, Fleming LE, Solo-Gabriele HM, Kish JK,
Sinigalliano CD, Plano L, Elmir SM, Wang JD, Withum
K, Shibata T, Gidley ML, Abdelzaher A, He G, Ortega C,
Zhu X, Wright M, Hollenbeck J, Backer LC. 2010.
The BEACHES Study: health effects and exposures from
non-point source microbial contaminants in subtropical
recreational marine waters. Int J Epidemiol 39:1291–1298.
24. Borchardt MA, Spencer SK, Kieke BA, Lambertini E,
Loge FJ. 2012. Viruses in nondisinfected drinking water
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-11
from municipal wells and community incidence of acute
gastrointestinal illness. Environ Health Perspect 120:
1272–1279.
25. Lambertini E, Borchardt MA, Kieke BA, Jr., Spencer SK,
Loge FJ. 2012. Risk of viral acute gastrointestinal illness
from nondisinfected drinking water distribution systems.
Environ Sci Technol 46:9299–9307.
26. Hellard ME, Sinclair MI, Forbes AB, Fairley CK. 2001. A
randomized, blinded, controlled trial investigating the
gastrointestinal health effects of drinking water quality.
Environ Health Perspect 109:773–778.
27. Colford JM, Jr., Hilton JF, Wright CC, Arnold BF, Saha
S, Wade TJ, Scott J, Eisenberg JN. 2009. The Sonoma
water evaluation trial: a randomized drinking water inter-
vention trial to reduce gastrointestinal illness in older
adults. Am J Public Health 99:1988–1995.
28. Colford JM, Jr., Saha SR, Wade TJ, Wright CC, Vu M,
Charles S, Jensen P, Hubbard A, Levy DA, Eisenberg
JN. 2005. A pilot randomized, controlled trial of an
in-home drinking water intervention among HIV + per-
sons. J Water Health 3:173–184.
29. Colford JM, Jr., Wade TJ, Sandhu SK, Wright CC, Lee S,
Shaw S, Fox K, Burns S, Benker A, Brookhart MA, van
der Laan M, Levy DA. 2005. A randomized, controlled
trial of in-home drinking water intervention to reduce gas-
trointestinal illness. Am J Epidemiol 161:472–482.
30. Fleisher JM, Kay D. 2006. Risk perception bias, self-
reporting of illness, and the validity of reported results in
an epidemiologic study of recreational water associated ill-
nesses. Mar Pollut Bull 52:264–268.
31. Dorevitch S, DeFlorio-Barker S, Jones RM, Liu L. 2015.
Water quality as a predictor of gastrointestinal illness fol-
lowing incidental contact water recreation. Water Res 83:
94–103.
32. Lin CJ, Wade TJ, Hilborn ED. 2015. Flooding and Clos-
tridium difficile infection: a case-crossover analysis. Int J
Environ Res Public Health 12:6948–6964.
33. Jagai JS, Li Q, Wang S, Messier KP, Wade TJ, Hilborn
ED. 2015. Extreme precipitation and emergency room vis-
its for gastrointestinal illness in areas with and without
combined wewer systems: an analysis of Massachusetts
Data, 2003–2007. Environ Health Perspect 123:873–879,
doi: 10.1289/ehp.1408971.
34. Carlton EJ, Eisenberg JN, Goldstick J, Cevallos W,
Trostle J, Levy K. 2014. Heavy rainfall events and diarrhea
incidence: the role of social and environmental factors.
Am J Epidemiol 179:344–352.
35. Drayna P, McLellan SL, Simpson P, Li SH, Gorelick MH.
2010. Association between rainfall and pediatric emer-
gency department visits for acute gastrointestinal illness.
Environ Health Perspect 118:1439–1443.
36. Bush KF, Fossani CL, Li S, Mukherjee B, Gronlund CJ,
O’Neill MS. 2014. Extreme precipitation and beach clo-
sures in the great lakes region: evaluating risk among the
elderly. Int J Environ Res Public Health 11:2014–2032.
37. Dorevitch S, Ashbolt NJ, Ferguson CM, Fujioka R,
McGee CD, Soller JA, Whitman RL. 2010. Meeting
report: knowledge and gaps in developing microbial criteria
for inland recreational waters. Environ Health Perspect 118:
871–876.
38. Stern EJ, Galloway R, Shadomy SV, Wannemuehler K,
Atrubin D, Blackmore C, Wofford T, Wilkins PP, Ari
MD, Harris L, Clark TA. 2010. Outbreak of leptospirosis
among Adventure Race participants in Florida, 2005. Clin
Infect Dis 50:843–849.
39. Morgan J, Bornstein SL, Karpati AM, Bruce M, Bolin
CA, Austin CC, Woods CW, Lingappa J, Langkop C,
Davis B, Graham DR, Proctor M, Ashford DA, Bajani
M, Bragg SL, Shutt K, Perkins BA, Tappero JW,
Leptospirosis Working Group. 2002. Outbreak of lepto-
spirosis among triathlon participants and community resi-
dents in Springfield, Illinois, 1998. Clin Infect Dis 34:
1593–1599.
40. Brockmann S, Piechotowski I, Bock-Hensley O, Winter
C, Oehme R, Zimmermann S, Hartelt K, Luge E, Nockler
K, Schneider T, Stark K, Jansen A. 2006. Outbreak of lep-
tospirosis among triathlon participants in Germany, 2006.
BMC Infect Dis 10:91.
41. Mori M, van Esbroeck M, Depoorter S, Decaluwe W,
Vandecasteele SJ, Fretin D, Reynders M. 2015. Outbreak
of leptospirosis during a scout camp in the Luxembourg Bel-
gian province, Belgium, summer 2012. Epidemiol Infect
143:1761–1766.
42. Radl C, Muller M, Revilla-Fernandez S, Karner-Zuser S,
de Martin A, Schauer U, Karner F, Stanek G, Balcke P,
Hallas A, Frank H, Furnschlief A, Erhart F, Allerberger
F. 2011. Outbreak of leptospirosis among triathlon partici-
pants in Langau, Austria, 2010. Wien Klin Wochenschr 123:
751–755.
43. U.S. EPA. 1983. Health Effects Criteria for Marine Rec-
reational Waters. http://nepis.epa.gov.
44. U.S. EPA. 1984. Health Effects Criteria for Fresh Recrea-
tional Waters. http://water.epa.gov/scitech/swguidance/
standards/upload/1999_11_03_criteria_healtheffects.pdf.
45. U.S. EPA. 1986. Ambient Water Quality Criteria for
Bacteria—1986. http://water.epa.gov/scitech/swguidance/
standards/upload/2001_10_12_criteria_ambientwqc_bac
teria1986.pdf.
46. Lee JV, Dawson SR, Ward S, Surman SB, Neal KR. 1997.
Bacteriophages are a better indicator of illness rates than
bacteria amongst users of a white water course fed by a low-
land river. Water Sci Technol 35:165–170.
47. van Asperen I, Medema G, Borgdorff MW, Sprenger
MJW, Havelaar AH. 1998. Risk of gastroenteritis among
triathletes in relation to fecal pollution of freshwaters. Int
J Epidemiol 27:308–315.
48. U.S. EPA. 2015. Review of Coliphages as Possible Indica-
tors of Fecal Contamination for Ambient Water Quality
820-R-15-098.
49. Jones F, Kay D, Stanwell-Smith R, Wyer M. 1991. Results
of the first pilot-scale controlled cohort epidemiological
investigation into possible health effects of bathing in sea-
water at Langland Bay, Swansea. J Inst Water Env Manage-
ment 91–98.
50. Dorevitch S, Dworkin MS, DeFlorio SA, Janda WM,
Wuellner J, Hershow RC. 2012. Enteric pathogens
in stool samples of Chicago-area water recreators with
new-onset gastrointestinal symptoms. Water Res 46:
4961–4972.
51. Yau V, Wade TJ, de Wilde CK, Colford JM. 2009. Skin-
related symptoms following exposure to recreational water:
a systematic review and meta-analysis. Water Qual Expo
Health 1:79–103.
52. MacKenzie WR, Schell WL, Blair KA, Addiss DG,
Peterson DE, Hoxie NJ, Kazmierczak JJ, Davis JP.
1995. Massive outbreak of waterborne cryptosporidium
infection in Milwaukee, Wisconsin: recurrence of illness
and risk of secondary transmission. Clin Infect Dis 21:
57–62.
53. U.S. EPA. 1998. National Primary Drinking Water Regu-
lations: Interim Enhanced Surface Water Treatment. Fed-
eral Register December 16, http://www.epa.gov/ogwdw/
mdbp/ieswtrfr.pdf.
54. U.S. EPA. 2002. National Primary Drinking Water
Regulations: Long Term 1 Enhanced Surface Water Treat-
ment Rule; Final Rule. Federal Register January 14, http://
www.epa.gov/fedrgstr/EPA-WATER/2002/January/Day-14/
w409.pdf.
3.1.4-12 ▪ WATER
55. U.S. EPA. 2006. National Primary Drinking Water Regu-
lations: Long Term 2 Enhanced Surface Water Treatment
Rule. Federal Register June 4, http://www.gpo.gov/fdsys/
pkg/FR-2006-01-05/pdf/06-4.pdf.
56. Marshall JK, Thabane M, Garg AX, Clark WF, Moayyedi
P, Collins SM. 2010. Eight year prognosis of postinfectious
irritable bowel syndrome following waterborne bacterial
dysentery. Gut 59:605–611.
57. Clark WF, Sontrop JM, Macnab JJ, Salvadori M, Moist L,
Suri R, Garg AX. 2010. Long term risk for hyperten-
sion, renal impairment, and cardiovascular disease after
gastroenteritis from drinking water contaminated with
Escherichia coli O157:H7: a prospective cohort study. BMJ
341:c6020.
58. U.S. EPA. 2006. National Primary Drinking Water Regu-
lations: Ground Water Rule. http://www.epa.gov/fedrgstr/
EPA-WATER/2006/November/Day-08/w8763.pdf.
59. Borchardt MA, Bradbury KR, Alexander EC, Jr., Kol-
berg RJ, Alexander SC, Archer JR, Braatz LA, Forest
BM, Green JA, Spencer SK. 2011. Norovirus outbreak
caused by a new septic system in a dolomite aquifer. Ground
Water 49:85–97.
60. Borchardt MA, Bertz PD, Spencer SK, Battigelli DA.
2003. Incidence of enteric viruses in groundwater from
household wells in Wisconsin. Appl Environ Microbiol 69:
1172–1180.
61. Lambertini E, Spencer SK, Kieke BA, Jr., Loge FJ,
Borchardt MA. 2011. Virus contamination from operation
and maintenance events in small drinking water distribu-
tion systems. J Water Health 9:799–812.
62. Roy SL, Scallan E, Beach MJ. 2006. The rate of acute gas-
trointestinal illness in developed countries. J Water Health 4
(Suppl 2):31–69.
63. Fraser DW, Tsai TR, Orenstein W, Parkin WE, Beecham
HJ, Sharrar RG, Harris J, Mallison GF, Martin SM,
McDade JE, Shepard CC, Brachman PS. 1977. Legion-
naires’ disease: description of an epidemic of pneumonia.
N Engl J Med 297:1189–1197.
64. McDade JE, Shepard CC, Fraser DW, Tsai TR, Redus
MA, Dowdle WR. 1977. Legionnaires’ disease: isolation
of a bacterium and demonstration of its role in other respi-
ratory disease. N Engl J Med 297:1197–1203.
65. Dondero TJ, Jr., Rendtorff RC, Mallison GF, Weeks RM,
Levy JS, Wong EW, Schaffner W. 1980. An outbreak
of Legionnaires’ disease associated with a contaminated
air-conditioning cooling tower. N Engl J Med 302:
365–370.
66. Palmore TN, Stock F, White M, Bordner M, Michelin A,
Bennett JE, Murray PR, Henderson DK. 2009. A cluster
of cases of nosocomial legionnaires disease linked to a con-
taminated hospital decorative water fountain. Infect Control
Hosp Epidemiol 30:764–768.
67. Burnsed LJ, Hicks LA, Smithee LM, Fields BS, Bradley
KK, Pascoe N, Richards SM, Mallonee S, Littrell L, Ben-
son RF, Moore MR. 2007. A large, travel-associated out-
break of legionellosis among hotel guests: utility of the
urine antigen assay in confirming Pontiac fever. Clin Infect
Dis 44:222–228.
68. CDC. 2011. Surveillance for waterborne disease outbreaks
associated with drinking water—United States, 2007–
2008. MMWR Surveill Summ 60:38–68.
69. Nguyen TM, Ilef D, Jarraud S, Rouil L, Campese C, Che
D, Haeghebaert S, Ganiayre F, Marcel F, Etienne J,
Desenclos JC. 2006. A community-wide outbreak of
Legionnaires disease linked to industrial cooling towers—
how far can contaminated aerosols spread? J Infect Dis
193:102–111.
70. Beumer A, King D, Donohue M, Mistry J, Covert T,
Pfaller S. 2010. Detection of Mycobacterium avium subsp.
paratuberculosis in drinking water and biofilms by quantita-
tive PCR. Appl Environ Microbiol 76:7367–7370.
71. Adjemian J, Olivier KN, Seitz AE, Holland SM, Prevots
DR. 2012. Prevalence of nontuberculous mycobacterial
lung disease in U.S. Medicare beneficiaries. Am J Respir
Crit Care Med 185:881–886.
72. Falkinham JO, III. 2015. Common features of opportunis-
tic premise plumbing pathogens. Int J Environ Res Public
Health 12:4533–4545.
73. Buse HY, Ashbolt NJ. 2011. Differential growth of Legion-
ella pneumophila strains within a range of amoebae at various
temperatures associated with in-premise plumbing. Lett
Appl Microbiol 53:217–224.
74. Thomas JM, Ashbolt NJ. 2011. Do free-living amoebae in
treated drinking water systems present an emerging health
risk? Environ Sci Technol 45:860–869.
75. Wang H, Edwards M, Falkinham JO, III, Pruden A.
2012. Molecular survey of the occurrence of Legionella
spp., Mycobacterium spp., Pseudomonas aeruginosa, and
amoeba hosts in two chloraminated drinking water distri-
bution systems. Appl Environ Microbiol 78:6285–6294.
76. Nishiuchi Y, Maekura R, Kitada S, Tamaru A, Taguri T,
Kira Y, Hiraga T, Hirotani A, Yoshimura K, Miki M, Ito
M. 2007. The recovery of Mycobacterium avium-intracel-
lulare complex (MAC) from the residential bathrooms
of patients with pulmonary MAC. Clin Infect Dis 45:
347–351.
77. Nishiuchi Y, Tamura A, Kitada S, Taguri T, Matsumoto
S, Tateishi Y, Yoshimura M, Ozeki Y, Matsumura N,
Ogura H, Maekura R. 2009. Mycobacterium avium
complex organisms predominantly colonize in the bath-
tub inlets of patients’ bathrooms. Jpn J Infect Dis 62:
182–186.
78. Falkinham JO, III. 2011. Nontuberculous mycobacteria
from household plumbing of patients with nontuberculous
mycobacteria disease. Emerg Infect Dis 17:419–424.
79. Joslin CE, Tu EY, Shoff ME, Anderson RJ, Davis FG.
2010. Shifting distribution of Chicago-area Acanthamoeba
keratitis cases. Arch Ophthalmol 128:137–139.
80. Michalak AM, Anderson EJ, Beletsky D, Boland S,
Bosch NS, Bridgeman TB, Chaffin JD, Cho K, Confesor
R, Daloglu I, Depinto JV, Evans MA, Fahnenstiel GL, He
L, Ho JC, Jenkins L, Johengen TH, Kuo KC, Laporte E,
Liu X, McWilliams MR, Moore MR, Posselt DJ,
Richards RP, Scavia D, Steiner AL, Verhamme E,
Wright DM, Zagorski MA. 2013. Record-setting algal
bloom in Lake Erie caused by agricultural and meteorolog-
ical trends consistent with expected future conditions. Proc
Natl Acad Sci USA 110:6448–6452.
81. Hilborn ED, Roberts VA, Backer L, Deconno E, Egan JS,
Hyde JB, Nicholas DC, Wiegert EJ, Billing LM, Diorio
M, Mohr MC, Hardy JF, Wade TJ, Yoder JS, Hlavsa
MC, CDC. 2014. Algal bloom-associated disease out-
breaks among users of freshwater lakes—United States,
2009–2010. MMWR Morb Mortal Wkly Rep 63:11–15.
82. Backer LC, Manassaram-Baptiste D, LePrell R, Bolton B.
2015. Cyanobacteria and algae blooms: review of health
and environmental data from the Harmful Algal Bloom-
Related Illness Surveillance System (HABISS) 2007–
2011. Toxins (Basel) 7:1048–1064.
83. Cheung MY, Liang S, Lee J. 2013. Toxin-producing cya-
nobacteria in freshwater: a review of the problems, impact
on drinking water safety, and efforts for protecting public
health. J Microbiol 51:1–10.
84. Stewart I, Webb PM, Schluter PJ, Fleming LE, Burns JW,
Jr., Gantar M, Backer LC, Shaw GR. 2006. Epidemiology
of recreational exposure to freshwater cyanobacteria—an
international prospective cohort study. BMC Public Health
6:93.
 3.1.4. Epidemiologic Aspects of Waterborne Infectious Disease ▪ 3.1.4-13
85. Jochimsen EM, Carmichael WW, An JS, Cardo DM,
Cookson ST, Holmes CE, Antunes MB, de Melo Filho
DA, Lyra TM, Barreto VS, Azevedo SM, Jarvis WR.
1998. Liver failure and death after exposure to microcystins
at a hemodialysis center in Brazil. N Engl J Med 338:
873–878.
86. CDC. 2008. Illness associated with red tide—Nassau
County, Florida, 2007. MMWR Morb Mortal Wkly Rep
57:717–720.
87. Cann KF, Thomas DR, Salmon RL, Wyn-Jones AP, Kay
D. 2013. Extreme water-related weather events and water-
borne disease. Epidemiol Infect 141:671–686.
88. Andre-Fontaine G, Aviat F, Thorin C. 2015. Waterborne
leptospirosis: survival and preservation of the virulence of
pathogenic Leptospira spp. in fresh water. Curr Microbiol
71:136–142.
89. Gaynor K, Katz AR, Park SY, Nakata M, Clark TA,
Effler PV. 2007. Leptospirosis on Oahu: an outbreak asso-
ciated with flooding of a university campus. Am J Trop Med
Hyg 76:882–885.
90. Naranjo M, Suarez M, Fernandez C, Amador N, Gonza-
lez M, Batista N, Gonzalez I, Valdes Y, Infante JF, Sierra
G. 2008. Study of a leptospirosis outbreak in Honduras fol-
lowing Hurricane Mitch and prophylactic protection of the
vax-SPIRAL(R) vaccine. MEDICC Rev 10:38–42.
91. Dechet AM, Parsons M, Rambaran M, Mohamed-
Rambaran P, Florendo-Cumbermack A, Persaud S,
Baboolal S, Ari MD, Shadomy SV, Zaki SR, Paddock
CD, Clark TA, Harris L, Lyon D, Mintz ED. 2012. Lep-
tospirosis outbreak following severe flooding: a rapid assess-
ment and mass prophylaxis campaign; Guyana, January–
February 2005. PLoS One 7:e39672.
92. Smith JK, Young MM, Wilson KL, Craig SB. 2013. Lep-
tospirosis following a major flood in Central Queensland,
Australia. Epidemiol Infect. 141:585–590.
93. Walsh J, Wuebbles D, Hayhoe K, Kossin J, Kunkel K,
Stephens G, Thorne P, Vose R, Wehner M. 2014. Our
Changing Climate. Climate Change Impacts in the United
States: The Third National Climate Assessment, p 19–67. In
Melillo J, Richmond TC, Yohe GW (eds), U.S. Global
Change Research Program.
94. Stow CA, Cha Y, Johnson LT, Confesor R, Richards RP.
2015. Long-term and seasonal trend decomposition of
Maumee River nutrient inputs to western Lake Erie. Envi-
ron Sci Technol 49:3392–3400.
95. Wade TJ, Lin CJ, Jagai JS, Hilborn ED. 2014. Flooding
and emergency room visits for gastrointestinal illness
in Massachusetts: a case-crossover study. PLoS One 9:
e110474.
96. Jagai JS, Griffiths JK, Kirshen PK, Webb P, Naumova
EN. 2012. Seasonal patterns of gastrointestinal illness
and streamflow along the Ohio River. Int J Environ Res Pub-
lic Health 9:1771–1790.
97. CDC. Climate and Health Resources. http://www.cdc.gov/
climateandhealth/site_resources.htm.
98. Charron D, Thomas M, Waltner-Toews D, Aramini J,
Edge T, Kent R, Maarouf A, Wilson J. 2004. Vulnerability
of waterborne diseases to climate change in Canada: a
review. J Toxicol Environ Health A 67:1667–1677.
99. Hsieh JL, Nguyen TQ, Matte T, Ito K. 2015. Drinking
water turbidity and emergency department visits for gastro-
intestinal illness in New York City, 2002–2009. PLoS One
10:e0125071.
100. Marx MA, Rodriguez CV, Greenko J, Das D,
Heffernan R, Karpati AM, Mostashari F, Balter S, Layton
M, Weiss D. 2006. Diarrheal illness detected through
syndromic surveillance after a massive power outage:
New York City, August 2003. Am J Public Health 96:
547–553.
101. Edge VL, Pollari F, Lim G, Aramini J, Sockett P, Martin
SW, Wilson J, Ellis A. 2004. Syndromic surveillance of
gastrointestinal illness using pharmacy over-the-counter
sales. A retrospective study of waterborne outbreaks
in Saskatchewan and Ontario. Can J Public Health 95:
446–450.
102. Edge VL, Pollari F, Ng LK, Michel P, McEwen SA, Wil-
son JB, Jerrett M, Sockett PN, Martin SW. 2006. Syn-
dromic surveillance of norovirus using over-the-counter
sales of medications related to gastrointestinal illness. Can
J Infect Dis Med Microbiol 17:235–241.
103. Ginsberg J, Mohebbi MH, Patel RS, Brammer L,
Smolinski MS, Brilliant L. 2009. Detecting influenza
epidemics using search engine query data. Nature 457:
1012–1014.
104. Becker SL, Chatigre JK, Gohou JP, Coulibaly JT, Leuppi
R, Polman K, Chappuis F, Mertens P, Herrmann M,
N’Goran EK, Utzinger J, von Muller L. 2015. Combined
stool-based multiplex PCR and microscopy for enhanced
pathogen detection in patients with persistent diarrhoea
and asymptomatic controls from Cote d’Ivoire. Clin Micro-
biol Infect 21:e591.
105. Moore NE, Wang J, Hewitt J, Croucher D, Williamson
DA, Paine S, Yen S, Greening GE, Hall RJ. 2015. Meta-
genomic analysis of viruses in feces from unsolved out-
breaks of gastroenteritis in humans. J Clin Microbiol 53:
15–21.
106. Griffin SM, Chen IM, Fout GS, Wade TJ, Egorov AI.
2011. Development of a multiplex microsphere immunoas-
say for the quantitation of salivary antibody responses
to selected waterborne pathogens. J Immunol Meth 364:
83–93.
107. Griffin SM, Converse RR, Leon JS, Wade TJ, Jiang X,
Moe CL, Egorov AI. 2015. Application of salivary anti-
body immunoassays for the detection of incident infections
with Norwalk virus in a group of volunteers. J Immunol
Meth 424:53–63, doi: 10.1016/j.jim.2015.05.001.
 Waterborne Enteric Viruses: Diversity,
Distribution, and Detection
MORTEZA ABBASZADEGAN AND ABSAR ALUM
3.1.5
Historically, human societies have flourished in areas with
plentiful water resources, and so waterborne viruses have
long been afflicting the human population (1–3). Egyptian
mummies showing shriveled legs chronicle the impact of
waterborne polioviruses in ancient times. The ancient
records chronicle sporadic incidence of waterborne ( polio)
virus (4). It was not until the 1800s that outbreaks of water-
borne polioviruses were recorded by noticing clusters of para-
lysis in infants and small children. In the United States the
first outbreak of poliovirus was recorded in the mid-1800s
in Louisiana (5). The first confirmed waterborne virus out-
break was reported in India in 1955–1956 resulting in
29,300 cases of jaundice (6). The outbreak has been retro-
spectively linked to hepatitis E virus.
DIVERSITY
A steady development in virological methods during the 20th
century has resulted in the identification and characterization
of several new waterborne viruses (Fig. 1; Table 1) (7, 8).
Poliovirus was identified in the early 1900s. Since then
many enteric viruses have been identified, and the most
recent additions to the list of viral gastroenteritis are the sali-
virus/klasseviruses (9).
Enteric viruses are a leading cause of gastroenteritis (10),
and they constitute a diverse group. The viruses commonly
implicated in gastroenteritis (cases or outbreaks) are classified
into the families Picornaviridae ( polioviruses, enteroviruses,
coxsackieviruses, and echoviruses), Adenoviridae (adeno-
virus 40 and adenovirus 41), Caliciviridae (noroviruses, sapo-
viruses), and Reoviridae (reoviruses). Although enteric virus
infections are mainly associated with diarrhea and self-
limiting gastroenteritis in healthy humans, they have also
been linked to aseptic meningitis, encephalitis, myocarditis,
and insulin-dependent diabetes (11, 12).
Currently, about 140 waterborne viruses are known to
infect humans (13). The waterborne viruses not only show
wide diversity in the illness (including diarrhea, fever, hepa-
titis, paralysis, meningitis, respiratory, and heart diseases; [8])
they also show broad diversity in their size, shape, infection
process, and replication mechanisms (Table 1) (14). Most
waterborne enteric viruses have an icosahedral capsid; how-
ever, the structural feature on the surface of capsids varies
among different virus groups (Fig. 2) (15). The rotavirus cap-
sid is composed of three concentric protein layers and the
doi:10.1128/9781555818821.ch3.1.5
3.1.5-1
outermost layer comprises the proteins VP4 and VP7. The
PV4 forms spike-like structures that are responsible for virus
attachment to the host cell. Adenovirus (Ad) has an icosahe-
dral capsid composed of hexon the major capsid protein. In
addition, the capsid has pentons with elongated fiber projec-
tions. The distal ends of these fibers have a globular “knob,”
which functions as the major attachment site for host cell
receptors. In general, the members of enteroviruses and cali-
civiruses/norovirus have capsids with less pronounced surface
features; members of these groups have significantly different
annotations of their capsids (14).
Viral diversity studies have been hampered by the lack of
effective methods for simultaneously detecting different types
of viruses and it has been mainly due to the limitation of cell
culture and PCR methods (16, 17). Recent developments in
metagenomic analyses of microorganisms have broadened the
scope of such studies. Such analyses have shown that coinfec-
tions (subjects simultaneously infected with multiple patho-
gens) are a more frequent phenomenon than previously
thought (18, 19). Population studies have shown coinfec-
tions of viral species particularly members of the Picornaviri-
dae (20–25), Astroviridae (18, 22, 26), Parvoviridae (26, 27),
and Circoviridae (28) families of enteric viruses.
In general enteric viruses are species-specific with narrow
host ranges (29); however, the application of advanced
molecular techniques has identified some enteric viruses
that can infect both humans and animals. A recent study
has shown that many circoviruses infecting farm animals
are also commonly detected in the feces of humans and
chimpanzees living in that environment (30). The enteric
virus host range or the host specificity is dictated by the epit-
opes on the viral capsid and their ability to recognize host
cell receptors (31). For example, coronaviruses are known
for their broad host range including humans, pigs, and cats.
These viruses use aminopeptidase N (APN) of their host
as receptor for infection initiation. The aminopeptidase N of
human (hAPN), pigs ( pAPN), and cats (fAPN) differ in
some areas of N-glycosylation. The levels of enterocyte APN
glycosylation control the binding of virus to their receptors at
the cellular level (32). In addition to APN, other molecules
such as CD55, CD155, ICAM-1, sialic acid, integrin, Ig-like
molecules on host cell surface have been widely reported as
receptors for enteric viruses (33). A slight variation in the
sequences of these molecules may change the susceptibility
status of the host cells for different viruses (34, 35).
3.1.5-2 ▪ WATER

FIGURE 1 Chronology of milestones in waterborne viruses. doi:10.1128/9781555818821.ch3.1.5.f1

The majority of human enteric viruses are RNA viruses,
and in general, RNA virus populations sustain high genetic
diversity due to the low fidelity of their polymerase, short
genome, high replication rates, and large population size
(36). In RNA viruses, the high mutation rate increases their
ability to adapt to diverse hosts and the potential to cause
new disease in humans or animals. RNA recombination is
a common mechanism for evolution of viruses, and genetic
recombination within the nonstructural regions of the
genome (P2 and P3) has been observed only among mem-
bers of the same species of enteric viruses (37). Genetic
recombination is particularly a major issue in viruses with
multisegment genome, such as rotaviruses with a genome
consisting 11 segments of double-stranded RNA coding six
structural and six nonstructural proteins (38). Given the
segmented nature of the genome, the genes encoding VP7
and VP4 can segregate independently and therefore serves
as the bases for a binary classification system. Rotaviruses
are serotyped based on antigenic differences in the VP7
and VP4 proteins, which provides the basis for strain classi-
fication into G (VP7, glycoprotein) and P (VP4, protease-
sensitive) serotypes (38). Among the enteric viruses with
nonsegmented genome, noroviruses show the highest level
of genetic mutation (39–41). In case of noroviruses GII.4,
the most prevalent genotype, the highest rate of mutation/
recombination has been reported in the P2 domain (42).
However, it is important to point out that in enteric viruses
new mutations in existing strains are “selected” from a preex-
isting pool, rather than through de novo mutations (31).
Viruses showing high mutation rates pose a unique challenge

TABLE 1 Genomic and morphological characteristics of major enteric viruses

Virus Genus, family Genome Capsid shape Capsid size (nm)
Aichi virus Kobuvirus, ssRNA + Icosahedral 27–30
Picornaviridae
Coxsackievirus Enterovirus, ssRNA + Icosahedral 27–30
Picornaviridae
Echovirus Enterovirus, ssRNA + Icosahedral 24–30
Picornaviridae
Hepatitis A virus Hepatovirus ssRNA + Icosahedral 28–30
Picornaviridae.
Hepatitis E virus Hepevirus, ssRNA + Icosahedral 30
Unassigned
Human adenovirus Mastadenovirus, dsDNA Icosahedral with 70–100
Adenoviridae protruding knob
Human astrovirus Mamastrovirus, ssRNA + Icosahedral star 27–32 (35)
Astroviridae
Norwalk virus Norovirus, ssRNA + Icosahedral 27–30
Caliciviridae
Poliovirus Enterovirus, ssRNA + Icosahedral 27–30
Picornaviridae
Rotavirus A, B, & C Rotavirus, dsRNA + Icosahedral with spokes 60–80
Reoviridae
Sapporo virus Sapovirus, ssRNA + Icosahedral 27–30
Caliciviridae
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-3
FIGURE 2 Shapes and structures of virions of major groups of enteric viruses. Source: (15) Images used with permission from Philippe Le
Mercier, SIB Swiss Institute of Bioinformatics, http://viralzone.expay.org. doi:10.1128/9781555818821.ch3.1.5.f2
for developing effective control measures and robust detec-
tion platforms.
DISTRIBUTION
Waterborne viruses known to infect humans are predomi-
nantly enteric in nature, which are invariably transmitted
through the fecal–oral route (8). Each gram of fecal material
from an infected person may contain billions of viral particles
(43). The massive quantities of viruses shed in the feces of
infected subjects may find their way to surface waters and fin-
ished drinking water by a variety of routes. Some of the water-
borne viruses can also be transmitted through other modes of
transmissions, such as foods and contaminated surface (44).
The distribution and modes of transmission of enteric viruses
are known to be impacted by the type of host population, san-
itary conditions, and living conditions. For example, rotavi-
rus infections are mostly prevalent in children, and human
norovirus outbreaks are predominantly associated with close,
crowded settings. Another example of limiting factor in viral
infection in humans is histo-blood group (HBG), as HBG
nonsecretors (individuals who do not secrete HBG) are not
prone to infection by human norovirus G1.1.
In the United States, a significant decrease in the pre-
valence of waterborne diseases has been recorded since
the introduction of water treatment technologies, but
still approximately 19.5 million cases occur annually (45).
Similar trends have been observed in other developed
countries. However, waterborne viral gastroenteritis is still
prevalent in developing and underdeveloped countries
that lack safe drinking water supplies and sanitary condi-
tions (46). A marked difference is seen in the types of water-
borne viruses prevalent in developed and underdeveloped
countries (47).
In a recent study, rotaviruses were found in each geo-
graphic region around the world; whereas noroviruses were
found across the world with the exception of four regions:
South Korea, Singapore, Sri Lanka, and Democratic Republic
of the Congo (Table 2) (47).
Major groups of waterborne viruses are known to have
multiple genotypes, and the pattern of their prevalence varies
globally. In general, enteroviruses have a global distribution,
with some levels of serotype/genotype variation. Some sero-
types may be endemic to a given geographic locality, where
it may persist for years (50). The viral serotype/genotype com-
plex may periodically change by the introductions of new
genotypes or mutations. Geographical variation in the prev-
alence of different enteric viruses and their genogroups is well
known. Rotavirus, the major cause of diarrhea in young
children, is more frequently detected in the western Pacific
region (33%) and less in the American region (23%) (51).
Aichi virus, a globally emerging enteric virus, was originally
detected in Japan in 1989. Since then it has been reported
in various countries across Asia (52); however, in recent years
this virus has been detected in sewage in Europe, South
America, and Africa (53–55). Hepatitis E virus (HEV)
3.1.5-4 ▪ WATER

TABLE 2 The relative frequencies (%) of rotaviruses and sporadic occurrence, low concentration, and coconcentra-
noroviruses detected in different countriesa tion of inhibitory substances during sample concentration.
Viral gastroenteritis is initiated by the consumption of conta-
Rotavirus Norovirus Unidentified minated water. The minimum number of viruses that can
South Korea 86 0 14 cause an infection in humans is called an infective dose. It
China/Hong Kong 43 12 45
varies by virus type and is usually lower in water than food
because waterborne viruses are quickly carried through the
Malaysia 45 7 48 stomach acid and are better dispersed, leading to the initial
Singapore 11 0 89 infection. In general, exposure to low numbers (1–10 infec-
Sri Lanka 44 0 56 tious units) of waterborne viruses may lead to virus infection
Uganda 17 8 75 and replication within the digestive tract (67). The infection
Congo 60 0 40
by majority of pathogenic waterborne viruses remains in the
gastrointestinal (GI) system; however, some viruses can cause
Senegal 30 20 50 infections that spread beyond the GI system and involve other
Tunisia 21 11 68 organs such liver (hepatitis A and E) and poliovirus (nervous
Central Africa 29 18 53 system). Waterborne viral infections may have wide range on
French Guiana 2 13 85 incubation time, duration of illness, and clinical symptoms
Iran 35 6 59
(Table 3) ranging from asymptomatic carrier states to life-
threatening illnesses (69). The severity of infection can be
Brazil 16 33 51 related to the number of viruses ingested. Low doses may
a
Based on data from (47–49). cause only asymptomatic infection, whereas heavy doses
may cause the worst illness. In addition to viral load, general
health condition and age of host also impact the outcome of
genotype 1 is commonly associated with sporadic cases of gas- infection, and low-dose infection can be self-limiting in
troenteritis in developing world and genotypes 2 and 4 are healthy young subjects compared with older subjects.
commonly associated with gastroenteritis outbreaks in devel- In developed countries approximately 100% of the urban
oping countries. HEV genotype 2 is mostly limited to Mexico population has access to sewage collection, yet the treatment
and western African region, whereas genotype 4 is prevalent of 100% collected sewage is not guaranteed. In Canada
in Asia (56–58). approximately 25% of collected sewage is not treated before
Globally the most prevalent genotype of rotavirus is G1P being discharged directly into rivers or oceans (70). In the
[8] followed by G3P[8]. Geographically, the rotavirus geno- United States, a sizable fraction of sewage receives little or
types can be either diverse or dominated by a single genotype no treatment before being discharged into waterways (71).
in any setting. Four rotavirus genotypes (G3P[8], G2P[nt], Under such circumstances enteric viruses have been detected
G5P[nt], and G9P[nt]) were detected in 32 samples from in surface waters several kilometers downstream from a single
Hong Kong, whereas only one genotype (G3P[8]) was sewage outlet (72, 73). In aquatic environments, some of the
detected in 19 samples from Tunisia (47). Similarly, genotype major waterborne viruses (such as adenoviruses, hepatitis A,
variations have been reported among the caliciviruses preva- and norovirus) are known to survive for long periods (61, 74).
lent around the globe, and norovirus GII.4 is more prevalent In groundwater under ambient temperature conditions, noro-
in the Western Hemisphere (59, 60). The other most com- virus can remain infectious for more than 2 months; however,
mon circulating norovirus genotypes include GII.2 and under similar conditions can be detected even after 3 years
GI.3; GIV.1 genotype is prevalent mostly in Uganda (47). using PCR (ClinicalTrials.gov identifier NCT00313404.).
A seasonal variation in the prevalence and distribution of High concentration of viruses might be detected at the sour-
waterborne enteric viruses has also been reported (61). In ces of contamination; however, virus concentration in waters
temperate climates an increased circulation of enteroviruses decreases due to the dilution effect as they drift and some are
is reported in summer and early fall (62). Noroviruses are inactivated in water sources. Therefore, efficient sample con-
more frequently found in water samples collected during win- centration methods are needed before application of the
ter compared with other seasons of the year (63, 64). In con- detection methodologies.
trast, the adenovirus prevalence in water is not related to the Since concentration of viruses in source waters is generally
season of the year, and their year-round occurrence has been low, direct analysis of water samples for the detection of
well documented (64, 65). The information of seasonal and viruses is not a practical strategy. Therefore a concentration
geographical variation in the occurrence of waterborne virus step is always required prior to the detection of viruses using
plays a critical role in reducing the burden of viral gastroenter- culture-based or molecular methods. General approaches
itis. Such information is collected through monitoring and used for virus concentration rely on virion characteristics,
epidemiological surveillance programs, which rely on rapid such as (a) ion charge/isoelectric point (adsorption elution
and sensitive detection methods. The Centers for Disease method), (b) particle size (ultrafiltration method), and (c)
Control and Prevention, the Environmental Protection density and sedimentation coefficient (ultracentrifugation
Agency (EPA), and the Council of State and Territorial Epi- method).
demiologists have maintained a database of disease outbreaks Over the course of several decades, a variety of filter devi-
in the United States since 1971 (66). ces have been reported to concentrate waterborne enteric
viruses. Techniques include gauze pads (75), ultrafiltration
(76, 77), glass wool (78), and electronegative (79) and elec-
VIRUS DETECTION METHODOLOGIES tropositive pleated filters (80). In general, efficacy of these
Detection of viruses is a complex task that involves techni- devices has been proven for the capture of viruses in water;
ques relying on the genetic, immunological, and host specif- however, several are impractical for wide-scale use because
icity features/characteristics of any virus. Great difficulties are of incompatibility with highly turbid natural waters (e.g.,
involved with detecting viruses in water because of their ultrafilters) or prohibitive cost of filter cartridge (e.g.,
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-5

TABLE 3 Pathobiology and prevalence of major waterborne enteric viruses

Incubation Duration of
Virus Prevalence/distribution Seasona Transmission
time illness
Adenovirus 8–10 days 5–12 days Worldwide No definite seasonality Person-to-person (fecal-oral,
respiratory?)
Aichi virus 39 h Worldwide incidence No specific season Fecal/oral route
(27–60 h) (0.9– 3.1%) of AiV
Astrovirus 1.5–2 days 1–4 days Worldwide Mostly winter Person-to-person (fecal-oral)
( protracted
diarrhea with
serotype 3)
Coxsackie 2–10 days Developing and Fecal-oral, food borne,
virus developed countries waterborne
Echovirus 2–10 days Developing countries Fecal-oral, food borne,
waterborne
Hepatitis A 15–50 days 2–12 weeks Worldwide No definite seasonality Person-to-person fecal-oral,
(av. 28) food borne, waterborne
Hepatitis E 15–60 days 4–10 weeks Developing countries No definite seasonality Person-to-person fecal-oral,
(av. 40) food borne, waterborne
and blood transfusion
Norovirus 1–3 days 4 Days Worldwide (with Mostly winter ( previously Person-to-person (fecal-oral,
(10–50 h) (shedding up genotype variation) called “winter vomiting droplet, fomites) shellfish,
to 8 weeks) disease”), but may occur cold food, drinking water
in warmer months as
well
Poliovirus 6–20 days Conflict zones: No definite seasonality Fecal-oral, food borne,
Afghanistan, Pakistan, waterborne
Africa, and Egypt
Rotavirus 2–6 days 5–7 Days Worldwide Winter Person-to-person (fecal-oral,
respiratory) food borne,
water
Sapporo virus 1–4 days 6 Days Far East Mostly winter Person-to-person (fecal-oral
or aerosol), shellfish, food,
water
a
A typical enterovirus season in the United States lasts from June through October (68).

1MDS electropositive filter, CUNO, Meriden, CT, ∼$180
each) (Table 4). The two major types of filters commonly
used are electronegative and electropositive. Electronegative
filters are not ideal for large-scale water sampling because they
require acidification of the water and the addition of multiva-
lent cationic salts to the water prior to filtration (81). These
techniques provide a wide range of overall procedural recov-
ery of viruses in water (Table 5). The concentration method
using filtration and elution has proven to be time consuming
and costly, especially for routine monitoring of water. A novel
method using surface display of virus receptor was recently
reported that provides a practical and inexpensive method
of viral concentration in 1 liter of water (90). The proposed
strategy fills the gap between the available methods and the
actual needs of the water industry for the cost, ease, and speed
of sample processing. Surface display is new paradigm for
development of simple and sensitive sample processing meth-
odologies for use in monitoring of waterborne viruses.
In general, the methods used for the detection of water-
borne viruses can be categorized as (a) cell culture infectivity
assays, and (b) molecular detection methods such as PCR and
nucleic acid hybridization (91, 92). Each has advantages and
disadvantages (Table 6). Historically, research into the detec-
tion of enteric viruses in water has been primarily limited to
the enteroviruses because of the availability of techniques
approved by regulatory authorities (93). Waterborne viruses
are detected using the plaque assay procedure as stated in
the EPA 40 CFR Part 503 subpart D regulations (94).
The conventional (and most commonly used) method for
the detection of waterborne enteric viruses is based on cell
culture assays (95). However, application of cell culture–
based assays can greatly increase the cost and time of the sam-
ple analysis. These assays rely on few established cell lines,
which have special cell surface receptors for attachment and
infection of certain group of viruses (Table 7) (96). The
attachment of virus particle to its specific receptor on host
cell surface initiates a series of events resulting in the entry
of virus in cell followed by pathogenesis. Pathogenesis also
is a complex process that involves multiple steps such as
host cell response, virus replication, cytopathogenicity, and
spread of infection, which may be elicited as plaque or cyto-
pathogenic effect (14).
Plaque assays using the BGMK (Buffalo green monkey kid-
ney) cell line have been one of the traditional methods, and
the most common, for monitoring enteroviruses in environ-
mental samples (97, 98). The BGMK cell is preferred over
others, including primary cells, because it provides high sen-
sitivity to natural isolates of enteroviruses (99). Its sensitivity
can be further enhanced by pretreatment of the cells with
enzymes or other substances (100). However, not all enteric
3.1.5-6 ▪ WATER

TABLE 4 Comparison of different types of filters used to concentrate waterborne viruses

Filter group Filter type Comments
Ultrafiltration Hollow fiber Pros No sample preconditioning needed, economical, good recoveries, simultaneous concentration
of multiple pathogens
Cons Slow filtration, high turbidity may cause clogging, not easy to deploy in field conditions
Electronegative GN-6 Metricel Pros Can handle high turbidity and large volumes, good recoveries
Cons Sample preconditioning is needed
Electropositive 1MDS Pros No sample preconditioning needed, can handle high turbidity and large volumes, good
recoveries
Cons Expensive
Glass wool Pros No sample preconditioning needed, economical, easy to deploy in field conditions
Cons Filter assembly and preparation, may clog with high-turbidity samples
Nanoceram Pros No sample preconditioning needed, economical, good recoveries
Cons May clog with high turbidity samples
ViroCap Pros No sample preconditioning needed, easy to deploy in field conditions, economical, good
recoveries
Cons Not suitable for large volumes

viruses can be detected by this method; hence, it may not be
an adequate means for assessing viral contamination in water.
Consequently, other cell lines are required to detect certain
groups of enteric viruses (101). For example, simian rotavirus
SA-11 was the first virus to be grown to high titers in the
MA-104 (rhesus monkey kidney) cell line (102). Different
cell lines have been compared for their efficacy to detect
human enteric viruses (103, 104). The greatest efficiency of
recovery of enteroviruses has been reported in BGMK cells,
followed by MA-104 cells (104). The BGM cell line, recom-
mended for the total culturable viral assay, is the standard
method for the enumeration of viruses from water (105).
Since different groups of enteric viruses attach to different
receptors, a single cell line may not have optimal receptors
for the detection of broad range of waterborne viruses. A com-
bination of cell lines such as A549 (lung carcinoma), Caco-2
(human colon adenocarcinoma), PLC/PRF/5 ( primary liver
carcinoma), Hela (epitheloid carcinoma), and FRhK-4 (rhe-
sus monkey kidney cell) may increase the efficacy and sensi-
tivity of cell culture assay for the detection of adenoviruses,
coxsackie A viruses, echoviruses, hepatitis A virus, and astro-
viruses or other viruses that do not grow effectively on BGM
cells (106–109). Despite the availability of a broad range of
mammalian cell lines showing diverse receptors, some human
enteric virus are difficult to culture or are not culturable in cell
culture system, and human norovirus is a classical example.
Many efforts have been made to develop cell culture models
for human norovirus. In a comprehensive study, 27 cell lines
(A549, AGS, Caco-2, CCD-18, CRFK, CR-PEC, Detroit
551, Detroit 562, FRhK-4, HCT-8, HeLa, HEC, HEp-2,
Ht-29, HuTu-80, I-407, IEC-6, IEC-18, Kato-3, L20B,
MA104, MDBK, MDCK, RD, TMK, Vero, and 293) were

TABLE 5 Recovery efficiency of enteric viruses from water

% Recovery efficiency for virusesa
Filter
Poliovirus Coxsackievirus Echovirus
Electropositive
1MDS 89–96 (82–84) 50–95 (82, 83) 92 (83)
MK 73 (82, 84) 33 (82) —
AMF Cuno Zeta plus 50S 56 (85) — —
Electronegative
Glass filter 37–100 (84) — —
Filterite filter 60 (85) — —
Cellulose nitrate 4 (86) — —
Cellulose acetate 51 (86) — —
Glass borosilicate filter 24 (86) — —
Zetapor glass filter 73 (84) — —
Double filtration using 0.22 µm 76.5 (87) — —
filters followed by 10,000 molecular weight cut-off
Ultrafiltration — — 49–97 (88)
Using gollow fiber 82–90 (88, 89) — —
Using tangential flow 34–95 (88, 89) — —
a
Recovery efficiency of poliovirus, coxsackievirus, and echovirus using different filters. (—) indicates that data are not available. Different filters can be found
at http://advancedfiltration.com/.
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-7

TABLE 6 Comparison of common methods for the detection of enteric viruses from environmental sourcesa
Method Advantages Disadvantages
Cell culture Infectivity can be determined; provides Long processing time (takes days to weeks);
quantitative data relatively more expensive than conventional
PCR; not all viruses can grow on cultured cells
PCR (RT-PCR) Rapid; increased sensitivity and specificity Presence or absence only (nonquantitative);
compared to cell culture inhibitors present in environmental samples
may interfere with PCR amplification;
infectivity cannot be determined
Nested PCR Increased sensitivity compared to conventional Potential risk of carryover contamination when
(semi/hemi-nested) PCR; can replace PCR confirmation steps, such transferring PCR products
as hybridization
Multiplex PCR Several types, groups, or species of viruses can be Difficult to achieve equal sensitivity for all targeted
detected in a single reaction; saves time and cost virus species, groups, or types; may produce
nonspecific amplification in environmental
samples
Real-time PCR Provides quantitative data; confirmation of PCR Expensive equipment; occasionally less sensitive
products is not required (saves time); can be than conventional PCR and nested PCR
done in a closed system, which reduces risk of
contamination compared to nested PCR
ICC-PCR Improves detection of infectious viral pathogens Less time-efficient and more costly than direct
compared to conventional cell culture; detects PCR detection; carryover detection of DNA of
viruses that do not produce CPE in cell culture; inactivated viruses inoculated onto cultured
provides results in half the time required for cells is possible
conventional cell culture
a
Source: (29).

evaluated under various culture conditions to grow human
noroviruses in cell culture system; all such efforts have been
unsuccessful (110, 111). More recently, possibility of norovi-
rus genome replication in a 3D culture of Caco2 cell line have
been reported; however, the proposed cell culture system is
not reliably reproducible (112–116).
Cell culture assay can detect infectious viruses in environ-
mental samples, without additional tests; however, no deter-
mination can be made as to the particular strain of the virus
present in the sample. Additionally, the length of time
needed to detect infection in the cell culture can vary greatly
from a few days to several weeks, depending on the type and
number of viruses present. Molecular techniques such as
PCR and real-time PCR have been reported as rapid methods
for the detection of viruses in sample concentrates (91, 92,
117, 118). The PCR-based assays can be used to amplify
nucleic acids associated with viruses in samples. During
PCR, a genome fragment of the target virus is amplified using
specific primers. Selection of appropriate primers is crucial
factor in the representativeness of data. For example, many
of the commonly used enterovirus PCR primers do not
include parechovirus; thus, this virus’s presence has not
been reflected in previous quantitative PCR enterovirus mon-
itoring (119).
Analysis of the genome sequences of all the prototype
strains of major groups of viruses has shown that the species
remain phylogenetically coherent in all genome regions
except in the 50 -nontranslated region (NTR) (37, 120,
121). The 30 -NTR of enteroviruses is highly conserved within
a species but highly divergent between species. Similarly,

TABLE 7 Major groups of enteric viruses and their receptorsa

Virus Receptors
Poliovirus PVR—Ig super family
Hepatitis A virus HAVcr-1, Ig super family, mucin-like
Coxsackie virus Integrins αγβ, CD54, CD55, CAR, Ig super family
Echovirus Integrins α2β1, CD55
Adenovirus subgroup A, C, D, E, F CAR, α2β1, α4β1
Human borovirus Histo-blood group antigens (HBGA) on type 1, 2, and 3 GSLs, heparan sulfate, sialic acid on SLex
Coronavirus serogroup 1 hAPN
Coronavirus serogroup 2 Sialic acid
Rotavirus Integrins α2β1, α4β1, αvβ3, αxβ2, hsc70
BK virus α2,8-linked disialic acid on GD1b and GT1b
JC virus Terminal α2,3-linked sialic acid on GT1b
a
Based on reference (96).
3.1.5-8 ▪ WATER
members of different groups of enteric viruses have genomic
segments that show divergence in species or genotypes (14).
This feature of enteric viruses allows designing species-
specific primers, which can be used to rapidly detect a collec-
tion of enterovirus isolates or specific or genotype specific
primers for in-depth characterization of viral isolates. In
the scientific literature, a wealth of information is available
regarding the primers for different groups of viruses including
group-specific primers for enteroviruses (91, 122, 123), HEV
(124), HAV (125), astroviruses (106), rotaviruses (92),
adenoviruses (123), coxsackievirus (126), and noroviruses
(127–129). Recent advances in molecular techniques have
enhanced the scope of virus detection techniques and PCR
mass spectrometry (16), microarray (130), biosensor (131),
and luminix-based assays (132) have been reported for the
detection of enteric viruses. In the context of diversity of
viruses in environmental sample multiplex-PCR assay have
been developed for the detection of broad range of viruses
that may be present in a sample (123, 133, 134). However,
this technique has its limitations, and recently metagenomic
analyses have been developed that allow the detection of
a much broader range of viruses. Metagenomics studies of
animal feces have similarly shown frequent coinfections
and yielded highly divergent viral species (7, 9, 2, 28, 30,
135–144).
Application of metagenomic analyses has allowed detec-
tion of coinfections of viral species of different families
including Picornaviridae (25), Astroviridae (18, 145), Parvovir-
idae (27, 145), and Circoviridae (28).
For RNA viruses, reverse transcription (RT) of the viral
RNA to a complimentary DNA strand (cDNA) is necessary
prior to PCR. Molecular methods offer higher sensitivity
than cell culture for the detection of target nucleic acids usu-
ally found in environmental samples (92). Although the
RT-PCR technique has been extensively employed for virus
detection from various environmental samples, the applica-
tion of it for the detection of viruses in environmental sam-
ples is an arduous task because of inhibitory substances to
gene amplification. Various strategies have been reported
to address these issues (146). A combination of the enzymatic
virus elution method with RT-PCR has been reported to
effectively detect viruses in environmental samples (147).
The method relies on specific enzyme to break down
the organic matter flocs to which viruses are adsorbed. Such
treatments help in elution and recovery of viruses from
environmental samples containing organic and inorganic
compounds (humic acids, polyphenols, and heavy metals)
(147). Such steps are helpful when large amounts of toxic
compound/elements are coconcentrated during the virus con-
centration step and are toxic and cause lysis in cell cultures.
These compounds are also liable to form complexes with
nucleic acids and inhibit nucleic acid amplification enzymes
(148). The results of cell culture analysis and PCR therefore
depend on the efficacy with which the viral extraction tech-
nique used removes such compounds. When carefully opti-
mized, these molecular assays are at least as sensitive as cell
culture systems for the detection of enteroviruses in environ-
mental samples (91, 122, 149–151).
The PCR-based techniques provide rapid and highly sen-
sitive methods for the detection of viruses; however, this fea-
ture may also increase the chances of false positive results due
to extremely low levels of contamination. To counter the vul-
nerability of PCR methods to contamination events, a strin-
gent quality control and quality assurance (QA/QC) program
needs to be instituted in any laboratory performing PCR
assays. Prevention of cross-contamination and QA/QC can
be assured by using aerosol-resistant pipette tips or positive-
displacement pipettors, decontamination of instruments
between experiments, and physical separation of pre- and
post-PCR products. To ensure the quality of PCR data, the
EPA has published guidelines for the laboratories performing
these analyses on environmental samples (152). The docu-
ment provides guidance on personnel, facilities, equipments,
reagents, work flow, methods, QA/QC, and data recoding.
The data analyses and reporting are critical steps in effectively
communicating the research findings. Over the past decade,
PCR and real-time PCR techniques have been widely
adopted by microbiology laboratories around the world. A
consensus has emerged to adopt standardized procedures for
analyzing, interpreting, and presenting data generated using
real-time PCR and a guideline on “Minimum Information
for Publication of Quantitative PCR Experiments” has been
published (153). Adoption of these guideline ensure the
quality of data presented by various research groups and
allow reasonable bases for comparison of data from different
studies performed in different laboratories. In the context
of the diversity of waterborne virus prevalent across the
globe, the adoption of standardized methods enhances the
values of such efforts toward the global management of viral
gastroenteritis.
REFERENCES
1. Wolf AT. 2007. Shared waters: conflict and cooperation.
Ann Rev Environ Resour 32:241–269.
2. Wittfogel KA. 1956. The Hydraulic Civilizations. University
of Chicago Press, Chicago, IL.
3. Delli Priscoli J. 2000. Water and civilization: using history
to reframe water policy debates and to build a new ecolog-
ical realism. Water Pol 1:623–636.
4. Robbins F, Daniel T. 1997. A history of poliomyelitis, p
5–22. In Daniel TM and Robbins EC (eds.), Polio, Univer-
sity of Rochester Press.
5. Trevelyan B, Smallman-Raynor M, Cliff AD. 2005. The
spatial dynamics of poliomyelitis in the United States: from
epidemic emergence to vaccine-induced retreat, 1910–
1971. Ann Assoc Am Geographers 95:269–293.
6. Viswanathan R. 1957. Epidemiology: infectious hepatitis
in New Delhi 1955–56. Indian J Med Res 45:1–29.
7. Enquist L. 2009. Virology in the 21st century. J Virol 83:
5296–5308.
8. Abbaszadegan M, Alum A. 2004. Microbiological contami-
nants and threats of concern, pp 2.1–2.12. In Water Supply
Systems Security, McGraw-Hill, New York, NY.
9. Shan T, Wang C, Cui L, Yu Y, Delwart E, Zhao W, Zhu
C, Lan D, Dai X, Hua X. 2010. Picornavirus salivirus/
klassevirus in children with diarrhea, China. Emerg Infect
Dis 16:1303–1305, doi:10.3201/eid1608.100087.
10. Hall AJ, Wikswo ME, Manikonda K, Roberts VA, Yoder
JS, Gould LH. 2013. Acute gastroenteritis surveillance
through the National Outbreak Reporting System, United
States. Emerg Infect Dis 19:1305–1309, doi:10.3201/
eid1908.130482.
11. Griffin DW, Donaldson KA, Paul JH, Rose JB. 2003.
Pathogenic human viruses in coastal waters. Clin Microbiol
Rev 16:129–143.
12. Kocwa-Haluch R. 2001. Waterborne enteroviruses as a
hazard for human health. Pol J Environ Stud 10:485–488.
13. Melnick JL. 1984. Etiologic agents and their potential for
causing waterborne virus diseases, pp 1–16. In Melnick JL
(ed.), Enteric Viruses in Water, vol. 15. Karger, Basel,
Switzerland.
14. Knipe D, Howley PM (eds.). 2013. Fields Virology, 6th ed.,
Lippincott Williams & Wilkins, Philadelphia, PA.
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-9
15. Hulo C, De Castro E, Masson P, Bougueleret L, Bairoch
A, Xenarios I, Le Mercier P. 2011. ViralZone: a knowl-
edge resource to understand virus diversity. Nucleic Acids
Res 39:D576–D582.
16. Piao J, Jiang J, Xu B, Wang X, Guan Y, Wu W, Liu L,
Zhang Y, Huang X, Wang P. 2012. Simultaneous detec-
tion and identification of enteric viruses by PCR-mass
assay. PLoS One 7:e42251.
17. Hamza IA, Jurzik L, Überla K, Wilhelm M. 2011. Meth-
ods to detect infectious human enteric viruses in environ-
mental water samples. Int J Hyg Environ Health 214:
424–436.
18. Finkbeiner SR, Kirkwood CD, Wang D. 2008. Complete
genome sequence of a highly divergent astrovirus isolated
from a child with acute diarrhea. Virol J 5:117.
19. Victoria JG, Kapoor A, Li L, Blinkova O, Slikas B, Wang
C, Naeem A, Zaidi S, Delwart E. 2009. Metagenomic
analyses of viruses in stool samples from children with acute
flaccid paralysis. J Virol 83:4642–4651.
20. Chiu CY, Greninger AL, Kanada K, Kwok T, Fischer KF,
Runckel C, Louie JK, Glaser CA, Yagi S, Schnurr DP.
2008. Identification of cardioviruses related to Theiler’s
murine encephalomyelitis virus in human infections. Proc
Natl Acad Sci 105:14124–14129.
21. Greninger AL, Runckel C, Chiu CY, Haggerty T, Par-
sonnet J, Ganem D, DeRisi JL. 2009. The complete
genome of klassevirus—a novel picornavirus in pediatric
stool. Virol J 6:82.
22. Holtz LR, Finkbeiner SR, Zhao G, Kirkwood CD,
Girones R, Pipas JM, Wang D. 2009. Klassevirus 1, a pre-
viously undescribed member of the family Picornaviridae, is
globally widespread. Virol J 6:422–426.
23. Jones MS, Lukashov VV, Ganac RD, Schnurr DP. 2007.
Discovery of a novel human picornavirus in a stool sample
from a pediatric patient presenting with fever of unknown
origin. J Clin Microbiol 45:2144–2150.
24. Kapoor A, Victoria J, Simmonds P, Slikas E, Chieochan-
sin T, Naeem A, Shaukat S, Sharif S, Alam MM, Angez
M. 2008. A highly prevalent and genetically diversified
Picornaviridae genus in South Asian children. Proc Natl
Acad Sci 105:20482–20487.
25. Li L, Victoria J, Kapoor A, Blinkova O, Wang C, Babrza-
deh F, Mason CJ, Pandey P, Triki H, Bahri O. 2009. A
novel picornavirus associated with gastroenteritis. J Virol
83:12002–12006.
26. Kapoor A, Li L, Victoria J, Oderinde B, Mason C, Pandey
P, Zaidi S, Delwart E. 2009. Multiple novel astrovirus spe-
cies in human stool. J Gen Virol 90:2965–2972.
27. Kapoor A, Mehta N, Esper F, Poljsak-Prijatelj M, Quan
P-L, Qaisar N, Delwart E, Lipkin WI. 2010. Identifica-
tion and characterization of a new bocavirus species in
gorillas. PLoS One 5:e11948.
28. Phan TG, Kapusinszky B, Wang C, Rose RK, Lipton HL,
Delwart EL. 2011. The fecal viral flora of wild rodents.
PLoS Pathog 7:e1002218.
29. Fong T-T, Lipp EK. 2005. Enteric viruses of humans and
animals in aquatic environments: health risks, detection,
and potential water quality assessment tools. Microbiol
Mol Biol Rev 69:357–371.
30. Li L, Kapoor A, Slikas B, Bamidele OS, Wang C, Shaukat
S, Masroor MA, Wilson ML, Ndjango J-BN, Peeters M.
2010. Multiple diverse circoviruses infect farm animals and
are commonly found in human and chimpanzee feces. J
Virol 84:1674–1682.
31. Borucki MK, Allen JE, Chen-Harris H, Zemla A, Vanier
G, Mabery S, Torres C, Hullinger P, Slezak T. 2013. The
role of viral population diversity in adaptation of bovine
coronavirus to new host environments. PLoS One 8:
e52752.
32. Wentworth DE, Holmes KV. 2001. Molecular determi-
nants of species specificity in the coronavirus receptor
aminopeptidase N (CD13): influence of N-linked glycosy-
lation. J Virol 75:9741–9752.
33. Schwegmann-Wessels C, Herrler G. 2008. Identification of
sugar residues involved in the binding of TGEV to porcine brush
border membranes, pp 319–329. SARS and Other Corona-
viruses, Springer, New York, NY.
34. Mathapati BS, Mishra N, Rajukumar K, Nema RK,
Behera SP, Dubey SC. 2010. Entry of bovine viral diarrhea
virus into ovine cells occurs through clathrin-dependent
endocytosis and low pH-dependent fusion. In Vitro Cell
Dev Biol-Anim 46:403–407.
35. Shakeel S, Seitsonen JJ, Kajander T, Laurinmäki P, Hyy-
piä T, Susi P, Butcher SJ. 2013. Structural and Functional
analysis of coxsackievirus A9 integrin αvβ6 binding and
uncoating. J Virol 87:3943–3951.
36. Elena SF, Carrasco P, Daròs JA, Sanjuán R. 2006. Mech-
anisms of genetic robustness in RNA viruses. EMBO Rep 7:
168–173.
37. Oberste MS, Maher K, Pallansch MA. 2004. Evidence for
frequent recombination within species human enterovirus
B based on complete genomic sequences of all thirty-seven
serotypes. J Virol 78:855–867.
38. Estes MK. 2001. Rotaviruses and their replication,
pp 1747–1786. In Knipe D, Howley PM (eds.), Fields Virol-
ogy, 4th ed, vol. 2. Lippincott Williams & Wilkins, Philadel-
phia, PA.
39. Xerry J, Gallimore CI, Iturriza-Gómara M, Allen DJ,
Gray JJ. 2008. Transmission events within outbreaks of
gastroenteritis determined through analysis of nucleotide
sequences of the P2 domain of genogroup II noroviruses.
J Clin Microbiol 46:947–953.
40. Xerry J, Gallimore CI, Iturriza-Gómara M, Gray JJ. 2009.
Tracking the transmission routes of genogroup II norovi-
ruses in suspected food-borne or environmental outbreaks
of gastroenteritis through sequence analysis of the P2
domain. J Med Virol 81:1298–1304.
41. Sukhrie FH, Beersma MF, Wong A, van der Veer B, Ven-
nema H, Bogerman J, Koopmans M. 2011. Using molec-
ular epidemiology to trace transmission of nosocomial
norovirus infection. J Clin Microbiol 49:602–606.
42. Vega E, Barclay L, Gregoricus N, Williams K, Lee D,
Vinjé J. 2011. Novel surveillance network for norovirus
gastroenteritis outbreaks, United States. Emerg Infect Dis
17:1389–1395, doi:10.3201/eid1708.101837.
43. Atmar RL, Opekun AR, Gilger MA, Estes MK, Crawford
SE, Neill FH, Graham DY. 2008. Norwalk virus shedding
after experimental human infection. Emerg Infect Dis 14:
1553–1557.
44. Sattar SA, Jason T, Bidawid S, Farber J. 2000. Foodborne
spread of hepatitis A: recent studies on virus survival, trans-
fer and inactivation. Can J Infect Dis 11:159.
45. Reynolds KA, Mena KD, Gerba CP. 2008. Risk of water-
borne illness via drinking water in the United States, Revi
Environ Contam Toxicol 192:117–158.
46. Choffnes ER, Mack A. 2009. Global issues in water, sanita-
tion, and health: workshop summary. National Academies
Press, Washington D.C.
47. Rackoff LA, Bok K, Green KY, Kapikian AZ. 2013. Epi-
demiology and evolution of rotaviruses and noroviruses
from an archival WHO Global Study in Children (1976–
79) with implications for vaccine design. PloS One 8:
e59394.
48. Jalilian S, Samarbaf-Zadeh AR, Mozhgani R, Hamid S,
Makvandi M, Parsa-nahad M, Pirmoradi R, Shamsi-
Zadeh A. 2012. Relative frequency of norovirus infection
in children suffering from gastroenteritis and Referred to
Aboozar Hospital, Ahvaz, Iran. Jund J Microbiol 5:355–358.
3.1.5-10 ▪ WATER
49. Ferreira MSR, Xavier MdPTP, Tinga ACDC, Rose TL,
Fumian TM, Fialho AM, De Assis RM, Costa FAC, De
Oliveira SA, Leite JPG. 2012. Assessment of gastroenteric
viruses frequency in a children’s day care center in Rio De
Janeiro, Brazil: a fifteen year study (1994–2008). PLoS One
7:e33754.
50. Pham NTK, Khamrin P, Nguyen TA, Kanti DS, Phan
TG, Okitsu S, Ushijima H. 2007. Isolation and molecular
characterization of Aichi viruses from fecal specimens col-
lected in Japan, Bangladesh, Thailand, and Vietnam. J Clin
Microbiol 45:2287–2288.
51. Lanata CF, Fischer-Walker CL, Olascoaga AC, Torres
CX, Aryee MJ, Black RE. 2013. Global causes of diarrheal
disease mortality in children <5 years of age: a systematic
review. PLoS One 8:e72788.
52. Yamashita T, Sakae K, Tsuzuki H, Suzuki Y, Ishikawa N,
Takeda N, Miyamura T, Yamazaki S. 1998. Complete
nucleotide sequence and genetic organization of Aichi
virus, a distinct member of the Picornaviridae associated
with acute gastroenteritis in humans. J Virol 72:
8408–8412.
53. Yamashita T, Kobayashi S, Sakac K, Nakata S, Chiba S,
Ishihara Y, Isomura S. 1991. Isolation of cytopathic small
round viruses with BS-Cl cells from patients with gastroen-
teritis. J Infect Dis 164:954–957.
54. Oh D-Y, Silva P, Hauroeder B, Diedrich S, Cardoso D,
Schreier E. 2006. Molecular characterization of the first
Aichi viruses isolated in Europe and in South America.
Arch Virol 151:1199–1206.
55. Sdiri-Loulizi K, Hassine M, Gharbi-Khelifi H, Sakly N,
Chouchane S, Guediche MN, Pothier P, Aouni M,
Ambert-Balay K. 2009. Detection and genomic character-
ization of Aichi viruses in stool samples from children in
Monastir, Tunisia. J Clin Microbiol 47:2275–2278.
56. Velazquez O, Stetler HC, Avila C, Ornelas G, Alvarez C,
Hadler SC, Bradley DW, Sepúlveda J. 1990. Epidemic
transmission of enterically transmitted non-A, non-B hep-
atitis in Mexico, 1986–1987. J Am Med Assoc 263:
3281–3285.
57. Maila HT, Bowyer SM, Swanepoel R. 2004. Identifica-
tion of a new strain of hepatitis E virus from an outbreak
in Namibia in 1995. J Gen Virol 85:89–95.
58. Wang Y, Ling R, Erker JC, Zhang H, Li H, Desai S,
Mushahwar IK, Harrison TJ. 1999. A divergent genotype
of hepatitis E virus in Chinese patients with acute hepatitis.
J Gen Virol 80:169–177.
59. Yu J, Kim M, Kim D, Kim S, Lee J, Park S, Song C, Shin
H, Seo K, Choi I. 2008. Prevalence of hepatitis E virus and
sapovirus in post-weaning pigs and identification of their
genetic diversity. Arch Virol 153:739–742.
60. Svraka S, Vennema H, van der Veer B, Hedlund K-O,
Thorhagen M, Siebenga J, Duizer E, Koopmans M.
2010. Epidemiology and genotype analysis of emerging
sapovirus-associated infections across Europe. J Clin Micro-
biol 48:2191–2198.
61. Abbaszadegan M, Lechevallier M, Gerba C. 2003. Occur-
rence of viruses in US groundwaters. J Am Water Works
Assoc 95:107–120.
62. Franco MA, Greenberg HB. 2012. Epidemiologic and
clinical features of rotaviruses, noroviruses, and other gastroin-
testinal viruses. pp. 2244–2247. In Goldman L, Schafer AI,
(eds.), Goldman-Cecil Medicine 25th Edition, Elsevier, Phil-
adelphia, PA.
63. Mounts AW, Ando T, Koopmans M, Bresee JS, Noel J,
Glass RI. 2000. Cold weather seasonality of gastroenteritis
associated with Norwalk-like viruses. J Infect Dis 181:
S284–S287.
64. Haramoto E, Katayama H, Oguma K, Ohgaki S. 2005.
Application of cation-coated filter method to detection
of noroviruses, enteroviruses, adenoviruses, and torque
teno viruses in the Tamagawa River in Japan. Appl Environ
Microbiol 71:2403–2411.
65. Pina S, Puig M, Lucena F, Jofre J, Girones R. 1998. Viral
pollution in the environment and in shellfish: human
adenovirus detection by PCR as an index of human viruses.
Appl Environ Microbiol 64:3376–3382.
66. Craun GF, Brunkard JM, Yoder JS, Roberts VA, Carpen-
ter J, Wade T, Calderon RL, Roberts JM, Beach MJ, Roy
SL. 2010. Causes of outbreaks associated with drinking
water in the United States from 1971 to 2006. Clin Micro-
biol Rev 23:507–528.
67. Bosch A. 2010. Human enteric viruses in the water envi-
ronment: a minireview. Int Microbiol 1:191–196.
68. Moore M. 1982. Enteroviral disease in the United States,
1970–1979. J Infect Dis 146:103–108.
69. Moe C. 2002. Prospective studies of endemic water-
borne disease in developing countries, pp 197–206. In
Hunter PR, Waite M, Rochi E (eds.), Drinking Water and
Infectious Disease: Establishing the Links. CRC Press, Boca
Raton, FL.
70. Boyd DR. 2001. Canada vs. the OECD: an environmental
comparison University of Victoria: Eco-Research Chair of
Environmental Law and Policy, University of Victoria, B.C.
http://bibvir2.uqac.ca/archivage/12536745.pdf.
71. U.S. EPA. 2003. Strategically Addressing Raw Sewage
Discharges Across Nation to Protect Public. EPA 300-N-
03-001. Office of Enforcement and Compliance Assurance.
72. Ferguson CM, Coote BG, Ashbolt NJ, Stevenson IM.
1996. Relationships between indicators, pathogens and
water quality in an estuarine system. Water Res 30:
2045–2054.
73. Jiménez-Clavero MA, Fernández C, Ortiz JA, Pro J,
Carbonell G, Tarazona JV, Roblas N, Ley V. 2003.
Teschoviruses as indicators of porcine fecal contami-
nation of surface water. Appl Environ Microbiol 69:
6311–6315.
74. Borchardt MA, Bertz PD, Spencer SK, Battigelli DA.
2003. Incidence of enteric viruses in groundwater from
household wells in Wisconsin. Appl Environ Microbiol 69:
1172–1180.
75. Berg G. 1967. Transmission of Viruses by the Water Route,
Interscience Publishers, New York, 484p.
76. Hill VR, Polaczyk AL, Hahn D, Narayanan J, Cromeans
TL, Roberts JM, Amburgey JE. 2005. Development of a
rapid method for simultaneous recovery of diverse microbes
in drinking water by ultrafiltration with sodium polyphos-
phate and surfactants. Appl Environ Microbiol 71:
6878–6884.
77. Polaczyk AL, Narayanan J, Cromeans TL, Hahn D, Rob-
erts JM, Amburgey JE, Hill VR. 2008. Ultrafiltration-
based techniques for rapid and simultaneous concentration
of multiple microbe classes from 100-L tap water samples.
J Microbiol Methods 73:92–99.
78. Lambertini E, Spencer SK, Bertz PD, Loge FJ, Kieke BA,
Borchardt MA. 2008. Concentration of enteroviruses,
adenoviruses, and noroviruses from drinking water by
use of glass wool filters. Appl Environ Microbiol 74:
2990–2996.
79. Gerba CP, Farrah SR, Goyal SM, Wallis C, Melnick JL.
1978. Concentration of enteroviruses from large volumes
of tap water, treated sewage, and seawater. Appl Environ
Microbiol 35:540–548.
80. Sobsey MD, Moore RS, Glass JS. 1981. Evaluating
adsorbent filter performance for enteric virus concentra-
tions in tap water. J Am Water Works Assoc 73:542–548.
81. Wallis C, Henderson M, Melnick JL. 1972. Enterovirus
concentration on cellulose membranes. Appl Microbiol
23:476–480.
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-11
82. Ma J-F, Naranjo J, Gerba CP. 1994. Evaluation of MK fil-
ters for recovery of enteroviruses from tap water. Appl Envi-
ron Microbiol 60:1974–1977.
83. Wen Li J, Wei Wang X, Yi Rui Q, Song N, Guang Zhang
F, Chuan Ou Y, Huan Chao F. 1998. A new and simple
method for concentration of enteric viruses from water.
J Virol Meth 74:99–108.
84. Haramoto E, Katayama H, Ohgaki S. 2004. Detection of
noroviruses in tap water in Japan by means of a new method
for concentrating enteric viruses in large volumes of fresh-
water. Appl Environ Microbiol 70:2154–2160.
85. Sobsey MD, Glass JS. 1980. Poliovirus concentration from
tap water with electropositive adsorbent filters. Appl Envi-
ron Microbiol 40:201–210.
86. Fuhrman JA, Liang X, Noble RT. 2005. Rapid detection
of enteroviruses in small volumes of natural waters by real-
time quantitative reverse transcriptase PCR. Appl Environ
Microbiol 71:4523–4530.
87. Wommack K, Hill R, Colwell R. 1995. A simple method
for the concentration of viruses from natural water samples.
J Microbiol Meth 22:57–67.
88. Olszewski J, Winona L, Oshima KH. 2005. Comparison
of 2 ultrafiltration systems for the concentration of seeded
viruses from environmental waters. Can JMicrobiol 51:
295–303.
89. Winona L, Ommani A, Olszewski J, Nuzzo J, Oshima K.
2001. Efficient and predictable recovery of viruses from
water by small scale ultrafiltration systems. Can J Microbiol
47:1033–1041.
90. Abbaszadegan M, Alum A, Abbaszadegan H, Stout V.
2011. Cell surface display of poliovirus receptor on Escher-
ichia coli, a novel method for concentrating viral particles
in water. Appl Environ Microbiol 77:5141–5148.
91. Abbaszadegan M, Huber M, Gerba CP, Pepper I. 1993.
Detection of enteroviruses in groundwater with the polymer-
ase chain reaction. Appl Environ Microbiol 59:1318–1324.
92. Abbaszadegan M, Stewart P, LeChevallier M. 1999. A
strategy for detection of viruses in groundwater by PCR.
Appl Environ Microbiol 65:444–449.
93. Green DH, Lewis GD. 1999. Comparative detection of
enteric viruses in wastewaters, sediments and oysters by
reverse transcription-PCR and cell culture. Water Res 33:
1195–1200.
94. U.S. EPA. 1992. Technical support document for Part 503
pathogen and vector attraction reduction requirements in
sewage sludge. NTIS No. PB89-136618. National Techni-
cal Information Service, Springfield, VA.
95. Abbaszadegan M. 2001. Advanced detection of viruses
and protozoan parasites in water. Rev Biol Biotech 1:
21–26.
96. Schneider-Schaulies J. 2000. Cellular receptors for viruses:
links to tropism and pathogenesis. J Gen Virol 81:
1413–1429.
97. Berg G, Saffermann R, Dahling D, Berman D, Hurst C.
1984. USEPA Manual of Methods for Virology. U.S. Envi-
ronmental Protection Agency. Cincinnati, OH.
98. Dahling DR, Wright BA. 1984. Processing and transport
of environmental virus samples. Appl Environ Microbiol
47:1272–1276.
99. Dahling DR, Wright BA. 1986. Recovery of viruses from
water by a modified flocculation procedure for second-step
concentration. Appl Environ Microbiol 51:1326–1331.
100. Benton WH, Hurst CJ. 1986. Evaluation of mixed cell
types and 5-iodo-20 -deoxyuridine treatment upon plaque
assay titers of human enteric viruses. Appl Environmen
Microbiol 51:1036–1040.
101. Smith EM, Gerba CP. 1982. Development of a method for
detection of human rotavirus in water and sewage. Appl
Environ Microbiol 43:1440–1450.
102. Whitaker A, Hayward C. 1984. The characterization of
three monkey kidney cell lines. Dev Biol Stand 60:125–131.
103. Chonmaitree T, Ford C, Sanders C, Lucia H. 1988. Com-
parison of cell cultures for rapid isolation of enteroviruses.
J Clin Microbiol 26:2576–2580.
104. Brashear DA, Ward RL. 1982. Comparison of methods for
recovering indigenous viruses from raw wastewater sludge.
Appl Environ Microbiol 43:1413–1418.
105. U.S. EPA. 1995. Part 503 implementation guidance. U.S.
EPA, Washington, DC.
106. Chapron CD, Ballester NA, Fontaine JH, Frades CN,
Margolin AB. 2000. Detection of astroviruses, enterovi-
ruses, and adenovirus types 40 and 41 in surface waters col-
lected and evaluated by the information collection rule and
an integrated cell culture-nested PCR procedure. Appl
Environ Microbiol 66:2520–2525.
107. Sedmak G, Bina D, MacDonald J. 2003. Assessment of an
enterovirus sewage surveillance system by comparison of
clinical isolates with sewage isolates from Milwaukee, Wis-
consin, collected August 1994 to December 2002. Appl
Environ Microbiol 69:7181–7187.
108. Sedmak G, Bina D, MacDonald J, Couillard L. 2005.
Nine-year study of the occurrence of culturable viruses in
source water for two drinking water treatment plants and
the influent and effluent of a wastewater treatment plant
in Milwaukee, Wisconsin (August 1994 through July
2003). Appl Environ Microbiol 71:1042–1050.
109. Vivier J, Ehlers M, Grabow W. 2004. Detection of enter-
oviruses in treated drinking water. Water Res 38:
2699–2705.
110. Duizer E, Schwab KJ, Neill FH, Atmar RL, Koopmans
MP, Estes MK. 2004. Laboratory efforts to cultivate noro-
viruses. J Gen Virol 85:79–87.
111. Wobus CE, Karst SM, Thackray LB, Chang K-O, Sos-
novtsev SV, Belliot G, Krug A, Mackenzie JM, Green
KY, Virgin IV HW. 2004. Replication of norovirus in
cell culture reveals a tropism for dendritic cells and macro-
phages. PLoS Biol 2:e432.
112. Papafragkou E, Hewitt J, Park GW, Greening G, Vinjé J.
2013. Challenges of culturing human norovirus in three-
dimensional organoid intestinal cell culture models. PLoS
One 8:e63485.
113. Alum A, Shieh YC Sun Y, Stewart DS, Reineke KF,
Tortorello ML. 2012. Method development for rapid concen-
tration of noroviruses in water and tomato eluate. Institute for
Food Safety and Health, Annual Review, Bedford Park, IL.
114. Straub TM, Bartholomew RA, Valdez CO, Valentine
NB, Dohnalkova A, Ozanich RM, Bruckner-Lea CJ,
Call DR. 2011. Human norovirus infection of Caco-2 cells
grown as a 3-dimensional tissue structure. J Water and
Health 9:225.
115. Straub TM, zu Bentrup KH, Orosz Coghlan P, Dohnal-
kova A, Mayer BK, Bartholomew RA, Valdez CO,
Bruckner-Lea CJ, Gerba CP, Abbaszadegan M, Nicker-
son CA. 2007. In vitro cell culture infectivity assay for
human noroviruses. Emerg Infect Dis 13:396–403.
116. Straub TM, Hutchison JR, Bartholomew RA, Valdez
CO, Valentine NB, Dohnalkova A, Ozanich RM,
Bruckner-Lea CJ. 2013. Defining cell culture conditions
to improve human norovirus infectivity assays. Water Sci
Technol 67:863–868.
117. Choi S, Jiang SC. 2005. Real-time PCR quantification of
human adenoviruses in urban rivers indicates genome prev-
alence but low infectivity. Appl Environ Microbiol 71:
7426–7433.
118. Parker J, Chang TY, Meschke J. 2011. Amplification of
viral RNA from drinking water using TransPlex™ whole-
transcriptome amplification. J Appl Microbiol 111:
216–223.
3.1.5-12 ▪ WATER
119. Wong K, Onan BM, Xagoraraki I. 2010. Quantification
of enteric viruses, pathogen indicators, and salmonella
bacteria in class B anaerobically digested biosolids by cul-
ture and molecular methods. Appl Environ Microbiol 76:
6441–6448.
120. Brown B, Oberste MS, Maher K, Pallansch MA. 2003.
Complete genomic sequencing shows that polioviruses
and members of human enterovirus species C are closely
related in the noncapsid coding region. J Virol 77:
8973–8984.
121. Oberste MS, Peñaranda S, Maher K, Pallansch MA. 2004.
Complete genome sequences of all members of the species
Human enterovirus A. J Gen Virol 85:1597–1607.
122. DeLeon R, Shieh C, Baric RS, Sobsey MD. 1990.
Detection of enteroviruses and hepatitis A virus in environ-
mental samples by gene probes and polymerase chain
reaction, pp 833–853. In Advances in Water Analysis and
Treatment. Proceedings, 1989 Water Quality Technology
Conference. American Water Works Association, Denver,
CO.
123. Xu W, McDonough MC, Erdman DD. 2000. Species-
specific identification of human adenoviruses by a multi-
plex PCR assay. J Clin Microbiol 38:4114–4120.
124. Dierssen U, Rehren F, Henke-Gendo C, Harste G, Heim
A. 2008. Rapid routine detection of enterovirus RNA in
cerebrospinal fluid by a one-step real-time RT-PCR assay.
J Clin Virol 42:58–64.
125. Schwab KJ, De Leon R, Baric RS, Sobsey MD. 1991
Detection of rotaviruses, enteroviruses, and hepatitis A
virus by reverse transcriptase-polymerase chain reaction,
p 475–491. In Advances in water analysis and treatment.
Proceedings of the AWWA Water Quality Technology
Conference, Orlando, FL, 10 to 14 November 1991 Amer-
ican Water Works Association, Denver, CO.
126. Ouyang X, Zhang H, Bayston TA, Archard LC. 1995.
Detection of coxsackievirus B3 RNA in mouse myocarditis
by nested polymerase chain reaction. Clin Diagn Virol 3:
233–245.
127. Kageyama T, Kojima S, Shinohara M, Uchida K, Fukushi
S, Hoshino FB, Takeda N, Katayama K. 2003. Broadly
reactive and highly sensitive assay for Norwalk-like viruses
based on real-time quantitative reverse transcription-PCR.
J Clin Microbiol 41:1548–1557.
128. Laverick M, Wyn-Jones A, Carter M. 2004. Quantitative
RT-PCR for the enumeration of noroviruses (Norwalk-like
viruses) in water and sewage. Lett Appl Microbiol 39:
127–136.
129. Jothikumar N, Lowther JA, Henshilwood K, Lees DN,
Hill VR, Vinjé J. 2005. Rapid and sensitive detection of
noroviruses by using TaqMan-based one-step reverse
transcription-PCR assays and application to naturally con-
taminated shellfish samples. Appl Environ Microbiol 71:
1870–1875.
130. Jääskeläinen AJ, Maunula L. 2006. Applicability of micro-
array technique for the detection of noro-and astroviruses.
J Virol Meth 136:210–216.
131. Connelly JT, Kondapalli S, Skoupi M, Parker JS, Kirby
BJ, Baeumner AJ. 2012. Micro-total analysis system for
virus detection: microfluidic pre-concentration coupled
to liposome-based detection. Anal Bioanal Chem 402:
315–323.
132. Liu Y, Xu Z-q, Zhang Q, Jin M, Yu J-m, Li J-s, Liu N, Cui
S-x, Kong X-y, Wang H. 2012. Simultaneous detection of
seven enteric viruses associated with acute gastroenteritis
by a multiplexed Luminex-based assay. J Clin Microbiol
50:2384–2389.
133. Rohayem J, Berger S, Juretzek T, Herchenröder O,
Mogel M, Poppe M, Henker J, Rethwilm A. 2004. A sim-
ple and rapid single-step multiplex RT-PCR to detect
norovirus, astrovirus and adenovirus in clinical stool sam-
ples. J Virol Meth 118:49–59.
134. Formiga-Cruz M, Hundesa A, Clemente-Casares P,
Albiñana-Gimenez N, Allard A, Girones R. 2005. Nested
multiplex PCR assay for detection of human enteric viruses
in shellfish and sewage. J Virolo Meth 125:111–118.
135. Blinkova O, Victoria J, Li Y, Keele BF, Sanz C, Ndjango
J-BN, Peeters M, Travis D, Lonsdorf EV, Wilson ML.
2010. Novel circular DNA viruses in stool samples of wild-
living chimpanzees. J Gen Virol 91:74–86.
136. Canuti M, Eis-Huebinger AM, Deijs M, de Vries M,
Drexler JF, Oppong SK, Mueller MA, Klose SM, Well-
inghausen N, Cottontail VM. 2011. Two novel parvovi-
ruses in frugivorous new and old world bats. PLoS One 6:
e29140.
137. Donaldson EF, Haskew AN, Gates JE, Huynh J, Moore
CJ, Frieman MB. 2010. Metagenomic analysis of the
viromes of three North American bat species: viral diversity
among different bat species that share a common habitat.
J Virol 84:13004–13018.
138. Ge X, Li J, Peng C, Wu L, Yang X, Wu Y, Zhang Y, Shi Z.
2011. Genetic diversity of novel circular ssDNA viruses in
bats in China. J Gen Virol 92:2646–2653.
139. Kapoor A, Simmonds P, Dubovi EJ, Qaisar N, Henriquez
JA, Medina J, Shields S, Lipkin WI. 2011. Characteriza-
tion of a canine homolog of human aichivirus. J Virol 85:
11520–11525.
140. Li L, Pesavento PA, Shan T, Leutenegger CM, Wang C,
Delwart E. 2011. Viruses in diarrhoeic dogs include novel
kobuviruses and sapoviruses. J Gen Virol 92:2534–2541.
141. Li L, Shan T, Wang C, Côté C, Kolman J, Onions D, Gul-
land FM, Delwart E. 2011. The fecal viral flora of Califor-
nia sea lions. J Virol 85:9909–9917.
142. Li L, Victoria JG, Wang C, Jones M, Fellers GM, Kunz
TH, Delwart E. 2010. Bat guano virome: predominance
of dietary viruses from insects and plants plus novel mam-
malian viruses. J Virol 84:6955–6965.
143. Reuter G, Nemes C, Boros Á, Kapusinszky B, Delwart E,
Pankovics P. 2012. Astrovirus in wild boars (Sus scrofa) in
Hungary. Arch Virol 157:1143–1147.
144. Reuter G, Pankovics P, Delwart E, Boros Á. 2012. Iden-
tification of a novel astrovirus in domestic sheep in Hun-
gary. Arch Virol 157:323–327.
145. Kapoor A, Slikas E, Simmonds P, Chieochansin T,
Naeem A, Shaukat S, Alam MM, Sharif S, Angez M,
Zaidi S. 2009. A newly identified bocavirus species in
human stool. J Infect Dis 199:196–200.
146. Rock C, Alum A, Abbaszadegan M. 2010. PCR inhibitor
levels in concentrates of biosolid samples predicted by a
new method based on excitation-emission matrix spectro-
scopy. Appl Environ Microbiol 76:8102–8109.
147. Sano D, Fukushi K, Yoshida Y, Omura T. 2003. Detec-
tion of enteric viruses in municipal sewage sludge by a com-
bination of the enzymatic virus elution method and
RT-PCR. Water Res 37:3490–3498.
148. Enterovirus R, Real-Time ST. 2000. RT-PCR. Biotechni-
ques 29:88–93.
149. Kopecka H, Dubrou S, Prevot J, Marechal J, Lopez-Pila
J. 1993. Detection of naturally occurring enteroviruses in
waters by reverse transcription, polymerase chain reac-
tion, and hybridization. Appl Environ Microbiol 59:
1213–1219.
150. Puig M, Jofre J, Lucena F, Allard A, Wadell G, Girones R.
1994. Detection of adenoviruses and enteroviruses in pol-
luted waters by nested PCR amplification. Appl Environ
Microbiol 60:2963–2970.
151. Tsai Y-L, Sobsey M, Sangermano L, Palmer C. 1993. Sim-
ple method of concentrating enteroviruses and hepatitis A
virus from sewage and ocean water for rapid detection by
 3.1.5. Waterborne Enteric Viruses: Diversity, Distribution, and Detection ▪ 3.1.5-13

reverse transcriptase-polymerase chain reaction. Appl Envi-
ron Microbiol 59:3488–3491.
152. U.S. EPA. 2004. Quality Assurance/Quality Control
Guidance for Laboratories Performing PCR Analyses on
Environmental Samples. EPA 815-B-04-001. U.S. EPA.
153. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J,
Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL.
2009. The MIQE guidelines: minimum information for
publication of quantitative real-time PCR experiments.
Clin Chem 55:611–622.
 Detection of Protozoa in Surface and Finished Waters
ABSAR ALUM, ERIC N. VILLEGAS, SCOTT P. KEELY, KELLY R. BRIGHT,
LAURA Y. SIFUENTES, AND MORTEZA ABBASZADEGAN
3.1.6
Humans are known to be the host to approximately 1,500
infectious agents, out of which 66 are protozoa and 287
are helminths (1, 2). From a global perspective, helminths
and protozoan parasites account for approximately one
fourth of the total infectious diseases of humans (Fig. 1). A
similar trend has been observed in waterborne infectious dis-
eases, among which a significant part is caused by enteric par-
asites (4).
Cryptosporidium and Giardia are the leading cause of water-
borne outbreaks of gastroenteritis across the globe (5, 6), and
are discussed in length in this chapter. These parasites are par-
ticularly suited for waterborne transmission as the environ-
mentally resistant cysts and oocysts, respectively, are shed in
large numbers in feces (108–109 oocysts/g), have a low infec-
tious dose, and are resistant to disinfection practices (7).
DESCRIPTION OF ETIOLOGICAL AGENTS
Cryptosporidium spp. were first described in 1907, but for the
next half century it was not recognized as a pathogen (8). The
first outbreak of Cryptosporidium-related diarrheal disease in
calves was reported in 1971 (9), and five years later Crypto-
sporidium was first reported as the causal agent of diarrhea in
children and immunosuppressed patients (10). Cryptospori-
dium is an apicomplexan protozoan (Phylum: Apicomplexa,
Class: Sporozoea, Subclass: Coccidia, Order: Eucoccidia,
Suborder: Eumeriina, Family Cryptosporidiidae, Genus:
Cryptosporidium) (11). To date 26 valid species of Cryptospori-
dium are known with more than 80 genotypes reported (Table
1). This parasite is capable of infecting humans and other
vertebrates, and different species/genotypes are known to
cause infection in various parts of the digestive tract (Table
2). Among the known Cryptosporidium species, C. hominis,
C. parvum, and C. meleagridis have caused over 90% of
the reported cases of cryptosporidiosis in humans (15, 16).
Rare cases of human infections caused by C. andersoni,
C. bovis, C. canis, C. cuniculus, C. fayeri, C. felis, C. muris,
C. scrofarum, C. suis, C. tyzzeri, C. ubiquitum, and C. viatorum
have also been reported, while the remaining species are only
found in animals.
Giardia was first described by Antonie van Leeuwenhoek
(1632–1723), and the description was not very precise. In
1859, Vilem Lambl (1824–1895) provided a more precise
description of the flagellate Giardia and the species name lam-
blia was given by Blanchard in 1888 (17). Giardia lamblia is a
doi:10.1128/9781555818821.ch3.1.6
3.1.6-1
unicellular diplomonadid flagellated protozoan parasite and
has two stages in its life cycle, the cyst and the trophozoite.
Originally, host specificity and morphological features were
used as criteria for speciation/classification, resulting in recog-
nition of more than 40 Giardia “species” (18). This practice
continued until 1950s, when it became clear that Giardia is
not strictly a host-specific pathogen and morphometric anal-
yses are not reliable criteria for speciation (19). Recent devel-
opment in biochemical and molecular techniques have
provided tools for better characterization of Giardia species
(20, 21). The new systematical classification is based on
genetic, structural, and biochemical data: Phylum: Metamo-
nada, Subphylum: Trichozoa, Superclass: Eopharyngia, Class:
Trepomonadea, Subclass: Diplozoa, Order: Giardiida, Family:
Giardiidae (22) (Table 3).
INFECTION CYCLE AND DISEASE SYMPTOMS
The Cryptosporidium and Giardia genera are known to infect a
wide range of hosts, including domestic and wild animals and
humans (Tables 1 and 3). After ingestion, the (oo)cysts go
through a series of developmental and morphological changes
as illustrated in Fig. 2. One of the major differences in the
infection process of these parasites lies in fact that all stages
of the Giardia life cycle are completed in the intestinal lumen
and do not involve any intracellular stage, whereas the Cryp-
tosporidium life cycle requires an intracellular stage (Fig. 3).
The prepatent period (time between the initial infection
and detection of oocysts in feces) of Cryptosporidium can
range from 2 to 10 days. In immunocompetent subjects,
asymptomatic cryptosporidiosis is possible. However, that
does not exclude the excretion of oocysts (20). The disease
is characterized by watery diarrhea, dehydration, fever, ano-
rexia, weight loss, weakness, abdominal cramps, vomiting,
lethargy, general malaise, and progressive deterioration of
overall condition (26). In general, immune status of subjects
can be a determining factor in disease sequelae. In healthy
subjects, the disease is self-limiting and the symptoms usually
last from several days to two weeks and in immunocompro-
mised subjects disease symptoms are more acute and may
lead to severe morbidity and even death (27). In case of Giar-
dia the prepatent period can range from one to three weeks.
Symptoms of giardiasis include; diarrhea, dehydration, nau-
sea, abdominal cramps, flatulence, and greasy stool (28).
The disease symptoms may last from 2 to 6 weeks.
3.1.6-2 ▪ WATER

TABLE 2 Selected species of Cryptosporidium and Giardia species

and their preferred site of infection in host digestive system

Site of infection Cryptosporidium and Giardia species

Small intestine C. bovis; C. canis; C. cuniculus; C. fayeri;
C. felis; C. hominis; C. macropodum;
C. meleagridis; C. parvum; C. ryanae;
C. scophthalmi; C. wrairi; C. tyzzeri;
C. ubiquitum; C. viatorum G. duodenalis
Stomach C. fragile; C. huwi; C. molnari; C. muris;
C. serpentis; C. varanii
Large intestine C. suis
FIGURE 1 Distribution of major etiological agents causing infec- Abomasum C. andersoni
tious diseases in humans. Reprinted from (3), with permission. Cloaca, bursa, trachea C. baileyi
doi:10.1128/9781555818821.ch3.1.6.f1
Proventriculus C. galli
Not known C. xiaoi; C. scrofarum
Since Cryptosporidium and Giardia infection in healthy
adults are self-limiting (29), a significant number of patients
may not seek medical advice; therefore, the epidemiological
data do not reflect actual disease prevalence, especially in water-transmitted infectious diseases. Cryptosporidium was
developing countries. The environmental infectious stages not considered a significant waterborne public health threat
of Cryptosporidium (oocyst) and Giardia (cyst) are excreted until 1993 when the largest documented cryptosporidiosis
in feces and are relatively resistant to environmental and outbreak of waterborne disease in the United States caused
chemical stressors used during water or sewage treatment an epidemic infecting 403,000 people and resulting in 112
(e.g., chlorination and UV irradiation), which make them deaths (30). Since then Cryptosporidium has been a pathogen
significant contributors in the total load of waterborne or of high interest among water industry stakeholders, regulatory
agencies, and the scientific community. This interest resulted
in the development of a series of methods that have been used
TABLE 1 Validated Cryptosporidium speciesa
for the detection of (oo)cysts in a variety of water matrices
Name Major hosts (31). The occurrence of Cryptosporidium oocysts and Giardia
cysts in different types of water is well documented, and a con-
C. andersoni Cattle siderable number of waterborne outbreaks of these parasites
C. baileyi Chicken and turkey has been reported around the world (4, 6). A worldwide over-
C. bovis Cattle view of the combined waterborne cryptosporidiosis and giar-
C. canis Dog, coyote, and fox diasis outbreaks is given in (6) and (4) (Fig. 4).
C. cuniculus Rabbit and humans
C. fayeri Red kangaroo
TABLE 3 Giardia species and assemblagesa
C. felis Cat
C. fragile Black spined toad Name Major hosts
C. galli Chicken G. agilis Amphibians
C. hominis Human G. ardeae Birds
C. huwi Guppy G. duodenalis, assemblage A Humans
C. macropodum Eastern grey kangaroo (G. duodenalis)*
C. meleagridis Turkey G. duodenalis, assemblage B Humans
C. molnari Fish (G. enterica)*
C. muris Rodents G. duodenalis, assemblage C Dogs
C. parvum Human and cattle (G. canis)*
C. ryanae Cattle G. duodenalis, assemblage D Dogs
(C. canis)*
C. scrofarum Pig
G. duodenalis, assemblage E Cattle and other hoofed
C. scophthalmi Fish (G. bovis)* animals
C. serpentis Snakes
G. duodenalis, assemblage F Cats
C. suis Pig (C. cati)*
C. tyzzeri Mice G. duodenalis, assemblage G Rodents
C. ubiquitum Sheep (G. simondi)*
C. varani Lizard G. duodenalis, assemblage H Seals
C. wrairi Guinea pig G. microti Muskrats and voles
C. viatorum Human G. muris Rodents
C. xiaoi Sheep G. psittaci Birds
a a
Data from (12–14). Data from (12, 23, 24), names in parentheses are proposed species.
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-3
FIGURE 2 Life cycle of Cryptosporidium and Giardia: a comparative illustration. Courtesy of the U.S. Centers for Disease Control and Pre-
vention. doi:10.1128/9781555818821.ch3.1.6.f2
DETECTION OF CRYPTOSPORIDIUM AND
GIARDIA IN WATER
Microbial quality of source and finished drinking water in the
United States is ascertained by complying with the total coli-
form rule. However, coliforms are unreliable indicators of the
presence or viability of parasitic (oo)cysts. Cryptosporidium
oocysts and Giardia cysts typically occur in source water in
low numbers (31–33), therefore, large volumes of water
must be processed and analyzed to reliably reach a conclusion
about the presence or absence of parasite (oo)cysts. These fea-
tures along with the small size of (oo)cysts make their detec-
tion a daunting task, requiring intensive work by well-trained
laboratory personnel. These circumstances stipulate that (oo)
cysts monitoring in surface and drinking waters be performed
using standardized techniques and the laboratory performing
FIGURE 3 Infection process of Cryptosporidium and Giardia.
Reprinted from (25), with permission.
doi:10.1128/9781555818821.ch3.1.6.f3
these tasks should have documented capability to perform
such assays. The methods recommended by regulatory agen-
cies and widely used by stakeholders are based on four sequen-
tially performed steps: (a): primary concentration—filtration
of enough volumes (1–1,000 liters) of water to maximize (oo)
cysts recovery; (b) secondary concentration of (oo)cysts and
purification using techniques such as immunomagnetic sepa-
ration (IMS); (c) staining of sample concentrate using spe-
cific fluorescent antibodies (FAs); and (d) microscopic
enumeration using fluorescence and differential interference
contrast microscopy. These steps have different levels of diffi-
culties and may take 0.5 to 4.5 h to complete with a broad
range of reported recoveries for these steps (Table 4).
Over the years, a variety of methods have been described
for (oo)cysts purification purposes, and these methods have
included the use of density gradient, saturated salt solution
centrifugation, continuous flow centrifugation, and flow
cytometry with cell sorting (38). In an effort to develop a sim-
pler, more efficient, standardized method, a variety of steps
from different methods have been combined in one protocol,
yielding wide range of results (11, 34, 39).
In 1996, following changes to the Safe Drinking Water
Act, the U.S. EPA began developing an improved method
for monitoring Cryptosporidium in water (28). In 1999, the
EPA published Method 1622 for the detection of Cryptospori-
dium in water. Since then several additional versions of this
method have been published (40). Most recently in 2012,
the EPA published Method 1623.1 for the detection and enu-
meration of Cryptosporidium and Giardia (41). The method
relies on filtration, IMS followed by immunoflorescent assay
microscopy. The viability of organisms could be ascertained
by staining with 40 ,6-diamidino-2-phenylindole (DAPI)
and differential interference contrast microscopy. Method
1622/1623.1 required a 10-liter grab sample concentrated
3.1.6-4 ▪ WATER
FIGURE 4 Distribution of waterborne outbreaks due to Cryptosporidium and Giardia in different continents. Reprinted from (2), with per-
mission. doi:10.1128/9781555818821.ch3.1.6.f4
using Envirochek HV Sampling Capsule (Pall Corporation),
or Filta-Max foam filter (IDEXX). The ISO 15553 standard
protocol (42) is equivalent of EPA Method 1623.
Over the years various types of filters have been developed
for the detection of waterborne (oo)cysts, and these methods
yield wide range of recovery efficiencies (Table 5). The hol-
low fiber ultrafilter offers much improved recovery rates for
Cryptosporidium and Giardia (oo)cysts (54–58), and a detailed
video describing the use of this approach in conjunction with
EPA Method 1623 has also been published (58). One major
advantage for using these filters is their capabilities to simul-
taneously concentrate protozoa, bacteria, and viruses (55–
60). Despite global acceptance of Methods 1622/1623 for
the detection of the detection of waterborne Cryptosporidium
and Giardia, these methods have been be criticized for their
low recovery efficiencies, cost, and the length of time to ana-
lyze the sample and recommendations have been made for
optimizing various detection methods (31).
Method 1623 is a significant improvement over all other
previous methods and is globally accepted for regulatory mon-
itoring of Cryptosporidium and Giardia in source and finished
waters, and these methods are part of the regulatory require-
ments in many countries. However, the method still has
many limitations, specifically its inability to provide informa-
tion on the viability/infectivity and species/genotypes of the
detected (oo)cysts. Since infectivity status of the detected
(oo)cysts is critical in accurately assessing public health risk
posed these waterborne parasites, this method is likely to
overestimate human health risk (61).
TABLE 4 Summary of steps for Cryptosporidium detectiona
Step Method
Primary concentration Flocculation
Filtration
Ultrafiltration
Continuous flow centrifugation
Secondary concentration Density gradient separation
and purification Immunomagnetic separation
Detection Immunofluorescence assay (IFA)
Flow cytometry
a
Based on (34–37).
METHODS FOR VIABILITY AND INFECTIVITY
DETERMINATION
Animal infectivity assay is the most direct method for assess-
ing the viability/infectivity of Cryptosporidium oocysts or
Giardia cysts and it is the gold standard for comparing the effi-
cacy of alternate methods. Animal infectivity assays for Cryp-
tosporidium are commonly performed using neonatal or
immunosuppressed rodents, typically CD-1 or BALB/c
strains (62, 63). However, infectious doses of different (oo)
cysts may vary among different animal models. For example,
up to 1,000 oocysts may be required to achieve 100% infec-
tion in an animal model (64), therefore this method is
impractical for the routine monitoring of infectious oocysts
present in environmental samples since the concentrations
found are very low.
Animal infectivity assays require specialized laboratory set-
tings and compliance with institutional review and approval
mechanism, therefore these assays are costly, labor intensive,
and time-consuming and pose many ethical issues. Since late
1980s efforts have been made to identify suitable in vitro sur-
rogate systems for animal infectivity. Early on, studies using
healthy/fresh stocks of parasites showed that excystation
assays can be a potentially useful surrogate for animal models
of infection (65, 66). Subsequent studies questioned earlier
findings and showed that excystation assays were not suitable
as a measure of viability/infectivity, especially in cases where
oocysts have been exposed to environmental stresses, UV
irradiation, or ozone treatment (62, 67–71).
Recovery Minimum
efficiency range (%) processing time (h)
45–75 4.5
15–80 0.5
47–113
30–85 1.0
60–70 0.3
85–95 1.5
95–99 2.5
95–99 1.5
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-5

TABLE 5 Recovery efficiencies of various filters used for the detection of Cryptosporidium oocysts in a variety of water matrices
Nominal pore Recovery (%)
Filter Type size (µm) Reference
Range Mean (SD)
Polypropylene cartridge filter 1.0 25.0–81.0 50a (43)
3.0–16.5 10.8b (44)
9.6–16.5 13.3a (45)
3.0–16.5 10.8c
15.0 4.5–16.1 11.2a (46)
3.9–14.2 9.4c
Polycarbonate membrane filter 2.0 18.6–34.3 26.2 (5.5)a (47)
4.4–7.0 6.0a (45)
3.6–7.2 5.6c
24.0–32.0 27.2a (46)
32.0–38.7 36.0c
Cellulose acetate membrane filter 1.2 24.5–55.7 37.8a (46)
23.6–33.0 26.6c
32.0–44.0 39.7a (46)
32.0–42.7 38.1c
12–19 16.2 (2.8)a (36)
0–14 3.2 (6.1)b
Cellulose nitrate membrane filter 1.2 30.8–52.2 42.1 (5.9)b (48)
3.0 19.3–31.6 24.1a (46)
5.4–8.4 7.1c
24.0–36.0 29.3a (46)
25.3–36.0 30.4c
5.0 6.1–14.6 10.3a (46)
3.6–5.4 4.6c
Versapore 1.2 10.2–36.0 22.5a (46)
29.3–41.3 n/ad
19.9–36.6 25.5 (4.7)a (49)
3.0 17.3–28.0 20.8a (46)
18.6–26.7 22.4c
Polyethylsulfone 0.8 21.3–26.7 24.0a (46)
18.7–25.3 21.8c
Capsule filter 4–44 15 (12)c (50)
37–43 n/a (51)
36–75 n/a
9–22.9 14.4 (7.5)a (52)
n/a 81 ± 15b (31)
n/a 37 ± 29a
n/a 29 ± 26c
Filta Max 1.0 76.7–106.7 90.4a (53)
n/a 28.2 (8.0)b (36)
n/a 19.4 (6.3)c
Hollow-fiber ultrafilters ∼30 KDa n/a 53 (19)a (54)
n/a 51 (22)a
n/a 47 (12)c (55)
n/a 54 (23)c
n/a 100 (4)c (56)
n/a 113 (15)a (57)
a
Tap water.
b
Distilled water.
c
Raw water.
d
n/a; not available.

Vital dyes have also been used as surrogates for viability
determination environmental studies. The fluorogenic dyes
such as propidium iodide (PI) and DAPI were some of the first
widely used techniques. These techniques essentially measure
the permeability of oocyst membrane (66). The cells which
are stained with both dyes are considered nonviable, whereas
cells which take up DAPI but are not stained with PI are con-
sidered viable. The accuracy of this dye permeability assays
has also been questioned (72, 63). Other selective exclusion
dyes such as SYTO-9 or SYTO-59 have also been investigated
as a surrogate of viability/infectivity (67, 73). Based on a com-
prehensive study, it has been shown that excystation, DAPI/
3.1.6-6 ▪ WATER
PI, SYTO-59, and SYTO-9 staining are not reliable predic-
tors of viability of disinfected Cryptosporidium oocysts or
Giardia cysts (68, 71). More recently, the use of propidium
monoazide in conjunction with PCR as an alternative
molecular-based viability assays for Cryptosporidium and Giar-
dia have been reported (74). Another variation of fluorescent
labeling technique is the fluorescence in situ hybridization
(FISH), which use fluorescently labeled oligonucleotide
probes to target specific sequences of ribosomal RNA
(rRNA). The oocysts are considered viable if they contained
the fluorescing sporozoites after hybridization with probes and
the nonfluorescing oocysts are considered nonviable (75). As
with fluorescent staining methods, the reliability of FISH for
viability determination of UV-inactivated cells has also been
questioned (76).
The evolution of molecular techniques has revolutionized
biological research in general, and viability testing is not an
exception. Since the 1990s, a large amount of scientific infor-
mation has been generated on the application of molecular
techniques for viability determinations. Many of these tech-
niques rely on detection of messenger RNA (mRNA), which
should be present only in viable organisms. One of the advan-
tages in the application of mRNA based techniques is the
instability of mRNA, which has a very short half-life, from
a few seconds to a few minutes (77). Viability studies have
used a variety of mRNA targets including: heat shock 70 pro-
tein (hsp70), β-tubulin, amyloglucosidase, the membrane pro-
tein CP2 and variant-specific protein (78–82). In several
studies, reverse transcriptase PCR (RT-PCR) amplification
platforms have been used to examine mRNA-based targets,
which require thermocycling machines to facilitate the reac-
tion. Baeumner et al. also reported the use of a single isother-
mal reaction called nucleic acid sequence-based
amplification (NASBA) for detecting viable oocysts (83).
Theoretically, these methods have tremendous potential for
direct applicability for detecting the low numbers of oocysts
commonly found in source and finished water, however
thus far the water industry has been reluctant in adopting
these methods for routine monitoring.
CELL CULTURE METHODS
A standardized cell culture system can be the best alternate
to animal infectivity studies for differentiating the infectious
TABLE 6 Cell line for infectivity studies of Cryptosporidium oocysts
Cell lines Cell line description
BALB/3T3 Mouse embryo
BT-549 Human breast infiltrating ductal carcinoma
Caco-2 Human colonic adenocarcinoma
HCT-8 Human ileocecal adenocarcinoma
HS-700 T Human pelvic adenocarcinoma
HT-29 Human colonic adenocarcinoma
HT-1080 Human fibrosarcoma
LS-174 T Human colonic adenocarcinoma
MDBK Madin-Darby bovine kidney
MDCK Madin-Darby canine kidney
RL95-2 Human endometrial carcinoma
FHs 74 Human small intestinal epithelial
a
ATCC; American Type Culture Collection.
from noninfectious oocysts. Chicken embryos were the first
system reported for supporting the complete development of
C. parvum and C. baileyi (84). Thereafter, the endodermal
cells of the chorioallantoic membrane of chicken embryos
were used to propagate the C. parvum. However, this system
was inefficient and difficult to perform under various param-
eters (85). In mid-1980s a cell culture system using human
rectal tumor cells was reported to support development of
C. parvum sporozoites into mature meronts (86, 87). Since
then a variety of cell line have been reported for partially
or fully supporting the propagation of Cryptosporidium in
an in vitro systems (88). The nature of cell line play a critical
role in such in vitro propagation systems and complete devel-
opment of C. parvum has been reported in several cell lines
with the production of some oocysts (84, 88–96); some of
these cell lines are listed in the Table 6.
Slow growing cell lines such as Caco2 can be beneficial in
developing cell culture assays for parasites, because minimal
host growth is desired during the infection process. The
Caco2 cell line has been described to display epithelial
cell–like features like brush border microvilli, tight junctions,
and active sodium chloride transport during growth (102).
Other cell lines such as BFTE, MDCK, and BS-C-1 have
also been shown to support the complete life cycle (sexual
and asexual stages) of this parasite (from sporozoites to sporu-
lated oocysts) (88). There are other cell lines that support a
partial life cycle of Cryptosporidium—for example, THP-1,
HRT, and LGA support the asexual phase only (87, 103–
105). Upton et al. (96) tested 11 cell lines for their ability
to support the growth of C. parvum and found that HCT-8
cells are the best for supporting nearly twice the number of
developmental stages of Cryptosporidium compared with
Caco-2 cells, MDCK cells, or any other line tested. In vitro
culturing of C. parvum in HCT-8 cells allow for greater para-
site development, raising the percentages of parasitized cells,
and it facilitates the observation of all developmental phases
of this parasite (92, 93). Recent studies also reported that
infected FHs 74 cells showed similar numbers of developmen-
tal stages but with higher infection rates (96). Additionally,
monolayers as old as 67 days have been shown to support
infection, and no difference in infectivity was observed
when evaluating fresh monolayers (2 days old) and aged
monolayers (8–22 days old) from repeated infectivity assays
on HCT-8 cells (106).
ATCCa # References
CCL-163 (96)
HTB-122 (96)
HTB-37 (96); (97)
CCL-244 (96); (98); (99)
HTB-147 (96)
HTB-38 (96)
CCL-121 (96)
CCL-188 (96)
CCL-22 (96)
CCL-34 (96)
CRL-1671 (96)
CCL-241 (100)

Over the years, many cell lines, various medium formula-
tions, different assay formats, and sample processing techni-
ques have been tested for application of cell culture
methods to determine the viability/infectivity of Cryptospori-
dium oocysts in water (97, 107–109). The cell culture assays
for waterborne oocysts are mostly based on HCT-8 cell line
and cell culture system and have been integrated with IFA,
PCR, or RT-PCR for improving the detection sensitivity
and specificity and reducing the assay time.
In contrast to Cryptosporidium, no cell culture system is
available for determining the infectivity of Giardia cysts,
and all the infectivity studies conducted to date used animal
models (110, 111). Although axenic culture of Giardia has
been reported, it has some limitations (112). Since there
are currently no effective purification procedures to com-
pletely remove fecal bacterial contaminants from Giardia
cysts, axenic cultures of cysts are very difficult to achieve
because the cultures are quickly overrun by bacteria (113).
In the absence of any cell culture systems, all the efforts
have been focused on viability studies using mRNA-based
molecular techniques (77) and other techniques involving
fluorescent tags (114). Recently, the first cell culture–based
system was reported for determining the viability/infectivity
of Giardia cysts (81). This in vitro system relies on the infection
mechanism of Giardia and measures the trans-cellular resist-
ance of Caco2 cells that had been exposed to excysted Giar-
dia. As a confirmatory step, RT-PCR is performed to
quantify the expression of variant specific protein on tropho-
zoites (81). This report opens a new research direction for the
development of cell culture models for parasites infectivity
determination.
MOLECULAR METHODS FOR DETECTION
AND CHARACTERIZATION OF
CRYPTOSPORIDIUM AND GIARDIA
Species and assemblages identification by morphologic and
morphometric analyses are very difficult to do given the
overlap in sizes, structures, and morphologies among Crypto-
sporidium and Giardia (oo)cysts. Advances in molecular tech-
niques allow for more specific genotyping of these organisms.
Various genotyping techniques have been crucial in under-
standing population structure and molecular epidemiology
of Cryptosporidium (12) and Giardia (52) species contaminat-
ing water supplies. Current drinking water regulations, like
the Long Term 2 Enhanced Surface Water Treatment Rule
(LT2) (115), require monitoring surface waters for Crypto-
sporidium and Giardia using EPA Method 1623 (40) or
1623.1 (41). However, these techniques alone only detect
down to the genus level and cannot determine species or
assemblages. In addition, molecular epidemiological studies
TABLE 7 Selected Cryptosporidium and Giardia loci used for species and subtype identification
Organism Gene
Cryptosporidium SSUrRNA
hsp70
COWP
gp60
Giardia β-giardin
SSUrRNA
tpi
gdh
a
Size correspond to the PCR product length amplified from different parts of the gene as originally described.
3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-7
have shown that the majority of the Cryptosporidium oocysts
detected in various surface waters were of animal origin, while
C. hominis or C. parvum were detected in less than 2% of sam-
ples (116, 117). Many molecular techniques have been used
for the detection of these parasites including PCR, RT-PCR,
and NASBA (77, 80, 81, 101, 118). Due to their sensitivity
and specificity, molecular techniques have been widely used
for characterization of the detected (oo)cysts to species and
genotype levels. The two most commonly used techniques
are PCR-restriction fragment length polymorphism (RFLP)
and single-locus sequence typing. PCR-RFLP analysis of
Cryptosporidium and Giardia has two parts. The first part
involves PCR amplification of the gene of interest, and the
second part uses specific restriction endonuclease to digest
the amplified products into distinct DNA fragments resolved
using agarose gel electrophoresis that can be used to differen-
tiate species of target microorganism. The most common gene
used for PCR-RFLP analysis of Cryptosporidium and Giardia
are presented in Table 7. The SSUrRNA is the most studied
gene for genotyping Cryptosporidium and Giardia (132).
This procedure uses a nested PCR that amplifies the
SSUrRNA gene using the primers listed in Table 8 (133,
134). The final nested PCR reaction is typically digested
with SspI and VspI enzymes that provides sufficient basis to
identify common Cryptosporidium species and genotypes. In
addition to SSUrRNA, Cryptosporidium oocyst wall protein
(COWP) (124), and heat shock protein 70 (hsp70) (135,
136) genes have also been used for genotyping of Cryptospori-
dium. Giardia genotyping can be accomplished using
β-giardin, triphosphate isomerase (tpi), glutamate dehydro-
genase (gdh), or SSUrRNA genes (52). Among these genes,
the β-giardin is the best choice gene because it has higher sub-
stitution rates than the SSUrRNA and thus has broader appli-
cations for species/assemblage identification and subtyping
of Giardia (52). Detailed procedure has been described by
Caccio and colleagues (130), and primers are listed in Table
8. The resulting PCR product is then digested using HaeIII
enzyme and analyzed by agarose gel electrophoresis.
Single-locus sequence typing of Cryptosporidium and Giar-
dia involves sequencing of hypervariable regions of the
genome and is considered the gold standard for species/geno-
type identification. This procedure is the same as the
PCR-RFLP except that the PCR product is sequenced by con-
ventional Sanger sequencing techniques instead of being
enzymatically digested. The SSUrRNA is also the most widely
studied gene for species/genotype identification using
sequencing (137). It is estimated that there are at approxi-
mately four to five copies (138) in the genome, resulting in
a more sensitive assay for detecting oocysts in water as com-
pared to other loci, like hsp70 or COWP (139). The Centers
for Disease Control and Prevention (CDC) has proposed
Size (bp)a Reference
298; 590; 660; 824–864; 1,056; 1,750 (23, 24, 119–121)
325; 1,950 (122, 123)
568; 769; 923 (124–126)
800–850 (127–129)
511; 753 (130, 131)
130; 292 (52)
532; 605 (52)
432; 530; 532; 754 (52)
3.1.6-8 ▪ WATER

TABLE 8 Selected Cryptosporidium and Giardia loci used for PCR-RFLP

Organism Gene Primer seq. Size (bp) Reference
0 0
Cryptosporidium SSUrRNA 1° F: 5 -TTC TAG AGC TAA TAC ATG CG-3 1,325 (133, 134)
1° R: 50 -CCC ATT TCC TTC GAA ACA GGA-30
2° F: 50 -GGA AGG GTT GTA TTT ATT AGA TAA AG-30 824–864a (133, 134)
2° R: 50 -AAG GAG TAA GGA ACA ACC TCC A-30
G. duodenalis β-giardin G7: 50 -AAG CCC GAC GAC CTC ACC CGC AGT GC–30 753 (130)
G759: 50 -GAG GCC GCC CTG GAT CTT CGA GAC GAC-30
GiarF: 50 -CTC GAC GAG CTT CGT TGT T-30 511 (131)
GiarR: 50 -CTC GAC GAG CTT CGT TGT T-30
a
Size vary depending on species/genotype.

using the SSUrRNA sequence typing as the standard method
for cryptosporidiosis outbreak investigations (140). One of
the most polymorphic markers described in the Cryptospori-
dium genome is Gp60, a gene expressed on the surface apical
region of the sporozoite that is involved in the invasion
process. It has been instrumental toward understanding intra-
species population structure, molecular epidemiology, trans-
mission dynamics, and virulence of this parasite (141, 142).
This gene is approximately 800–850 bp in size (Table 7). It
contains variable tandem trinucleotide repeat sequences
coding for the serine amino acid (TCA, TCG, or TCT)
and nonrepeat regions with extensive sequence variability.
The sequence polymorphisms in the nonrepeat regions are
used to categorize C. parvum and C. hominis into different
subtype families, while the trinucleotide repeats are used to
distinguish the different members within each subtype family.
The procedure for PCR-sequencing of the gp60 is described
in detail elsewhere (23, 24, 143).
The four most commonly used markers to identify Giardia
species and assemblages are listed in Table 7 (52). They pro-
vide varying levels of resolution with some markers only iden-
tifies Giardia at the genus or species level, while others can
resolve down to the assemblage and subtype levels. The
SSUrRNA gene is used for genus- and species-level identifica-
tion because it contains both conserved and hypervariable
regions that are most useful for genotype classification. The
SSUrRNA gene can easily classify Giardia at the species level,
but assemblage typing of G. duodenalis is more difficult. To
gain a better understanding the molecular epidemiology of
Giardia, gdh and tpi are more appropriate for species, assem-
blage, and subtype identification (52). Among the different
loci used to date, the β-giardin sequence database contained
the largest number of submitted sequences and is the preferred
marker used for Giardia species/assemblage identification.
There are two options that allow for genotyping protozoa
detected using EPA Methods 1623/1623.1. The first option is
extracting DNA immediately after the IMS step called
“off-the-bead” (O-T-B) genotyping (116, 133, 144), and
the second is after microscopic analysis of the slide samples
“off-the-slide” (O-T-S) genotyping (145, 146). Both were
shown to be effective at identifying down to species and sub-
type resolution. The O-T-B technique examines a portion of
the packed water concentrate pellet processed through the
IMS step as described in EPA Method 1623/1623.1 (40, 41,
147). This procedure subjects the bead:oocyst/cyst complex
through a series of freeze-thaw cycles followed by a DNA
extraction procedure (133, 144, 148). The O-T-S procedure,
on the other hand, extracts Cryptosporidium and Giardia DNA
directly from the microscopic slides, which is more compati-
ble with EPA Methods 1623/1623.1 than the O-T-B
procedure since DNA is collected after microscopic enumer-
ation. One advantage of the O-T-S procedure is that it can
determine concentration of these parasites in the water as
well as genotype those detected by microscopy. The original
report by Nichols (149), modified by DiGiovanni (146),
describes the O-T-S technique. Ware and colleagues further
modified the procedure to allow for genotyping Giardia cysts
that are also captured on the microscopic slides (145). A
detailed video of the O-T-S procedure can be found online
(https://www.youtube.com/watch?v=_JyK9WwIp4w).
Once the DNA has been extracted, PCR-RFLP and
sequence typing procedures can be used for molecular-based
identification of Cryptosporidium species/genotypes and Giar-
dia species/assemblages as well as the study of environmental
transmission, occurrences of human and zoonotic species, and
molecular epidemiology of Cryptosporidium and Giardia in
surface and drinking waters (149, 150). There is also a simpler
and more rapid technique that is compatible with the O-T-S
method to address more specific issues, like differentiating
human versus animal infectious forms of Cryptosporidium.
DiGiovanni and colleagues developed a conventional PCR
assay or a quantitative real-time PCR high-resolution melt
curve (qPCR-HRM) analysis to differentiate the most com-
mon human infectious species, C. parvum, C. hominis, and
C. meleagridis, from other Cryptosporidium spp. (146). The
conventional PCR procedure is designed as duplex qPCR
assay which amplifies the SSUrRNA and hsp70 genes. The
appearance of the predicted SSUrRNA PCR amplicon
(435 bp) indicates the presence of Cryptosporidium spp.
oocysts, while the presence of the hsp70 PCR amplicon
(346 bp) differentiates oocysts as either C. hominis,
C. meleagridis, C. parvum, or C. cuniculus in a given sample.
For the qPCR-HRM assay, either the hsp70 or SSUrRNA
gene is amplified to distinguish C. hominis, C. meleagridis,
C. parvum, and C. ubiquitum from each other and from other
animal-specific Cryptosporidium species based on their unique
melt curve profile. This method was specifically developed as
a user-friendly and rapid procedure that provides water util-
ities additional information about the species identity of
the oocysts (e.g., human infectious versus animal species)
detected using Method 1623/1623.1. There are some limita-
tions to this approach that are worth noting. First, it cannot
differentiate novel Cryptosporidium species or genotypes that
have not been previously described. Second, as it was initially
described, it does not offer the ability to also genotype Giardia
cysts. Last, no sequence information is collected, which can
be useful for identifying novel species/genotypes or subtypes
of Cryptosporidium oocysts in the sample.
While the O-T-S has been modified to allow for species
and G. duodenalis assemblage identification (145), it has yet
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-9
to be used for molecular typing in conjunction with Method
1623/1623.1. Ware and colleagues estimated that the limits of
detection at 95% probability of detection required at least
three or seven replicates to detect at least 2 Cryptosporidium
or Giardia (oo)cyst in the slides, respectively. This O-T-S
approach is very promising not only because it is compatible
with the approved method used for surface water quality mon-
itoring in various parts of the world, but it is relatively sensi-
tive at detecting low numbers of (oo)cysts that are typically
found in surface waters (145).
Occurrences, Global Distribution, and Seasonal
Variations
Cryptosporidium and Giardia genera contain many species,
assemblages (for Giardia), genotypes, and subtypes that infect
a wide range of animal hosts from wild animals, livestock, and
pets to humans (12). For Cryptosporidium, there are currently
27 valid species with more than 80 genotypes identified from
clinical and environmental samples (12, 13, 151) (Table 1).
Phylogenetic analysis of all the known species of Cryptospori-
dium can be loosely divided into gastric and intestinal dwell-
ing parasites with the latter infecting humans (Fig. 5). Many
have been found in surface and drinking water sources at vary-
ing frequencies, and waterborne cryptosporidiosis outbreaks
have been reported worldwide (6). Among the human infec-
tious species, transmission dynamics, parasite virulence, and
zoonoses of these pathogens have been linked to intraspecies
FIGURE 5 Phylogenetic tree analyses of the valid Cryptosporidium species. Accession numbers of the reference sequences for each of the
species analyzed are listed. doi:10.1128/9781555818821.ch3.1.6.f5
variations (119, 120, 141). For example, C. hominis has been
described to have at least six subtype families with multiple
members in each subtype. Similarly, C. parvum has at least
11 subtype families also containing multiple members. It is
very likely that new subtype families and additional members
for each family will be discovered, particularly for the recently
identified zoonotic C. ubiquitum (120) and C. viatorum spe-
cies (13).
Cryptosporidium Geographic Distribution and
Seasonal Variations
Genotypes associated with human cryptosporidiosis were
observed in 35 countries (Fig. 6) with more than 73% of
the genotype observed in 7 countries: UK (28.7%), Australia
(19.7%), and India (6.7%) followed by Malaysia (6.0%),
Uganda (5.7%), China (3.7%), and the United States
(2.9%) (Fig. 7). The worldwide frequencies of C. parvum,
C. hominis and C. meleagridis were 54.7%, 39.0%, and
6.2%, respectively (Fig. 8, “All”). The frequency of
C. parvum, in these seven countries was the greatest for
United States and Malaysia (91.7 and 79.3%, respectively),
followed by Uganda (57.9%), UK (49.1%), Australia
(46.1%), India (25.0%), and China (24.2%). The C. hominis
frequencies were the inverse of C. parvum, where India
(65.6%), China (54.4%), and Australia (53.3%) have the
greatest values followed by the UK (42.8%), Uganda
(36.8%), Malaysia (17.0%), and United States (8.3%). The
3.1.6-10 ▪ WATER

FIGURE 6 Global distribution of human associated Cryptosporidium species. Frequency of clinical isolates identified by DNA sequences
found at http://cryptodb.org/cryptodb/. doi:10.1128/9781555818821.ch3.1.6.f6

FIGURE 7 Occurrences of the three most common human infectious Cryptosporidium species: C. parvum, C. hominis, and C. meleagridis based
on the number of filtered accession numbers found at http://cryptodb.org/cryptodb/. Black bars, C. parvum; open bars, C. hominis; gray bars,
C. meleagridis. doi:10.1128/9781555818821.ch3.1.6.f7
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-11
FIGURE 8 The life cycle and route of infection of Naegleria fowleri. Courtesy of the U.S. Centers for Disease Control and Prevention.
doi:10.1128/9781555818821.ch3.1.6.f8
greatest frequencies for C. meleagridis in these seven countries
were China (21.2%), India (9.4%), UK (8.1%), Uganda
(5.3%), Malaysia (3.8%), and Australia (0.6%). C. meleagri-
dis was not detected in the United States.
A seasonal pattern of cryptosporidiosis is also evident, par-
ticularly for C. parvum and C. hominis (7, 152). In the United
States, most of the reported water-related cryptosporidiosis
outbreaks occurred in the late summer season (7). This
increase was attributed to transmission of oocysts through
communal bathing areas like recreational beaches and swim-
ming pools (7). Molecular genotyping of the oocysts collected
from patients revealed C. hominis as the primary cause of
infection. Similarly, studies by Sopwith and colleagues
reported temporal variation in the occurrences of cryptospor-
idiosis in the UK between 1996 and 2000 (152). Their results
revealed two distinct peaks of waterborne cryptosporidiosis
outbreaks, one during the spring and the other in late
summer. The report showed that C. parvum was responsible
for the springtime cryptosporidiosis outbreaks while
C. hominis was associated with the late summer outbreaks.
The significance of these outbreaks forced regulatory changes
in drinking water quality management in the UK, one of
which restricted access of livestock to the drinking water
reservoirs. Once implemented, the springtime outbreaks
were substantially reduced, demonstrating that calf-derived
C. parvum oocyst contamination of the drinking water reser-
voir was the predominant cause and further implied that the
summer outbreaks were due to human-to-human transmis-
sion. A study by Fournet and colleagues further showed that
the late summer outbreaks, at least those reported in 2012,
were indeed attributed to C. hominis and the main risk factor
was traveling abroad, which may have also resulted from
ingesting inadequately treated drinking water while on
travel (121).
Giardia Species, Geographic Distribution, and
Seasonal Variation
The taxonomic nomenclature for Giardia is just as complex as
Cryptosporidium but less well understood. Currently there are
six valid species identified: G. agilis, G. ardeae, G. microti,
G. muris, and G. psittaci (12, 23) (Table 3), all of which infect
only animals, while G. duodenalis infects animals and
humans. G. duodenalis is subdivided into eight assemblages
3.1.6-12 ▪ WATER
(A–H). Assemblages A and B infect animals and are also the
only two known to infect humans. Assemblages C and D
infect dogs, assemblage E infects cattle and other hoofed ani-
mals, assemblage F infects cats, assemblage G infects rodents
(12, 23), and assemblage H infects seals (24). Given the host
specificity of G. duodenalis assemblages, species names have
been proposed to each of the assemblages and are listed in
Table 2: Assemblages A (G. duodenalis), B (G. enterica), C
and D (G. canis), E (G. bovis), F (G. cati), and G
(G. simondi). A more detailed listings of G. duodenalis
assemblages and phylogenetic analyses, are reviewed by
Feng and Xiao (52).
Feng and Xiao (52) reported that Giardia is ubiquitous and
has been found in many countries. Karanis et al. also summar-
ized that at least 132 outbreaks of giardiasis have been
reported worldwide, with 104 and 18 of the reported out-
breaks were drinking water or recreational water associated,
respectively (6). In addition, most of the reported giardiasis
outbreaks were caused by G. duodenalis assemblages A or B.
Data on the seasonal variation of waterborne giardiasis are
very limited, however, Robertson et al. reported that in
autumn and winter 2004, a giardiasis outbreak in Bergen,
Norway, occurred with over 1,500 people becoming infected
with G. duodenalis assemblage B. The suspected route of
transmission is through drinking water contaminated with
sewage that leaked from residential pipes (143). In the
summer of 2007, an outbreak of giardiasis in a private recrea-
tional camp in California was also reported. The probable
source of contamination was from a newly installed slow-sand
filtration system that was used to treat spring water for their
drinking water. G. duodenalis assemblage B was also identified
to cause this outbreak (153). A third drinking water related
outbreak occurred in New Hampshire in September 2007,
where G. duodenalis assemblage B was again identified as
the etiologic agent of the giardiasis outbreak (154).
NAEGLERIA FOWLERI: DESCRIPTION OF
ETIOLOGICAL AGENT
Naegleria fowleri is a pathogenic free-living amoeba found in
the environment in both water and soil belonging to the fam-
ily Vahlkampfiidae (155). There have been over 40 species of
Naegleria described to date, but only N. fowleri is pathogenic
to humans; N. australeinsis has been shown to be pathogenic
in mice (156). N. fowleri was first identified as a human
pathogen in 1965 in Australia (157, 158). The first case in
the United States was reported in 1966 and was described
as primary meningoencephalitis (159). Prior to this docu-
mented case, free-living amoebae were not considered patho-
genic. Pathogenic N. fowleri is not easily differentiated from
other Naegleria species due to similarities, including common
morphology when observed microscopically and indistin-
guishable behavior in cell culture (155).
N. fowleri is considered to be a frank pathogen, as it is able
to infect healthy children and young adults (160, 161). It pri-
marily infects individuals in younger age groups (children and
teenagers) who are more likely to engage in recreational water
activities that involve submersion of the head under water.
This amoeba is thermophilic, able to proliferate in tempera-
tures of up to 45°C, but prefers temperatures ranging from
30°C to 45°C (162).
Infection Cycle and Disease Symptoms
N. fowleri is an amoeboflagellate with three life stages: the
trophozoite, cyst, and flagellate stages. N. fowleri can trans-
form back and forth between the trophozoite and the
flagellate stage. The flagellate form is a transitional stage,
only lasting an hour or less, in which the organism can neither
feed nor divide (162). Flagellates are pear-shaped and range
in size from 10 to 16 µm (162), usually with two flagella
(Fig. 8). The trophozoite is the feeding stage of the organism
and its infectious form, causing primary amoebic meningoen-
cephalitis (PAM) in humans (163). Trophozoites usually
measure from 10 to 25 µm in size and reproduce by binary
fusion. Locomotion is achieved by the use of lobopodia, hem-
ispherical bulges, which allow for rapid movements (162).
The trophozoite can transition into a cyst form when the
environmental conditions change or become unfavorable,
such as when the food supply diminishes (162).
N. fowleri mostly exists as a cyst in the environment,
whether in soil or water. Cysts are resistant to different forms
of environmental stress, such as desiccation and changes in
temperature. The cyst is spherical and is composed of a double
cell wall containing a thick endocyst and a thin ectocyst
(160), measuring 8–12 µm in diameter (164). The cysts can
also revert (excyst) to the trophozoite form when the envi-
ronmental conditions become suitable (162, 165, 166).
The flagellate and the cyst forms are known to contain a
nucleus and a well-defined nucleolus (167).
N. fowleri trophozoites replicate by promitosis (nuclear
membrane remains intact). N. fowleri is found in freshwater,
soil, thermal discharges of power plants, heated swimming
pools, hydrotherapy and medicinal pools, aquariums, and
sewage. Trophozoites can turn into temporary nonfeeding
flagellated forms that usually revert back to the trophozoite
stage. Trophozoites infect humans or animals by penetrating
the nasal mucosa and migrating to the brain via the olfactory
nerves, causing PAM trophozoites to be found in CSF and tis-
sue, while flagellated forms are occasionally found in cerebro-
spinal fluid. Cysts are not seen in brain tissue (168).
N. fowleri causes PAM, a fulminating and hemorrhagic
encephalitis (163). Humans and other mammals most often
come in contact with N. fowleri during water-related recrea-
tional activities such as swimming, bathing, or, in the case
of cattle and domesticated animals, drinking from water sour-
ces containing N. fowleri. Infections are most commonly
reported in children and young adults swimming in natural
springs or warm water lakes. Another possible route of expo-
sure to is inhalation from dust in arid regions (155). The
trophozoite is the infective stage and is the form found in cer-
ebrospinal fluid and in tissues (168) (Fig. 1). The time from
exposure to the first onset of illness is 5 to 7 days (162), but
can be as quick as 24 h (169). The most common symptoms
of the disease at presentation are headache, stiff neck, seizures,
and coma (162).
The organism is acquired through the forceful inhalation
through the nose to the olfactory neuroepithelium. The
amoeba is phagocytosed in the nasopharyngeal mucosa and
then migrates via the olfactory nerve, where it eventually
invades the brain through the cribriform plate (162).
N. fowleri has the ability to evade the host immune system.
It is able to attach to the nasal mucosa, move quickly through
locomotion, and release cytolytic molecules (170). N. fowleri
causes a cytopathogenic effect thought to occur through the
production of amoebostomes that eat away and destroy tissues
(162). The immune response leads to swelling of the brain
and PAM (171). Death usually follows within 7 to 10 days
of infection (158, 160, 164, 170) and occurs due to an
increase in brain pressure and subsequent herniation, which
leads to cardiopulmonary arrest and pulmonary edema (158,
172). Because death occurs so quickly, the exact immune
response is not easy to detect (162). Resistance to infection
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-13
may be attributed to the host innate immunity rather than
acquired immunity, as seen by the activity of complement,
neutrophils, and macrophages (170). There is some evidence
that the antibody response generated is IgM. Over 80% of
patients hospitalized with PAM have some IgG, but mostly
IgM (162).
Early diagnosis and treatment is critical to successful treat-
ment of PAM; however, most cases are diagnosed postmortem
due to the rapid progression of the disease. Although there is
no standard treatment, effective treatments for PAM can be
administered early if the patient is known to have engaged
in a recreational water activity and exhibits the symptoms
of an infection. The similarity of the early symptoms to the
common cold makes the disease difficult to diagnose and treat
in a rapid manner. Another issue is that it is often mistaken for
bacterial meningitis and thus is not treated properly. As such,
there have been very few known survivors (162). One survi-
vor was treated with a combination of intravenous and intra-
thecal amphotericin B and miconazole and oral rifampin
(173), and he remained healthy up to a four-year follow-up
(174). Other patients successfully treated have shown signs
of physical or cognitive impairment despite surviving the ill-
ness (162). The drug of choice for treatment of PAM is
amphotericin B (168), and the combination of amphotericin
B and miconazole are thought to have a synergistic effect
against the amoeba as demonstrated by laboratory studies
(175). A comparison of amphotericin B, miltefosine, and
chloropromazine determined that chloropromazine was the
most effective against N. fowleri. This could be a possible
effective treatment in humans since it resulted in a 75% sur-
vival rate in mice infected with PAM (176). In the summer of
FIGURE 9 Number of case reports of primary amoebic meningoencephalitis by state of exposure: United States, 1962–2012. Reprinted from
(195). doi:10.1128/9781555818821.ch3.1.6.f9
2013, a 12-year-old girl contracted PAM and was treated with
a combination of antifungals, antibiotics, and the investiga-
tional drug miltefosine within 36 h of the onset of symptoms.
In addition, her body temperature was lowered to below nor-
mal to help with brain swelling. This girl survived with no
apparent adverse cognitive effects (177).
Occurrence of Naegleria fowleri in Water
The occurrence is worldwide (178) and the organism can be
found in warm, fresh, or brackish water, including swimming
pools (163), hot tubs (179), domestic water supplies (180,
181) and premise plumbing (181–183), well water (181,
184), ponds, lakes, streams, rivers, natural hot springs (185–
188), and thermally polluted run-off from industrial zones
(189–193). Water sources do not have to be contaminated
to contain N. fowleri (194) since it is a water-based, free-
living organism.
N. fowleri has been found primarily in surface waters in the
Southern tier of the United States (195) (Fig. 9). It is thermo-
philic, preferring temperatures of 30°C and higher (162); sur-
face waters in warm regions are therefore believed to be an
ideal habitat for the amoeba (196). In the United States,
the occurrence has been documented in surface waters in
Texas, Virginia, Oklahoma, Florida, New Mexico, Arizona,
California, Pennsylvania, and Georgia, among other states
(181, 185, 186, 197–200). Most of these waters were either
naturally warmed or thermally polluted (185, 198, 201).
There have been over 310 cases of PAM documented
worldwide (202). The majority of cases have occurred in
countries with tropical and subtropical climates (196, 199,
203–206). The CDC documented 128 cases of PAM in the
3.1.6-14 ▪ WATER
United States between 1962 and 2012 (195) (Fig. 9).
Approximately 50% of these cases occurred in Texas and
Florida (30 and 29, respectively) (207). Recent cases have
been reported in Minnesota and Kansas (169, 208, 209).
These were the first documented cases in these states and
are significant because these locations were not previously
considered to be areas at risk for N. fowleri infection.
Of the 128 documented PAM cases in the United States as
of 2012, only 27 have been in women (195). It is thought that
young men are more likely to engage in recreational activities
such as water sports, or other activities that involve submerg-
ing the head under water. A few PAM cases may have resulted
from nasal exposure of tap waters via bathing in bathtubs,
pools, or hot tubs. The only cases associated with drinking
water have occurred in Australia, Pakistan, Arizona, Louisi-
ana, and the U.S. Virgin Islands (165, 181, 210–213). The
two cases that occurred in 2002 in Arizona were in young
children who had been exposed to the same water supply
(210). These cases were eventually linked to the presence
of N. fowleri in the drinking water (obtained from ground-
water) (181) from a well that was not routinely chlorinated.
Recently, two cases in adults in Louisiana were attributed to
chloraminated drinking water being used for nasal rinses via
neti pots (213). Another case of PAM was acquired through
nasal rinsing in the U.S. Virgin Islands in 2013 (212) and
recent cases in Pakistan (also primarily in young adults)
have also been linked to ritual cleansing that involve taking
water into the nose (211). In 2013, a young boy in Louisiana
became ill and subsequently died after playing on a outdoor
water slide hooked up to a hose with water from one of the
chloraminated systems previously linked to a neti pot death
(214).
Detection of Naegleria fowleri in Water
Cultural methods are often used to differentiate N. fowleri
from other free-living amoebae. The organism is concen-
trated from water then added to non-nutrient agar (NNA)
in petri dishes inoculated with a bacterial lawn and incubated
at 42°C to 45°C (189). Culture results are often confirmed by
more specific secondary tests. A summary of the most com-
monly used methods to concentrate and detect N. fowleri
from water is presented in Table 9.
Concentration of Water Samples for the
Detection of N. fowleri
Though there are currently no standard protocols, there are a
variety of detection methods that have been developed to
detect N. fowleri in water and the environment. The concen-
tration of large volumes is possible using the Envirocheck HV
Sampling Capsule (Pall Corp., East Hills, NY), which was
developed for the concentration of Cryptosporidium and Giar-
dia. Large volumes have also been collected using Micro-
Wynd filters (Cuno, Meriden, CT). For instance, following
two cases of PAM in children being linked to bathing in
untreated tap water (with a groundwater source), 16 gal of
water from the household bath tubs were filtered through a
1 µm pore size polypropylene Micro-Wynd II filter (grade
Y) at a flow rate of 16 gal/min. The same water was recycled
through the filter an additional two times. All of these sam-
ples tested positive for the presence of N. fowleri DNA by
PCR (181).
Smaller water volumes may also be concentrated by mem-
brane filtration (e.g., membrane with a pore size of 0.4–2 µm)
(188, 192, 215, 216). The membrane filter is then placed in
Page’s amoeba saline and vortexed thoroughly to recover
the organism from the filter. Painter et al. (123, 216) reported
polycarbonate filters (47 mm diameter and 0.4 µm pore size)
for concentrating 40 ml lake water samples. The nucleic acids
are extracted directly from the filter using a silica extraction
technique with bead beating, and the extracts were quantified
for the presence of N. fowleri DNA in a subsequent qPCR
assay. Filter membrane with different pore size such as 2 µm
have also been reported to concentrate N. fowleri from 1-liter
grab samples from various recreational surface waters in Ari-
zona. The filter concentrates can then be used in viable amoe-
bae and nested PCR assays (188).
Small water volumes may also be concentrated by centri-
fugation. Following centrifugation, the pelleted cells are
resuspended in a small volume (e.g., 1–5 ml) of Page’s
amoeba saline (181, 184). If there is no visible pellet, the cen-
trifugal supernatant is carefully removed via pipetting and dis-
carded until only 1–5 ml remains. In one study, 1-liter well
water samples were first concentrated via centrifugation
(4,640 × g for 15 min) with resuspension of the pellet in 5 ml
Page’s saline. The samples were then further concentrated
via membrane filtration using polyethylene filters with a
pore size of 2 µm with resuspension in 1 ml Page’s saline
(184). Pernin et al. (215) found that filtration (1.2 mm
pore size cellulose acetate filters) and centrifugation
(1,000 × g for 15 min for 10-ml volumes; 3,000 × g for 10 min
for 100-ml volumes) worked equally well for Naegleria
cysts (53% ± 21% versus 57% ± 25%, respectively), but
centrifugation worked significantly better for trophozoites
(5% ± 5% for filtration versus 22% ± 5% for centrifugation;
P < 0.001). More such comparisons are needed to determine
the efficiencies of various methods for the concentration of
N. fowleri in water.
Recently, an IMS procedure for capturing and concentrat-
ing N. fowleri from small-volume water samples using
N. fowleri–specific antibodies bound to magnetic beads was
developed (231). This procedure was reported to have an
overall recovery efficiency of 75 ± 17.7%. N. fowleri tropho-
zoites were recovered more efficiently than cysts (88 ± 8.4%
versus 61 ± 13.2%, respectively; P < 0.0001). There were no
significant differences between the recovery efficiencies for
four different N. fowleri genotypes; however, there was a small
amount of nonspecific binding to other free-living amoebae
including other Naegleria species (0.6–14% recovery). To
concentrate N. fowleri from 1-liter samples, the samples
were first concentrated via centrifugation at 1,500 × g for
15 min followed by resuspension of the pellet in 2.5 ml saline.
A volume of 1.5 ml was further concentrated using the IMS
procedure down to a volume of 100 µl (231).
N. fowleri Viability Assay
The concentrated water sample can then be used to inoculate
NNA containing either heat-killed Escherichia coli (200, 236)
or lawns of live E. coli (186, 188) to determine amoebae via-
bility. Other species have been used, but N. fowleri has dem-
onstrated preference for E. coli over other bacteria (237).
Direct plating of water samples onto NNA with E. coli
(NNA-EC) without any prior concentration is also some-
times performed. Plates for N. fowleri culture can be incubated
at 37°C or up to 44°C to inhibit the growth of other amoebae
(167, 188, 200). Amoebic plaques (clearings in the bacterial
lawns caused by amoebic predation) are then observed, and
the amoebae are harvested from each of the plaques from
the advancing edge using a sterile cell scraper and examined
for the transformation of amoebae to flagellates under a light
microscope (181, 188). The identity of viable N. fowleri is
typically confirmed via PCR.

TABLE 9 Summary of concentration and detection methods for Naegleria fowleri
Source (reference)
Thermally polluted lake waters
(185)
Spiked river samples (215),
industrial cooling waters (192),
surface waters (188, 216)
Drinking water (181), freshwater
ponds (200), groundwaters
(184)
Spiked river water samples (215)
Freshwater ponds (186, 200),
surface waters (188)
Spiked river waters (215),
industrial cooling waters (192)
Industrial cooling waters (191)
Spiked river water (217)
Pure cultures (192, 218–221)
Pure cultures (222, 223)
Artificially heated river water
(224)
Drinking water (181)
Industrial cooling waters (192)
Industrial cooling waters (190),
hot springs (187)
Pure cultures and spiked tap, river,
and lake water (225)
Pure cultures (226)
Pure cultures (227)
3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-15
Method(s)
Sand column filtration, Indirect fluorescent
Antibody-microscopy (IFA), Mouse
inoculation
Membrane filtration
Centrifugation
Centrifugation vs. membrane filtration
Viable amoebae assay followed by light
microscopy
Most probable number (MPN)
Immunofluorescent assay combined with flow
cytometry
Fountain flow cytometry
Isoelectric focusing, isoenzyme analyses
Restriction fragment length polymorphism
(RFLP) analysis
Enzyme-linked immunosorbent assay
(ELISA)
Viable amoebae assay, flagellation test,
Western immunoblot, nested PCR
ELISA, isoenzyme electrophoresis, PCR
Multiplex PCR, RFLP or
sequencing analysis
Nested PCR
PCR
Flagellation test, ELISA, PCR
Observations
• Large volumes filtered (50–100 gal).
• Some cross-reactivity between pathogenic and
nonpathogenic strains for IFA.
• Mouse inoculation used to determine
pathogenicity.
• Variable membrane pore sizes used. Can be used
with volumes up to a few liters.
• Only useful for small volumes.
• Centrifugation: 53 ± 21% recovery efficiency for
cysts, 22 ± 5% for trophozoites.
• Filtration: 57 ± 25% recovery efficiency for
cysts, 5 ± 5% for trophozoites.
• Incubation at 37°C–44°C (higher temperatures
will inhibit most other free-living amoebae
species).
• Requires a confirmation step (e.g., PCR).
• Modification of the viable amoebae assay.
• Requires a concentration step for
volumes > 1 ml.
• Requires a confirmation step (e.g., PCR).
• Detection limit: 200 cells per liter.
• Detection limit: 0.6 amoebae per ml at a flow
rate of 15 ml per minute.
• Method performed with N. lovaniensis.
• Requires pure cultures.
• Labor and time intensive.
• N. fowleri found to have more homogeneous
isoenzyme profiles than other Naegleria species.
• Labor and time intensive.
• Can be used to differentiate between Naegleria
species.
• Detection limit: 2,000 N. fowleri per ml.
• 97% specificity, 97.4% sensitivity.
• Detection limit for PCR: 5 amoebae in 50 ml.
• ELISA: Similar results to the PCR.
• Isoenzyme electrophoresis: Similar results to
ELISA and PCR, but more labor intensive and
requires pure cultures.
• Detection limit: 5 intact trophozoites or cysts.
• Detection limit: 5 pg N. fowleri DNA or 5 intact
N. fowleri.
• More sensitive than PCR (detection limit of 0.5
ng).
• Cannot determine organism viability.
• Detection limit: 1 pg of N. fowleri DNA (∼6
trophozoites).
• Flagellation tests had 21% false-negative rate
when compared to ELISA and PCR tests.
(Continued on next page)
3.1.6-16 ▪ WATER

TABLE 9 Summary of concentration and detection methods for Naegleria fowleri (Continued )
Source (reference)
Industrial cooling waters (193)
Industrial cooling waters (228)
Mouse brain tissue (229)
Spiked cerebral spinal fluid (230)
Spiked lake water (231)
Spiked swimming pool samples
(232)
Pure cultures (233)
Pure cultures (234)
Drinking water (235)
Method(s) Observations
Flagellation test vs. nested one-step PCR • Detection limit: Between >1 pg to 10 pg of
N. fowleri DNA (>5 to 50 cells).
• Culture/flagellation tests were all negative,
whereas the nested one-step PCR detected
N. fowleri in 18 of 119 samples.
Duplex qPCR vs. MPN • qPCR detection limit: 3 N. fowleri.
• Whole method detection limit: 65 cells per liter.
• qPCR: 100% specificity.
• Primers based on previously described sequence
(225).
• MPN method: detected N. fowleri in 17% of
samples that tested negative by qPCR.
qPCR • Detection limit: 1 trophozoite.
• Genus specific, not species specific. Therefore
useful for clinical samples, but not
environmental samples.
Multiplex qPCR • Detection limit: 1 amoeba.
• Does not detect other Naegleria species.
Immunomagnetic separation (IMS) • IMS: 61 ± 13.2% recovery efficiency for cysts,
and qPCR 88 ± 8.4% for trophozoites; 0.6–14% binding of
other Naegleria species.
• IMS can only be used with small
sample volumes.
• Detection limit for combined method: 14
amoebae per liter.
qPCR-melt-curve analysis • Detection limit: 1 copy of Mp2Cl5 DNA
sequence.
• Primers based on previously described sequence
(225), but more rapid than original PCR
method.
qPCR-melt-curve analysis • Detection limit: 0.1 cells due to the high copy
number (∼4,000 per cell) of the target gene in
Naegleria species.
qPCR-melt-curve analysis • Detection limit: 0.1–0.2 cells.
• qPCR is not specific to N. fowleri so requires
pure culture.
Viable amoebae assay vs. • Amoebae assay detection limit: 1–10 cells in
qPCR-melt-curve analysis biofilms; 10 cells in water.
• qPCR-melt-curve detection limit: 1–5 cells in
biofilms; 5–10 cells in water.
• qPCR is specific to N. fowleri. Does not require
culture of the sample.

This viability assay can be modified to incorporate sev-
eral dilutions of a sample to generate a most probable num-
ber (MPN) to estimate the number of viable N. fowleri
present. For instance, one group (192) used five replicates
of 10-fold dilutions of water samples ranging from 100 ml
to 0.1 ml. The 100 ml and 10 ml samples were concen-
trated via membrane filtration and the filter was placed
onto NNA-EC plates. The 1 ml and 0.1 ml volumes were
added directly to the NNA-EC plates. The 20 plates
were incubated at 44°C for 3–10 days and examined for
flagellates under a microscope. N. fowleri identification
was confirmed via enzyme-linked immunosorbent assay
(ELISA) and PCR methods. The numbers of positive rep-
licate plates per dilution were used to calculate the MPN of
N. fowleri in the sample. Similarly, in another study (215),
researchers used 10 replicates of 100 ml and 10 ml samples,
which were concentrated via membrane filtration and the
filter placed onto NNA-EC plates for incubation at 42°C
for 8–9 days.
Serological Assays to Detect N. fowleri
Several studies have employed indirect fluorescent antibody
microscopy for the identification of N. fowleri. In one such
study, some cross-reactivity was observed between pathogenic
and nonpathogenic Naegleria; however, each had their own
specific set of antigens as well (185). ELISA methods for
the specific identification of N. fowleri for water samples are
reliable but time consuming. One ELISA method for
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-17
N. fowleri was reported to have a specificity of 97% and a sen-
sitivity of 97.4%, with a detection limit of 2,000 N. fowleri per
ml (224). An ELISA specific for N. fowleri antigens (called
Nf ELISA) is commercially available through Indicia Diag-
nostic (Oullins, France). This test yielded similar results to
conventional PCR, with 578 of 825 thermal water samples
testing positive for N. fowleri versus 574 testing positive by
PCR (192).
N. fowleri can also be identified in samples using monoclo-
nal or polyclonal antibodies. Polyclonal antibodies have been
developed but exhibit cross-reactivity with other Naegleria
species (238). For instance, a Western immunoblot analysis
using polyclonal antibodies was used to detect N. fowleri in
one study. Although this method was successful in some sam-
ples, the method was unable to distinguish between among
Naegleria species. The results were therefore confirmed using
a monoclonal antibody (5D12) specific to N. fowleri (181). A
solid phase cytometry method combined with an immuno-
fluorescent assay was developed to detect N. fowleri in water.
This method used a monoclonal antibody that recognized
both cysts and trophozoites. The method was rapid (results
within 3 h) and had a detection limit of 200 cells per liter
of water (191).
Molecular Detection Methods for N. fowleri
Enzyme-Based Assays
Isoenzyme analyses have been developed to identify N. fowleri
in environmental samples including soil and water. The
enzyme aspartate aminotransferase was used in cellulose ace-
tate zymograms to distinguish between N. lovaniensis and
N. fowleri strains isolated from hot springs following a case
of PAM in the United Kingdom (218). Agarose isoelectric
focusing with the isoenzyme patterns have been used to study
Naegleria isolates (217, 220). N. fowleri was found to be the
most homogeneous of the Naegleria species. Application of
this technique for pathogenic and nonpathogenic Naegleria
species have resulted in delineating two primary phylogenetic
groups in the genus (219).
RFLP analyses have also been used to distinguish N. fowl-
eri from other Naegleria species and other free-living amoebae
(222). Application of RFLP analysis, isoenzyme analyses, and
monoclonal antibody for typing of Naegleria species yield
well-correlated results (223). Isoenzyme and RFLP analyses
usually require strains to be adapted to axenic culture to
obtain pure cultures for the analyses (226). As such, these
analyses are both time and labor intensive.
Conventional PCR/Nested PCR
PCR has provided an effective alternative for the specific
identification of N. fowleri that is both rapid and sensitive
for environmental samples. Some of the PCR assays devel-
oped are able to detect one N. fowleri cell (226, 239). The
only drawback to some of these methods is that they often
require the amoebae to be cultured prior to performing the
assays.
The sensitivity of conventional PCR detection can be
increased by using nested PCR to detect as few as five
N. fowleri cells (225). This assay has been used for the iden-
tification of N. fowleri from river water, soil (200), and domes-
tic water implicated in two deaths attributed to PAM (181).
Quantitative PCR.
Several qPCR assays have been developed for the detection
of N. fowleri in recent years (227–230, 231–234). Most
qPCR method are based on detection of the ribosomal
small unit (18S) rRNA gene, with SYBR green 1 labeling.
Other genes that have been used include Mp2C15 gene
and this assay is rapid and more sensitive with a detection
limit of one copy of the plasmid Mp2Cl5 DNA (225).
Mp2C15 gene has also been used to develop a duplex
qPCR assay for the quantitative detection of N. fowleri
in water samples with a detection limit of three cells
(228). Other popular targets for developing qPCR-based
methods include 18S rRNA, 5.8S rRNA, and internal
transcribed spacers 1 and 2 (ITS1 and ITS2) (230, 233,
234). TaqMan technology–based multiplex qPCR assay
has been reported for identification of different genera of
free-living amoebae (N. fowleri, Balamuthia mandrillaris,
and Acanthamoeba spp.) (230). A combined method using
IMS for the concentration of the sample followed by qPCR
was designed to analyze water samples with a detection
limit of 14 amoebae per liter and a 46% recovery efficiency
(231). For these types of analyses, it is necessary to obtain a
pure culture prior to the analysis (235). Nonpathogenic
Naegleria spp. can outgrow N. fowleri on NNA-EC plates,
leading to false-negative results (235). However, this can
be addressed by developing a direct total DNA extraction
procedure followed by N. fowleri-specific qPCR (to the
ITS region) melt curve analysis (235). This method had
a lower detection limit than the viability assay using
NNA-EC plates.
Genotyping
Genotyping of has been performed, sequencing the mito-
chondrial small subunit (5.8S) ribosomal RNA gene and
ITS1 and ITS2 (208, 240). There are six known genotypes
of based on the number of repeats of motifs 1 and 2 in ITS
1 (240). Combined qPCR and melting curve analysis can
also potentially be used for genotyping (233, 234).
CONCLUSIONS
Future of Genotyping
Ideally, genotyping environmental samples, which may con-
tain mixed species/genotype, should be done at the single
(oo)cyst level. Even then, intraoocyst sequence variation
have been reported (241). In addition, there are many typing
methods and markers used for species, assemblage, and sub-
type identification for Cryptosporidium and Giardia. Although
genotyping environmental isolates using single locus
sequence typing has been useful, it does not provide the
most accurate or complete genotyping information. A multi-
locus sequence typing is used for more accurate and thorough
species, assemblage, and subtype identification and for phylo-
genetic analyses. A standardized, robust, and sensitive
method to unequivocally identify Cryptosporidium and Giar-
dia in the water has not yet been established. Recent progress
has been made toward narrowing the list of potential markers
and developing a standardized method for genotyping Cryp-
tosporidium (137, 242) and Giardia (243). Alternatively,
with recent advances in next-generation high-throughput
sequencing technologies, sequencing entire genomes is
becoming more feasible. It is likely that whole genome
sequencing of environmental samples will be the method of
choice in the near future. Recently, the C. meleagridis genome
was reported to be completely sequenced using the Illumina
MiSeq sequencing platform (244). Moreover, by combining
this technology with flow cytometry, laser capture microdis-
section, and microfluidics, single cell whole genome sequenc-
ing will also become more feasible and applicable for routine
3.1.6-18 ▪ WATER
monitoring of Cryptosporidium and Giardia species in drinking
water systems.
Cultural methods for the detection of N. fowleri are often
time and labor intensive. In addition, a subsequent more spe-
cific test is usually necessary for confirmation of the result
(e.g., ELISA or PCR). The major advantages of molecular
and serological methods for the detection of N. fowleri in
water samples are that they are rapid, sensitive, and specific.
However, for some assays, it is still necessary to first obtain a
pure culture before the analysis, increasing the amount of
time required. Many assays also require DNA extraction fur-
ther increasing the time needed to identify N. fowleri from
environmental samples. N. fowleri-specific ELISA methods
are an alternative to molecular methods such as PCR in loca-
tions where funds and access to expensive equipment and
facilities are limited.
Nevertheless, molecular and serological methods provide
attractive alternatives to culture/microscopy alone since there
needs to be greater discrimination between pathogenic and
nonpathogenic species of Naegleria and other free-living
amoebae species. The use of more efficient extraction meth-
ods in combination with PCR methods have been shown to
successfully in identifying N. fowleri from environmental
samples (228); however, water samples may contain very
low levels of N. fowleri that may be below the assay’s limit
of detection. Another limitation of PCR and ELISA methods
is that they are not able to determine N. fowleri viability or
infectivity when not performed in tandem with culture meth-
ods. Therefore, a combination of a viability assay with a more
specific molecular or serological assay still appears to be the
best approach for detecting N. fowleri in water. Although
some progress has been made in recent years, more research
is needed to develop reliable quantitative, sensitive, and spe-
cific methods for the detection of N. fowleri cysts and troph-
ozoites in water and other environmental samples (rather
than just clinical samples).
The U.S. Environmental Protection Agency through its Office of
Research and Development partially funded and managed the research
described here. It has been subjected to the agency’s administrative review
and approved for publication. Mention of trade names or commercial prod-
ucts does not constitute endorsement or recommendation for use.
REFERENCES
1. Chomel BB. 2008. Control and prevention of emerging
parasitic zoonoses. Int J Parasitol 38:1211–1217.
2. Taylor LH, Latham SM, Mark E. 2001. Risk factors for
human disease emergence. Philos Trans R Soc Lond B Biol
Sci 356:983–989.
3. Alum A, Rubino JR, Ijaz MK. 2010. The global war
against intestinal parasites—should we use a holistic
approach? Int J Infect Dis 14:e732–e738.
4. Baldursson S, Karanis P. 2011. Waterborne transmission
of protozoan parasites: review of worldwide outbreaks–an
update 2004–2010. Water Res 45:6603–6614.
5. Mircean V, Györke A, Cozma V. 2012. Prevalence and
risk factors of Giardia duodenalis in dogs from Romania.
Vet Parasitol 184:325–329.
6. Karanis P, Kourenti C, Smith H. 2007. Waterborne trans-
mission of protozoan parasites: a worldwide review of out-
breaks and lessons learnt. J Water Health 5:1–38.
7. Yoder JS, Beach MJ. 2010. Cryptosporidium surveillance and
risk factors in the United States. Exp Parasitol 124:31–39.
8. Tyzzer E. 1907. A sporozoan found in the peptic glands of
the common mouse. Exp Biol Med 5:12–13.
9. Panciera R, Thomassen R, Garner F. 1971. Cryptospori-
dial infection in a calf. Vet Pathol Online 8:479–484.
10. Meisel J, Perera D, Meligro C, Rubin C. 1976. Overwhelm-
ing watery diarrhea associated with a cryptosporidium in an
immunosuppressed patient. Gastroenterology 70:1156–1160.
11. Plutzer J, Ongerth J, Karanis P. 2010. Giardia taxonomy,
phylogeny and epidemiology: facts and open questions. Int J
Hyg Environ Health 213:321–333.
12. Xiao L, Fayer R. 2008. Molecular characterisation of spe-
cies and genotypes of Cryptosporidium and Giardia and
assessment of zoonotic transmission. Int J Parasitol 38:
1239–1255.
13. Elwin K, Hadfield SJ, Robinson G, Crouch ND, Chalm-
ers RM. 2012. Cryptosporidium viatorum n. sp.(Apicom-
plexa: Cryptosporidiidae) among travellers returning to
Great Britain from the Indian subcontinent, 2007–2011.
Int J Parasitol 42:675–682.
14. Ren X, Zhao J, Zhang L, Ning C, Jian F, Wang R, Lv C,
Wang Q, Arrowood MJ, Xiao L. 2012. Cryptosporidium
tyzzeri n. sp.(Apicomplexa: Cryptosporidiidae) in domestic
mice (Mus musculus). Exp Parasitol 130:274–281.
15. Nichols GL. 2008. Epidemiology, pp. 79–118. In Xiao RFL
(ed.), Cryptosporidium and Cryptosporidiosis, 2nd ed, CRC
Press and IWA Publishing, Boca Raton, FL.
16. Chalmers RM. 2012. Waterborne outbreaks of cryptospor-
idiosis. Ann dell’Istit super sanità 48:429–446.
17. Meyer E, Jarroll EL. 1980. Giardiasis. Am J Epidemiol 111:
1–12.
18. U.S. EPA. 1984. Criteria document on giardia. ICAIR,Life
Systems, Washington, DC.
19. Filice FP. 1952. Studies on the Cytology and Life History of a
Giardia from the Laboratory Rat. University of California
Press, Berkeley, CA.
20. Arrowood M. 1997. Diagnosis. In Fayer R (ed.), Crypto-
sporidium and Cryptosporidiosis, . CRC Press, New York, NY.
21. Adam RD. 2001. Biology of Giardia lamblia. Clin Microbiol
Rev 14:447–475.
22. Cavalier-Smith T. 2003. Protist phylogeny and the high-
level classification of Protozoa. Eur J Protistol 39:338–348.
23. Thompson R, Smith A. 2011. Zoonotic enteric protozoa.
Vet Parasitol 182:70–78.
24. Lasek-Nesselquist E, Welch DM, Sogin ML. 2010. The
identification of a new Giardia duodenalis assemblage in
marine vertebrates and a preliminary analysis of
G. duodenalis population biology in marine systems. Int J
Parasitol 40:1063–1074.
25. Yoshida N, Tyler KM, Llewellyn MS. 2011. Invasion
mechanisms among emerging food-borne protozoan para-
sites. Trends Parasitol 27:459–466.
26. Hunter PR, Hughes S, Woodhouse S, Nicholas R, Syed
Q, Chalmers RM, Verlander NQ, Goodacre J. 2004.
Health sequelae of human cryptosporidiosis in immuno-
competent patients. Clin Infect Dis 39:504–510.
27. Clifford C, Crook D, Conlon C, Fraise A, Day D, Peto T.
1990. Impact of waterborne outbreak of cryptosporidiosis
on AIDS and renal transplant patients. Lancet 335:
1455–1456.
28. U.S. EPA. 1997. Cryptosporidium and Giardia in water
by filtration/IMS/FA. Draft method 1622. EPA-821-R-97-
023. Office of Research and Development, U.S. Govern-
ment Printing Office, Washington, DC.
29. Petry F, Jakobi V, Tessema TS. 2010. Host immune
response to Cryptosporidium parvum infection. Exp Parasitol
126:304–309.
30. Wobus CE, Karst SM, Thackray LB, Chang K-O, Sos-
novtsev SV, Belliot G, Krug A, Mackenzie JM, Green
KY, Virgin HW IV. 2004. Replication of norovirus in
cell culture reveals a tropism for dendritic cells and macro-
phages. PLoS Biol 2:e432.
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-19
31. Ryu H, Abbaszadegan M. 2008. Long-term study of Cryp-
tosporidium and Giardia occurrence and quantitative micro-
bial risk assessment in surface waters of Arizona in the
USA. J Water Health 6:263–273.
32. Smith H, Rose J. 1990. Waterborne cryptosporidiosis. Par-
asitol Today 6:8–12.
33. Gold D, Smith HV. 2002. Pathogenic Protozoa and Drinking
Water, pp. 143–162. In Palumbo F, Ziglio G, Beken AV
(eds), Detection Methods for Algae, Protozoa and Helmints
in Fresh and Drinking Water. Wiley & Sons, Chichester,
West Sussex, England.
34. Wohlsen T, Katouli M. 2008. A review of the existing
methods for detection, enumeration and inactivation of
Cryptosporidium in surface waters. Journal of Water Supply:
Research and Technology-AQUA 57:65–77.
35. Quintero-Betancourt W, Peele ER, Rose JB. 2002. Cryp-
tosporidium parvum and Cyclospora cayetanensis: a review of
laboratory methods for detection of these waterborne para-
sites. J Microbiol Meth 49:209–224.
36. Wohlsen T, Bates J, Gray B, Katouli M. 2004. Evaluation
of five membrane filtration methods for recovery of Crypto-
sporidium and Giardia isolates from water samples. Appl
Environ Microbiol 70:2318–2322.
37. Zarlenga DS, Trout JM. 2004. Concentrating, purifying
and detecting waterborne parasites. Vet Parasitol 126:
195–217.
38. Wallis P, Erlandsen S, Isaac-Renton J, Olson M, Robert-
son W, Van Keulen H. 1996. Prevalence of Giardia cysts
and Cryptosporidium oocysts and characterization of Giardia
spp. isolated from drinking water in Canada. Appl Environ
Microbiol 62:2789–2797.
39. Koompapong K, Sutthikornchai C, Sukthana Y. 2009.
Cryptosporidium oocyst detection in water samples: floata-
tion technique enhanced with immunofluorescence is as
effective as immunomagnetic separation method. Korean
J Parasitol 47:353–357.
40. U.S. EPA. 2005. Method 1623: Cryptosporidium and Giar-
dia in water by filtration/IMS/FA. EPA 815-R-05-002.
Office of Research and Development, Government Print-
ing Office, Washington, DC.
41. U.S. EPA. 2012. Method 1623.1: Cryptosporidium
and Giardia in water by filtration/IMS/FA. EPA
816-R-12-001. Research and Development, Government
Printing Office, Washington, DC.
42. ISO. 2006. Water quality—isolation and identification of
Cryptosporidium oocysts and Giardia cysts from water. ISO
15553. ISO, Geneva, Switzerland.
43. Rose JB, Cifrino A, Madore M, Gerba CP, Sterling CR,
Arrowood MJ. 1986. Detection of Cryptosporidium from
wastewater and freshwater environments. Water Sci Technol
18:233–239.
44. Hsu B-m, Huang C, Hsu Y-F, Jiang G-Y. 2001. Evalua-
tion of two concentration methods for detecting Giardia
and Cryptosporidium in water. Water Res 35:419–424.
45. Shepherd K, Wyn-Jones A. 1995. Evaluation of different
filtration techniques for the concentration of Cryptospori-
dium oocysts from water. Water Sci Technol 31:425–429.
46. Shepherd K, Wyn-Jones A. 1996. An evaluation of meth-
ods for the simultaneous detection of Cryptosporidium
oocysts and Giardia cysts from water. Appl Environ Microbiol
62:1317–1322.
47. Hansen JS, Ongerth JE. 1991. Effects of time and water-
shed characteristics on the concentration of Cryptospori-
dium oocysts in river water. Appl Environ Microbiol 57:
2790–2795.
48. Falk CC, Karanis P, Schoenen D, Seitz HM. 1998. Bench
scale experiments for the evaluation of a membrane filtra-
tion method for the recovery efficiency of Giardia and Cryp-
tosporidium from water. Water Res 32:565–568.
49. Dawson D, Maddocks M, Roberts J, Vidler J. 1993.
Evaluation of recovery of Cryptosporidium parvum oocysts
using membrane filtration. Lett Appl Microbiol 17:
276–279.
50. Simmons OD, Sobsey MD, Heaney CD, Schaefer FW,
Francy DS. 2001. Concentration and detection of Crypto-
sporidium oocysts in surface water samples by method 1622
using ultrafiltration and capsule filtration. Appl Environ
Microbiol 67:1123–1127.
51. Digiorgio CL, Gonzalez DA, Huitt CC. 2002. Cryptospori-
dium and Giardia recoveries in natural waters by using Envi-
ronmental Protection Agency method 1623. Appl Environ
Microbiol 68:5952–5955.
52. Feng Y, Xiao L. 2011. Zoonotic potential and molecular
epidemiology of Giardia species and giardiasis. Clin Micro-
biol Rev 24:110–140.
53. Sartory D, Parton A, Parton A, Roberts J, Bergmann K.
1998. Recovery of Cryptosporidium oocysts from small and
large volume water samples using a compressed foam filter
system. Lett Appl Microbiol 27:318–322.
54. Hill VR, Polaczyk AL, Kahler AM, Cromeans TL, Hahn
D, Amburgey JE. 2009. Comparison of hollow-fiber ultra-
filtration to the USEPAVIRADEL technique and USEPA
method 1623. J Environ Qual 38:822–825.
55. Francy DS, Stelzer EA, Brady AM, Huitger C, Bushon
RN, Ip HS, Ware MW, Villegas EN, Gallardo V, Lind-
quist HD. 2013. Comparison of filters for concentrating
microbial indicators and pathogens in lake water samples.
Appl Environ Microbiol 79:1342–1352.
56. Liu P, Hill VR, Hahn D, Johnson TB, Pan Y, Jothikumar
N, Moe CL. 2012. Hollow-fiber ultrafiltration for simulta-
neous recovery of viruses, bacteria and parasites from
reclaimed water. J Microbiol Meth 88:155–161.
57. Rhodes ER, Hamilton DW, See MJ, Wymer L. 2011.
Evaluation of hollow-fiber ultrafiltration primary concen-
tration of pathogens and secondary concentration of
viruses from water. J Virol Meth 176:38–45.
58. Rhodes ER, Hamilton DW, See MJ, Wymer L. 2011.
Evaluation of hollow-fiber ultrafiltration primary concen-
tration of pathogens and secondary concentration of
viruses from water. J Virol Meth 176:38–45.
59. Hill VR, Polaczyk AL, Hahn D, Narayanan J, Cromeans
TL, Roberts JM, Amburgey JE. 2005. Development of a
rapid method for simultaneous recovery of diverse microbes
in drinking water by ultrafiltration with sodium poly-
phosphate and surfactants. Appl Environ Microbiol 71:
6878–6884.
60. Polaczyk AL, Narayanan J, Cromeans TL, Hahn D, Rob-
erts JM, Amburgey JE, Hill VR. 2008. Ultrafiltration-
based techniques for rapid and simultaneous concentration
of multiple microbe classes from 100-L tap water samples.
J Microbiol Meth73:92–99.
61. Ruecker N. 2013. Genotyping Cryptosporidium from
water to source track fecal contamination in agricultural
watersheds. Thesis in Agriculture Studies. University of
Calgary, Alberta Canada. http://theses.ucalgary.ca/bit
stream/11023/506/2/ucalgary_2013_ruecker_norma.pdf
62. Finch G, Daniels C, Black E, Schaefer F, Belosevic M.
1993. Dose response of Cryptosporidium parvum in outbred
neonatal CD-1 mice. Appl Environ Microbiol 59:3661–3665.
63. Korich D, Mead J, Madore M, Sinclair N, Sterling CR.
1990. Effects of ozone, chlorine dioxide, chlorine, and
monochloramine on Cryptosporidium parvum oocyst viabil-
ity. Appl Environ Microbiol 56:1423–1428.
64. Peeters JE, Mazas EA, Masschelein WJ, De Maturana
IVM, Debacker E. 1989. Effect of disinfection of drinking
water with ozone or chlorine dioxide on survival of Crypto-
sporidium parvum oocysts. Appl Environ Microbiol 55:
1519–1522.
3.1.6-20 ▪ WATER
65. Blewett D. 1988. Disinfection and oocysts, pp. 107–115. In
Angus KW, Blewett DA (eds), Cryptosporidiosis.
Edinburgh, United Kingdom, Animal Disease Research
Association.
66. Campbell A, Robertson L, Smith H. 1992. Viability of
Cryptosporidium parvum oocysts: correlation of in vitro
excystation with inclusion or exclusion of fluorogenic vital
dyes. Appl Environ Microbiol 58:3488–3493.
67. Belosevic M, Guy R, Taghi-Kilani R, Neumann N,
Gyürék L, Liyanage L, Millard P, Finch G. 1997. Nucleic
acid stains as indicators of Cryptosporidium parvum oocyst
viability. Int J Parasitol 27:787–798.
68. Bukhari Z, Marshall M, Korich D, Fricker C, Smith H,
Rosen J, Clancy J. 2000. Comparison of Cryptosporidium
parvum viability and infectivity assays following ozone
treatment of oocysts. Appl Environ Microbiol 66:
2972–2980.
69. Chauret CP, Radziminski CZ, Lepuil M, Creason R,
Andrews RC. 2001. Chlorine dioxide inactivation of
Cryptosporidium parvum oocysts and bacterial spore indica-
tors. Appl Environ Microbiol 67:2993–3001.
70. Kato S, Bowman DD. 2002. Using flow cytometry to deter-
mine the viability of Cryptosporidium parvum oocysts
extracted from spiked environmental samples in chambers.
Parasitol Res 88:326–331.
71. Clancy JL, Bukhari Z, McCuin R, Clancy TP, Marshall
MM, Korich DG, Fricker CR, Sykes N, Smith HV,
O’Grady JE, Sobrinho J, Schaefer FW. 2000. Cryptospori-
dium viability and infectivity methods. American Water
Works Association, Denver, CO.
72. O’Donoghue P. 1999. Isolation, propagation and charac-
terisation of Cryptosporidium. Int J Parasitol 29:1379–1413.
73. Neumann NF, Gyürek LL, Gammie L, Finch GR,
Belosevic M. 2000. Comparison of animal infectivity
and nucleic acid staining for assessment of Cryptosporidium
parvum viability in water. Appl Environ Microbiol 66:
406–412.
74. Brescia CC, Griffin SM, Ware MW, Varughese EA,
Egorov AI, Villegas EN. 2009. Cryptosporidium propidium
monoazide-PCR, a molecular biology-based technique for
genotyping of viable Cryptosporidium oocysts. Appl Environ
Microbiol 75:6856–6863.
75. Vesey G, Ashbolt N, Wallner G, Dorsch M, Williams
KL, Veal DA. 1995. Assesssing Cryptosporidium parvum
Oocyst Viability with Fluorescent In-Situe Hybridization Using
RIbosomal RNA Probes and Flow Cytometry. Royal Society
of Chemistry, Cambridge, UK.
76. Smith JJ, Gunasekera TS, Barardi CR, Veal D, Vesey G.
2004. Determination of Cryptosporidium parvum oocyst via-
bility by fluorescence in situ hybridization using a ribosomal
RNA-directed probe. J Appl Microbiol 96:409–417.
77. Abbaszadegan M, Huber MS, Gerba CP, Pepper IL. 1997.
Detection of viable Giardia cysts by amplification of
heat shock-induced mRNA. Appl Environ Microbiol 63:
324–328.
78. Jenkins MC, Trout J, Abrahamsen MS, Lancto CA,
Higgins J, Fayer R. 2000. Estimating viability of Crypto-
sporidium parvum oocysts using reverse transcriptase-poly-
merase chain reaction (RT-PCR) directed at mRNA
encoding amyloglucosidase. J Microbiol Meth 43:97–106.
79. Stinear T, Matusan A, Hines K, Sandery M. 1996. Detec-
tion of a single viable Cryptosporidium parvum oocyst in
environmental water concentrates by reverse transcrip-
tion-PCR. Appl Environ Microbiol 62:3385–3390.
80. Alum A, Rubino JR, Khalid Ijaz M. 2011. Comparison of
molecular markers for determining the viability and infec-
tivity of Cryptosporidium oocysts and validation of molecu-
lar methods against animal infectivity assay. Int J Infect Dis
15:e197–e200.
81. Alum A, Sbai B, Asaad H, Rubino JR, Khalid Ijaz M.
2012. ECC–RT-PCR: a new method to determine the via-
bility and infectivity of Giardia cysts. Int J Infect Dis 16:
e350–e353.
82. Lee S-U, Joung M, Ahn M-H, Huh S, Song H, Park W-Y,
Yu J-R. 2008. CP2 gene as a useful viability marker for
Cryptosporidium parvum. Parasitol Res 102:381–387.
83. Baeumner AJ, Humiston MC, Montagna RA, Durst RA.
2001. Detection of viable oocysts of Cryptosporidium par-
vum following nucleic acid sequence based amplification.
Anal Chem 73:1176–1180.
84. Current WL, Long PL. 1983. Development of human and
calf Cryptosporidium in chicken embryos. J Infect Dis 148:
1108–1113.
85. Current WL, Garcia LS. 1991. Cryptosporidiosis. Clin
Microbiol Rev 4:325.
86. Woodmansee DB, Pohlenz JFL. 1983. Development of
Cryptosporidium sp. in a human rectal tumor cell line, pp.
306–319. In Veterinary Infectious Disease Organization
(ed), Proceedings of the Fourth International Symposium
on Neonatal Diarrhea, October 3–5, 1983. Veterinary
Infectious Disease Organization, Saskatoon, Canada.
87. Woodmansee DB. 1986. Isolation, in vitro excystation,
and in vitro development of Cryptosporidium sp. from
calves. PhD dissertation. Iowa State University.
88. Hijjawi NS. 2003. In vitro cultivation and development of
Cryptosporidium in cell culture, pp. 233–253. In Thompson
RCA, Armson A, Ryan UM (eds), Cryptosporidium from
Molecules to Disease. Division of Veterinary and Biomedical
Sciences, Murdoch University, Murdoch, WA, Australia.
89. Buraud M, Forget E, Favennec L, Bizet J, Gobert J-G,
Deluol A-M. 1991. Sexual stage development of crypto-
sporidia in the Caco-2 cell line. Infec Immu 59:4610–4613.
90. Rasmussen KR, Larsen NC, Healey MC. 1993. Complete
development of Cryptosporidium parvum in a human endo-
metrial carcinoma cell line. Infect Immun 61:1482–1485.
91. Eggleston MT, Tilley M, Upton SJ. 1994. Enhanced
development of Cryptosporidium parvum in vitro by removal
of oocyst toxins from infected cell monolayers. J Helminthol
Soc Wash 61:122–125.
92. Upton SJ. 1997. In vitro cultivation, pp. 181–207. In Fayer
R (ed.), Cryptosporidiosis of Humans and Animals. CRC
Press, Boca Raton, FL.
93. Hijjawi N, Meloni B, Morgan U, Thompson R. 2001.
Complete development and long-term maintenance of
Cryptosporidium parvum human and cattle genotypes in
cell culture. Int J Parasitol 31:1048–1055.
94. Hijjawi N, Meloni B, Ryan U, Olson M, Thompson R.
2002. Successful in vitro cultivation of Cryptosporidium
andersoni: evidence for the existence of novel extracellular
stages in the life cycle and implications for the classification
of Cryptosporidium. Int J Parasitol 32:1719–1726.
95. Lacharme L, Villar V, Rojo-Vazquez FA, Suárez S.
2004. Complete development of Cryptosporidium parvum
in rabbit chondrocytes (VELI cells). Microbes Infect 6:
566–571.
96. Upton SJ, Tilley M, Brillhart DB. 1994. Comparative
development of Cryptosporidium parvum (Apicomplexa)
in 11 continuous host cell lines. FEMS Microbiol Lett
118:233–236.
97. Rochelle PA, Ferguson DM, Handojo TJ, De Leon R,
Stewart MH, Wolfe RL. 1997. An assay combining cell
culture with reverse transcriptase PCR to detect and deter-
mine the infectivity of waterborne Cryptosporidium parvum.
Appl Environ Microbiol 63:2029–2037.
98. Hijjawi N, Meloni B, Ng’anzo M, Ryan U, Olson M, Cox
P, Monis P, Thompson R. 2004. Complete development
of Cryptosporidium parvum in host cell-free culture. Int J Par-
asitol 34:769–777.
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-21
99. Joachim A, Eckert E, Petry F, Bialek R, Daugschies
A. 2003. Comparison of viability assays for Cryptospori-
dium parvum oocysts after disinfection. Vet Parasitol
111:47–57.
100. Zhang H, Guo F, Zhu G. 2015. Cryptosporidium Lactate
Dehydrogenase Is Associated with the Parasitophorous
Vacuole Membrane and Is a Potential Target for Develop-
ing Therapeutics. PLoS Pathog 11(11):e1005250.
101. Stroup SE, Roy S, Mchele J, Maro V, Ntabaguzi S, Siddi-
que A, Kang G, Guerrant RL, Kirkpatrick BD, Fayer R.
2006. Real-time PCR detection and speciation of Crypto-
sporidium infection using Scorpion probes. J Med Microbiol
55:1217–1222.
102. Favennec L. 1997. Physiopathologic and therapeutic stud-
ies in in vitro and in vivo models of Cryptosporidium parvum
infection. J Eukaryot Microbiol 44:69s–70s.
103. Bonnin A, Salimbeni I, Dubremetz J, Harly G, Chavanet
P, Camerlynck P. 1990. Mise au point d’un modèle expér-
imental de culture in vitro des stades asexués de Cryptospori-
dium sp. Annal parasitol hum comp 65:41–43.
104. Lawton P, Naciri M, Mancassola R, Petavy AF. 1997. In
vitro cultivation of Cryptosporidium parvum in the non-
adherent human monocytic THP-1 cell line. J Eukaryot
Microbiol 44:66s–66s.
105. Wiest PM, Johnson JH, Flanigan TP. 1993. Microtubule
inhibitors block Cryptosporidium parvum infection of a
human enterocyte cell line. Infect Immu 61:4888–4890.
106. Sifuentes LY, Di Giovanni GD. 2007. Aged HCT-8 cell
monolayers support Cryptosporidium parvum infection.
Appl Environ Microbiol 73:7548–7551.
107. Di Giovanni GD, Lechevallier MW. 2005. Quantitative-
PCR assessment of Cryptosporidium parvum cell culture
infection. Appl Environ Microbiol 71:1495–1500.
108. Garvey M, Farrell H, Cormican M, Rowan N. 2010.
Investigations of the relationship between use of
in vitro cell culture-quantitative PCR and a mouse-based
bioassay for evaluating critical factors affecting the disinfec-
tion performance of pulsed UV light for treating Crypto-
sporidium parvum oocysts in saline. J Microbiol Meth 80:
267–273.
109. Johnson AM, Di Giovanni GD, Rochelle PA. 2012. Com-
parison of assays for sensitive and reproducible detection of
cell culture-infectious Cryptosporidium parvum and Crypto-
sporidium hominis in drinking water. Appl Environ Microbiol
78:156–162.
110. Belosevic M, Faubert G, Maclean J, Law C, Croll N.
1983. Giardia lamblia infections in Mongolian gerbils: an
animal model. J Infect Dis 147:222–226.
111. Hill DR, Guerrant RL, Pearson RD, Hewlett EL. 1983.
Giardia lamblia infection of suckling mice. J Infect Dis
147:217–221.
112. Meyer EA. 1976. Giardia lamblia: isolation and axenic
cultivation. Exp Parasitol 39:101–105.
113. Erlandsen SL, Bemrick WJ, Wells CL, Feely DE,
Knudson L, Campbell SR, Van Keulen H, Jarroll EL.
1990. Axenic culture and characterization of Giardia
ardeae from the great blue heron (Ardea herodias).
J Parasitol:717–724.
114. Schupp DG, Erlandsen SL. 1987. A new method to deter-
mine Giardia cyst viability: correlation of fluorescein diace-
tate and propidium iodide staining with animal infectivity.
Appl Environ Microbiol 53:704–707.
115. U.S. EPA. 2006. Long term 2 enhanced surface water treat-
ment rule. 40 CFR Parts 9, 141, and 142. Fed Reg71. Reg-
ulations NPDW.
116. Yang W, Chen P, Villegas EN, Landy RB, Kanetsky C,
Cama V, Dearen T, Schultz CL, Orndorff KG, Prelewicz
GJ. 2008. Cryptosporidium source tracking in the Potomac
River watershed. Appl Environ Microbiol 74:6495–6504.
117. Ruecker NJ, Matsune JC, Wilkes G, Lapen DR, Topp E,
Edge TA, Sensen CW, Xiao L, Neumann NF. 2012.
Molecular and phylogenetic approaches for assessing sour-
ces of Cryptosporidium contamination in water. Water Res
46:5135–5150.
118. Zheng G-X, Zhang X-M, Yang Y-S, Zeng S-R, Wei J-F,
Wang Y-H, Li Y-J. 2014. An integrated microfludic device
for culturing and screening of Giardia lamblia. Exp Parasitol
137:1–7.
119. Insulander M, Silverlås C, Lebbad M, Karlsson L, Matts-
son J, Svenungsson B. 2013. Molecular epidemiology and
clinical manifestations of human cryptosporidiosis in Swe-
den. Epidemiol Infect 141:1009–1020.
120. Li N, Xiao L, Alderisio K, Elwin K, Cebelinski E, Chalm-
ers R, Santin M, Fayer R, Kvac M, Ryan U. 2014. Subtyp-
ing Cryptosporidium ubiquitum, a zoonotic pathogen
emerging in humans. Emerg Infect Dis 20:217.
121. Fournet N, Deege MP, Urbanus AT, Nichols G, Rosner
BM, Chalmers RM, Gorton R, Pollock KG, van der
Giessen JW, Wever PC, Dorigo-Zetsma JW, Mulder B,
Mank TG, Overdevest I, Kusters JG, van Pelt W,
Kortbeek LM. 2013. Simultaneous increase of Cryptospori-
dium infections in the Netherlands, the United Kingdom
and Germany in late summer season, 2012. Euro Surveill
18: http://www.eurosurveillance.org/ViewArticle.aspx?
ArticleId=20348.
122. Sulaiman IM, Morgan UM, Thompson RA, Lal AA, Xiao
L. 2000. Phylogenetic relationships of Cryptosporidium par-
asites based on the 70-kilodalton heat shock protein
(HSP70) gene. Appl Environ Microbiol 66:2385–2391.
123. Morgan UM, Monis PT, Xiao L, Limor J, Sulaiman I,
Raidal S, O’donoghue P, Gasser R, Murray A, Fayer R.
2001. Molecular and phylogenetic characterisation of
Cryptosporidium from birds. Int J Parasitol 31:289–296.
124. Pedraza-Díaz S, Amar C, Nichols GL, Mclauchlin J.
2001. Nested polymerase chain reaction for amplification
of the Cryptosporidium oocyst wall protein gene. Emerg
Infect Dis 7:49.
125. Spano F, Putignani L, Mclauchlin J, Casemore DP, Cri-
santi A. 1997. PCR-RFLP analysis of the Cryptosporidium
oocyst wall protein (COWP) gene discriminates between
C. wrairi and C. parvum, and between C. parvum isolates
of human and animal origin. FEMS Microbiol Lett 150:
209–217.
126. Homan W, Van Gorkom T, Kan YY, Hepener J. 1999.
Characterization of Cryptosporidium parvum in human
and animal feces by single-tube nested polymerase
chain reaction and restriction analysis. Parasitol Res 85:
707–712.
127. Peng MM, Matos O, Gatei W, Das P, Stantic-Pavlinic M,
Bern C, Sulaiman IM, Glaberman S, Lal AA, Xiao L.
2001. A comparison of Cryptosporidium subgenotypes
from several geographic regions. J Eukaryot Microbiol 48:
28s–31s.
128. Alves M, Xiao L, Sulaiman I, Lal AA, Matos O, Antunes
F. 2003. Subgenotype analysis of Cryptosporidium isolates
from humans, cattle, and zoo ruminants in Portugal. J
Clin Microbiol 41:2744–2747.
129. Glaberman S, Sulaiman IM, Bern C, Levior J, Peng MM,
Morgan U, Gilman R, Lal AA, Xiao L. 2001. A multilocus
genotypic analysis of Cryptosporidium meleagridis. J Eukaryot
Microbiol 48:19s–22s.
130. Cacciò SMDG, Marzia Pozio, Edoardo. 2002. Sequence
analysis of the β-giardin gene and development of a poly-
merase chain reaction–restriction fragment length poly-
morphism assay to genotype Giardia duodenalis cysts from
human faecal samples. Int J Parasitol 32:1023–1030.
131. Lalle M, Jimenez-Cardosa E, Cacciò SM, Pozio E. 2005.
Genotyping of Giardia duodenalis from humans and dogs
3.1.6-22 ▪ WATER
from Mexico using a β-giardin nested polymerase chain
reaction assay. J Parasitol 91:203–205.
132. Xiao L, Lal AA, Jiang J. 2004. Detection and differentia-
tion of Cryptosporidium oocysts in water by PCR-RFLP,
pp. 163–176. In Spencer JFT, Ragout AL (eds), Public
Health Microbiology. Humana Press Inc., Totowa, NJ.
133. Jiang J, Alderisio KA, Singh A, Xiao L. 2005. Develop-
ment of procedures for direct extraction of Cryptosporidium
DNA from water concentrates and for relief of PCR inhib-
itors. Appl Environ Microbiol 71:1135–1141.
134. Xiao L, Singh A, Limor J, Graczyk TK, Gradus S, Lal A.
2001. Molecular characterization of cryptosporidium oocysts
in samples of raw surface water and wastewater. Appl Envi-
ron Microbiol 67:1097–1101.
135. Gobet P, Toze S. 2001. Sensitive genotyping of
Cryptosporidium parvum by PCR-RFLP analysis of the
70-kilodalton heat shock protein (HSP70) gene. FEMS
Microbiol Lett 200:37–41.
136. Sturbaum GD, Klonicki PT, Marshall MM, Jost BH,
Clay BL, Sterling CR. 2002. Immunomagnetic separation
(IMS)-fluorescent antibody detection and IMS-PCR
detection of seeded Cryptosporidium parvum oocysts in nat-
ural waters and their limitations. Appl Environ Microbiol 68:
2991–2996.
137. Robinson G, Chalmers RM. 2012. Assessment of poly-
morphic genetic markers for multi-locus typing of Crypto-
sporidium parvum and Cryptosporidium hominis. Exp
Parasitol 132:200–215.
138. Le Blancq SM, Khramtsov NV, Zamani F, Upton SJ, Wu
TW. 1997. Ribosomal RNA gene organization in Crypto-
sporidium parvum. Mol Biochem Parasitol 90:463–478.
139. Khramtsov NV, Tilley M, Blunt DS, Montelone BA,
Upton SJ. 1995. Cloning and analysis of a Cryptosporidium
parvum gene encoding a protein with homology to cyto-
plasmic form Hsp70. J Eukaryot Microbiol 42:416–422.
140. CDC. 2014. CryptoNet: molecular-based tracking to better
understand U.S. cryptosporidiosis transmission. National
Center for Emerging and Zoonotic Infectious Diseases,
CDC, Atlanta, GA. http://www.cdc.gov/parasites/crypto/
pdf/cryptonet_fact_sheet508c.pdf. Accessed 12/22/2015
141. Xiao L. 2010. Molecular epidemiology of cryptosporidiosis:
an update. Exp Parasitol 124:80–89.
142. O’Connor RM, Wanyiri JW, Cevallos AM, Priest JW,
Ward HD. 2007. Cryptosporidium parvum glycoprotein
gp40 localizes to the sporozoite surface by association
with gp15. Mol Biochem Parasitol 156:80–83.
143. Robertson L, Hermansen L, Gjerde B, Strand E, Alvsvåg
J, Langeland N. 2006. Application of genotyping during an
extensive outbreak of waterborne giardiasis in Bergen, Nor-
way, during autumn and winter 2004. Appl Environ Micro-
biol 72:2212–2217.
144. Xiao L, Alderisio K, Singh A. 2006. Development and
standardization of a Cryptosporidium genotyping tool for water
samples, vol. 91101. American Water Works Association,
Denver, CO.
145. Ware MW, Keely SP, Villegas EN. 2013. Development
and evaluation of an off-the-slide genotyping technique
for identifying Giardia cysts and Cryptosporidium oocysts
directly from US EPA Method 1623 slides. J Appl Microbiol
115:298–309.
146. Di Giovanni G, Hoffman R, Sturbaum G. 2010. Crypto-
sporidium genotyping method for regulatory microscope
slides. Project 4099 Report. Water Research Foundation,
Denver, CO, 56.
147. U.S. EPA. 1992. Standards for the disposal of sewage sludge,
vol. R-93/003, pp. 9387–9404. EPA, Washington, DC.
148. Xiao L, Fayer R, Ryan U, Upton SJ. 2004. Cryptospori-
dium taxonomy: recent advances and implications for
public health. Clin Microbiol Rev 17:72–97.
149. Nichols RA, Campbell BM, Smith HV. 2006. Molec-
ular fingerprinting of Cryptosporidium oocysts isolated
during water monitoring. Appl Environ Microbiol 72:
5428–5435.
150. Nichols R, Connelly L, Sullivan C, Smith H. 2010. Iden-
tification of Cryptosporidium species and genotypes in Scot-
tish raw and drinking waters during a one-year monitoring
period. Appl Environ Microbiol 76:5977–5986.
151. Chalmers RM, Katzer F. 2013. Looking for Cryptospori-
dium: the application of advances in detection and diagno-
sis. Trends Parasitol 29:237–251.
152. Sopwith W, Osborn K, Chalmers R, Regan M. 2005. The
changing epidemiology of cryptosporidiosis in North West
England. Epidemiol Infect 133:785–793.
153. Karon A, Hanni K, Mohle-Boetani J, Beretti R, Hill V,
Arrowood M, Johnston S, Xiao L, Vugia D. 2011. Giardia-
sis outbreak at a camp after installation of a slow-sand
filtration water-treatment system. Epidemiol Infect 139:
713–717.
154. Daly E, Roy S, Blaney D, Manning J, Hill V, Xiao L, Stull
J. 2010. Outbreak of giardiasis associated with a community
drinking-water source. Epidemiol Infect 138:491–500.
155. Ma P VG, Martinez AJ, Theodore FH, Daggett PM, Saw-
yer TK. 1990. Naegleria and Acanthoamoeba infections. Rev
Infect Dis 12: 490–513.
156. De Jonckheere JF. 2004. Molecular definition and the
ubiquity of species in the genus Naegleria. Protist 155:
89–103.
157. Stevens A, De Jonckheere J, Willaert E. 1980. Naegleria
lovaniensis new species: isolation and identification of six
thermophilic strains of a new species found in association
with Naegleria fowleri. Int J Parasitol 10:51–64.
158. Martinez AJ. 1985. Free-Living Amoebas; Natural History,
Prevention, Diagnosis, Pathology and Treatment of Disease.
CRC Press, Boca Raton, FL.
159. Fowler M, Carter R. 1965. Acute pyogenic meningitis
probably due to Acanthamoeba sp.: a preliminary report.
British Med J 2:734.732.
160. Barnett ND, Kaplan AM, Hopkin RJ, Saubolle MA,
Rudinsky MF. 1996. Primary amoebic meningoencephali-
tis with Naegleria fowleri: clinical review. Pediatr Neurol 15:
230–234.
161. Cervantes-Sandoval I, De Jesus Serrano-Luna J, Tapia-
Malagón JL, Pacheco-Yepez J, Silva-Olivares A, Galindo-
Gomez S, Tsutsumi V, Shibayama M. 2007. Characteriza-
tion of Naegleria fowleri strains isolated from human cases of
primary amoebic meningoencephalitis in Mexico. Rev
Invest Clin 59:342.
162. Visvesvara GS, Moura H, Schuster FL. 2007. Pathogenic
and opportunistic free-living amoebae: Acanthamoeba spp.,
Balamuthia mandrillaris, Naegleria fowleri, and Sappinia dip-
loidea. FEMS Immunol Med Microbiol 50:1–26.
163. Parija S, Jayakeerthee S. 1999. Naegleria fowleri: a free
living amoeba of emerging medical importance. J Commun
Dis 31:153.
164. Martinez AJ, Visvesvara GS. 1997. Free-living, amphizoic
and opportunistic amebas. Brain Pathol 7:583–598.
165. Dorsch MM, Cameron AS, Robinson BS. 1983. The epi-
demiology and control of primary amoebic meningoence-
phalitis with particular reference to South Australia.
Trans Royal Soc Trop Med Hyg 77:372–377.
166. Esterman A, Dorsch M, Cameron S, Roder D, Robinson
B, Lake J, Christy P. 1984. The association of naegleria
fowleri with the chemical, microbiological and physical
characteristics of South Australian water supplies. Water
Res 18:549–553.
167. Schuster FL, Visvesvara GS. 2004. Free-living amoebae as
opportunistic and non-opportunistic pathogens of humans
and animals. Int J Parasitol 34:1001–1027.
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-23
168. CDC. 2012. http://www.cdc.gov/parasites/naegleria/state-
map.html. Accessed April 29.
169. Katranjidan, Olivia. 2011. Kansas resident dies from brain-
eating amoeba after swimming in warm lake. http://
abcnews.go.com/Health/kansas-resident-dies-brain-eating-
amoeba-swimming-warm/story?id=14489468. Accessed
05/6/2014.
170. Marciano-Cabral F, Cabral GA. 2007. The immune
response to Naegleria fowleri amebae and pathogenesis of
infection. FEMS Immu Med Microbiol 51:243–259.
171. Marshall MM, Naumovitz D, Ortega Y, Sterling CR.
1997. Waterborne protozoan pathogens. Clin Microbiol
Rev 10:67–85.
172. Visvesvara G, Schuster F. 2008. Opportunistic free-
living amebae, part I. Clin Microbiol Newsletter 30:
151–158.
173. Parija SC BM. 2014. http://emedicine.medscape.com/
article/223910-treatment.
174. Seidel K. 1984. Bacteriological problems of bathing waters
with special reference to whirlpool baths and “psycho-
tanks.” Schriftenreihe des Vereins fur Wasser-, Boden-und
Lufthygiene 65:303–310.
175. Kim J-H, Jung S-Y, Lee Y-J, Song K-J, Kwon D, Kim K,
Park S, Im K-I, Shin H-J. 2008. Effect of therapeutic
chemical agents in vitro and on experimental meningoen-
cephalitis due to Naegleria fowleri. Antimicrob Agents Che-
mother 52:4010–4016.
176. Visvesvara GS. 2010. Amebic meningoencephalitides
and keratitis: challenges in diagnosis and treatment. Curr
Opin Infect Dis 23:590–594.
177. CDC. 2014a. Naegleria fowleri—primary amebic menin-
goencephalitis (PAM). Treatment. http://www.cdc.gov/par
asites/naegleria/treatment.html. CDC.
178. Cabanes P-A, Wallet F, Pringuez E, Pernin P. 2001.
Assessing the risk of primary amoebic meningoencephalitis
from swimming in the presence of environmental Naegleria
fowleri. Appl Environ Microbiol 67:2927–2931.
179. Č erva L. 1971. Studies of limax amoebae in a swimming
pool. Hydrobiologia 38:141–161.
180. Rivera F, Ramirez E, Bonilla P, Calderon A, Gallegos E,
Rodriguez S, Ortiz R, Zaldivar B, Ramirez P, Durán A.
1993. Pathogenic and free-living amoebae isolated from
swimming pools and physiotherapy tubs in Mexico. Environ
Res 62:43–52.
181. Marciano-Cabral F, Maclean R, Mensah A, Lapat-
Polasko L. 2003. Identification of Naegleria fowleri in
domestic water sources by nested PCR. Appl Environ Micro-
biol 69:5864–5869.
182. Stockman LJ, Wright CJ, Visvesvara GS, Fields BS,
Beach MJ. 2011. Prevalence of Acanthamoeba spp. and
other free-living amoebae in household water, Ohio,
USA—1990–1992. Parasitol Res 108:621–627.
183. Buse HY, Lu J, Struewing IT, Ashbolt NJ. 2013. Eukary-
otic diversity in premise drinking water using 18S rDNA
sequencing: implications for health risks. Environ Sci Pollut
Res 20:6351–6366.
184. Bright KR, Marciano-Cabral F, Gerba CP. 2009.
Occurrence of Naegleria fowleri in Arizona drinking water
supply wells (PDF). J-Am Water Works Assoc 101:43–50.
185. Wellings F, Amuso P, Chang S, Lewis A. 1977. Isolation
and identification of pathogenic Naegleria from Florida
lakes. Appl Environ Microbiol 34:661–667.
186. John DT, Howard MJ. 1995. Seasonal distribution of
pathogenic free-living amebae in Oklahoma waters. Parasi-
tol Res 81:193–201.
187. Sheehan KB, Fagg JA, Ferris MJ, Henson JM. 2003. PCR
detection and analysis of the free-living amoeba Naegleria
in hot springs in Yellowstone and Grand Teton National
Parks. Appl Environ Microbiol 69:5914–5918.
188. Sifuentes LY CB, Gerba CP, Bright KR. 2014. The occur-
rence of Naegleria fowleri in recreational waters in Arizona.
Environ Sci Health A 49(11):41–49.
189. De Jonckheere J, Voorde Hd. 1977. The distribution of
Naegleria fowleri in man-made thermal waters. Am J Trop
Med Hyg 26:10–15.
190. Pélandakis M, Pernin P. 2002. Use of multiplex PCR
and PCR restriction enzyme analysis for detection and
exploration of the variability in the free-living amoeba
Naegleria in the environment. Appl Environ Microbiol 68:
2061–2065.
191. Pougnard C, Catala P, Drocourt J-L, Legastelois S, Per-
nin P, Pringuez E, Lebaron P. 2002. Rapid detection
and enumeration of Naegleria fowleri in surface waters by
solid-phase cytometry. Appl Environ Microbiol 68:
3102–3107.
192. Behets J, Declerck P, Delaedt Y, Verelst L, Ollevier F.
2007. Survey for the presence of specific free-living amoe-
bae in cooling waters from Belgian power plants. Parasitol
Res 100:1249–1256.
193. Ahmad AF, Lonnen J, Andrew PW, Kilvington S. 2011.
Development of a rapid DNA extraction method and one-
step nested PCR for the detection of Naegleria fowleri from
the environment. Water Res 45:5211–5217.
194. Heggie TW. 2010. Swimming with death: Naegleria fowleri
infections in recreational waters. Travel Med Infect Dis 8:
201–206.
195. CDC. 2013. Number of case-reports by age group and gen-
der. http://www.cdc.gov/parasites/naegleria/graphs.html#
casereports. Accessed May 6, 2015.
196. Yoder JS, Blackburn BG, Craun GF, Hill V, Levy DA,
Chen N, Lee SH, Calderon RL, Beach MJ. 2004. Surveil-
lance for waterborne-disease outbreaks associated with rec-
reational water—United States, 2001–2002. MMWR
Surveill Summ 53:1–22.
197. Griffin JL. 1983. The pathogenic amoeboflagellate Naegle-
ria fowleri: environmental isolations, competitors, ecologic
interactions, and the flagellate-empty habitat hypothesis. J
Protozool 30:403–409.
198. Sykora J, Keleti G, Martinez A. 1983. Occurrence and
pathogenicity of Naegleria fowleri in artificially heated
waters. Appl Environ Microbiol 45:974–979.
199. Lee SH, Levy DA, Craun GF, Beach MJ, Calderon RL.
2002. Surveillance for waterborne-disease outbreaks—
United States, 1999–2000. Morbidity and Mortality Weekly
Report Surveillance Summaries 51:1–47.
200. Maclean RC, Richardson DJ, Lepardo R, Marciano-
Cabral F. 2004. The identification of Naegleria fowleri
from water and soil samples by nested PCR. Parasitol Res
93:211–217.
201. De Jonckheere JF. 2011. Origin and evolution of the
worldwide distributed pathogenic amoeboflagellate Naegle-
ria fowleri. Infect Genet Evol 11:1520–1528.
202. Gautam PL, Sharma S, Puri S, Kumar R, Midha V,
Bansal R. 2012. A rare case of survival from primary
amebic meningoencephalitis. Indian J Crit Care Med
16:34.
203. Dziuban EJ, Liang JL, Craun GF, Hill V, Yu PA,
Painter J, Moore MR, Calderon RL, Roy SL, Beach
MJ. 2006. Surveillance for waterborne disease and outbreaks
associated with recreational water—United States, 2003–
2004. CDC, U.S. Department of Health and Human
Services.
204. Yoder J, Roberts V, Craun GF, Hill V, Hicks L,
Alexander NT, Radke V, Calderon RL, Hlavsa MC,
Beach MJ. 2008. Surveillance for waterborne disease and
outbreaks associated with drinking water and water not
intended for drinking—United States, 2005–2006.
MMWR Surveill Summ 57:39–62.
3.1.6-24 ▪ WATER
205. Yoder JS, Craun GF, Calderon RL. 2008. Surveillance for
waterborne disease and outbreaks associated with recreational
water use and other aquatic facility-associated health events–
United States, 2005–2006. Coordinating Center for Health
Information and Service, CDC, U.S. Department of
Health and Human Services.
206. Hlavsa MC, Brunkard JM. 2011. Surveillance for water-
borne disease outbreaks and other Health events associated
with recreational Water—United States, 2007–2008. CDC,
U.S. Department of Health and Human Services.
207. Yoder J, Eddy B, Visvesvara G, Capewell L, Beach M.
2010. The epidemiology of primary amoebic meningoence-
phalitis in the USA, 1962–2008. Epidemiol Infect 138:
968–975.
208. Kemble SK, Lynfield R, Devries AS, Drehner DM, Pom-
putius WF, Beach MJ, Visvesvara GS, Da Silva AJ, Hill
VR, Yoder JS. 2012. Fatal Naegleria fowleri infection
acquired in Minnesota: possible expanded range of a deadly
thermophilic organism. Clin Infect Dis cir961.
209. Louisiana Public Health. 2011. A primary amebic menin-
goencephalitis (PAM) fatality-Louisiana. Louisiana Office of
Public Health, Health LOoP.
210. Okuda DT, Hanna HJ, Coons SW, Bodensteiner JB.
2004. Naegleria fowleri hemorrhagic meningoencephalitis:
report of two fatalities in children. J Child Neurol 19:
231–233.
211. Shakoor S, Beg MA, Mahmood SF, Bandea R, Sriram R,
Noman F, Ali F, Visvesvara GS, Zafar A. 2011. Primary
amebic meningoencephalitis caused by Naegleria fowleri,
Karachi, Pakistan. Emerg Infect Dis 17:258.
212. CDC. 2013. Notes from the field: primary amebic menin-
goencephalitis associated with ritual nasal rinsing
—St. Thomas, US Virgin islands, 2012. MMWR 62:903.
213. Yoder JS, Straif-Bourgeois S, Roy SL, Moore TA, Visves-
vara GS, Ratard RC, Hill VR, Wilson JD, Linscott AJ,
Crager R. 2012. Primary amebic meningoencephalitis
deaths associated with sinus irrigation using contaminated
tap water. Clin Infect Dis 55:e79–e85.
214. CDC. 2014. Naegleria fowleri in Louisiana Public
Water Systems. http://www.cdc.gov/parasites/naegleria/
public-water-systems-louisiana.html. Accessed June 20,
2014.
215. Pernin P, Pélandakis M, Rouby Y, Faure A, Siclet F. 1998.
Comparative recoveries of Naegleria fowleri amoebae from
seeded river water by filtration and centrifugation. Appl
Environ Microbiol 64:955–959.
216. Stephanie MP, Russell SP, Jeff AB, Anne MSM. 2013.
Quantitative assessment of Naegleria fowleri and Escherichia
coli concentrations within a Texas reservoir. Journal of
Water and Health 112:346–357.
217. Johnson P, Deromedi A, Lebaron P, Catala P, Havens C,
Pougnard C. 2007. High throughput, real-time detection
of Naegleria lovaniensis in natural river water using
LED-illuminated Fountain FlowTM cytometry. J Appl
Microbiol 103:700–710.
218. Kilvington S, Mann P, Warhurst D. 1984. Differentiation
between Naegleria fowleri and N. lovaniensis using isoen-
zyme electrophoresis of aspartate aminotransferase. Trans
R Soc Trop Med Hyg 78:562–563.
219. Pernin P, Cariou ML, Jacquier A. 1985. Biochemical
identification and phylogenetic relationships in free-
living amoebas of the genus Naegleria. J Protozool 32:
592–603.
220. De Jonckheere JF. 1982. Isoenzyme patterns of pathogenic
and non-pathogenic Naegleria spp. using agarose isoelectric
focusing. Ann Microbiol 133A:319–342.
221. Pernin P. 1984. Isoenzyme patterns of pathogenic and non-
pathogenic thermophilic Naegleria strains by isoelectric
focusing. Int J Parasitol 14:459–465.
222. De Jonckheere JF. 1987. Characterization of Naegleria spe-
cies by restriction endonuclease digestion of whole-cell
DNA. Mol Biochem Parasitol 24:55–66.
223. Mclaughlin G, Brandt F, Visvesvara G. 1988. Restriction
fragment length polymorphisms of the DNA of selected
Naegleria and Acanthamoeba amebae. J Clin Microbiol 26:
1655–1658.
224. Reveiller FL, Varenne MP, Pougnard C, Cabanes PA,
Pringuez E, Pourima B, Legastelo S, Pernin P. 2003.
An enzyme-linked immunosorbent assay (ELISA) for the
identification of Naegleria fowleri in environmental water
samples. J Eukaryot Microbiol 50:109–113.
225. Reveiller FL, Cabanes P-A, Marciano-Cabral F. 2002.
Development of a nested PCR assay to detect the patho-
genic free-living amoeba Naegleria fowleri. Parasitol Res
88:443–450.
226. Kilvington S, Beeching J. 1995. Development of a PCR for
identification of Naegleria fowleri from the environment.
Appl Environ Microbiol 61:3764–3767.
227. Behets J, Seghi F, Declerck P, Verelst L, Duvivier L, Van
Damme A, Ollevier F. 2003. Detection of Naegleria spp.
and Naegleria fowleri: a comparison of flagellation tests,
ELISA and PCR. Water Sci Technol 47:117–122.
228. Behets J, Declerck P, Delaedt Y, Verelst L, Ollevier F.
2007. A duplex real-time PCR assay for the quantitative
detection of Naegleria fowleri in water samples. Water Res
41:118–126.
229. Schild M, Gianinazzi C, Gottstein B, Müller N. 2007.
PCR-based diagnosis of Naegleria sp. infection in formalin-
fixed and paraffin-embedded brain sections. J Clin Microbiol
45:564–567.
230. Qvarnstrom Y, Visvesvara GS, Sriram R, Da Silva AJ.
2006. Multiplex real-time PCR assay for simultaneous
detection of Acanthamoeba spp., Balamuthia mandrillaris,
and Naegleria fowleri. J Clin Microbiol 44:3589–3595.
231. Mull BJ, Narayanan J, Hill VR. 2013. Improved method
for the detection and quantification of Naegleria fowleri
in water and sediment using immunomagnetic separation
and real-time PCR. J Parasitol Res 2013. Article ID
608367, http://dx.doi.org/10.1155/2013/608367.
232. Maďarová L, Trnková K, Feiková S, Klement C,
Obernauerová M. 2010. A real-time PCR diagnostic
method for detection of Naegleria fowleri. Exp Parasitol
126:37–41.
233. Behets J, Declerck P, Delaedt Y, Verelst L, Ollevier F.
2006. Quantitative detection and differentiation of free-
living amoeba species using SYBR green–based real-
time PCR melting curve analysis. Curr Microbiol 53:
506–509.
234. Robinson BS, Monis PT, Dobson PJ. 2006. Rapid, sensi-
tive, and discriminating identification of Naegleria spp. by
real-time PCR and melting-curve analysis. Appl Environ
Microbiol 72:5857–5863.
235. Puzon GJ, Lancaster JA, Wylie JT, Plumb JJ. 2009. Rapid
detection of Naegleria fowleri in water distribution pipeline
biofilms and drinking water samples. Environ Sci Technol
43:6691–6696.
236. John DT, Howard MJ. 1996. Techniques for isolating
thermotolerant and pathogenic free-living amebae. Folia
Parasitol 43: 267–271.
237. Marciano-Cabral F, Cline M. 1987. Chemotaxis by Nae-
gleria fowleri for bacteria. J Protozool 34:127–131.
238. Visvesvara G, Peralta M, Brandt F, Wilson M, Aloisio C,
Franko E. 1987. Production of monoclonal antibodies to
Naegleria fowleri, agent of primary amebic meningoence-
phalitis. J Clin Microbiol 25:1629–1634.
239. Sparagano O. 1993. Detection of Naegleria fowleri cysts in
environmental samples by using a DNA probe. FEMS
Microbiol Lett 112:349–351.
 3.1.6. Detection of Protozoa in Surface and Finished Waters ▪ 3.1.6-25
240. Zhou L, Sriram R, Visvesvara GS, Xiao L. 2003. Genetic
variations in the internal transcribed spacer and mitochon-
drial small subunit rRNA gene of Naegleria spp. J Eukaryot
Microbiol 50:522–526.
241. Riner DK, Nichols T, Lucas SY, Mullin AS, Cross JH,
Lindquist HA. 2010. Intragenomic sequence variation
of the ITS-1 region within a single flow-cytometry-
counted Cyclospora cayetanensis Oocysts. J Parasitol 96:
914–919.
242. Ferguson C, Deere D, Sinclair M, Chalmers RM, Elwin
K, Hadfield S, Xiao L, Ryan U, Gasser R, El-Osta YA.
2006. Meeting report: application of genotyping methods
to assess risks from Cryptosporidium in watersheds. Environ
Health Perspect 114:430.
243. Wielinga C, Thompson R. 2007. Comparative evaluation
of Giardia duodenalis sequence data. Parasitology 134:
1795–1821.
244. Weiss LM, Cushion MT, Didier E, Xiao L, Marciano-
Cabral F, Sinai AP, Matos O, Calderon EJ, Kaneshiro
ES. 2013. The 12th International workshops on oppor-
tunistic Protists (IWOP-12). J Eukaryot Microbiol 60:
298–308.
 Drinking Water Microbiology
MARYLYNN V. YATES
3.1.7
One of the Millennium Development Goals is to halve,
between 1990 and 2015, the proportion of the population
that does not have sustainable access to safe drinking water,
which would increase the coverage from 76% to 88% (1).
This goal was met in 2010, with 116 countries meeting the
target (2). Since 1990, more than two billion people have
gained access to improved drinking water sources, with 89%
of the population covered. There were only three countries
(Democratic Republic of the Congo, Mozambique, and
Papua New Guinea) where less than half the population
had access to an improved drinking water source, and 35
countries (26 of which are in sub-Saharan Africa) in which
only 50–75% of the population has access to an improved
source of drinking water. Although this represents a substan-
tial improvement, more than 700 million people still lack
access to improved sources of drinking water, almost half of
which are in sub-Saharan Africa (2). It is predicted that there
will still be more than half a billion people without access to
an improved drinking water supply by 2015.
In the United States and other countries where drinking
water treatment is routinely practiced, waterborne disease
outbreaks still occur (see, e.g., 3). Although people generally
assume that all drinking water in developed countries is
treated, that is not the case. For example, in the United
States, the Environmental Protection Agency (EPA) esti-
mates that 15% of the population obtains their drinking water
from untreated private wells or other untreated sources (4, 5);
these sources are not subject to regulation under the Safe
Drinking Water Act. Most of the drinking water–related
disease outbreaks in the United States are associated with
untreated or inadequately treated ground water, Legionella
bacteria, and distribution system deficiencies (6). Epidemio-
logical investigations have estimated that between 12 million
and 19.5 million waterborne illnesses occur annually in the
United States (7). While these illnesses rarely result in death
in developed countries, millions of deaths occur every year
in developing countries where adequate access to medical
care is limited.
CLASSIFICATION OF WATER-RELATED
DISEASES
The term “waterborne disease” is commonly used to refer to
all diseases that may result from exposure to or consumption
of water. However, there are numerous ways to classify
doi:10.1128/9781555818821.ch3.1.7
3.1.7-1
microorganisms that are associated with water and cause
human disease. White et al. (8) proposed four classifications,
commonly called the Bradley Classifications:
• Waterborne: an enteric microorganism enters the water
via fecal contamination and is transmitted through inges-
tion of the contaminated water. Examples include typhoid
and cholera.
• Water-washed: an enteric microorganism is transmitted
through ingestion of fecal material, but the presence of
fecal material is due to a lack of water for bathing, washing
of hands, utensils, and so on. In these cases, the pathogen
is not present in the water, but on hands or utensils,
and transferred to water, food, and so on that is then con-
sumed. Waterborne pathogens can also be transmitted
through the water-washed route.
• Water-based: a worm spends a part of its life cycle in the
aquatic environment. Water-based infections are of two
types: those acquired via ingestion (such as is the case
with the guinea worm, Dranunculus medinensis), and those
acquired by contact with the water (such as with Schisto-
soma, the causative agent of schistosomiasis).
• Water-related: an insect that transmits the pathogen
breeds in water. Examples include malaria and Dengue
fever.
Various modifications of this classification scheme have
been proposed. For example, Bradley himself proposed the
addition of two new classifications (9): “water-aerosol” and
“water source crowding transmission.” The water aerosol class
is used for those diseases, such as legionellosis, that are trans-
mitted via aerosols. The water source crowding is meant for
those diseases that are the result of multiple species (such as
livestock, wildlife, and humans) congregating at waterholes,
thereby facilitating the spread of zoonotic diseases. Yang
et al. (10) proposed the addition of “water-dispersed” diseases
to refer to those caused by microorganisms that proliferate in
fresh water and enter the human body through the respiratory
tract. An example of this is Legionella. Bartram (11) recently
proposed a modification of Bradley’s scheme in which the
term “water-washed” is replaced by “water access-related”
and a new classification, “engineered water system associated”
is added. The latter is intended for those disease agents that
arise within engineered systems (such as Legionella, which
are naturally present in water but can multiply to harmful
3.1.7-2 ▪ WATER
levels in engineered water systems, or certain nontuber-
culous mycobacteria that can flourish in water supply system
biofilms).
MICROORGANISMS AS CAUSATIVE AGENTS
OF DRINKING WATER–RELATED DISEASE
Traditionally, the microorganisms of most concern as drink-
ing water contaminants have been those of enteric origin.
Hundreds of different types of microorganisms may be present
in the fecal material of infected humans and animals. The
fecal material of infected individuals is the major source of
pathogenic microorganisms in domestic wastewater; however,
urine may also be a source of certain pathogenic microorgan-
isms, such as Salmonella enterica and polyomaviruses. The
concentrations and types of pathogens found in domestic
wastewater vary over time, depending on the disease inci-
dence in the population producing the wastewater, season
of the year, amount of water being used, the economic status
of the population, and quality of the drinking water. If the
fecal material or wastewater enters a water body that is used
as a source of potable water, there is a potential for disease
transmission. The major groups of pathogenic microorgan-
isms transmitted by fecally contaminated water and food
include bacteria, viruses, and parasites.
Other nonfecal microorganisms also cause waterborne
disease from the consumption or use of potable water; these
are typically microorganisms indigenous to the water or
sediment. For example, Legionella, a bacterium transmitted
via aerosols, is a frequent cause of drinking water–associated
disease. Indeed, in 2001 the Centers for Disease Control
and Prevention (CDC) began including reports of Legionella-
caused outbreaks with the drinking water surveillance data
(12). A sediment-dwelling Amoeba, Naegleria fowleri, has
also been reported as the cause of several waterborne disease
outbreaks (12). The World Health Organization (13, 14)
has compiled a set of fact sheets that describe the human
health effects, source and occurrence, routes of exposure,
and significance in drinking water for many waterborne
pathogens.
Bacteria
The bacteria of most concern with respect to transmission
from domestic wastewater are the enteric bacteria, which
are those that infect the gastrointestinal tract of humans
and are shed in the fecal material. Enteric bacteria are well
adapted to the conditions present in the gastrointestinal tract,
which include high organic carbon concentrations, low pH,
and a relatively high temperature (∼37°C). When these
organisms are introduced to the wastewater, water, or soil
environment, the conditions are generally very different
from those in the gastrointestinal tract, in that nutrients
tend to be present in low concentrations or forms that are dif-
ficult to metabolize, the temperature is lower, and so on. As a
result, the enteric bacteria typically do not compete well with
the indigenous soil or water bacteria for the scarce nutrients
available. Therefore, their ability to reproduce and even
survive without reproducing in the environment is typically
limited to a few days to weeks, depending on the environmen-
tal conditions.
Historically, the bacterial pathogens have been the focus
of waterborne disease because they were among the first
microorganisms recognized. Bacteria have caused a number
of diseases, such as cholera, typhoid, and dysentery that
have been responsible for large epidemics (and, in the case
of cholera, pandemics) all over the world. In general, many
bacteria pathogens are relatively easy to remove through
wastewater and drinking water treatment processes and are
therefore of less concern than the other groups of pathogens
in many industrialized countries. However, these pathogens
are still of great concern in many of the less industrialized
countries because of the lack of adequate sanitary treatment
processes in those countries.
Some of the most common bacterial pathogens found in
human and animal waste that are transmitted by contami-
nated water and food are listed in Table 1. Several relevant
characteristics, such as infectivity, survival in water, resist-
ance to chlorine, and health impacts are also provided.
Parasites
Parasites are another group of pathogenic microorganisms
that can be transmitted through contaminated drinking
water. Enteric parasites that are pathogenic to humans gener-
ally can be classified into two groups: the protozoa and the
helminths. Protozoa are single-celled microorganisms whose
life cycles include at least one vegetative and at least one rest-
ing stage. The resting stage of the organism (e.g., cyst, ovum)
is generally relatively resistant to inactivation due to environ-
mental stresses and during conventional drinking and waste-
water treatment processes. Most of the intestinal protozoa
are transmitted by fecally contaminated water, food, or other
materials. The infective stage of the protozoa varies by organ-
ism. In the case of Cryptosporidium, the resting (i.e., oocyst)
form must be ingested to initiate infection; for others, the
vegetative form must be ingested for infection to occur.
Some of the most common protozoan pathogens found in
human and animal waste that are transmitted by contami-
nated water and food are listed in Table 1. Several relevant
characteristics, such as infectivity, survival in water, resist-
ance to chlorine, and health impacts are also provided.
Viruses
Although feces always contain high concentrations of
nonpathogenic bacteria, pathogenic viruses are only present
in fecal material if the individual is infected. There are
more than 100 types of enteric viruses that can be present
in domestic wastewater. A list of several of the pathogenic
human enteric viruses, along with several relevant charac-
teristics, such as infectivity, survival in water, resistance to
chlorine, and health impacts, is shown in Table 1. The man-
ifestations that can be caused by enteric virus infections range
widely—from asymptomatic infections to relatively mild, self-
limiting diarrhea, to more severe illnesses, including respira-
tory illness, infectious hepatitis, paralysis, encephalitis, and
myocarditis.
Indicator Microorganisms
Because feces, and therefore fecally contaminated drinking
water, may contain so many different types of enteric patho-
gens, it is not technically or economically feasible to test
wastewater for the presence of all different types of pathogens.
Therefore, the microbiological quality of water has been
assessed by measuring the concentration of indicator bacteria
it contains. Indicator bacteria are always present in fecal
material. The logic in using fecal indicator bacteria is that if
they are found to be present, the water contains fecal material
and therefore may also contain pathogenic microorganisms.
In contrast, if fecal indicators are absent, the water is consid-
ered not to be contaminated by fecal material, and therefore
no fecal pathogens are present. The most common groups of
indicator organisms used are the total coliform bacteria, the
fecal coliform bacteria (also called thermotolerant coliform
TABLE 1 Selected pathogens that can be transmitted by water
Incubation Health Persistence in Resistance to Relative
Disease Duration of illness
period significancea waterb chlorinec infectivityd
Burkholderia pseudomallei Pneumonia High May multiply Low Low
Campylobacter Gastroenteritis 3–5 days 1–4 days High Moderate Low Moderate
Legionella Pontiac fever 5–66 hours 2–7 days High May multiply Low Moderate
Pneumonia 2–14 days Weeks to months High May multiply Low Moderate
Leptospira Weil’s disease (headache, chills, 2–20 days 3 days to 3 weeks High Long Low High
fever, nausea, neck or joint pain)
Shiga-toxin-producing Gastroenteritis, hemolytic uremic 12 h–8 days 1 day to 3 weeks High Moderate Low Low to high,
E. coli syndrome, kidney failure depending on strain
Salmonella enterica serovar Typhoid fever 7–28 days Weeks to months High Moderate Low Low
typhi
Salmonella Salmonellosis 8–48 h 3–5 days High Moderate Low Low
Shigella Bacillary dysentery 1–7 days 4–7 days High Short Low High
Vibrio cholerae O1 Profuse, watery diarrhea, vomiting, 9–72 h 3–4 days High Short to long Low Low
rapid dehydration
Vibrio cholerae non-O1 Watery diarrhea 1–5 days 3–4 days High Short to long Low Low
Yersinia enterocolitica Gastroenteritis 2–7 days 1–21 days Moderate Long Low Low
Cryptosporidium parvum Diarrhea 1–2 weeks 4–21 days High Long High High
Cyclospora cayatanensis Watery diarrhea alternating 2–11 days Days High Long High High
with constipation

3.1.7. Drinking Water Microbiology ▪ 3.1.7-3

Entamoeba histolytica Amoebic dysentery 2–4 weeks Weeks to months High Moderate High High
Giardia intestinalis Diarrhea, malabsorption 5–25 days Weeks to months High Moderate High High
Microsporidium Chronic diarrhea, weight loss
Naegleria fowleri Primary amoebic meningoencephalitis Minutes to hours High May multiply Low Moderate
Toxoplasma gondii Toxoplasmosis High Long High High
Adenovirus Respiratory illness, conjunctivitis, 1–4 days 2–3 days Moderate Long Moderate High
vomiting, diarrhea
Astrovirus Vomiting, diarrhea 3–4 days 2–7 days Moderate Long Moderate High
Calicivirus Vomiting, diarrhea Moderate Long Moderate High
Coronavirus Vomiting, diarrhea Moderate Long Moderate High
Enterovirus: 3–14 days Variable
Poliovirus Paralysis, meningitis, fever High Long Moderate High
Coxsackie A Meningitis, fever, herpangina, High Long Moderate High
respiratory illness, paralysis
(Continued on next page)
 3.1.7-4 ▪ WATER
TABLE 1 Selected pathogens that can be transmitted by water (Continued )
Disease Incubation Duration of illness Health Persistence in Resistance to Relative
period significancea waterb chlorinec infectivityd
Coxsackie B Myocarditis, congenital heart High Long Moderate High
anomalies,
rash, fever, meningitis, respiratory
illness, pleurodynia
Echovirus Meningitis, encephalitis, respiratory High Long Moderate High
illness,
rash, diarrhea, fever, myocarditis,
endocarditis
Enterovirus 68–71 Meningitis, encephalitis, High Long Moderate High
respiratory illness, acute
hemorrhagic conjunctivitis, fever
Hepatitis A virus Infectious hepatitis 15–50 days 1 week to several months High Long Moderate High
Hepatitis E virus Hepatitis 15–65 days 1 week to several months High Long Moderate High
Norovirus Epidemic vomiting and diarrhea 1–3 days 1–3 days High Long Moderate High
Reovirus Respiratory illness Moderate Long Moderate High
Rotavirus Diarrhea, vomiting 1–3 days 3–7 days High Long Moderate High
Sapoviruses Gastroenteritis High Long Moderate High
Sources: 14, 15, 16, 17.
a
Health significance relates to the severity of impact, including association with outbreaks.
b
Detection period for infective stage in water at 20°C: short, up to 1 week; moderate, 1 week to 1 month; long, over 1 month.
c
When the infective stage is freely suspended in water treated at conventional doses and contact times and pH between 7 and 8. Low means 99% inactivation at 20°C generally in <1 min, moderate 1–30 min and high
>30 min. It should be noted that organisms that survive and grow in biofilms, such as Legionella and mycobacteria, will be protected from chlorination.
d
From experiments with human volunteers, from epidemiological evidence, and from animal studies. High means infective doses can be 1–100 organisms or particles, moderate 100–10,000, and low >10,000.

bacteria), the enterococci (also called fecal streptococci),
and E. coli.
However, in industrialized countries, many bacterial dis-
eases have been controlled through the use of drinking water
and wastewater treatment, including disinfection. As this has
occurred, more of the focus in drinking water and wastewater
treatment has been on controlling the concentrations of
parasitic and viral pathogens. Unfortunately, the bacterial
indicators do not always correlate well with the presence or
behavior of the nonbacterial pathogens. Efforts are under
way to identify more appropriate indicators for these patho-
gens. A more complete discussion of this issue can be found
in the section on Microbial Source Tracking.
A National Academies of Science committee (18) devel-
oped a series of criteria that they recommend when assessing
the appropriateness of any indicator or group of indicators.
These are based on criteria developed more than 40 years
ago and separate the biological attributes of the indicators
from the attributes of the methods used to detect them.
The desirable biological attributes of indicators, as developed
by the NRC are:
• Correlated to health risk
• Similar (or greater) survival to pathogens under environ-
mental conditions
• Similar (or greater) transport to pathogens
• Present in greater numbers than pathogens
• Specific to a fecal source or identifiable as to source of
origin
The desirable attributes of the methods to detect the
indicators, according to the NRC, are:
• Specificity (independent of matrix effects)
• Broad applicability
• Precision
• Adequate sensitivity
• Rapidity of results
• Quantifiable
• Measures viability or infectivity
• Logistical feasibility
The most important characteristic of any indicator or indi-
cator system is its ability to indicate possible health risk to
individuals exposed to the water or food.
BURDEN OF DISEASE
The most significant health risk from drinking water is from
fecal contamination (19), often due to untreated or inad-
equately treated water supplies. The World Health Organiza-
tion (2) estimates that almost 25% of the people in the world
do not have access to adequately treated drinking water.
According to data compiled by WHO/UNICEF (20), in
2010, 783 million people continued to use unimproved
sources to meet their drinking water needs, and 2.5 billion
people continued to use an unimproved sanitation facility
or defecate in the open. The presence of human and/or ani-
mal fecal material in drinking water can lead to a number of
diseases, with diarrhea among the most common. In countries
where people do not have access to medical care, diarrhea can
lead to death from dehydration, especially among the very
young and the elderly. Nearly a billion people worldwide,
mostly children in the developing world, suffer from diseases
caused by pathogenic microorganisms in water. Rotavirus, the
most common cause of severe diarrheal disease in infants and
3.1.7. Drinking Water Microbiology ▪ 3.1.7-5
young children globally, results in an estimated 527,000
deaths annually (85% occur in Africa and Asia) (21, 22).
WHO/UNICEF (20) estimates that the annual global eco-
nomic losses associated with inadequate water supply and
sanitation were US$260 billion.
According to UNICEF (2012), diarrhea is the second
most common cause of death in children under the age of
five; in 2012, this resulted in 1.2 million deaths worldwide.
In addition, WHO estimated that 10% of the people in less
industrialized countries are infected with parasitic worms as
a result of the improper management of waste. Among the
most important risk factors for these diseases are access to
improved drinking water supplies and access to improved
sanitation. In high-income areas, 99–100% of the population
has this access, regardless of whether they live in urban or
rural areas. In low-income areas, however, the situation is
markedly different. It is estimated that only 78% (89% urban,
73% rural) of the population has access to improved drinking
water sources and 34% (53% urban, 26% rural) has access to
improved sanitation.
While the impacts of diarrhea may be most severe in less
industrialized countries, they can have a significant impact
in industrialized countries as well. For example, in the United
States, the CDC estimates for 2006 were that there were 2.1
million cases of rotavirus gastroenteritis in children younger
than five years every year. These cases result in more than
400,000 outpatient or doctors’ office visits, and between
205,000 and 272,000 emergency room visits. Of these chil-
dren, 55,000 to 70,000 require hospitalization, and 20–60
children die every year as a result of rotavirus gastroenteritis.
The introduction of rotavirus vaccines has decreased the
number of these cases dramatically.
Waterborne disease continues to occur in the United
States and other countries with drinking water and waste-
water treatment systems, albeit at much lower levels than in
countries without such systems.
It must be noted that determining the magnitude of dis-
ease caused by the consumption of contaminated water
is extremely difficult. One reason is that many infected indi-
viduals never exhibit signs or symptoms of disease. Another
reason is that many of the enteric pathogens cause gastroen-
teritis. If the illness is mild, as it is in many cases, no medical
care is sought and thus there is no record of the illness. In the
case of most enteric pathogens, the disease is not reportable
in any event. Compounding the issue is the fact that many
of the enteric pathogens are also transmitted by contaminated
foods, making it difficult to determine whether drinking
water was the source of the infection.
It is difficult to get accurate information on the true
burden of waterborne disease even in developed countries.
In the United States, state laws require reporting of only a
few pathogens, so the total number of cases of illness is
believed to be underestimated. Scallan et al. (23) estimated
that illnesses from pathogens associated with water are under-
reported by an average of 43-fold and may be underreported
by up to 143 times for certain Vibrio species. Furthermore, it
has been estimated that waterborne pathogens cause 8.5–
12% of annual acute gastrointestinal illness cases, affecting
between 12 and 19 million people annually in the United
States (7, 24, 25).
Investigations of suspected waterborne disease outbreaks
are typically conducted by local or state public health
departments. In some cases, resources are too scarce to permit
a thorough investigation. In other cases, however, consider-
able time and effort is put into investigating the outbreak,
and in some cases, the CDC is enlisted to assist. The CDC
3.1.7-6 ▪ WATER

FIGURE 1 Etiology of 885 drinking water-associated outbreaks, by year—United States, 1971–2012.

doi:10.1128/9781555818821.ch3.1.7.f1

is dependent on local and state health departments for the
voluntary reporting of outbreaks, so the numbers presented
are believed to represent only a fraction of the disease out-
breaks that actually occur.
The CDC compiles reports of waterborne disease out-
breaks and reports on them approximately every two years.
The latest report, published in 2015 (6), covers the years
2011–2012. A summary of general causes of the reported
outbreaks from 1971–2012 is shown in Fig. 1. The reported
etiologic agents in those outbreaks with a microbiological
cause are shown in Fig. 2.
DRINKING WATER REGULATIONS
In the United States, the EPA is required, under the Safe
Drinking Water Act, to set regulatory limits on the amount
of certain contaminants that water provided by public water
systems may contain. These regulations are established for

FIGURE 2 Reported causes of outbreaks in drinking water in the United States (data from the CDC, excludes Legionella).
doi:10.1128/9781555818821.ch3.1.7.f2

both chemical and microbiological contaminants. For micro-
biological contaminants, rather than setting maximum allow-
able concentrations, the EPA relies primarily on monitoring
for indicator microorganisms (e.g., E. coli, total coliform bac-
teria, fecal coliform bacteria, enterococci, coliphages) and
treatment requirements to control the presence of pathogenic
microorganisms in drinking water. A list of the pathogens
regulated by the EPA, their maximum contaminant level
goal, and the regulations that describe the treatment require-
ments is provided in Table 2 (26).
The EPA is also required to ensure that new or newly
recognized drinking water contaminants are considered for
regulation. This is done through the development, every
five years, of the Contaminant Candidate List (CCL). The
CCL is “a list of contaminants that are currently not subject
to any proposed or promulgated national primary drinking
water regulations, but are known or anticipated to occur in
public water systems. Contaminants listed on the CCL may
require future regulation under the Safe Drinking Water
Act” (26). The fourth CCL is currently under review; the
microorganisms proposed for inclusion on the CCL4 are
shown in Table 3.
The effect of establishing drinking water treatment
requirements based on pathogen removals appear to have
had an impact on the number of reported waterborne out-
breaks associated with drinking water. As seen in Fig. 1, the
number of reported outbreaks has decreased by approximately
50% since 1971.
CLIMATE CHANGE AND WATERBORNE
DISEASE
Predicted Changes in Climate
In its latest reports, the Intergovernmental Panel on Climate
Change (27, 28) made several predictions regarding changes
in climate that are likely to have an impact on diseases trans-
mitted via water. By the end of this century, they predict that:
• Changes in mean annual precipitation will vary
• The high latitudes and the equatorial Pacific Ocean are
likely to experience an increase in annual mean
precipitation
• In many mid-latitude and subtropical dry regions, mean
precipitation will likely decrease
• In many mid-latitude wet regions, mean precipitation
will likely increase
• Over most of the mid-latitude landmasses and over wet
tropical regions, extreme precipitation events will very
likely become more intense and more frequent as global
mean surface temperature increases
• It is likely that the area encompassed by monsoon systems
will increase. While monsoon winds are likely to weaken,
monsoon precipitation is likely to intensify, and the mon-
soon season in many regions will lengthen, due to sooner
beginnings and later endings
• There is high confidence that the El Niño–Southern
Oscillation (ENSO) will remain the dominant mode of
interannual variability in the tropical Pacific. ENSO-
related precipitation variability on regional scales will
likely intensify
• Renewable surface water and groundwater resources in
most dry subtropical regions are predicted to decrease,
increasing competition for the already scarce resource
• The frequency of droughts in presently dry regions will
likely increase
3.1.7. Drinking Water Microbiology ▪ 3.1.7-7
• Water resources at high latitudes are predicted to
increase
• The interaction of increased temperature; increased
sediment, nutrient, and pollutant loadings from heavy
rainfall; increased concentrations of pollutants during
droughts; and disruption of treatment facilities during
floods will reduce raw water quality and pose risks to
drinking water quality
• The fractions of the global population that will experience
water scarcity and be affected by major river floods are
projected to increase
In the United States, extreme precipitation events are pre-
dicted to increase everywhere in the country (29), although
the increases will be greater along the coasts. The magnitude
of the increase will depend on the amount of emissions
reductions that occur: under rapid reduction scenarios,
the frequency of extreme precipitation events may double
by the end of the century (compared with the frequency
from 1981 to 2000). However, if no reductions in emissions
occur, the frequency of extreme precipitation events is
predicted to increase by up to five times that of the late
20th century.
Effects of Changing Climate on Waterborne Disease
Several climate-related factors are known to impact disease
transmission; these include precipitation patterns, extreme
rainfall events, air and water temperatures, and seasonal var-
iations (30, 31). These factors can impact disease transmis-
sion in a number of ways, for example, by affecting the
survival, growth, transmission, or virulence of the microor-
ganism. Climate can also indirectly influence disease trans-
mission by altering the habitat of the species that act as
reservoirs for the pathogens.
Pathogens that can be affected by climate change include
enteric organisms that are transmitted by the fecal-oral route
and bacteria and protozoa that occur naturally in aquatic
systems. Soller et al. (32) found that 97% of all water-
borne illness in the United States was caused by eight
pathogens that can be considered “climate sensitive.” These
include noroviruses, rotaviruses, and adenoviruses; the
bacteria Campylobacter jejuni, E. coli O157:H7, Salmonella
enterica; and the protozoa Cryptosporidium spp. and
Giardia spp.
As discussed previously, although most drinking water in
the United States is treated, and subject to regulation by
the EPA, waterborne disease outbreaks continue to occur
(see, e.g., 6). Interestingly, most outbreaks are associated
with untreated or inadequately treated groundwater, Legion-
ella bacteria, and distribution system deficiencies (3, 6), rather
than water that is contaminated when it leaves the drinking
water treatment plant. Between 2011 and 2012, almost
50% of the illnesses (201/431 cases) were associated with
consumption of contaminated, untreated groundwater (6),
and most of the remaining cases were associated with contam-
ination of the distribution system. Legionella were responsible
for 25.3% of those cases of illness.
The CDC has reported an increased incidence of reported
waterborne outbreaks in times of extreme precipitation (33).
This is likely due to the increased numbers of pathogens in
treated drinking water that occur following these events
(34). The effects appear to be especially noticeable among
children (35, 36).
Current drinking water treatment processes in developed
countries are very effective in preventing waterborne disease,
TABLE 2 Drinking water pathogens regulated by the EPA
Pathogen Definition and health effects
Cryptosporidium Cryptosporidium is a single-celled protozoan parasite commonly
found in lakes and rivers, especially when the water is
contaminated with sewage and animal waste. Cryptosporidium
can cause gastrointestinal illness (e.g., diarrhea, vomiting,
cramps).
Giardia lamblia Giardia lamblia is a single-celled protozoan parasite that lives in the
intestine of infected humans or animals. It is found on surfaces or
in soil, food, or water that has been contaminated with the feces
from infected humans or animals. G. lamblia can cause symptoms
such as nausea, cramps, diarrhea, and associated headaches.
Legionella Legionella bacteria are found naturally in the environment, usually
in water. The bacteria grow best in warm water, like the kind
found in hot tubs, cooling towers, hot water tanks, large
plumbing systems, or parts of the air conditioning systems of large
buildings. Legionella bacteria in water are a health risk if the
bacteria are aerosolized (e.g., in an air conditioning system or a
shower) and then inhaled. Inhalation can result in a type of
pneumonia known as Legionnaire’s disease.
Viruses Enteroviruses are small viruses that live in the intestines of infected
(enteric) humans or animals. This group includes the polioviruses,
coxsackieviruses, echoviruses, and other enteroviruses. In
addition to the 3 different polioviruses, there are 62 nonpolio
enteroviruses that can cause disease in humans: 23 coxsackie A
viruses, 6 coxsackie B viruses, 28 echoviruses, and 5 other
enteroviruses. Illness from viruses ranges from gastroenteritis
caused by viruses such as rotavirus and norovirus (Norwalk-like
virus) to meningitis caused by echovirus to myocarditis caused by
coxsackie B.
Source: https://safewater.zendesk.com/hc/en-us/sections/203279767.
3.1.7-8 ▪ WATER
Treatment methods
Maximum
Treatment technique proven to be effective
contaminant
and regulation(s) for removal or
level goal
inactivation
Zero Surface Water Treatment Rule requirements: systems using surface Disinfection with
water or ground water under the direct influence of surface water ultraviolet light
(GWUDI) must disinfect and filter their water so that 99% of or ozone and/or
Cryptosporidium oocysts are removed or inactivated (killed). filtration
Unfiltered systems (systems that meet criteria for avoiding
filtration) are required to include Cryptosporidium in their existing
watershed control provisions.
Zero Surface Water Treatment Rule requirements: systems using surface Disinfection and/or
water or GWUDI must disinfect and filter their water so that filtration
99.9% of G. lamblia is removed or inactivated. Unfiltered systems
(systems that meet criteria for avoiding filtration) are also required
to include G. lamblia in their existing watershed control
provisions.
Zero Surface Water Treatment Rule requirements: systems using surface Disinfection and/or
water or GWUDI must (1) disinfect their water, and (2) filter filtration
their water or meet criteria for avoiding filtration.
There is no limit specific to Legionella, but EPA believes that if
G. lamblia and viruses are removed/inactivated according to the
treatment techniques in the surface water treatment rules,
Legionella will be controlled.
Zero Surface Water Treatment Rule requirements: systems using surface Disinfection and/or
water or GWUDI must disinfect and filter their water so that filtration
99.99% of viruses are removed or inactivated.
Ground Water Rule: public water systems that use ground water
must take corrective action if a sufficient deficiency is identified,
or if the initial source sample (if required by the state) or one of
the five additional groundwater source samples tests positive for
fecal contamination (E. coli, Enterococci, or coliphage). The
systems must implement at least one of the following corrective
actions:
• Correct all significant deficiencies
• Provide an alternate source of water
• Eliminate the source of contamination
• Provide treatment that reliably achieves at least 4-log treatment
of viruses (using inactivation, removal, or a state-approved
combination of 4-log virus inactivation and removal) before or
at the first customer for the groundwater source.
 3.1.7. Drinking Water Microbiology ▪ 3.1.7-9

TABLE 3 Draft CCL 4 microbial contaminants (February 2015) (26)

Microbial contaminant
Adenovirus
Caliciviruses
Campylobacter jejuni
Enterovirus
Escherichia coli (0157)
Helicobacter pylori
Hepatitis A virus
Legionella pneumophila
Mycobacterium avium
Naegleria fowleri
Salmonella enterica
Shigella sonnei
Source: http://www2.epa.gov/ccl/microbial-contaminants-ccl-4.
Diseases and infections
Respiratory illness and occasionally gastrointestinal illness
Mild self-limiting gastrointestinal illness
Mild self-limiting gastrointestinal illness
Mild respiratory illness
Gastrointestinal illness and kidney failure
Capable of colonizing human gut, and can cause ulcers and cancer
Liver disease and jaundice
Lung diseases when inhaled
Lung infection in those with underlying lung disease, and disseminated
infection in the severely immunocompromised
Primary amoebic meningoencephalitis
Mild self-limiting gastrointestinal illness
Mild self-limiting gastrointestinal illness and bloody diarrhea

as documented by recent waterborne disease statistics. It is Comprehensive Reviews

unlikely that climate change will significantly increase the
risk of contracting a water-related illness. There is, however,
concern in areas in which the drinking water infrastructure
is older, as aging infrastructure is especially susceptible to fail-
ure (e.g., main breaks, low-pressure events) (37, 38). It is
believed that climate change will place additional stresses
on the capacity of drinking water treatment facilities (e.g.,
during extreme precipitation events the treatment capacity
may be exceeded) and may increase the risk of failure of or
damage to infrastructure for drinking water (37–39). If the
capacity of the drinking water treatment systems is exceeded
(thereby rendering insufficient pathogen removal via filtra-
tion and disinfection), and the distribution system is vulner-
able to intrusion or increased biofilm growth and release,
there could be an increase in the numbers of waterborne dis-
ease outbreaks in the future.
The largest documented waterborne disease outbreak
in U.S. history, which caused an estimate of more than
400,000 cases of illnesses and resulted in 50 deaths (40),
occurred in Milwaukee, Wisconsin, in 1993 and was preceded
by the heaviest rainfall event in 50 years in the adjacent
watersheds (41). Several other investigators have noted the
strong relationship between extreme precipitation and water-
borne disease (42–45).
As stated previously, in recent years, many of the water-
borne disease outbreaks in the United States have been asso-
ciated with the consumption of contaminated untreated
drinking water. Small groundwater systems are more likely
not to be disinfected than are surface water systems. Accord-
ing to the EPA, only about half of the community water sys-
tems and 20% of the noncommunity water systems practice
disinfection (46).
Untreated drinking water systems are particularly vulner-
able to contamination after extreme precipitation events
(47). For groundwater systems, the situation is exacerbated
if the well and wellhead are not properly constructed and/or
maintained. That was the situation in an outbreak in Walker-
ton, Ontario, Canada, in 2000. Heavy rains carried livestock
waste in runoff to a drinking water well that was not properly
operated and maintained. In this outbreak, more than 2000
people became ill and 7 died as a result of infection by
E. coli O157:H7 and/or Campylobacter (43, 44, 48).
A comprehensive study of the association between extreme
precipitation events and waterborne disease outbreaks in
the United States was performed by Curriero et al. (49).
Using the EPA waterborne disease database, they examined
548 reported outbreaks that occurred between 1948 and
1994. Precipitation data from the National Climatic
Data Center were used for the analysis. They found a statis-
tically significant relationship between reported water-
borne disease outbreaks and precipitation events: 51% of
the outbreaks were preceded by a precipitation event that
was above the 90th percentile ( p = 0.002) and 68% were
preceded by events above the 80th percentile ( p = 0.001).
Surface water–related outbreaks showed the strongest asso-
ciation with extreme precipitation during the month of the
outbreak; a 2-month lag applied to groundwater contamina-
tion events.
Cann et al. (50) analyzed reports of waterborne disease
obtained from several global databases. They examined
approximately 300 reports, between 1910 and 2010, 65 of
which met their criteria, in which an extreme weather event
(including flooding, extreme precipitation, and drought) pre-
ceded an outbreak. The most commonly reported pathogens
in these outbreaks were Vibrio spp. Among the parasites,
Cryptosporidium were the most commonly reported, while
noroviruses were the most commonly reported viruses to
occur in the waterborne outbreaks (see Table 4). They found
that 55.2% of the outbreaks were preceded by extreme precip-
itation events, and flooding preceded 52.9%. Vibrio were
the most common bacterial pathogens reported (21.6%), fol-
lowed by Leptospira spp. (12.7%).
An analysis of waterborne disease outbreaks in England
and Wales found that while some outbreaks were preceded
by extreme precipitation events, others were associated
with low cumulative rainfall (51). Therefore, drought, which
may cause pathogens present in water to become concen-
trated, may also be of concern.
Semenza et al. (52) conducted an analysis of food- and
waterborne disease outbreaks for the period 1998–2009 by
searching the PubMed and ScienceDirect databases. They
specifically searched for terms related to climate change and
food- and waterborne pathogens; in total more than 700
articles were used to generate more than 6000 data points.
3.1.7-10 ▪ WATER

TABLE 4 Waterborne pathogens implicated in outbreaks following extreme water-related

weather events identified from the scientific literature and ProMED reports (50)

No. (%)a of times reported

Waterborne pathogen
Scientific literature ProMED reports
Bacteria
Vibrio spp. 21 (28.4) 145 (68.7)
Vibrio cholerae 20 (27.0) 137 (64.9)
Other Vibrio spp. 2 (2.7) 8 (3.8)
Leptospira spp. 13 (17.6) 36 (17.1)
Campylobacter spp. 10 (13.5) 3 (1.4)
Escherichia coli 9 (12.2) 9 (4.3)
Shigella spp. 4 (5.4)
Salmonella spp. 3 (4.1) 5 (2.4)
Salmonella typhi 1 (1.4) 4 (1.9)
Salmonella sp. unknown 2 (2.7) 1 (0.5)
Burkholderia pseudomallei 3 (4.1) 9 (4.3)
Yersinia enterocolitica 2 (2.7)
Aeromonas spp. 1 (1.4)
Parasites
Cryptosporidium spp. 9 (12.2) 3 (1.4)
Giardia lamblia 5 (6.8)
Acanthamoeba spp. 1 (1.4)
Cyclospora spp. 1 (1.4)
Viruses
Hepatitis A virus 4 (5.4) 2 (0.9)
Hepatitis E virus 2 (2.7) 1 (0.5)
Norovirus 6 (8.1) 1 (0.5)
Rotavirus 3 (4.1) 1 (0.5)
Adenovirus 2 (2.7)
Enterovirus 1 (1.4)
a
Percentage of either 74 outbreak accounts or 211 ProMED reports reporting the pathogens involved.

Unlike the previously discussed studies, which examined the
association between climate and waterborne pathogens in
general, these researchers generated results showing asso-
ciations between individual climate-related factors and
individual pathogens. Selected results for some waterborne
pathogens are shown in Table 5.
In another study, Yang et al. (10) searched the
Global Infectious Disease and Epidemiology Network
for water-associated outbreaks reported between 1991 and
2008. They combined these data with a GIS database
containing georeferenced socioenvironmental information
including population density (2000), annual accumulated

TABLE 5 Selected pathogens with environmental/climatic variables (and factors associated with climate) from the climate change
knowledge (52)

Base for food- and waterborne diseases, 1998–2009

Variable Campylobacter Salmonella Vibrio Cryptosporidium Norovirus
Temperature Impact Impact Impact Impact Impact
Extreme temperature Impact impact unknown Impact Impact Impact unknown
Precipitation Impact Impact Impact Impact Impact unknown
Precipitation pattern Impact Impact unknown Impact unknown Impact Impact
Extreme precipitation Impact Impact unknown Impact Impact Impact
Seasonality Impact Impact Impact Impact Impact
Floods Impact Impact Impact Impact Impact
Drought Impact Impact unknown No impact Impact Impact unknown
Storms Impact unknown Impact unknown Impact Impact unknown Impact unknown

temperature, surface water area, and average annual precipita-
tion. Based on an analysis of more than 1400 reported out-
breaks, they concluded that outbreaks of water-associated
diseases are significantly correlated with socioenvironmental
factors. Specifically, they found that
• All water-associated diseases were significantly correlated
with population density
• Waterborne and water-related (e.g., vectorborne) out-
breaks are inversely related to average annual rainfall
• Water-related outbreaks are associated with accumulated
temperature (i.e., the sum [in degrees] by which the air
temperature rises above a given threshold value and the
number of days during which this increase is maintained)
• Water-washed diseases (e.g., conjunctivitis) are inversely
related to surface water area.
The model predicted that the risks of water-associated dis-
eases varied geographically:
• For waterborne diseases, West Europe, central Africa, and
north India were at higher risk
• For water-washed diseases, West Europe, north Africa, and
Latin America were at higher risk
• For water-based diseases, risks were higher in east Brazil,
northwest Africa, central Africa, and southeast of China
• For water-related diseases, central Africa ( particularly
particular Ethiopia and Kenya) and north India were at
higher risk
• For water-dispersed diseases, West Europe was at
higher risk.
Climate Change and Enteric Diseases
It is well documented that there are seasonal patterns in infec-
tion by enteric viruses. For example, enterovirus infections in
the United States peak in summer and fall (53). Rotavirus
infections tend to peak during cooler, drier months (54).
In Singapore, it has been shown that the incidence of
hand, foot, and mouth disease (caused by coxsackievirus
A16 and enterovirus 71) has a linear relationship with tem-
perature, with a rapid rise in incidence when the temperature
exceeds 32°C (55). Rates of diarrhea have been associated
with high temperatures (56).
Rotavirus is of particular concern, as on a worldwide basis
in 2008, this virus caused about 450,000 deaths in children
younger than 5 years (57). Therefore, Moors et al. (58) con-
ducted an in-depth analysis on the impacts of climate change
in northern India for rotavirus, which is responsible for
approximately 39% of diarrhea cases worldwide (59, 60),
and annually causes approximately 122,000–153,000 deaths
in children below the age of 5 in India (61). The impact
of increasing or decreasing precipitation was negligible. Inter-
estingly, the impact of increasing temperature shows that diar-
rhea incidences caused by rotavirus might decrease by 1–2%
because of increasing temperatures in the Ganges region
during the winter period. The average effect of humidity on
rotavirus diarrhea (an increase of 1.5%) would cancel the
decrease due to temperature.
Chou et al. (62) studied the relationship between several
climate factors (relative humidity, maximum temperature,
and the numbers of extreme rainfall events) and the numbers
of diarrhea cases in Taiwan between 1996 and 2007. They
found that the climate actors had different effects on the inci-
dence of diarrhea among different age groups. The impact
of maximum temperature on diarrhea cases was strongest in
children (0–14 years) and older adults (40–64 years) and
3.1.7. Drinking Water Microbiology ▪ 3.1.7-11
had less of an effect on adults between the ages of 15 and
39 years. Relative humidity and extreme rainfall days also
contributed to the incidence of diarrhea, although the effects
were smaller.
Martin et al. (63) found that temperature is associated
with risk of enteric disease in Arctic communities, due to
the melting of permafrost that transports sewage into ground-
water, drinking water sources, or other surface waters (63).
It is also believed that thawing may damage drinking water
intake systems (64).
Rainfall has also been associated with enteric infections
(65). Higher concentrations of enteric viruses have been
reported frequently in drinking water and recreational water
following heavy rainfall (66). As discussed previously, heavy
rains preceded the outbreaks in Walkerton and Milwaukee.
A number of studies have attempted to project the impacts
of anticipated future climate change on waterborne disease in
the mid-21st century or beyond. Moors et al. (58) examined
the relationship between four climate factors (temperature,
increase precipitation/extreme rainfall, decreased preci-
pitation/drought, and air humidity) and the incidence of diar-
rhea in northern India. They projected diarrhea from all
causes by the year 2040. They attributed the increase mainly
to increased temperature (10.1%), with smaller effects due to
humidity (1.8%), increased rainfall (0.14%), and decreased
rainfall (1.1%).
Kolstad and Johansson (56) projected an increase of
8–11% in the risk of diarrhea in the tropics and subtropics
in 2039 due to climate change, using the A1B scenario devel-
oped by the IPCC (which assumes very rapid economic
growth, global population that peaks in mid-century and
then declines, the rapid introduction of new and more effi-
cient technologies, and energy sources that are balanced
[i.e., not relying too heavily on one particular energy source]).
Patz et al. (41) performed a very localized projection for the
city of Chicago, predicting that sewage overflows into the
city’s watersheds would increase by 50–120% by 2100 as a
result of more frequent and intense rainfall. Alexander et al.
(67) predicted an increase in the incidence of diarrhea in
Botswana as a consequence of hot, dry conditions beginning
earlier in the year and lasting for a longer time. They also
predicted a decrease in the incidence of diarrhea in the wet
season.
Climate Change and Vibrio spp.
Perhaps one of the earliest and best studied relationships
between climate change and microorganisms is that for Vibrio
spp. Vibrio is a genus of bacteria native to the marine environ-
ment. There are several pathogenic species of Vibrio, the most
well known of which is toxigenic V. cholera, the causative
agent of cholera. According to the CDC, “Cholera is an acute
intestinal infection causing profuse watery diarrhea, vomit-
ing, circulatory collapse and shock. Many infections are asso-
ciated with milder diarrhea or have no symptoms at all. If
left untreated, 25–50% of severe cholera cases can be fatal.”
Cholera is transmitted by consumption of contaminated
water or food; large epidemics have occurred as a result of fecal
contamination of water supplies (68).
In 1996, Colwell published one of the earliest papers
describing the relationship between climate factors and infec-
tious disease, using cholera as the example. She graphically
showed the relationship between the incidence of cholera
in Bangladesh and the mean sea surface (69) temperature
over the course of 1994. This relationship has been further
substantiated in other countries in which cholera is endemic
(70–72). A study of DNA in water samples collected over a
3.1.7-12 ▪ WATER
TABLE 6 Influence of environment, climate, and weather on cholera (74)
Factor Climate driver
Temperature Seasons, interannual variability
Salinity Seasons, monsoons, ENSO, sea level rise
Sunlight Seasons, monsoons, interannual variability
pH Seasons, interannual variability ( phytoplankton
growth)
Fe3+ Precipitation (runoff ), atmospheric deposition
Products of algal Seasons, monsoons, interannual variability in
growth light, nutrients
Chitin Seasons, monsoons, zooplankton blooms
44-year period established an unequivocal positive relation-
ship between Vibrio numbers and sea surface temperature in
the North Sea (73).
The relationship between V. cholera and environmental
factors is complex, involving abiotic factors, such as salinity,
nutrients, and pH; climate factors, such as rainfall and tem-
perature; and biotic factors, such as the type and presence of
marine plants, phytoplankton, zooplankton, and crustaceans.
These factors were discussed by Lipp et al. (74) and are sum-
marized in Table 6.
This chapter has reviewed the pathogens involved in
waterborne disease, as well as many of the possible reasons
for outbreaks and the latest literature on the subject. It is clear
that even under the best circumstances, outbreaks of disease
will continue to occur. The best drinking water treatment
plants are still subject to occasional problems: a rapid increase
in the turbidity of influent waters or environmental fluc-
tuations due to climate change that affect treatment process
efficacy, for example. Therefore, it would not be realistic to
expect a 0% probability of risk from the consumption of
drinking water. In addition, the prevalence of enteric diseases
in the population and the incidence of new cases may not
necessarily be related to the microbiological water quality or
the quality of treatment, but to unforeseen events due to con-
tamination after treatment such as bacterial proliferation in
point-of-use treatment devices, by food handlers, and so on.
Most cases of gastroenteritis are not reported, and thus we
are just looking at the tip of the iceberg; it may be more effec-
tive to take a more proactive approach to protect public
health through education. Regardless of the best preventive
measures, a relatively high prevalence of enteric diseases is
to be expected in any given population. It is natural to focus
on reducing disease that is transmitted through drinking
water, as we have a great deal of control over how extensively
it is treated to reduce microbial risks. We can continue to
reduce those risks by taking a more holistic approach to man-
agement of our source waters and the surrounding watersheds
to protect from microbiological contamination with anthro-
ponotic and zoonotic enteric pathogens.
REFERENCES
1. United Nations. 2000. Resolution adopted by the General
Assembly [without reference to a Main Committee (A/55/
L.2)]. 55/2. United Nations Millennium Declaration.
2. World Health Organization. 2014. Progress on Drinking
Water and Sanitation: 2014 Update. WHO Press, Geneva,
Switzerland.
Influence
Growth of V. cholerae, phytoplankton blooms
Growth of V. cholerae, expression of cholera toxin
Survival of V. cholerae, phytoplankton blooms, induction of
toxin-inducing phage
Growth of V. cholerae
Growth of V. cholerae, expression of cholera toxin
Survival of V. cholerae
Growth of V. cholerae, attachment to exoskeletons
3. Craun GF, Brunkard JM, Yoder JS, Roberts VA, Carpen-
ter J, Wade T, Calderon RL, Roberts JM, Beach MJ, Roy
SL. 2010. Causes of outbreaks associated with drinking water
in the United States from 1971 to 2006. Clin Microbiol Rev
23:507–528.
4. US Environmental Protection Agency. 2015. Private
Drinking Water Wells. http://water.epa.gov/drink/info/
well/index.cfm.
5. US Environmental Protection Agency. 2015. Public
Drinking Water Systems Programs: Overview. http://water.
epa.gov/infrastructure/drinkingwater/pws/index.cfm.
6. Beer KD, Gargano JW, Roberts VA, Hill VR, Garrison
LE, Kutty PK, Hilborn ED, Wade TJ, Fullerton KE, Yoder
JS. 2015. Surveillance for waterborne disease outbreaks asso-
ciated with drinking water — United States, 2011–2012.
MMWR Morb Mortal Wkly Rep 64:849–851.
7. Reynolds KA, Mena KD, Gerba CP. 2008. Risk of water-
borne illness via drinking water in the United States. Rev
Environ Contam Toxicol 192:117–158.
8. White GF, Bradley DJ, White AU. 1972. Drawers of Water:
Domestic Water Use in East Africa. University of Chicago
Press, Chicago, IL.
9. IOM (Institute of Medicine). 2009. The spectrum of water-
related disease transmission processes. In Global Issues in
Water, Sanitation, and Health. National Academies Press,
Washington, DC.
10. Yang K, LeJeune J, Alsdorf D, Lu B, Shum CK, Liang S.
2012. Global distribution of outbreaks of water-associated
infectious diseases. PLoS Negl Trop Dis 6:e1483.
11. Bartram J. 2015. Bradley classification of disease transmis-
sion routes for water-related hazards. In Bartram J (ed.),
Routledge Handbook of Water and Health. Routledge,
New York, NY.
12. Yoder JS, Blackburn BG, Craun GF, Hill V, Levy DA,
Chen N, Lee SH, Calderon RL, Beach MJ. 2004. Surveil-
lance for waterborne-disease outbreaks associated with
recreational water—United States, 2001–2002. MMWR
Surveill Summ 22:1–22.
13. World Health Organization. 2011. Microbial fact sheets.
In Guidelines for Drinking-Water Quality, 4th ed. WHO Press,
Geneva, Switzerland.
14. World Health Organization. 2011. Microbial aspects. In
Guidelines for Drinking-Water Quality, 4th ed. WHO Press,
Geneva, Switzerland.
15. Moe CL. 2007. Waterborne transmission of infectious
agents, p. 222–248. In Hurst CJ, Crawford RL, Garland JL,
Lipson DA, Mills AL, Stetzenbach LD (eds.), Manual
of Environmental Microbiology, 3rd ed, ASM Press,
Washington, DC.
16. Yates MV. 2014. Astroviruses, p. 479–492. In Percival SL,
Yates MV, Williams DL, Chalmers RM, Gray NF (eds.),

Microbiology of Waterborne Diseases, 2nd ed, Elsevier, Lon-
don, UK.
17. Chalmers RM. 2014. Cyclospora, p 327–354. In Percival
SL, Yates MV, Williams DL, Chalmers RM, Gray NF
(eds.), Microbiology of Waterborne Diseases, 2nd ed, Elsevier,
London, UK.
18. Committee on Indicators for Waterborne Pathogens. 2004.
Indicators for Waterborne Pathogens. National Academies
Press, Washington, DC.
19. World Health Organization. 2006. Recommendations, 3rd
ed. WHO Press, Geneva, Switzerland. Guidelines for
drinking-water quality: incorporating first addendum, 1.
20. World Health Organization and UNICEF. 2012. Pneumo-
nia and diarrhoea: tackling the deadliest diseases for the world’s
poorest children. Statistics and Monitoring Section, Division
of Policy and Strategy. New York, NY.
21. World Health Organization. 2009. Rotavirus infections.
http://www.who.int/nuvi/rotavirus/en/.
22. World Health Organization. 2012. Immunization Monthly
Update in the African Region. http://www.afro.who.int/en/
tanzania/tanzania-publications/2486-rotavirus-surveillance-
in-the-african-region.html.
23. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV,
Widdowson MA, Roy SL, Jones JL, Griffin PM. 2011.
Foodborne illness acquired in the United States—major
pathogens. Emerg Infect Dis 17:7–15.
24. Colford JM Jr , Roy S, Beach MJ, Hightower A, Shaw SE,
Wade TJ. 2006. A review of household drinking water inter-
vention trials and an approach to the estimation of endemic
waterborne gastroenteritis in the United States. J Water
Health 4(Suppl 2):71–88.
25. Messner M, Shaw S, Regli S, Rotert K, Blank V, Soller J.
2006. An approach for developing a national estimate
of waterborne disease due to drinking water and a national
estimate model application. J Water Health 4(Suppl 2):
201–240.
26. US Environmental Protection Agency. 2015. Draft Candi-
date Contaminant List 4—CCL4. http://www2.epa.gov/ccl/
draft-contaminant-candidate-list-4-ccl-4.
27. Intergovernmental Panel on Climate Change. 2013. Cli-
mate change 2013: the physical science basis. In Stocker
TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung
J, Nauels A, Xia Y, Bex V, Midgley PM (eds.), Contribution
of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, UK.
28. Intergovernmental Panel on Climate Change. 2014:
Climate change 2014: synthesis report. In Pachauri RK,
Meyer LA (eds.), Contribution of Working Groups I, II and
III to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change. IPCC, Geneva, Switzerland.
29. Luber G, Knowlton K, Balbus J, Frumkin H, Hayden M,
Hess J, McGeehin M, Sheats N, Backer L, Beard CB,
Ebi KL, Maibach E, Ostfeld RS, Wiedinmyer C, Zielinski-
Gutiérrez E, Ziska L. 2014. Human health. In Melillo JM,
Richmond TC, Yohe GW (eds.), Climate change impacts in
the United States: the third national climate assessment. U.S.
Global Change Research Program, Published by the U.S.
Government Printing Office, Washington DC, 220–256.
30. ECDC. 2012. Assessing the potential impacts of climate change
on food- and waterborne diseases in Europe. European Centre
for Disease Prevention and Control, Stockholm.
31. Semenza JC, Suk JE, Estevez V, Ebi KL, Lindgren E. 2012.
Mapping climate change vulnerabilities to infectious dis-
eases in Europe. Environ Health Perspect 120:385–392.
32. Soller JA, Bartrand T, Ashbolt NJ, Ravenscroft J, Wade
TJ. 2010. Estimating the primary etiologic agents in recrea-
tional freshwaters impacted by human sources of faecal
contamination. Water Res 44:4736–4747.
3.1.7. Drinking Water Microbiology ▪ 3.1.7-13
33. Brunkard JM, Ailes E, Roberts VA, Hill V, Hilborn ED,
Craun GF, Rajasingham A, Kahler A, Garrison L, Hicks
L, Carpenter J, Wade TJ, Beach MJ, Yoder Msw JS,
CDC. 2011. Surveillance for waterborne disease outbreaks
associated with drinking water—United States, 2007–
2008. MMWR Surveill Summ 60:38–68.
34. Bradbury KR, Borchardt MA, Gotkowitz M, Spencer SK,
Zhu J, Hunt RJ. 2013. Source and transport of human
enteric viruses in deep municipal water supply wells. Environ
Sci Technol 47:4096–4103.
35. Uejio CK, Yale SH, Malecki K, Borchardt MA,
Anderson HA, Patz JA. 2014. Drinking water systems,
hydrology, and childhood gastrointestinal illness in
central and northern Wisconsin. Am J Public Health
104:639–646.
36. Drayna P, McLellan SL, Simpson P, Li S-H, Gorelick MH.
2010. Association between rainfall and pediatric emergency
department visits for acute gastrointestinal illness. Environ
Health Perspect 118:1439–1443.
37. Beach MJ, Roy S, Brunkard J, Yoder J, Hlavsa MC.
2009. Vulnerable infrastructure and waterborne disease
risk, p. 159–169. In National Research Council (ed.), Global
Issues in Water, Sanitation, and Health: Workshop Summary.
National Academies Press, Washington DC.
38. Levin RB, Epstein PR, Ford TE, Harrington W, Olson E,
Reichard EG. 2002. U.S.drinking water challenges in the
twenty-first century. Environ Health Perspect 110(Suppl 1):
43–52.
39. Whitehead PG, Wilby RL, Battarbee RW, Kernan M,
Wade AJ. 2009. A review of the potential impacts of climate
change on surface water quality. Hydrol Sci Journal 54:
101–123.
40. Hoxie NJ, Davis JP, Vergeront JM, Nashold RD, Blair
KA. 1997. Cryptosporidiosis-associated mortality following
a massive waterborne outbreak in Milwaukee, Wisconsin.
Am J Public Health 87:2032–2035.
41. Patz JA, Vavrus SJ, Uejio CK, McLellan SL. 2008. Climate
change and waterborne disease risk in the Great Lakes region
of the U.S. Am J Prev Med 35:451–458.
42. Jofre J, Blanch AR, Lucena F. 2010. Water-Borne infec-
tious disease outbreaks associated with water scarcity and
rainfall events. Water Scarc Mediterr Perspect Under Glob
Change 8:147–159.
43. Auld H, MacIver D, Klaassen J. 2004. Heavy rainfall and
waterborne disease outbreaks: the Walkerton example.
J Toxicol Environ Health A 67:1879–1887.
44. Salvadori MI, Sontrop JM, Garg AX, Moist LM, Suri RS,
Clark WF. 2009. Factors that led to the Walkerton tragedy.
Kidney Int Suppl 75:S33–S34.
45. Fong T-T, Mansfield LS, Wilson DL, Schwab DJ, Molloy
SL, Rose JB. 2007. Massive microbiological groundwater
contamination associated with a waterborne outbreak in
Lake Erie, South Bass Island, Ohio. Environ Health Perspect
115:856–864.
46. US Environmental Protection Agency. 1990. Strawman
rule for ground water disinfection. http://nepis.epa.gov/
Exe/ZyPURL.cgi?Dockey=10003I0N.txt.
47. Kozlica J, Claudet AL, Solomon D, Dunn JR, Carpenter
LR. 2010. Waterborne outbreak of Salmonella I 4,[5],
12:i:-. Foodborne Pathog Dis 7:1431–1433.
48. Clark CG, Price L, Ahmed R, Woodward DL, Melito PL,
Rodgers FG, Jamieson F, Ciebin B, Li A, Ellis A. 2003.
Characterization of waterborne outbreak-associated Campy-
lobacter jejuni, Walkerton, Ontario. Emerg Infect Dis 9:
1232–1241.
49. Curriero FC, Patz JA, Rose JB, Lele S. 2001. The associa-
tion between extreme precipitation and waterborne disease
outbreaks in the United States, 1948–1994. Am J Public
Health 91:1194–1199.
3.1.7-14 ▪ WATER
50. Cann KF, Thomas DR, Salmon RL, Wyn-Jones AP, Kay
D. 2013. Extreme water-related weather events and water-
borne disease. Epidemiol Infect 141:671–686.
51. Nichols G, Lane C, Asgari N, Verlander NQ, Charlett A.
2009. Rainfall and outbreaks of drinking water related dis-
ease and in England and Wales. J Water Health 7:1–8.
52. Semenza JC, Herbst S, Rechenburg A, Suk JE, Höser C,
Schreiber C, Kistemann T. 2012. Climate change impact
assessment of food- and waterborne diseases. Crit Rev Environ
Sci Technol 42:857–890.
53. Khetsuriani N, Lamonte-Fowlkes A, Oberst S, Pallansch
MA, Centers for Disease Control and Prevention.
2006. Enterovirus surveillance—United States, 1970–
2005. MMWR Surveill Summ 55(SS–8):1–20.
54. D’Souza RM, Hall G, Becker NG. 2008. Climatic
factors associated with hospitalizations for rotavirus diar-
rhoea in children under 5 years of age. Epidemiol Infect
136:56–64.
55. Hii YL, Rocklöv J, Ng N. 2011. Short term effects of weather
on hand, foot and mouth disease. PLoS One 6:e16796.
56. Kolstad EW, Johansson KA. 2011. Uncertainties associated
with quantifying climate change impacts on human health:
a case study for diarrhea. Environ Health Perspect 119:
299–305.
57. Tate JE, Burton AH, Boschi-Pinto C, Steele AD, Duque J,
Parashar UD, WHO-Coordinated Global Rotavirus
Surveillance Network. 2012. 2008 estimate of worldwide
rotavirus-associated mortality in children younger than 5
years before the introduction of universal rotavirus vaccina-
tion programmes: a systematic review and meta-analysis.
Lancet Infect Dis 12:136–141.
58. Moors E, Singh T, Siderius C, Balakrishnan S, Mishra A.
2013. Climate change and waterborne diarrhoea in northern
India: impacts and adaptation strategies. Sci Total Environ
(Suppl:S139–51):468–469.
59. Jagai JS, Sarkar R, Castronovo D, Kattula D, McEntee J,
Ward H, Kang G, Naumova EN. 2012. Seasonality of
rotavirus in South Asia: a meta-analysis approach assessing
associations with temperature, precipitation, and vegetation
index. PLoS One 7:e38168.
60. Levy K, Hubbard AE, Eisenberg JN. 2009. Seasonality of
rotavirus disease in the tropics: a systematic review and meta-
analysis. Int J Epidemiol 38:1487–1496.
61. Tate JE, Panozzo CA, Payne DC, Patel MM, Cortese MM,
Fowlkes AL, Parashar UD. 2009. Decline and change in
seasonality of US rotavirus activity after the introduction
of rotavirus vaccine. Pediatrics 124:465–471.
62. Chou WC, Wu J-L, Wang Y-C, Huang H, Sung F-C,
Chuang C-Y. 2010. Modeling the impact of climate variabil-
ity on diarrhea-associated diseases in Taiwan (1996–2007).
Sci Tot Environ 409:43–51.
63. Martin D, Belanger B, Gosselin P, Brazeau J, Furgal C,
Dery S. 2007. Drinking water and potential threats to
human health in Nunavik: adaptation strategies under
climate change conditions. Arctic 60:195–202.
64. Hess JJ, Malilay JN, Parkinson AJ. 2008. Climate
change: the importance of place. Am J Prev Med 35:
468–478.
65. Smith KR, Woodward A, Campbell-Lendrum D, Chadee
DD, Honda Y, Liu Q, Olwoch JM, Revich B, Sauerborn
R. 2014. Human health: impacts, adaptation, and
co-benefits, p. 709–754. In Field CB, Barros VR, Dokken
DJ, Mach KJ, Mastrandrea MD, Bilir TE, Chatterjee M,
Ebi KL, Estrada YO, Genova RC, Girma B, Kissel ES,
Levy AN, MacCracken S, Mastrandrea PR, White LL
(eds.), Climate Change 2014: Impacts, Adaptation, and Vul-
nerability. Part A: Global and Sectoral Aspects. Contribution
of Working Group II to the Fifth Assessment Report of the Inter-
governmental Panel on Climate Change. Cambridge University
Press, Cambridge, UK.
66. Delpla I, Jung AV, Baures E, Clement M, Thomas O.
2009. Impacts of climate change on surface water quality
in relation to drinking water production. Environ Int 35:
1225–1233.
67. Alexander KA, Carzolio M, Goodin D, Vance E. 2013.
Climate change is likely to worsen the public health threat
of diarrheal disease in Botswana. Int J Environ Res Public
Health 10:1202–1230.
68. CDC. 2015. Cholera - Vibrio cholerae infection. http://www.
cdc.gov/cholera/infection-sources.html.
69. Colwell RR. 1996. Global climate and infectious disease:
the cholera paradigm. Science 274:2025–2031.
70. Islam MS, Sharker MA, Rheman S, Hossain S, Mahmud
ZH, Islam MS, Uddin AM, Yunus M, Osman MS, Ernst
R, Rector I, Larson CP, Luby SP, Endtz HP, Cravioto A.
2009. Effects of local climate variability on transmission
dynamics of cholera in Matlab, Bangladesh. Trans R Soc
Trop Med Hyg 103:1165–1170.
71. Paz S. 2009. Impact of temperature variability on cholera
incidence in southeastern Africa, 1971–2006. Eco Health
6:340–345.
72. Reyburn R, Kim DR, Emch M, Khatib A, von Seidlein L,
Ali M. 2011. Climate variability and the outbreaks of chol-
era in Zanzibar, East Africa: a time series analysis. Am J Trop
Med Hyg 84:862–869.
73. Vezzulli L, Brettar I, Pezzati E, Reid PC, Colwell RR,
Höfle MG, Pruzzo C. 2012. Long-term effects of ocean
warming on the prokaryotic community: evidence from
the vibrios. ISME J 6:21–30.
74. Lipp EK, Huq A, Colwell RR. 2002. Effects of global cli-
mate on infectious disease: the cholera model. Clin Microbiol
Rev 15:757–770.
 Introduction to Aerobiology
PAULA KRAUTER AND LINDA D. STETZENBACH
3.2.1
Aerobiology encompasses the study of airborne human
pathogens, plant pathogens, plant products, opportunistic
and nonpathogenic organisms, and aerosolized microbial
by-products. The organisms of interest include airborne
culturable and nonculturable bacteria, saprophytic fungi,
free-living parasites, viruses, and algae. Microbial metabo-
lites, toxins, and fragments of microbial agents that may
result in adverse health effects or environmental impact are
also contained within the field of aerobiology. Other aspects
of aerobiology include the study of the source, dispersion, and
effects of airborne biomaterials.
This chapter introduces the study of airborne micro-
organisms and their by-products, discusses indoor and an-
thropogenic outdoor sources of airborne microorganisms
that affect human health and the environment, reviews the
association of bioaerosols and indoor environmental quality,
presents background information on airborne microorgan-
isms as potential bioterrorism agents, and describes categories
of microbial agents common within bioaerosols. Subsequent
chapters are focused on specific topics concerning airborne
microorganisms, including sampling methods, analysis, fate,
and transport; fungi and mycotoxins; bacteria and endo-
toxins; Legionellae; viruses; and agricultural pathogens.
BIOAEROSOLS: COMPOSITION
AND SOURCES
Bioaerosols are suspensions of airborne particles that contain
living organisms or derive from living organisms. These par-
ticles or droplets may include fungal spores, fungal mycelium
fragments, pollen, bacteria, viruses, or protozoa (e.g., amoe-
bas). The particles can be fragments of plant materials,
algae, or mammalian (skin scales, hair, or other components)
excreta, or fragments of insects (dust mites). Microbial metab-
olites may include volatile organic compounds or residues
or products of organisms, such as bacterial lipopolysacchar-
ides, endotoxins, or fungal mycotoxins. Most bioaerosols
are harmless constituents of normal environments; how-
ever, some bioaerosol particles may be infectious agents or
allergens, or they may carry toxic or irritant components or
metabolites. To be infectious, an organism must be viable,
but to cause allergenic or toxic effects viability is not a pre-
requisite. Dead cells as well as cell residues may affect human
health.
doi:10.1128/9781555818821.ch3.2.1
3.2.1-1
A variety of bioaerosol sources are found in the environ-
ment, including fresh and marine surface waters, soil, and
plants. Natural bioaerosols can cause health problems—
for example, some viruses and bacteria can cause infection
or allergies. Bioaerosols include different compositions based
on their source and environmental conditions. One com-
mon constituent is endospores that are typically resistant to
environmental stressors. Air currents can carry away these
particles due to their single-cell size (0.5–30 μm). Another
example of a bioaerosol source is that of Legionella species,
which may be found in water. Some species are found in
warm-water systems and, if aerosolized, can cause disease in
humans (1). Actinomycetes are a type of bacteria that live
in soil and may form spores. Their spores may contribute
to occupational exposure in agricultural workers (2). Other
sources are bacterial endotoxins (lipopolysaccharides speci-
fic to the cell wall of Gram-negative bacteria) that are found
in agricultural environments, some industries, and in humidi-
fied indoor air (3).
In general, bioaerosols are generated as a mixture of par-
ticles or droplets, consisting of different sizes ranging from
0.5 to 30 μm in diameter (4) but may vary from 0.01 up to
100 μm (5). Fragmentation, agglomeration, or formation of
droplet nuclei can alter the size of the bioaerosol particle.
Bioaerosols that are generated from water sources during
splash and wave action are usually formed with a thin layer
of moisture surrounding the microbial cell and often consist
of aggregates of several organisms in a packet (6). Some
bioaerosols can include viable but nonculturable bacteria
(VBNC). For more than a decade, VBNCs have been
investigated and appear to be common to many bacteria,
especially those in aquatic habitats (7). Among bacteria
that may become VBNCs, some are significant human patho-
gens. New methodologies have advanced the study of VBNCs
(see chapters 3.2.2 and 3.2.3 for detailed information).
Bioaerosols released into the air from soil or dry surfaces
are often composed of single cells or units, or the microorgan-
isms are associated with particles, a condition referred to as
“rafting” (8). Aerosols generated from soil are characterized
by the type of soil (9) or by the presence of vegetation (10).
The transport and ultimate settling of a bioaerosol is affected
by its physical properties and the environmental conditions it
encounters while airborne. The most important physical
characteristics are the size, density, and shape of the droplets
3.2.1-2 ▪ AIR
or particles while the most significant environmental condi-
tions are the temperature, relative humidity, and magni-
tude of air currents (11–14). Temperature and relative
humidity also contribute to the generation of airborne
microorganisms as increased concentrations of some fungal
spores (e.g., Nigrospora, Cladosporium) in outdoor air (15)
and increased numbers of bacteria released from plant surfa-
ces (10) have been associated with high temperatures and
low relative humidity. An annual and diurnal variation
in the concentration of airborne microorganisms outdoors
has been observed and correlated with environmental condi-
tions (11).
The survival of airborne microorganisms is also influenced
by environmental factors. Survival of microorganisms in
the air depends on several factors, including the intensity of
ultraviolet and possibly other microbicidal light waves in
sunlight. Harsh environmental conditions tend to decrease
the numbers of viable airborne organisms, but there is varia-
bility in survival between groups of microorganisms and
within genera. In general, fungal spores, enteric viruses, and
amoebic cysts are more resistant to environmental stresses
encountered during transport through the air. Bacteria and
algae are more susceptible to stresses, although bacterial
endospores (e.g., Bacillus spp., Gram-positive cocci) are quite
resistant to environmental conditions (16). It is generally
believed that microbial numbers do not increase during
transport, but a doubling of airborne bacterial cell numbers
was demonstrated in a laboratory setting using a rotating
drum aerosol chamber with saturated humidity conditions
and tryptone added to the cell suspension prior to aero-
solization (17). A detailed discussion of the factors affect-
ing the survival and transport of bioaerosols is presented in
chapter 3.2.4.
Air is the mode of transport for the dispersal of bioaerosols
from one location to another, and the composition and con-
centration of bioaerosols vary with the source(s) and the dis-
persal mechanism (13). The movement of a particle is
governed by the principles of gravitation, electromagnetism,
turbulence, and diffusion. The kinds and numbers of micro-
organisms in the air vary with the environment, for example,
meteorological conditions and the type of terrain influence
microorganisms in outside air. Seasonal variations were
observed for some bioaerosol components (18). Furthermore,
TABLE 1 Adverse effects associated with exposure to airborne microorganisms
Microbial Associated adverse effect on
agent Human health
Algae Allergic reactions
Bacteria Hypersensitivity pneumonitis infection, mucous
membrane irritation
Endotoxin Cough, headache, fever, malaise, muscle aches,
nausea, respiratory distress
Fungi Allergic reactions, exacerbation of asthma,
dermal irritation, hypersensitivity pneumonitis,
infections, mucous membrane irritation
Mycotoxin Headache, muscle problems, neurologic disorders,
respiratory distress, toxicosis
Protozoa Encephalitis, hypersensitivity reactions, infections
Virus Infections
quantity and type of bacteria varies by location and time of
day (19).
Airborne microorganisms can transmit disease from one
region to another and impact agriculture and human health.
Microorganisms as bioaerosols may result in the spread of
plant diseases and losses of agricultural productivity and can
impact worker health (Table 1). For example, a study to assess
airborne microorganisms, endotoxin, and (1→3)β-D-glucan
exposure in greenhouses indicated that workers could be
exposed to elevated levels of inhalable culturable microor-
ganisms (20). Microbial aerosols from anthropogenic and
environmental sources in some cases may be associated
with public health concerns (Table 1; (21–29)). A detailed
discussion of airborne agricultural pathogens is presented in
chapter 3.2.8.
Respiratory symptoms and lung function impairment
are among the most important bioaerosol-associated health
effects. Bioaerosols that cause human disease have specific
chemical composition, biological characteristics, particle
size distribution, and inhalation concentration requirements.
The average human inhales approximately 10 m3 of air
per day (13). Large airborne particles (>10 μm) are lodged
in the upper respiratory tract (nose and nasopharynx) (30,
196). Particles <5 μm in diameter are removed by sneezing
and blowing or wiping the nose, and those particles deposited
in the pharynx (2–5 μm in diameter) are removed to the
pharynx by mucociliary action and then swallowed (31,
32). Particles 1–5 μm in diameter can be transported to the
lung, but the greatest retention in the alveoli are the 1–
2 μm particles (33, 34). Exhaled droplets of healthy people
have been measured from 0.3 to 8 μm although few droplets
were >2 μm (35).
An example of a bioaerosol that can affect the respiratory
system is generated from the red tide blooms (RTBs) that
periodically occurs in the Gulf Coast (USA). The RTB is
generally a large bloom of algae that is often dangerous to
local aquatic life. When the algal cells are broken up by the
energy associated with the waves, the toxin is released into
the air as nonviable bioaerosols. Karenia brevis (dinoflagel-
late) produces neurotoxins. Droplets containing the toxin
can cause extreme eye and respiratory irritation to beachgoers,
boaters, and coastal residents, particularly those with
compromised respiratory systems (36). Nonviable agents of
Environment
Odor problems
Deterioration of building materials, loss of agricultural
productivity (crop and livestock diseases), odor problems
None reported
Deterioration of building materials, loss of agricultural
productivity (crop and livestock diseases), odor problems
Loss of agricultural productivity (disease in livestock)
May afford protection from biocide treatment to other microorganisms,
loss of agricultural productivity (disease in livestock)
Loss of agricultural productivity (crop and livestock diseases)

bioaerosols such as toxins and antigens can induce sensitiza-
tion or toxic diseases (36). Another example of a bioaerosol
inhalation hazard occurs during intense and continuous
exposure of dairy farmers to airborne Saccharopolyspora recti-
virgula. Workers may be at risk to develop S. rectivirgula–
specific humoral immune response (37).
SOURCES: MAN-MADE BIOAEROSOLS IN
OUTDOOR ENVIRONMENTS
Numerous anthropogenic activities serve as the origin of
bioaerosols in outdoor environments, especially agricultural
practices and wastewater treatment processes (Table 2).
Other sources of bioaerosols include industrial operations,
such as those at sawmills, livestock feed lots, metal industrial
sites, and commercial horticulture sites (20, 38). Increases in
airborne concentrations of microorganisms during harvest-
ing operations have been documented (20, 39, 40), and
several investigators have reported the presence of airborne
bacteria and viruses resulting from wastewater treatment
(41, 42), sanitary landfill operations (43), and reuse-water
irrigation practices (44, 45). Recycling facilities also generate
bioaerosols (46).
The release of biotechnology products (e.g., genetically
engineered microorganisms [GEMs], microbial pest control
agents [MPCAs]) developed to enhance agricultural pro-
ductivity, mineral recovery, oil spill cleanup, and toxic waste
disposal can also be a source of airborne microorganisms.
The application of biotechnology products to crops by aerosol
mist increases the possibility for transport of the microbial
product from the target crop to surrounding areas. Aerosolized
genetically modified cells have been monitored in a barn
setting (47) and at biotechnology-based fermenters (48).
SOURCES: BIOAEROSOLS AND INDOOR
ENVIRONMENTS
Deterioration of building materials, offensive odors, and
adverse human health effects are associated with microbial
contamination of indoor environments. Residences, offices,
schools, health care facilities, enclosed agricultural struc-
tures (e.g., barns and crop storage areas), pharmaceutical
and industrial facilities, food processing plants, and recycling
facilities are among the indoor environments where airborne
microorganisms have been studied (Table 2). Sources and
reservoirs of microorganisms are present within these settings,
including building materials and furnishings, pets, plants,
and air conditioning systems (8, 49). Wallboard, ceiling tiles,
carpeting and vinyl flooring, painted surfaces, upholstery
and drapery, wallpaper, plastics, wood, cement, and bricks
are examples of building materials and components that can
serve as sites for microbial colonization and subsequent
dispersal into the air (Table 2).
Bacteria and algae generally grow in areas with standing
water, such as air-handling system components (e.g., water
spray humidification systems, condensate pans), and sites
where water intrusion or leaking (e.g., flooding, condensa-
tion) has occurred. However, with the exception of viruses
that require a living host cell for replication, microorganisms
will colonize virtually any surface where there is sufficient
moisture. Fungi, which have lower water activity (aw) require-
ments than other microorganisms, tend to colonize a wide
variety of building materials (126). Penicillium spp. and Asper-
gillus versicolor are primary colonizers that are often isolated on
wallpaper and drier margins of wetted walls, while Cladospo-
rium spp. proliferate as secondary colonizers. With variation
3.2.1. Introduction to Aerobiology ▪ 3.2.1-3
by organism, the aw for primary and secondary colonizers
on building materials at room temperature ranges from 0.79
to >0.90 (127). Tertiary colonizers (e.g., Ulocladium spp.
and Stachybotrys chartarum) proliferate when aw becomes
>0.93 (126).
Water intrusion resulting from plumbing and roof leaks,
flooding, and condensation may provide conditions favor-
able for microbial growth in buildings. Once moisture has
accumulated on building surfaces, biocontaminants may
proliferate on surfaces and then be dispersed as bioaerosols
(Table 2). This enrichment potential on building materials
not designed for water exposure may increase indoor bioaero-
sol levels (88). Components of heating, ventilation, and air
conditioning (HVAC) systems may serve as amplification
sites, and these systems have been associated with the disper-
sal of contaminants indoors (69, 128). The cycling operation
of the HVAC may result in the release of surface bio-
contaminants present on duct materials into the occupied
space served by the system (69). Naturally ventilated build-
ings are also affected by bioaerosols, as organisms can be
transported via drafts through open windows and doors (8).
When microbial amplification occurs, the indoor environ-
ment then becomes a potential source of bioaerosol exposure
to the occupants. This exposure may be important because
people spend approximately 22 h/day in indoor environ-
ments (129).
Building occupants are a major source of bacterial aero-
sols, with human-to-human transmission often occurring in
high-density indoor environments, such as correctional facili-
ties (130), military training centers (131), and dormitories.
Coughing and loud talking are reported to release appro-
ximately 105 droplets/m3 with a mean droplet size <1 μm
and 104 droplets/m3 of droplets >1 μm (132). Microorganisms
are also dispersed from surfaces as a result of occupant activity
(133, 134). The Institute of Medicine (23) reports an associ-
ation of exposure to damp indoor environments and mold to
exacerbation of asthma in sensitized people, but also points to
the lack of sufficient evidence to associate other symptoms
to indoor exposures, thereby highlighting the need for addi-
tional research in this area. No regulations regarding bioaero-
sol concentrations are currently mandated for residential,
office, or classroom environments, but a variety of guidelines
predicated on baseline data have been proposed (135). It is
generally accepted that indoor sources of bioaerosols may be
significant when differences are noted between indoor and
outdoor concentrations and/or populations, but the lack of
exposure/dose response data has precluded the establishment
of bioaerosol threshold limit values (136).
Constituents of industrial indoor bioaerosols will vary
depending on the environmental conditions and the source.
For example, in grain processing plants, bioaerosols contain
the grain itself and a variety of other components, including
nongrain plant matter; molds and spores (e.g., Aspergillus and
Cladosporium spp.); in humid grain, thermophilic Actinomy-
cete spp.; mycotoxins (e.g., aflatoxin, zearalenone, vomitoxin,
ochratoxin, and T2 toxin); bacteria and their biochemi-
cal components and excretions (e.g., endotoxins, peptido-
glycans, and proteolytic enzymes); mites (e.g., Lepidoglyphus
destructor and Tyrophagus putrescentiae); insects (e.g., grain
weevil); and other animal matter, including parts of insects,
rodents, and birds and their excreta (36).
SOURCES: BIOLOGICAL WEAPON AGENTS
A dangerous anthropomorphic bioaerosol is one that is
generated from a biological weapon agent (BWA), such as
TABLE 2 Sources and amplification sites of indoor biocontaminants and associated airborne or surface concentrations

3.2.1-4 ▪ AIR
Site
Microbial agent(s) Concentration/unit vol.a Reference(s)
Category Facility/activity/material
Agricultural Animal facilities Actinomycetes (thermophilic) 100–101 CFU/m3 (50)
Bacteria (not specified) 103–105 CFU/m3 (50, 51)
Bacteria (Archaea) 106–108 16S rRNA gene copies/m3 (37, 52)
β-D-glucan 8–35 ng/mg (53)
Endotoxin 0.1–0.5 ng/m3 (54)
0.3–41 ng/mg (53)
Fungi (not specified) 101–104 EU/m3 (50, 55)
102–107 CFU/m3 (50, 51, 55–57)
Composting Actinomycetes (thermophilic) 104–105 CFU/m3 (46)
Bacteria (Legionellae) 103 CFU/m3 (59)
Fungi (thermophilic) 103–106 CFU/g (60)
103–105 CFU/m3 (58)
102–106 CFU/g (60)
Farming, harvesting, baling, grain Actinomycetes (thermophilic) 102–103CFU/m3 (61)
storage Bacteria
Gram-negative 104–105 CFU/m3 (61)
Mesophilic 105 CFU/m3 (61)
Not specified 101–104 CFU/m3 (40, 62)
Fungi (not specified) 103– 109 CFU/m3 (27, 40, 56, 57, 61–64)
Endotoxin 16–172 µg/m3 (27)
Mycotoxin 101–103 mg/kg (55, 65)
Air- handling systems HVAC systems Amoebic cysts NQb (66)
Bacteria (not specified) 102–103 CFU/m3 (66)
102–107 CFU/cm2 (67)
Fungi
Penicillium 102–107 CFU/g (68)
P. chrysogenum 104 CFU/m3 (69)
Micropolyspora faeni NQ (70)
Thermophilic 105 spores/m3 (71)
Cooling towers, water spray systems Actinomycetes (thermophilic) <dlc (72)
Bacteria
Heterotrophic <dl–106 CFU/ml (72)
Human source <dl–104 CFU/ml (72)
Fungi
Aureobasidium pullulans NQ (29)
not specified <dl–102 CFU/ml (72)
Portable humidifiers Bacteria
Legionella pneumophila 102–104 CFU/ml (73)
Pseudomonas 102–104 CFU/h (74)
Fungi
Several genera 101–102 CFU/ml (75)
Thermoactinomyces vulgaris NQ (75, 76)
Yeast 101–103 CFU/ml (75)
Public buildings Day care center Endotoxin 0.43–5.7 ng/m3 (77, 78)
β-D-glucan 0.157–0.20 ng/m3 (77, 78)
Museum Fungi NQ (68)
<dl–101CFU/m3 (79)
Office building Endotoxin 3.7 ng/m3 (78)
β-D-glucan 58 pg/m3 (78)
Post office Endotoxin 0.19 ng/m3 (77)
β-D-glucan 0.06 ng/m3 (77)
School Endotoxin 0.21–0.26 ng/m3 (77)
β-D-glucan 0.49–0.55 ng/m3 (77)
Building materials and Ceiling tile, insulation, painted Fungi
furnishings surfaces, wallpaper Aspergillus spp. 102 CFU/m3 (80)
A. versicolor 105 CFU/g (81)
Cladosporium spp. 102 CFU/m3 (80)
Penicillium spp. 103 CFU/m3 (80)
105 CFU/g (81)
Stachybotrys chartarum NQ (82, 83)
104–107 CFU/cm2 (84)
102 CFU/m3 (84)
106 CFU/g (81)
Carpet Bacteria (not specified) 104 CFU/in2 (85)
Fungi
Alternaria spp. 102 CFU/m3 (80)
Mesophilic 102 CFU/m3 (80)
Stachybotrys chartarum NQ (80)
Gypsum wallboard Bacteria (Gram-negative) 106 CFU/g (53)
β-D-glucan 0.4–210 µg/g (53)
Endotoxin 0.41–17 µg/g (53)
Fungi
Aspergillus versicolor 104–107 CFU/cm2 (84)
Mesophilic 103 CFU/g (86)

3.2.1. Introduction to Aerobiology ▪ 3.2.1-5

Mycobacteria 106 CFU/g (87)
Stachybotrys chartarum 106 CFU/g (87)
Total spores 107 spores/g (86)
Yeasts 103 CFU/g (86)
Xerophilic 102 CFU/g (86)
Remediated home ( post-flood) Bacteria (culturable) 5.7 × 102 avg CFU/m3 (88)
Fungi (culturable) 2.2 × 103 avg CFU/m3 (88)
Hot water heaters, hot water systems Bacteria
Legionella spp. 100–101 CFU/m3 (89)
L. pneumophila NQ (90)
House dust Bacteria
Bacillus 104–105 CFU/g (91)
Pseudomonas 104–104 CFU/g (91)
Streptomyces 104–106 CFU/g (91)
β-D-glucan 0.5–1.4 ng/mg (87)
Endotoxin 0.2–0.3 ng/mg (87)
0.7–18 ng/mg (92)
(Continued on next page)
TABLE 2 Sources and amplification sites of indoor biocontaminants and associated airborne or surface concentrations (Continued )

3.2.1-6 ▪ AIR
Site
Microbial agent(s) Concentration/unit vol.a Reference(s)
Category Facility/activity/material
Fungi
Alternaria 103–104 CFU/g (6, 93)
Aspergillus fumigatus 104 CFU/g (91)
A. niger 103 CFU/g (91)
Cladosporium 103–106 CFU/g (6, 91, 93)
Mixed species 104 CFU/g (94)
Penicillium 103–105 CFU/g (93, 95)
Rhodotorula 103–106 CFU/g (91, 95)
Yeasts (not specified) 103–106 CFU/g (91, 93, 95)
House plants Fungi (not specified) 103–104 CFU/m3 (96)
(greenhouses)
Remediation/demolition Fungi
Not specified 103– >109 CFU/m3 (97, 98)
Penicillium 105 CFU/m3 (97)
Total spores 105–106 counts/m3 (97)
Health care facilities Operating room Bacteria (not specified) 102 CFU/m3 (99)
102–105 particles/m3 (100)
Patient rooms Pneumocystis carinii NQ (101)
Virus (varicella-zoster) NQ (102)
Industrial Cellulose, Fungi
wood chip factory, sawmill Aspergillus fumigatus 102–105 CFU/m3 (103)
Penicillium spp. 101–105 CFU/m3 (103)
Not specified 104–108 CFU/m3 (56, 104)
Food processing, storage, warehouse Endotoxin 0.0125–54.9 µg/m3 (105)
Fungi
Cladosporium herbarium 103 CFU/m3 (106)
Not specified 102–104 CFU/m3 (104, 107)
Total 101–103 CFU/m3 (106)
Manufacturing (cotton mill, Bacteria
tobacco processing, paper Gram-negative bacilli 101–104 CFU/m3 (108–110)
processing, equipment cleaning, Total 101–103 CFU/m3 (111)
wood factory) Endotoxin 0.0042–3.6 μg/m3 (110, 111)
Fungi
Aspergillus fumigatus 101–102 CFU/m3 (112)
NQ (113)
Not specified 102–105 CFU/m3 (2, 57)
Total spores 102–104 spores/m3 (112)
Mycotoxin NQ (113)
Packing boxes Fungi (Aureobasidium pullulans) NQ (29)
Recycling plant Bacteria 105 CFU/m3 (46)
Fungi (Aspergillus fumigatus) 104 CFU/m3 (46)
Sanitary landfill Actinomycetes (thermophilic) <dl–103 CFU/m3 (43)
Bacteria
 Culturable 101–104 CFU/m3 (43)
Total coliforms <dl–103 CFU/m3 (43)
Fecal streptococci <dl–104 CFU/m3 (43)
Fungi
Mesophilic 103–104 CFU/m3 (43)
Thermophilic 100–104 CFU/m3 (43)
Waste handling Bacteria
Culturable 104 CFU/m3 (114)
Gram-negative bacilli 103 CFU/m3 (114)
Fungi 105 CFU/m3 (114)
104–105 CFU/m3 (115)
Transportation Private automobiles Bacteria (mixed biofilms) NQ (116)
Fungi NQ (116, 117)
Public buses Fungi (not specified) 102–105 CFU/m3 (118)
Wastewater treatment Activated sludge Bacteria
processing Coliforms 0.27–5.17 CFU/m3 (41)
Escherichia coli 102 CFU/m3 (42)
Gram-negative bacilli 101–105 CFU/m3 (119)
Endotoxin 0.1–350 ng/m3 (119)
Fungi
Mesophilic <dl–103 CFU/m3 (120)
Thermophilic <dl–102 CFU/m3 (120)
Aeration tanks Actinomycetes (thermophilic) <dl–101 CFU/m3 (121)
Bacteria
Fecal streptococci 102 CFU/m3 (82)
Gram-negative 101–102 CFU/m3 (121)
Mesophilic 102 CFU/m3 (121)
Streptococcus faecalis <dl–102 CFU/m3 (82)
Fungi 101–102 CFU/m3 (121)
Virus

3.2.1. Introduction to Aerobiology ▪ 3.2.1-7

Animal virus NQ (122)
Coliphage <10 PFU/m3 (122)
Effluent irrigation Bacteria
Escherichia coli 101–103 CFU/m3 (45)
Fecal coliforms 102 CFU/m3 (44)
Fecal streptococci 102 CFU/m3 (44)
Total coliforms 102 CFU/m3 (44)
Virus (enteric) NQ (44, 123)
Sludge compost Bacteria
Coliforms <dl (120)
not specified 100–103 CFU/m3 (120)
Trickling filter Bacteria
Coliforms 102–104 CFU/m3 (124)
Total 101–103 CFU/m3 (125)
a
Concentration/unit volume or area
b
NQ = not quantitated
c
dl = detection limit
3.2.1-8 ▪ AIR
Bacillus anthracis. Aerosolized BWA may be the most danger-
ous method of BWA dissemination because of the potential
for rapid spread of the agent throughout a building and the
difficulty in detecting the agent (137). Between 2 October
and 2 November 2001, B. anthracis spores were disseminated
from letters mailed through the U.S. Postal Service. The
subsequent cleanup of the Hart Senate Office Building in
Washington, D.C., cost more than $27 million. When used
as a weapon of mass destruction, a BWA is dispersed in par-
ticles less than 5 μm in diameter, a size that allows penetration
into the pulmonary alveoli (138).
Ventilation systems can provide a conduit to contami-
nate a building and the surrounding area. Such systems can
become entry points or distribution systems for hazardous
contaminants, including BWAs (139). Air circulation in
ordinary buildings can aid the spread of airborne disease
and disperse contaminants (137, 140). Re-aerosolization
of these hazardous bioparticles deposited onto surfaces can
be a continuing source of contamination. For example, a
study of re-aerosolization of B. anthracis after dispersal in a
postal facility found that a mail sorter remained contaminated
many days after processing contaminated letters (141). Par-
ticles may be deposited onto and re-aerosolized from sur-
faces at different rates, depending on particle size, velocity,
physical configuration of bioparticulate, surface, and other
environmental factors, such as humidity, dirt, and biofilm for-
mation (137, 142–145).
Infectious and toxigenic biological agents released in aero-
sols under favorable meteorological conditions for warfare
or terrorism may result in severe illness for military and civil-
ian populations (146). The Centers for Disease Control and
Prevention (CDC) has compiled a listing of microbial agents
that are of most concern and has classified them as category
A, B, or C agents (147). Category A agents are those that
have the greatest potential for adverse public health impact
with mass casualties, and most of the agents in this category
require broad-based public health preparedness efforts (e.g.,
specific or specialized surveillance, laboratory diagnosis, and
stockpiling of specific medications). Category A agents also
have a moderate to high potential for large-scale dissemina-
tion or a heightened general public awareness that could
cause mass public fear and civil disruption. Category B agents
are the second highest priority. They are moderately easy to
disseminate, cause moderate morbidity and low mortality,
and require specific diagnostic capacity/disease surveillance.
Category C agents are emerging pathogens that could be
engineered for mass dissemination, for example, the Nipah
and Hanta viruses (148). Their availability, ease of produc-
tion and dissemination, potential for high morbidity/mor-
tality, and major health impact place them on the list. A
listing of the CDC biothreat agents is presented in Table 3.
In addition to the CDC select agent list, multidrug-
resistant tuberculosis is a serious concern for health care facili-
ties and could be considered a potential biothreat agent.
MICROBIAL AGENTS
The environmental effects and human health complaints
resulting from airborne microorganisms have renewed interest
in a wide variety of microorganisms. The discovery of
Legionella pneumophila as the cause of the outbreak of Legion-
naires’ disease in Philadelphia in 1976 increased the aware-
ness of diseases caused by bacterial aerosols (28). A detailed
discussion of Legionella spp. and Legionnaires’ disease is pre-
sented in chapter 3.2.11. Similarly, the increased reporting
of tuberculosis in developing and industrialized countries
has prompted renewed interest in the genus Mycobacterium
and its airborne transmission. There is increasing concern
in the former Soviet Union of multidrug-resistant tuberculo-
sis (149). M. tuberculosis is spread via aerosols from an
infected person by coughing, sneezing, or talking and is
recognized as a significant public health concern because of
the low infectious dose (150). Nontuberculosis mycobacteria
have also been associated with respiratory illness (151, 152).
However, the significance of commonly isolated airborne
bacteria and fungi in offices, schools, residences, and outdoor
environments has not been determined. This is due partly
to the isolation of numerous airborne Gram-positive cocci
and Gram-positive bacilli in the absence of adverse health
effects. High ratios of airborne bacteria isolated from indoor
air compared with those from outdoor air have been used
as an indication of high occupancy rate, poor ventilation,
or inadequate building maintenance (136), but additional
research in this area is needed. Indoor sources for bioaerosols
at residences may include human occupants, pets, house dust,
organic waste, and HVAC system.
Airborne transmission of bacteria and other micro-
organisms in health care facilities can cause nosocomial infec-
tions (153). To mitigate possible infections, most hospitals
adhere to infection control procedures that include monitor-
ing air filtration, airflow direction and pressure, air changes
per hour, humidity, ventilation system cleaning and mainte-
nance, and using cleaners and disinfectants. However,
increasing numbers of elderly people are living in nursing
homes and home health care facilities where maintenance
and decontamination practices may not be strictly followed.
Therefore, the risk of exposure to opportunistic pathogens is
increased.
Archaebacteria are now included in bioaerosol assess-
ments. Advances in molecular phylogeny have been in-
strumental in the study of archaebacteria (methanogens,
halophiles, thermoacidophiles). Many archaebacteria are
difficult to study in the laboratory by culture methods
because some are killed by oxygen, some require extremely
high salt concentrations for growth, while others grow at tem-
peratures exceeding that of boiling water. Since DNA
sequencing has become a routine laboratory procedure,
archaebacteria are now included in bioaerosol assessments.
Archaebacteria are often analyzed using 16S rRNA techni-
que. This group is a dominant population in some agricul-
turally generated bioaerosols, such as those found in dairy or
swine facilities (37). Archaebacteria also make up a substan-
tial portion of the total microbial biomass in the oceans of
the world (154).
Exposure to airborne lysed bacterial cells may also result
in the inhalation of endotoxin, a lipopolysaccharide (LPS)
found in the cell walls of Gram-negative bacteria and
blue-green algae. Exposure to endotoxin may result in fever,
cough, headache, respiratory impairment, and exacerbation
of asthma (78, 155, 156). LPS in the bloodstream can lead
to fever, leukopenia and hypoglycemia, hypotension and
shock, and death from massive organ dysfunction. Airborne
endotoxin may be a major cause of illness in enclosed agri-
cultural settings, such as silage facilities, poultry processing
houses, and cotton mills (157, 158). Waste handlers have
reported increased nausea and gastrointestinal problems
that were associated with endotoxin exposure (81). Chilled
water spray humidification systems used in textile manufac-
turing facilities have also been implicated as sources of
aerosolized endotoxin and associated with cases of humidifier
fever and hypersensitivity pneumonitis (111). A review of
airborne endotoxin is presented in chapter 3.2.6.

TABLE 3 Classification of biological agents with potential use in bioterrorism (modified from references (44, 147))
Biothreat level Microorganism
Bacteria
A Bacillus anthracis
Francisella tularensis
Yersinia pestis
B Brucella spp.
Burkholderia mallei
Burkholderia pseudomallei
Coxiella burnetii
Escherichia coli O157:H7
Salmonella spp.
Shigella dysenteriae
Rickettsia prowazekii
Clamydia psittaci
Vibrio cholerae
Viruses
A Arenaviruses (e.g., Lassa, Machupo)
Filoviruses (e.g., Ebola, Marburg)
Junin
Variola major
B Eastern equine encephalomyelitis
Venezuelan equine encephalomyelitis
Western equine encephalomyelitis
C Nipah
Hantavirus
Tickborne virus
Yellow fever virus
Protozoa
C Cryptosporidium parvum
Toxin sources
A Clostridium botulinum
B Clostridium perfringens
Staphyloccus aureus
Ricinus communis (castor bean)
Although viruses do not replicate outside of a susceptible
host cell, they are readily transported through the air. Numer-
ous human viruses are transmitted via droplets and spread by
the respiratory route from one person to another in an indoor
environment (30). Enteric virus bioaerosols are produced at
sewage treatment facilities (124), and zoonotic transmission
of viruses is also of concern because serious illness resulting
from exposure to hantavirus aerosolized from rodent feces
and urine has been reported (21). Rodent-borne hemorrhagic
fevers include Korean (Hantaan virus), South American
(Junin and Machupo viruses), and Lassa fevers. The natural
reservoirs of Marburg and Ebola viruses (African hemorrhagic
fever) are not known, but it is suspected that they are har-
bored by rodents or bats (149). An extensive review of
airborne viruses is presented in chapter 3.2.7.
Airborne fungi have been the focus of much concern
because of potential for serious respiratory infection and
allergic reactions. Although few fungi actually cause infection
in healthy individuals, immunosuppressed host defenses
resulting from organ transplantation, cancer, therapy (e.g.,
antibiotics, steroids, drugs), or the presence of another
disease-causing agent may increase the likelihood of fungal
infection (159). More commonly, exposure to fungi can
initiate adverse health effects in the absence of infection by
3.2.1. Introduction to Aerobiology ▪ 3.2.1-9
Associated disease or toxin
Anthrax
Tularemia
Plague
Brucellosis
Glanders
Melioidosis
Q fever
Hemorrhagic colitis (food contaminant)
Salmonellosis (food contaminant)
Shigellosis (food contaminant)
Typhus fever (food contaminant)
Psittacosis
Cholera (water contaminant)
Lassa fever
Hemorrhagic fever
Argentine hemorrhagic fever
Small pox
Encephalomyelitis
Encephalomyelitis
Encephalomyelitis
Encephalomyelitis
Hemorrhagic fever, hantavirus pulmonary syndrome
Hemorrhagic fever and/or encephalitis
Yellow fever
Cryptosporidiosis (water safety)
Botulinum toxin
Epsilon toxin
Enterotoxin B
Ricin toxin
serving as aeroallergens, resulting in exacerbation of asthma,
allergic rhinitis, and respiratory distress (2, 67, 93, 160).
Exposure to the yeast Rhodotorula rubra was reported as the
cause in a case of extrinsic allergic alveolitis (161), while
the molds Penicillium and Aspergillus have been identified as
risk factors for asthma and atopy, respectively, in children
(162). Children exposed to damp houses were reported to
be at higher risk for colds, sore throats, and ear infections,
and mold in the home was a significant risk factor for otitis
and bronchitis (163). Kilpelainen et al. (164) reported an
association between colds and visible mold, dampness, or
water damage when surveying college students. Ebbehøj
et al. (165) reported that successive remediation of a water-
damaged, moldy building was needed to minimize occupant
symptoms of elevated peak flow variability, irritation,
rash, headache, dizziness, and difficulty in concentration.
Symptoms had persisted after removal of visible mold, but
additional remediation resulted in peak flow variability
returning to normal and other symptoms subsiding.
Numerous surveys have been conducted to determine air-
borne and surface concentrations of fungi present in a variety
of outdoor and indoor environments (Table 2). Horner et al.
(94) demonstrated the presence of leaf surface fungi (e.g.,
Cladosporium, Alternaria, Epicoccum, and Curvularia spp.)
3.2.1-10 ▪ AIR
as ≥20% of the culturable fungi and a low proportion of soil
fungi (e.g., Aspergillus, Penicillium, Emercella, and Paecilo-
myces spp.) in dust samples from nonproblem residences.
Fungi that indicate water damage, accumulation, or intrusion
(e.g., Stachybotrys, Chaetomium, and Ulocladium spp.) were
absent. The indoor airborne culturable data also confirmed
that fungal populations in non–water-damaged buildings
resemble outdoor airborne data. These data support the ear-
lier report by Samson et al. (166) that background popula-
tions of common fungi are present in indoor environments,
but water intrusion results in proliferation of soil fungi;
when conditions persist, the water indicator fungi colonize
and grow. Unfortunately, dose-response information for fun-
gal exposure is limited and many studies do not quantitatively
assess the fungal populations present in conjunction with
health data. Therefore, the Institute of Medicine (23) con-
cluded that there are insufficient data of a statistical associa-
tion between the presence of mold in damp environments
and upper respiratory tract symptoms. Additional studies are
needed to correlate mold exposure and the variety of reported
symptoms by occupants of damp buildings.
Exposure to mycotoxins produced by some fungi can result
in adverse human health effects. Exposure of agricultural
workers to mycotoxin-containing dust (2) and industrial
exposure to mycotoxins produced by A. fumigatus (113)
have been reported. Growth of fungi and the presence of
mycotoxins in indoor environments have been associated
with water-damaged building materials (49, 167). Fungi iso-
lated in indoor environments—such as Alternaria alternata
(168), Aspergillus fumigatus (169), A. flavus and A. parasiticus
(170), Fusarium gramineatum and F. sporotrichioides (171), and
Stachybotrys (27)—produce spores that contain mycotoxins,
and toxin production by isolates from residences of ill
children has been reported (172). Studies with laboratory
mice demonstrated dose-dependent inflammatory changes
following intranasal challenge to Stachybotrys spores (83),
and dose-response pulmonary inflammation and injury have
been documented in laboratory rats exposed to S. chartatum
spores (173). A report (174) also demonstrated the presence
of ergot alkaloids in the conidia of A. fumigatus that could
be respired and result in negative health effects. Wady and
Larsson (175) reported the presence of microbial volatile
organic compounds associated with respirable dust could
result in respiratory illness, and Fischer et al. (176) reported
microbial volatile organic compounds produced by fungi at
a compost facility. An extensive review of fungal bioaerosols
and mycotoxin is found in chapter 3.2.5. Additional infor-
mation on agricultural fungal pathogens is presented in
chapter 3.2.8.
Airborne actinomycetes have also been cited as the
cause of adverse health effects. Spores of Streptomycetes spp.
from damp, moldy houses have been shown to increase the
production of inflammatory mediators (e.g., macrophages
that produce tumor necrosis factor and interleukin-6) (92).
Exposure to thermophilic actinomycetes, whose optimal
growth temperature is >40°C, may occur near compost or
municipal landfills and when handling decomposing organic
matter in agricultural facilities (e.g., milling, sugarcane
bagasse) (67). Diagnostics for the detection of (1→3)
β-D-glucan, which form part of the cell wall of certain
medically important fungi, such as Aspergillus and Agaricus
species have been developed. See chapter 3.2.5 for a detailed
discussion.
Blowing dust was reported as the source of 62 genera of
airborne algae (177), and eutrophic lakes are also contri-
butors to outdoor airborne algal concentrations. Dust (54)
and aeration of fountains and aquariums (4) have been pro-
posed as possible indoor sources. Although exposure
to algal extracts has been associated with adverse human
health effects (54), the extent of allergic reactions due to algal
bioaerosols has not been fully investigated. Additional
research is needed to determine the environmental and
human health effects of airborne algae.
Free-living amoebae (e.g., Naegleria fowleri, Acantha-
moeba) are indigenous to soil and water and can be aerosolized
from natural and artificially heated waters such as power plant
discharges, lakes, and hot springs. Airborne Acanthamoeba
have been detected in the nasal passages of children in
Africa (178) and have been observed in air samples from
Mexico City (179). Cooling system waters have also been
reported as potential sources of aerosolized amoebae (13).
Humidifiers and ceiling dust were cited as an indoor source
of airborne amoebic antigen (127). Although severe health
effects can be elicited by exposure to these organisms (180),
insufficient information is currently available on airborne
protozoa.
Universal to the field of aerobiology is the need for meas-
urement methods to detect and identify microorganism(s)
of interest (181). Classical microbiological methods for
measurement of airborne microorganisms rely on culture
and/or microscopic assay using forced airflow sampling (136).
Culture-based methods require appropriate conditions for
growth of culturable organisms, but airborne microorganisms
are stressed during transport and collection and may not
respond to incubation conditions in the laboratory (182).
Total count procedures, using direct microscopic enumera-
tion, are tedious and often fail to distinguish genera and/or
species. Measurement methods to determine endotoxin,
ergosterol, and ß,1–3 D-glucan (183) are used as alternatives
to culturable methods and polymerase chain reaction
(PCR) has been shown to be an effective means to detect
and identify airborne bacteria (5, 184, 185), Mycoplasma
(186), Pneumocystis carinii (101), Stachybotrys chartarum
(187), and virus (102). Chapters 3.2.2 and 3.2.3 discuss sam-
pling and analysis methods, respectively.
In summary, interest in the populations of airborne micro-
organisms in agricultural and industrial settings, health care
facilities, residences, offices, and classroom environments
has increased in recent years. The threat of purposeful release
of microorganisms as bioterrorism agents has prompted
renewed interest in aerobiology, and research activity in this
area of environmental microbiology has rapidly expanded.
The following chapters focus on specific topics related to
bioaerosols and provide detailed information on sampling
methods, analysis, fate and transport, fungi and mycotoxins,
bacteria and endotoxins, Legionellae, viruses, and agricul-
tural pathogens.
REFERENCES
1. Keleti G, Shapiro MA, Snyder IS. 1987. Legionella and the
environment. Crit Rev Env Sci and Tec 17(2):133–185.
2. Lacey J, Crook B. 1988. Fungal and actinomycete spores as
pollutants of the workplace and occupational allergens.
Ann Occup Hyg 32(4):515–533.
3. Jacobs RR. 1989. Airborne endotoxins: an association with
occupational lung disease. Appl Ind Hyg 4(2):50–56.
4. Lighthart B. 1994. Physics of microbial bioaerosols, p 5–27.
In Lighthart B, Mohr A (ed), Atmospheric microbial aerosols,
theory and applications. Chapman and Hall, New York.

5. Alvarez AJ, Buttner MP, Stetzenbach LD. 1995. PCR
for bioaerosol monitoring: sensitivity and environmental
interference. Appl Environ Microb 61(10):3639–3644.
6. Wickman HH. 1994. Deposition, adhesion, and release of
bioaerosols, p 99–165. In Lighthart B, Mohr AJ. (ed),
Atmospheric microbial aerosols, theory and applications.
Chapman and Hall, New York.
7. Oliver J. 2000. The public health significance of Viable but
nonculturable bacteria, p 277–300. In Colwell R, Grimes
DJ (ed), Nonculturable microorganisms in the environment.
Springer, New York.
8. Lighthart B, Stetzenbach LD. 1994. Distribution of micro-
bial bioaerosols, p 68–98. In Lighthart B, Mohr AJ (ed),
Atmospheric microbial aerosols. Theory and Applications
Chapman and Hall, New York.
9. Gillette DA, Blifford Jr IH, Fenster CR. 1972. Measure-
ments of aerosol size distributions and vertical fluxes of
aerosols on land subject to wind erosion. J Appl Meteorol
11:977–987.
10. Lindemann J, Constantinidou HA, Barchet WR,
Upper CD. 1982. Plants as sources of airborne bacteria,
including ice nucleation-active bacteria. Appl Environ
Microb 44(5):1059–1063.
11. Lighthart B. 2000. Mini-review of the concentration
variations found inthe alfresco atmospheric bacterial popu-
lations. Aerobiologia 16(1):7–16.
12. Lighthart B, Mohr A. 1987. Estimating downwind con-
centrations of viable airborne microorganisms in dynamic
atmospheric conditions. Appl Environ Microb 53(7):
1580–1583.
13. Lynch JM, Poole NJ. 1979. Aerial dispersal and the devel-
opment of microbial communities, p 140–170. In Lynch
JM, Poole NJ. (ed), Microbial ecology: a conceptual approach.
John Wiley and Sons, New York.
14. Pedgley DE. 1991. Aerobiology: the atmosphere as a source
and sink for microbes, p 43–59. In Andrews JH, Hirano SS
(ed), Microbial Ecology of Leaves. Springer, New York.
15. Cammack RH. 1955. Seasonal changes in three common
constituents of the air spora of Southern Nigeria. Nature
176:1270–1272.
16. Knudsen G, Spurr Jr H. 1987. Field persistence and effi-
cacy of five bacterial preparations for control of peanut
leaf spot. Plant Dis 71:442–445.
17. Dimmick RL, Wolochow H, Chatigny MA. 1979. Evi-
dence that bacteria can form new cells in airborne particles.
Appl Environ Microb 37(5):924–927.
18. Adhikari A, Reponen T, Rylander R. 2013. Airborne
fungal cell fragments in homes in relation to total fungal
biomass. Indoor Air 23(2):142–147.
19. Shaffer BT, Lighthart B. 1997. Survey of culturable
airborne bacteria at four diverse locations in Oregon:
urban, rural, forest, and coastal. Microb Ecol 34(3):
167–177.
20. Adhikari A, Gupta J, Wilkins JR, Olds RL, Indugula R,
Cho KJ, Li C, Yermakov M. 2011. Airborne microorgan-
isms, endotoxin, and (1→3)-β-D-glucan exposure in green-
houses and assessment of respiratory symptoms among
workers. Ann Occup Hyg 55(3):272–285.
21. Diglisic G, Rossi C, Doti A, Walshe D. 1999. Seropreva-
lence study of hantavirus infection in the community based
population. Maryland Med J 48(6):303–306.
22. Flaherty DK, Deck F, Cooper J, Bishop K, Winzenburger
P, Smith L, Bynum L, Witmer W. 1984. Bacterial endo-
toxin isolated from a water spray air humidification system
as a putative agent of occupation-related lung disease. Infect
Immun 43(1):206–212.
23. Institute of Medicine of the National Academies 2004.
Damp indoor spaces and health. National Academies Press,
Washington, DC.
3.2.1. Introduction to Aerobiology ▪ 3.2.1-11
24. Rubin SA. 1997. Tuberculosis and atypical mycobacterial
infections in the 1990s. Radiographics 17(4):1051–1059.
25. Rylander R, Etzel R. 1999. Introduction and summary:
workshop on children’s health and indoor mold exposure.
Environ Health Perspect 107(Suppl 3):465–468.
26. Seuri M, Husman K, Kinnunen H, Reiman M, Kreus R,
Kuronen P, Lehtomäki K, Paananen M. 2000. An
outbreak of respiratory diseases among workers at a
water-damaged building—a case report. Indoor Air 10(3):
138–145.
27. Sorenson W, Frazer DG, Jarvis BB, Simpson J, Robinson
V. 1987. Trichothecene mycotoxins in aerosolized
conidia of Stachybotrys atra. Appl Environ Microb 53(6):
1370–1375.
28. Winn WC, Jr, 1988 Legionnaires disease: historical per-
spective. Clin Microbiol Rev 1(1):60–81.
29. Woodard ED, Friedlander B, Lesher RJ, Font W, Kinsey
R, Hearne FT. 1988. Outbreak of hypersensitivity pneu-
monitis in an industrial setting. J Am Med Assoc 259
(13):1965–1969.
30. Zeterberg J. 1973. A review of respiratory virology and
the spread of virulent and possibly antigenic viruses via
air conditioning systems. I Ann Allergy 31(5):228.
31. Slote L. 1976. Viral aerosols—a potential occupationally
related health threat in aerated wastewater treatment sys-
tems. J Environ Health 38(5):310.
32. Blais Lecours P, Duchaine C, Taillefer M, Tremblay C,
Veillette M, Cormier Y, Marsolais D. 2011. Immunogenic
properties of archaeal species found in bioaerosols. PLoS
One 6(8):e23326.
33. Salem H, Gardner DE. 1994. Health aspects of bioaero-
sols, p 304–330. In Lighthart B, Mohr AJ (ed), Atmospheric
microbial aerosols, theory and applications. Chapman and
Hall, New York.
34. Sattar SA, Ijaz MK. 1987. Spread of viral infections by
aerosols. Crit Rev Environ Contr 17(2):89–131.
35. Papineni RS, Rosenthal FS. 1997. The size distribution of
droplets in the exhaled breath of healthy human subjects.
J Aerosol Med 10(2):105–116.
36. American Thoracic Society 1998. Supplement: American
Thoracic Society: respiratory health hazards in agriculture.
Am J Resp Crit Care Med 158(5Part 2):S1–S79.
37. Blais Lecours P, Veillette M, Marsolais D, Duchaine C.
2012. Characterization of bioaerosols from dairy barns:
reconstructing the puzzle of occupational respiratory
diseases by using molecular approaches. Appl Environ
Microbiol 78(9):3242–3248.
38. Park H, Park H, Lee I. 2010. Microbial exposure
assessment in sawmill, livestock feed industry, and metal
working fluids handling industry. Saf Health Work 1(2):
183–191.
39. Abdel-Hafez S, Moubasher A, Shoreit A, Ismail M. 1990.
Fungal flora associated with combine harvester wheat and
sorghum dusts from Egypt. J Basic Microb 30(7):467–479.
40. Lighthart B. 1984. Microbial aerosols: estimated contribu-
tion of combine harvesting to an airshed. Appl Environ
Microb 47(2):430–432.
41. Fannin KF, Vana SC, Jakubowski W. 1985. Effect of an
activated sludge wastewater treatment plant on ambient
air densities of aerosols containing bacteria and viruses.
Appl Environ Microb 49(5):1191–1196.
42. Randall CW, Ledbetter JO. 1966. Bacterial air pollution
from activated sludge units. Am Ind Hyg Assoc J 27(6):
506–519.
43. Rahkonen P, Ettala M, Loikkanen I. 1987. Working
conditions and hygiene at sanitary landfills in Finland.
Ann Occup Hyg 31(4a):505–513.
44. Applebaum J, Guttman-Bass N, Lugten M, Teltsch B,
Fattal B, Shuval H. 1984. Dispersion of aerosolized enteric
3.2.1-12 ▪ AIR
viruses and bacteria by sprinkler irrigation with wastewater.
Monogr Virol 15:193–201.
45. Teltsch B, Katzenelson E. 1978. Airborne enteric bacteria
and viruses from spray irrigation with wastewater. Appl
Environ Microbiol 35(2):290–296.
46. Reinthaler FF, Haas D, Feierling G, Schlacher R, Pichler-
Semmelrock FP, Kock M, Wust G, Feenstra O, Marth E.
1999. Comparative investigations of airborne culturable
microorganisms in selected waste treatment facilities and
in neighbouring residential areas. Int J Hyg Environ Med
202(1):1–17.
47. Marshall B, Flynn P, Kamely D, Levy S. 1988. Survival of
Escherichia coli with and without ColE1: Tn5 after aerosol
dispersal in a laboratory and a farm environment. Appl
Environ Microb 54(7):1776–1783.
48. Juozaitis A, Huang Y-L, Willeke K, Donnelly J, Kalatoor
S, Leeson A, Wyza R. 1994. Dispersion of respirable aero-
sols in a fermenter and their removal in an exhaust system.
Appl Occup Environ Hyg 9(8):552–559.
49. Gravesen S, Nielsen PA, Iversen R, Nielsen KF. 1999.
Microfungal contamination of damp buildings—examples
of risk constructions and risk materials. Environ Health Persp
107(Suppl 3):505.
50. Duchaine C, Grimard Y, Cormier Y. 2000. Influence of
building maintenance, environmental factors, and seasons
on airborne contaminants of swine confinement buildings.
J Am Ind Hyg Assoc 61(1):56–63.
51. Thorne PS, Kiekhaefer MS, Whitten P, Donham KJ.
1992. Comparison of bioaerosol sampling methods in
barns housing swine. Appl Environ Microbiol 58(8):2543–
2551.
52. Nehmé B, Gilbert Y, Létourneau V, Forster RJ, Veillette
M, Villemur R, Duchaine C. 2009. Culture-independent
characterization of archaeal biodiversity in swine con-
finement building bioaerosols. Appl Environ Microb 75
(17):5445–5450.
53. Andersson M, Nikulin M, Köljalg U, Andersson M,
Rainey F, Reijula K, Hintikka E, Salkinoja-Salonen M.
1997. Bacteria, molds, and toxins in water-damaged build-
ing materials. Appl Environ Microb 63(2):387–393.
54. Bernstein IL, Safferman R. 1973. Clinical sensitivity to
green algae demonstrated by nasal challenge and in vitro
tests of immediate hypersensitivity. J Allerg Clin Immun
51(1):22–28.
55. Kullman G, Thorne P, Waldron P, Marx J, Ault B, Lewis
D, Siegel P, Olenchock S, Merchant J. 1999. Organic dust
exposures from work in dairy barns. Am Ind Hyg Assoc 59
(6):403–413.
56. Eduard W, Heederik D, Duchaine C, Green BJ. 2012.
Bioaerosol exposure assessment in the workplace: the
past, present and recent advances. J Environ Monitor
14(2):334–339.
57. Lundholm IM. 1982. Comparison of methods for quan-
titative determinations of airborne bacteria and evalua-
tion of total viable counts. Appl Environ Microb 44(1):
179–183.
58. Herr C, Zur Nieden A, Jankofsky M, Stilianakis N, Boe-
deker R, Eikmann T. 2003. Effects of bioaerosol polluted
outdoor air on airways of residents: a cross sectional study.
Occup Environ Med 60(5):336–342.
59. Hughes MS, Steele TW. 1994. Occurrence and distribu-
tion of Legionella species in composted plant materials.
Appl Environ Microbiol 60(6):2003–2005.
60. Straatsma G, Samson RA, Olijnsma TW, Op Den Camp
HJ, Gerrits JP, Van Griensven LJ. 1994. Ecology of
thermophilic fungi in mushroom compost, with emphasis
on Scytalidium thermophilum and growth stimulation
of Agaricus bisporus mycelium. Appl Environ Microbiol 60
(2):454–458.
61. Krysiń ska-Traczyk E, Skorska C, Prażmo Z, Sitkowska J,
Cholewa G, Dutkiewicz J. 2004. Exposure to airborne
microorganisms, dust and endotoxin during flax scutching
on farms. Ann Agric Environ Med 11:309–317.
62. Swan JR, Crook B. 1998. Airborne microorganisms asso-
ciated with grain handling. Ann Agric Environ Med 5(1):
7–15.
63. Burge HA, Solomon WR, Muilenberg ML. 1982. Evalu-
ation of indoor plantings as allergen exposure sources.
J Allerg Clin Immun 70(2):101–108.
64. Pasanen AL, Yli-Pietila K, Pasanen P, Kalliokoski P,
Tarhanen J. 1999. Ergosterol content in various fungal
species and biocontaminated building materials. Appl Envi-
ron Microbiol 65(1):138–142.
65. Nordby K-C, Halstensen AS, Elen O, Clasen P-E, Lang-
seth W, Kristensen P, Eduard W. 2004. Tricho-
thecene mycotoxins and their determinants in settled
dust related to grain production. Ann Agr Environ Med 11
(1):75–83.
66. Hugenholtz P, Fuerst JA. 1992. Heterotrophic bacteria in
an air-handling system. Appl Environ Microbiol 58(12):
3914–3920.
67. Banaszak EF, Barboriak J, Fink J, Scanlon G, Schlueter
DP, Sosman A, Thiede W, Unger G. 1974. Epidemiologic
studies relating thermophilic fungi and hypersensitivity
lung syndromes. Am Rev Resp Dis 110(5):585.
68. Krake AM, Worthington KA, Wallingford KM, Martinez
KF. 1999. Evaluation of microbiological contamination in
a museum. Appl Occup Environ Hyg 14(8):499–509.
69. Buttner MP, Cruz-Perez P, Garrett PJ, Stetzenbach LD.
1999. Dispersal of fungal spores from three types of air
handling system duct material. Aerobiologia 15(1):1–8.
70. Fink JN, Banaszak EF, Thiede WH, Barboriak JJ.
1971. Interstitial pneumonitis due to hypersensitivity to
an organism contaminating a heating system. Ann Intern
Med 74(1):80–83.
71. Warhurst D. 1977. Humidifier fever. Amoebae and aller-
gic lung disease. Thorax 32:653–663.
72. Burkhart JE, Stanevich R, Kovak B. 1993. Microorgan-
ism contamination of HVAC humidification systems: case
study. Appl Occup Environ Hyg 8(12):1010–1014.
73. Zuravleff JJ, Yu VL, Shonnard JW, Rihs JD, Best M.
1983. Legionella pneumophila contamination of a hospital
humidifier. Demonstration of aerosol transmission and
subsequent subclinical infection in exposed guinea pigs.
Am Rev Respir Dis 128(4):657–661.
74. Covelli H, Kleeman J, Martin J, Landau W, Hughes R.
1973. Bacterial emission from both vapor and aerosol
humidifiers. Am Rev Resp Dis 108(3):698.
75. Burge HA, Muilenberg ML, Chapman JA. 1991. Crop
plants as a source of fungus spores of medical importance,
p 222–236, Microbial ecology of leaves. Springer, Berlin.
76. Tourville DR, Weiss WI, Wertlake PT, Leudemann GM.
1972. Hypersensitivity pneumonitis due to contamina-
tion of home humidifier. J Allerg Clin Immunol 49(4):
245–251.
77. Rylander R, Persson K, Goto H, Yuasa K, Tanaka S.
1992. Airborne beta-1, 3-glucan may be related to symp-
toms in sick buildings. Indoor Environ 1(5):263–267.
78. Wan GH, Li CS. 1999. Indoor endotoxin and glucan in
association with airway inflammation and systemic symp-
toms. Arch Environ Health 54(3):172–179.
79. Brimblecombe P, Blades N, Camuffo D, Sturaro G,
Valentino A, Gysels K, Grieken R, Busse HJ, Kim O,
Ulrych U. 1999. The indoor environment of a modern
museum building, the Sainsbury Centre for Visual Arts,
Norwich, UK. Indoor Air 9(3):146–164.
80. Kozak Jr P, Gallup J, Cummins L, Gillman S. 1980.
Currently available methods for home mold surveys. II.

Examples of problem homes surveyed. Ann Allerg 45(3):
167.
81. Johanning E, Landsbergis P, Gareis M, Yang CS, Olmsted
EA. 1999. Clinical experience and results of a Sentinel
Health Investigation related to indoor fungal exposure.
Environ Health Persp 107(Suppl 3):489.
82. Crawford GV, Jones PH. 1979. Sampling and differentia-
tion techniques for airborne organisms emitted from waste-
water. Water Res 13(4):393–399.
83. Nikulin M, Pasanen A, Berg S, Hintikka E. 1994. Stachy-
botrys atra growth and toxin production in some building
materials and fodder under different relative humidities.
Appl Environ Microbiol 60(9):3421–3424.
84. Hodgson MJ, Morey P, Leung WY, Morrow L, Miller D,
Jarvis BB, Robbins H, Halsey JF, Storey E. 1998.
Building-associated pulmonary disease from exposure to
Stachybotrys chartarum and Aspergillus versicolor. J Occup
Environ Med/Am Coll Occup Environ Med 40(3):
241–249.
85. Anderson RL. 1969. Biological evaluation of carpeting.
Appl Microbiol 18(2):180–187.
86. Pasanen AL, Kalliokoski P, Pasanen P, Salmi T, Tossa-
vainen A. 1989. Fungi carried from farmers’ work into
farm homes. Am Ind Hyg Assoc J 50(12):631–633.
87. Andersson A, Weiss N, Rainey F, Salkinoja-Salonen M.
1999. Dust-borne bacteria in animal sheds, schools
and children’s day care centres. J App Microbiol 86(4):
622–634.
88. Fabian M, Miller S, Reponen T, Hernandez M. 2005.
Ambient bioaerosol indices for indoor air quality assess-
ments of flood reclamation. J Aerosol Sci 36(5):763–783.
89. Bollin G, Plouffe J, Para M, Hackman B. 1985. Aerosols
containing Legionella pneumophila generated by shower
heads and hot-water faucets. Appl Environ Microb 50(5):
1128–1131.
90. Helms CM, Massanari RM, Zeitler R, Streed S, Gilchrist
MJ, Hall N, Hausler WJ, Sywassink J, Johnson W, Win-
termeyer L. 1983. Legionnaires’ disease associated with a
hospital water system: a cluster of 24 nosocomial cases.
Ann Intern Med 99(2):172–178.
91. Nilsson A, Kihlström E, Lagesson V, Wessén B, Szponar
B, Larsson L, Tagesson C. 2004. Microorganisms and vol-
atile organic compounds in airborne dust from damp resi-
dences. Indoor Air 14(2):74–82.
92. Hirvonen MR, Nevalainen A, Makkonen N,
Mönkkönen J, Savolainen K. 1997. Induced production
of nitric oxide, tumor necrosis factor, and interleukin-6
in RAW 264.7 macrophages by streptomycetes from indoor
air of moldy houses. Arch Environ Health: An Int J 52(6):
426–432.
93. Miller J, Laflamme A, Sobol Y, Lafontaine P, Greenhalgh
R. 1988. Fungi and fungal products in some Canadian
houses. Int Biodeterior 24(2):103–120.
94. Horner WE, Worthan AG, Morey PR. 2004. Air- and
dustborne mycoflora in houses free of water damage
and fungal growth. Appl Environ Microbiol 70(11):
6394–6400.
95. Wickman M, Gravesen S, Nordvall SL, Pershagen G,
Sundell J. 1992. Indoor viable dust-bound microfungi in
relation to residential characteristics, living habits, and
symptoms in atopic and control children. J Allerg Clin
Immunol 89(3):752–759.
96. Burge HA, Solomon WR, Boise JR. 1980. Microbial prev-
alence in domestic humidifiers. Appl Environ Microb 39(4):
840–844.
97. Rautiala S, Reponen T, Hyvärinen A, Nevalainen A,
Husman T, Vehviläinen A, Kalliokoski P. 1996. Exposure
to airborne microbes during the repair of moldy buildings.
Am Ind Hyg Assoc 57(3):279–284.
3.2.1. Introduction to Aerobiology ▪ 3.2.1-13
98. Rautiala S, Reponen T, Nevalainen A, Husman T,
Kalliokoski P. 1998. Control of exposure to airborne viable
microorganisms during remediation of moldy buildings;
report of three case studies. Am Ind Hyg Assoc 59(7):
455–460.
99. Tjade OH, Gabor I. 1980. Evaluation of airborne operat-
ing room bacteria with a Biap slit sampler. J Hyg 84(1):
37–40.
100. Seal D, Clark R. 1990. Electronic particle counting
for evaluating the quality of air in operating theatres: a
potential basis for standards? J Appl Microbiol 68(3):
225–230.
101. Olsson M, Lidman C, Latouche S, Bjorkman A, Roux P,
Linder E, Wahlgren M. 1998. Identification of Pneu-
mocystis carinii f. sp. hominis gene sequences in filtered
air in hospital environments. J Clin Microbiol 36(6):
1737–1740.
102. Sawyer MH, Chamberlin CJ, Wu YN, Aintablian N,
Wallace MR. 1994. Detection of varicella-zoster virus
DNA in air samples from hospital rooms. J Infect Dis 169
(1):91–94.
103. Kotimaa MH. 1990. Occupational exposure to fungal and
actinomycete spores during the handling of wood chips.
Grana 29(2):153–156.
104. Blomquist G, Ström G, Strömquist L-H. 1984. Sampling
of high concentrations of airborne fungi. Scand J Work,
Environ Health (10):109–113.
105. Dutkiewicz J, Krysiń ska-Traczyk E, Skórska C, Sitkow-
ska J, Prażmo Z, Urbanowicz B. 2000. Exposure of agricul-
tural workers to airborne microorganisms and endotoxin
during handling of various vegetable products. Aerobiologia
16(2):193–198.
106. Jain AK. 2000. Survey of bioaerosol in different indoor
working environments in central India. Aerobiologia 16
(2):221–225.
107. Jensen PA, Todd WF, Harth ME, Mickelseri RL,
O’Brien DM. 1993. Evaluation and control of worker
exposure to fungi in a beet sugar refinery. Am Ind Hyg Assoc
J 54(12):742–748.
108. Lugauskas A, Krikstaponis A, Sveistyte L. 2003. Airborne
fungi in industrial environments—potential agents ofrespira-
tory diseases. Ann Agr Environ Med 11(1):19–25.
109. Prazmo Z, Krysinska-Traczyk E, Skorska C, Sitkowska J,
Cholewa G, Dutkiewicz J. 2003. Exposure to bioaerosols
in a municipal sewage treatment plant. Ann Agric Environ
Med 10(2):241–248.
110. Prażmo Z, Dutkiewicz J, Cholewa G. 2000. Gram-negative
bacteria associated with timber as a potential respiratory haz-
ard for woodworkers. Aerobiologia 16(2):275–279.
111. Walters M, Milton D, Larsson L, Ford T. 1994. Airborne
environmental endotoxin: a cross-validation of sampling
and analysis techniques. Appl Environ Microbiol 60(3):
996–1005.
112. Huuskonen MS, Husman K, Jarvisalo J, Korhonen O,
Kotimaa M, Kuusela T, Nordman H, Zitting A, Manty-
jarvi R. 1984. Extrinsic allergic alveolitis in the tobacco
industry. Br J Ind Med 41(1):77–83.
113. Land C, Hult K, Fuchs R, Hagelberg S, Lundström H.
1987. Tremorgenic mycotoxins from Aspergillus fumigatus
as a possible occupational health problem in sawmills.
Appl Environ Microb 53(4):787–790.
114. Kiviranta H, Tuomainen A, Reiman M, Laitinen S,
Nevalainen A, Liesivuori J. 1999. Exposure to airborne
microorganisms and volatile organic compounds in differ-
ent types of waste handling. Ann Agr Environ Med 6(1):
39–44.
115. Nielsen EM, Breum NO, Nielsen BH, Würtz H, Poulsen
OM, Midtgaard U. 1997. Bioaerosol exposure in waste
collection: a comparative study on the significance of
3.2.1-14 ▪ AIR
collection equipment, type of waste and seasonal variation.
Ann Occup Hyg 41(3):325–344.
116. Simmons RB, Rose LJ, Crow SA, Ahearn DG. 1999. The
occurrence and persistence of mixed biofilms in auto-
mobile air conditioning systems. Curr Microbiol 39(3):
141–145.
117. Kumar P, Marier R, Leech SH. 1984. Respiratory allergies
related to automobile air conditioners. New Eng J Med 311
(25):1619–1621.
118. Van Netten C, Brands R, Dill B. 1997. Investigation and
remediation of diesel converted trolley buses associated
with extensive fungal growth and health complaints. Am
Ind Hyg Assoc J 58(10):726–731.
119. Laitinen S, Kangas J, Kotimaa M, Liesivuori J, Martikai-
nen PJ, Nevalainen A, Sarantila R, Husman K. 1994.
Workers’ exposure to airborne bacteria and endotoxins at
industrial wastewater treatment plants. Am Ind Hyg Assoc
55(11):1055–1060.
120. Jones B, Cookson J. 1983. Natural atmospheric microbial
conditions in a typical suburban area. Appl Environ Microb
45(3):919–934.
121. Prazmo Z, Dutkiewicz J, Skorska C, Sitkowska J,
Cholewa G. 2003. Exposure to airborne Gram-negative
bacteria, dust and endotoxin in paper factories. Ann Agric
Environ Med 10(1):93–100.
122. Brenner K, Scarpino P, Clark C. 1988. Animal viruses,
coliphages, and bacteria in aerosols and wastewater at
a spray irrigation site. Appl Environ Microb 54(2):409–
415.
123. Fattal B, Katzenelson E, Guttman-Bass N, Sadovski A.
1984. Relative survival rates of enteric viruses and bacterial
indicators in water, soil, and air. Enteric Viruses in Water
(15):184–192.
124. Adams D, Spendlove J, Spendlove R, Barnett B. 1982.
Aerosol stability of infectious and potentially infectious
reovirus particles. Appl Environ Microb 44(4):903–908.
125. Adams AP, Spendlove JC. 1970. Coliform aerosols emit-
ted by sewage treatment plants. Science 169(3951):1218–
1220.
126. Grant C, Hunter C, Flannigan B, Bravery A. 1989. The
moisture requirements of moulds isolated from domestic
dwellings. Int Biodeterior 25(4):259–284.
127. Eduard W, Lacey J, Karlsson K, Palmgren U, Strom G,
Blomquist G. 1990. Evaluation of methods for enumerat-
ing microorganisms in filter samples from highly contami-
nated occupational environments. Am Ind Hyg Assoc J
51(8):427–436.
128. Bernstein RS, Sorenson WG, Garabrant D, Reaux C,
Treitman RD. 1983. Exposures to respirable, airborne Pen-
icillium from a contaminated ventilation system: clinical,
environmental and epidemiological aspects. Am Ind Hyg
Assoc J 44(3):161–169.
129. Spengler JD, Sexton K. 1983. Indoor air pollution: a public
health perspective. Science 221(4605):9–17.
130. Hoge CW, Reichler MR, Dominguez EA, Bremer JC,
Mastro TD, Hendricks KA, Musher DM, Elliott JA,
Facklam RR, Breiman RF. 1994. An epidemic of pneumo-
coccal disease in an overcrowded, inadequately ventilated
jail. New Eng J Med 331(10):643–648.
131. Brundage JF, Scott RM, Lednar WM, Smith DW, Miller
RN. 1988. Building-associated risk of febrile acute respira-
tory diseases in Army trainees. J Am Med Assoc 259(14):
2108–2112.
132. Lindow SE, Knudsen GR, Seidler RJ, Walter MV,
Lambou VW, Amy PS, Schmedding D, Prince V, Hern
S. 1988. Aerial dispersal and epiphytic survival of Pseudo-
monas syringae during a pretest for the release of genetically
engineered strains into the environment. Appl Environ
Microb 54(6):1557–1563.
133. Buttner MP, Stetzenbach LD. 1993. Monitoring airborne
fungal spores in an experimental indoor environment to
evaluate sampling methods and the effects of human activ-
ity on air sampling. Appl Environ Microb 59(1):219–226.
134. Reynolds S, Streifel A, McJilton C. 1990. Elevated
airborne concentrations of fungi in residential and office
environments. Am Ind Hyg Assoc J 51(11):601–604.
135. Rao CY, Burge HA, Chang JC. 1996. Review of quantita-
tive standards and guidelines for fungi in indoor air. J Air &
Waste Manage Assoc 46(9):899–908.
136. American Conference of Governmental Industrial Hygien-
ists. 1999. Bioaerosols: assessment and control. ACGIH,
Cincinnati, OH.
137. Kowalski WJ. 2002. Immune building systems technology.
McGraw-Hill Professional, New York.
138. Oncü S, Sakarya S. 2003. Anthrax–an overview. Medical
Sci Mon: Int Med J Exp Clin Res 9(11):RA276.
139. National Institute for Occupational Safety and Health
2002. Guidance for protecting building environments from
airborne chemical, biological, or radiological attacks NIOSH
publication no 2002–139 National Institute for Occupational
Safety and Health Centers for Disease Control and Prevention,
Department of Health and Human Services, Cincinnati,
OH.
140. Sippola MR, Nazaroff WW. 2003. Modeling particle
loss in ventilation ducts. Atmos Environ 37(39):
5597–5609.
141. Dull PM, Wilson KE, Kournikakis B, Whitney E, Boulet
CA, Ho J, Ogston J, Spence MR, Mckenzie MM, Phelan
MA. 2002. Bacillus anthracis aerosolization associated with a
contaminated mail sorting machine. Emerg Infect Dis 8
(10):1044–1047.
142. Ibrahim A, Dunn P, Brach R. 2003. Microparticle detach-
ment from surfaces exposed to turbulent air flow: controlled
experiments and modeling. J Aerosol Sci 34(6):765–782.
143. Krauter P, Biermann A. 2007. Reaerosolization of fluid-
ized spores in ventilation systems. Appl Environ Microb 73
(7):2165–2172.
144. Krauter P, Biermann A, Larsen LD. 2005. Transport
efficiency and deposition velocity of fluidized spores in
ventilation ducts. Aerobiologia 21(3–4):155–172.
145. Sehmel G. 1980. Particle resuspension: a review. Environ
Int 4(2):107–127.
146. Wiener SL. 2000. Biological warfare defense, p 119–146.
In Zilinskas RA (ed), Biological warfare modern offense and
defense, Lynne Rienner, Boulder, CO.
147. Snyder JW, Weissfeld AS. 2003. Laboratory detection of
potential agents of bioterrorism, p 121–128. In Murray
PR, Baron EJ, Jorgensen JH, Pfaller MA, Yolken RH
(ed), Manual of clinical microbiology, 8th ed. ASM Press,
Washington, DC.
148. Centers for Disease Control and Prevention 2/2002,
Emergency Preparedness and Pesponse. Bioterrorism
Agents/Diseases. CDC. www.bt.cdc.gov/agent/agentlist-
category
149. Morse SA, Butel JS, Brooks GF, Carroll KC. 2007.
Jawetz, Melnick, and Adelberg’s medical microbiology series,
24th ed. McGraw-Hill, New York.
150. Kaufmann SH, van Embden JD. 1993. Tuberculosis: a
neglected disease strikes back. Trends Microbiol 1(1):2–5.
151. Contreras MA, Cheung OT, Sanders DE, Goldstein RS.
1988. Pulmonary infection with nontuberculous mycobac-
teria. Am Rev Respir Dis 137:149–152.
152. Kirschner RA Jr., Parker BC, Falkinham III, JO 1992.
Epidemiology of infection by nontuberculous mycobacte-
ria. Am Rev Respir Dis 145:271–275.
153. Schaal K. 1991. Medical and microbiological problems
arising from airborne infection in hospitals. J Hosp Infec
18:451–459.

154. Olsen GJ. 1994. Archaea, Archaea, every where. Nature
371(6499):657–658.
155. Michel O. 2001. Role of house-dust endotoxin exposure in
aetiology of allergy and asthma. Mediat Inflamm 10(6):301.
156. Rylander R, Fogelmark B. 1994. Inflammatory responses
by inhalation of endotoxin and (1→3)-beta-D-glucan.
Am J Ind Med 25(1):101–102.
157. Castellan RM, Olenchock SA, Hankinson JL, Millner
PD, Cocke JB, Bragg CK, Perkins HH, Jacobs RR.
1984. Acute bronchoconstriction induced by cotton dust:
dose-related responses to endotoxin and other dust factors.
Ann Intern Med 101(2):157–163.
158. Rylander R, Vesterlund J. 1982. Airborne endotoxins
in various occupational environments. Prog Clin Biol Res
93:399–409.
159. Mishra S, Ajello L, Ahearn D, Burge H, Kurup V,
Pierson D, Price D, Samson R, Sandhu R, Shelton B.
1992. Environmental mycology and its importance to pub-
lic health. Med Mycol 30(S1):287–305.
160. Gravesen S. 1979. Fungi as a cause of allergic disease.
Allergy 34(3):135–154.
161. Siersted H, Gravesen S. 1993. Extrinsic allergic alveolitis
after exposure to the yeast Rhodotorula rubra. Allergy 48
(4):298–299.
162. Garrett M, Rayment P, Hooper M, Abramson M, Hooper
B. 1998. Indoor airborne fungal spores, house dampness
and associations with environmental factors and respiratory
health in children. Clin Exp Allerg 28(4):459–467.
163. Rylander R, Megevand Y. 2000. Environmental risk fac-
tors for respiratory infections. Arch Environ Health 55(5):
300–303.
164. Kilpeläinen M, Terho E, Helenius H, Koskenvuo M.
2001. Home dampness, current allergic diseases, and respi-
ratory infections among young adults. Thorax 56(6):
462–467.
165. Ebbehøj NE, Hansen MØ, Sigsgaard T, Larsen L. 2002.
Building-related symptoms and molds: a two-step interven-
tion study. Indoor Air 12(4):273–277.
166. Samson RA, Houbraken J, Summerbell RC, Flannigan B,
Miller JD. 2001. Common and important species of fungi
and actinomycetes in indoor environments, p 287–490.
In Flannigan B, Samson RA, Miller JD (ed), Microorgan-
isms in home and indoor work environments. Taylor & Francis,
New York, NY.
167. Bloom E, Nyman E, Must A, Pehrson C, Larsson L. 2009.
Molds and mycotoxins in indoor environments—a survey
in water-damaged buildings. J Occup Environ Hyg 6(11):
671–678.
168. Häggblom P. 1987. De novo synthesis of alternariol in
conidia of Alternaria alternata. J Gen Microbiol 133(12):
3527–3529.
169. Palmgren MS, Lee LS. 1986. Separation of mycotoxin-
containing sources in grain dust and determination of their
mycotoxin potential. Environ Health Persp 66:105.
170. Wicklow DT, Shotwell OL. 1983. Intrafungal distribution
of aflatoxins among conidia and sclerotia of Aspergillus fla-
vus and Aspergillus parasiticus. Can J Microbiol 29(1):1–5.
171. Miller JD. 1992. Fungi as contaminants in indoor air.
Atmos Environ Part A Gen Top 26(12):2163–2172.
3.2.1. Introduction to Aerobiology ▪ 3.2.1-15
172. Jarvis BB, Sorenson W, Hintikka E-L, Nikulin M, Zhou
Y, Jiang J, Wang S, Hinkley S, Etzel RA, Dearborn D.
1998. Study of toxin production by isolates of Stachybotrys
chartarum and Memnoniella echinata isolated during a study
of pulmonary hemosiderosis in infants. Appl Environ Microb
64(10):3620–3625.
173. Rao CY, Brain JD, Burge HA. 2000. Reduction of pulmo-
nary toxicity of Stachybotrys chartarum spores by methanol
extraction of mycotoxins. Appl Environ Microb 66(7):
2817–2821.
174. Panaccione DG, Coyle CM. 2005. Abundant respirable
ergot alkaloids from the common airborne fungus Aspergil-
lus fumigatus. Appl Environ Microb 71(6):3106–3111.
175. Wady L, Larsson L. 2005. Determination of microbial vol-
atile organic compounds adsorbed on house dust particles
and gypsum board using SPME/GC-MS. Indoor Air 15
(Suppl 9):27–32.
176. Fischer G, Schwalbe R, Möller M, Ostrowski R, Dott W.
1999. Species-specific production of microbial volatile
organic compounds (MVOC) by airborne fungi from a
compost facility. Chemosphere 39(5):795–810.
177. Brown JH, Cook KM, Ney FG, Hatch T. 1950. Influence
of particle size upon the retention of particulate matter in
the human lung. Am J Public Health and the Nations Health
40(4):450–480.
178. Lawande R, Abraham S, John I, Egler L. 1979. Recovery
of soil amebas from the nasal passages of children during the
dusty harmattan period in Zaria. Am J Clin Pathol 71(2):
201–203.
179. Rivera F, Roy-Ocotla G, Rosas I, Ramirez E, Bonilla P,
Lares F. 1987. Amoebae isolated from the atmosphere of
Mexico City and environs. Environ Res 42(1):149–154.
180. Burrell R. 1991. Microbiological agents as health risks in
indoor air. Environ Health Persp 95:29.
181. Griffiths W, DeCosemo G. 1994. The assessment of bio-
aerosols: a critical review. J Aerosol Sci 25(8):1425–1458.
182. Kingston D. 1971. Selective media in air sampling: a
review. J Appl Microbiol 34(1):221–232.
183. Rao C, Cox-Ganser J, Chew G, Doekes G, White S. 2005.
Use of surrogate markers of biological agents in air and set-
tled dust samples to evaluate a water-damaged hospital.
Indoor Air 15(s9):89–97.
184. Ranalli G, Principi P, Sorlini C. 2000. Bacterial aero-
sol emission from wastewater treatment plants: culture
methods and bio-molecular tools. Aerobiologia 16(1):39–
46.
185. Alvarez A, Buttner M, Toranzos G, Dvorsky E, Toro A,
Heikes T, Mertikas-Pifer L, Stetzenbach L. 1994. Use of
solid-phase PCR for enhanced detection of airborne micro-
organisms. Appl Environ Microb 60(1):374–376.
186. Stark KD, Nicolet J, Frey J. 1998. Detection of Myco-
plasma hyopneumoniae by air sampling with a nested PCR
assay. Appl Environ Microbiol 64(2):543–548.
187. Vesper S, Dearborn DG, Yike I, Allan T, Sobolewski J,
Hinkley SF, Jarvis BB, Haugland RA. 2000. Evaluation
of Stachybotrys chartarum in the house of an infant with
pulmonary hemorrhage: quantitative assessment before,
during, and after remediation. J Urban Health: Bull NY
Acad Med 77(1):68–85.
 Sampling for Airborne Microorganisms
SERGEY A. GRINSHPUN, MARK P. BUTTNER, GEDIMINAS MAINELIS,
AND KLAUS WILLEKE
3.2.2
Microbiologists have confronted the challenges of sampling
and analysis of airborne microorganisms since the early
20th century. Today, the concentration and composition
of airborne microorganisms are of interest in diverse areas
such as agricultural and industrial settings, hospitals, home
and office environments, and military installations. In all of
these applications, the term “bioaerosol” is used to refer to
airborne biological particles, such as bacterial cells, fungal
spores, viruses, and pollen grains, and to their fragments
and by-products. A wide variety of bioaerosol sampling and
analysis methods have been used, and new methods are being
developed (1–42). However, several problems remain to be
solved. For instance, no single sampling method is suitable
for the collection and analysis of all types of bioaerosols,
and no standardized protocols are currently available. There-
fore, data from different studies are often difficult to compare
because of differences in sampler designs, collection times,
airflow rates, collection media, and analysis methods. In addi-
tion, human exposure limits have not been established for
bioaerosols because of the lack of exposure, dose, and response
data. This complicates the use of sampling results for risk
assessment.
Measurement of airborne microorganisms with a bioaero-
sol sampler often aims at documenting the presence of specific
sources (34, 43–68). However, when no biological particles
are collected and identified by air sampling, one cannot
definitely conclude that such particles and their sources are
absent. This lack of certainty always exists, particularly due
to the small air sampling volume of conventional bioaerosol
samplers over a limited period of time. The sampling time
and the sampler’s location may not coincide with the time
and place of bioaerosol release from the source. However,
the presence of a source does not necessarily imply that
the bioaerosol level is elevated because release from sources
is often sporadic. It is important to carefully consider the
objectives of sampling before any samples are taken. After
determining what information is desired, an appropriate sam-
pling and analysis method can be selected and incorporated
into the study design. The purpose of this chapter is to present
various bioaerosol sampling methods that would allow an
intelligent selection of instrumentation and techniques.
The principles of bioaerosol sampling are presented, followed
by a review of sampling methods and techniques currently
available, including the results of performance evaluations
of the various sampler types. Equipment calibration and air
doi:10.1128/9781555818821.ch3.2.2
3.2.2-1
sampling considerations such as collection times and the
number of samples are discussed. Analysis methods are
addressed in another chapter, beginning with traditional cul-
ture and total count methods and concluding with some more
recently developed biochemical and molecular techniques
(5, 10, 13, 22, 23, 25, 27, 47, 69–79).
BIOAEROSOL SAMPLING
The collection of bioaerosol samples is performed through
active air sampling. The advantages and disadvantages
of active sampling have been discussed in the literature
(11, 33, 35, 80). The objective of active bioaerosol sampling
is the efficient removal and collection of representative bio-
logical particles from the air in a manner that does not affect
the ability to detect the organisms (e.g., no alteration in
culturability or biological integrity) (81). This ability is
dependent on the physical and biological characteristics of
the organisms and on the selected sampling method, includ-
ing the physical features of the sampling instrument (11, 82).
The motion of a bioaerosol particle in the air is governed
by the same laws of physics that apply to a biologically inert
particle. The principal collection methods used in quanti-
tative bioaerosol sampling are impaction, impingement, and
filtration. Electrostatic precipitation has also been actively
explored as a method for bioaerosol sampling due to its poten-
tial to better preserve microorganism characteristics. Gravita-
tional settling has been employed or explored for bioaerosol
sampling (8, 24), but this method is not quantitative and
often nonrepresentative.
Methods of Collection
Impaction
The impaction method separates particles from the air stream
by utilizing the inertia of the particles to force their deposition
onto a solid or semisolid collection surface. The collection
surface is usually an agar medium for culture-based analysis
or an adhesive-coated surface that can be analyzed micro-
scopically. The impaction process depends on the inertial
properties of the particle (e.g., density, diameter, and veloc-
ity) and on the physical parameters of the impactor (e.g.,
the inlet nozzle dimensions and the airflow pathway) (33).
Air enters the sampler through one or more inlet nozzle(s),
and as the air stream is directed at the collection surface, it
3.2.2-2 ▪ AIR
is forced to make a turn. Particles with sufficient inertia are
not able to follow the turning air stream and are impacted.
Particles with lower inertia remain airborne with the airflow
(Fig. 1a). Centrifugal impaction also uses inertial forces to
separate the particles from the air stream, but in a radial geo-
metry (Fig. 1b). Due to a limited collection surface area,
impactors usually operate at relatively short sampling periods
(sometimes as short as minutes). The sampling time can be
extended if employing impactors with a rotating sampling
surface.
Liquid Impingement
Liquid impingement is similar to impaction in that the iner-
tial force acting on the particle is the principal force removing
it from the air. However, the collection medium is a liquid
(usually a dilute buffer solution) and the collected micro-
organisms move around freely in the bubbling liquid
(Fig. 1c). As a result, aggregates of cells may be broken apart.
Particles remaining in the air stream may diffuse to the surface
of a bubble and be transferred to the collection buffer. The
collection of bioaerosol particles in a liquid medium allows
division of the sample and the potential application of several
analysis methods, as described later.
Filtration
Filtration achieves the separation of particles from the air
stream by passage of the air through a porous medium, such
as membrane or fiber filter. The collection of particles
depends on their physical properties (e.g., size, shape, and
density), the filter pore size, and the airflow rate (83). Inertial
forces and other mechanisms such as interception, diffusion,
and electrostatic attraction result in the collection of particles
on the surface of the filter (Fig. 1d). Simultaneous action of
all of these forces removes particles smaller than the pore
FIGURE 1 Mechanisms of collection utilized in bioaerosol sam-
pling. (a) Solid plate impaction; (b) centrifugal impaction; (c) liquid
impingement; (d) filtration. Fin or Finertial, inertial force. (Adapted
from Nevalainen et al. (29) with kind permission from Elsevier
Science Ltd., Kidlington, UK).
doi:10.1128/9781555818821.ch3.2.2.f1
size of the filter (16). The collected microorganisms can be
extracted from the filter for subsequent analyses. Direct
examination of the filter by microscopy has also been used.
Electrostatic Precipitation
Commonly used sampling methods (e.g., inertia-based
impaction and impingement and filter-based methods)
have been shown to adversely affect the viability of microor-
ganisms being sampled (84, 85). Therefore, other methods
are being explored for bioaerosol sampling. Among those,
electrostatic precipitation has gained increased attention
over the past few years due to its potential to preserve micro-
bial viability and low power requirements for its operation.
The electrostatic precipitation method has been extensively
used for air quality control purposes, and current research
shows that it could be successfully adapted for bioaerosol
sampling as well. In this technique, air stream with biological
particles is drawn into a sampler where the particles are elec-
trically charged and then subjected to an electrostatic field
that deposits them onto a collection surface. The deposited
microorganisms can then be extracted from the surface and
analyzed by various techniques (24). The deposition velocity
is much lower compared to inertia-based sampling techni-
ques, thus resulting in lower sampling stress (86).
Gravitational Settling
Gravitational settling, or depositional sampling, is a non-
quantitative collection method in which an agar medium
is exposed to the environment and airborne organisms are
collected primarily by gravity. This method is often used
because it is inexpensive and easily performed. However,
the collection of airborne microorganisms using this method
is affected by the sizes and shapes of the particles and by the
motion of the surrounding air (33). As a result, large particles
are more likely to be deposited onto the collection surface (8).
This can lead to misrepresentation of the prevalence of air-
borne microorganisms and the exclusion of smaller particles
from the collection (62). In addition, the concentration of
the airborne microorganisms cannot be determined by gra-
vity sampling because the volume of air from which the
particles originate is unknown. Sampling based on gravita-
tional settling has been compared with various methods
that pass a known volume of air to the collection medium.
The results show that the airborne-microorganism concentra-
tions derived from gravity sampling are not qualitatively or
quantitatively accurate and do not compare favorably with
those obtained by other sampling methods (61, 62, 87, 88).
Sampler Types
There is a wide variety of commercially available bioaerosol
samplers (11, 16, 33). The selection of a sampler depends
on a number of factors, such as sampler performance, expec-
ted bioaerosol concentration, and the analysis method.
Prior to initiating an investigation, there should be an under-
standing of the specific objectives of air monitoring and the
limitations of the various sampling methods. Although stand-
ardized sampling protocols are not available at the present
time, guidelines have been published to assist in the selection
of bioaerosol sampling protocols (11, 16). Some of the more
widely used bioaerosol sampling methods are listed in Table 1
and discussed in the following sections. Also included are
some novel sampling concepts that have not yet translated
into commercial products, but represent advancement in
the field.
 3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-3

TABLE 1 General characteristics of several commercially available bioaerosol samplers

Airflow rate(s) Sample analysis
Sampler(s) Collection medium(a) Comments
(liters/min) method(s)
Impaction
Andersen viable impactors, 90 mm agar platesa 28.3 Culture Particle size discrimination
1-, 2-, and 6-stage (2- and 6-stage models);
Andersen-type samplers vacuum pump required;
featuring the same counts corrected for multiple
configuration impaction
BioCassette 90 mm agar plate 28.3 Culture Disposable impactor; pump
required; counts corrected for
multiple impaction.
Biotest RCS, RCS Plus, Agar wells in 40 (RCS), 50 Culture Portable; battery operated
RCS High Flow plastic strips (RCS Plus), 100
(RCS High Flow)
MAS-100 90 mm agar plates 100 Culture Portable; battery operated;
counts corrected for multiple
impaction
Microflow 90 mm agar plates 30–120 Culture Same as above
BioCulture 90 mm agar plates 120 Culture Same as above
SMA MicroPortable 90 mm agar plates 28–141 Culture Same as above
Millipore Air Tester Plates supplied by 140–180 Culture Same as above
manufacturer
Sampl’Air 90 mm agar plates 100 Culture Same as above
SAS Super 100, SAS 55 mm or 84 mm 100 (Super 100) Culture Same as above
Super 180 contact plates 180 (Super 180)
(RODAC), 90 mm
agar plates
Burkard portable air 90 mm agar plates 10 or 20 Culture Same as above
sampler for agar plates
SpinAir 60 mm RODAC plates, 60–100 Culture Portable; battery operated
90 mm agar plates
Time-resolved slit impactors
Mattson/Garvin 220 and Agar; 150-mm-diam 28.3 Culture Time discrimination up to 1 h
P-320 petri dishes
New Brunswick slit-to-agar, Agar; 150-mm-diam 15–50 (STA-203), Culture Time discrimination up to 1 h;
STA-203, and STA-204 petri dishes 15–30 (STA-204) vacuum pump required for
some models
BIAP Slit Sampler 140 mm agar plate 100 Agar Time discrimination; 10, 30, 60,
180 min sampling time
Spore traps
Burkard 7-day Recording Adhesive-coated surface, 10 Microscopy Determination of total fungal
Volumetric Spore Trap; tape, or glass slide spores and pollen; motor or
Burkard Personal pump is part of the apparatus
Volumetric Spore Trap.
Via-Cell Adhesive-coated surface 15 Microscopy, Determination of viable and
culture, and total fungal spores;
PCR analysis disposable; pump required
Air-O-Cell, MoldSNAP Adhesive-coated surface 5 (MoldSNAP) Microscopy Determination of total fungal
sampling cassettes 15 (Air-O-Cell) spores and pollen; disposable;
pump required
Micro-5 Microcell cassette Adhesive-coated surface 5 Microscopy Same as above
Allengenco D cassettes Adhesive-coated surface 15 Microscopy Same as above
cyclex-d cassette Adhesive-coated surface 20 Microscopy Same as above
VersaTrap cassette Adhesive-coated surface 5–30 (15 is nominal) Microscopy Same as above
BioSIS Slit Impaction Air Adhesive-coated glass Microscopy Determination of total fungal
Sampler slide spores and pollen; re-usable;
pump required.
(Continued on next page)
3.2.2-4 ▪ AIR

TABLE 1 General characteristics of several commercially available bioaerosol samplers (Continued )

Airflow rate(s) Sample analysis
Sampler(s) Collection medium(a) Comments
(liters/min) method(s)
Liquid-based samplers
All-glass impingers Liquid, 20 ml 12.5 Culture, microscopy, High and low bioaerosol
(AGI-30 and AGI-4) biochemical assay, concentrations; vacuum
PCR, immunoassay pump required
BioSampler Liquid, 5 or 20 ml 12.5 Same as above High and low bioaerosol
concentrations; short- and
long-term sampling; viscous
and nonviscous fluids;
vacuum pump required
Burkard multistage liquid Liquid, 6 ml for 20 Same as above High and low bioaerosol
impinger each stage concentrations; particle size
discrimination in 3 fractions;
vacuum pump required
BioGuardian Air Sampler Liquid, 10–15 ml 100–1000 Same as above High and low bioaerosol
concentrations; vacuum
pump included; replacement
of liquid lost to evaporation
SpinCon Advanced Air Liquid, 10 ml 450 Same as above Same as above
Sampler
OMNI 3000 Liquid, 9 ml 300 Same as above Same as above
Coriolis Delta Liquid, 10–15 ml 100–300 Same as above Same as above
SASS 2300 and SASS Liquid, 4–6 ml 390 (SASS 2300) Same as above Same as above
2400 multistage 40 (SASS 2400)
wetted-wall cyclone
BioCapture 650 Liquid, 2–5 ml 200 Same as above Same as above; portable; uses
disposable cartridges with
collection liquid
Filtration
25-, 37-, or 47-mm pore Filter membrane 1–50 Microscopy, Loss of culturable vegetative
size filter cassettes biochemical assay, cells; portable; useful for
PCR, immunoassay, personal monitoring; vacuum
culture (spores) pump required
Sartorius AirPort MD 8 Gelatin membrane 42–133 Culture Gelatin filter reduces
air sampler filter desiccation stress,
high-volume sampling; virus
collection
25 mm gelatin filter Gelatin membrane User and sampler Culture, PCR Gelatin filter reduces
filter dependant (inhibitions desiccation stress
reported)
a
In United States, 90-mm plates are sometimes replaced by 100-mm plates.

Impactor Samplers a multinozzle impactor plate resembles a sieve, it is sometimes

Impaction is the most commonly used method of collection
for airborne microorganisms because of its ease of use and abil-
ity to collect microorganisms directly onto agar growth
medium without the need for postcollection sample process-
ing. A variety of impactor samplers are commercially available
(Table 1). They differ in the dimensions of their nozzles and
jet-to-plate distances, the shapes of the nozzles, the numbers
of nozzles, the numbers of stages, and sampling flow rates. Sin-
gle nozzle impactors feature both rectangular and circular
inlet nozzles. If a nozzle is rectangular, the impactor is referred
to as a slit sampler. The nozzles of multinozzle impactors are
usually circular in shape and can number from 12 (SMA
MicroPortable, Veltek Associates, Phoenixville, PA) to
1,000 (Millipore Air Tester, Millipore, Billerica, MA). Since
referred to as a sieve sampler. If there are several stages with
successively smaller nozzles, the sampler is referred to as a
cascade impactor.
The Andersen multistage impactor sampler (Thermo Sci-
entific, Franklin, MA; formerly Graseby-Andersen, Smyrna,
GA) was introduced in 1958 (89) and remains one of
the most commonly used devices for the collection of cul-
turable airborne microorganisms (89). It has been recom-
mended and used as a reference sampler in bioaerosol
studies (11, 90, 91). This sampler is designed to operate at
28.3 liters/min and collect particles of ≤10 µm in diameter.
A six-stage version consists of six stages with 400 precision-
drilled holes in each stage, with decreasing nozzle diameters,
such that successive stages collect progressively smaller
particles, for example, stage 1 captures particles >7 µm, and

stage 6 captures particles >0.65 µm. The stages are designed
to mimic the particle deposition in different regions of the
human respiratory system. In addition, multiple stages can
also provide size distribution of bioaerosol particles. Theoret-
ical and experimental parameters of each stage are provided in
Table 2. The original six-stage Andersen impactor later was
simplified by using the last (sixth) stage only (90), which
allowed collecting all particles above its cut-off size on a single
agar plate. A two-stage model of the impactor has also been
used (17, 92, 93). As patent restrictions expired, Andersen-
type impactors have been offered by other manufacturers,
including one-, two-, and six-stage versions by Tisch Envi-
ronmental (Cleves, OH), 200- and 400-nozzle single-stage
versions by SKC (Eighty Four, PA), the E6 Bioaerosol Sam-
pler by Environmental Monitoring Systems (Charleston,
SC), and others. Recently, a disposable version of the Ander-
sen sampler, the BioCassette (EMLab P&K, Cherry Hill, NJ),
was developed and used in bioaerosol studies (94).
Andersen-type viable impactors require external
vacuum source for their operation, which may limit their
application under certain circumstances. Therefore, battery-
powered multinozzle portable microbial samplers are becom-
ing increasingly popular due to their ease of use, portability,
and ability to achieve high sampling flow rates. The Portable
Air Sampler (Burkard Manufacturing Company, Hertford-
shire, UK), MAS-100 (EMD Chemicals, Gibbstown, NJ),
Microflow (Aquaria srl, Lacchiarella, Italy), BioCulture
(A.P. BUCK, Orlando, FL), SMA MicroPortable (Veltek
Associates, Phoenixville, PA), Millipore Air Tester, Sampl’
Air (AES Chemunex, Cranbury, NJ), SAS Super 180 (Bio-
science International, Rockville, MD), Spin Air (IUL Instru-
ments, Barcelona, Spain) and others presented in Table 1
operate at flow rates ranging from 10 liters/min (Burkard)
to 180 (SAS Super 180) liters/min and collect bioaerosol
particles onto single agar plates. The Reuter Centrifugal
Sampler (RCS), RCS plus, and RCS High Flow (Biotest
AG, Dreieich, Germany) portable samplers centrifugally
collect particles onto agar strips. The single-stage Andersen-
type impactor can also be used as a portable bioaerosol sam-
pler by deploying a suitable battery-operated pump (95).
Due to power requirements, multistage impactors are not
available in portable configurations. The MAS-100 sampler
has been adapted to collect bioaerosols into liquid as well
(96, 97). Unlike other multinozzle impactors, SpinAir
sampler features a rotation mechanism to increase the collec-
tion surface.
Slit impactors deposit the bioaerosol onto an agar surface
for the estimation of numbers of viable cells. Examples are
the Mattson/Garvin air sampler (Barramundi, Homosassa,
FL), the New Brunswick slit-to-agar sampler (New Bruns-
wick Scientific, Edison, NJ), and the BIAP Slit Sampler
(Scantago, Denmark). All these samplers feature a rotation
mechanism to provide even distribution of colonies on a plate
and also allow temporal analysis of collected culturable
microorganisms.
A separate class of microbial impactors features an
adhesive-coated surface for an improved retention of the col-
lected particles, usually fungal spores or pollen grains, which
are subjected to microscopic enumeration. Examples of this
type of sampler are the Air-O-Cell, Via-Cell, MoldSnap
(Zefon International, Ocala, FL), Micro5, Allergenco D,
cyclex-d, BioSIS Slit Impaction Air Sampler (EMS), Versa-
Trap (SKC), and Personal Volumetric Air Sampler (Burkard
Manufacturing). The latter sampler has a built-in pump,
while the other samplers require an external pump to provide
flow rates ranging from 5 liters/min (Micro5, MoldSNAP)
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-5
to 30 liters/min (maximum for BioSIS and VersaTrap).
The Air-O-Cell, VersaTrap, Allergenco D, BioSIS, and Per-
sonal Volumetric Air Sampler are slit impactors, whereas
the cyclex-d and Micro-5 are circular impactors. The BioSIS
and Personal Volumetric Air Sampler can be used for multi-
ple samples by changing the sampling medium, while the
other impactors on the list, also called “spore traps,” are dis-
posable. The slit glass-slide impactors, such as the Personal
Volumetric Air Sampler and the Air-O-Cell sampling cas-
sette, are typically used for indoor spore sampling (98–100).
The Burkard 7-day Volumetric Spore Trap with a single-stage
rotating drum can be used to sample outdoor fungi and pollen
for up to 7 days (101, 102).
Liquid-Based Samplers
Liquid impinger samplers are commonly used for the collec-
tion of bioaerosol particles over a wide range of airborne-
particle concentrations (Table 1). For analysis, the liquid
sample can be concentrated by filtration or diluted by liquid
addition, depending on the concentration of microorganisms.
Several culture media can be inoculated with aliquots of the
collection medium for the assessment of groups of microor-
ganisms with different culture requirements. The liquid sam-
ples may also be analyzed by biochemical, immunological,
and molecular biological assays to detect the presence of spe-
cific microorganisms, culturable or nonculturable.
Design and adaptation of impingers for bioaerosol sam-
pling was mentioned as early as 1947 (103). The original
design lead to the development of the AGI-30 all-glass
impinger (Ace Glass, Vineland, NJ) which features the dis-
tance of 30 mm between the sampling nozzle and the bottom
of the collection vessel and operates at 12.5 liters/min. The
AGI-4 model features a distance of 4 mm between the sam-
pling nozzle and the vessel’s bottom to improve the collection
efficiency. However, added sampling stress may result from
impaction against the glass bottom of the sampler, leading
to a loss of viability of cells. Both samplers have a curved
inlet tube designed to simulate the nasal passage, making
this sampler useful for studying the respiratory infection
potential of bioaerosols (16, 104, 105). For other applica-
tions, the inlet tube is washed with a known volume of col-
lection fluid to recover nonrespirable airborne particles.
The Burkard multistage liquid impinger is a stainless steel
sampler that collects particles in three size fractions (i.e.,
>10, 4–10, and <4 µm).
While liquid impingement has numerous advantages,
most of the currently available impingers that are efficient
in collecting small bioaerosol particles require a very high
sampling velocity, which usually results in a violent motion
of the collection fluid. The latter causes reaerosolization
of the initially collected bioparticles and stress that leads
to the viability loss (106–109). Conventional impingers
can be used only with water or liquids that have about
the same viscosity as water. The BioSampler (SKC) com-
bines impingement into a liquid with centrifugal motion
(39). As a result, this sampler can be used with buffer or
with viscous collection fluids (e.g., heavy white mineral
oil). Having the same inlet geometry and the same airflow
rate of 12.5 liters/min as AGI samplers, the BioSampler
achieves particle collection by drawing aerosol through
three nozzles that are directed at an angle toward the inner
sampler wall. During normal operation, the liquid swirls
upward on the sampler’s inner wall and removes collected
particles. Depending on the model, the BioSampler can be
operated with 5 or 20 ml of collection fluid at the same
sampling flow rate of 12.5 liters/min.
3.2.2-6 ▪ AIR

TABLE 2 Calculated and reported cutoff diameters (d50s) for several commercially available bioaerosol samplers
d50 (µm)
Reference(s) for performance
Samplers Calculated Reported evaluation studies
(reference) (reference[s])
Impaction
Andersen 6-stage viable impactor (9, 17, 63, 87, 89, 90, 92, 137, 186, 196–198)
Stage 1 0.24 (16), 6.61 (29) 7.0
Stage 2 4.21 4.7
Stage 3 2.86 3.3
Stage 4 1.84 2.1
Stage 5 0.94 1.1
Stage 6 0.58 0.65
Andersen 2-stage viable impactor (17, 82, 92, 96, 97, 137, 199)
Stage 0 6.28 8.0
Stage 1 0.83 0.95
Andersen 1-stage viable impactor 0.58 0.65 (17, 82, 90, 96, 97)
BioCassette (200)
Biotest RCS 7.5 3.8 (17, 32, 201, 202)
Biotest RCS Plus 0.82 (12)
Biotest RCS High Flow 1.7 (139) 1.2 (139) (119, 139, 150)
MAS-100 1.5 (139) 1.7–2.5 (139) (96, 97, 119, 139, 150)
Microflow 17.5 (139) >10 (139) (119, 139, 150)
BioCulture 8.1 (139) 7 (139) (119, 139, 150)
SMA MicroPortable 13.7 (139) >10 (139) (119, 139, 150)
Millipore Air Tester 0.9–1 (139) 2.3–2.5 (139) (119, 139, 150)
Sampl’Air (203)
SAS Super 180 1.3 (139) 2.1–3.0 (139) (119, 139, 150)
Burkard portable air sampler for agar plates 4.18 (204), 4 (205)
SpinAir (203)
Time-resolved slit impactors
Mattson/Garvin 220 and P-320 0.53 (17) (17, 32)
New Brunswick slit-to-agar, STA-203, and (168)
STA-204
Spore traps
Burkard spore traps, 24-h, 7-day (87, 198, 206)
Standard nozzle 3.70
High-efficiency nozzle 2.17
Burkard Personal Volumetric Spore Trap 2.52 2.5 (33)
Air-O-Cell 2.3 (147), 2.6 (205) (3, 147)
Micro-5 Microcell ∼1 (148) (148)
cyclex-d cassette ∼1 (148), 2 (28, 205) (148)
Liquid-based samplers
All-glass impinger AGI-30 0.30 (9, 17, 37, 82, 89, 96, 97, 137, 169, 186, 196)
BioSampler <0.3 (39) (39, 96, 97, 107, 108, 169)
Burkard multistage liquid impinger
Stage 1 10 (199, 207)
Stage 2 4–10
Stage 3 4
SpinCon Advanced Air Sampler (208)
OMNI 3000 (209)
Coriolis Delta 1–2 (210)
SASS 2300 and SASS 2400 wetted-wall (167)
cyclone
Sartorius AirPort MD8 (209, 211, 212)

Another approach to bioaerosol sampling with a liquid
utilizes a porous medium submerged into a liquid layer so
that the aspirated air is blown through it. As a result, the aero-
sol flow is split into many very small bubbles and particulates
are effectively removed on the walls of the impinger’s vessel.
A prototype sampler has been developed and evaluated (1, 2,
110) and used in combination with real-time PCR to rapidly
detect airborne viruses (111), but it is not yet commercially
available.
In recent years, high-volume liquid-based bioaerosol
samplers have been developed. InnovaTek (Richland,
WA) introduced the BioGuardian Air Sampler, which
operates from 100 to 1,000 liters/min and collects samples
into 10–15 ml of liquid. The SpinCon Advanced Air
Sampler, supported by InnovaPrep (Drexel, MO), samples
at 450 liters/min for up to 6 h and concentrates samples
into 10 ml of liquid. Another sampler supported by the same
company, the OMNI 3000, has been designed for portability
and collects biological particles at a flow rate of 300 liters/min
into 9 ml of collection fluid or less for up to 6 h. A somewhat
recent addition to the high-flow-rate liquid samplers for
bioaerosols is the Coriolis family of samplers (Bertin Technol-
ogies, France). Coriolis Delta operates at flow rates from 100
to 300 liters/min and samples particles into 10–15 ml of
liquid for up to 6 h. For long-term operation, the SASS
2300 wetted-wall cyclone (Research International, Monroe,
WA) collects particles into 4–6 ml of liquid at a flow rate of
390 liters/min and can be operated for several days. It has
been successfully used to detect the airborne viral pathogens
that cause exotic Newcastle disease and hoof-and-mouth
disease, as well as strains of avian flu virus in California
commercial poultry flocks (112). A 40 liters/min version of
the sampler (SASS 2400) is also available. The BioCapture
650 (FLIR Systems, Wilsonville, OR) is a portable sampler
that uses a sampling flow rate of 200 liters/min and collects
particles into cartridges with 5 ml of liquid.
Recently a microcentrifuge-tube based personal bioaerosol
cyclone has been developed and evaluated (113–116). In this
sampler, the particles are collected into a dry cyclone and
then removed by washing off the cyclone. The cyclone was
shown to be suitable for collection of ambient fungal spores,
detection of infectious airborne influenza virus, and sample
analysis by microscopy and PCR (116, 117). The sampler’s
aspiration efficiency at 4 liter/min has been found to follow
the inhalable convention of the American Conference of
Governmental Industrial Hygienists within the tested
particle size range (up to 10 µm) (42).
Filtration Sampling
The collection of airborne microorganisms onto a filter
material is used in bioaerosol monitoring due to its simplicity,
low cost, and versatility (Table 1). Air samples are usually
collected on 25-, 37-, or 47-mm diameter filter membranes
housed in various filter holders, or sampling heads. A variety
of filtration devices are available from several manufacturers
(e.g., Gelman Sciences, Ann Arbor, MI; Millipore, Bedford,
MA; Nuclepore, Corning Costar, Cambridge, MA; Poretics,
Livermore, CA; and SKC). Polycarbonate, mixed cellulose
ester, or polyvinyl chloride filter material may be used
depending on the nature of the bioaerosol and the method
of sample analysis (16, 83). Filter membrane pore sizes range
from 0.01 to 10 µm, with air sample flow rates ranging from 1
to 50 liters/min. Filtration sampling is adaptable to a variety
of assays (Table 1), but loss of viability of vegetative cells
may occur, presumably due to desiccation stress during sam-
pling (17, 82, 85, 118, 119). The AirPort MD8 air sampler
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-7
(Sartorius AG, Göttingen, Germany) collects airborne
microorganisms on a gelatin filter to reduce desiccation stress.
A 25-mm gelatin filter (SKC) can be utilized for the same
purpose with an appropriate filter holder or sampler, such as
the Button Aerosol Sampler or IOM cassette (96, 97, 119,
120). The gelatin membrane can be directly incubated onto
the agar medium of choice for culture analysis or dissolved
in a liquid for culture or other analyses.
The use of a size-selective filter housing as well as a size-
selective adaptor allows one to collect bioaerosol size fractions
of interest, such as inhalable particles only (3). The use of
size-selective polyurethane foams has been explored to sam-
ple thoracic and respirable bioaerosol fractions (121, 122).
Electrostatic Precipitation Samplers
Attempts to utilize electrostatic precipitation for bioaerosol
sampling were made more than 60 years ago (123). In the
1960s, a large-volume electrostatic air sampler operating at
a flow rate of 1,000 liters/min was developed to recover
airborne viruses (124). Due to the advancement of other
sampling techniques, the electrostatic method for bioaerosol
collection has not been widely utilized. However, concerns
over the negative effects of inertia-based sampling on micro-
bial viability and interest in low-power sampling methods
led to a renewed interest in applying electrostatic collection
for bioaerosol sampling. An existing Electrostatic Aerosol
Sampler (model 3100, TSI) was successfully modified to col-
lect bacteria on agar (86) and the principles were applied to
develop a stand-alone electrostatic collector for bioaerosols
for laboratory and field studies (24, 125, 126). In an effort
to simplify sampler design, native electrostatic charges carried
by microorganisms were successfully utilized in a sampler to
collect them on agar without the need for any additional
charging (127). In the past few years, several new electrostatic
sampler concepts have been developed, including one where
a narrow collection electrode covered by a superhydrophobic
(“lotus leaf” type) substance and a specifically shaped ground
electrode allows for efficient collection, removal, and con-
centration of bacteria and fungi in liquid droplets as small
as 5 µl (40, 41, 128). An electrostatic collector for bioaerosols
featuring circulating liquid has also been proposed (129).
All of these recently developed conceptual models feature
relatively low sampling flow rates. However, an electrostatic
bioaerosol collector with a sampling flow rate as high as
3,500 liters/min has also been presented (130, 131). How-
ever, these models are not yet commercially available.
Zaromb Research Corporation (Burr Ridge, IL) offers a com-
mercially available wet electrostatic precipitator (WEP-2)
capable of sampling biological agents at a rate of about 500
liters/min. Advantages of electrostatic collectors are not only
low microorganism deposition velocity leading to an im-
proved preservation of their properties but also low power
consumption. Similar to other liquid samplers, wet electro-
static precipitators offer the ability to analyze samples using
a variety of techniques, including microscopy, quantitative
PCR, ATP-based bioluminescence (40, 41), and PCR–dena-
turing gradient gel electrophorises (132). Among their disad-
vantages are limited availability and more complicated
operation compared to more common sampling methods.
Ozone production during their operation and potential
effects on microorganism viability could also be considered
as a concern, but very limited data are presently available
on this issue.
There have also been attempts to incorporate electrostatic
effects into passive bioaerosol sampling techniques, leading to
electrostatic dustfall collectors (133–135). It was concluded
3.2.2-8 ▪ AIR
that this method allows evaluating settled dust from longer
periods of exposure.
Sampler Performance
Bioaerosol sampler performance has been reviewed in several
papers (11, 16, 29, 105, 136–139), and numerous laboratory
and field studies comparing the utilities of various sampling
methods have been performed (Table 2). Data from these
studies are often difficult to compare because of differences
in samplers, the lengths of sampling time, the volumes of
air sampled, the sample analysis methods, and the character-
istics of the bioaerosols being measured. Although no single
sampler type currently available is ideally suited for the collec-
tion and analysis of all types of bioaerosols, consideration of
the theoretical aspects of air sampling and the experimental
results of performance studies may facilitate the selection of
the appropriate sampling method for a particular monitoring
situation.
The performance of bioaerosol samplers can be divided
into physical and biological components. Physical parameters
include inlet sampling efficiency and collection efficiency,
whereas biological sampling efficiency reflects the effect of
sampling on the biological properties of the microorganisms,
that is, culturability or biological integrity, allowing their
identification and enumeration. Inlet sampling efficiency
refers to the ability of the sampler inlet to extract particles
from the ambient environment without bias due to the sizes,
shapes, or densities of the particles. The inlet characteristics
of several bioaerosol samplers have been calculated for dif-
ferent types of bioaerosol particles sampled under various
conditions (105). Depending on the external wind direction
and velocity relative to the inlet geometry and flow character-
istics, increased or, more commonly, decreased particle con-
centration measurements may be obtained relative to the
true concentration in the environment (140, 141).
Collection efficiency is the ability of the sampler to
remove particles from the air stream and transfer them to
the collection medium. In an impactor sampler, the physical
characteristics of the impaction nozzle(s) and the airflow rate
are used to calculate the cutoff size (d50) in the impaction
stage. The d50 is defined as the particle diameter at which
50% of the particles are collected. However, because of the
sharp cutoff characteristics of impactor samplers, the d50 is
generally considered to be the particle diameter above which
all particles are collected, while all those with diameters
below the d50 pass through (29, 142). Theoretical d50s have
been calculated for several bioaerosol samplers (16, 29, 139)
and are shown in Table 2. For efficient collection, it is impor-
tant to choose an impactor that has the d50 below the mean
size of the microorganism being sampled. For membrane fil-
ters, the collection efficiency is approximately 100% for par-
ticles larger than the pore size (143). However, re-emission of
up to 0.63% of bacteria collected on a filter has been reported
with rod-shaped bacteria showing lower reemission rates
compared to spherical bacteria (144). Biological particles,
such as viruses, are usually smaller than the common filter
pore sizes, but can also be collected by filters with efficiency
dependant on filter material (145). However, elution of
microorganisms collected on filters could be a source of sam-
ple loss; thus, elution efficiency should be taken into account
when determining the bioaerosol concentration.
In an inertial device (e.g., an impactor), the “aerody-
namic” diameter of the particle is the diameter of a unit
density sphere that has the same gravitational settling velo-
city as the particle in question (146). A nonspherical micro-
organism is thus described by a single dimension. The d50
depends on several parameters, including the ratio of the
jet-to-plate distance (S) to the impactor’s nozzle size (W).
Most commonly used bioaerosol impactors meet the conven-
tional Marple’s design criteria so that S/W is greater than the
established threshold (1.5 for rectangular nozzles and 1 for the
circular ones) (142). Studies have shown that these samplers
underestimate the concentration of some bioaerosol particles
that they are designed to measure (e.g., fungal spores) (147).
The d50s of several of the most commonly used single-stage
impaction-based spore collectors (e.g., the Air-O-Cell cas-
sette and the Burkard Personal Volumetric Air Sampler)
are about 2.5 µm or greater. At the same time, some fungal
species produce spores of 1.8–2.5 µm in aerodynamic diame-
ter. An increase in the sampling flow rate would help reduce
the d50s of these impactors (3, 33, 142, 147, 148). However,
high sampling flow rates require a rather powerful pump
which limits the use of such a sampler in the field. In addition,
a high impaction velocity may cause particle bounce that
decreases the actual collection efficiency, particularly for
spores. A high-velocity impaction also affects the viability
of stress-sensitive microorganisms. An alternative approach
was developed in which the nondimensional jet-to-plate dis-
tance was proposed to be very small, that is, ≪1 (148). The
feasibility of such impactors with S/W values outside of Mar-
ple’s design criteria for the total spore collection and enumer-
ation has been demonstrated through laboratory and field
evaluations (148). Several impactors with very small
jet-to-plate distances became commercially available in
2004 and 2005, including the Cyclex bioaerosol impact sam-
pler, the Cyclex-d cassette, and the Micro-5 Microcell
(EMS). The increase in collection efficiency as a result of
decreasing the S/W ratio has also been demonstrated for port-
able agar samplers (139).
Biological sampling efficiency differs among bioaerosol
sampler types. Ideally, each sampler should collect all air-
borne microorganisms without altering the culturability or
the biological integrity required for the detection and/or
quantification of the microorganisms. Biological effects of
sampling are difficult to assess because of the heterogeneity
of bioaerosols with respect to particle size, biological compo-
sition, and environmental factors. Airborne microorganisms
are subjected to a variety of environmental stressors (e.g.,
UV radiation, chemical pollutants, desiccation, and temper-
ature extremes). Consequently, many organisms may be in a
nonculturable state (149). In addition, the stress of impac-
tion may injure the collected microorganisms, depending
on their physiological characteristics and impactor design
parameters (84, 150). Although several analytical techniques
do not make a distinction between viable and nonviable cells
(151), others rely on specific properties of the collected
microorganisms, such as culturability or structural integrity.
A recent study showed that conventional methods of bioaer-
osol sampling, such as filtration, impaction, and impingement
cause a certain amount of membrane damage in sensitive
microorganisms (152). Therefore, sampling stresses, which
can result in a reduction in the culturability or other proper-
ties of airborne microorganisms that introduce bias into their
enumeration or identification, are important in assessing the
overall bioaerosol sampler performance. For example, filtra-
tion sampling, though highly efficient for the collection of
airborne microorganisms (3, 153–155), has the disadvantage
of viability losses for vegetative cells, presumably as a result of
desiccation (17, 82, 118). Therefore, filtration sampling in
combination with culture analysis is generally used when
bioaerosol concentrations are very high and sampling times
are short. The use of gelatin filters seems to alleviate that

concern, but only partially, depending on the species col-
lected and filter postsampling processing method (119, 132,
156). Filtration sampling is also used in monitoring
desiccation-resistant forms such as fungal spores and bacterial
endospores or in combination with a total count method of
analysis (30).
The stress imposed by sampling may also result in a loss of
culturability when agar or liquid collection media are used.
However, it has been observed that the recovery of culturable
cells can be improved by using certain culture media (67, 157,
158) or by adding certain compounds to the collection
medium, for example, osmoprotectants, which aid in the
resuscitation of stressed or damaged cells (159). Impaction
into different media also permits the differentiation of meta-
bolic from structural injuries (84). The length of collection
time (discussed in the following section) also has a major
role in the efficacy of air sampling for the retrieval of cultur-
able microorganisms.
Most of the bioaerosol sampler performance evaluations
have not distinguished between physical and biological
parameters in assessing efficiency, focusing instead on an
overall measure of sampling efficacy. Although there is no sin-
gle sampling method that is appropriate for all bioaerosols, the
Andersen six-stage viable impactor and the AGI-30 sampler
have been suggested as reference methods (11), and many
sampler efficiency studies have included these samplers in
side-by-side comparisons (Table 2). The principal advantage
of the agar impactor samplers is the collection of organisms
directly onto the culture medium. However, one of the dis-
advantages of this sampling method is the nonculturable
component of the bioaerosol, which is not measured and
can be a significant percentage of the total composition
(30). Other problems with agar impactor samplers that can
produce erratic results and underestimation of the true bio-
aerosol concentration are (a) particle bounce (e.g., particles
may rebound off the collection surface or another particle
and reenter the air stream [33]), (b) aggregation or clumping
of microorganisms (e.g., multiple organisms impact at the
same place on the agar surface and are enumerated as a single
colony), (c) electrostatic forces (e.g., biological particles are
attracted to the plastic rims or other surfaces of the collection
device [89]), and (d) overloading (the overall performance of
agar impactors can be adversely affected by overloading when
bioaerosol concentrations are high, resulting in agar plates or
strips that contain colonies that overlap and are too numerous
to count [160, 161]). In an impinger, microorganisms are
collected by impaction onto a liquid-wetted surface and are
also subjected to culturability losses due to a sampling stress
(89, 162). The microorganisms collected into a liquid (e.g.,
distilled water or phosphate buffer) may become entrained
in bubbles and reaerosolize from the collection fluid in a
manner similar to the aerosolization of microorganisms from
bubbling liquids such as whirlpools and fermenters (52,
163, 164). The reaerosolization effect has been shown to
decrease the collection efficiencies of AGI-4 and AGI-30
impingers (106, 109). Reaerosolization has also been repor-
ted as a mode of loss in airborne virus sampling (165). In
addition, the collection efficiencies of conventional liquid
impingers may decrease significantly due to liquid evapora-
tion over time (109, 166), thus limiting the use of impingers
to relatively short sampling times (t ≤ 30 min). Once the
bottom of an impinger’s collection vessel becomes dry, the
particles may bounce from it, which further decreases the effi-
ciency of collection (106). This problem may be minimized
by periodically stopping the sampling process and refilling
the impingers or using a liquid-based sampler with automatic
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-9
regulation of liquid level (167, 168). However, even with
these adjustments, one also has to take into account the pos-
sibility of internal losses of collected particles, which in some
cases could be substantial (166, 169).
Effects of particle bounce, reaerosolization, and evapo-
ration of the collection fluid could be minimized in the Bio-
Sampler, when it is operated with nonevaporating and highly
viscous liquids (39). The overall design of the BioSampler’s
collection unit minimizes impaction stress. When used with
heavy white mineral oil, the BioSampler can maintain
microbial viability and high physical collection efficiency
for several hours (107, 108). The prototype of another
liquid sampler—using the bubbling technique (mentioned
earlier)—has been demonstrated to maintain high physical
collection efficiency and viability for a broad spectrum of
microorganisms, including viruses, bacteria (vegetative cells
and spores), and fungi (1, 2, 110).
Collection Time
Sample collection time is an integral part of the bioaerosol
sampling design, and guidelines for the selection of optimal
sampling times for various bioaerosol samplers have been
published previously (22, 33). Parameters that must be con-
sidered are the expected bioaerosol concentration, the quan-
titation range of the sampler, and the effect of sampling stress
on the overall collection efficiency. For each sample, the
sampling period must be sufficiently long to obtain a represen-
tative sample of the airborne microorganisms present without
exceeding the upper quantitation limit of the sampler or
causing losses in the culturability of airborne organisms.
The selection of a sample collection time is complicated by
the fact that bioaerosol concentrations may vary greatly
over time, often by several orders of magnitude within the
same environment. Thus, air sample periods of short duration
provide only a brief temporal and spatial glimpse of the envi-
ronment and several samples may be required to determine
the average bioaerosol concentrations. Furthermore, the sam-
pling environment may range from relatively low bioaerosol
concentrations (≤102 CFU/m3) in clean rooms and hospital
operating rooms (170) to exceptionally high concentrations
(105–1010 CFU/m3) found in certain industrial and agricul-
tural settings or after major flooding events (171–173). It
was observed when sampling with Andersen N6 sampler
that concentrations of culturable bacteria and fungi declined
when sampling duration exceeded 3 min (174). However,
a sampling time of up to 6 min was recommended when
collecting airborne fungi at low concentrations with this sam-
pler (175).
While the ambient bioaerosol concentration is an
unknown quantity, the quantitation range of a bioaerosol
sampler can be determined if the airflow rate and the collec-
tion time are known (9). For a specific air volume collected,
the lower quantitation limit (LQL) is obtained by assuming
the detection of a single organism, whereas the upper quan-
titation limit (UQL) is based on the maximum number of
organisms that can be enumerated from a sample. When
the collection medium is a membrane filter from which
organisms can be eluted or a liquid, the sample can be serially
diluted prior to analysis. For these methods, there is essen-
tially no UQL, unless a sample volume is extremely small
(41). If a filter sample is to be examined under a microscope,
its loading density is an important factor influencing the limit
of quantitation. For an impactor sampler, the UQL is deter-
mined by the area of the collection surface. In the case of
an impactor with several nozzles in its stage(s), the UQL
depends on the number of sampling nozzles. For example,
3.2.2-10 ▪ AIR
the UQL of the Andersen single-stage sampler is reached
when a CFU develops under each of the 400 sampling nozzles.
The UQL is higher for samplers featuring higher number of
nozzles, for example, 1,000 in the Millipore Air Tester.
When aerosol samples are collected to levels near the UQL,
errors in analysis (e.g., uncertainty associated with “positive-
hole” corrections) may occur as a result of overcrowding on
the impaction surface. Other problems include difficulties
in resolving distinct particles or colonies (160) and inhibition
of the growth of microorganisms as a result of competition for
space and nutrients. When a limited amount of overlap
occurs, the representative bioaerosol concentration can be
statistically calculated from the number of colonies on the
agar (160, 176). Conversely, air samples collected at levels
near the LQL of a sampler may contain an insufficient num-
ber of particles to accurately represent the actual bioaerosol
concentration.
One way to predict the optimal sampling time for a partic-
ular sampler is to determine the ideal surface density of micro-
organisms on the collection area and assume the order of
magnitude of the bioaerosol. Nevalainen et al. (29) calcu-
lated the optimal sampling time for five bioaerosol samplers
by using the following formula:
t ¼ (d)(A)=(Ca )(Q) (1)
where t is the sampling time, δ is the desired surface density, A
is the area of the sampling surface, Ca is the average expected
bioaerosol concentration, and Q is the sampler flow rate.
Figure 2 illustrates the sampling times for an ideal surface
density on a culture plate, δmacro, of 1 colony per cm2, or 104
particles per cm2 for microscopic counts, δmicro, with a sam-
pler collection efficiency of 100%. The indicated sampling
times for an airborne concentration of 103 particles per m3
illustrate that the optimal sampling time depends on the sam-
pling method and on the anticipated bioaerosol concentra-
tion. For example, a high-volume sampler (e.g., the SAS
FIGURE 2 Collection times for selected bioaerosol samplers.
BURK, Burkard spore trap ( personal sampler); AND, Andersen six-
stage viable impactor sampler; AND-VI, sixth stage of the Andersen
six-stage impactor used as a separate sampler; MK-II, Casella MK-II
sampler; SAS, surface air system high-flow sampler. (Adapted from
Nevalainen et al. (29) with kind permission from Elsevier Science
Ltd., Kidlington, UK). doi:10.1128/9781555818821.ch3.2.2.f2
sampler) may be best suited for sampling environments
with low bioaerosol concentrations, and the Burkard personal
sampler may be most effective with relatively high concen-
trations. Because the bioaerosol concentration is unknown
and can only be estimated, often more than one collection
time is employed for sample collection to enhance the like-
lihood of obtaining useful data.
Another factor that influences the selection of the sample
collection time is sampling stress, which can result in the loss
of culturability of airborne microorganisms. It has been
observed that increased sampling time has resulted in
decreased viability for aerosolized vegetative bacterial cells
(9, 104, 138, 177, 178). As air flows over the nutrient agar
surface of an impactor, the agar may lose water content, result-
ing in a harder surface. Subsequently, sampled micro-
organisms may rebound from the hardened surface and not
be collected (137). If they are collected, they may be less
embedded in the agar, exposing the microorganisms to desic-
cation stresses as continued air flows through the sampler.
Thus, a doubling of sampling time may not result in a dou-
bling of CFUs, depending on the stress tolerance of the air-
borne microorganisms being sampled. A sampling time as
short as reasonably possible is recommended (178); conse-
quently, the investigator could consider taking several con-
secutive samples of short duration rather than a few samples
for a long time interval.
Concentration Rate
Collection time limitations usually apply to agar-based sam-
plers. Liquid-based samplers sometimes face the opposite
problem: the selected or feasible sampling time is not long
enough to fill the collection liquid with a sufficient number
of microorganisms to exceed an LQL for a particular analyti-
cal method. This especially becomes an issue when samplers
with a low sampling flow rate and relatively high sample
volume are used in low-concentration environments. One
parameter that allows one to estimate the ability of a particu-
lar liquid sampler to collect a desired bioaerosol amount is the
concentration rate: the rate at which particles present in an air
volume are concentrated in a liquid volume per time period:
Concentrationrate, RC (t1 )
Airborne particle concentration (L1 )
¼
Particle concentration in liquid (L1 )
Q(L= min)
) h, (2)
v(L)
where Q is the sampling flow rate, v is the liquid sample vol-
ume, and η collection efficiency. High concentration rates
reduce the sampling time needed to detect airborne particles
and enable detection of lower particle concentrations (179).
Traditional liquid samplers operate at flow rates up to 20
liters/min and feature low sample concentration rates, for
example, up to 2,500 for the BioSampler (SKC) operating
at 12.5 liters/min and sampling into 5 ml of liquid. After
the anthrax attacks of 2001, there has been a focus to develop
samplers with higher concentration rates. The BioGuardian
Air Sampler (InnovaTek) uses flow rates from 100 to 1,000
liters/min and samples into 10–15 ml of liquid. The SpinCon
air sampler (InnovaPrep) samples at 450 liters/min and col-
lects particles into 10 ml of liquid. The BioCapture 650
(FLIR Systems) is a portable sampler that achieves a sampling
flow rate of 200 liters/min and collects particles into 2–5 ml of
liquid. The concentration rates for these samplers are tens of
thousands. The Lawrence Livermore National Laboratory

(Livermore, CA) has developed a stationary Autonomous
Pathogen Detection System (APDS) that combines a virtual
impactor and a wetted-wall cyclone and is capable of con-
tinuous and fully autonomous monitoring for multiple bio-
warfare organisms using a variety of analytical techniques,
including PCR and immune essays (180–183). The APDS
operates at collection flow rates up to 3,750 liters/min with
4–5 ml of collection liquid and can achieve concentration
rates in hundreds of thousands. Another way to achieve
very high concentration rates is to use very small amounts
of collection fluid as was demonstrated in a new electrostatic
collector (40, 41, 128).
If a collected sample does not have sufficient concen-
tration, one possible solution could be to use a hydrosol con-
centration system developed by InnovaPrep (184, 185).
However, this approach has not yet been widely applied.
Number of Samples
Given the heterogeneity and temporal variability exhibited
by bioaerosols and the relatively short collection times for
many air sampling methods, multiple samples are often taken
to more accurately determine the concentration and compo-
sition of a bioaerosol. The number of air samples required in
a specific investigation depends on the statistical methods
employed to analyze the data, the length of the sample collec-
tion time, and the variability of the bioaerosol. Sometimes
pilot sampling studies are performed in a particular environ-
ment to arrive at the final sampling design for a full-fledged
study. Simultaneous paired samples are often taken because
of the high degree of variability that has been demonstrated
between data from paired samplers of the same type (87,
186). Because of the changes in bioaerosol concentrations
over time and the small volume of air from which the sample
is taken, a single air sample at a discrete point in time and
space usually has limited value in the assessment of the air-
borne microorganisms in an environment of concern. The
aerosolization of fungal spores and other microorganisms
may be sporadic, and the air environment may not be per-
fectly mixed. Therefore, a negative result obtained from an
air sample is not definitive and additional testing of surfaces
is used to provide more information (see “Surface Sampling”).
Methods to estimate the required number of airborne samples
have been presented in literature (80).
Other Performance Variables
The selected sampling locations during a bioaerosol investi-
gation could also have an effect on study outcomes. It was
reported that ground-level sampling (breathing height) com-
pared to roof-level sampling could have a significantly higher
concentration of some important fungal aeroallergens but
lower concentrations of others (187). Another study found
that the level of endotoxin concentration measured close to
the carpeted flooring was significantly higher compared to
that measured at 1 m height, most likely due to bioaerosol
resuspension from carpet (188). Sampler orientation, espe-
cially when sampling outdoors, could also have a substantial
effect on the results as was reported for pollen sampling (141).
With the advancement and increased application of
molecular analysis tools in bioaerosol field, the selection of
a bioaerosol sampler is important because the sampling proc-
ess may affect the utility of those tools. As was described ear-
lier, many sampling techniques could be interfaced with
molecular analysis tools. However, a sampler-related bias
regarding species diversity has been reported when using
quantitative PCR (189). Another study indicated that opera-
tional taxonomic units occurring in low abundance (few
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-11
sequences) may differ considerably depending on the sampler
that was used (190).
Sampler Calibration
The concentration of airborne microorganisms is calculated
by dividing the number of collected microorganisms by the
sampled air volume (e.g., CFU or total count per cubic
meter). Therefore, the volumetric flow rate of each sampler
should be calibrated and adjusted to a desired level, if neces-
sary. This is particularly important for impactors and
impingers as the d50 is a function of the flow rate through
the nozzle(s). A decreased flow rate (e.g., due to a weakened
battery) increases the d50, shifting collection bias toward
larger particles (142). If some of the nozzles in a multinozzle
impactor are clogged, the flow velocity through the remaining
nozzles is increased, resulting in a lower d50 with increased
collection efficiency and the potential for increased stress
on the microorganisms. In impactor sampling, the recom-
mended distance between the exit plane of the impactor
nozzle(s) and the top of the nutrient surface should be main-
tained. For example, the manufacturer’s recommended vol-
ume for plastic petri plates is 40 ml of agar in the Andersen
single-stage impactor sampler. Otherwise, the impactor’s col-
lection characteristics ( primarily the d50) may be changed.
The flow rate of a battery-powered sampler is usually cali-
brated before and after sample collection. If the flow rate has
decreased by less than 10%, the average of the two flow rate
measurements is generally used for the calculation of the con-
centration of airborne microorganisms. If the difference
exceeds 10%, the sample is considered invalid. Various cali-
bration methods that measure the flow rate through a soap
bubble meter or an electronic calibration instrument are
available (191–193).
Surface Sampling
Surface sampling is often used in conjunction with air sam-
pling in the indoor environment to provide information
about microbial sources. Direct source evaluation through
surface sampling is used to locate and identify potential
bioaerosol hazards and predict the bioaerosol dispersal and
deposition (43, 194, 195). Qualitative and quantitative
information on the concentration and composition of
surface-associated microorganisms can be obtained with sur-
face sampling. A variety of methods have been used to collect
microorganisms from smooth and porous surfaces, for numer-
ous applications in diverse areas of microbiology. Surface
sampling methods and procedures are discussed elsewhere
in this book.
REFERENCES
1. Agranovski IE, Agranovski V, Grinshpun SA, Reponen
T, Willeke K. 2002. Collection of airborne microorganisms
into liquid by bubbling through porous medium. Aerosol Sci
Tech 36:502–509.
2. Agranovski IE, Agranovski V, Reponen T, Willeke K,
Grinshpun S. 2002. Development and evaluation of a
new personal sampler for culturable airborne microorgan-
isms. Atmos Environ 36:889–898.
3. Aizenberg V, Reponen T, Grinshpun SA, Willeke K.
2000. Performance of Air-O-Cell, Burkard, and button
samplers for total enumeration of airborne spores. Am Ind
Hyg Assoc J 61:855–864.
4. Alvarez AJ, Buttner MP, Toranzos GA, Dvorsky EA,
Toro A, Heikes TB, Mertikas LE, Stetzenbach LD.
1994. The use of solid-phase polymerase chain reaction
3.2.2-12 ▪ AIR
for the enhanced detection of airborne microorganisms.
Appl Environ Microb 60:374–376.
5. Alvarez A, Buttner M, Stetzenbach L. 1995. PCR for bio-
aerosol monitoring: sensitivity and environmental interfer-
ence. Appl Environ Microb 61:3639–3644.
6. An HR, Mainelis G, Yao M. 2004. Evaluation of a high
volume portable bioaerosol sampler in laboratory and field
environments. Indoor Air 14:385–393.
7. Beuchat LR. 1992. Media for detecting and enume-
rating yeasts and moulds. Int J Food Microbiol 17:
145–158.
8. Burge HA, Solomon WR. 1987. Sampling and analysis of
biological aerosols. Atmos Environ 21:451–456.
9. Buttner MP, Stetzenbach LD. 1991. Evaluation of four
aerobiological sampling methods for the retrieval of aero-
solized Pseudomonas syringae. Appl Environ Microbiol 57:
1268–1270.
10. Chapman MD. 1995. Analytical methods: immunoassays.
In Burge HA (ed), Bioaerosols. CRC Press, Boca Raton, FL.
11. Chatigny MA, Macher JM, Burge HA, Solomon WR.
1989. Samplingairborne microorganisms and aeroaller-
gens, p. 199–220. In Hering SV (ed), Air Sampling Instru-
ments for Evaluation of Atmospheric Contaminants, 7th ed.
ACGIH, Cincinnati, OH.
12. Eduard W, Heederic D. 1998. Methods for quantitative
assessment of airborne levels of noninfectious microorgan-
isms in highly contaminated work environments. Am Ind
Hyg Assoc J 59:113–127.
13. Flannigan B. 1997. Air sampling for fungi in indoor Envi-
ronments. J Aerosol Sci 28:381–392.
14. Grinshpun SA, Clark JM. 2005. Guest editorial for the
special issue on measurement and characterization of bio-
aerosols. J Aerosol Sci 36:553–555.
15. Hobbie JE, Daley RJ, Jasper S. 1977. Use of nuclepore
filters for counting bacteria by fluorescence microscopy.
Appl Environ Microb 33:1225–1228.
16. Jensen PA, Lighthart B, Mohr AJ, Shaffer BT. 1994.
Instrumentation used with microbial bioaerosol,
p. 226–284. In Lighthart B, Mohr AJ (eds), Atmospheric
Microbial Aerosols, Theory and Applications. Chapman and
Hall, New York.
17. Jensen PA, Todd WF, Davis GN, Scarpino PV. 1992.
Evaluation of eight bioaerosol samplers challenged with
aerosols of free bacteria. Am Ind Hyg Assoc J 53:
660–667.
18. Lange J, Thorne P, Lynch N. 1997. Application of flow
cytometry and fluorescent in situ hybridization for assess-
ment of exposures to airborne bacteria. Appl Environ Microb
63:1557–1563.
19. Lee KS, Bartlett KH, Brauer M, Stephens GM, Black
WA, Teschke K. 2004. A field comparison of four samplers
for enumerating fungal aerosols. I. Sampling characteristics.
Indoor Air 14:360–366.
20. Lee KS, Teschke K, Brauer M, Bartlett KH. 2004. A field
comparison of four fungal aerosol sampling instruments:
inter-sampler calibration and caveats. Indoor Air 14:
367–372.
21. Leser TD, Boye M, Hendriksen NB. 1995. Survival and
activity of Pseudomonas sp. strain B13(FR1) in a marine
microcosm determined by quantitative PCR and an
rRNA-targeting probe and its effect on the indigenous bac-
terioplankton. Appl Environ Microb 61:1201–1207.
22. Macher JM. 1999. Bioaerosols: Assessment and Control.
American Conference of Governmental Industrial Hygien-
ists, Cincinnati, OH.
23. Madelin TM, Madelin MF. 1995. Biological analysis
of fungi and associated molds, p. 361–386. In Cox CS,
Wathes CM (eds), Bioaerosols Handbook. CRC Press,
Boca Raton, FL.
24. Mainelis G, Willeke K, Adhikari A, Reponen T, Grinsh-
pun SA. 2002. Design and collection efficiency of a new
electrostatic precipitator for bioaerosol collection. Aerosol
Sci Tech 36:1073–1085.
25. Miller JD, Young JC. 1997. The use of ergosterol to meas-
ure exposure to fungal propagules in indoor air. Am Ind Hyg
Assoc J 58:39–43.
26. Milton DK. 1995. Endotoxin, p. 77–86. In Burge HA (ed),
Bioaerosols. CRC Press, Boca Raton, FL.
27. Milton DK, Gere RJ, Feldman HA, Greaves IA. 1990.
Endotoxin measurement: aerosol sampling and application
of a new Limulus method. Am Ind Hyg Assoc J 51:
331–337.
28. Muilenberg M. 2003. Sampling devices. Immunol Allergy
Clin 23:337–355.
29. Nevalainen A, Pastuszka J, Liebhaber F, Willeke K. 1992.
Performance of bioaerosol samplers: collection characteris-
tics and sampler design considerations. Atmos Environ
26A:531–540.
30. Palmgren U, Strom G, Blomquist G, Malmberg P. 1986.
Collection of airborne micro-organisms on Nuclepore fil-
ters, estimation and analysis-CAMNEA method. J Appl
Bacteriol 61:401–406.
31. Pasanen AL, Nikulin M, Tuomainene M, Berg S, Parikka
P, Hintikka EL. 1993. Laboratory experiments on
memberane filter sampling of airborne mycotoxins pro-
duced by Stachybotrys atra Corda. Atmos Environ 27A:
9–13.
32. Placencia AM, Peeler JT, Oxborrow GS, Danielson JW.
1982. Comparison of bacterial recovery by Reuter centrifu-
gal air sampler and slit-to-agar sampler. Appl Environ
Microb 44:512–513.
33. Reponen T, Willeke K, Grinshpun SA, Nevalainen A.
2001. Biological particle sampling, p. 751–777. In Willeke
K, Baron PA (eds), Aerosol Measurement: Principles, Techni-
ques, and Applications, 2nd ed. John Wiley & Sons,
New York.
34. Smid T, Schokkin E, Boleij JSM, Heederik D. 1989. Enu-
meration of viable fungi in occupational environments: a
comparison of samplers and media. Am Ind Hyg Assoc
J 50:235–239.
35. Wagner J, Marcher JM. 2003. Comparison of a passive
aerosol sampler to size selective pump samplers in indoor
environments. Am Ind Hyg Assoc J 64:630–639.
36. Walters M, Milton D, Larsson L, Ford T. 1994. Airborne
environmental endotoxin: a cross validation of sampling
and analysis techniques. Appl Environ Microb 60:
996–1005.
37. White LA, Hadley DJ, Davids DE, Naylor R. 1975.
Improved large-volume sampler for the collection of bacte-
rial cells from aerosol. Appl Microbiol 29:335–339.
38. Willeke K, Baron PA. 1993. Aerosol Measurement: Princi-
ples, Techniques and Applications. Van Nostrand Reinhold,
New York.
39. Willeke K, Lin X, Grinshpun SA. 1998. Improved aerosol
collection by combined impaction and centrifugal motion.
Aerosol Sci Tech 28:439–456.
40. Han T, An HR, Mainelis G. 2010. Performance of an elec-
trostatic precipitator with superhydrophobic surface when
collecting airborne bacteria. Aerosol Sci Tech 44:
339–348.
41. Han T, Nazarenko Y, Lioy PJ, Mainelis G. 2011. Collec-
tion efficiencies of an electrostatic sampler with superhy-
drophobic surface for fungal bioaerosols. Indoor Air 21:
110–120.
42. Su W-C, Tolchinsky AD, Chen BT, Sigaev VI, Cheng
YS. 2012. Evaluation of physical sampling efficiency for
cyclone-based personal bioaerosol samplers in moving air
environments. J Environ Monitor 14:2430–2437.

43. Buttner MP, Cruz-Perez P, Stetzenbach LD, Garrett PJ,
Luedtke AE. 2002. Measurement of airborne fungal spore
dispersal from three types of flooring materials. Aerobiologia
18:1–11.
44. Castellan RM, Olenchock SA, Kinsley KB, Hankinson
JL. 1987. Inhaled endotoxin and decreased spirometric
values: an exposure-response relation for cottondust.
N Engl J Med 317:605–610.
45. Croft WA, Jarvis BB, Yatawara CS. 1986. Airborne out-
break of trichothecene toxicosis. Atmos Environ 20:
549–552.
46. DeBlay F, Chapman MD, Platts-Mills TAE. 1991. Air-
borne cat allergen Feld I: environmental control with the
cat in situ. Am Rev Respir Dis 143:1334.
47. Flannigan B. 1987. Mycotoxins in the air. Int Biodeterior
23:73–78.
48. Higgins JA, Cooper M, Schroeder-Tucker L, Black S,
Miller D, Karns JS, Manthey E, Breeze R, Perdue ML.
2003. A field investigation of Bacillus anthracis contamina-
tion of U.S. Department of Agriculture and other Wash-
ington D.C., buildings during the anthrax attack of
October 2001. Appl Environ Microb 69:593–599.
49. Hirsch SR, Sosman JA. 1976. A one-year survey of mold
growth inside twelve homes. Ann Allergy 36:30–38.
50. Hocking AD, Pitt JI. 1980. Dichloran-glycerol medium for
enumeration of xerophilic fungi from low-moisture foods.
Appl Environ Microb 39:488–492.
51. Horner WE, Wortham AG, Morey PR. 2004. Air and dust
borne mycoflora in houses free of water damage and fungal
growth. Appl Environ Microb 70:6394–6400.
52. Huang YL, Willeke K, Juozaitis A, Donnelly J, Leeson
A, Wyza R. 1994. Fermentation process monitoring
through measurement of aerosol release. Biotechnol Prog
10:32–38.
53. Hunter CA, Grant C, Flannigan B, Bravery AF. 1988.
Mould in buildings: the air spora of domestic buildings.
Int Biodeterior 24:81–101.
54. Jacobs RR. 1989. Airborne endotoxins: an association
with occupational lung disease. Appl Ind Hyg 4:50–56.
55. Jarvis BB, Sorenson WG, Hintikka E, Nijulin M, Zhou
Y, Jiang J, Wang S, Hinkley S, Etzel RA, Dearbom D.
1998. Study of toxin production by isolates of Stachybotrys
chartarum and Memnoniella echinata isolated during a study
of pulmonary hemosiderosis in infants. Appl Environ
Microb 64:3620–3625.
56. Johanning E, Biagini R, Hull D, Morey P, Jarvis B, Land-
sbergis P. 1996. Health and immunology study following
exposure to toxigenic fungi (Stachybotrys chartarum) in a
water-damaged office environment. Int Arch Occup Envi-
ron Health 68:207–218.
57. Luczynska CM, Li Y, Chapman MD, Platts-Mills TAE.
1990. Airborne concentration and particle size distribution
of allergen derived from domestic cats (Felis domesticus):
measurements using cascade impactor, liquid impinger
and a two site monoclonal antibody assay for Fel d I. Am
Rev Respir Dis 141:361–367.
58. Miller JD, Laflamme AM, Sobol Y, Lafontaine P, Green-
halgh R. 1988. Fungi and fungal products in some Cana-
dian houses. Int Biodeterior 24:103–120.
59. Platts-Mills TAE, Thomas WR, Aalbersee RC, Vervloet
D, Chapman MD. 1992. Dust mite allergens and asthma:
report of a second international workshop. J Allergy Clin
Immunol 89:1046.
60. Sawyer MH, Chamberlin CJ, Wu YN, Aintablian N,
Wallace MR. 1994. Detection of varicella-zoster virus
DNA in air samples from hospital rooms. J Infect Dis
169:91–94.
61. Sayer WJ, MacKnight MN, Wilson HW. 1972. Hospital
airborne bacteria as estimated by the Andersen sampler
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-13
versus the gravity settling culture plate. Am J Clin Pathol
58:558–562.
62. Solomon WR. 1975. Assessing fungus prevalence in
domestic interiors. J Allergy Clin Immunol 56:235–242.
63. Sorber CA, Bausum HT, Schaub SA, Small MJ. 1976. A
study of bacterial aerosols at a wastewater irrigation site.
J Water Pollut Control Fed 48:2367–2379.
64. Sorenson WG, Frazer DG, Jarvis BB, Robinson VA.
1987. Trichothecene mycotoxins in aerosolized conidia
of Stachybotrys atra. Appl Environ Microb 53:1370–1375.
65. Stärk KDC, Nicolet J, Frey J. 1998. Detection of Myco-
plasma hyopneumoniae by air sampling with a nested PCR
assay. Appl Environ Microb 64:543–548.
66. Strachan DP, Flannigan B, McCabe EM, McGarry E.
1990. Quantification of airborne moulds in the homes of
children with and without wheeze. Thorax 45:382–387.
67. Verhoeff AP, Winjen JH, Boleij JSM, Brunekreef B, Van
Reenen-Hoekstra BS, Samson RA. 1990. Enumeration
and identification of airborn viable mould propagules in
house. Allergy 45:275–284.
68. Wallace L. 1996. Indoor particles: a review. J Air Waste
Manag Assoc 46:98–126.
69. Cruz P. 2002. Identification of airborne fungi,
p. 1647–1661. In Bitton G (ed), Encyclopedia of Environ-
mental Microbiology. John Wiley & Sons, New York.
70. Desjardin LE, Chen Y, Perkins MD, Teixeira L, Cave
MD, Eisenach KD. 1998. Comparison of the ABI 7700 sys-
tem (TaqMan) and competitive PCR for quantification of
IS6110 DNA in sputum during treatment of tuberculosis.
J Clin Microbiol 36:1964–1968.
71. Douwes J, Doekes G, Montijn R, Heederik D, Brunek-
reef B. 1996. Measurement of beta(1→3)-glucans in
occupational and home environments with an inhibi-
tion enzyme immunoassay. Appl Environ Microb 62:
3176–3182.
72. Ferre F. 1992. Quantitative or semi-quantitative PCR: real-
ity versus myth. PCR Meth Appl 2:1–9.
73. Fogelmark B, Goto H, Yuasa K, Marchat B, Rylander R.
1992. Acute pulmonary toxicity of inhaled beta-1,3-glucan
and endotoxin. Agents Actions 35:50–56.
74. Goto H, Yuasa K, Rylander R. 1994. (1→3)-Beta-D-
glucan in indoor air, its measurement and in vitro activity.
Am J Ind Med 25:81–83.
75. Josephson KL, Gerba CP, Pepper IL. 1993. Polymerase
chain reaction detection of nonviable bacterial pathogens.
Appl Environ Microb 59:3513–3515.
76. Picard C, Ponsonnet C, Paget E, Nesme X, Simonet P.
1992. Detection and enumeration of bacteria in soil by
direct DNA extraction and polymerase chain reaction.
Appl Environ Microb 58:2717–2722.
77. Saiki RK, Scharf S, Faloona F, Mullis KB, Hom GT,
Erlich H, Amheim N. 1985. Enzymatic amplification of
beta-globin genomic sequences and restriction site analysis
for diagnosis of sickle cell anemia. Science 230:
1350–1354.
78. Wilson IG. 1997. Inhibition and facilitation of nucleic
acid amplification. Appl Environ Microb 63:3741–3751.
79. Yike I, Allan T, Sorenson WG, Dearborn DG. 1999.
Highly sensitive protein translation assay for trichothecene
toxicity in airborne particulates: comparison with cytotox-
icity assays. Appl Environ Microb 65:88–94.
80. Burge HA. 1995. Bioaerosol investigations, p. 1–23. In
Burge HA (ed), Bioaerosols. CRC Press, Boca Raton, FL.
81. Atlas RM, Bartha R. 1993. Microbial ecology: fundamentals
and applications, 3rd ed. Benjamin/Cummings, Redwood
City, CA.
82. Thorne PS, Kiekhaefer MS, Whitten P, Donham KJ.
1992. Comparison of bioaerosol sampling methods in barns
housing swine. Appl Environ Microb 58:2543–2551.
3.2.2-14 ▪ AIR
83. Lee KW, Mukund R. 2001. Filter collection, p. 197–228.
In Baron PA, Willeke K (eds), Aerosol Measurement: Prin-
ciples, Techniques and Applications. John Wiley & Sons,
New York.
84. Stewart SL, Grinshpun SA, Willeke K, Terzieva S,
Ulevicius V, Donnelly J. 1995. Effect of impact stress on
microbial recovery on an agar surface. Appl Environ
Microb 61:1232–1239.
85. Wang Z, Reponen T, Grinshpun SA, Gorny RL, Willeke
K. 2001. Effect of sampling time and air humidity on the
bioefficiency of filter samplers for bioaerosol collection.
J Aerosol Sci 32:661–674.
86. Mainelis G, Grinshpun SA, Willeke K, Reponen T,
Ulevicius V, Hintz PJ. 1999. Collection of airborne micro-
organisms by electrostatic precipitation. Aerosol Sci Tech
30:127–144.
87. Buttner MP, Stetzenbach LD. 1993. Monitoring airborne
fungal spores in an experimental indoor environment to
evaluate sampling methods and the effects of human activ-
ity on air sampling. Appl Environ Microb 59:219–226.
88. Sayer WJ, Shean DB, Ghosseiri J. 1969. Estimation of air-
borne fungal flora by the Andersen sampler versus the grav-
ity settling plate. J Allergy 44:214–227.
89. Andersen AA. 1958. New sampler for the collection, siz-
ing, and enumeration of viable airborne particles.
J Bacteriol 76:471–484.
90. Jones W, Marring K, Morey P, Sorenson W. 1985. Evalu-
ation of the Andersen viable impactor for single stage sam-
pling. Am Ind Hyg Assoc J 46:294–298.
91. Yao MS, Mainelis G. 2007. Use of portable microbial sam-
plers for estimating inhalation exposure to viable biological
agents. J Expo Sci Env Epid 17:31–38.
92. Gillespie VL, Clark CS, Bjomson HS, Samuels SJ,
Holland JW. 1981. A comparison of two stage and six-stage
Andersen impactors for viable aerosols. Am Ind Hyg
Assoc J 42:858–864.
93. Mehta SK, Mishra SK, Pierson DL. 1996. Evaluation of
three portable samplers for monitoring airborne fungi.
Appl Environ Microb 62:1835–1838.
94. Li CM, Pan LK, Luong JHT. 2006. Capacitance immuno-
sensors based on an array biotape. Analyst 131:788–790.
95. Zhen SQ, Li KJ, Yin LH, Yao MS, Zhang HL, Chen LS,
Zhou MH, Chen XD. 2009. A comparison of the efficien-
cies of a portable BioStage impactor and a Reuter centrifu-
gal sampler (RCS) High Flow for measuring airborne
bacteria and fungi concentrations. J Aerosol Sci 40:
503–513.
96. Chang CW, Hung PY. 2012. Evaluation of sampling tech-
niques for detection and quantification of airborne legion-
ellae at biological aeration basins and shower rooms.
J Aerosol Sci 48:63–74.
97. Chang CW, Chou FC. 2011. Assessment of bioaerosol
sampling techniques for viable Legionella pneumophila by
ethidium monoazide quantitative PCR. Aerosol Sci
Tech 45:343–351.
98. Levetin E, Shaughnessy R, Fisher E, Ligman B, Harrison
J, Brennan T. 1995. Indoor air quality in schools: exposure
to fungal allergens. Aerobiologia 11:27–34.
99. Adhikari A, Sen MM, Gupta-Bhattacharya S, Chanda S.
2000. Incidence of allergenically significant fungal aerosols
in a rural bakery of West Bengal, India. Mycopathologia 149:
35–45.
100. DeKoster JA, Thorne PS. 1995. Bioaerosol concentra-
tions in noncomplaint, complaint, and intervention homes
in the Midwest. Am Ind Hyg Assoc J 56:573–580.
101. Mitakakis TZ, Tovey ER, Xuan W, Marks GB. 2000. Per-
sonal exposure to allergenic pollen and mould spores in
inland New South Wales, Australia. Clin Exp Allergy
30:1733–1739.
102. Emberlin J, Newman T, Bryant R. 1995. The incidence of
fungal spores in the ambient air and inside homes: evidence
from London. Aerobiologia 11:253–258.
103. Rosebury T. 1947. Experimental Airborne Infection.
Williams and Wilkins, Baltimore, MD.
104. Cox CS. 1987. The aerobiological pathway of microorganisms.
John Wiley & Sons, Chichester, UK.
105. Grinshpun SA, Chang C-W, Nevalainen A, Willeke K.
1994. Inlet characteristics of bioaerosol samplers.
J Aerosol Sci 25:1503–1522.
106. Grinshpun SA, Willeke K, Ulevicius V, Juozaitis A,
Terzieva S, Donnelly J, Stelma GN, Brenner KP. 1997.
Effect of impaction, bounce and reaerosolization on the
collection efficiency of impingers. Aerosol Sci Tech 26:
326–342.
107. Lin X, Reponen T, Willeke K, Wang Z, Grinshpun SA,
Trunov M. 2000. Survival of airborne microorganisms dur-
ing swirling aerosol collection. Aerosol Sc Technol 32:
184–196.
108. Lin X, Reponen T, Willeke K, Grinshpun SA, Foarde
KK, Ensor DS. 1999. Long-term sampling of airborne bac-
teria and fungi into a non-evaporating liquid. Atmos Envi-
ron. 33:4291–4298.
109. Lin X, Willeke K, Ulevicius V, Grinshpun SA. 1997.
Effect of sampling time on the collection efficiency of all-
glass impingers. Am Ind Hyg Assoc J 58:480–488.
110. Agranovski IE, Safatov AS, Borodulin AI, Pyankov OV,
Petrishchenko VA, Sergeev AN, Sergeev AA, Agranov-
ski V, Grinshpun SA. 2005. New personal sampler for via-
ble airborne viruses: feasibility study. J Aerosol Sci 36:
609–617.
111. Pyankov OV, Agranovski IE, Pyankova O, Mokhonova
E, Mokhonov V, Safatov AS, Khromykh AA. 2007.
Using a bioaerosol personal sampler in combination with
real-time PCR analysis for rapid detection of airborne
viruses. Environ Microbiol 9:992–1000.
112. Hietala SK, Hullinger PJ, Crossley BM, Kinde H, Ardans
AA. 2005. Environmental air sampling to detect exotic
Newcastle disease virus in two california commercial poul-
try flocks. J Vet Diagn Invest 17:198–200.
113. Macher J, Chen B, Rao C. 2008. Chamber evaluation of a
personal, bioaerosol cyclone sampler. J Occ Environ Hyg
5:702–712.
114. Su WC, Tolchinsky AD, Sigaev VI, Cheng YS. 2012. A
wind tunnel test of newly developed personal bioaerosol
samplers. J Air Waste Manage 62:828–837.
115. Lindsley WG, Schmechel D, Chen BT. 2006. A two-stage
cyclone using microcentrifuge tubes for personal bioaerosol
sampling. J Environ Monitor 8:1136–1142.
116. Macher J, Chen B, Rao C. 2008. Field evaluation of a per-
sonal, bioaerosol cyclone sampler. J Occ Environ Hyg 5:
724–734.
117. Cao G, Noti JD, Blachere FM, Lindsley WG, Beezhold
DH. 2011. Development of an improved methodology to
detect infectious airborne influenza virus using the NIOSH
bioaerosol sampler. J Environ Monitor 13:3321–3328.
118. Lundholm IM. 1982. Comparison of methods for quantita-
tive determinations of airborne bacteria and evaluation of
total viable counts. Appl Environ Microb 44:179–183.
119. Yao MS, Mainelis G. 2007. Analysis of portable impactor
performance for enumeration of viable bioaerosols. J Occ
Environ Hyg 4:514–524.
120. Wu Y, Shen FX, Yao MS. 2010. Use of gelatin filter and
BioSampler in detecting airborne H5N1 nucleotides, bac-
teria and allergens. J Aerosol Sci 41:869–879.
121. Haatainen S, Laitinen J, Linnainmaa M, Reponen T, Kal-
liokoski P. 2009. The suitability of the IOM foam sampler
for bioaerosol sampling in occupational environments.
J Occ Environ Hyg 7:1–6.

122. Kenny LC, Bowry A, Crook B, Stancliffe JD. 1999. Field
testing of a personal size-selective bioaerosol sampler. Ann
Occup Hyg 43:393–404.
123. Berry CM. 1941. An electrostatic method for collecting
bacteria from air. PhD dissertation, State University of Iowa.
124. Gerone PJ, Couch RB, Keefer GV, Douglas RG, Derren-
bacher EB, Knight V. 1966. Assessment of experimental
and natural viral aerosols. Bacteriol Rev 30:576–588.
125. Lee SA, Willeke K, Mainelis G, Adhikari A, Wang HX,
Reponen T, Grinshpun SA. 2004. Assessment of electrical
charge on airborne microorganisms by a new bioaerosol
sampling method. J Occ Environ Hyg 1:127–138.
126. Mainelis G, Adhikari A, Willeke K, Lee S-A, Reponen T,
Grinshpun SA. 2002. Collection of airborne microorgan-
isms by a new electrostatic precipitator. J Aerosol Sci 33:
1417–1432.
127. Yao MS, Mainelis G. 2006. Utilization of natural electrical
charges on airborne microorganisms for their collection by
electrostatic means. J Aerosol Sci 37:513–527.
128. Han T, Mainelis G. 2008. Design and development of an
electrostatic sampler for bioaerosols with high concentra-
tion rate. J Aerosol Sci 39:1066–1078.
129. Tan M, Shen F, Yao M, Zhu T. 2011. Development of an
automated electrostatic sampler (AES) for bioaerosol
detection. Aerosol Sci Tech 45:1154–1160.
130. Madsen AM, Sharma AK. 2008. Sampling of high
amounts of bioaerosols using a high-volume electrostatic
field sampler. Ann Occup Hyg 52:167–176.
131. Sharma KA, Wallin H, Jensen AK. 2007. High volume
electrostatic field-sampler for collection of fine particle
bulk samples. Atmos Environ 41:369–381.
132. Xu Z, Yao M. 2011. Analysis of culturable bacterial and
fungal aerosol diversity obtained using different samplers
and culturing methods. Aerosol Sci Tech 45:1143–1153.
133. Liebers V, van Kampen V, Buenger J, Dueser M, Stubel
H, Bruening T, Raulf-Heimsoth M. 2012. Assessment of
airborne exposure to endotoxin and pyrogenic active dust
using electrostatic dustfall collectors (EDCs). J Toxicol
Env Heal A 75:501–507.
134. Noss I, Wouters IM, Visser M, Heederik DJJ, Thorne
PS, Brunekreef B, Doekes G. 2008. Evaluation of a low-
cost electrostatic dust fall collector for indoor air endotoxin
exposure assessment. Appl Environ Microb 74:
5621–5627.
135. Frankel M, Timm M, Hansen EW, Madsen AM. 2012.
Comparison of sampling methods for the assessment of
indoor microbial exposure. Indoor Air 22:405–414.
136. Grinshpun SA, Reponen T. 2004. Sampling of biological
particles from ambient environment: physical principles,
efficiency, and exposure assessment. Int Aerobiol Newsl
59:1–2.
137. Juozaitis A, Willeke K, Grinshpun SA, Donnelly J. 1994.
Impaction onto a glass slide or agar versus impingement
into a liquid for the collection and recovery of airborne
microorganisms. Appl Environ Microb 60:861–870.
138. Thompson MW, Donnelly J, Grinshpun SA, Juozaitis A,
Willeke K. 1994. Method and test system for evaluation of
bioaerosol samplers. J Aerosol Sci 25:1579–1593.
139. Yao MS, Mainelis G. 2006. Investigation of cut-off sizes
and collection efficiencies of portable microbial samplers.
Aerosol Sci Tech 40:595–606.
140. Brockmann JE. 2001. Sampling and transport of aerosols,
p. 143–195. In Baron P, Willeke K (eds), Aerosol Measure-
ment: Principles, Techniques and Applications. John Wiley &
Sons, New York.
141. Michel D, Rotach MW, Gehrig R, Vogt R. 2012. On the
efficiency and correction of vertically oriented blunt bio-
aerosol samplers in moving air. Int J Biometeorol. 56:
1113–1121.
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-15
142. Marple VA, Rubow KL, Olson BA. 2001. Inertial, gravita-
tional, centrifugal and thermal collection techniques,
p. 229–260. In Baron PA, Willeke K (eds), Aerosol Meas-
urement: Principles, Techniques and Applications. John Wiley
& Sons, New York.
143. Lippmann M. 1989. Sampling aerosols by filtration,
p. 305–336. In Hering SV (ed), Air Sampling Instruments
for Evaluation of Atmospheric Contaminants, 7th ed. Ameri-
can Conference of Governmental Industrial Hygienists,
Cincinnati, OH.
144. Miaskiewicz-Peska E, Lebkowska M. 2012. Comparison
of aerosol and bioaerosol collection on air filters. Aerobiolo-
gia 28:185–193.
145. Burton NC, Grinshpun SA, Reponen T. 2007. Physical
collection efficiency of filter materials for bacteria and
viruses. Ann Occup Hyg 51:143–151.
146. Baron PA, Willeke K. 2001. Gas and particle motion,
p. 61–82. In Baron PA, Willeke K (eds), Aerosol Measure-
ment: Principles, Techniques and Applications. John Wiley
& Sons, New York.
147. Trunov M, Trakumas S, Willeke K, Grinshpun SA,
Reponen T. 2001. Collection of bioaerosols particles by
impaction: effect of fungal spore agglomeration and
bounce. Aerosol Sci Tech 34:490–498.
148. Grinshpun SA, Mainelis G, Trunov M, Górny RL,
Sivasubramani SK, Adhikari A, Reponen T. 2005.
Collection of airborne spores by circular single-stage impac-
tors with small jet-to-plate distance. J Aerosol Sci 36:
575–591.
149. Cox CS. 1989. Airborne bacteria and viruses. Sci Prog 73:
469–500.
150. Yao MH, Mainelis G. 2006. Effect of physical and biolog-
ical parameters on enumeration of bioaerosols by portable
microbial impactors. J Aerosol Sci 37:1467–1483.
151. Martinez K, Rao C, Burton N. 2004. Exposure assessment
and analysis for biological agents. Grana 43:193–208.
152. Zhen H, Han T, Fennell D, Mainelis G. 2013. Release of
free DNA molecules by the membrane-impaired bacteria
during aerosolization and sampling. Appl Environ Microb
77:7780–7789.
153. Adhikari A, Martuzevicius D, Reponen T, Grinshpun
SA, Cho S-H, Sivasubramani SK, Zhong W, Levin L,
Kelley AL, St. Clair HG, LeMasters G. 2003. Perform-
ance of the Button personal inhalable sampler for the meas-
urement of outdoor aeroallergens. Atmos Environ 37:
4723–4733.
154. Adhikari A, Reponen T, Lee SA, Grinshpun SA. 2004.
Assessment of human exposure to airborne fungi in
agricultural confinements: personal inhalable sampling
versus stationary sampling. Ann Agr Env Med 11:
269–277.
155. Burton NC, Adhikari A, Grinshpun SA, Hornung R,
Reponen T. 2005. The effect of filter material on bioaero-
sol collection of Bacillus subtilis spores used as a Bacillus
anthracis simulant. J Environ Monitor 7:475–480.
156. Dabisch P, Yeager J, Kline J, Klinedinst K, Welsch A, Pitt
ML. 2012. Comparison of the efficiency of sampling devi-
ces for aerosolized Burkholderia pseudomallei. Inhal Toxicol
24:247–254.
157. Burge HP, Solomon WR, Boise JR. 1977. Comparative
merits of eight popular media in aerometric studies of fungi.
Allergy Clin Immunol 60:199–203.
158. Morring KL, Sorenson WG, Attfield MD. 1983. Sam-
pling for airborne fungi: a statistical comparison of media.
Am Ind Hyg Assoc J 44:662–664.
159. Marthi B. 1994. Resuscitation of microbial bioaerosols,
p. 192–208. In Lighthart B, Mohr, AJ (eds), Atmospheric
Microbial Aerosols, Theory and Applications. Chapman &
Hall, New York.
3.2.2-16 ▪ AIR
160. Chang C-W, Hwang Y-H, Grinshpun SA, Macher JM,
Willeke K. 1994. Evaluation of counting error due to col-
ony masking in bioaerosol sampling. Appl Environ Microb
60:3732–3738.
161. Chang C-W, Grinshpun SA, Willeke K, Macher JM,
Donnelly J, Clark S, Juozaitis A. 1995. Factors affecting
microbiological colony count accuracy for bioaerosol sam-
pling and analysis. Am Ind Hyg Assoc J 56:979–986.
162. Tyler ME, Shipe EL. 1959. Bacterial aerosol samplers.
I. Development and evaluation of the all-glass impinger.
Appl Microbiol 7:337–349.
163. Baron PA, Willeke K. 1986. Respirable droplets from whir-
pools: measurements of size distribution and estimation of
disease potential. Environ Research 39:8–18.
164. Juozaitis A, Huang YL, Willeke K, Donnelly J, Kalatoor
S, Leeson A, Wyza R. 1994. Dispersion of respirable aero-
sols in a fermenter and their removal in an exhaust system.
Appl Occup Environ Hyg 9:552–559.
165. Riemenschneider L, Woo MH, Wu CY, Lundgren D,
Wander J, Lee JH, Li HW, Heimbuch B. 2010. Character-
ization of reaerosolization from impingers in an effort to
improve airborne virus sampling. J Appl Microbiol 108:
315–324.
166. Han TW, Mainelis G. 2012. Investigation of inherent and
latent internal losses in liquid-based bioaerosol samplers.
J Aerosol Sci 45:58–68.
167. Chang CW, Hung PY. 2012. Methods for detection and
quantification of airborne Legionellae around cooling tow-
ers. Aerosol Sci Tech 46:369–379.
168. Blatny JM, Reif BAP, Skogan G, Andreassen O, Hoiby
EA, Ask E, Waagen V, Aanonsen D, Aaberge IS, Cau-
gant DA. 2008. Tracking airborne Legionella and Legionella
pneumophila at a biological treatment plant. Environ Sci
Technol 42:7360–7367.
169. Kesavan J. 2012. Sampling and retention efficiencies of
batch-type liquid-based bioaerosol samplers. Aerosol Sci
Tech 44:817–829.
170. Favero MS, Puleo JR, Marshall JH, Oxborrow GS. 1966.
Comparative levels and types of microbial contamination
detected in industrial clean rooms. Appl Microbiol 14:
539–551.
171. Eduard W, Lacey J, Karlsson K, Palmgren U, Strom G,
Blomquist G. 1990. Evaluation of methods for enumerat-
ing microorganisms in filter samples from highly contami-
nated occupational environments. Am Ind Hyg Assoc
J 51:427–436.
172. Kotimaa M. 1990. Occupational exposure to spores in the
handling of wood chips. Grana 29:153–156.
173. Chew GL, Wilson J, Rabito FA, Grimsley F, Iqbal S,
Reponen T, Muilenberg ML, Thorne PS, Dearborn
DG, Morley RL. 2006. Mold and endotoxin levels in the
aftermath of hurricane Katrina: a pilot project of homes
in New Orleans undergoing renovation. Environ Health
Perspect 114:1883–1889.
174. Godish DR, Godish TJ. 2007. Relationship between sam-
pling duration and concentration of culturable airborne
mould and bacteria on selected culture media. J Appl
Microbiol 102:1479–1484.
175. Saldanha R, Manno M, Saleh M, Ewaze JO, Scott JA.
2008. The influence of sampling duration on recovery of
culturable fungi using the Andersen N6 and RCS bioaero-
sol samplers. Indoor Air 18:464–472.
176. Macher JM. 1989. Positive-hole correction of multiplejet
impactors for collecting viable microorganisms. Am Ind
Hyg Assoc J 50:561–568.
177. Hatch MT, Wolochow H. 1969. Bacterial survival: conse-
quences of the airborne state. In Dimmick RL, Akers AB
(eds), An Introduction to Experimental Aerobiology. John
Wiley & Sons, New York.
178. Mainelis G, Tabayoyong M. 2010. The effect of sampling
time on the overall performance of portable microbial
impactors. Aerosol Sci Tech 44:75–82.
179. Haglund JS. 2003. Two linear slot nozzle virtual impactors for
concentration of bioaerosols, College Station, TX.
180. Hindson BJ, Brown SB, Marshall GD, McBride MT,
Makarewicz AJ, Gutierrez DM, Wolcott DK, Metz TR,
Madabhushi RS, Dzenitis JM, Colston BW Jr. 2004.
Development of an automated sample preparation module
for environmental monitoring of biowarfare agents. Anal
Chem 76:3492–3497.
181. Hindson BJ, Makarewicz AJ, Setlur US, Henderer BD,
McBride MT, Dzenitis JM. 2005. APDS: the autonomous
pathogen detection system. Biosens Bioelectron 20:
1925–1931.
182. Hindson BJ, McBride MT, Makarewicz AJ, Henderer
BD, Setlur US, Smith SM, Gutierrez DM, Metz TR,
Nasarabadi SL, Venkateswaran KS, Farrow SW, Colston
BW, Dzenitis JM. 2005. Autonomous detection of aerosol-
ized biological agents by multiplexed immunoassay with
polymerase chain reaction confirmation. Anal Chem
77:284–289.
183. Langlois RG, Brown S, Colston B, Jones L, Masquelier
D, Meyer P, McBride M, Nasarabi S, Ramponi AJ, Ven-
kateswaran K, Milanowich F. 2000. Development of an
autonomous pathogen detection system, p. 227–234. In
Abstracts of the First Joint Conference on Point Detection,
Williamsburg, VA.
184. Alburty D, Murowchick P, Packingham Z, Page A, Davis
T, Wheeldon D II. 2010. Integration of a hydrosol concen-
trator with an aerosol to hydrosol sampler for fluid reuse and
improved concentration factors. Abstract 446, AAAR
29th Annual Conference, Portland, OR.
185. Hunter DM, Leskinen SD, Magaña S, Schlemmer SM,
Lim DV. 2011. Dead-end ultrafiltration concentration
and IMS/ATP-bioluminescence detection of Escherichia
coli O157:H7 in recreational water and produce wash.
J Microbiol Meth 87:338–342.
186. Lembke LL, Kniseley RN, van Nostrand RC, Hale MD.
1981. Precision of the all-glass impinger and the
andersen microbial impactor for air sampling in solid-
waste handling facilities. Appl Environ Microb 42:
222–225.
187. Khattab A, Levetin E. 2008. Effect of sampling height on
the concentration of airborne fungal spores. Ann Allerg
Asthma Im 101:529–534.
188. Wang Z, Shalat SL, Black K, Lioy PJ, Stambler AA,
Emoekpere OH, Hernandez M, Han T, Ramagopal M,
Mainelis G. 2012. Use of a robotic sampling platform to
assess young children’s exposure to indoor bioaerosols.
Indoor Air 22:159–169.
189. Li KJ. 2011. Molecular comparison of the sampling effi-
ciency of four types of airborne bacterial samplers. Sci Total
Environ 409:5493–5498.
190. Fahlgren C, Bratbak G, Sandaa R-A, Thyrhaug R, Zwei-
fel UL. 2011. Diversity of airborne bacteria in samples col-
lected using different devices for aerosol collection.
Aerobiologia 27:107–120.
191. Chen BT. 1993. Instrument calibration, p. 493–520. In
Willeke K, Baron PA (eds), Aerosol Measurement: Princi-
ples, Techniques and Applications. Van Nostrand Reinhold,
New York.
192. Lippmann M. 1973. The Industrial Environment—Its Evalu-
ation and Control, p. 101–122. National Institute for Occu-
pational Safety and Health, Government Printing Office,
Washington, DC.
193. Lippmann M. 1989. Calibration of air sampling instru-
ments, p. 73–110. In Hering SV (ed), Air Sampling Instru-
ments for Evaluation of Atmospheric Contaminants, 7th ed.

American Conference of Governmental Industrial Hygien-
ists, Cincinnati, OH.
194. Martyny JW, Martinez KF, Morey PR. 1999. Source sam-
pling. In Macher J (ed), Bioaerosols: Assessment and Control.
American Conference of Governmental Industrial Hygien-
ists, Cincinnati, OH.
195. Weis CP, Intrepido AJ, Miller AK, Cowin PG, Durno
MA, Gebhardt JS, Bull R. 2002. Secondary aerosolization
of viable Bacillus anthracis spores in a contaminated US
Senate office. J Am Med Assoc 288:2853–2858.
196. Bausum HT, Schaub SA, Kenyon KF, Small MJ. 1982.
Comparison of coliphage and bacterial aerosols at a waste-
water spray irrigation site. Appl Environ Microb 43:
28–38.
197. Blomquist G, Strom G, Stromquist LH. 1984. Sampling of
high concentrations of airborne fungi. Scand J Work Envi-
ron Health 10:109–113.
198. Burge HA, Boise JR, Rutherford JA, Solomon WR.
1977. Comparative recoveries of airborne fungus spores
by viable and non-viable modes of volumetric collection.
Mycopathologia 61:27–33.
199. Zimmerman NJ, Reist PC, Turner AG. 1987. Com-
parison of two biological aerosol sampling methods. Appl
Environ Microb 53:99–104.
200. Gallup D, Purves J, Burge H. 2004. A disposable sampler
for collecting volumetric air samples onto agar media.
J Allergy Clin Immun 113:S138.
201. Kaye A. 1988. Efficiency of “Biotest RCS” as a sampler of
airborne bacteria. PDA J Pharm Sci Technol 42:147–152.
202. Macher JM, First MW. 1983. Reuter centrifugal air sam-
pler: measurement of effective airflow rate and collection
efficiency. Appl Environ Microbial. 45:1960–1962.
203. Sanchez-Munoz M, Munoz-Vicente M, Cobas G, Portela
R, Amils R, Sanchez B. 2012. Comparison of three high-
flow single-stage impaction-based air samplers for bacteria
quantification: DUO SAS SUPER 360, SAMPL’AIR
and SPIN AIR. Anal Meth 4:399–405.
3.2.2. Sampling for Airborne Microorganisms ▪ 3.2.2-17
204. Jensen P, Schafer M. 1998. Sampling and characterization
of bioaerosols, p. 82–112. In NIOSH Manual of Analytical
Methods. National Institute for Occupational Safety and
Health, Atlanta, GA.
205. Levetin E. 2004. Methods for aeroallergen sampling. Curr
Allergy Asthm R 4:376–383.
206. Solomon WR, Burge HA, Boise JR, Becker M. 1980.
Comparative particle recoveries by the retracting rotorod,
rotoslide and Burkard spore trap sampling in a compact
array. Int J Biometeorol 24:107–116.
207. Crook B, Higgins S, Lacey J. 1987. Methods for sampling
microorganisms at solid waste disposal sites, p. 791–797. In
Houghton DR, Smith RN, Eggins HOW (eds), Biodeterio-
ration 7. Elsevier, London.
208. Cage BR, Schreiber K, Barnes C, Portnoy J. 1996. Eval-
uation of four bioaerosol samplers in the outdoor environ-
ment. Ann Allerg Asthma Im 77:401–406.
209. Zhao Y, Aarnink AJA, Doornenbal P, Huynh TTT,
Koerkamp PWGG, Landman WJM, de Jong MCM.
2011. Investigation of the efficiencies of bioaerosol sam-
plers for collecting aerosolized bacteria using a fluorescent
tracer. II: Sampling efficiency and half-life time. Aerosol
Sci Tech 45:432–442.
210. Carvalho E, Sindt C, Verdier A, Galan C, O’Donoghue
L, Parks S, Thibaudon M. 2008. Performance of the Cori-
olis air sampler, a high-volume aerosol-collection system for
quantification of airborne spores and pollen grains. Aerobio-
logia 24:191–201.
211. Zhao Y, Aarnink AJA, Doornenbal P, Huynh TTT,
Koerkamp PWGG, de Jong MCM, Landman WJM.
2011. Investigation of the efficiencies of bioaerosol sam-
plers for collecting aerosolized bacteria using a fluorescent
tracer. I: Effects of non-sampling processes on bacterial cul-
turability. Aerosol Sci Tech 45:423–431.
212. Parks S, Bennet AM, Speight SE, Benbough JE. 1996. An
assessment of the Sartorius MD8 microbiological air sam-
pler. Appl Microb 80:529–534.
 Analysis of Bioaerosol Samples
PATRICIA CRUZ AND MARK P. BUTTNER
3.2.3
Respiratory exposure to certain pathogenic or toxigenic
microorganisms and/or elevated concentrations of environ-
mental organisms could result in health effects, such as aller-
gic reactions, irritant responses, toxicosis, and respiratory
illness. Determination of the concentration and composition
of bioaerosols in indoor environments is necessary for assess-
ment of contamination levels and to estimate potential expo-
sure of occupants. The need for accurate measurement of
bioaerosols has received increased attention in recent years
owing to concerns with mold contamination in indoor envi-
ronments and the threat of bioterrorism. Sample analysis
methods include culture, microscopic, biochemical, immu-
nological, or molecular biological assays. Traditionally, air-
borne microorganisms have been analyzed by culturable
and microscopic total count determinations. However, limi-
tations of these methods have led to the development of tech-
niques that can increase the sensitivity and accuracy of
bioaerosol monitoring. The selection of an analysis method
is a critical component of a bioaerosol sampling plan, and it
should be designated before air sampling is conducted. Factors
that influence the choice of an analytical method include the
cost and length of time required for analysis, the sensitivity
and specificity of the analysis method, the sampling methods
to be utilized, and the expected characteristics of the bioaer-
osol of interest. The purpose of this chapter is to present an
overview of available methods for the analysis of bioaerosols.
In addition, the potential use of enhanced monitoring of bio-
aerosols with PCR, biochemical, and immunological assays is
discussed.
CULTURE METHODS
Culture Media and Incubation
Many of the currently available bioaerosol sampling methods
rely on culture for the quantification and characterization of
airborne bacteria and fungi. Microorganisms that are col-
lected on a nutrient agar surface by impaction can be cultured
directly, whereas organisms that are collected in a liquid or on
a filter are transferred to a culture medium. Only those cells
that survive and reproduce under the culture conditions to
form visible colonies are enumerated. However, nonviable
microorganisms and/or their by-products that may cause
health effects (1–3) are not detected. Because microorgan-
isms exhibit a wide range of nutritional requirements, no
doi:10.1128/9781555818821.ch3.2.3
3.2.3-1
microbiological media formulation is capable of culturing
every type of organism. Therefore, a common strategy in bio-
aerosol monitoring is to choose general media which promote
the growth of the greatest diversity of species. This approach is
especially useful in survey efforts directed toward the charac-
terization of the airborne microbiota. Because of differences
in temperature, incubation time, and nutritional require-
ments among microorganisms, no single culture medium is
satisfactory for the simultaneous isolation of airborne bacteria
and fungi. Generally, it is necessary to perform replicate sam-
pling using different culture media or to divide samples for
inoculation onto multiple types of nutrient media to detect
fungi and bacteria.
Several conditions, such as pH, temperature, water activ-
ity, nutrients, antibiotics, light, and aeration, can be manip-
ulated to favor the growth of a selected group of organisms
(4). For example, incubation at a temperature of 55°C may
be used to select for the growth of thermophilic actinomy-
cetes, or antibiotic resistance markers may be used for the
selective culture of a target species (5–7). However, the
growth of microorganisms on selective media may be hin-
dered compared with that on general media. This is especially
of concern when culturing airborne microorganisms, which
are already stressed or damaged by aerosolization and sam-
pling. For this reason, general media are often used for the ini-
tial culture of bioaerosol samples, followed by replication onto
differential or selective media for identification (8).
Several broad-spectrum media have been evaluated for
their utility in the retrieval of culturable airborne fungi (9–
12). The media discussed below have been recommended,
although the results vary between investigations. Malt extract
agar (MEA), rose bengal–containing agars (e.g., rose bengal–
streptomycin and dichloran–rose bengal–chloramphenicol),
and dichloran glycerol-18 agar have been suggested for the
isolation of airborne fungi. MEA may be the most widely
used fungal isolation medium. Of the formulations available,
unamended 2% MEA was found to promote sporulation bet-
ter than MEA amended with glucose and peptone (13, 14).
Rose bengal– and dichloran-containing media have the
advantage of inhibiting the spread of rapidly growing fungal
genera (e.g., Rhizopus and Mucor), allowing the enumeration
and identification of other slower-growing fungi in the sam-
ple. However, rose bengal is photoactivated in direct sunlight,
forming cytotoxic products that may inhibit fungal growth
(15). Dichloran glycerol-18 is a low water activity medium
3.2.3-2 ▪ AIR
(aw = 0.955) developed for the isolation of xerophilic fungi
(16), and it compares favorably with other media for culturing
mesophilic airborne fungi (12). Other commonly used
fungal media include potato dextrose agar and Sabouraud
agar (17, 18).
Incubation periods for fungi typically range from 3 to 7
days; however, fastidious xerophiles may require several weeks
for mycelial development (19). Most airborne fungi are
mesophilic and grow well at temperatures of 20°C to 25°C,
and medically important fungal pathogens grow well at
30°C (20). Incubation temperature can also be used to select
for certain species.
Aspergillus fumigatus is a thermotolerant species that can be
cultured at 37°C to 45°C, above the temperature range of
mesophilic fungi.
For the culture of bacteria, several broad-spectrum
media, such as tryptic soy agar, nutrient agar, and casein soy
peptone agar, may be used (4, 8). These media are often
amended with antibiotics to restrict the growth of fungi.
Blood agar has also been used for the collection of air samples
in hospitals and other clinical settings (3). Incubation tem-
peratures from 28°C to 35°C for 1–7 days are usually used
for environmental and human source bacteria; an important
exception being thermophilic actinomycetes, which are cul-
tured at 55°C.
Enumeration
Viable counts are determined after the appropriate incuba-
tion period by enumerating the colony forming units
(CFUs). For volumetric sampling the concentration of cul-
turable airborne microorganisms (CFU per m3) is determined
by dividing the number of CFUs per sample by the volume of
air sampled. When microorganisms are collected using
multiple-jet impactor samplers (e.g., Andersen and SAS;
Thermo Scientific, Waltham, MA, and Bioscience Interna-
tional Rockville, MD, respectively), positive-hole correc-
tions are generally applied to the data (21–23). These
corrections are estimations that account for the probability
of multiple impactions of microorganisms through the same
sampling jet, which results in the enumeration of only a single
colony at that impaction site. The magnitude of the correc-
tion increases with the number of CFUs, and the uncertainty
of the estimate increases near the upper quantification limit of
the sampler (21). The corrected CFU count for a sample is
used to calculate the airborne concentration of microorgan-
isms. When organisms are collected through a single nozzle
onto a stationary or moving surface, colonies may overlap
each other, but the resulting reduced number of CFUs can
be corrected (24).
Enumeration errors are associated with colony counts that
are very low or very high. Colony counts that are too low can
be nonrepresentative of the population and exhibit high var-
iability. As colony counts increase, counting errors increase
because of overlap of colonies and inhibitory effects of micro-
organisms on one another. Researchers have proposed that
upper limits for efficient detection be determined by colony
size and the resolution ability of the enumeration system
(24). When filtration or impingement sampling is used, the
wash solution or collection buffer can be serially diluted to
obtain counts within an acceptable range. However, when
collecting air samples directly onto an agar surface without
knowledge of the bioaerosol concentration being sampled,
too few or too many colonies may result on a plate. These dif-
ficulties can be lessened by setting more than one sample col-
lection time.
Identification
Fungal identification is based largely on the morphological
characteristics of spores and spore-bearing structures by using
direct microscopy. Misidentification of fungi is a significant
source of error in culturable sample analysis; therefore, accu-
rate fungal identification requires training and experience.
Bacterial isolates may be identified by a variety of methods.
Classical biochemical methods may be employed for identifi-
cation to the genus and species level, or procedures such as
Gram staining may be used to differentiate bacterial groups.
A variety of identification systems that classify bacteria
according to various biochemical characteristics, such as sub-
strate utilization and cellular fatty acid profiles, are available.
However, many of the databases used to identify microorgan-
isms are composed mainly of clinical isolates; therefore, the
identification of environmental microorganisms may be
problematic.
Microscopy
Microscopic enumeration methods are used to obtain
an estimate of the total number of microorganisms pre-
sent in a sample. In contrast with culture techniques, micro-
scopic analysis allows enumeration of both culturable and
nonculturable microorganisms. However, identification of
microorganisms to the species level is usually not possible
without the aid of a taxon-specific technique, such as immu-
nospecific fluorescence staining (described shortly). A variety
of stains may be utilized to differentiate biological particles
from nonbiological material and respiring cells from nonres-
piring cells.
Impaction onto a glass slide or tape or filtration sampling
methods are generally used for the determination of total air-
borne fungal spores. Air samples obtained with Burkard spore
traps are analyzed via light microscopy by examining the col-
lection surface. Often the entire spore trap trace is analyzed;
alternatively, several traverses of the collection slide or tape
are enumerated and the total number of microorganisms is
estimated by extrapolation. The volume of air represented
by a traverse can be calculated, and the average number of
fungal spores per traverse is used to determine the total num-
ber of spores per cubic meter. When the sample is collected
onto a moving slide, time discrimination of fungal spore con-
centration is possible by enumerating traverses between speci-
fied points along the collection surface. Although staining is
not required for enumeration, several stains, including lacto-
phenol cotton blue, phenosaffranin, and basic fuchsin, are
commonly used to facilitate discrimination of spores from
debris (4). Accurate identification to the genus level is possi-
ble for a limited number of fungal spore types. Data are usually
reported as total number of spores per cubic meter of air or the
percent composition of fungal genera or groups. Fungal spores
collected by filtration may be washed from the filter mem-
brane, stained with acridine orange, and enumerated by epi-
fluorescence microscopy (25). However, some fungal spore
types may resist staining or have dark pigmentation which
masks fluorescence (4).
For the determination of total airborne bacteria by micro-
scopic analysis, liquid impingement or filtration sampling is
used. Aliquots of collection buffer or the filter membrane
wash solution may be stained for epifluorescence microscopy
by the acridine orange direct count method (25–27); how-
ever, more than 104 cells per sample are required to obtain
meaningful results with this method. A field study using swirl-
ing liquid impingers demonstrated that direct epifluorescence
microscopy detected total airborne microbial concentrations

that were 3 to 1000 times higher than those obtained by con-
ventional culture analysis (28). Other researchers have used
epifluorescence microscopy to detect the fluorescent signal
given by the DNA probes used in fluorescent in situ hybridiza-
tion assays. This method has been used for the detection of
microorganisms in air samples from a wastewater treatment
facility, and the results were comparable to those obtained
with culturable counts and 40 6-diamidino-2-phenylindole
(DAPI) staining (29). Cells and endospores may also be
enumerated by bright-field or phase-contrast microscopy
using a hemocytometer or counting chamber.
The advantage of microscopic analysis of air samples is the
determination of total airborne microorganisms, both cultur-
able and nonculturable. Because a significant fraction of air-
borne microorganisms may be nonculturable, microscopy is
often used in conjunction with culture methods to analyze
bioaerosol samples. The major disadvantage is the tedious
and time-consuming nature of microscopic enumeration,
although computerized image analysis systems may automate
the process. Another drawback is the level of expertise
required to identify fungal spores. Misidentification of fungal
spores and the inability to distinguish microorganisms from
nonbiological particles are common sources of error using
this method (4).
FLOW CYTOMETRY
Flow cytometry offers an alternative to microscopic enumer-
ation of total cells. A target-specific fluorescent probe is
added to a liquid sample and subjected to a laser beam
(30). Cell concentrations can be measured by light scattering
and fluorescence emitted from fluorochromes bound to cells.
An advantage of flow cytometry is its ability to be used for the
simultaneous detection of several targets; however, assay
optimization can be tedious (30). In one study, impinger sam-
ples labeled with DAPI yielded comparable results when
enumerated by microscopy or flow cytometry (31). Similarly,
in another study, laboratory and field impinger samples were
collected; the data showed that flow cytometry results
obtained for the detection of airborne fungi were in close
agreement with epifluorescence microscopy counts and that
flow cytometry data were more precise and reliable at fungal
concentrations between 103 and 106 spores/ml (32). Other
researchers have used flow cytometry to determine total and
viable bacterial concentrations (33). In a different applica-
tion of flow cytometry, researchers coupled sandwich immu-
noassay utilizing microsphere beads with flow cytometry
for the detection of biothreat agents (34). Following the
immunoassay, the bound analytes were detected with a com-
mercially available flow cytometer by using secondary anti-
bodies labeled with the fluorescent dye phycoerythrin. Flow
cytometry coupled with a fluorochrome has also been inves-
tigated as a real-time method for detection of airborne fungi
and bacteria (35).
IMMUNOASSAYS
Immunoassay methods, used extensively in biomedical
research, have been recognized for their potential application
to bioaerosol analysis (36). Immunoassays rely on the binding
of antibodies to a specific target antigen. Target antigens may
be (a) cell surface-associated proteins or polysaccharides or
(b) human allergens. The development of an antibody is a
critical component that affects the sensitivity and specificity
of the assay; therefore, monoclonal antibodies are often used.
Among the methods that may be applied to bioaerosol
3.2.3. Analysis of Bioaerosol Samples ▪ 3.2.3-3
analysis are fluorescence immunoassay, enzyme immunoassay,
and radioimmunoassay.
Fluorescence immunoassay consists of staining samples
with a fluorescently labeled antibody that binds specifically
to the antigens on the surfaces of the target organisms and
enumeration by epifluorescence microscopy. Fluorescent
labeling dyes include fluorescein, fluorescein isothiocyanate,
and rhodamine isothiocyanate (8). Appropriate controls
are necessary to determine levels of background and nonspe-
cific fluorescence. An enzyme-linked immunosorbent assay
(ELISA)–based flow-through capture instrument was devel-
oped for the detection of bacterial endospores and cells and
is capable of analyzing large sample volumes in a short period
of time (37). Other researchers have reported that the detec-
tion of satratoxins by an ELISA was comparable to results
obtained with high-performance liquid chromatography
(HPLC) (38). The immunoassay technique has also been
utilized in the development of hand-held assays (HHAs) or
lateral flow assays for the detection of biothreat agents (30).
HHAs are disposable devices that employ the same technol-
ogy as that of home pregnancy test kits and consist of card-
board tickets containing immobilized antibodies specific to
a target organism (39–41). The HHA has a positive control
well and a test well. Application of a liquid sample initiates
an immunological reaction when the target organism is
present. If a colored line develops in both wells, the result is
considered positive; if a line develops only in the control
well, the result is considered negative. Limitations of these
devices include false-positive results and poor sensitivity
(42). Although HHAs were designed to provide data quickly,
the U.S. Department of Defense has determined that they are
effective only when employed as part of a methodical detec-
tion approach that provides additional levels of confirmation
(41). Improved sensitivity and quantitative detection with
this assay have been explored by incorporating fluorescent
microspheres or paramagnetic beads and analysis by a compat-
ible reader (30).
Radioimmunoassay consists of combining a specific
antibody with a radioactive label, while enzyme immunoassay
utilizes binding of the antibody or the antigen to an enzyme.
The concentration of antigen is measured by radioactivity or
enzyme activity. These methods have been applied for the
measurement of airborne allergens, such as dust mite allergen,
animal dander, and ß-1,3-glucan (43–47). Impaction,
impingement, and filtration sampling methods are compati-
ble with the assay (48). The advantages of these methods
for quantification of airborne allergen are their specificity
and sensitivity. For example, the lower detection limit of an
ELISA for most household allergens can be 0.8 ng/ml (49).
The major limitation of immunoassays is that specific anti-
gens for microorganisms are difficult to define and standardize
(20). As more microbial antigenic compounds are character-
ized, immunoassay analysis methods may be used to provide
data necessary for assessing environmental exposure to
allergens.
BIOCHEMICAL ASSAYS
Biochemical methods of analysis are used to measure a biolog-
ical compound of interest, such as endotoxins or mycotoxins.
Because inhalation of these compounds may produce adverse
health effects (50, 51), there have been efforts to measure
these compounds directly to relate environmental exposure
to human response. Endotoxins are lipid polysaccharides
(LPSs) in the cell walls of Gram-negative bacterial cells and
cyanobacteria (52). Airborne endotoxin is widespread
3.2.3-4 ▪ AIR
because of the ubiquity of Gram-negative bacteria in the
environment, and elevated levels of endotoxin have been
measured in bioaerosol samples from a variety of agricultural,
industrial, and office environments (53, 54). The most widely
used method for measurement of endotoxin is the Limulus
amebocyte lysate (LAL) test. The lysate of amebocytes from
the horseshoe crab, Limulus polyphemus, gels in the presence
of LPS. This reaction forms the basis of endotoxin quantifica-
tion, and a variety of test systems are commercially available.
Airborne endotoxin data are expressed as ng per m3 or endo-
toxin units (EUs) per m3. EU is defined as the potency of
0.10 ng of a reference standard endotoxin (55). Because
endotoxin potency varies among Gram-negative bacteria,
data are often expressed as EUs to facilitate comparisons
between studies. The lower detection limit of the LAL test
is 0.5 pg/ml (49). Filtration sampling is most often used to col-
lect airborne endotoxin; however, various filter materials
have been shown to inhibit the LAL assay (56). Differences
in collection and analysis methods between laboratories has
made comparison of results difficult. Recently, Duquenne
et al. (52) published a literature review that captures the cur-
rent status of standardization for measuring airborne endo-
toxin and constitutes a reference document of available
quantification methods. Another limitation observed with
endotoxin analysis is interference with the LAL reaction by
inhibitors present in environmental samples (49, 55). Other
biochemical methods that may be used in the analysis of bac-
teria in indoor environments include branched-chain fatty
acids and muramic acid assays (57).
Mycotoxins associated with dusts from fungus-
contaminated grains have long been recognized as a source
of illness (58). Exposure to Stachybotrys chartarum mycotoxins
has been associated with toxicoses in contaminated home
and office environments (59–61). Filtration sampling using
glass-fiber filters or polycarbonate membranes has been used
to measure aerosolized mycotoxins of S. chartarum in the
laboratory (62, 63) and in the field (64). Field sampling
conducted in Finnish water-damaged buildings utilized high-
pressure liquid chromatography, mass spectrometry, and elec-
trospray ionization to analyze bulk samples for the presence
of mycotoxins (65). This study demonstrated that mycotox-
ins were found in more than 40% of the 79 bulk samples ana-
lyzed and that some toxigenic fungal species were usually
present in mycotoxin-positive samples even if the species
responsible for mycotoxin production was not isolated. Other
researchers have demonstrated the presence of ergot alka-
loids, a mycotoxin, associated with conidia of Aspergillus
fumigatus (66). In addition, HPLC coupled to mass spectrom-
etry (HPLC-MS) has also been used to characterize Aspergil-
lus isolates (i.e., A. flavus, A. parasiticus, and A. fumigatus) as
toxigenic or nontoxigenic. The study found that 51% of the
strains isolated produced at least one mycotoxin after culture
in the laboratory (67).
Other biochemical methods which may be used as surro-
gates of airborne fungal biomass exposure include ergosterol
(68, 69), mold extracellular polysaccharides (70), and
ß-1,3-glucan assays (49, 57, 71–74). An alternative to ergo-
sterol content as a means to estimate fungal biomass is the flu-
orogenic detection of beta-N-acetylhexosaminidase activity
(75). Bauer et al. (76) observed a high correlation between
the sugar alcohols arabitol and mannitol and fungal spore
concentrations, and they suggest the use of these markers
for quantification of fungal spores in PM10. Microbial volatile
organic compounds (MVOCs) are products of fungal and bac-
terial metabolism that may cause toxic and allergic reactions.
Volumetric air samples collected onto activated carbon
media are analyzed for MVOCs by gas chromatography/
mass spectrometry. A study involving bioaerosol monitoring
of residences cleaned and reoccupied following a major flood
found that MVOCs were detected in over half of the houses
tested (28). Although MVOCs cannot be used to quantify
fungi, they can serve as good indicators of fungal amplifica-
tion in indoor environments (77), provided that artificial
sources are considered and indoor/outdoor ratios are estab-
lished (28). Studies have shown that clean building materials
may produce some of the same VOCs emitted by microorgan-
isms (77). Therefore, background VOCs need to be deter-
mined in nonproblem buildings prior to utilizing MVOC
data as an indicator of microbial growth in indoor environ-
ments (77). Chen and Hildemann (70) point out that there
are several limitations to the use of biomarkers. Namely, the
methods cannot differentiate between viable and nonviable
microorganisms, genera and/or species cannot be identified,
different microbial species have different amounts of bio-
marker per cell, and there is a variation in specificity of the
biomarkers.
PCR
PCR is a procedure used to rapidly amplify specific DNA
sequences (78). Conventional PCR involves the separation
or melting of DNA into strands, annealing of oligonucleotide
primers, and elongation by means of the DNA polymerase
(79). The assay involves temperature cycling and results in
the production of multiple copies of the target DNA. This
technique has been used successfully to enhance the detec-
tion of microorganisms in a variety of matrices, including
air samples (5, 80–84). Application of the PCR technique
to bioaerosol sampling can result in the rapid, sensitive detec-
tion of the target microorganism(s). PCR is particularly suited
for detection of specific microorganisms that are difficult to
culture, grow very slowly, or have never been cultured in vitro.
The method also provides an increase in sensitivity over tradi-
tional culture methods (5, 85–87). The amplified DNA can
be visualized, quantified, or further analyzed by genetic fin-
gerprinting or microarrays (79).
Another advantage of PCR is the rapidity with which
results are obtained compared with culture counts. It is possi-
ble to obtain results within hours of sample collection
with PCR, compared with days or weeks for culture methods.
It should be noted that the PCR can be used for the direct
detection of microorganisms provided that unique, non–
cross-hybridizing sequences are available for the specific
microorganisms to serve as primers for amplification. The
18S or 16S ribosomal RNA genes are commonly used for
the design of genus-specific primers and probes because
they contain highly conserved sequences between members
of the same genus, but the sequences are variable among dif-
ferent genera. The internal transcribed spacer regions and
intergenic spacer of the nuclear rRNA gene are often the tar-
get of species-specific primers and probes, because these are
highly variable areas within a genus or among populations
(88). The design of species-specific primers and probes is dif-
ficult due to the homology in the DNA of closely related spe-
cies. Although a sequence comparison may indicate that the
designed primer set will amplify only the target organism, it is
crucial that the designed primers and probes be thoroughly
tested in the laboratory with DNA extracted from both the
target and other (nontarget) organisms to confirm that they
amplify the DNA from all target strains and will not cross-
react with nontarget strains (89). Researchers have evaluated
published primers for the specific detection of common
 3.2.3. Analysis of Bioaerosol Samples ▪ 3.2.3-5

airborne fungi and verified the specificity of 28 sets of pri-
mers (90). With conventional PCR, the authenticity of the
amplified DNA can be confirmed by sequencing, restriction
analysis, or gene probe hybridization. A source of published
gene probes is probeBase, an online database of probes target-
ing rRNA genes (http://www.microbial-ecology.net/probe
base/) (91). If there are multiple organisms of interest in a
sample, it may be possible to use multiplex PCR in the
same reaction mix to simultaneously amplify several DNA
targets of interest (92–94).
A limitation of the PCR assay is the inability to provide a
measure of viability as obtained with culture analysis. Techni-
cal expertise and equipment cost can also be considered a lim-
itation (3). Quantification of the concentration of the DNA
segments that serve as templates for PCR has been reported
for several sample matrices (95–97). The major difficulty
associated with PCR as a quantitative method is that no other
technique yet matches its sensitivity and can be used to
compare results. Thus, any data obtained by quantitative
PCR should be supported by carefully designed controls and
compared with reliable standards (98). Guidelines that
describe the minimum information essential for publica-
tion/evaluation of quantitative PCR (qPCR) experiments
are available (99).
Sample Processing
Selection of environmental sampling and processing proto-
cols must consider methods that yield a liquid sample that
is compatible with PCR analysis. A concentration step is
required for PCR to be effectively applied to air sample anal-
ysis. Two sample collection methods, liquid impingement
and filtration, have been utilized successfully for air sample
collection, with subsequent PCR detection of the target
microorganism (5, 82, 83, 100).
Inhibition
Another limitation of PCR is the presence of exogenous sub-
stances or inhibitors in environmental samples and in pure
cultures (5, 101–105). The presence of PCR inhibitors in
environmental samples may reduce sensitivity or result in
false negatives (39, 85, 101–103, 106). PCR inhibition
can result from the sequestration or degradation of the tar-
get DNA (including the primers), or decreased activity of
the DNA polymerase, etc. (79). A nucleic acid purification
step or dilution of the sample may be required to remove envi-
ronmental interference to PCR amplification; however, these
procedures can reduce the sensitivity of detection (84, 100,
101). Additives, such as bovine serum albumin, have been
successfully utilized to minimize PCR inhibition (107). An
essential component of a PCR experiment should be the
inclusion of dilution or a universal inhibition assay (99); in
their absence, the results obtained should include a qualifier
(105). An internal positive control (IPC) should be incorpo-
rated into the PCR assay to determine whether the samples
contain PCR inhibitors, thus avoiding the risk of reporting
false negative results. An IPC consists of a known quantity
of control DNA, primers, and probe. Inhibition of the PCR
assay is observed by reduced amplification results of the IPC
DNA compared with control samples (105).
Quantitative PCR
Two primary approaches exist that are capable of providing
accurate, reproducible measurement of initial target DNA
concentrations in samples: competitive PCR and qPCR using
fluorogenic probes (95, 108). The first approach, competitive
PCR, utilizes a known amount of internal control DNA com-
petitor in each reaction (Fig. 1a). A suitable competitor acts
as an internal standard and has the same primer binding sites
as the target DNA. Quantification is achieved by comparing
the concentration of the target product to that of the compet-
itor. An important consideration is that the competitor must
amplify with approximately the same efficiency as the target
template; therefore, a significant effort must be applied in
the development and testing of the competitor (109). After
amplification, a variety of post-PCR methods such as gel
electrophoresis followed by ethidium bromide staining or

(a) (b)
Target template Target template
+
Internal control DNA competitor
Oligonucleotide primers
and fluorescent probe
Oligonucleotide
primers QPCR

cPCR Monitoring of
fluorescence in real time
Post PCR
manipulations

Target template Internal control

DNA competitor

FIGURE 1 Diagrams illustrating competitive polymerase chain reaction (cPCR) and quantitative PCR (qPCR). (a) In cPCR, the target
template and the internal control DNA compete for the oligonucleotide primers during the PCR assay. Post-PCR manipulations are required
to separate the target and competitor products and determine the ratio between them (i.e., final product concentrations). (b) In qPCR, a fluo-
rescent probe anneals to the target DNA between the specific primer binding sites. As new DNA is synthesized, the probe is cleaved by the Taq
polymerase, causing it to fluoresce, and the amount of fluorescence is used to measure the initial target DNA concentration.
doi:10.1128/9781555818821.ch3.2.3.f1
3.2.3-6 ▪ AIR
radioactivity measurement (if radioactive deoxynucleoside
triphosphates are used), quantification of fluorescence with
image analysis systems, or HPLC can be used to determine
the final product concentrations (108).
The second approach, qPCR, is a real-time assay that
measures target product accumulation with fluorogenic
probes (Fig. 1b). The TaqMan QPCR (Applied Biosystems;
Life Technologies, Grand Island, NY) uses a fluorescent oli-
gonucleotide probe that anneals to the target DNA between
the specific primer binding sites. As new DNA is synthesized
the probe is cleaved by the Taq polymerase, causing it to flu-
oresce, and the amount of fluorescence is used to measure the
initial DNA concentration. The probe also provides confir-
mation of the product, as it fluoresces only in the presence
of the target template. The TaqMan assay has been demon-
strated to accurately and rapidly detect bacteria (95, 110–
112), fungi of indoor air quality significance (102, 103,
113–115), and viruses (116) and also to detect bacteria serv-
ing as simulants for biological warfare agents (39, 40, 85).
Other qPCR instruments have been developed that use
the fluorescent dye SYBR green or two fluorescently labeled
probes that serve as hybridization probes to the double-
stranded amplified DNA (LightCycler; Roche, Indianapolis,
IN; R.A.P.I.D. System; Idaho Technology, Salt Lake City,
UT). These instruments have been utilized in the detection
of fungi such as Aspergillus fumigatus and Candida albicans
(117), and bacteria such as Staphylococcus aureus (118) and
Bacillus anthracis (119).
EMERGING TECHNOLOGIES
Recent advances in microarray technology and microfluidics
are being utilized in the development of fast and sensitive
sensing systems that are of particular importance in the detec-
tion of biothreat agents (120), and several microarray-based
systems are commercially available. Microarrays allow for
thousands of hybridizations to be analyzed simultaneously
(79). Nanotechnology has been used for detection of targets
in liquid samples, and microfluidics has been applied to the
detection of viruses (49). Hybrid technologies are being
developed that combine more than one analysis method to
maximize the accuracy of the results obtained (120). One
such system utilizes immunoassay, flow cytometry, and PCR
analysis in various combinations for the continuous monitor-
ing of airborne biological threat agents (34, 121). Other
researchers have coupled PCR and enzyme immunoassay for
the differentiation of Aspergillus species from other fungi
(122). Automation has been explored to increase sample
throughput, maximize accuracy, and reduce costs in the
detection of biological threat agents (123). Plasma-assisted
fragmentation coupled with plasma electrochemistry has
been proposed as a sensitive method to detect airborne par-
ticles (124), and pyrosequencing has been utilized to charac-
terize (fungal and bacterial) bioaerosols in a poultry
production facility (125).
With the wide variety of bioaerosol sampling and analysis
methods in use, the investigator must have a clear under-
standing of the objectives of bioaerosol sampling before sam-
pling is conducted. This information will assist in the
selection of the appropriate sampling and analysis methodol-
ogies for collecting meaningful bioaerosol data. As existing
methods are improved and new techniques are developed,
more accurate bioaerosol information may be obtained,
which will further the understanding of human exposure to
airborne microorganisms and aid in the development of
standardized monitoring protocols.
REFERENCES
1. Eduard W, Heederik D. 1998. Methods for quantitative
assessment of airborne levels of noninfectious microorgan-
isms in highly contaminated work environments. Am
Indust Hyg Assoc J 59:113–127.
2. Levetin E. 1995. Fungi, p. 87–120. In Burge H (ed), Bioaer-
osols. Lewis Publishers, Boca Raton, FL.
3. D’Arcy N, Canales M, Spratt DA, Lai K. 2012. Healthy
schools: standardisation of culturing methods for seeking
airborne pathogens in bioaerosols emitted from human
sources. Aerobiologia 28:413–422.
4. Burge HA. 1995. Bioaerosol investigations, p. 1–23. In
Burge HA (ed.), Bioaerosols. CRC Press, Boca Raton, FL.
5. Alvarez AJ, Buttner MP, Toranzos GA, Dvorsky EA,
Toro A, Heikes TB, Mertikas LE, Stetzenbach LD.
1994. The use of solid-phase polymerase chain reaction
for the enhanced detection of airborne microorganisms.
Appl Environ Microbiol 60:374–376.
6. Buttner MP, Stetzenbach LD. 1991. Evaluation of four
aerobiological sampling methods for the retrieval of aero-
solized Pseudomonas syringae. Appl Environ Microbiol 57:
1268–1270.
7. Walter MV, Marthi B, Fieland VP, Ganio LM. 1990.
Effect of aerosolization on subsequent bacterial survival.
Appl Environ Microbiol 56:3468–3472.
8. Jensen PA, Lighthart B, Mohr AJ, Shaffer BT. 1994.
Instrumentation used with microbial bioaerosol. In Light-
hart B, Mohr AJ (eds), Atmospheric microbial aerosols, theory
and applications. Chapman and Hall, New York.
9. Burge HP, Solomon WR, Boise JR. 1977. Comparative
merits of eight popular media in aerometric studies of fungi.
J Allergy Clin Immunol 60:199–203.
10. Morring KL, Sorenson WG, Attfield MD. 1983. Sam-
pling for airborne fungi: a statistical comparison of media.
Am Ind Hyg Assoc J 44:662–664.
11. Smid T, Schokkin E, Boleij JSM, Heederik D. 1989. Enu-
meration of viable fungi in occupational environments: a
comparison of samplers and media. Am Ind Hyg Assoc J
50:235–239.
12. Verhoeff AP, van Wijnen JH, Boleij JSM, Brunekreef B,
van Reenen-Hoekstra ES, Samson RA. 1990. Enumera-
tion and identification of airborne viable mould propagules
in houses. Allergy 45:275–284.
13. Hunter CA, Grant C, Flannigan B, Bravery AF. 1988.
Mould in buildings: the air spora of domestic buildings.
Intl Biodet 24:81–101.
14. Strachan DP, Flannigan B, McCabe EM, McGarry F.
1990. Quantification of airborne moulds in the homes
of children with and without wheeze. Thorax 45:382–
387.
15. Banks JG, Board RG, Carger J, Dodge AD. 1985. The
cytotoxic and photodynamic inactivation of microorgan-
isms by rose bengal. J Appl Bacteriol 58:392–400.
16. Hocking AD, Pitt JI. 1980. Dichloran-glycerol medium for
enumeration of xerophilic fungi from low-moisture foods.
Appl Environ Microbiol 39:488–492.
17. Hirsch SR, Sosman JA. 1976. A one-year survey of mold
growth inside twelve homes. Ann Allergy 36:30–38.
18. Solomon WR. 1975. Assessing fungus prevalence in
domestic interiors. J Allergy Clin Immunol 56:235–242.
19. Beuchat LR. 1992. Media for detecting and enumerating
yeasts and moulds. Intl J Food Microbiol 17:145–158.
20. Madelin TM, Madelin MF. 1995. Biological analysis of
fungi and associated molds, p. 361–386. In Cox CS, Wathes
CM (eds), Bioaerosols Handbook. CRC Press, Boca Raton,
FL.
21. Andersen AA. 1958. New sampler for the collection, siz-
ing, and enumeration of viable airborne particles. J Bacteriol
76:471–484.

22. Buttner MP, Stetzenbach LD. 1993. Monitoring air-
borne fungal spores in an experimental indoor environ-
ment to evaluate sampling methods and the effects of
human activity on air sampling. Appl Environ Microbiol
59:219–226.
23. Macher JM. 1989. Positive-hole correction of multiple-jet
impactors for collecting viable microorganisms. Am Ind
Hyg Assoc J 50:561–568.
24. Chang CW, Hwang YH, Grinshpun SA, Macher JM,
Willeke K. 1994. Evaluation of counting error due to col-
ony masking in bioaerosol sampling. Appl Environ Microbiol
60:3732–3738.
25. Palmgren U, Strom G, Blomquist G, Malmberg P. 1986.
Collection of airborne micro-organisms on Nucleopore
filters, estimation and analysis—CAMNEA method.
J Appl Bacteriol 61:401–406.
26. Hobbie JE, Daley RJ, Jasper S. 1977. Use of nucleopore
filters for counting bacteria by fluorescence microscopy.
Appl Environ Microbiol 33:1225–1228.
27. Walters M, Milton D, Larsson L, Ford T. 1994. Airborne
environmental endotoxin: a cross validation of sampling
and analysis techniques. Appl Environ Microbiol 60:
996–1005.
28. Fabian MP, Miller SL, Reponen T, Hernandez MT. 2005.
Ambient bioaerosol indices for indoor air quality assess-
ments of flood reclamation. J Aerosol Sci 36:763–783.
29. Patentalakis N, Pantidou A, Kalogerakis N. 2008. Deter-
mination of enterobacteria in air and wastewater samples
from a wastewater treatment plant by epi-fluorescence
microscopy. Water Air Soil Poll Focus 8:107–115.
30. Andreotti PE, Ludwig GV, Peruski AH, Tuite JJ, Morse
SS, Peruski LF Jr, 2003. Immunoassay of infectious agents.
Bio Techn 35:850–859.
31. Lange JL, Thorne PS, Lynch N. 1997. Application of flow
cytometry and fluorescent in situ hybridization for assess-
ment of exposures to airborne bacteria. Appl Environ Micro-
biol 63:1557–1563.
32. Prigione V, Lingua G, Filipello Marchisio V. 2004. Devel-
opment and use of flow cytometry for detection of airborne
fungi. Appl Environ Microbiol 70:1360–1365.
33. Rule AM, Kesavan J, Schwab KJ, Buckley TJ. 2007.
Application of flow cytometry for the assessment of
preservation and recovery efficiency of bioaerosol samplers
spiked with Pantoea agglomerans. Environ Sci Technol 41:
2467–2472.
34. McBride MT, Masquelier D, Hindson BJ, Makarewicz
AJ, Brown S, Burris K, Metz T, Langlois RG, Tsang
KW, Bryan R, Anderson DA, Venkateswaran KS, Mila-
novich FP, Colston BW Jr. 2003. Autonomous detection
of aerosolized Bacillus anthracis and Yersinia pestis. Anal
Chem 75:5293–5299.
35. Chen PS, Li CS. 2005. Bioaerosol characterization by flow
cytometry with fluorochrome. J Environ Monit 7:950–959.
36. Burge HA, Solomon WR. 1987. Sampling and analysis of
biological aerosols. Atmos Environ 21:451–456.
37. Weimer BC, Walsh MK, Beer C, Koka R, Wang X. 2001.
Solid-phase capture of proteins, spores, and bacteria. Appl
Environ Microbiol 67:1300–1307.
38. Chung Y-J, Jarvis BB, Tak H, Pestka JJ. 2003. Immuno-
chemical assay for satratoxin G and other macrocyclic tri-
chothecenes associated with indoor air contamination by
Stachybotrys chartarum. Tox Mech Meth 13:247–252.
39. Buttner MP, Cruz P, Stetzenbach LD, Klima-Comba AK,
Stevens VL, Cronin TD. 2004. Determination of the effi-
cacy of two building decontamination strategies by surface
sampling with culture and quantitative PCR analysis. Appl
Environ Microbiol 70:4740–4747.
40. Buttner MP, Cruz P, Stetzenbach LD, Klima-Comba AK,
Stevens VL, Emanuel PA. 2004. Evaluation of the
3.2.3. Analysis of Bioaerosol Samples ▪ 3.2.3-7
biological sampling kit (BiSKit) for large-area surface sam-
pling. Appl Environ Microbiol 70:7040–7045.
41. Emanuel PA, Chue C, Kerr L, Cullin D. 2003. Validating
the performance of biological detection equipment: the
role of the federal government. Biosecur Bioterror 1:
131–137.
42. Centers for Disease Control and Prevention (CDC).
2012. Emergency preparedness and response Anthrax Q &
A: laboratory testing http://www.bt.cdc.gov/agent/anthrax/
faq/labtesting.asp.
43. DeBlay F, Chapman MD, Platts-Mills TA. 1991. Air-
borne cat allergen Feld I: environmental control with the
cat in situ. Am Rev Respir Dis 143:1334–1339.
44. Douwes J, Doekes G, Montijn R, Heederik D, Brunek-
reef B. 1996. Measurement of β(1-3)-glucans in occupa-
tional and home environments with an inhibition
enzyme immunoassay. Appl Environ Microbiol 62:
3176–3182.
45. Luczynska CM, Li Y, Chapman MD, Platts-Mills TAE.
1990. Airborne concentration and particle size distribution
of allergen derived from domestic cats (Felis domesticus):
measurements using cascade impactor, liquid impinger
and a two site monoclonal antibody assay for Fel d I. Am
Rev Respir Dis 141:361–367.
46. Platts-Mills TAE, Thomas WR, Aalbersee RC, Vervloet
D, Chapman MD. 1992. Dust mite allergens and asthma:
report of a second international workshop. J Allergy Clin
Immunol 89:1046–1060.
47. Sander I, Fleischer C, Borowitzki G, Brüning T, Raulf-
Heimsoth M. 2008. Development of a two-site enzyme
immunoassay based on monoclonal antibodies to measure
airborne exposure to (1-3)-beta-D-glucan. J Immunol
Meth 337:55–62.
48. Chapman MD. 1995. Analytical methods: immunoassays.
In Burge HA (ed), Bioaerosols. CRC Press, Boca Raton, FL.
49. Xu Z, Wu Y, Shen F, Chen Q, Tan M, Yao M. 2011. Bio-
aerosol science technology and engineering: past, present,
and future. Aerosol Sci Technol 45:1337–1349.
50. Castellan RM, Olenchock SA, Kinsley KB, Hankinson
JL. 1987. Inhaled endotoxin and decreased spirometric
values: an exposure-response relation for cotton dust.
N Eng J Med 317:605–610.
51. Crook B, Higgins S, Lacey J. 1987. Methods for sampling
microorganisms at solid waste disposal sites, p. 791–797. In
Houghton DR, Smith RN, Eggins HOW (eds), 7th Intl Bio-
deterioration Symposium, vol. 7. Elsevier, London.
52. Duquenne P, Marchand G, Duchaine C. 2012. Measure-
ment of endotoxins in bioaerosols at workplace: a critical
review of literature and a standardization issue. Ann Occup
Hyg 57:137–172.
53. Jacobs RR. 1989. Airborne endotoxins: an association
with occupational lung disease. Appl Ind Hyg 4:50–56.
54. Liu J, Pankhurst LJ, Deacon LJ, Abate W, Hayes ET,
Drew GH, Longhurst PJ, Pollard S, Longhurst J, Tyrrel
SF, Jackson SK. 2011. Evaluation of inflammatory effects
of airborne endotoxin emitted from composting sources.
Environ Toxicol Chem 30:602–606.
55. Milton DK. 1995. Endotoxin, p. 77–86. In Burge HA (ed),
Bioaerosols CRC Press, Boca Raton, FL.
56. Milton DK, Gere RJ, Feldman HA, Greaves IA. 1990.
Endotoxin measurement: aerosol sampling and applica-
tion of a new Limulus method. Am Ind Hyg Assoc J 51:
331–337.
57. Sebastian A, Larsson L. 2003. Characterization of the
microbial community in indoor environments: a
chemical-analytical approach. Appl Environ Microbiol 69:
3103–3109.
58. Flannigan B. 1987. Mycotoxins in the air. Intl Biodet 23:
73–78.
3.2.3-8 ▪ AIR
59. Croft WS, Jarvis BB, Yatawara CS. 1986. Airborne out-
break of trichothecene toxicosis. Atmos Environ 20:
549–552.
60. Jarvis BB, Sorenson WG, Hintikka E, Nikulin M, Zhou
Y, Jiang J, Wang S, Hinkley S, Etzel RA, Dearborn D.
1998. Study of toxin production by isolates of Stachybotrys
chartarum and Memnoniella echinata isolated during a study
of pulmonary hemosiderosis in infants. Appl Environ Micro-
biol 64:3620–3625.
61. Johanning E, Biagini R, Hull D, Morey P, Jarvis B, Land-
sbergis P. 1996. Health and immunology study following
exposure to toxigenic fungi (Stachybotrys chartarum) in a
water-damaged office environment. Int Arch Occup Environ
Health 68:207–218.
62. Pasanen A, Nikulin M, Tuomainen M, Berg S, Parikka P,
Hintikka E. 1993. Laboratory experiments on membrane
filter sampling of airborne mycotoxins produced by Stachy-
botrys atra corda. Atmos Environ 27A:9–13.
63. Sorenson WG, Frazer DG, Jarvis BB, Simpson J, Robin-
son VA. 1987. Trichothecene mycotoxins in aerosolized
conidia of Stachybotrys atra. Appl Environ Microbiol 53:
1370–1375.
64. Yike I, Allan T, Sorenson WG, Dearborn DG. 1999.
Highly sensitive protein translation assay for trichothecene
toxicity in airborne particulates: comparison with cytotox-
icity assays. Appl Environ Microbiol 65:88–94.
65. Tuomi T, Reijula K, Johnsson T, Hemminki K, Hintikka
E-L, Lindroos O, Kalso S, Koukila-Kähkölä P, Mussalo-
Rauhamaa H, Haahtela T. 2000. Mycotoxins in crude
building materials from water-damaged buildings. Appl
Environ Microbiol 66:1899–1904.
66. Panaccione DG, Coyle CM. 2005. Abundant respirable
ergot alkaloids from the common airborne fungus Aspergil-
lus fumigatus. Appl Environ Microbiol 71:3106–3111.
67. Garon D, Kaddoumi AE, Carayon A, Amiel C. 2010.
FT-IR spectroscopy for rapid differentiation of Aspergillus
flavus, Aspergillus fumigatus, Aspergillus parasiticus and char-
acterization of aflatoxigenic isolates collected from agricul-
tural environments. Mycopathologia 170:131–142.
68. Miller JD, Laflamme AM, Sobol Y, Lafontaine P, Green-
halgh R. 1988. Fungi and fungal products in some Cana-
dian houses. Intl Biodet 24:103–120.
69. Miller JD, Young YC. 1997. The use of ergosterol to meas-
ure exposure to fungal propagules in indoor air. Am Ind Hyg
Assoc J 58:39–43.
70. Chen Q, Hildemann LM. 2009. Size-resolved concentra-
tions of particulate matter and bioaerosols inside versus out-
side of homes. Aerosol Sci Technol 43:699–713.
71. Douwes J, Thorne P, Pearce N, Heederik D. 2003. Bio-
aerosol health effects and exposure assessment: progress
and prospects. Ann Occup Hyg 47:187–200.
72. Fogelmark B, Goto H, Yuasa K, Marchat B, Rylander R.
1992. Acute pulmonary toxicity of inhaled ß-1,3-glucan
and endotoxin. Agent Actions 35:50–56.
73. Goto H, Yuasa K, Rylander R. 1994. (1-3)-ß-D-glucan in
indoor air, its measurement and in vitro activity. Am J Ind
Med 25:81–83.
74. Rylander R. 1999. Indoor air-related effects and airborne
(1-3)-ß-D-glucan. Environ Health Perspect 107:501–503.
75. Reesley M, Miller M, Nielsen KF. 2003. Quantifying mold
biomass on gypsum board: comparison of ergosterol and
beta-N-acetylhexosaminidase as mold biomass parameters.
Appl Environ Microbiol 69:3996–3998.
76. Bauer H, Claeys M, Vermeylen R, Schueller E, Weinke
G, Berger A, Puxbaum H. 2008. Arabitol and mannitol
as tracers for the quantification of airborne fungal spores.
Atmos Environ 42:588–593.
77. Korpi A, Pasanen A-L, Pasanen P. 1998. Volatile com-
pounds originating from mixed microbial cultures on
building materials under various humidity conditions.
Appl Environ Microbiol 64:2914–2919.
78. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT,
Erlich H, Arnheim N. 1985. Enzymatic amplification of B-
globin genomic sequence and restriction site analysis for
diagnosis of sickle cell anemia. Science 230:1350–1354.
79. Dungan RS, Leytem AB. 2009. Qualitative and quantita-
tive methodologies for determination of airborne microor-
ganisms at concentrated animal-feeding operations. World
J Microbiol Biotechnol 25:1505–1518.
80. Myatt TA, Johnston SL, Zuo Z, Wand M, Kebadze T,
Rudnick S, Milton DK. 2004. Detection of airborne rhi-
novirus and its relation to outdoor air supply in office envi-
ronments. Am J Respir Crit Care Med 169:1187–1190.
81. Olsson M., Lidman C, Latouche S, Bjorkman A, Roux P,
Linder E, Wahlgren M. 1998. Identification of Pneumocys-
tis carinii f. sp. hominis gene sequences in filtered air in hos-
pital environments. J Clin Microbiol 36:1737–1740.
82. Sawyer MH, Chamberlain CJ, Wu YN, Aintablian N,
Wallace MR. 1994. Detection of varicella-zoster virus
DNA in air samples from hospital rooms. J Infect Dis
169:91–94.
83. Stärk KDC, Nicolet J, Frey J. 1998. Detection of Myco-
plasma hypopneumoniae by air sampling with a nested
PCR assay. Appl Environ Microbiol 64:543–548.
84. Zeng Q-Y, Westermark S-O, Rasmuson-Lestander Å,
Wang X-R. 2004. Detection and quantification of
Wallemia sebi in aerosols by real-time PCR, conventional
PCR, and cultivation. Appl Environ Microbiol 70:7295–
7302.
85. Buttner MP, Cruz-Perez P, Stetzenbach LD. 2001.
Enhanced detection of surface-associated bacteria in
indoor environments by quantitative PCR. Appl Environ
Microbiol 67:2564–2570.
86. Josephson KL, Gerba CP, Pepper IL. 1993. Polymerase
chain reaction detection of nonviable bacterial pathogens.
Appl Environ Microbiol 59:3513–3515.
87. Rinsoz T, Duquenne P, Greff-Mirguet G, Oppliger A.
2008. Application of real-time PCR for total airborne bac-
terial assessment: comparison with epifluorescence micro-
scopy and culture-dependent methods. Atmos Environ
42:6767–6774.
88. White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification
and direct sequencing of fungal ribosomal RNA genes for
phylogenetics. p. 315–322. In Innis MA, Gelfand DH,
Sninsky JJ, White TJ (eds), PCR Protocols: A Guide to
Methods and Applications. Academic Press, San Diego, CA.
89. Cruz P, Stetzenbach LD. 2004. Specific detection of fungi
associated with sick building syndrome when using quanti-
tative polymerase chain reaction. In Straus DC (ed),
Advances in Applied Microbiology, vol. 55: Sick Building Syn-
drome. Elsevier Academic Press, San Diego, CA.
90. Wu Z, Wang XR, Blomquist G. 2002. Evaluation of PCR
primers and PCR conditions for specific detection of com-
mon airborne fungi. J Environ Monit 4:377–382.
91. Loy A, Horn M, Wagner M. 2003. probeBase: an online
resource for rRNA-targeted oligonucleotide probes. Nuc
Acids Res 31:514–516.
92. Chamberlain JS, Gibbs RA, Ranier JE, Nguyen PN,
Cashey CT. 1988. Deletion screening of the Duchenne
muscular dystrophy locus via multiplex DNA amplifica-
tion. Nucl Acids Res 16:11141–11156.
93. Dean TR, Roop B, Betancourt D, Menetrez MY. 2005. A
simple multiplex polymerase chain reaction assay for the
identification of four environmentally relevant fungal con-
taminants. J Microbiol Meth 61:9–16.
94. Luo G, Mitchell TG. 2002. Rapid identification of patho-
genic fungi directly from cultures by using multiplex PCR.
J Clin Microbiol 40:2860–2865.

95. Desjardin LE, Chen Y, Perkins MD, Teixeira L, Cave
MD, Eisenach KD. 1998. Comparison of the ABI 7700
System (TaqMan) and competitive PCR for quantification
of IS6110 DNA in sputum during treatment of tuberculosis.
J Clin Microbiol 36:1964–1968.
96. Leser TD, Boye M, Hendriksen NB. 1995. Survival and
activity of Pseudomonas sp. strain B13(FR1) in a marine
microcosm determined by quantitative PCR and a rRNA-
targeting probe and its effect on the indigenous bacterio-
plankton. Appl Environ Microbiol 61:1201–1207.
97. Picard C, Ponsonnet C, Paget E, Nesme X, Simonet P.
1992. Detection and enumeration of bacteria in soil by
direct DNA extraction and polymerase chain reaction.
Appl Environ Microbiol 58:2717–2722.
98. Ferre F. 1992. Quantitative or semi-quantitative PCR: real-
ity versus myth. PCR Meth Appl 2:1–9.
99. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J,
Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL,
Vandesompele J, Wittwer CT. 2009. The MIQE guide-
lines: minimum information for publication of quantitative
real-time PCR experiments. Clin Chem 55:611–622.
100. Yamamoto N, Kimura M, Matsuki H, Yanagisawa Y.
2010. Optimization of a real-time PCR assay to quantitate
airborne fungi collected on a gelatin filter. J Biosci Bioeng
109:83–88.
101. Alvarez AJ, Buttner MP, Stetzenbach LD. 1995. PCR for
bioaerosol monitoring: sensitivity and environmental
interference. Appl Environ Microbiol 61:3639–3644.
102. Cruz-Perez P, Buttner MP, Stetzenbach LD. 2001. Detec-
tion and quantitation of Aspergillus fumigatus in pure cul-
ture using polymerase chain reaction. Mol Cell Probes 15:
81–88.
103. Cruz-Perez P, Buttner MP, Stetzenbach LD. 2001. Spe-
cific detection of Stachybotrys chartarum in pure culture
using quantitative polymerase chain reaction. Mol Cell
Probes 15:129–138.
104. Cruz P, Buttner MP. 2008. Development and evaluation
of a real-time quantitative PCR assay for Aspergillus flavus.
Mycologia 100:683–690.
105. McDevitt JJ, Lees PSJ, Merz WG, Schwab KJ. 2007.
Inhibition of quantitative PCR analysis of fungal conidia
associated with indoor air particulate matter. Aerobiologia
23:35–45.
106. Wilson IG. 1997. Inhibition and facilitation of nucleic
acid amplification. Appl Environ Microbiol 63:3741–3751.
107. Vadrot C, Bex V, Mouilleseaux A, Squinazi F, Darbord
J-C. 2004. Detection of Mycobacterium tuberculosis com-
plex by PCR in hospital air samples. J Hosp Infect 58:
262–267.
108. Orlando C, Pinzani P, Pazzagli M. 1998. Developments in
quantitative PCR. Clin Chem Lab Med 36:255–269.
109. Heid CA, Stevens J, Livak KJ, Williams PM. 1996. Real
time quantitative PCR. Genome Res 6:986–994.
110. Bassler HA, Flood SJA, Livak KJ, Marmaro J, Knorr R,
Batt CA. 1995. Use of a fluorogenic probe in a PCR-based
assay for the detection of Listeria monocytogenes. Appl Envi-
ron Microbiol 61:3724–3728.
111. Nogva HK, Lillehaug D. 1999. Detection and quantifica-
tion of Salmonella in pure cultures using 50 -nuclease poly-
merase chain reaction. Intl J Food Microbiol 51:191–196.
112. Oppliger A, Charrière N, Droz P-O, Rinsoz T. 2008.
Exposure to bioaerosols in poultry houses at different stages
3.2.3. Analysis of Bioaerosol Samples ▪ 3.2.3-9
of fattening; use of real-time PCR for airborne bacterial
quantification. Ann Occup Hyg 52:405–412.
113. Haugland RA, Vesper SJ, Wymer LJ. 1999. Quantitative
measurement of Stachybotrys chartarum conidia using real
time detection of PCR products with the TaqManTM fluo-
rogenic probe system. Mol Cell Probes 13:329–340.
114. Leenders ACAP, Van Belkum A, Behrendt M, Luijen-
dijk A, Verbrugh HA. 1999. Density and molecular epi-
demiology of Aspergillus in air and relationship to
outbreaks of Aspergillus infection. J Clin Microbiol 37:
1752–1757.
115. Vesper S, Dearborn DG, Yike I, Allan T, Sobolewski J,
Hinkley SF, Jarvis BB, Haugland RA. 2000. Evaluation
of Stachybotrys chartarum in the house of an infant with pul-
monary hemorrhage: quantitative assessment before, dur-
ing, and after remediation. J Urban Health 77:68–84.
116. Pyankov OV, Agranovski IE, Pyankova O, Mokhonova
E, Mokhonov V, Safatov AS, Khromykh AA. 2007.
Using a bioaerosol personal sampler in combination with
real-time PCR analysis for rapid detection of airborne
viruses. Environ Microbiol 9:992–1000.
117. Loeffler J, Henke N, Hebart H, Schmidt D, Hagmeyer L,
Schumacher U, Einsele H. 2000. Quantification of fungal
DNA by using fluorescence resonance energy transfer and
the Light Cycler system. J Clin Microbiol 38:586–590.
118. Hein I, Lehner A, Rieck P, Klein K, Brandl E, Wagner M.
2001. Comparison of different approaches to quantify
Staphylococcus aureus cells by real-time quantitative PCR
and application of this technique for examination of
cheese. Appl Environ Microbiol 67:3122–3126.
119. Higgins JA, Cooper M, Schroeder-Tucker L, Black S,
Miller JD, Karns JS, Manthey E, Breeze R, Perdue ML.
2003. A field investigation of Bacillus anthracis contamina-
tion of US Department of Agriculture and other Washing-
ton D.C. buildings during the anthrax attack of October
2001. Appl Environ Microbiol 69:593–599.
120. Ivnitski D, O’Neil DJ, Gattuso A, Schlicht R, Calidonna
M, Fisher R. 2003. Nucleic acid approaches for detection
and identification of biological warfare and infectious
agents. BioTechniques 35:862–869.
121. Hindson BJ, Brown SB, Marshall GD, McBride MT,
Makarewicz AJ, Gutierrez DM, Wolcott DK, Metz TR,
Madabhushi RS, Dzenitis JM, Colston BW Jr. 2004.
Development of an automated sample preparation module
for environmental monitoring of biowarfare agents. Anal
Chem 76:3492–3497.
122. de Aguirre L, Hurst SF, Choi JS, Shin JH, Hinrikson HP,
Morrison CJ. 2004. Rapid differentiation of Aspergillus spe-
cies from other medically important opportunistic molds
and yeasts by PCR-enzyme immunoassay. J Clin Microbiol
42:3495–3504.
123. Emanuel PA, Fruchey IR, Bailey AM, Dang JL, Niyogi
K, Roos JW, Cullin D, Emanuel DC. 2005. Automated
screening for biological weapons in homeland defense. Bio-
secur Bioterror 3:39–50.
124. Caruana DJ. 2011. Detection and analysis of airborne par-
ticles of biogical origin: present and future. Analyst 136:
4641–4652.
125. Nonnenmann MW, Bextine B, Dowd SE, Gilmore K,
Levin JL. 2010. Culture-independent characterization of
bacteria and fungi in a poultry bioaerosol using pyrose-
quencing: a new approach. J Occup Environ Hyg 7:693–699.
 Fate and Transport of Microorganisms in Air
GARY S. BROWN AND ALAN JEFF MOHR
3.2.4
The fate and transport of microorganisms in the atmosphere
is a complicated issue involving many physical, biological,
and chemical factors. Microorganisms including fungi, bacte-
ria, viruses, and phages are typically transported as bioaerosols
in combination with inorganic particulates and/or moisture
(1, 2). Bioaerosol particles can be either solid or liquid and
can come from a number of natural and anthropogenic sour-
ces. Potential anthropogenic sources include animal, com-
posting, harvesting, and grain storage facilities; carpets,
house plants, operating rooms, saw mills, aeration systems,
and activated sludge facilities; as well as accidental or inten-
tional release of biowarfare agents. The composition, size, and
concentration of the microbial populations comprising the
bioaerosols vary with the source, dispersal mechanism, and
most importantly the environmental conditions prevailing
at a particular source site (3).
The dispersal mechanism of bioaerosols is primarily gov-
erned by hydrodynamic and kinetic factors while fate is
dependent on the specific bioaerosol biological and chem-
ical composition and environmental and meteorological
conditions encountered during dispersal and deposition (4).
A thorough understanding of spatial and temporal atmos-
pheric motion variables (such as wind speed, drag, and tur-
bulence) are required to fully anticipate the transport of
microorganisms in the atmosphere as transport is influenced
by features of large-scale flow fields, geographical location,
and local topography. Atmospheric turbulence, responsible
for aerosol diffusion during transport by the mean wind (that
is, atmospheric motion averaged over a short time period), is
strongly influenced by local atmospheric conditions and the
diurnal variation of solar irradiance reaching the ground.
Additionally, bioaerosol transport is generally down the ther-
mal gradient from regions of warmer air to cooler regions.
The most significant environmental factors influencing
microorganism fate are relative humidity (RH), water con-
tent, solar irradiance, temperature, and oxygen concentra-
tion. The vast majority of airborne microorganisms are
immediately inactivated due to environmental stresses such
as desiccation, temperature, and oxygen concentration that
act to alter the makeup of the outer surface, which in turn
affects cell viability. Additional influences are exerted
through air ions and open air factors (OAFs), such as reac-
tions between ozone and olefins (5, 6), and a combination
doi:10.1128/9781555818821.ch3.2.4
3.2.4-1
of factors including pollutant concentration, relative humid-
ity, and pressure fluctuations.
Many of the physical and chemical processes that describe
more general classifications of aerosols also apply to bioaero-
sols. Classic texts, which should be studied by all individuals
interested in aerosols, include works by Fuchs (7), Hinds (8),
and Hidy (9). Significant sources which address bioaerosols
include texts by Cox (6, 10), Anderson and Cox (11), and
Lighthart and Mohr (12). An additional list of valuable
references is compiled in the Journal of Aerosol Science and
Technology (14[1]: 1–4 [1991]).
FATE OF BIOAEROSOLS
Physical and Chemical Factors Influencing Viability
Studies to determine the factors that influence viability in
the airborne state have been performed for decades. Many
of these studies have yielded varying results, primarily due
to the different methodologies used during the evaluations.
Cox (6) outlines several factors that exert stresses on airborne
microorganisms and lists water content or relative humidity
(RH) as being of primary interest. The effects of temperature
on aerosol stability are less well understood. Marthi (13)
presents his survey of pertinent stresses as being either
“primary” or “secondary” in effect. He lists RH, temperature,
radiation, and OAFs as important primary stressors. For artifi-
cially generated aerosols, he lists secondary stressors that influ-
ence inactivation or protection provided against the primary
stressors. Important secondary stressors include the method
of aerosol generation, makeup of generation fluid, the type
of aerosol sampling method used, and the type of medium
used to enumerate the collected microorganisms. It has
been shown that the type of assay medium can significantly
influence the results for bioaerosol studies (13).
Water Content (RH)
Although water content is related to RH, it does not necessa-
rily describe the same environmental condition. Relative
humidity is expressed as a percentage and describes the amount
of water vapor contained in a volume of air at a specific temper-
ature relative to the maximum amount of water vapor that air
can hold at that temperature. At higher temperatures, air can
3.2.4-2 ▪ AIR
hold much more water vapor than at lower temperatures, thus
any discussion of the relationship of RH to microbiological
viability must also include temperature. Absolute humidity,
which is independent of temperature and is an actual measure
of water content, may be the environmental variable of most
import and should be used in future studies of bioaerosol–water
content relationships.
The state of water and the water content associated with
bioaerosols are fundamental factors influencing the fate or
viability of these microorganisms. As the atmospheric water
content decreases, so does the water available to the exterior
environment of the microorganism. Loss of water can cause
desiccation, resulting in inactivation of many microorgan-
isms. Of all of the measurable meteorological parameters,
RH and the associated water vapor content are the most
important with respect to aerosol stability (6). Israeli et al.
(14) concluded that the mechanism of cell death in bioaero-
sols was due to conformation changes in the biomembrane
phospholipid bilayers from crystalline to gel phases as a result
of water loss affecting cellular macromolecules, such as pro-
teins and nucleic acids. However, the effect of water vapor
content on airborne microorganisms as a function of RH is
not consistent across all classes of microorganisms or even
within classes.
Of the bacteria, some studies on airborne Gram-negative
bacteria indicate an increased death rate at intermediate to
high relative humidity of 50–90% (15, 16). Won and Ross
(16) surmised that an increased death rate observed in for Pas-
teurella pestis was possibly due to continuing metabolic activ-
ity and altered cell membrane permeability at elevated RH,
which allows lethal material, found in the free atmosphere,
to enter the cell and cause death. However, other studies
demonstrate relative stability of Gram-negative bacteria at
intermediate to high RH (17, 18) as expected. Increased
cell death rate for Gram-positive airborne bacteria at inter-
mediate RH has also been observed for Staphylococcus, Strep-
tococcus, and Bacillus species (15, 16).
Water content conditions at dissemination and aeroso-
lization may also impact survival rates for bacteria. Cox
(19) found that dry-aerosolized organisms tended to absorb
water from the environment and partially rehydrate, and
wet-aerosolized organisms tended to lose water to the envi-
ronment and desiccate. The differential aerosolized water
content affected final survival such that wet-aerosolized Pas-
teurella tularensis maximum death occurred at 50–55% RH
and dry-aerosolized Pasteurella maximum death occurred at
75% RH (20).
Some generalizations can be made concerning virus stabil-
ity and survivability as related to RH or water content. Typ-
ically, viruses with higher lipid content tend to be more
persistent at lower relative humidity, and those with lower
lipid content are more stable at higher relative humidity
(21). The non–lipid enveloped viruses tend to survive longer
at RH in the range of 70–90% (19, 22, 23) and viruses with
lipid envelopes tend to survive longer at 20–30% RH (24–
26). Minimal survival for both enveloped and nonenveloped
viruses appears to generally occur at an intermediate RH
range of 40–70% (23).
Not only is the water content of aerosolized microorgan-
isms a principal factor contributing to aerosol viability (15,
27, 28), but the RH of a system will directly affect the density
of the bioaerosol unit. The size, shape, and density of an aero-
solized particle are directly related to the aerodynamic diam-
eter, which determines settling velocity and location of
deposition in the respiratory tract and will influence the col-
lection efficiency of aerosol samplers.
The effects of RH can be influenced by the content of the
suspension fluid used before aerosolization (29, 30), the con-
tent of the collection fluid (31), and prehumidification (32).
For some microorganisms, shifts in RH after aerosolization
have a more profound effect on aerosol stability than does
constant RH (33).
Temperature
Studies to determine the effect of temperature on bacterial
stability have generally shown that increases in temperature
tend to decrease the viability of airborne microorganisms
(34). A review of previous studies indicates airborne bacteri-
al survival decreases at temperatures above of 24°C for
Gram-negative and Gram-positive organisms (35–41). Addi-
tionally, frozen cells tend to lose cellular proteins that result in
elevated aerosol inactivation rates.
Temperature is considered the major factor determining
virus inactivation in the environment. In general, virus sur-
vival as an aerosol varies inversely with temperature, but
the overall effect of temperature depends on virus type and
RH (42). Viruses containing DNA are more stable than
RNA viruses over a similar temperature range, but high tem-
peratures eventually affect viral DNA integrity as well. Virus
inactivation is faster at temperatures ≥50°C than at temper-
atures <50°C, but there was also a significant temperature-
matrix effect (43). Virus inactivation appears to occur faster
in complex matrices rather than simple matrices (43).
In general, bacteriophages appear to be highly persistent
whatever the matrix or the temperature, which makes them
useful indicators for virus inactivation studies (43).
Oxygen Concentration
Oxygen concentration can have a pronounced effect on the
aerosol stability and infectivity of some bacteria through
inactivation by free radicals of oxygen (44–46). Cox (45)
has observed an inverse relationship between oxygen con-
centration and viability. Israeli et al. (14) note that oxygen
susceptibility increases with dehydration, increased oxygen
concentration, and time of exposure.
Viruses tend to be more resistant to the effects of oxygen
toxicity than are bacteria because structurally and biochemi-
cally they are far less complex (47). Benbough (48) also found
that atmospheric oxygen had no effect on aerosolized viruses
and Webb (49) determined oxygen concentration had little
effect on survival of bacteriophages in aerosols.
Other Factors
Solar Irradiance
Bioaerosol inactivation caused by solar irradiance has been
shown to be dependent on both wavelength and intensity
of the radiation. Shorter wavelengths contain more energy
and are generally more deleterious to aerosolized microorgan-
isms. The mechanism of solar radiation–induced cell death
depends on the presence of sensitizers and chromophores,
which absorb the radiation, resulting directly in cell damage
particularly in DNA or indirectly through the production of
superoxides (50, 51). Bellair et al. (52) observed a diurnal var-
iation in fecal coliform die-off rate and determined an inverse
relationship between bacterial number and solar radiation.
Other studies have shown that samplers monitoring bacteria
from airborne emissions at sewage plants consistently yielded
greater numbers at night (53). Relative humidity (11), water
activity (54, 55), oxygen concentration (54), age of the aero-
sol (6), and the presence of other gases all influence the effect
radiation exerts on airborne microorganisms. The targets of
 3.2.4. Fate and Transport of Microorganisms in Air ▪ 3.2.4-3
radiation induced inactivation appear to be phospholipids,
proteins, and nucleic acids.
OAF
Open air has bacterial, viral, and phage inactivation proper-
ties attributed to OAFs (1, 5, 56–59). The term OAF was
introduced in studies to explain the observed significantly
lower survival of many biological aerosols in open outdoor
air compared with filtered, clean, inert laboratory-supplied
air (1, 59, 60). Whereas Cox (6) suggests that the primary rea-
son for inactivation is caused by reactions between ozone and
olefins, bacterial OAF inactivation is probably caused by a
combination of factors including pollutant concentration,
RH, pressure fluctuations, and air ions (1, 5, 6, 59, 60). The
targets of OAF-induced inactivation appear to be phospholi-
pids, proteins, and nucleic acids.
Microbial Factors
The type, species, or strain of an organism will affect its air-
borne survival. Bacteria certainly differ in their responses to
all types of stresses. The atmospheric environment is hostile
to all microorganisms. No species has evolved to fill a niche
in the open atmosphere. However, there is some evidence
suggesting that bacteria can form new cells in the airborne
state (61). Gram-negative bacteria contain more phospho-
lipids than Gram-positive bacteria, resulting in greater sus-
ceptibility to inactivation through Maillard reactions,
dehydration, and osmotic shock. Bacteria grown on enriched
media or when collected during different phases of develop-
ment display widely varying aerosol stability rates (62, 63).
Perhaps cells grown during the log phase are more active met-
abolically and may be more resistant to stresses than those in
the stationary phase of growth. The physiological age of a bac-
terial population will also influence its stability as old cells die
off at higher rates than newly grown populations. The type of
culture media used for enumeration of bacteria also influences
the quantitative results from aerosol stability studies. Some
media provide constituents that others may not and enable
stressed microorganisms to survive.
Miscellaneous
Review articles by Cox (6), Strange and Cox (64), and
Spendlove and Fannin (65) address the influences that pollu-
tants, pressure fluctuations, air ions (66), and season may
have on aerosol inactivation rates. Some atmospheric pollu-
tants that have been studied include NO2, SO2, and O3.
The deleterious effects of these pollutants are greatly influ-
enced by RH. Some pollutants may react with water to
form acids (67), resulting in inactivation of aerosolized micro-
organisms. Other pollutants have been studied, including
HCOH, CO, HCl, HF, C2H2, C2H4, and C2H8 (68, 69),
which in the concentrations studied appear to produce
much less bacterial inactivation than what is observed with
typical outdoor OAF.
The effect of electrical charges on the viability of airborne
bacteria was recently studied by Mainelis et al. (70). They
found that Pseudomonas fluorescens viability declined rapidly
as a function of the net charge. Viability between 40% and
60% was observed in cells carrying a net charge from 4,100
negative to 30 positive elementary charges compared with
less than 1.5% viability when the cells possessed >2,700 pos-
itive charges. Changes occurring in the membrane potential
during aerosolization appear to influence the ultimate viabil-
ity of the bioaerosols and may also explain why recovery effi-
ciencies of bioaerosols can vary significantly, depending on
the local atmospheric conditions (3).
A review (71) on the diurnal and annual variations of bio-
aerosol concentrations details some interesting phenomena.
For a high desert climate, the upward flux of bacteria was
shown to follow the solar irradiance/sensible heat cycle
(72) with the maximum occurring at solar noon. It was also re-
ported that even though the upward bacterial flux decreased
after solar noon, the bacterial concentration increased due to
a reduction in solar output, a function of solar angle, which
resulted in lower rates of bacterial inactivation. Another study
(73) performed over a grass seed field showed lower numbers
of bacteria associated with the upward flux and that variations
in airborne concentrations were linked to the local diurnal
sea breezes. Annual variations in bioaerosol concentrations
have been shown to exhibit minimums in the winter and
a primary peak in the late spring (71) with a secondary
peak observed during midsummer. These maximums and
minimums are related to favorable growth conditions, such
as moderate temperatures, humidity, and solar irradiance,
respectively.
Microbial Aerosol Stability
The experimental determination of aerosol stability is
dependent on the ability to assay an entity for biological
activity. Many sensitive immunological procedures are avail-
able that take advantage of the specificity of antigen/antibody
reactions. The problem with these assays is that they reveal
very little about viability of the microorganism in question.
Spendlove and Fannin (65) define virus aerosol survival stat-
ing that stability in the aerosol particle is really a measure of
the ability of the virus to infect tissue culture cells of living
hosts. For this reason, infectivity and stability can be used
interchangeably for collected viral bioaerosols.
Data are presented in this section relating aerosol stability
or infectivity studies with bacteria, animal viruses, bacterial
viruses ( phages), and other microorganisms. Cox (6) and
Mohr (59) discuss in greater detail individual bacteria and
viruses and provide a rationale for the demonstrated stability.
The results for aerosol stability tests performed on the same
microorganism often differ because testing techniques have
not been standardized. Harper (74) noted that variations in
results may be caused by:
1. Method of aerosol generation, storage, and sampling
procedure,
2. Method of assay,
3. Differentiation of total, physical, and viability decay,
4. Presence or absence of light,
5. Methods and extent to which relative humidity and tem-
perature are controlled,
6. Growth stage,
7. Culture media, and
8. Method of data presentation.
In addition, variations can occur because of choice of sus-
pending fluid, collection fluid, and local constituents in the
atmosphere, such as presence or absence of oxygen and other
gases.
Bacteria
Review papers addressing the aerosol characteristics of bacte-
ria are available (6, 11, 12, 64, 75, 76). More variation in
aerosol test results has been observed for bacteria than viruses,
phages, and other microorganisms. Table 1 shows some of the
parameters that have been shown to influence the viability of
selected bacteria. This is due not only to differences in test
3.2.4-4 ▪ AIR

TABLE 1 Aerosol stability parameters for selected bacteria

Bacteria Stability parameters Reference
Bacillus subtilis Death rate (15, 27)
B. prodigious Ambient temperature, bacterial viability (15, 77)
B. patchiness Ambient temperature, bacterial viability (15, 27)
B. violaceous Ambient temperature, bacterial viability (15, 27)
Chlamydia pneumoniae RH, temperature (41)
Cytophaga allerginae RH, temperature (78)
Enterobacter ccoacae RH, temperature, CO2 (79–82)
Erwinia herbicola RH, temperature (29, 81, 83)
Escherichia coli RH, temperature, O2, wet, dry (45, 84–89)
Klebsiella planticola RH, temperature (82)
Legionella pneumonphilia RH (63)
Mycoplasma RH (90)
M. pneumoniae RH, temperature, solar (40, 90–93)
Pasteurella tularensis RH, wet and dry generation, solar (20, 31, 94)
Serratia marcescens RH, O2, freeze, time (33, 44, 46, 68, 95)
Staphylococcus albus Time, Ambient temperature (15, 27)
S. aureus RH, temperature (96)

procedures and data presentation but also to the greater struc-
tural and metabolic complexity of bacteria, such as cell walls,
membranes, and metabolism. It has been observed (63) for
some bacteria that the growth phase and type of media used
for production play major roles in the demonstrated aerosol
stability of the bacteria. Most of the published data available
cannot be used for comparing aerosol stability because of the
presence of several stresses that could have acted synergisti-
cally in a detrimental manner or possibly enhanced survival
in some unknown way. Hess (95) performed an elaborate ser-
ies of studies with Serratia marcescens and showed that loss of
viability was caused by desiccation and the presence of oxy-
gen. Some generalities concerning bacterial aerosol stability
can be made, including (1) loss of aerosol viability may be
caused by desiccation and oxygen toxicity, (2) many bacteria
have complicated RH inactivation profiles in inert atmos-
pheres, and (3) survival can be greatly increased by the addi-
tion of sugars that act as osmoprotectants. Additionally,
aerosolized spores of many bacteria are extremely resistant
to inactivation by solar radiation, oxygen concentration,
RH, and temperature (57).
Viruses
Several review papers addressing the aerosol characteristics of
viruses are available (6, 11, 66, 97). Table 2 shows some of the
parameters that have been shown to influence the viability of
selected viruses. Viruses are normally very resistant to inacti-
vation by oxygen. This and the relative simplicity of their
structure explain why results of aerosol inactivation studies
are more consistent for viruses than for bacteria.
Phages
Phages, like other viruses, are not inactivated by oxygen.
Table 3 shows some of the parameters that are associated
with the viability of selected phages. Most of the work has
been performed on the T series of coliphages, the phages
that infect Escherichia coli. Generally, it has been determined
that NaCl is toxic to phages, but the effect can be reversed
by addition of peptone to the suspending media. Aerosol
inactivation of complex phages has been shown to be caused
by the breakup of the head-tail complex (123, 127). Aerosol
stability is high for T phages above 75% RH, but at below
75% RH the lack of water vapor is sufficient to weaken the
head-to-tail bonds and promote breakage and inactivation
of the phages (32, 84, 126). Inactivation may also occur
with rigorous bioaerosol sampling, which has been shown
to break up the head-to-tail complex of some phages, but pre-
wetting of the sampling surface with a device such as a
humidifier bulb has been shown to reduce sampling inactiva-
tion (32, 126).
INACTIVATION MECHANISMS
Bioaerosols are subject to inactivation and transport the
moment they become airborne. Desiccation of the droplet
is the main factor responsible for inactivation in liquid bioaer-
osols. Particle sizes for these droplets are usually 2–10 μm and
they tend to follow the streamlines of the local wind. For par-
ticles larger than 5.0 µm, gravitational settling and impaction
are the leading causes of loss of particles during transport.
Some relationships between aerosol stability and the bio-
logical composition of microorganisms have been identified
(14, 129). Bacterial aerosol stability is considerably more
complex than that observed for viruses. Cox (85–87) set up
an elaborate series of stability experiments to examine the fac-
tors that influenced inactivation of E. coli B. When E. coli B
was aerosolized from a suspension of distilled water into highly
purified nitrogen, argon, or helium atmospheres, survival was
virtually complete at low RH but was critically dependent on
RH above values of 80%. Under these conditions, Cox (6)
speculates that the gases slightly modified the water structure
through gas hydrate formation and water lattice modification,
which affected the stability of biological structures. Cox (6)
then explained the events that may take place as precursors
to aerosol inactivation. As E. coli is aerosolized into inert
atmospheres at mid to high RH, the biological membrane
constituents become destabilized through loss of water mole-
cules. Additives, such as polyhydroxyl compounds, which
supersaturate, can stabilize these structures. Polyhydroxyl
compounds, by binding to sites on proteins, cause conforma-
tional changes and thereby stabilize proteins, making them
 3.2.4. Fate and Transport of Microorganisms in Air ▪ 3.2.4-5

TABLE 2 Aerosol stability parameters for selected viruses

Virus Stability parameters Reference
Adenovirus RH (98)
Bovine rhinotraheitis RH, temperature, media (99, 100)
Columbia SK virus RH, temperature (101)
Coxsackie UV light (102)
Encephalomyocarditis (ENC) RH, rehumidification infectivity (103, 104)
Enteric viruses RH (83)
Foot & mouth virus Radiation, RH, temperature, weather factor (29, 105–107)
Influenza, virus RH (108)
Langat virus A, B RH, temperature (48)
Newcastle disease RH, temperature (100)
Picornavirus 37A RH (34, 109)
Pigeon pox RH, inositol (110)
Pono virus Rh, temperature, environmental, (24, 26, 48, 74, 103, 111, 112)
seasonal inactivation
REO RH, temperature (97, 113)
Rotavirus RH, temperature (26, 114–116)
Rouse sarcoma RH, inositol (110)
Semliki forest virus RH (48, 117)
Simian virus 40 RH, temperature (109)
St. Louis encephalitis RH temperature (118)
Venezuelan equine encephalomy RH, temperature, solar (67, 119)
Vesicular somatitis RH, temperature, O3 (100, 120)
Yellow fever RH, temperature (121)
General RH, drying (122)

less susceptible to denaturation. This is convincing evidence
that the state of proteins on the outer membrane of some
microorganisms is critical to the resultant stability profile.
Cox (6) also explains that Maillard reactions between
reducing sugars and amino acids are responsible for some
denaturation and unfolding of the ovalbumin α-helix, which
occurs during drying. Additionally, he explained that dried
proteins might be stabilized by similar reactions between other
sugars and sodium glutamate. Soddu and Vieth (130) showed
that sucrose could bind to collagen membranes with different
affinities and produce conformational changes in the protein
structure. Cox’s (6) explanation of the sequence of events
is as follows. During the desiccation process, polyhydroxyl
compounds and amino acids react together causing confor-
mational changes that strengthen the overall protein struc-
ture. The presence of sugar additives causes conformational
changes in the coat proteins, and the new-configuration
coat proteins do not react or react more slowly with the poly-
hydroxyl coat moieties. In the absence of these sugar additives
and free molecules, the coat proteins may react irreversibly,
through Maillard reactions, with polyhydroxyl coat moieties
and cause loss of viability. In addition, the sugar additives
could compete with the polyhydroxyl coat moieties for the
reaction sites of the coat proteins or physically hinder those
reactions’ molecular collisions. The result in each case would
be more aerostable microorganisms. When in their more nor-
mal aqueous environment, the Maillard reactions leading to
this inactivation may not occur because the reaction sites
are separated either by bulk water molecules or by water mol-
ecules bound at the reactive sites. Removal of these water
molecules by evaporation leads to the events just proposed.
A possible mechanism for aerosol inactivation has been
presented here by inference from mostly unrelated work.
There have been many studies on solute concentration
effects, which may also be applied to help model the molec-
ular events that take place during inactivation. In a practical
sense, the process may be viewed as the increase of solute con-
centration, which concomitantly occurs during drying of the
aerosol droplet.
Freeze-drying, freezing, and aerosolization of micro-
organisms all act to remove water from the system. Elaborate
studies have been performed using freeze-dried micro-
organisms as models to study the processes responsible for
inactivation. Israeli et al. (14) studied freeze-dried micro-
organisms and showed the importance of water content to
the viability of microorganisms. They concluded that some
areas of the biomembrane, specifically the phospholipid
bilayers, go through conformational changes from crystalline
to gel phases as a result of water loss. These factors caused
damage to cell proteins, which in turn resulted in a loss of
viability.
BIOAEROSOL STABILIZERS
Several compounds have been shown to provide protection to
aerosolized microorganisms. Generation fluid additives, such
as inositol and bovine serum albumin, have been shown to
increase the recovery of some bacteria (30). Some polyhydric
compounds such as raffinose, dextran, glycerol, glutamate,
and inositol have been shown to increase recovery of bacte-
ria, but the results are dependent on the RH of the system
(30, 110, 129). Cox (6) hypothesizes that these compounds
bind with available proteins and stabilize them against dena-
turation. Marthi and Lighthart (79) added betaine to the
3.2.4-6 ▪ AIR

TABLE 3 Aerosol stability parameters for selected phages

Phages Stability parameter Reference
T series coliphages RH (84)
T1 Air ions (123, 124)
T2, T3, T7 Mechanism of inactivation (48)
T3, T7 Freeze (125)
T3 RH, sample prehumidification (126)
Phages Sample prehumidification (32)
MS 2 Prehumidification, RH, inactivation, temperature (32, 118, 123, 127)
Pasteurella pestis phage Prehumidification (126)
S13 Prehumidification, RH, inactivation (32, 123, 127)
φX174 RH, chemicals (69, 128)

generating and collection fluid during field studies collecting
airborne bacteria and observed increases from 21.6% to
61.3% when compared with samples taken in the absence
of betaine. The recovery was dependent on the concentration
of betaine, but they found the optimum concentration to be
between 2 and 5 mM. Betaine acts as an osmoprotectant.
Additionally, trehalose has been shown to afford protection
to both freeze-dried (14) and aerosolized bacteria. Trehalose
is known to stabilize lipids, proteins, and phospholipids,
important constituents of microorganisms.
REPAIR MECHANISMS AND RESUSCITATION
Microorganisms in the environment are constantly subjected
to forces that cause inactivation. The consequences of these
forces are not always lethal. Marthi (13) and Cox (6) present
interesting details involved with sublethal damage to micro-
organisms. Marthi (13) states that perhaps the most impor-
tant effect of sublethal stress is the inability of stressed
microorganisms to grow on different media, both selective
and nonselective. The fact that microorganisms are present
but cannot be detected by standard enumeration procedures
will require additional planning by the investigator to con-
firm that the reported results are representative of the sample
that has been collected. Cox (6) points out that although the
microorganisms are stressed and not detectable by normal
methods, they may be able to recover from the injury and ini-
tiate disease. Marthi (13) points out that the primary sites of
initial stress include the outer membrane, cell wall, cytoplas-
mic membrane, RNA and ribosomes, and DNA. When inju-
ries are inflicted on the bacterial cell constituents, the
metabolic activities of microorganisms are concomitantly
influenced. Cox (6) addresses repair mechanisms associated
with surface structures, transport functions, and radiation
damage in his review article.
VIABILITY MODELS
Exponential Decay Model
The loss of viability of aerosolized microorganisms is caused
by complex physical, meteorological, and cellular interac-
tions. Early attempts to explain aerosol viability relied on
the exponential decay model with mixed results. Even
though exponential decay has been shown to be a simplifica-
tion, application of intricate mathematical expressions often
delivered conclusions that were no more accurate than simple
expressions. If all of the known parameters were applied to a
mathematical expression there would be around 20 inputs,
many of which would depend on other parameters during spe-
cific periods of flight. The expression for exponential decay is:
Vt ¼ V0 eK (1)
where,
Vt = viability at time t,
V0 = viability at time zero, and
K = decay rate constant.
Kinetic Model
Cox, in a series of articles (6), presents what he calls the
kinetic model. He was dissatisfied with the lack of explana-
tion of the exponential model, which did not account for
the time-dependent decay observed for most microorganisms.
Cox explains his kinetic model, which supposes that micro-
organisms contain a molecular species B(n)H2O, the bio-
logical activity of which is essential for a microbial cell to
replicate or be infectious. B(n)H2O, when exposed to an
environment of lowered water activity, forms a series of
hydrates similar to other biomolecules; some of these hydrates
are unstable and spontaneously denature through a first-order
process, that is,
dx=dt ¼ kx (2)
where x is the concentration of the species that denatures.
The model form is then
kþ
BðnÞH2 O $ Bðn  xÞH2 O þ xH2 O $
k
Bðn  x  yÞH2 O þ yH2 O $ B þ iH2 O (3)
where B(n − x)H2O is the denatured form with a rate con-
stant kx and B(n − x − y)H2O is the denatured form with
a rate constant ky.
Cox (6) then applies probability theory to evaluate the
likelihood of death, which is related to percent viability and
sets up appropriate boundary conditions and integrates the
results. When denaturation follows a first-order pattern the
final form of the equation is written:


ln V ¼ K1 ½Bðn  xÞH2 O0 ekt  1 þ ln 100 (4)
where k is a first-order denaturation constant, t is time, K1 is
the probability constant, and V is viability. Cox (6) has ana-
lyzed several hundred viability-time curves and has found
very good agreement with experimental results for dehydra-
tion inactivation.
 3.2.4. Fate and Transport of Microorganisms in Air ▪ 3.2.4-7
Catastrophe Model
Catastrophe theory was formulated to explain the nature of
discontinuous events, such as the breaking of waves on a
beach, the crash of the stock market, or sudden aggression dis-
played by an animal. Overall reactions involving large num-
bers of molecules appear to behave in a continuous fashion
because discontinuities are smoothed, but on an individual
level, the reactions are discontinuous. If the number of reac-
tions is relatively low, as is the case when aerosol viability is
being considered, discontinuity is appropriately represented
and aerosol inactivation rates may be predicted. On an indi-
vidual level, the loss of infectivity of an airborne microorgan-
ism is a sudden discontinuous event, and this change in state
is termed the catastrophe. The bases of catastrophe theory are
related to the potential energy and therefore equilibrium of
the system. The potential energy is a function of what are
termed control parameters, which govern the equilibrium of
an event. Within a certain range, equilibrium will not be
influenced by variations in the control parameters, but
some small critical change can cause a shift in the potential
energy and result in inactivation. Cox (6) combined catastro-
phe theory and kinetics to explain loss of viability caused by
desiccation, temperature, oxygen, and OAF. The calculated
viability curves agreed very well for inactivation caused by
oxygen concentration and OAF and were more accurate
than predictions based on probability theory for denaturation
induced by desiccation.
Dispersion Models
Models have been developed to predict dispersion and depo-
sition of microbial aerosols with respect to penetration of
structures (131), predict infectious microbial concentrations
downwind from known sources (91, 132–136), and evaluate
the spread of plant pathogens outdoors (137). Many of the
early models were based on the treatments of atmospheric dif-
fusion, which were created to predict the fate of air pollutants.
The inert particle dispersion model by Pasquill (138) is
empirical and based on observations of the deposition of
inanimate particles. Results from its application yield average
distributions of airborne microorganisms, have a somewhat
limited downwind range, and are applied to flat terrain with
steady wind conditions. Assumptions that are made for this
model include (1) Gaussian distribution of the plume in
the horizontal and vertical plane; (2) total reflection of
particles from the ground; (3) particles emitted from the
source at a constant rate; (4) wind velocity and direction
are constant; (5) flat terrain; (6) particles smaller than 20
μm, making gravitational effects negligible; and (7) diffusion
downwind is negligible as is the difference in wind velocity
between the source and real wind. The classical form of the
inert particle dispersion model is:
Xðx,y,z:HÞ ¼ Q=½2pðsy sz Þu  EXP½0:5ðy=sy Þ2 
 fEXP½0:5ððZ  HÞ=sy Þ2 
þ ½0:5ððZ þ HÞ=sz Þ2 g (5)
where
X = the number of particles per cubic meter at downwind
location x,y,z.
Q = the number of particles emitted from the source per
second
u = the mean air speed in m/s
σy = the standard deviation of the horizontal concentration
(at the downwind distance)
σz = the standard deviation of the vertical concentration (at
the downwind distance)
H = the height of the source plus the plume rise.
The equation can be simplified for ground level (z = 0)
and becomes:
Xðx,y,0:HÞ ¼ Q=psy sz u EXP
½ðy2 =2sy 2 þ H2 =2sz 2 Þ (6)
For a ground-level source (H = 0) and analysis along the
center line ( y = 0) the equation becomes:
Xðx,y,z:HÞ ¼ Q=psy sz u (7)
As explained earlier, these equations are for inanimate
particles and do not take into account microbial viability
decay. Lighthart and Frisch (139) expanded on the inert par-
ticle diffusion model by adding a biological death constant
(α) and creating a graphic method that estimates ground-
level concentrations when microbial death rate, mean wind
speed, atmospheric stability class, source height, and down-
wind sampled distance are known. Microbial death rate is
determined under laboratory conditions and applied to the
model. The equation for biological death (BD), after modifi-
cation becomes:
Xðx,y,z:HÞBD ¼ Xðx,y,z:HÞEXPðatÞ (8)
where t is approximated by x/u.
Biological decay and the rate constants associated with
inactivation must be determined under dynamic conditions
in the laboratory. Teltsch et al. (135) presented methodology
to estimate the death constant under field conditions to deter-
mine inactivation and final concentrations downwind from
sprinklers using wastewater. Death constants must be eval-
uated over a wide range of environmental conditions for accu-
rate results with the determined decay constants not to be
applied for predictive use because of the study-specific mete-
orological conditions.
Lighthart and Mohr (140) applied best-fit laboratory-
generated decay constants to the Gaussian plume model using
microbial source strength and local hourly mean weather data
to drive the model through a summer and winter day cycle.
For near-source locations, higher wind speeds or short travel
times exerted a major modulating effect during the day
because time was inadequate for inactivation. Additionally,
as travel time increased, because of low wind speed or long dis-
tances, modulation was influenced more by solar irradiance,
RH, and temperature than by wind speed.
Spendlove (131) applied various models—box and test
tube—to determine the penetration of structures by microbial
aerosols. Models showed that for a single-story dwelling with-
out air conditioning, the aerosol dose received was the same
inside as outside and that air conditioning units would act
to moderately decrease inside concentrations depending on
ventilation rates.
Lighthart and Kim (141) have presented a simulation
model that describes the dispersion of individual droplets
of water containing viable microbes. By repeating the
modeling process many times for individual droplets, an
aerosol cloud can be simulated. The model accounts for the
physical, chemical, biological, and measured meteorological
parameters of each droplet for many time increments. The
droplet model is separated into five submodels, which include
aerosol generation, evaporation, dispersion, deposition, and
3.2.4-8 ▪ AIR
microbial death, all of which are calculated at chosen time
intervals in the trajectory for each drop. The results show
that evaporation is an important factor in determining depo-
sition sites because of particle size dependence, chemical
reactions, and protection offered by large droplets. Wind
gust data are required because average wind velocity data
tends to “smooth” and oversimplify what is occurring on a
micrometeorological scale.
A variety of additional dispersion models have been devel-
oped to deal with short- and long-range bioaerosol transport,
including Lagrangian, computational fluid dynamic, and
modified Gaussian plume and puff models (142–144).
TRANSPORT OF BIOAEROSOLS
Transport of bioaerosols and aerosols in general is usually
described in terms of the processes involved in the movement
of particles from one surface or location to another: adhesion,
release, and deposition.
Adhesion
The primary adhesive forces are van der Waals force, electro-
static force, and the surface tension of adsorbed liquid films.
These forces are affected by:
1. particle composition, shape, size, roughness, and
contamination;
2. temperature;
3. relative humidity;
4. duration of contact; and
5. initial contact velocity (8).
Wickman (129) describes adhesive forces as being domi-
nated by the molecular structure and organization near the
contact surfaces. The two electrodynamic forces, London–
van der Waals and electrostatic, are primarily responsible
for at-a-distance interactions occurring between aerosol par-
ticles and surfaces. If the surface is wet, surface tension will
tend to retain the particle (145). Adhesion is also governed
by the atomic attraction between surfaces that is, Hamaker
constants and electronic frequency and geometrical factors
such as surface shape and configuration.
Release
Bioaerosols are released into the atmosphere naturally by
mechanical action, electrodynamic action, or biochemical
processes. Mechanical action release may be initiated by
aerodynamic drag, similar electric charge, impact of other
particles, or inertia of rest in the event of sudden substrate
movement. Electrodynamic action may take the form of
repulsion by similar electric charge between substrate and par-
ticle. Natural mechanical action may be generated by wind,
rain, or breaking water. Bioaerosols in the form of pollen and
spores with possible bacterial or viral organisms attached are
released to the atmosphere through metabolic release proc-
esses. It must be noted that as the size of the particle decreases,
it becomes increasingly difficult to remove them from a sur-
face, because adhesive forces are proportional to the particle
diameter and removal forces are proportional to second and
third powers of the particle diameter (8). Bioaerosols are
also released to the atmosphere through anthropogenic activ-
ities associated with soil disturbance and spraying in indus-
tries such as animal husbandry, composting, harvesting and
grain storage facilities, saw mills, aeration systems, and acti-
vated sludge facilities. Additionally, biowarfare agents in
the form of bioaerosols may also be released to the atmosphere
either accidentally through negligence or intentionally as an
act of terror.
Deposition
Deposition is the process by which aerosol particles collect
on solid surfaces and is accomplished through dry or wet
processes.
Dry Deposition
There are five basic mechanisms by which aerosol particles
are dry deposited: (1) interception, (2)inertial impaction,
(3) diffusion or Brownian motion, (4) gravitational settling,
and (5) electrostatic attraction (8). Deposition by intercep-
tion occurs when a particle follows a gas streamline that passes
within one particle radius of the surface. Due to its size and
proximity within one radius, the particle touches the surface,
resulting in capture. For a given particle size there are unique
streamlines that result in capture. Interception is the only
mechanism in which the particle does not leave the gas
streamline until capture. Inertial impaction occurs when a
particle, due to inertia, is unable to adjust to changing gas
streamlines near a surface, crosses streamlines, and is captured
as it hits the surface. Three primary factors affect the probabil-
ity of inertial impaction (1) particle mass, (2) aerodynamic
particle size, and (3) velocity differential between the particle
and surface. Diffusion or Brownian motion provides an
increased probability of small particles colliding with and
being captured by a surface while moving past on a noninter-
cepting streamline. Deposition by gravitational settling
occurs when the force of gravity is greater than particle air
resistance or drag, and the particles fall to the surface. Electro-
static attraction deposition occurs as charged particles are
attracted to an oppositely charged surface by coulombic
attraction. Charged particles may also be attracted to an
uncharged surface at close range by image forces in which
charged particles induce equal and opposite charges in the
surface, creating an attraction field (8).
Wet Deposition
In wet deposition, atmospheric hydrometeors, such as rain,
snow, fog, or clouds, formed from the condensation of water
vapor in the atmosphere, scavenge aerosol particles. This
means that wet deposition is basically gravitational settling
with components of diffusion and turbulent coagulation
with water droplets or snow. There are three basic mecha-
nisms by which aerosol particles are scavenged from the
atmosphere and wet deposited. Below-cloud or precipitation
scavenging occurs as falling raindrops or snowflakes collide
with and capture suspended aerosol particles. In-cloud scav-
enging occurs when atmospheric aerosol particles convec-
tively moving among clouds collide with and are captured
by suspended water droplets, such as inside a fog. Nucleation
scavenging occurs when small aerosol particles function as
cloud condensation nuclei and are carried with the water
droplets as they grow larger and ultimately fall as rain.
CONCLUSION
The study of bioaerosols has recently gained more attention
due to the recognized importance of indoor and outdoor
events that have been shown to affect applicable populations.
Significant additions to the mechanisms governing the fate
and transport of bioaerosols have been identified. Hydrody-
namic and physical factors have been studied in increasing
detail so that the biochemical changes that are responsible
for the inactivation of aerosolized microorganisms can be
 3.2.4. Fate and Transport of Microorganisms in Air ▪ 3.2.4-9
defined. Methods have been developed to study the mecha-
nisms involved with the deposition, adhesion, and release
of biological particles. Studies have indicated a high degree
of variability in survival capability for microorganisms spe-
cific to environmental conditions and species. The realiza-
tion and application of standards for generating, storing,
sampling, and presenting data will allow better comparison
and reproducibility of results. Before a bioaerosol project is
initiated, a substantial amount of study must be completed
to determine methods to maximize the recovery of aerosolized
microorganisms. These studies should include effects con-
cerning sampling media, counting media, generation media,
and the makeup of test atmospheres.
REFERENCES
1. Cox CS, Wathes CM. 1995. Bioaerosols handbook. Lewis
Publishers, Boca Raton, FL.
2. Maier RM, Pepper IL, Gerba CP. 2000. Environmental
microbiology. Academic Press, San Diego, CA.
3. Pillai SD. 2007. Bioaerosolsfrom land-applied biosolids:
issues and needs. Water Environ Res 79:270–278.
4. Brooks JP, Gerba CP, Pepper IL. 2004. Aerosol emission,
fate and transport from municipal and animal wastes.
J Residual Sci Technol 1:13–25.
5. de Mik G, de Groot I. 1977. The germicidal effect of the
open air in different parts of the Netherlands. J Hyg 78:
175–180.
6. Cox CS. 1987. The aerobiological pathway of microorganisms.
John Wiley & Sons, New York.
7. Fuchs NA. 1964. The mechanics of aerosols. Dover Publica-
tions, New York.
8. Hinds WC. 2003. Aerosol technology: properties, behavior,
and measurement of airborne particles. John Wiley & Sons,
New York.
9. Hidy GM. 1984. Aerosols. Academic Press, New York.
10. Cox CS, Wathes CM. 1995. Bioaerosols in the environ-
ment. In Cox CS, Wathes CM (eds), Bioaerosols handbook.
CRC Press, Boca Raton, FL.
11. Anderson JD, Cox CS. 1976. Microbial survival,
p 203–226. In Gray TR, Postgate JR (eds), The survival
of vegetative microbes. Cambridge University Press,
Cambridge.
12. Lighthart B, Mohr AJ. 1994. Atmospheric microbial aero-
sols. Theory and applications. Chapman & Hall, New York.
13. Marthi B. 1994. Resuscitation of microbial bioaerosols, p
192–225. In Lighthart B, Mohr AJ (eds), Atmospheric
microbial aerosols. Chapman & Hall, New York.
14. Israeli E, Gitelman J, Lighthart B. 1994. Death mecha-
nisms in bioaerosols, p 166–191. In Lighthart B, Mohr AJ
(eds), Atmospheric microbial aerosols. Chapman & Hall,
New York.
15. Webb SJ. 1959. Factors affecting the viability of air-borne
bacteria. I. Bacteria aerosolized from distilled water.
Can J Microbiol 5:649–669.
16. Won WD, Ross H. 1966. Effect of diluents and relative
humidity in the survival of airborne Pasteurella pestis. Appl
Microbiol 14:742–745.
17. Bolister NJ, Johnson HE, Eathes CM. 1992. The ability of
airborne Klebsiella pneumoniae to colonize mouse lungs. Epi-
demiol Infect 109:121–131.
18. Jericho KW, Langford EV, Pantekoek J. 1977. Recovery
of Pasteurella hemolytica from aerosols at differing tempera-
ture and humidity. Can J Comp Med 41:211–214.
19. Cox CS. 1989. Airborne bacteria and viruses. Sci Prog 73:
469–499.
20. Cox CS. 1971. Aerosol survival of Pasteurella tularensis dis-
seminated from the wet and dry states. Appl Microbiol 21:
482–486.
21. Assar SK, Block SS. 2000. Survival of microorganisms in
the environment. In Block SS (ed), Disinfection, steriliza-
tion, and preservation. Lippincott Williams & Wilkins, Phil-
adelphia, PA.
22. Karim YG, Ijaz MK, Sattar SA, Johnson-Lussenburg
CM. 1985. Effect of relative humidity on the air-
borne survival of rhinovirus 14. Can J Microbiol 31:
1058–1061.
23. Arundel AV, Sterling EM, Biggin JH, Sterling TD. 1986.
Indirect health effects of relative humidity in indoor envi-
ronments. Environ Health Perspect 65:351–361.
24. Harper GJ. 1961. Airborne microorganisms: survival tests
with four viruses. J Hyg Camb 59:479–486.
25. Schaffer FL, Soergl ME, Straube DC. 1976. Survival of
airborne influenza virus: effects of propagating host, relative
humidity, and composition of spray fluids. Arch Virol 51:
263–273.
26. Ijaz MK, Sattar SA, Johnson-Lussenburg CM, Spring-
thorpe VS. 1985. Comparison of the airborne survival of
calf rotavirus and poliovirus type 1 (sabin) aerosolized as
a mixture. Appl Environ Microbiol 49:289–293.
27. Webb SJ. 1960. Factors affecting the viability of
air-borne bacteria. III. The role of bonded water and pro-
tein structure in the death of air-borne cells. Can J Microbiol
6:89–105.
28. Webb SJ. 1965. The role of bound water in the maintenance of
the integrity of a cell or virus. Charles C. Thomas, Pub.,
Springfield, Ill.
29. Barlow DF. 1972. The effects of various protecting
agents on the inactivation of foot-and-mouth disease virus
in aerosols and during freeze-drying. J Gen Virol 17:
281–288.
30. Goldberg LJ, Ford I. 1973. The function of chemical addi-
tives in enhancing microbial survival in aerosols, p 86–89.
In Hers JFP, Winkler KC (eds), Airborne transmission and
infection. John Wiley & Sons, New York.
31. Cabelli VJ. 1962. The rehydration of aerosolized bacteria:
compounds which enhance the survival of rehydrated Pas-
teurella tularensis. Technical Report # 314, U.S. Army, Dug-
way Proving Ground, Utah.
32. Warren JC, Akers TG, Dubovi EJ. 1969. Effect of prehu-
midification on sampling of selected airborne viruses. Appl
Microbiol 18:893–896.
33. Hatch MT, Dimmick RL. 1966. Physiological responses of
airborne bacteria to shifts in relative humidity. Bacteriol Rev
30:597–603.
34. Dimmock NL. 1967. Differences between the thermal
inactivation of picornaviruses at “high” and “low” temper-
atures. Virology 31:338–353.
35. Handley BA, Webster AJ. 1993. Some factors affecting
airborne survival of Pseudomonas fluorescens indoors.
J Appl Bacteriol 79:358–378.
36. Handley BA, Webster AJ. 1995. Some factors affecting
airborne survival of bacteria indoors. J Appl Bacteriol 75:
35–42.
37. Ehrlich R, Miller S. 1973. Survival of airborne Pasteurella
tularensis at different atmospheric temperatures. Appl Micro-
biol 25:369–372.
38. Ehrlich R, Miller S, Walker RL. 1970. Relationship
between atmospheric temperature and survival of airborne
bacteria. Appl Microbiol 19:245–249.
39. Wathes CM, Howard K, Webster AJ. 1986. The survival
of Escherichia coli in an aerosol at air temperatures of 15 and
30 degrees C and a range of humidities. J Hyg 97:489–496.
40. Wright DN, Bailey GD. 1969. Effect of relative humidity
on the stability of Mycoplasma pneumoniae exposed to
3.2.4-10 ▪ AIR
simulated solar ultraviolet and to visible radiation. Can J
Micro 15:1449–1452.
41. Theunissen HJ, Lemmens-den Toom NA, Burggraaf A,
Stolz E, Michel MF. 1993. Influence of temperature and
relative humidity on the survival of Chlamydia pneumoniae
in aerosols. Appl Environ Microbiol 59:2589–2593.
42. Sobsey MD, Meschke JS. 2003. Virus survival in the envi-
ronment with special attention to survival in sewage drop-
lets and other environmental media of fecal or respiratory
origin. Report, World Health Organization. http://www.
unc.edu/courses/2008spring/envr/421/001/WHO_Virus
SurvivalReport_21Aug2003.pdf.
43. Bertrand I, Schijven JF, Sánchez G, Wyn-Jones P, Otto-
son J, Morin T, Muscillo M, Verani M, Nasser A, de Roda
Husman AM, Myrmel M, Sellwood J, Cook N, Gantzer
C. 2012. The impact of temperature on the inactivation
of enteric viruses in food and water: a review. Appl Microbiol
112:1059–1074.
44. Cox CS, Heckly RJ. 1973. Effects of oxygen upon freeze-
dried and freeze-thawed bacteria: viability and free radical
studies. Can J Microbiol 19:189–194.
45. Cox CS, Baldwin F. 1967. The toxic effect of oxygen upon
the aerosol survival of Escherichia coli B. J Gen Microbiol 49:
115–117.
46. Cox CS, Gagen SJ, Baxter J. 1974. Aerosol survival
of Serratia marcescens as a function of oxygen concentra-
tion, relative humidity and time. Can J Microbiol 20:
1529–1534.
47. Dungan RS. 2010. Board invited review: fate and transport
of bioaerosols associated with livestock operations and
manures. J Animal Sci 88:3693–3706.
48. Benbough JE. 1971. Some factors affecting the survival of
airborne viruses. J Gen Virol 10:209–220.
49. Webb SJ. 1967. The influence of oxygen and insitol on the
survival of semidried microorganisms. Can J Microbiol 13:
733–742.
50. Eisenstark A. 1971. Mutagenic and lethal effects of visible
and near-ultraviolet light on bacterial cells. Adv Genet 16:
167–198.
51. Rastogi RP, Richa AK, Tyagi MB, Sinha RP. 2010.
Molecular mechanisms of ultraviolet radiation-induced
DNA damage and repair. J Nucl Acids 2010:1–32. doi:
10.4061/2010/592980.
52. Bellair JT, Smith GAP, Wallis IG. 1977. Significance of
diurnal variations in fecal coliform die-off rates in the
design of ocean outfalls. J Water Pollut Control Fed 49:
2022–2030.
53. Goff GD, Spendlove JC, Adams AP, Nicholes PS. 1973.
Emission of microbial aerosols from sewage treatment
plants that use tricking filters. Health Serv Rep 88:640–652.
54. Bridges BA. 1976. Survival of bacteria following exposure
to ultraviolet and ionizing radiation, p 183–208. In Gray
TRG, Postgate JR (eds), Twenty-Sixth Symposium of the
Society for General Microbiology. Cambridge University
Press, Cambridge.
55. Krinsky NI. 1976. Cellular damage initiated by visible
light, p 209–239. In Gray TRG, Postgate JR (eds), Twenty-
Sixth Symposium of the Society for General Microbiology.
Cambridge University Press, Cambridge.
56. Druett HA, May KR. 1968. Unstable germicidal pollutant
in rural air. Nature 220:395–396.
57. May KR, Druett HA, Packman LP. 1969. Toxicity of
open air to a variety of microorganisms. Nature 221:
1146–1147.
58. Benbough JE, Hood AM. 1971. Viricidal activity of open
air. J Hyg Camb 69:619–626.
59. Mohr AJ. 1991. Development of models to explain the sur-
vival of viruses and bacteria in aerosols, p 160–190. In
Hurst CJ (ed), Modeling the environmental fate of
microorganisms. American Society for Microbiology, Wash-
ington, DC.
60. Druett HA. 1973. The open air factor, p 141–151. In Hers
JFP, Winkler KC (eds), Airborne transmission and infection.
John Wiley & Sons, New York.
61. Dimmick RL, Wolochow H, Chatigny MA. 1979. Evi-
dence that bacteria canform new cells in airborne particles.
Appl Environ Microbiol 37:924–927.
62. McDermid AS, Lever MS. 1966. Survival of Salmonella
enteriditis Pt4 and Salmonella typhimurium Swindon in aero-
sols. Lett Appl Microbiol 23:107–109.
63. Hambleton P, Broster MG, Dennis PJ, Henstridge R,
Fitzgeorge R, Conlan JW. 1983. Survival of virulent
Legionella pneumophilia in aerosols. J Hyg Lond 90:
451–460.
64. Strange RE, Cox CS. 1976. Survival of dried and airborne
bacteria, p 111–154. In Gray TRG, Postgate JR (eds),
Twenty-Sixth Symposium of the Society for General
Microbiology. Cambridge University Press, Cambridge.
65. Spendlove JC, Fannin KF. 1982. Methods of characteriza-
tion of virus aerosols, p 261–329. In Gerba CP, Goyal SM
(eds), Methods in environmental virology. Marcel Dekker,
New York.
66. Phillips G, Harris GJ, Jones MW. 1963. The effect of ions
on microorganisms. Int J Biometeor 8:27–37.
67. Ehrlich R, Miller S. 1971. Effect of relative humidity and
temperature on airborne Venezuelan equine encephalitis
virus. J Appl Microbiol 22:194–200.
68. Cox CS, Baxter J, Maidment BJ. 1973. A mathematical
expression foroxygen-induced death in dehydrated bacte-
ria. J Gen Microbiol 75:179–185.
69. de Mik G, de Groot I. 1973. Effect of gases on the aerosol
stability of various microorganisms, p 155–158. In Hers JFP,
Winkler KC (eds), Airborne transmission and infection. John
Wiley & Sons, New York.
70. Mainelis G, Gorney RL, Reponan T, Trunov M, Grinsh-
pun SA, Baron P, Yadav J, Willeke K. 2002. Effect of elec-
trical charges and fields on injury of airborne bacteria.
Biotech Bioengin 79:229–240.
71. Stetzenbach L, Lighthart B. 1994. Distribution of bioaer-
osols, p 68–98. In Lighthart B, Mohr AJ (eds), Atmospheric
microbial aerosols. Chapman & Hall, New York.
72. Lighthart B, Shaffer BT. 1994. Bacterial flux from chapar-
ral into the atmosphere in mid-summer at a high desert
location. Atmos Environ 28:1267–1274.
73. Lighthart B, Shaffer BT. 1995. Airborne bacteria in
the atmospheric surface layer: temporal distribution
above a grass seed field. Appl Environ Microbiol 61:1492–
1496.
74. Harper GJ. 1963. The influence of environment on the
survival of airborne virus particles in the laboratory. Arch
Gesamte Virusforsh 13:64–71.
75. Akers TG. 1973. Survival, damage and inactivation in
aerosols, p 73–86. In Hers JF, Winkler KC (eds), Airborne
transmission and infection. John Wiley & Sons, New York.
76. Hatch MT, Wolochow H. 1969. Bacterial aerosols, p
267–295. In Dimmick RL, Akers AB (eds), An introduc-
tion to experimental aerobiology. Wiley Interscience,
New York.
77. Ward RL, Ashley CS. 1977. Inactivation of enteric viruses
in wastewater sludge through dewatering by evaporation.
Appl Environ Microbiol 34:564.
78. Heidelberg, JF, Shahamat M, Levin M, Rahman I, Stelma
G, Colwell RR. 1997. Effect of aerosolization on culturabil-
ity and viability of gram-negative bacteria. Appl Environ
Micrbiol 63:3585–3588.
79. Marthi B, Lighthart B. 1990. Effects of betaine on enumer-
ation of airborne bacteria. Appl Environ Microbiol 56:
1286–1289.
 3.2.4. Fate and Transport of Microorganisms in Air ▪ 3.2.4-11
80. Lighthart B. 1973. Survival of airborne bacteria in a high
urban concentration of carbon monoxide. Appl Microbiol
25:86–91.
81. Lighthart B, Hiatt VE, Rossano AT, Jr. 1971. The survi-
val of airborne Serratia marcescens in urban concentrations
of sulfur dioxide. J Air Pol Cont Assoc 21:539–642.
82. Marthi B, Fieland VP, Walter M, Seidler RJ. 1990. Sur-
vival of bacteria during aerosolization. Appl Environ Micro-
biol 56:3436–3467.
83. Walter MV, Marthi B, Fieland VP, Ganio LM. 1990.
Effect of aerosolization on subsequent bacterial survival.
Appl Environ Microbiol 56:3468–3472.
84. Cox CS, Baldwin F. 1966. The use of phage to study causes
of loss of viability of Escherichia coli in aerosols. J Gen Micro-
biol 44:15–22.
85. Cox CS. 1966. The survival of Escherichia coli atomized into
air and into nitrogen from distilled water and from solution
of protecting agents as a function of relative humidity. J Gen
Microbiol 43:383–399.
86. Cox CS. 1968. The aerosol survival of Escherichia coli B in
nitrogen, argon, and helium atmospheres and the influence
of relative humidity. J Gen Microbiol 50:139–147.
87. Cox CS. 1970. Aerosol survival of Escherichia coli B disse-
minated from the dry state. Appl Microbiol 19:604–607.
88. Webb SJ. 1969. The effects of oxygen on the possible repair
of dehydration damage by Escherichia coli. J Gen Microbiol
58:317–326.
89. Webb SJ, Booth AD. 1969. The effect of radiation on
Escherichia coli. Nature 222:1199–1200.
90. Wright DN, Bailey GD, Hatch MT. 1968. Survival of air-
borne Mycoplasma as affected by relative humidity.
J Bacteriol 95:251–252.
91. Lembke LL, Kniseley RN. 1980. Coliforms in aerosols
generated by a municipal solid waste recovery system.
Appl Environ Microbiol 40:888–891.
92. Kethley TW, Crown WB, Fincher EL. 1957. The nature
and composition of experimental bacterial aerosols. Appl
Microbiol 5:1–17.
93. Wright DN, Bailey GD, Goldberg LJ. 1969. Effect of tem-
perature on survival of airborne Mycoplasma pneumoniae.
J Bacteriol 99:491–495.
94. Cox CS, Goldberg LJ. 1972. Aerosol survival of Pasteurella
tularensis and the influence of relative humidity. Appl
Microbiol 23:1–3.
95. Hess GE. 1965. Effects of oxygen on aerosolized Serratia
marcescens. Appl Microbiol 13:781–787.
96. Wells WF. 1934. Droplet and droplet nuclei. Amer J Hyg
20:611–627.
97. Mohr AJ. 1984. Aerosol stability of reovirus. Dissertation,
Utah State University, Logan, UT.
98. Miller WS, Artenstein MS. 1966. Aerosol stability of three
acute respiratory disease viruses. Proc Soc Exp Biol Med 123:
222–227.
99. Elazhary MASY, Derbyshire JB. 1977. Effect of tempera-
ture, relative humidity and medium on the aerosol stability
of infectious bovine rhinotracheitis virus. Can J Comp Med
43:158–167.
100. Songer JR. 1966. Influence of relative humidity on the sur-
vival of some airborne viruses. Appl Microbiol 15:1–16.
101. Akers TG, Bond S, Goldberg LJ. 1966. Effect of temper-
ature and relative humidity on the survival of airborne
Colombia sk group viruses. Appl Microbiol 14:361–365.
102. Jensen MM. 1964. Inactivation of airborne viruses by ultra-
violet irradiation. Appl Microbiol 12:418–412.
103. de Jong JC. 1970. On the mechanism of the decay of polio-
myelitis virus and encephalomyocarditis virus in aerosols,
p 210–211. In Silver IH (ed), Aerobiology, Third Interna-
tional Symposium on Aerobiology, Brighton, 1969. Aca-
demic Press, London.
104. de Jong JC. 1970. Decay mechanism of polio and emc
viruses in aerosols, p 210–245. In Silver IH (ed), Third
International Symposium on Aerobiology. Academic
Press, London.
105. Donaldson AI. 1972. Effect of radiation on selected virus.
Vet Bull 48:83–94.
106. Donaldson AI, Ferris NP. 1975. The survival of foot-and-
mouth disease virus in open air conditions. J Hyg Camb 74:
409–415.
107. Smith LP, Hugh-Jones ME. 1969. The weather factor in
foot-and-mouth diseaseepidemics. Nature 223:712–715.
108. Loosli CG, Lemon HM, Robertson OH, Appel E. 1943.
Experimental air-borne influenza infection. I. Influence
of humidity on survival of virus in air. Proc Soc Exp Biol
Med 53:205–206.
109. Akers TG, Hatch MT. 1968. Survival of a picornavirus
and its infectious rna after aerosolization. Appl Microbiol
16:1811–1813.
110. Webb SJ, Bather R, Hodges RW. 1963. The effect of rel-
ative humidity and inositol on air-borne viruses.
Can J Microbiol 9:87–94.
111. Hemmes JH, Winkler KC, Kool SM. 1960. Virus survival
as a seasonal factor in influenza and poliomyelitis. Nature
188:430–438.
112. de Jong JC, Winkler KC. 1968. The inactivation of polio-
virus in aerosols. J Hyg Camb 66:557–565.
113. Adams DJ, Spendlove JC, Spendlove RS, Barnett BB.
1982. Aerosol stability of infectious and potentially infec-
tious reovirus particles. Appl Environ Microbiol 44:
903–908.
114. de Jong JC, Trouwborst T, Winkler KC. 1973. The mech-
anism of virus decay in aerosols. In Hers JFP, Winkler C
(eds), Airborne transmission and infection. John Wiley &
Sons, New York.
115. Moe K, Harper GJ. 1983. The effect of relative humidity
and temperature on the survival of calf rotavirus in aerosol.
Arch Virol 76:211–216.
116. Sattar SA, Ijaz MK, Johnson-Lussenburg CM, Spring-
thorpe VS. 1984. Effect of relative humidity on the air-
borne survival of rotavirus SA11. Appl Environ Microbiol
47:879–881.
117. de Jong JC, Harnsen M, Plantinga AD, Trouwborst T.
1976. Aerosol stability functions of Semliki Forest virus.
Appl Environ Microbiol 32:315–319.
118. Rabey F, Janssen RJ, Kelley LM. 1969. Stability of
St. Louis encephalitis virus in the airborne state. Appl Envi-
ron Microbiol 18:880–882.
119. Berendt RF, Dorsey EL. 1971. Effect of simulated solar
radiation and sodiumfluorescein on the recovery of Vene-
zuelan equine encephalomyelitis virus from aerosols. Appl
Microbiol 21:447–450.
120. Fairchild GA. 1974. Ozone effect on respiratory deposition
of vesicular stomatitis virus aerosols. Amer Rev Resp Dis
109:446–451.
121. Mayhew CJ, Zimmerman WD, Hahon N. 1968. Assess-
ment of aerosol stability of yellow fever virus by fluorescent-
cell counting. Appl Microbiol 16:263–266.
122. Buckland FE, Tyrrell DA. 1964. Loss of infectivity on dry-
ing various viruses. Nature 195:1063–1064.
123. Trouwborst T, de Jong JC, Winkler KC. 1972. Mecha-
nism of inactivation in aerosols of phage T1. J Gen Virol
15:235.
124. Happ JW, Harstad JB, Buchanan LM. 1951. Effect of air
ions on submicron T1 bacteriophage aerosols. Appl Micro-
biol 14:888–891.
125. Cox CS, Harris WJ, Lee J. 1974. Viability and electron
microscope studies of phages T3 and T7 subjected to freeze-
drying, freeze-thawing and aerosolization. J Gen Microbiol
81:207–215.
3.2.4-12 ▪ AIR
126. Hatch MT, Warren JC. 1969. Enhanced recovery of
airborne T3 coliphage and Pasteurella pestis bacteriophage
by means of a presampling humidification technique.
Appl Microbiol 17:685–689.
127. Dubovi EJ, Akers TG. 1970. Airborne stability of tailless
bacterial viruses S-13 and MS-2. J Appl Microbiol 19:
624–630.
128. Dubovi EJ. 1971. Biological activity of the nucleic acid
extracted from two bacterial viruses. Appl Microbiol 21:
624–630.
129. Wickman HH. 1994. Deposition, adhesion, and release, p
99–165. In Lighthart B, Mohr AJ (eds), Atmospheric micro-
bial aerosols. Chapman & Hall, New York.
130. Soddu A, Vieth WR. 1980. The effect of sugars on mem-
branes. J Mol Catal 7:491–500.
131. Spendlove JC. 1975. Penetration of structures by microbial
aerosols, 427–436. In Underkofler LA (ed), Developments in
industrial microbiology. American Institute of Biological
Sciences, Washington, DC.
132. Adams AP, Spendlove JC. 1970. Coliform aerosols emit-
ted by sewage treatment plants. Science 169:1218–1220.
133. Lighthart B. 1984. Microbial aerosols: estimated contribu-
tion of combine harvesting to an airshed. Appl Environ
Microbiol 47:430–432.
134. Sorber CA, Moore BE, Johnson DE, Harding HJ,
Thomas RE. 1984. Microbiological arosols from the appli-
cation of liquid sludge to land. J Water Pollut Control Fed 56:
830–836.
135. Teltsch B, Shuval HI, Tadmor J. 1980. Die-away kinetics
of aerosolized bacteria from sprinkler application of waste-
water. Appl Environ Microbiol 39:1191–1197.
136. Zeterberg JM. 1973. A review of respiratory virology
and the spread of virulent and possibly antigenic
viruses via air conditioning systems. Ann Allergy 31:
228–234.
137. Cook RJ, Baker KF. 1983. The nature and practice of biolog-
ical control of plant pathogens. American Phytopathological
Society Press, St. Paul, MN.
138. Pasquill F. 1961. The estimation of the dispersion of wind
borne material. Meteor Mag 90:33–49.
139. Lighthart B, Frisch AS. 1976. Estimation of viable
airborne microbes downwind from a point source. Appl
Environ Microbiol 31:700–704.
140. Lighthart B, Mohr AJ. 1987. Estimating downwind con-
centrations of viable airborne microorganisms in dynamic
atmospheric conditions. Appl Environ Microbiol 53:
1580–1583.
141. Lighthart B, Kim J. 1989. Simulation of airborne microbial
transport. Appl Environ Microbiol 55:2349–2355.
142. Mayer D, Reiczigel J, Rubel F. 2008. A Lagrangian particle
model to predict the airborne spread of foot-and-mouth dis-
ease virus. Atmos Environ 42:466–479.
143. Minyoung K, Thurston JA, Hagen LJ. 2007. Computa-
tional fluid dynamic (CFD) modeling to predict bioaerosol
transport behavior during center pivot wastewater irriga-
tion. In American Society Agricultural Biology Engineer-
ing Annual International Meeting, Technical Paper No.
074063, 8, 1–8.
144. Nimmermark SA, Jacobson LD, Schmidt DR, Gay SW.
2005. Predictions by the odor from feedlots, setback
estimation tool (OFFSET) compared with observations
by neighborhood monitors. J Air Waste Manage 55:
1306–1314.
145. Jones AM, Harrison RM. 2003. The effects of meterolog-
ical factors on atmospheric bioaerosols concentrations—a
review. Sci Total Environ 326:151–180.
 Airborne Fungi and Mycotoxins
DE-WEI LI, ECKARDT JOHANNING, AND CHIN S. YANG
3.2.5
Fungi are a heterogeneous group of organisms including true
fungi (Kingdom Eumycota or Fungi), lichens (a fungus and
an alga in symbiotic relationship), the true slime molds (Myx-
omycetes, Kingdom Protozoa), and the water molds (Oomy-
cetes, Kingdom Chromista). At present the latter two are
sometimes referred to as fungi-like organisms due to their phy-
logenetic differences from the true fungi. Frequently, they are
still called fungi, since they are traditionally studied by mycol-
ogists and covered in mycology (1). Fungi as a group inhabit a
wide range of niches and environments from plants, plant
debris, soils, and animals to exposed rock (some lichens), riv-
ers and lakes (aquatic fungi), the sea (marine fungi), the
North Pole, and the tropics (2). They have developed
many different modes of obtaining nutrients. Fungi function
in both beneficial and detrimental ways from a human per-
spective. Some fungi, such as powdery mildews and rusts,
are obligate parasites of plants. Some fungi exist in symbiotic
relationships with plant roots to form mycorrhizae or with
algae to form lichens. Some fungi are able to break down or
detoxify wastes and other pollutants (3). A large number of
fungi survive as saprobes recycling nutrients in ecosystems,
such as carbon, nitrogen, and sulfur. Unfortunately, some of
these saprobes have been found to grow in the indoor envi-
ronments of buildings and have led to building related
complaints and illnesses (4–7). Samson (8) estimated that
100–149 species occurs indoors, while Miller and Day (9)
found approximately 270 species recovered from the dust
indoors (10). Li and Yang (11) found that over 600 species
were identified from samples collected from indoor environ-
ments in North America based on the information in a
database of a commercial laboratory. However, the exact
number still remains unknown. Likely it will exceed 600 spe-
cies, as occasionally a new record is discovered from samples
collected from the indoor environment. A majority of indoor
fungi are anamorphic fungi.
Fungi develop and release their spores into the air for dis-
persal (12). Human beings are often exposed to fungi by inha-
lation of airborne fungal spores, hyphal fragments, and fungal
by-products, especially in indoor environments. Certain
fungi have been associated with asthma and respiratory con-
ditions (13–15).
Fungi produce a variety of secondary metabolites, includ-
ing mycotoxins and some fungal volatile organic com-
pounds (VOC’s), also known as microbial VOCs (MVOCs)
(3, 16–18). Mycotoxins are harmful to animals and humans
doi:10.1128/9781555818821.ch3.2.5
3.2.5-1
(19–21) and are suggested as a major possible human health
risk factor in buildings with mold problems (22, 23). In addi-
tion to mycotoxins, some MVOCs produced by actively grow-
ing fungi as both primary and secondary metabolites are
known irritants or hazardous chemicals (18, 24). They may
pose a health risk to building occupants and workers handling
fungal matter (5, 6, 25–27).
This chapter reviews the existing literature on airborne
fungi with an emphasis on indoor fungal growth and contam-
ination as well as the principal human health effects of expo-
sure to fungi, mycotoxins, and other by-products. It is
important to note that bacteria (including actinomycetes),
mites, insects, and other microbes may also grow in a wet,
damp indoor environment.
A wealth of literature on outdoor airborne fungi can be
found in reviews by Gregory (28), Flannigan et al. (29), Lacey
(30, 31), and Levetin (32, 33). It is important to note that
outdoor airborne spores are often the source of indoor fungal
spora. Their impact on indoor airborne fungal populations
could be immediate or delayed until they have settled and
colonized an indoor environment. Sampling of airborne
and indoor fungi is not covered in this chapter. For any issues
or questions related to sampling of indoor and airborne
fungi, refer to chapter 3.2.2 in this book and several guides/
books published by AIHA in the past 10 years (34–36).
AIRBORNE FUNGAL POPULATIONS
It must be emphasized that it is to the selective advantage of
fungi to release, disperse, and disseminate their spores and
occasionally hyphal fragments or partial conidiophores in
air from one location to others. They cannot survive and
complete their life cycles by staying afloat in air for an indef-
inite period of time. The dispersal of fungal spores can be
short or long distance (37, 38). The majority disperse their
spores over a short distance. Therefore, when discussing fun-
gal contamination, identifying and locating the source of fun-
gal colonization is often of higher importance than assessing
airborne fungi data (39, 40). In addition to airborne dispersal,
some fungi rely on running water, insects, and animals for dis-
persal of their spores (41, 42).
A large collection of literature on assessing indoor fun-
gal populations has been accumulated. The majority of
the literature was based on air sampling data. These include
hospitals and health care facilities (43–48), residential
3.2.5-2 ▪ AIR
dwellings (49–55), schools (56–58), and office buildings (59,
60). The focus of hospital sampling was often on Aspergillus
species, including A. fumigatus, an opportunistic human
pathogen (43). General fungal populations were identified
in nonhospital sampling.
A comprehensive assessment of fungal contamination in
the indoor environment should include consideration of fac-
tors such as outdoor air, air conditioning, heating and venti-
lation systems, ventilation mode, heating, occupant density,
ventilation rate, moisture (including water damage, high
relative humidity [RH] in the air, and dampness), mainte-
nance, on-site inspection, air sampling, surface and source
sampling, sample analysis, risk analysis, and finally remedial
actions (49, 61). Unfortunately, the majority of investigations
fail to follow the approach of comprehensive assessment,
often due to insufficient strategy, labor, time, and funds as well
as understanding of environments and ecosystems indoors.
An alternate approach to identifying indoor fungal contami-
nation is to focus on inspection and surface/source sampling.
Factors Affecting Airborne Fungal Populations
Three important factors that directly affect airborne fungal
populations are the availability of food/substrates and free
water for fungal growth and the methods of spore dispersal.
Other physical, chemical, and biological parameters affecting
fungal growth, and subsequent airborne fungal populations,
can be found in recent references (2, 3, 8, 41).
Substrates (Including Water Activity)
Fungi are achlorophyllous and heterotrophic, take up
nutrients by absorption from substrates, and require simple
sugars, carbohydrates, and other organics, such as vitamins,
amino acids, and essential mineral elements to survive (3).
In the natural environment, fungi have developed a number
of ways to obtain these nutrients (3), such as necrotrophic,
symbiotic, and saprotrophic relationships.
Humans share food, living space, environments, and
resources with fungi. We utilize fungi to produce bread; man-
tou (steamed bun); cheese; edible mushrooms; alcoholic bev-
erages; and useful by-products, such as antibiotics, enzymes,
and organic acids (60). Some fungi cause food spoilage or
make food toxic to humans. Botrytis cinerea is a well-known
pathogenic fungus causing gray mold disease on grapevines,
strawberries, and many other fruits and produce. Species of
Penicillium and Aspergillus often cause spoilage in foodstuffs
and make them inconsumable to animals and humans. On
the other hand, Penicillium camemberti and P. roqueforti are
used in cheese production. Fungi are also known to cause
wood stains, wood decay, biodeterioration, and biodegrada-
tion of polymers, carpet, plaster, drywalls, wallpaper, paints
and organic coatings, fuels and lubricants, leather, fabric
products, paper, and wood products (4, 62–65). These refer-
ences underscore the very likelihood that fungi can and will
grow in artificial environments. Consequently, controlling
nutrient sources to limit fungal proliferation/growth is practi-
cally impossible.
One of the critical factors affecting indoor fungal growth is
water. There are a number of ways to measure water availabil-
ity in materials. Water or moisture content of a material is
expressed as a percentage of the oven dry weight (65). How-
ever, water content does not suggest the actual availability of
free water in the material to fungi. A better measurement of
water availability to fungi is water activity. Water activity is
numerically equal to equilibrium relative humidity (ERH)
expressed as a decimal. If a sample of substrate is held at
constant temperature in a sealed enclosure until the water
in the sample reaches equilibrium with the water vapor in
the enclosure, then aw = ERH/100. Another expression is:
vapor pressure of water in substrate
aw ¼
vapor pressure of pure water
A detailed discussion of water activity of fungi in food and
materials is presented by Gravesen et al. (4), Hocking and
Miscamble (66), Li and Yang (40), Smith and Onion (67),
Troller and Scott (68), Adan et al. (69), and Huinink and
Adan (70).
Many common indoor fungi are hydrophilic and require
aw near 1 for growth. Some xerophilic fungi, however, have
optimal water activity ranging from 0.65 to 0.90 (4). Both
mesophilic and xerophilic fungi can be found in the indoor
environment. Table 1 lists some hydrophilic, mesophilic,
and xerophilic fungi and their minimum aw. It should be
noted that a very short period of peak humidity, even below
saturation RH, may lead to fungal growth. Adan et al. (69)
showed that it took only 69 h for Penicillium chrysogenum
Thom to develop from spore germination to sporulation
and 73 h to develop mycelial mass on pure gypsum at 21°C
and 97% RH observed with cryo-SEM.
Fungal Spore Discharge Mechanisms
Fungal spores are released by two basic mechanisms: (1)
active spore discharge and (2) passive spore release. Concen-
trations of certain airborne fungal spores have been known to
peak during certain hours of the day or night. This periodicity
is related to spore discharge mechanisms and environmental
factors in nature (28, 30, 42). Details of these two mecha-
nisms and environmental factors affecting spore release are
presented by Lacey (30) and Levetin (33).
Fungi with active spore discharge include such common
airborne fungi as Sporobolomyces, Epicoccum, Nigrospora, and
some smut-like yeasts. Many ascospores and basidiospores
also have active discharge mechanisms (33). Sporobolomyces
and some basidiospores are usually most abundant at night
or in the predawn hours. Their spore release requires the
absorption of moisture to build-up release pressure or forming
a droplet at the hilar area. Dry spore fungi, such as Aspergillus,
Penicillium, and Cladosporium, are often hydrophobic. They
become airborne by passive force, such as air movement or
rain droplets (71, 72). Cladosporium usually dominates the air-
borne spore population during the day. Its spores stay airborne
owing to the buoyancy of warmer air.
Some fungi, such as puffballs (Calvatia, Lycoperdon, Sclero-
derma) of basidiomycetes, are able to produce a huge number
of spores and release their spores in “spore clouds or puffs”
when affected by raindrops, humans, or small animals (33,
71–73). The spore clouds may persist for a period of time until
air mixing and dilution disperse them. Results of air sampling
can be greatly affected by whether the spore cloud has dis-
persed (31).
Many fungi that are frequently detected indoors and out-
doors produce spores in a slimy mass. These include such
common indoor contaminants as Acremonium (although
some species of Acremonium produce dry spores), Aureobasi-
dium, Fusarium, Phoma, Stachybotrys, Trichoderma, and yeasts.
Slimy spores may be released into the air when they become
dry, disturbed, or attached to other particles. Their dissemina-
tion is often assisted by insects, mites, small animals, or water
(42). Because slimy spores do not become airborne easily,
their detection indoors should be considered significant.
Any detection of Stachybotrys in air samples taken indoors
 3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-3

TABLE 1 Selected fungi and their aw b

aw T (°C) aw T (°C)
Absidia corymbifera 0.88 25 Paecilomyces variotii 0.84 25
Alternaria citri 0.84 25 Penicillium brevicompactum 0.81 23
Aspergillus candidus 0.75 25 P. chrysogenum 0.79 25
A. flavus 0.78 33 P. citrinum 0.80 25
A. fumigatus 0.82 25 P. corylophilum 0.80 25
A. niger 0.77 35 P. expansum 0.83 23
A. ochraceus 0.77 25 P. frequentans 0.81 23
A. restrictus 0.75 25 P. griseofulvum 0.81 23
A. sydowii 0.78 25 P. spinulosum 0.80 25
A. terreus 0.78 37 Phoma herbarum 0.93 25
A. versicolor 0.78 37 Rhizopus microsporus 0.90 25
A. wentii 0.84 25 R. stolonifer 0.84 25
Cladosporium cladosporioides 0.84 25 R. oryzae 0.88 25
Cladosporium sphaerospermum 0.84 25 Scopulariopsis brevicaulis 0.90 NA
Eurotiuma amstelodami 0.70 25 Sistotrema brinkmannii 0.97 25
Eurotium chevalieri 0.71 33 Stachybotrys chartarum 0.93 25
Emericellaa nidulans 0.78 37 Syncephalastrum racemosum 0.84 25
Fusarium moniliforme 0.89 25 Trichoderma harzianum 0.91 25
Geomyces pannorum 0.89 25 Wallemia sebi 0.70 25
Mucor circinelloides 0.90 25 Ulocladium chartarum 0.89 25
M. plumbeus 0.93 25 U. consortiale 0.89 25
a
Both genera are teleomorphs of Aspergillus.
b
NA, not available. Sources: 40, 65–67; For calculation of aw, see text.

should trigger further investigation and search for the con-
firmed presence of this fungus.
Airborne Fungal Populations
Fungi as a group produce a number of different spores, both
sexual and asexual. Many types of spores are capable of
becoming airborne. Some spores, such as chlamydospores
(asexual) and zygospores (sexual), are not designed to be dis-
persed by air transmission, and there has been no report of
recovering these spores in air or they are rarely observed in
the air, such as spores of Glomus species (Klironomos, unpub-
lished data). Hypogeous fungi (including such well-known
fungi as truffles) produce subterranean sporocarps and dis-
perse their spores through different mechanisms. However,
both sexual and asexual spores of four major groups (Myxo-
mycetes, zygomycetous fungi, Ascomycetes including ana-
morphic fungi [formerly classified as Deuteromycetes], and
Basidiomcetes) have been isolated and reported from air.
Spores of Ascomycetes and Basidiomycetes have fre-
quently been recovered from air samples using spore trap sam-
plers. Levetin (41) reported 18 genera of basidiospores from
the atmosphere in Tulsa, Oklahoma. Furthermore, many spe-
cies of basidiospores have been demonstrated to be allergenic
(74–80). The importance of basidiomycetes in the indoor
environment depends on the building construction. Wood-
inhabiting basidiomycetes, such as polypores, are associated
with wood decay and may be found in buildings constructed
of wood. In fact, wood decay caused by polypores is a signifi-
cant problem in the United Kingdom (81) and in the United
States (65). A number of Coprinus species and several other
mushrooms were found growing in buildings with water dam-
age (82). Unfortunately, identification of airborne basidio-
mycetes collected using culture techniques can be difficult
and they are often missed or misidentified. A real-time
PCR test has been developed for the detection of dry-rot
fungi, Meruliporia incrassata and Serpula lacrymans (83).
Ascospores have been identified from air samples using
spore traps (84). Many Ascomycetes, however, do not pro-
duce ascomata and ascospores in culture. This makes it diffi-
cult to determine the frequency of occurrence of ascospores in
air. Some ascomycetes, such as Ascotricha, Chaetomium,
Emericella, and Eurotium (both Emericella and Eurotium are
teleomorphs of some Aspergillus species) are commonly
formed in cultures. Ascomata of species of Ascotricha, Peziza,
and Pyronema are frequently observed or reported from indoor
environments (11, 85). This demonstrates that ascospores
may become airborne. Ascospores are suspected to be aller-
genic. Four species of Chaetomium were listed as licensed
by the FDA for commercial production as allergens (84).
Rivera-Mariani et al. (86) demonstrated that out of 33 aller-
gic rhinitis and asthmatic patients in Puerto Rico, 31 reacted
to ascospores, 29 to basidiospores, 19 to hyphae/fungal frag-
ments, and 12 to mitospores. They stated that sensitization
to airborne spores of basidiomycetes, ascomycetes, and fungal
fragments seem to be more prevalent than to similar numbers
of conidia in patients with active allergies, suggesting a possi-
ble role in exacerbations of respiratory allergies in tropical
environments. Otherwise, little documentation is available
on the allergenicity of ascospores.
The majority of airborne fungi collected on samplers and
grown on agar media are anamorphic fungi and zygomycetous
fungi. Most anamorphic fungi are asexual states of Ascomy-
cota and a minority Basidiomycota. Asexual spores of ana-
morphic fungi are called conidia, and of zygomycetous
fungi, sporangiospores. Many of these spores are known aller-
gens. Some of them have been prepared into allergen extracts
3.2.5-4 ▪ AIR
and approved by the FDA for medical uses (84). Common
anamorphic fungi found in air include Alternaria, Aspergillus,
Cladosporium, Epicoccum, and Penicillium. Mucor, Rhizopus,
and Syncephalastrum, all zygomycetous fungi, are also fre-
quently isolated from air.
Another group capable of producing and releasing spores
into the air is the Myxogastria (formerly Myxomycota) or
the true slime molds. Slime molds, as a group, are polyphy-
letic. They are considered to have similarities to both true
fungi and animals (1) and have been placed in the Kingdom
Protista (33, 87). Spores of slime molds have been docu-
mented from air samples (28, 87). Occasionally, species of
Stemonitis have been found growing in environments such
as a cellar, basement, and window sill underneath a leaking
window air conditioning unit in buildings with water issues
( personal observation). Allergic reactions to extracts of slime
molds have been reported. Giannini et al. (77) reported that
15.4% of patients tested showed positive skin reactions to
extracts of Fuligo septica, Lycogala epidendrum, and Stemonitis
ferruginea. Benaim-Pinto (74) and Santilli et al. (79) found
that patients yielded positive skin responses to spore extracts
of Fuligo septica.
Outdoor airborne fungal populations may directly or indi-
rectly affect the indoor populations as the pathways of infiltra-
tion are often suspected to be from leaks and cracks or through
doors, windows, and air intake systems. Therefore, it is not
surprising that common outdoor fungal taxa are often the
predominant fungal types detected indoors (42, 42, 88). In
a review of the literature, some agreements and disagreements
exist on the predominant fungi identified indoors. Yang et al.
(88), based on the culture of over 2,000 Andersen samples
collected outdoors and in nonresidential buildings in the
United States, found that Cladosporium, Penicillium, Aspergil-
lus, Basidiomycetes, and Alternaria were the top five fungal
taxa found indoors as well as outdoors in frequency of occur-
rence. All five fungal taxa were detected in less than 40% of
indoor samples. However, Cladosporium was found in over
80% of outdoor samples, and Penicillium was detected in
58%. These suggest that both Cladosporium and Penicillium
are common in outdoor air. The results were somewhat in
agreement with those reported by Strachan et al. (52) from
British homes and by Womble et al. (89) in 86 office build-
ings in the continental United States. Strachan et al. (52)
found that Penicillium, Cladosporium, and Basidiomycetes
(including Sistotrema brinkmanii) were the common types of
mold isolated as well as the predominant mold concentrations
measured. Womble et al. (90) found, in rank order, nonspor-
ulating fungi, Cladosporium, Penicillium, yeast, and Aspergillus
were the most common fungi indoors. Using a number of
different types of samplers and sampling media, VerHoeff
et al. (53) found that species of Cladosporium, Penicillium,
and Aspergillus (including teleomorph Eurotium) were com-
mon in homes in the Netherlands. In a survey of 10 elemen-
tary schools in southern California using an Andersen
sampler, Dungy et al. (56) found that Cladosporium, Alterna-
ria, Penicillium, sterile mycelia, and Epicoccum were the top
five fungal groups isolated indoors. The predominant fungi
detected outdoors were slightly different in that Aureobasi-
dium was more frequently encountered than Epicoccum. How-
ever, using a Roto Rod sampler, they found that spores of
Alternaria, rust fungi, Cladosporium, Epicoccum, and smut
fungi were predominant both indoors and outdoors. Through
further comparison of airborne fungal populations in the same
study, the authors found that the top seven fungal types were
identical at schools and at homes. In a hospital sampl-
ing, Solomon et al. (45) found Aspergillus fumigatus, A. niger,
Mucor, Paecilomyces, and yeasts were the five most common
fungi recovered in culture at 37°C. The agreements and dis-
agreements in the findings may be attributed to differences
in sampling techniques, isolation media used, incubation
temperatures, and geographical areas (42, 88).
Shelton et al. (91) evaluated 12,026 fungal air samples
(9,619 indoor samples and 2,407 outdoor samples), collected
from 1,717 buildings located across the United States using
Andersen N6 single-stage samplers. Ninety-nine percent of
the samples were collected with rose bengal agar and the
other 1% with malt extract agar. The culturable airborne fun-
gal concentrations in indoor air were lower than those in out-
door air. However, Stachybotrys chartarum was identified in
the air in 6% of the buildings studied and 1% of the outdoor
samples. The fungal levels were highest in the fall and summer
and lowest in the winter and spring. Geographically, the high-
est fungal levels were found in the southwest, far west, and
southeast. Because different fungal isolation media are known
to have different selective effects, the combined use of the
data derived from two different media is inappropriate. The
reliability of fungal identification, data, and sampling quality
control of such a large project must be scrutinized before the
results and conclusions are fully accepted.
In a study of 50 single-family detached homes built since
1945 with less than 0.18 m2 (or 2 ft2) of known water damage,
and located within a central city census tract in the metropol-
itan Atlanta city (DeKalb and Fulton Counties) of Georgia,
air and dust samples were collected for assessment to establish
a baseline of “normal and typical” types and concentrations of
fungi in urban homes (92). The homes were predetermined
not to have noteworthy moisture problems or indoor fungal
growth. The homes were sampled twice (summer and winter)
within a calendar year. Air samples were collected with a Spi-
ral Air System at 180 lpm onto MEA plates. Positive-hole cor-
rection was applied. Dust samples were sieved and inoculated
by the “direct plating” method onto MEA and DG18 media.
Cladosporium cladosporioides, Cladosporium sphaerospermum,
and Cladosporium spp. were the top three fungal species and
group in both indoor and outdoor air samples. The findings
included that rankings by prevalence and abundance of the
types of airborne and dustborne fungal spores did not differ
from winter to summer, nor did the rankings differ when air
samples taken indoors were compared with those taken out-
doors. Water indicator fungi (such as Chaetomium, Stachybo-
trys, and Ulocladium) were essentially absent from both air and
dust samples.
Indoor Sources of Fungi
The detection of airborne fungi does not necessarily suggest
growth and reproduction of fungi indoors. However, it is
believed that actively growing fungi in the indoor environ-
ment are the primary cause of adverse health effects due to
exposure to indoor fungal allergens, mycotoxins, and fungal
MVOCs. It is therefore important to identify and detect infes-
tation sites of fungi indoors. Due to the limitation of correct
identification of fungi, some of the health effects of indoor
fungi are arguably considered not necessarily species-specific,
nevertheless, correct identification of fungi may be important
for practical and research issues.
Many common indoor fungi are strong biodeteriorating
agents and have been reported from various building materi-
als and systems. Raper and Fennell in their classic publication
(93) reported various Aspergilli from building materials, such
as wallpaper and paper products, textiles, jute, insulation
materials, and fabrics. Two species of Aspergillus were isolated
from and found to grow on glass (94). Many species of the

genus Penicillium, commonly detected in indoor air sampling,
were frequently referred to as food spoilage and biodeteriorat-
ing agents (4, 64, 85, 93). Penicillium chrysogenum is reported
to be the most common fungus indoors (95). However, Scott
et al. (95) found that P. chrysogenum is rare in outdoor air.
P. chrysogenum was further studied by Houbraken et al. (96)
using partial β-tubulin, calmodulin, and RPB2 data sets. Hou-
braken et al. (96) found that Fleming’s penicillin-producing
strain is not Penicillium chrysogenum but P. rubens. It is, there-
fore, P. rubens that is one of most common fungi in indoor
environments, not P. chrysogenum s. str. Gravensen et al.
(4) included a list of 13 fungal species as important molds
in damp buildings. Samson et al. (7, 74) described some com-
mon fungal species from indoor environments.
Morgan-Jones and Jacobsen (63) studied moldy carpets,
plasterboard, and wallpaper from three hotels in Florida and
Georgia. Their brief literature review suggested that many
fungi had been reported to cause biodeterioration of paper,
textiles, and plaster. The genera of fungi most often identified
were the ascomycete genus Chaetomium; dematiaceous
hyphomycete genera Alternaria, Cladosporium, Stachybotrys,
and Ulocladium; the moniliaceous hyphomycete genera Acre-
monium, Aspergillus, and Penicillium; and the pycnidial genus
Phoma. In the study, 14 species, including 2 new species of
Cladosporium, in 11 genera were isolated and identified. In a
study of toxicity of moldy building materials, Johanning
et al. (97) not only detected cytotoxicity of the materials to
cell cultures but also identified satratoxin H and spirolac-
tone/lactams and several groups of fungi. The fungi were iso-
lated from gypsum wallboards and other building materials.
The fungi identified included those described by Morgan-
Jones and Jacobsen and additional species of Aspergillus, Pae-
cilomyces, and Trichoderma.
Li and Yang (11) described seven new records or note-
worthy fungi isolated from indoor environments. The
seven species are Ascotricha chartarum, A. erinacea, Memno-
niella echinata, Sporoschisma saccardoi, Stachybotrys microspora,
S. nephrospora, and Zygosporium masonii. All species were
recovered from water-damaged building materials, such as
drywall, wallpaper, or wood. Four species were reported for
the first time from the United States. Li et al. (98) described
a new species from indoor environments in the United States
and Canada. Corda (99) described Stachybotrys chartarum as
S. atra from wallpaper of a residence in Prague in 1837.
In the past several years, mass sequencing techniques
were used to study indoor fungi from dust samples collected
around the world in several studies (100). Amend et al.
(100) found that fungal diversity in the temperate region is
greater than that in the tropical region and building function
does not have significant effect on indoor fungal composi-
tion, despite differences between architecture and materials
of some buildings in close vicinity based on 72 dust samples
collected from six continents. However, Andersen et al.
(101) showed that indoor mycota are associated with differ-
ent building materials: (a) Acremonium spp., Penicillium chrys-
ogenum, Stachybotrys spp., Ulocladium spp. and gypsum and
wallpaper; (b) Arthrinium phaeospermum, Aureobasidium pul-
lulans, Cladosporium herbarum, Trichoderma spp., yeasts and
woods and plywood; and (c) Aspergillus fumigatus, Aspergillus
melleus, Aspergillus niger, Aspergillus ochraceus, Chaetomium
spp., Mucor racemosus, Mucor spinosus, and concrete and
other floor-related materials in Denmark and Greenland.
In addition to building materials, fungi have been known
to grow in the heating, ventilating, and air-condition-
ing system (HVAC) (54, 102). Heinemann et al. (103)
studied contamination of fungi, bacteria, and thermophilic
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-5
actinomycetes. They sampled surfaces of filters and fans with
RODAC contact plates and water from humidifiers. The
serial dilution method was used for analyzing humidifier
water. A wide variety of fungi were identified. However,
some of the identified fungi were likely spore contaminants
rather than the result of fungal growth. Kemp et al. (104)
studied fungal growth on filters of the HVAC system and
reported isolation of Aspergillus niger, A. fumigatus, Alternaria,
Cladosporium, Mucor sp., Aspergillus sp., and Penicillium sp.
However, they could only confirm growth of Aspergillus sp.,
Cladosporium sp., and Penicillium sp. when filters were directly
examined under the microscope. Buttner et al. (105) reported
a controlled study using three air duct materials (i.e., gal-
vanized metal, rigid fibrous glass ductboard, and fiberglass
duct liner) and Penicillium chrysogenum spores. They found
that fungal growth might occur on a variety of duct materials,
including bare metal, provided soiling and moisture were
present. The results showed that contaminated air ducts
might expose building occupants to high concentrations of
spores dispersed from fungal colonies growing on duct materi-
als during normal operation of the system. Yang (102) exam-
ined and cultured 1,200 fiberglass insulation liner samples
from HVAC systems in the United States and found fungal
colonization and growth in approximately 50% of the samples
studied. Fungal taxa were differentiated based on water and
humidity conditions. Species of Cladosporium and Penicillium
were primarily from areas with high RH, whereas species of
Acremonium, Aureobasidium, Exophiala, Fusarium, Paecilomy-
ces, Phoma, Rhodotorula, Sporobolomyces, and yeasts were sus-
pected in areas subjected to frequent wetting. Cladosporium
cladosporioides, C. herbarum, and C. sphaerospermum were
the primary species identified.
All fungi found to colonize building materials are saprobic
and biodeteriorating agents. Some fungi, such as species of
Chaetomium, Aspergillus, Stachybotrys, and Trichoderma, are
known to be capable of degrading cellulose fibers. Although
at least 270 fungal species have been reported from indoor
environments in the literature (10), it is very likely that any
saprobic, biodeteriogenic, or cellulolytic fungi can potentially
grow indoors if opportunity arises (4, 39).
Although species of anamorphic fungi are commonly
detected in moldy building materials, Ascotricha chartarum,
A. erinacea, Chaetomium species, Peziza spp., and Pyronema
domesticum of the Ascomycetes are occasionally found on
damp materials in buildings (11, 39, 79). The asexual state
of basidiomycetes, such as species of Cryptococcus, Rhodotor-
ula, and Sporobolomyces, are also common in indoor envi-
ronments (39) and Wallemia sebi was often reported from
buildings with dampness or water damage problems. In addi-
tion, Yang (unpublished data) has seen and identified fruiting
structures of a slime mold, Stemonitis sp., and basidiomata
(fruiting bodies of basidiomycetes) of Coprinus spp., Pleurotus,
and Poria from various building materials, from ceiling tiles to
wood products. Samson et al. (85) reported that Sistotrema
brinkmanii, a wood decay fungus, is commonly isolated from
wet, decaying window and door joinery. Mycelia and hyphae
with clamp connections, indicating basidiomycetes, are fre-
quently detected colonizing water-damaged wood structures.
Not surprisingly, wood-inhabiting basidiomycetes are often
wood decay fungi. The wood decay, dry rot fungus Serpula
lacrymans is the most common indoor basidiomycete in
central Europe, while Meruliporia incrassata pendant to S.
lacrymans also received considerable attention in North
America (106). Specimens of S. lacrymans from California
collected in nature were ascribed to var. shastensis while all
other collections, mainly from buildings in Europe, were
3.2.5-6 ▪ AIR
ascribed to var. lacrymans (107). Meruliporia incrassata
(syn. Poria incrassata) and Serpula lacrymans were detected
in building wood samples collected from the United States
by real-time PCR (83). Approximately 80 species of wood
decay fungi have been found in buildings in northern Ger-
many (82, 106). Twenty-nine species or genera of basidiomy-
cetes were identified from 3,434 decay fungi occurrences in
Norwegian houses from 2001 to 2003 (108). One hundred
fifty-two species of wood decay fungi were reported from
wood products in the United States (109). It should be
pointed out that these fungi were not isolated from an indoor
environment. The exact number of wood decay fungi from
indoor environments in the United States remains unknown.
FUNGI AND HUMAN HEALTH
Fungi are known in veterinary and human medicine to be a
source of infections, allergies, and irritant-toxic health reac-
tions primarily with symptoms and disorders of the skin,
mucous membranes, or internal organs. Current knowledge
and key concepts are summarized and discussed below (57,
110, 111). Expert reviews of reported health problems associ-
ated with building dampness and biological agents, such as
the Institute of Medicine (IOM) (2004) or the WHO-EU
(2009) and others concluded, based on their reviews mainly
of the English-language literature and epidemiological stud-
ies, that dampness-related fungi are associated with allergies,
respiratory symptoms or diseases, such as asthma, and changes
of the immunological system (112–116). In addition, there
are several clinical studies and case reports of adverse health
reactions that include primarily nonallergic adverse effects
to the lungs (bleeding in infants; allergic alveolitis), neuro-
logical system (headaches and cognitive dysfunction), endo-
crine and reproductive organs (thyroid hormonal changes
and menstrual disorders in women), and rheumatological dis-
orders ( joint pain), and an increased risk of cancer has been
also explored. Some of the fungi (maybe in combination with
bacteria) produce chemicals that are known genotoxins and
carcinogens (117–123). However, these are findings that
are difficult to document and validate in epidemiological or
experimental studies and have therefore been considered
debatable by some. Further evidence needs to be obtained.
Fungi and their by-products, such as (1–3)-ß-D-glucan,
mycotoxins, and MVOCs, have been implicated in adverse
health reactions and diseases (7,124–127). Americans spend
up to 90% of their time indoors, where contaminants often
are at higher levels than they are in the ambient air. It is
not uncommon that exposure duration and concentrations
of atypical fungi (fungi associated with excessive dampness)
are greater indoors than outdoors because of occupant life-
styles, building conditions, and materials that lead to fungal
growth and accumulation. Additionally, exposure to biologi-
cal contaminants of all kinds, but particularly molds and bac-
teria, can be high when buildings have moisture problems or
water damage (128). It is estimated that more than 1one third
of buildings in the United States and Western Europe have
severe moisture problems that result in significant fungal
contamination in indoor environments (129–131). Exposure
to high levels of indoor dampness and mold has been associ-
ated with upper and lower respiratory symptoms, including
nasal and sinus irritation, congestion and inflammation,
sore throats and chest burning sensation, cough, wheeze,
chest tightness, and exertional dyspnea in people, according
to several large epidemiological studies cited by the IOM,
WHO, and others (112, 113, 116, 132, 133). Clinical case
studies and research have shown that nonallergic health
TABLE 2 Complaints and symptoms reported by patients with
exposure to excessive fungal growth
Headaches
Runny nose or nasal congestion
Burning sensation and watery eyes
Sore throat and hoarseness
Sneezing or irritant-dry cough, chest tightness and burning chest
sensation, shortness of breath, wheezing
Unusual nosebleeds and coughing up blood (rare)
Skin and mucous membrane irritation (occasionally hair loss)
Dizziness, concentration and memory problems; cognitive
dysfunction
Severe fatigue and exhaustion ( physical and/or mental);
Nausea, (vomiting) and gastrointestinal problems (loose stools,
stomachaches)
Feverish feeling
Joint and muscle ache (flu-like reaction)
reactions are common in individuals with atypical fungal
indoor exposure. Typical health complaints of patients living
in indoor environments with excessive fungal exposure are
listed in Table 2.
The medical conditions and illnesses associated with fun-
gal indoor exposure include a spectrum of infectious, skin and
respiratory disorders, allergy, and irritant/toxic health reac-
tions (112, 114, 125, 134). Most of the reported adverse
health reactions are normally of short duration and reversible
after the exposure has been stopped or controlled. In some
cases, the adverse health consequences can be more serious
or may be irreversible, requiring symptomatic treatment and
careful avoidance of microbial triggers (134). Medical condi-
tions are listed in Table 3 and described later in more detail.
Infections
Infections caused by fungi are called mycoses. They are cate-
gorized as endemic mycoses and opportunistic mycoses.
Opportunistic fungal pathogens have a great public health
importance, especially in persons with HIV, with organ fail-
ure, and or receiving organ transplants (135, 136). Endemic
mycoses are related to the geographical distribution of certain
fungal pathogens. These types of infection are caused by the
inhalation of airborne spores or conidia found in certain
regions where there is a higher frequency of such fungi
because of unique soil and plant/flora conditions (30, 111,
137). Table 4 lists several important fungi and the infections
that occur through air transmission, the diseases they cause,
and clinical manifestations.
Opportunistic infections are secondary complications that
occur in patients with an altered or weakened immune sys-
tem. Patients at risk for fungal infections usually have major
systemic diseases or health suppressed conditions such as
complicated diabetes mellitus, cancer, HIV/AIDS, severe
liver or kidney diseases, organ transplantation, and burn
injury or may be on immune-suppressive medication treat-
ment. An endemic outbreak of fungal meningitis, spinal
infections, and other serious health complications in 2012
appeared to be caused by injecting directly into the spinal
fluid of patients a steroid medication for pain control, meth-
ylprednisolone acetate, from a compounding pharmacy
that reportedly was contaminated by Exserohilum rostratum,
Aspergillus fumigatus, Stachybotrys chartarum, Cladosporium,
 3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-7

TABLE 3 Medical conditions associated with indoor fungal exposure

Organ system
Upper airways: Nose, sinuses, and
throat
Lower airways: lung with bronchial
system and alveoli
Combined upper and lower airway
Skin and mucous membrane
Other organs: central nervous
system, immune system, liver,
kidney, endocrine system
Clinical effect Exposure
Rhinitis, sinusitis, laryngitis Fungi, allergens, irritants,
MVOCs, particles
Bronchitis, asthma, bronchiolitis, asthma, Fungi, allergens, fungal
allergic bronchopulmonary aspergillosis, by-products. Fine, ultra-fine
allergic extrinsic alveolitis (a.k.a. particles
hypersensitivity pneumonitis); toxic
alveolitis and pneumonitis
Aspergillosis Fungal rhinosinusitis Fungi, particles
Urticaria, dermatitis (allergic vs. irritant), Fungal irritants, allergens,
conjunctivitis
Overlapping diagnoses, differential Fungi, organic dusts, microbial
diagnoses (exclusion of unrelated by-products
conditions)

and several other fungi (139). The outbreak was associated
with more than 39 deaths out of 620 cases in multiple states
by mid-December 2012 (140). These iatrogenic fungal
infections were likely the result of contamination during
the medication production process in an unhygienic indoor
environment. These infections are not contagious, and the
fungi are not considered obligatory pathogens. Secondary
fungal infections and medical complications related to air-
borne fungal contamination in hospitals and transplant units
have been reported. Immunocompromised patients may be at
an increased risk for opportunistic infections if pathogenic
fungi become airborne and are significantly elevated in
indoor air surveys. Among the fungi of concern are Aspergillus
spp., such as A. fumigatus, A. flavus, and A. niger. Soil, bird
and bat droppings, water-damaged materials, or organic-rich
substrates in buildings may be a reservoir for these fungi
(39, 57, 141). Other clinically important fungal infections
are candidiasis with local mucocutaneous or disseminated
systemic organ manifestations and skin mycoses, such as
dermatophytoses, keratomycosis, tinea nigra, piedra, and
malassezia-caused dermatitis. Invasive fungal diseases of the
paranasal sinuses may also be associated with allergic sinusitis
in atopic patients (142). Aspergillus species are often
involved. Noninvasive forms may colonize preexisting body
cavities and may be asymptomatic as long as some immuno-
logical resistance can be maintained. Chronic rhinosinusitis
with eosinophilic inflammation of the airways has been
linked to dampness-related fungi from indoor environments
and may also be related to the development of asthma
(143–145). The beneficial use of antifungals (i.e., itracona-
zole) has been observed in treatment of respiratory conditions
such as chronic sinusitis and asthma, suggesting a combined
effect of infections and inflammation (146–149).
The prevention, diagnosis, and therapy of opportunistic
infections may be difficult for those who are not well trained
and experienced in this field. Early recognition, preventive
building engineering, hygiene, and public health interven-
tion can reduce the incidence of mycosis, especially the insti-
tutional or iatrogenic acquisition in care facilities.
Allergy and Respiratory Diseases
Fungi are known to cause an immune pathology with an exag-
gerated or inappropriate immune response, called hypersensi-
tivity reaction or common allergy (114). The important types
of the allergic immune reactions according to the Coomb-
Gell classification are listed in Table 5, but in the clinical
context these often occur with overlapping presentation.
The fungal spore is a known cause of allergic diseases (150–
152) and was identified as one of the major indoor allergens
(114, 153). However, there are still significant methodologi-
cal problems in the production of reliable allergen extract
from fungi compared to cats, dust mites, and other better char-
acterized allergens. Extracts that were available often corre-
sponded poorly with the fungi frequentlyfound in indoor
surveys (154). Many extracts of common indoor fungi are
not available for clinical allergy testing. Several recent epide-
miological studies have shown that long-duration indoor
exposure to certain fungi can result in hypersensitivity reac-
tion and chronic diseases. Mold levels and fungi comparable
to outside background levels and types are usually well toler-
ated by most people. Normal or “typical” indoor molds may
vary depending on climate variations and geographical

TABLE 4 Diseases transmitted by airborne fungi and affected tissuesa

Fungi Disease Affected organ/tissues
Histoplasma capsulatum Histoplasmosis Lung, eye (skin and bone)
Cryptococcus neoformans Cryptococcosis Lung, central nervous system, meninges, skin, and viscera
Coccidioides immitis Coccidioidomycosis Lung, multi-organ dissemination (skin, bone, meninges, joints)
Blastomyces dermatitidis Blastomycosis Lung, skin and mucous membrane, bone, joints
Aspergillus spp., particularly A. fumigatus Aspergillosis Lung, bronchial airways and sinus cavities, ear canal, eye (cornea)
Sporotrix schenkii Sporotrichosis Granulomatous pneumonitis (rare), skin, joints, central nervous system, eyes
Mucorales, zygomycetes Mucormycosis Nose, sinuses, eye, lung (brain and other organs), gastrointestinal system
a
Sources: 112, 138.
3.2.5-8 ▪ AIR

TABLE 5 Immunopathological responses caused by fungal hypersensitivity

Type
I: immediate hypersensitivity
III: immune complex
mediated
IV: delayed-type
hypersensitivity
Immune response
IgE, mast cell
IgG, antigen/antibody complexes deposition
in the blood vessels and tissues
Antigen-sensitized T-lymphocyte
Diseases
Asthma, rhinitis, eczema, and hay fever
Hypersensitivity pneumonitis, Arthus
reaction, extrinsic allergic alveolitis
Allergic contact dermatitis, pneumonitis

regions. However, when types of mold and levels that are
“atypical” in the indoor environment increase because of
recurrent water leaks, home dampness, and high humidity,
the prevalence of allergy and respiratory problems also rises
(29, 39, 40, 52, 155–161). Dampness and mold have, in
many epidemiological studies, been shown to be associated
with cough, environmental lung disease, and asthma (116,
162–166). Molds found on wet building materials and known
to be associated with allergy and respiratory problems are spe-
cies of Alternaria, Aspergillus, Aureobasidium, Cladosporium,
Fusarium, Paecilomyces, Phoma, Penicillium, Rhizopus, Stachy-
botrys, Trichoderma, and others (7, 79, 97, 101, 167, 168).
The prevalence of allergy to fungi among atopic patients is
estimated to be around 30% and in the general population up
to 6%. Exposure to molds during childhood is suspected to be
a risk factor for development of allergic respiratory disease
(169, 170), although a possible protective effect has been
reported concerning children’s microbial exposure in farm
environments (171–173). New onset of asthma in children
after prolonged moisture and mold exposure has been demon-
strated in a prospective study (170). However, respiratory
problems can occur in atopic and nonatopic individuals
(atopy is a genetic trait of increased allergen sensitivity).
The reported percentages of populations allergic to molds
vary from 2% to 18%. Approximately 80% of asthmatics were
reported to be allergic to molds (29). In a 2002 study, up to
35% of newly diagnosed asthma cases were attributable to
workplace mold exposure (172). A recent study in Europe
found that workplace exposure to dampness and molds is
associated with the occurrence of new-onset asthma. In
addition, exposed workers suffering from asthma-like symp-
toms are subject to an increased risk for the development of
asthma (174).
The incidence and prevalence of allergic diseases is on the
rise (114). Many patients with chronic rhinosinusitis have a
very high incidence of positive fungal cultures (up to 96%),
and it is often associated with allergic fungal sinusitis (144).
Park et al. (145) found building-related (BR) rhinosinusitis
symptoms were a risk factor for the onset and development
of BR asthma symptoms and exposure to molds in water dam-
aged buildings is an increased risk for the development of BR
asthma symptoms among the individuals with BR rhino-
sinusitis symptoms. Murr et al. (175) found that Alternaria
alternata, Cladosporium cladosporioides, Cladosporium herba-
rum, Penicillium brevicompactum, Penicillium crustosum, and
Penicillium chrysogenum were at very high concentrations
ranging from 1,451 to 2,867,839 cells equivalence/sample
in the sinus samples of some chronic rhinosinusitis (CRS)
patients. However, the Environmental Relative Moldiness
Index (ERMI) results did not show significant difference in
fungi in the dust samples between homes of CRS and
non-CRS patients.
ERMI is a DNA-based method, based on a collection
of approximately 1,000 dust samples collected from U.S.
houses, developed to screen indoor environments for molds
(176). This method provided alternative tools to the
morphology-based methods for indoor mold research and
investigation. However, ERMI is still subject to debate and
further evaluation and validation. It is questionable whether
ERMI should be applied to nonresidential environments. It
has not been widely accepted as a valid exposure marker at
present.
In clinical allergy studies, patients can be tested for specific
mold allergy using skin or serological tests (IgE-RAST), and
appropriate advice and treatment can then be prescribed. Due
to the low sensitivity of some of the commercially available
mold extract tests, false negative results are not uncommon.
Patients with an atopy are frequently allergic to multiple fun-
gal species and manifest type I reactions (Table 5). Another
practical problem is that available mold extracts at present
only cover a very small portion of molds we are exposed to
and that are common in buildings with dampness and water
damage.
Most of the fungi in bioaerosols may be allergenic depend-
ing on the exposure situation and doses (114), although the
sensitivity of clinical tests may vary with the study population
and individual immune system characteristics. Atopic indi-
viduals typically have a higher rate of positive skin reactions
after provocation tests and serological allergy tests measuring
antibody precipitins (IgE). Diseases such as allergic broncho-
pulmonary aspergillosis (138) and allergic fungal sinusitis
possibly require additional host factors which are not well
documented (144), and may be the result of a combined reac-
tion of allergenic inflammation and the immunotoxic effect
of fungal metabolites. The relevant route of exposure is inha-
lation. Fungal by-products (i.e., mycotoxins) have ciliostatic
effects in the respiratory tract (177), which can be one of the
important pathological mechanisms causing diminished
mucociliary clearing and local inflammatory effects in the air-
ways and sinuses. In general, the adverse effects of fungal
inhalation are related to duration and intensity of fungal
exposure. However, typical for allergic reactions is that once
an individual develops an allergy to certain fungi, even small
airborne concentrations can trigger an asthma attack or other
allergic reactions. This is principally different from fungal
toxic-inflammatory health reactions, which depend on air-
borne concentrations and will be similar for most people,
whether or not they are sensitized. Allergy “threshold levels”
to common mold have been reported (151), but variations in
sampling strategies and methodological limitations make
these very unreliable in practical settings (178, 179). There-
fore, the consensus is that acceptable safe threshold limits for
fungal indoor exposure cannot be established (17, 180), and it
is generally recommended to avoid or minimize unnecessary
indoor fungal exposures (181).
Although a low rate of IgE-mediated allergic responses to
the toxigenic fungi Stachybotrys chartarum have been reported
in some studies, it is unlikely a strong allergen in the clini-
cal setting, and the toxic-irritant effects appear to be more
important (182). Chun et al. (183) established a suggested
 3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-9

TABLE 6 Fungal agents of HP and occupational dust exposure

Fungal agent Sources Disease
Aspergillus clavatus Moldy malt Malt worker’s lung
Aureobasidium pullulans Steam Sauna-taker’s lung
Alternaria spp. Wood Wood worker’s lung
Botrytis cincerea Moldy fruits Winegrowers’ lung
Cryptostroma corticale Wood Maple bark stripper’s lung
Farnai rectivirgula Straw Potato riddler’s lung
Serpula (Merulius) lacrymans Moldy building Dry rot lung
Penicillium spp. Cork Suberosis, woodman’s disease
Penicillium casei Cheese Cheese worker’s lung
Mucor stolonifer Moldy paprika Paprika worker’s lung
Trichosporon cutaneum House dust Japan summer pneumonitis

threshold dose (10 μg) for S. chartarum allergy induction by
comparing the allergenicity of S. chartarum to house dust
mite extracts in a mouse model. They opined that exposure
to S. chartarum might be easily over the sensitization thresh-
old for a susceptible population in buildings with damp or
water-damaged conditions.
Nagayoshi et al. (184) reported for the first time that
inhalation exposure to the conidia of S. chartarum resulted
in the remodeling of pulmonary arteries and pulmonary
hypertension in mice. Yike and Dearborn (185) considered
it a significant step to understand the pathologic effects of
S. chartarum. Rakkestad et al. (186) demonstrated that heat-
treated S. chartarum conidia induced cell death (apoptosis)
within 3–6 h due to DNA damage.
Hypersensitivity Pneumonitis and Organic
Dust Toxic Syndrome
The clinical features, biochemistry, and pathophysiology of
allergic or inflammatory-toxic reactions to airborne microbial
exposure are difficult to separate (187, 188). Hypersensitivity
pneumonitis (HP), also called extrinsic allergic alveolitis, is a
well-recognized occupational disease. Table 6 lists several
fungi, their sources, and the occupational hypersensitivity
diseases they cause. Organic dust toxic syndrome (ODTS),
also called toxic pneumonitis, is a nonallergic, noninfectious
form of an acute inflammatory lung reaction to high fungal
dust exposure (89, 189, 190). The differences between HP
and ODTS can be difficult to distinguish. Table 7 lists com-
parative features of HP and ODTS. The significance of
ODTS in occupational health is such that preventive meas-
ures have been recommended for certain occupations
(agriculture) by the National Institute for Occupational
Safety and Health. The measures include the use of industrial
hygiene controls, special protective equipment, ventilation,
and respiratory protection (192). Although ODTS is more
likely to occur in settings where large amounts of organic
waste are handled (such as agriculture and composting
facilities), it may also happen in office and domestic environ-
ments (189).
Large-scale composting of organic waste is a new, growing
technology in municipal waste management (191). Environ-
mental monitoring suggests that this procedure involves risks
of high levels of exposure to several pathogenic fungi and
bacteria. Immunological blood changes (such as IgG anti-
body elevation) can be observed in waste-handling workers
or other occupations with high fungal exposure (193).
Cladosporium cladosporioides, Cladosporium sp., and Fusa-
rium napiforme were reported to cause HP in residences
(194–196). HP caused by fungi was recently reviewed by Sel-
man et al. (197). Bhan et al. (198) indicated that Stachybotrys
chartarum is a potential cause of HP which is TLR9-depend-
ent in mouse model.
Mycotoxins and Human Health
Fungi produce toxic chemicals, such as the poisonous
compounds found in some mushrooms and the toxic metab-
olites of some species of microfungi. Many of the mushroom
poisons are polypeptides or amino acid–derived toxins (42).
Poisoning due to ingestion of poisonous mushrooms is
excluded from the scope of this chapter, and the reader
should consult suggested references on mushroom toxins
(199, 200).

TABLE 7 Comparative features of HP and ODTSa

HP (extrinsic allergic alveolitis) OTDS (toxic pneumonitis)
Immune responses Type IV delayed hypersensitivity, cell-mediated Nonallergenic, noninfectious, lack of IgG.
immune reaction
Affected tissue/organ Lung alveoli, forming granulomas Inflammatory lung reaction
Exposure levels 106–1010 CFU/m3 of thermophilic actinomycetes or High concentrations of fungi, >109 spores/m3,
fungi >1–2 µg/m3 of endotoxins, or (1–3)-ß-D-glucan.
Clinical features Dypnea, cough, fatigue, poor appetite, weight loss, Dypnea, cough, headaches, fever, chills, malaise,
abnormal chest X ray, abnormal pulmonary acute inflammatory lung reaction, negative chest X
functions, high antibody precipitins; may cause ray; may recover after exposure cessation
pulmonary fibrosis long term
a
Source: 191.
3.2.5-10 ▪ AIR
Some fungi have been known to produce secondary
metabolites called mycotoxins that are harmful to animals
and humans (19–21), and specifically when ingested (201,
202), inhaled (5, 156, 203, 204), or in contact with the
skin (205–207). Mycotoxins in food and feed may lead to
cancer and mutagenicity, and estrogenic, gastrointestinal,
and kidney disorders. Some mycotoxins are immunosuppres-
sive and compromise the resistance to infectious disease. In
food safety mycotoxins are well recognized and regulated as
potential disease agents affecting human and animal health
(see “scientific opinion” at http://www.efsa.europa.eu/en/
topics/topic/mycotoxins.htm).
A recent study found that settled dust collected from
moisture-damaged, damp schools contained larger numbers
of microbial secondary metabolites at higher levels than the
dust samples from schools without moisture damage and
dampness (208). Mycotoxin production is species specific
(22, 209). Nielsen and Frisvad (22) pointed out the six chal-
lenges faced in mycotoxin research at present. Among the
challenges, flaws in methodology and misidentification of
mycotoxigenic fungi or using sequences generated from mis-
identified cultures had led to some questionable results,
such as false negative or positive results. The most important
one among these hurdles is expertise.
These mycotoxins belong chemically to the alkaloids,
cyclopeptides, and coumarins (3). Early reports of the delete-
rious and poisonous effects of mycotoxins on human health
goes back to the 1100s on ergotism (holy fire or St. Anthony’s
fire) caused by consumption of rye bread contaminated by the
alkaloid-containing ergot developed by Claviceps purpurea
(210). At present, more than 400 mycotoxins have been dis-
covered (207). The total number of mycotoxins remains
unknown but is believed to be in the thousands.
The effects of toxins produced by molds in humans have
been mostly described and researched in relationship to food-
borne diseases affecting animals or regional human disease
outbreaks (202, 211). The earliest known mycotoxin pro-
ducers, primarily Claviceps purpurea, produce the substance
ergot, which causes ergotism. Ergot toxins caused food poi-
soning outbreaks due to the consumption of contaminated
rye bread. The toxin was associated with bizarre behaviors
(known as “dancing plague,” “holy fire,” or “St. Vitus’s
dance”) and may have contributed to the population decline
in Europe from the fourteenth to the eighteenth centuries
(210, 212).
There has been a debate regarding the public health
importance of “toxic mold” in enclosed indoor environments
and its impact on the occupants’ health (213, 214). Aspergil-
lus versicolor and species of Penicillium, Fusarium, Trichoderma,
Cephalosporium, Chaetomium, and Stachybotrys are known to
produce naturally potent mycotoxins, depending on available
nutrients, favorable environmental conditions, or their life
cycle. Health complaints and clinical findings in patients liv-
ing or working in wet and moldy buildings often cannot be
explained as allergic reactions in otherwise healthy individu-
als. An overview of clinically important health disorders
based on various case reports and results of disease cluster
investigations are presented for the most important myco-
toxin producers (Table 8) (215–217).
The toxicological knowledge of such mycotoxins pri-
marily stems from the veterinary and food-safety science
(218). A limited number of the more than 400 known myco-
toxins have been studied and found to have important geno-
toxic, mutagenic, cytotoxic, carcinogenic, nephrotoxic,
pseudo-estrogenic, immunosuppressive, protein synthesis
inhibitor, or other toxic properties (19–21, 219, 220). The
knowledge of these adverse health effects has led internation-
ally to regulatory efforts to protect humans from excess expo-
sure in food and agricultural products based in many cases on
a “precautionary principle,” in part because the data are still
limited and definite dose-response models have not yet
been established for these agents. Toxic molds can induce
abortions and reproductive abnormalities in animals. There
is a concern among environmental health clinicians about
similar effects in humans, however adequate human studies
are lacking. The International Agency for Research on Can-
cer (IARC) classified aflatoxin produced by Aspergillus flavus,
A. niger, and A. parasiticus as having “sufficient evidence” for
human and animal liver carcinogenicity (217). Ochratoxin A
is a potent nephrotoxin with immunosuppressive, terato-
genic, and carcinogenic properties and has been classified
as a possible carcinogen to humans (85, 221). Sterigmatocys-
tin is also a carcinogen (224). Whether patulin is carcino-
genic is subject to further research. A recent study showed
the possible role of free radicals in patulin-mediated dermal
tumorigenicity involving mitogen-activated protein kinase
(222). Fumonisins (B1 and B2) are cancer-promoting metab-
olites (223, 224). Many of the other mycotoxins have not
been classified as carcinogens, but carcinogenicity cannot
be ruled out due to a lack of appropriate studies.
Human cases of true mycotoxicosis appear to be rare and
in the past were thought to be mostly related to ingestion of
contaminated grain products. However, possible occupa-
tional or environmental inhalation exposures have been
described in recent case studies and epidemiological surveys.
Typical nonallergic symptoms of patients in which myco-
toxin exposure was either confirmed or highly suspected are
recurrent cold and flu–like symptoms, extreme fatigue, con-
stant sore throat or skin irritation, severe and unusual head-
aches, neuromuscular and neurocognitive dysfunction
(tremor and shakes, unusual memory and concentration prob-
lems), bleeding disorders of the lung in infants, irregular
menses, diarrhea, dermatitis and irritation of skin, and
impaired immune function (23, 225–229). Mycotoxins may
also be involved in occupational diseases and respiratory can-
cers among food and grain workers (230–232). Environmen-
tal sentinel investigations in water-damaged buildings have
shown detectable levels of airborne mycotoxins from Stachy-
botrys chartarum and others that may be of concern (233–
236). This is important in clean-up and remediation projects,
as the removal of toxins from water-damaged and moldy
building materials is difficult (237). Further studies are
needed to improve our understanding of mycotoxins found
in the indoor environment and possible adverse human
health effects.
Some mycotoxins, such as lysergic acid, are derivatives of
amino acids (such as tryptophan). Others derived from other
precursors are grouped into aromatic and phenolic–related
toxins and terpenoid toxins. Some well-known and potent
mycotoxins in the aromatic and phenolic–related toxin group
are aflatoxin, zearalenone, and griseofulvin. The terpenoid
toxins include trichothecenes and fusidanes (3). There
were more than 200 mycotoxins produced by a variety of com-
mon fungi according to the WHO Environmental Health
Criteria 105 on mycotoxins, published in 1990 (238). Sam-
son (127) suggested that there were more than 400 toxic
metabolites in 1992. It is likely that the number of recorded
toxic metabolites will increase over time because of new dis-
coveries (216, 239, 240). These alcohol- and water-soluble
toxins can be attached to spores, mycelia, or dust particles
and are sufficiently small in size (2–10 micron) to be inhaled
into the human lungs. Some mycotoxins are lipid soluble and
 3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-11

TABLE 8 Some toxigenic fungi, fungal chemical metabolites, and their health effects
Fungi Chemical metabolites Health effects
Penicillium (>200 species) Patulin Hemorrhage of lung, brain disease
citrinin Renal damage, vasodilatation, bronchial constriction,
increased muscular tone
ochratoxin A Nephrotoxic, hepatotoxic
citroviridin Neurotoxic
emodin Reduced cellular oxygen uptake
gliotoxin Lung disease
verruculogen Neurotoxic: trembling in animal
secalonic acid D Lung, teratogenic in rodents
Aspergillus spp. Patulin Hemorrhage of lung, brain disease
A. clavatus Aflatoxin B1 Liver cancer, resp. cancer, cytochrome P-450
monooxygenase disorder.
A. flavus and A. parasiticus Sterigmatocystin Carcinogen
A. versicolor ochraceus Ochratoxin A Nephrotoxic, hepatotoxic
Stachybotrys chartarum, Trichothecenesa (more than 50 derivatives Immune suppression and dysfunction, cytotoxic,
Fusarium species, known): T-2, nivalenol, deoxynivalenol, bleeding, dermal necrosis; high dose ingestion = lethal
Trichoderma species diacetoxyscirpenol, satratoxin H, G, (human case reports); low dose, chronic = potentially
other macrocyclic trichothecenes lethal); teratogenic, abortogenic (in animals).
Spirolactone Hemorrhage
Zearalenone “Alimentary toxic aleukia” (ATA) reported in Russia
and Siberia
“Staggering wheat” in Siberia
“Red mold disease” in Japan
Neurotoxic/nervous behavior abnormality
Co-carcinogen/chemotoxic (?)
Anticomplement function.
Phytoestrogen may alter immune function; stimulates
growth of uterus and vulva, atrophy of ovary
Claviceps spp. Ergot alkaloids Prolactin inhibitor, vascular constriction, uterus
contraction promoted.
a
Trichothecenes are also produced by Myrothecium, Trichothecium, and Gibberella (teleomorph of some Fusarium species).

may be absorbed via skin. Building materials contaminated
with the above-mentioned fungi have been shown to produce
detectable levels of mycotoxins on the materials (97, 167,
168, 241).
Aflatoxin may be involved in occupational respiratory
cancers among food and grain workers (242). Aflatoxin-
induced disease has been well reviewed (217, 243, 244). Tri-
chothecene toxins (T-2 toxin, Fusarium toxins) are listed
with “limited evidence” for animals and “inadequate evi-
dence” (no data available) for humans (217). Macrocyclic
trichothecenes, such as satratoxin H, have not been classified.
Epidemiological studies suggest a higher rate of upper respira-
tory tract and lung cancer in workers in the grain and food-
handling industry with high fungal product inhalation risk
(230). The high rate of lung cancer among uranium miners
in Slesien (Schneeberg disease), may be related to combined
effects of high radon and Aspergillus exposure in the under-
ground mines (244).
Research has shown that water-damaged building materi-
als are often contaminated with fungi that produce detectable
levels of mycotoxins (157, 167, 245, 246), which may
become airborne and further contribute to indoor air pollu-
tion (101, 241). From a public health point of view, probable
important toxigenic fungi are Aspergillus species, Penicillium
species, Fusarium species, S. chartarum (syn. S. atra), Paecilo-
myces species, and Trichoderma species. These fungi have
been associated with adverse health effects in humans and
animals resulting in typical organ damage and disease, which
is often neither allergic nor infectious in nature. In several dis-
ease outbreaks, human and animal death has been linked to
exposure to toxigenic fungi, typically through ingestion. Cur-
rent research, however, indicates that inhalation of certain
mycotoxins has even stronger effects (247) and may be fre-
quently associated with health complaints and human dis-
ease. The occurrence of mycotoxins in the environment,
their chemistry and related adverse health effects have been
reviewed (5, 20, 21, 127, 206, 216, 248). Several mycotoxins,
even in low concentrations, were observed to be cytotoxic,
interfere with DNA and RNA synthesis, inhibit protein syn-
thesis, and cause apoptosis of cells of different body organs.
These toxic effects may cause a variety of short-term as well
as long-term adverse health effects in animals and humans
(21, 168, 248–250). Samson (127) divided the effects into
four basic categories: acute, chronic, mutagenic, and terato-
genic. Symptoms thought to be due to mycotoxins or toxin-
containing spores ( particularly those of S. chartarum) include
dermatitis, recurring cold and flu–like symptoms, burning
sore throat, headaches and excessive fatigue, diarrhea, and
impaired or altered immune function (206). The ability of
the body to resist infectious diseases may be weakened, result-
ing in opportunistic infections. Certain mycotoxins, such as
zearalenone, have been found to cause infertility and still-
births in pigs (201). Low-level, complex exposures from a
mixture of mycotoxins, as would be typically encountered
in real-life situations, may have synergistic effects, which
may result in central neuroendocrine-immune changes and
3.2.5-12 ▪ AIR
consequently in complex effects of the endocrine and nerv-
ous system (121).
Nonallergic complaints from patients in which myco-
toxin-producing fungal exposure was either confirmed or
strongly suggested include recurrent cold and flu–like symp-
toms, extreme fatigue, constant sore throat or skin irritation,
severe and unusual headaches, neuromuscular and neurocog-
nitive dysfunction (tremor and shakes, unusual memory and
concentration problems), bleeding disorders of the lung in
infants, irregular menses, diarrhea, dermatitis and irritation
of skin, and impaired immune function. Mycotoxins may
also be involved in occupational respiratory cancers among
food and grain workers (230), however, typical home or office
indoor environments have not been studied. Better con-
trolled studies are needed to improve our understanding of
mycotoxins found in the indoor environment and possible
adverse human health effects.
Historically, mycotoxins have been a problem to farmers
and food industries and in Eastern European and developing
countries (251). The large-dose exposure to fungi and myco-
toxins encountered by farmers and in food industries was gen-
erally considered unlikely to occur in nonfarming activities.
However, many toxigenic fungi, such as S. chartarum and spe-
cies of Aspergillus, Penicillium, and Fusarium have been found
to infest buildings with known indoor air and building-related
problems and illnesses (29, 156, 203, 252). It has been sug-
gested that inhalation exposure to mycotoxin-containing
fungal spores is significant in the reported cases of building-
related mycotoxicoses (252). Croft et al. (203) reported sev-
eral cases of mycotoxicoses caused by airborne exposure to the
toxigenic fungus S. chartarum in a residential building. Addi-
tional cases of office building–associated Stachybotrys myco-
toxicosis were reported by Hodgson et al. (226), Johanning
(253), and Johanning et al. (156). Satratoxin H was detected
in the fungus isolated from the contaminated building. Hem-
orrhagic lung disease in infants was highly associated with
indoor S. chartarum exposure in a case cluster investigation
in Cleveland (254) and in a case-home investigation in
the U.S. Midwest (255). Subsequently consultants for the
Centers for Disease Control and Prevention called for more
research to prove the causal relationship of Stachybotrys
and idiopathic pulmonary hemorrhage in infants on review
of the Cleveland study (256). After more than 10 years,
more cases were reported, increasing from 9 cases in the orig-
inal study to 52 cases at present. Among the cases, 91% of
patients were living in residences with S. chartarum (257).
Studies with toxic Stachybotrys fungi in mice showed similar
effects (inflammation and hemorrhage) (258). S. chartarum
was isolated from brochoalveolar lavage fluid of a child
with pulmonary hemorrhage (259), and S. atra exposure
was found in an infant who developed laryngeal spasm and
hemorrhage during general anesthesia (260). In an epide-
miological study a high prevalence of pulmonary diseases
of office workers in Florida court buildings were reported
after prolonged indoor exposure to primarily S. chartarum
and A. versicolor (226).
Mycotoxins generally have low volatility; therefore, inha-
lation of volatile mycotoxins is not very likely (206). Rather,
the toxins are an integral part of the fungus. Sorenson et al.
(261) demonstrated in the laboratory that aerosolized conidia
of S. atra contained trichothecene mycotoxins. The most
common toxin was satratoxin H. Lesser amounts of satratoxin
G and trichoverrols A and B were also detected but less fre-
quently. They also found that most of the airborne particles
were within respirable range. Similar experiments, conducted
by Pasanen et al. (262), demonstrated that trichothecene
mycotoxins were in airborne fungal propagules of S. atra
and could be collected on membrane filters. Conidia of
A. flavus and A. parasiticus were reported to contain aflatoxins
(263). Miller (5) also reported detection of two mycotoxins,
deoxynivalenol and T-2 toxin, in conidia of Fusarium grami-
nearum and F. sporotrichioides, respectively. These references
suggest that inhalation exposure to conidia may also increase
the chance of exposure to mycotoxins. Corey et al. (123)
found that satratoxin G from S. chartarum induced rhinitis
and apoptosis of olfactory sensory neurons in the nasal airways
of rhesus monkeys.
Although relationships were established to link inhalation
exposure to mycotoxin-containing fungal spores and symp-
toms of mycotoxicoses in fungi-infested indoor environments
(155, 203, 261, 262), other possible exposure routes such as
ingestion and skin contact are likely. Because fungal spores
are ubiquitous in a contaminated environment, the chance
of ingesting toxin-containing spores is likely to increase
through eating, drinking, and smoking. It is prudent to limit
exposure to such potent toxic chemicals (124), particularly
when significant fungal growth and amplification is found
indoors (134).
The toxigenic fungus frequently detected in “problem
buildings” is S. chartarum, which produces a series of potent
cytotoxins (trichothecenes, satratoxins, and spirolactones)
as well as a variety of other compounds affecting the immune
system (240, 264, 265). Case studies of health effects and
immunological laboratory changes related to indoor exposure
to trichothecenes and possibly other mycotoxins, disorders of
the respiratory and central nervous system were noted (215,
266, 267). Abnormal test results of the cellular and humoral
immune system were found (152). In earlier cases in Eastern
Europe, typically in an agricultural setting, marked leukope-
nia or acute “radiation-mimetic” effects on the blood cell sys-
tem with subsequent sepsis-like opportunistic infections after
trichothecene ingestion were reported (268, 269).
Trichothecenes are considered to be the most potent
small molecule inhibitor of protein synthesis, acting through
inhibition of the peptidyl transferase activity (219, 270).
These toxins can cause alveolar macrophage defects and
may affect phagocytosis. They have been investigated for
use in cancer treatment (271), but also in chemical-biological
warfare. The presence of fungal chemical metabolites has
been reported in several cases of animal and human ingestion-
related mycotoxicosis, resulting sometimes in death (252,
272). Mycotoxins, such as satratoxin H of the trichothecene
group, have been shown to cause depressed T or B lymphocyte
activity, suppressed immunoglobulin and antibody produc-
tion, reduced complement or interferon activity, and
impaired macrophage-effector cell function of human neutro-
phils (20).
Laboratory changes of immunoglobulins (IgA, IgE, IgG,
and IgM) in workers handling mycotoxin-contaminated
foodstuffs, primarily deoxynivalenol (vomitoxin), have
been reported (273). An increase of IgA production and
IgA nephropathy and a decrease of IgG and IgM after inges-
tion of vomitoxin were reported in a mice experiment (274).
Renal failure and IgG deposition in the glomeruli after inha-
lation of ochratoxin produced by Aspergillus ochraceus was
found in the case of a farmer (275).
A WHO task group concluded that an association
between trichothecene exposure and human disease episodes
is possible; however, only limited data are available (238).
Immunotoxicological effects principally depend on the expo-
sure conditions, dose, and timing. Some immunological
effects may only be transient, of short duration, and difficult

to detect in routine medical tests. Medical findings are often
nonspecific and other systemic diseases or causes need to be
ruled out by the experienced clinician. The treating physician
often does not recognize mycotoxicosis, especially because
exposure circumstances and presence of certain mycotoxins
are unknown. Advanced fungal exposure characterization
and sampling techniques now available should improve the
chances for better medical detection of mycotoxicosis.
Analytical methods involving immunoassays and cell line
cytotoxicity analysis are able to provide relatively rapid and
easy screening tests to detect the presence of mycotoxins in
fungal-contaminated materials (157, 250, 276).
Mycotoxin research faces methodological challenges,
such as misidentification of mycotoxigenic fungi or using
sequences generated from misidentified cultures (22). This
may led to erroneous results, such as false negative or positive
results.
Volatile Organic Compounds Produced by Fungi
Fungi in active growth produce VOCs also known as
MVOCs, which typically are noticed as a musty, moldy
odor. Over 200 MVOC compounds have been identified
from different fungi (277). Indoor measured VOC levels,
however, are typically low and any serious health risks are
uncertain (278, 279). Possibly, related mucous membrane
and olfactory irritations may trigger an “unpleasant odor reac-
tion” and annoyance. Measurement of VOCs may be an indi-
cator of excessive indoor fungal growth (280). A number of
VOCs have been identified from fungi common in indoor
contamination. Most of these fungal VOCs are derivatives
of alcohols, aldehydes, amines, ketones, terpenes, esters,
hydrocarbons, aromatics, and sulfur-containing compounds
(281, 282). The in vitro production of fungal volatiles from
47 Penicillium taxa were made up of alcohols, ketones, esters,
small alkenes, monterpenes, sesquiterpenes, and aromates
(283). However, aldehydes were not detected.
Some of the fungal VOCs have an unpleasant odor (4),
and other fungi (such as mushrooms) produce VOCs of pleas-
ant odors and flavors. 1-Octen-3-ol, one of the major fungal
VOCs, has a characteristic mushroom odor. The musty,
moldy, and earthy odors are likely to come from 2-octen-1-ol
and geosmin (1,10-dimethyl-9 decalol) (4, 29). Ezeonu et al.
(27) identified ethanol, 2-ethyl hexanol, cyclohexane, and
benzene from fiberglass air duct liners colonized by
A. versicolor, Acremonium obclavatum, and Cladosporium her-
barum. Acetone and 2-butanone were only detected on agar
plate samples of A. versicolor and A. obclavatum. Other fungal
VOCs associated with two common indoor fungi, Penicillium
and Aspergillus, have been identified. They are 2-methyl-
isoborneol, 2-methyl-1-propanol, 3-methyl-1-butanol, and
3-octanone (4). Additional fungal VOCs are compiled and
listed by Ammann (25) and Batterman (26). Almost all of
the published information regarding fungal VOC’s concerns
Penicillium spp. and Aspergillus spp. Little is known about
VOCs of other common indoor fungal contaminants. A
recent study found MVOCs: anisole, 3-octanone, styrene,
3-methyl-anisole and 4-methyl-anisole emitted by S. charta-
rum (284).
CONCLUSIONS
Fungi, fungal spores, hyphae, and by-products are ubiquitous
in nature and in indoor environments. Although fungal
growth is found throughout nature, dampness and molds
should not be allowed in indoor environments such as homes,
offices, public buildings, and health care facilities. Fungi in
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-13
damp and water-damaged buildings play an important role
in public health and disease prevention. There is a growing
consensus among experts that fungi associated with dampness
leads to preventable health problems, primarily of the respira-
tory organs and allergies. They are allergenic and irritant/
toxic agents that typically cause or aggravate airways, or are
associated with infectious and mainly respiratory diseases in
exposed people. Research findings indicate that they are a
major problem in buildings where moisture control is poor
or where water intrusion is common. Human exposures in
these situations typically are a mixture of different fungi and
bacteria. Synergistic inhalation effects of fungal by-products,
such as mycotoxins in fungal spores, β-glucans, or likely fun-
gal MVOCs released into the surroundings are potentially
irritating, toxic, teratogenic, carcinogenic, and immune-
suppressive. Clinical diagnoses of mold allergies and fungal
infections are generally easier and less complicated than
emerging health concerns of such fungal metabolites as
(1-3)-ß-D-glucan, airborne mycotoxins, and fungal MVOCs.
Furthermore, risk assessment of human exposure to these
fungi and their by-products is complex, because multiple
agents, hypersensitivity reactions, and different disease out-
comes are involved. Human susceptibility to them varies
from individual to individual. Some of the health implica-
tions from inhalation exposure of fungi are undergoing further
research, particularly at low exposure concentrations that are
likely different from agricultural settings studied in the past.
In most cases, diligent exposure cessation and control leads
to symptom reversal and health improvement. Little is known
concerning the consequences of short-term and long-term
environmental exposures to mycotoxins and whether all of
the health effects are reversible. However, based on what is
known at present regarding fungal exposures and potential
adverse health effects, it is prudent to avoid or minimize expo-
sure to infectious, allergenic, and toxic fungi and to control
indoor growth conditions. Furthermore, fungal growth
indoors suggests water infiltration and damage to building
structures and material, that ought to be corrected to avoid
further decay.
REFERENCES
1. Alexopoulos CJ, Mims CM, Blackwell M. 1996. Introduc-
tory Mycology, p. 4. John Wiley & Sons, New York.
2. Carroll GC. 1992. The Fungal Community: its Organization
and Role in the Ecosystem. CRC Press, Boca Raton, FL.
3. Griffin DH. 1996. Fungal Physiology. Wiley-Liss,
New York.
4. Gravesen S, Frisvad JC, Samson RA. 1994. Microfungi.
Munksgaard International Publishers, Copenhagen.
5. Miller JD. 1992. Fungi as contaminants in indoor air.
Atmos Environ A Gen Top 26(12):2163–2172.
6. Miller JD. 1992. Fungi and the Building Engineer.
7. Samson RA, Flannigan B, Flannigan M, Verhoeff A,
Adan O, Hoekstra E. 1994. Health Implications of Fungi
in Indoor Environments Elsevier Science, Amsterdam,
Netherlands.
8. Samson RA. 2011. Ecology and general characteristics of
indoor fungi, pp. 101–116. In Adan OC, Samson RA
(eds), Fundamentals of Mold Growth in Indoor Environments
and Strategies for Healthy Living. Wageningen Academic
Press, Wageningen.
9. Miller J, Day J. 1997. Indoor mold exposures: epidemiol-
ogy, consequences and immune therapy. Can J Allergy
Clin Immunol 2:25–32.
10. Miller JD. 2011. Health effects from mold and dampness
in housing in western societies: early epidemiology studies
3.2.5-14 ▪ AIR
and barriers to further progress, p. 183–210. In Adan OC,
Samson RA (eds), Fundamentals of Mold Growth in Indoor
Environments and Strategies for Healthy Living. Wageningen
Academic Press, Wageningen.
11. Li DW, Yang CS. 2004. Notes on indoor fungi I: new
records and noteworthy fungi from indoor environments.
Mycotaxon 89(2):473–488.
12. Kendrick B. 2001. The Fifth Kingdom. Mycologue Publish-
ing, Sydney, British Columbia.
13. O’Driscoll B, Powell G, Chew F, Niven R, Miles J, Vyas
A, Denning D. 2009. Comparison of skin prick tests with
specific serum immunoglobulin E in the diagnosis of fungal
sensitization in patients with severe asthma. Clin Exper
Allergy 39(11):1677–1683.
14. Denning DW, O’Driscoll BR, Powell G, Chew F, Athe-
rton GT, Vyas A, Miles J, Morris J, Niven RM. 2009.
Randomized controlled trial of oral antifungal treatment
for severe asthma with fungal sensitization the fungal
asthma sensitization trial (FAST) study. Am J Respir Crit
Care Med 179(1):11–18.
15. Ogawa H, Fujimura M, Takeuchi Y, Makimura K. 2012.
Role of fungal colonization for sensitization in asthma. Clin
Exper Allergy 42(10):1540–1541.
16. Cole RJ, Cox RH. 1981. Handbook of Toxic Fungal Metab-
olites. Academic Press, New York.
17. Macher J. 1999. Bioaerosols: Assessment and Control. Amer-
ican Conference of Governmental Industrial Hygienists
(ACGIH), Cincinnati, OH.
18. Korpi A, Järnberg J, Pasanen A-L. 2009. Microbial
volatile organic compounds. Crit Rev Toxicol 39(2):
139–193.
19. Betina V. 1989. Mycotoxins. Chemical, Biological and Envi-
ronmental Aspects. Elsevier, Amsterdam.
20. Wyllie TD, Morehouse LG. 1977. Mycotoxic Fungi and
Chemistry of Mycotoxins. Marcel Dekker, New York.
21. Sharma RP, Salunkhe DK. 1991. Mycotoxins and Phytoa-
lexins. CRC Press, Boca Raton, FL.
22. Nielsen KF, Frisvad JC. 2011. Mycotoxins on building
materials, p. 245–275. In Adan O, Samson R (eds), Funda-
mentals of Mold Growth in Indoor Environments and Strategies
for Healthy Living. Wageningen Academic Publishers,
Wageningen.
23. Johanning E, Biagini R, Hull DL, Morey P, Jarvis B,
Landsbergis P. 1996. Health and immunology study fol-
lowing exposure to toxigenic fungi (Stachybotrys chartarum)
in a water-damaged office environment. Int Arch Occup
Environ Health 68(4):207–218.
24. Moularat S, Hulin M, Robine E, Annesi-Maesano I, Cail-
laud D. 2011. Airborne fungal volatile organic compounds
in rural and urban dwellings: detection of mould contami-
nation in 94 homes determined by visual inspection and
airborne fungal volatile organic compounds method. Sci
Tot Environ 409(11):2005–2009.
25. Ammann H. 1999. Microbial volatile organic compounds.
26. Batterman SA. 1995. Sampling and analysis of biolog-
ical volatile organic compounds, p. 249–268. In Cox C,
Wathes C (eds), Bioaerosols Handbook. CRC Press, Inc.,
Boca Raton, Florida.
27. Ezeonu I, Price D, Simmons R, Crow S, Ahearn D. 1994.
Fungal production of volatiles during growth on fiberglass.
App Environ Microbiol 60(11):4172–4173.
28. Gregory P. 1973. The Microbiology of the Atmosphere.
Plant Science Monographs 377.
29. Flannigan B, McCabe E, McGarry F. 1991. Allergenic
and toxigenic micro-organisms in houses. J Appl Bacteriol
70(S):61S–73S.
30. Lacey J. 1981. The aerobiology of conidial fungi, p.
373–416. In Cole GT, Kendrick B (eds), Biology of Conidial
Fungi, vol 1. Academic Press, New York.
31. Lacey J. 1991. Aerobiology and health: the role of airborne
fungal spores in respiratory disease, p. 157–185. In Frontiers
in Mycology CAB International, Oxon, UK.
32. Levetin E. 1991. Identification and concentration of air-
borne basidiospores. Grana 30(1):123–128.
33. Levetin E. 1995. Fungi, p. 87–120. In Burge H (ed), Bioaer-
osols. Lewis Publishers, Boca Raton, FL.
34. Dotson KB, Patton LE, Ryan TJ, Throckmorton JV,
Weekes DM. 2004. Assessment, Remediation, and Post-
Remediation Verification of Mold in Buildings, Assess-
ment, Remediation, and Post-Remediation Verification
of Mold in Buildings. AIHA, Fairfax, VA.
35. Prezant B, Weekes DM, Miller JD. 2008. Recognition,
Evaluation, and Control of Indoor Mold. AIHA, Fairfax,
VA.
36. Dillon HK, Heinsohn PA, Miller JD. 2005. Field Guide
for the Determination of Biological Contaminants in Envi-
ronmental Samples. AIHA, Fairfax, VA.
37. Nagarajan S, Singh D. 1990. Long-distance dispersion of
rust pathogens. Annu Rev Phytopathol 28(1):139–153.
38. Asai G. 1960. Intra-and inter-regional movement of uredo-
spores of black stem rust in the upper Mississippi river val-
ley. Phytopathology 50(7):535–541.
39. Burge HA. 1990. The fungi, p. 136–162. In Morey P,
Feeley J, Otten J (eds). Biological Contaminants in Indoor
Environments. American Society for Testing and Materials,
Philadelphia.
40. Burge H. 1990. Bioaerosols: prevalence and health effects
in the indoor environment. J Allergy Clin Immunol 86(5):
687–701.
41. Li DW, Yang CS. 2004. Fungal contamination as a major
contributor to sick building syndrome. Adv Appl Microbiol
55:31–112.
42. Yang CS. 1995. Understanding the biology of fungi found
indoors, p. 131–137. In Johanning E, Yang CS (eds), Fungi
and Bacteria in Indoor Environments: Health Effects, Detec-
tion and Remediation. Eastern New York Occupational
Health Program, Albany, NY.
43. Morrison J, Yang C, Lin KT, Haugland R, Neely A, Ves-
per S. 2004. Monitoring Aspergillus species by quantitative
PCR during construction of a multi-storey hospital build-
ing. J Hosp Infect 57(1):85–87.
44. Noble W, Clayton YM. 1963. Fungi in the air of hospital
wards. Jf Gen Microbiol 32(3):397–402.
45. Solomon WR, Burge HP, Boise JR. 1978. Airborne Asper-
gillus fumigatus levels outside and within a large clinical
center. J Allergy Clin Immunol 62(1):56–60.
46. Streifel A, Rhame F. 1993. Hospital air filamentous fungal
spore and particle counts in a specially designed hospital.
Proc Indoor Air 4:161–163.
47. Brun C, Miron D, SILLA L, Pasqualotto A. 2012. Fungal
spore concentrations in two haematopoietic stem cell
transplantation (HSCT) units containing distinct air con-
trol systems. Epidemiol Infect 1(1):1–5.
48. Okten S, Asan A. 2012. Airborne fungi and bacteria in
indoor and outdoor environment of the Pediatric Unit of
Edirne Government Hospital. Environ Monit Assess 184
(3):1739–1751.
49. Burge HA. 1995. Bioaerosols in the residential environ-
ment, p. 579–597. In Cox C, Wathes C (eds), Bioaerosols
Gandbook, CRC Press, Boca Raton, FL.
50. DeKoster JA, Thorne PS. 1995. Bioaerosol concentra-
tions in noncomplaint, complaint, and intervention homes
in the Midwest. Am Ind Hyg Assoc 56(6):573–580.
51. Macher JM, Huang FY, Flores M. 1991. A two-year study
of microbiological indoor air quality in a new apartment.
Arch Environ Health 46(1):25–29.
52. Strachan DP, Flannigan B, McCabe EM, McGarry F.
1990. Quantification of airborne moulds in the homes

of children with and without wheeze. Thorax 45(5):
382–387.
53. Verhoeff A, Wijnen JH, Boleij J, Brunekreef B, Reenen-
Hoekstra ES, Samson R. 1990. Enumeration and identifi-
cation of airborne viable mould propagules in houses.
Allergy 45(4):275–284.
54. Li DW, Kendrick B. 1995. A year-round comparison
of fungal spores in indoor and outdoor air. Mycologia
190–195.
55. Tischer CG, Heinrich J. 2013. Exposure assessment of
residential mould, fungi and microbial components in rela-
tion to children’s health: Achievements and challenges.
Int J Hyg Environ Health 216(2):109–114.
56. Dungy C, Kozak P, Gallup J, Galant S. 1986. Aeroaller-
gen exposure in the elementary school setting. Ann Allergy
56(3):218–221.
57. Levetin E, Shaughnessy R, Fisher E, Ligman B, Harrison
J, Brennan T. 1995. Indoor air quality in schools: exposure
to fungal allergens. Aerobiologia 11(1):27–34.
58. Hussin NHM, Sann LM, Shamsudin MN, Hashim Z.
2011. Characterization of bacteria and fungi bioaerosol in
the indoor air of selected primary schools in Malaysia.
Indoor Built Environ 20(6):607–617.
59. Morey P, Hodgson M, Sorenson W, Kullman G, Rhodes
W, Visvesvara G. 1984. Environmental studies in moldy
office buildings: biological agents, sources and preventive
measures. ACGIH 10(1):21–35.
60. Bluyssen PM, Aries M, van Dommelen P. 2011. Comfort
of workers in office buildings: the European HOPE project.
Build Environ 46(1):280–288.
61. Burge HA. 1993. Characterization of Bioaerosols in Build-
ings in the United States. American Society of Heating,
Refrigerating, and Air-conditioning Engineers, Atlanta,
GA.
62. Llewellyn GC, O’Rear CE. 1990. Biodeterioration Research
3: Mycotoxins, Biotoxins, Wood Decay, Air Quality, Cultural
Properties, General Biodeterioration, and Degradation. Ple-
num Press, New York.
63. Morgan-Jones G, Jacobsen B. 1988. Notes on hyphomy-
cetes. LVIII. Some dematiaceous taxa, including two
undescribed species of Cladosporium, associated with biode-
terioration of carpet, plaster and wallpaper. Mycotaxon 32:
223.
64. Seifert KA, Frisvad JC. 2000. Penicillium on solid wood
products, p. 285–298. In Samson RA, Pitt JI (eds), Integra-
tion of modern taxonomic methods from Penicillium and
Aspergillus classification. Harwood Academic Publishers,
Amsterdam, The Netherlands.
65. Zabel RA, Morrell JJ. 1992. Wood Microbiology: Decay and
its Prevention. Academic Press, New York.
66. Hocking AD, Miscamble BF. 1995. Water relations of
some Zygomycetes isolated from food. Mycol Res 99(9):
1113–1118.
67. Smith D, Onions AHS. 1994. The Preservation and Main-
tenance of Living Fungi, 2nd ed. CAB International, Wall-
ingford, Oxon, UK.
68. Troller J, Scott V. 1992. Measurement of water activity
(aw) and acidity, p. 135–151. In Vanderzant C, Splitt-
stoesser DF (eds), Compendium of Methods for the Microbio-
logical Examination of Foods. American Public Health
Association, Washington, DC.
69. Adan OCG, Huinink HP, Bekker M. 2011. Water rela-
tions of fungi in indoor environments, p. 41–65. In Adan
OC, Samson RA (eds), Fundamentals of mold growth in
indoor environments and strategies for healthy living. Wagenin-
gen Academic Publishers, Wageningen, The Netherlands.
70. Huinink HP, Adan OCG. 2011. Fungal growth and
humidity fluctuations: a toy model, p. 67–82. In Adan
OCG, Samson R (eds), Fundamentals of mold growth in
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-15
indoor environments and strategies for healthy living. Wagenin-
gen Academic Publishers, Wageningen, The Netherlands.
71. Ingold CT. 1965. Spore liberation. Clarendon Press,
Oxford, UK.
72. Ingold CT. 1971. Their liberation and dispersal. Fungal
spores, Clarendon Press, Oxford, UK.
73. Li DW. 2011. Five trillion basidiospores in a fruiting body
of Calvatia gigantea. Mycosphere 2(4):457–462.
74. Benaim-Pinto C. 1992. Sensitization to basidiomycetes
and to Fuligo septica (Myxomycete) in Venezuelan atopic
patients suffering from respiratory allergy. J Allergy Clin
Immunol (89):282–289.
75. Butcher BT, O’Neil CE, Reed MA, Altman LC, Lopez
M, Lehrer SB. 1987. Basidiomycete allergy: measurement
of spore-specific IgE antibodies. J Allergy Clin Immunol 80
(6):803–809.
76. Davis W, Horner W, Salvaggio J, Lehrer S. 1988. Basidio-
spore allergens: analysis of Coprinus quadrifidus spore, cap,
and stalk extracts. Clin Exper Allergy 18(3):261–267.
77. Giannini E, Northey W, Leathers C. 1975. The allergenic
significance of certain fungi rarely reported as allergens.
Ann Allergy 35(6):372.
78. Lehrer SB, Lopez M, Butcher BT, Olson J, Reed M, Sal-
vaggio JE. 1986. Basidiomycete mycelia and spore-allergen
extracts: skin test reactivity in adults with symptoms of res-
piratory allergy. J Allergy Clin Immunol 78(3):478–485.
79. Santilli J, Rockwell W, Collins R. 1985. The significance
of the spores of the basidiomycetes (mushrooms and their
allies) in bronchial asthma and allergic rhinitis. Ann Allergy
55(3):469–71.
80. Rivera-Mariani FE, Nazario-Jiménez S, López-Malpica F,
Bolaños-Rosero B. 2011. Skin test reactivity of allergic
subjects to basidiomycetes’ crude extracts in a tropical envi-
ronment. Med Mycol 49(8):887–891.
81. Singh J. 1994. Nature and extent of deterioration in build-
ings due to fungi, p. 34–53. In Singh J (ed), Building Mycol-
ogy. Chapman and Hall, London.
82. Schmidt O, Huckfeldt T. 2011. Characteristics and iden-
tification of indoor wood-decaying basidiomycetes, p.
117–180. In Adan OC Samson RA (eds), Fundamentals
of Mold Growth in Indoor Environments and Strategies for
Healthy Living. Wageningen Academic Press, Wageningen.
83. Lin K-T, Li D-W, Denis DA, Woodcock R, Yang CS.
2005. Qualitative identification of Meruliporia incrassata
using real time polymerase chain reaction (PCR), p.
335–342. In Johanning E (ed), Bioaerosols, Fungi, Bacteria,
Mycotoxins and Human Health: Patho-physiology, Clinical
Effects, Exposure Assessment, Prevention and Control in
Indoor Environments and Work. Fungal Research Group
Foundation, Albany, NY.
84. Smith EG. 1990. Sampling and Identifying Allergenic Pollens
and Molds. Blewstone Press, San Antonio, TX.
85. Samson R, Houbraken J, Summerbell R, Flannigan B,
Miller J. 2001. Common and important species of fungi
and actinomycetes in indoor environments, p. 287–292.
In Flannigan B, Samson RA, Miller JD (eds), Microogran-
isms in Home and Indoor Work Environments. Taylor & Fran-
cis, London.
86. Rivera-Mariani FE, Nazario-Jiménez S, López-Malpica F,
Bolaños-Rosero B. 2011. Sensitization to airborne asco-
spores, basidiospores, and fungal fragments in allergic rhini-
tis and asthmatic subjects in san juan, puerto rico. Int Arch
Allergy Immunol 155(4):322–334.
87. Blackwell M. 1990. Air dispersed spores of myxomycetes,
p. 51–52. In Smith EG (ed), Sampling and Identifying
Allergenic Pollens and Molds. Blewstone Press, San Antonio,
TX.
88. Yang CS, Hung L-L, Lewis FA, Zampiello FA. 1993. Air-
borne fungal populations in non-residential buildings in the
3.2.5-16 ▪ AIR
United States. Helsinki University of Technology, Espoo,
Finland.
89. Von Essen S, Robbins RA, Thompson AB, Rennard SI.
1990. Organic dust toxic syndrome: an acute febrile reac-
tion to organic dust exposure distinct from hypersensitivity
pneumonitis. Clin Toxicol 28(4):389–420.
90. Womble S, Burton L, Kolb L, Girman J, Hadwen G, Car-
penter M, McCarthy J. 1999. Prevalence and concentra-
tions of culturable airborne fungal spores in 86 office
buildings from the Building Assessment Survey and Evalu-
ation (BASE) study. Indoor Air 99:261–266.
91. Shelton BG, Kirkland KH, Flanders WD, Morris GK.
2002. Profiles of airborne fungi in buildings and outdoor
environments in the United States. Appl Environ Microbiol
68(4):1743–1753.
92. Horner WE, Worthan AG, Morey PR. 2004. Air-
and dustborne mycoflora in houses free of water damage
and fungal growth. Appl Environ Microbiol 70(11):6394–
6400.
93. Raper KB, Fennell DI. 1965. The genus Aspergillus.
Williams & Wilkins, Philadelphia, PA.
94. Ohtsuki T. 1962. Studies on the glass mould. V. On two
species of Aspergillus isolated from glass. Bot Mag Tokyo
75:436–442.
95. Scott J, Untereiner WA, Wong B, Straus NA, Malloch D.
2004. Genotypic variation in Penicillium chysogenum from
indoor environments. Mycologia 96(5):1095–1105.
96. Houbraken J, Frisvad JC, Samson RA. 2011. Fleming’s
penicillin producing strain is not Penicillium chrysogenum
but P. rubens. IMA Fungus Global Mycol J 2(1):87.
97. Johanning E, Gareis M, Yang CS, Hintikka EL, Nikulin
M, Jarvis B, Dietrich R. 1999. Toxicity screening of mate-
rials from buildings with fungal indoor air quality problems
(Stachybotrys chartarum). Mycotox Res 14(2):60–73.
98. Li DW, Kendrick B, Spero D, Macdonald C. 2008. Bal-
aniopsis triangularis sp. nov. from indoor environments.
Mycotaxon 105:105.
99. Corda ACJ. 1837. Icones fungorum hucusque cognitorum.
Calve, Prague.
100. Amend AS, Seifert KA, Samson R, Bruns TD. 2010.
Indoor fungal composition is geographically patterned
and more diverse in temperate zones than in the tropics.
Proc Natl Acad Sci 107(31):13748–13753.
101. Andersen B, Frisvad JC, Søndergaard I, Rasmussen IS,
Larsen LS. 2011. Associations between fungal species
and water-damaged building materials. Appl Environ Micro-
biol 77(12):4180–4188.
102. Yang CS. 1996. Fungal colonization of HVAC fiber-glass
air-duct liner in the U.S. Institute of Health, Minatoku,
Tokyo.
103. Heinemann S, Begiun H, Nolard N, Samson R, Flanni-
gan B, Flannigan M, Verhoeff A, Adan O, Hoekstra E.
1994. Biocontamination in air-conditioning. Elsevier,
Amsterdam.
104. Kemp SJ, Kuehn TH, Pui DYH, Vesley D, Streifel AJ.
1995. Growth of microorganisms on HVAC filters under
controlled temperature and humidity conditions. ASH-
RAE Trans 101(1):305–316.
105. Buttner MP, Cruz-Perez P, Garrett PJ, Stetzenbach LD.
1999. Dispersal of fungal spores from three types of air han-
dling system duct material. Aerobiologia 15(1):1–8.
106. Schmidt O. 2007. Indoor wood-decay basidiomycetes:
damage, causal fungi, physiology, identification and char-
acterization, prevention and control. Mycol Prog 6(4):
261–279.
107. Kauserud H, Knudsen H, Högberg N, Skrede I. 2012.
Evolutionary origin, worldwide dispersal, and population
genetics of the dry rot fungus Serpula lacrymans. Fung Biol
Rev 26(2–3):84–93.
108. Alfredsen G, Solheim H, Jenssen K. 2005. Evaluation
of decay fungi in Norwegian buildings. IRG/WP 10562:
1–12.
109. Duncan C. 1965. Fungi associated with principal decays in
wood products in the United States. Res. Pap. WO-4, U.S.
Forest Service, Wisconsin.
110. Kwon-Chung KJ, Bennett JE. 1992. Medical mycology.
Lea & Febiger, Philadelphia, PA.
111. Pfaller M, Fromtling R. 1999. Mycology, p. 1159–1326. In
Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH
(eds), Manual of Clinical Microbiology, 7th ed. American
Society for Microbiology Press, Washington, DC.
112. Institute of Medicine. 2000. Clearing the Air: Asthma and
Indoor Air Exposures. National Academies Press, Washing-
ton, DC.
113. Institute of Medicine. 2004. Damp Indoor Spaces and Health.
National Academies Press, Washington, DC.
114. Pope AMP, Patterson R, Burge HA. 1993. Indoor Aller-
gens: Assessing and Controlling Adverse Health Effects.
National Academies Press, Washington, DC.
115. WHO. 2009. WHO Guidelines for Indoor Air Quality:
Dampness and Mould. World Health Organization,
Regional Office for Europe, Copenhagen.
116. Mendell MJ, Mirer AG, Cheung K, Douwes J. 2011. Res-
piratory and allergic health effects of dampness, mold, and
dampness-related agents: a review of the epidemiologic evi-
dence. Environ Health Perspect 119(6):748.
117. Koskinen OM, Husman TM, Meklin TM, Nevalainen
AI. 1999. Adverse health effects in children associated
with moisture and mold observations in houses.
Int J Environ Health Res 9(2):143–156.
118. Olsen J, Dragsted L, Autrup H. 1988. Cancer risk and
occupational exposure to aflatoxins in Denmark.
Br J Cancer 58(3):392.
119. WHO & National Agency For Research On Cancer (Iarc).
1993. IARC Monographs on the Evaluation of Carconogenic
Risks to Humans. World Health Organization (IARC),
Lyon, France.
120. Johanning E, Landsbergis P, Gareis M, Yang CS, Olmsted
E. 1999. Clinical experience and results of a Sentinel
Health Investigation related to indoor fungal exposure.
Environ Health Perspect 107(Suppl 3):489.
121. Ammann HM. 1999. IAQ and human toxicosis: empirical
evidence and theory, p. 84–93. In Johanning E (ed), Bioaer-
osols, Fungi and Mycotoxins: Health Effects, Assessment, Pre-
vention and Control. Eastern New York Occupational and
Environmental Health Center, Albany, NY.
122. Lorenz W, Sigrist G, Shakibaei M, Mobasheri A,
Trautmann C. 2006. A hypothesis for the origin and
pathogenesis of rheumatoid diseases. Rheumatol Int 26(7):
641–654.
123. Carey SA, Plopper CG, Hyde DM, Islam Z, Pestka JJ,
Harkema JR. 2012. Satratoxin-G from the black mold Sta-
chybotrys chartarum induces rhinitis and apoptosis of olfac-
tory sensory neurons in the nasal airways of rhesus monkeys.
Toxicol Pathol 40(6):887–898.
124. Burge H, Ammann H. 1999. Fungal toxins and β
(1→3)-D-glucans. In Macher J, Ammann H, Burge H, Mil-
ton D, Morey P (eds), Bioaerosols: Assessment and Control.
ACGIH, Cincinnati, OH.
125. Husman T. 1996. Health effects of indoor-air microorgan-
isms. Scand J Work, Environ Health 22(1):5–13.
126. Putus T, Johanning E. 2005. Health effects of moisture
damage associated microbes, Proceedings of the 5th inter-
national conference: Bioaerosols, fungi, bacteria, mycotox-
ins and human health. Fungal Research Group Foundation,
Inc. Albany, NY: Boyd publishing, 94–107.
127. Samson R. 1992. Mycotoxins: a mycologist’s perspective.
Med Mycol 30(S1):9–18.

128. Stetzenbach L, Amman H, Johanning E, King G, Shaugh-
nessy R. 2004. Microorganisms, mold and indoor air qual-
ity. American Society for Microbiology, Washington, DC.
129. Brunekreef B, Dockery DW, Speizer FE, Ware JH, Spen-
gler JD, Ferris BG. 1989. Home dampness and respiratory
morbidity in children. Am J Respir Crit Care Med 140(5):
1363–1367.
130. Hyvärinen A, Reponen T, Husman T, Nevalainen A.
2001. Comparison of the indoor air quality in mould dam-
aged and reference buildings in a subarctic climate. Cent
Eur J Public Health 9(3):133.
131. Nevalainen A, Partanen P, Jääskeläinen E, Hyvärinen
A, Koskinen O, Meklin T, Vahteristo M, Koivisto J,
Husman T. 2010. Prevalence of moisture problems in Fin-
nish houses. Indoor Air 8(S4):45–49.
132. Bonham RE. 2005. A review of the IOM report on mold.
Tex Med 100(12):55–57.
133. Fisk WJ, Lei-Gomez Q, Mendell MJ. 2007. Meta-
analyses of the associations of respiratory health effects
with dampness and mold in homes. Indoor Air 17(4):
284–296.
134. Rylander R. 1999. Effects after mold exposure—which
are the causative agents? p. 28–32. In Johanning E (ed),
Bioaerosols, Fungi and Mycotoxins: Health Effects, Assess-
ment, Prevention and Control. Eastern New York Occupa-
tional and Environmental Health Center, Albany, NY.
135. Pfaller MA. 1992. Opportunistic fungal infections, In Last
JM, Wallace RB (eds), Maxcy-Rosenau Public Health and
Preventive Medicine, 13th ed. Appleton and Lange, East
Norwalk, CT.
136. Kaplan JE, Jones JL, Dykewicz CA. 2000. Protists as
opportunistic pathogens: public health impact in the
1990s and beyond1. J Eukaryot Microbiol 47(1):15–20.
137. Pitt JI. 1979. The genus Penicillium and its teleo-
morphic states. Eupenicillium and Talaromyces, Academic
Press, London, UK.
138. Kurup VP. 1999. Allergens, diagnosis, and pathogenesis,
p. 620–625. In Johanning E (ed), Bioaerosols, Fungi and
Mycotoxins: Health Effects, Assessment, Prevention and Con-
trol. Eastern New York Occupational and Environmental
Health Center, Albany, NY.
139. CDC. 2012. Multistate outbreak of fungal infection associ-
ated with injection of methylprednisolone acetate solution
from a single compounding pharmacy—united states,
2012. Morb Mortal Wkly Rep 61(41):839–842.
140. CDC. 2012. Multistate Fungal Neningitis Outbreak Investiga-
tion. CDC, Atlanta, GA.
141. Benenson AS. 1990. Control of Communicable Disease in
Man. American Public Health Association, Washington,
DC.
142. Fatterpekar G, Mukherji S, Arbealez A, Maheshwari S,
Castillo M. 1999. Fungal Diseases of the Paranasal Sinuses.
Elsevier, Amsterdam.
143. Kern EB, Ponikau JU, Sherris DA, Kita H. 2012. Fungi
and chronic rhinosinusitis (CRS): cause and effect. In
Johanning EPM, Auger P (eds), Bioaerosols—6th Interna-
tional Scientific Conference on Bioaerosols, Fungi, Bacte-
ria, Mycotoxins in Indoor and Outdoor Environments and
Human Health. Fungal Research Group Foundation,
Albany, NY.
144. Ponikau JU, Sherris DA, Kern EB, Homburger HA,
Frigas E, Gaffey TA Roberts GD. 1999. The diagnosis
and incidence of allergic fungal sinusitis. Mayo Clin Proc
74(9):877–884.
145. Park JH, Kreiss K, Cox-Ganser JM. 2012. Rhinosinusitis
and mold as risk factors for asthma symptoms in occupants
of a water-damaged building. Indoor Air 22:396–404.
146. Salez F, Lamblin C, Wallaert B. 2000. Allergic broncho-
pulmonary aspergillosis. Rev Mal Resp 17(1 Pt 2):265– 278.
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-17
147. Wark PAB, Gibson PG. 2001. Allergic bronchopulmo-
nary aspergillosis: new concepts of pathogenesis and treat-
ment. Respirology 6(1):1–7.
148. Bellanti JA. 2006. Literature review: The best new articles
in the specialty of allergy, asthma, and immunology, 2004–
2005. Allergy and Asthma Proceedings 27(3):186–196.
149. Denning D, O’driscoll B, Hogaboam C, Bowyer P,
Niven R. 2006. The link between fungi and severe
asthma: a summary of the evidence. Eur Respir J 27(3):
615–626.
150. Chapman JA. 1999. Update on airborne mold and mold
allergy. Allergy and Asthma Proceedings 20(5):289–292.
151. Gravesen S. 1979. Fungi as a cause of allergic disease.
Allergy 34(3):135–154.
152. Horwitz R, Bush R. 1997. Allergens and other factors
important in atopic disease, p. 75–129. In Patterson R,
Grammer LC, Greenberger PA (eds), Allergic Diseases,
Diagnosis and Management, 5th ed. Lippincott-Raven,
Philadelphia.
153. Burr M. 1999. Indoor fungal exposure—does it matter, and
what can be done about it? Clin Exper Allergy J Br Soc
Allergy Clin Immunol 29(11):1442–1444.
154. Horner W, Lehrer S. 1999. Why are there still problems
with fungal allergen extracts? p. 313–319. In Johanning E
(ed), Bioaerosols, Fungi and Mycotoxins: Health Effects,
Assessment, Prevention and Control. Eastern New York
Occupational and Environmental Health Center, Albany,
NY.
155. Dales RE, Burnett R, Zwanenburg H. 1991. Adverse
health effects among adults exposed to home dampness
and molds. Am J Respir Crit Care Med 143(3):505–509.
156. Johanning E, Morey PR, Jarvis BB. 1993.
Clinical-epidemiological investigation of health effects caused
by Stachybotrys atra building contamination: V.1 Health
effects. Helsinki University of Technology, Espoo, Finland.
157. Ruotsalainen R, Jaakkola N, Jaakkola J. 1995. Dampness
and molds in day-care centers as an occupational
health problem. Int Arch Occup Environ Health 66(6):
369–374.
158. Smith JE, Anderson JG, Lewis CW, Murad YM. 1992.
Cytotoxic fungal spores in the indoor atmosphere of the
damp domestic environment. FEMS Microbiol Lett 100
(1):337–343.
159. Spengler JD, Neas L, Nalai S, Dockery D, Speizer F,
Ware J, Raizenne M. 1993. Respiratory symptoms and hous-
ing characteristics. Vol. 1 Health effects. Helsinki University
of Technology, Espoo, Finland.
160. Storey E, Dangman K, Schenck P, DeBernardo R, Yang
C, Bracker A, Hodgson M. 2004. Guidance for clinicians
on the recognition and management of health effects related to
mold exposure and moisture indoors. University of Connecti-
cut Health Center, Division of Occupational and Environ-
mental Medicine, Center for Indoor Environments and
Health, Farmington, CT.
161. Strachan DP, Elton RA. 1986. Relationship between res-
piratory morbidity in children and the home environment.
Fam Pract 3(3):137–142.
162. Bornehag CG, Blomquist G, Gyntelberg F, Jarvholm B,
Malmberg P, Nordvall L, Nielsen A, Pershagen G, Sun-
dell J. 2001. Dampness in buildings and health. Indoor
Air 11(2):72–86.
163. Cox-Ganser JM, White SK, Jones R, Hilsbos K, Storey E,
Enright PL, Rao CY, Kreiss K. 2005. Respiratory morbid-
ity in office workers in a water-damaged building. Environ
Health Perspect 113(4):485–490.
164. Dangman KH, Bracker AL, Storey E. 2005. Work-related
asthma in teachers in Connecticut: association with
chronic water damage and fungal growth in schools.
Conn Med 69(1):9–17.
3.2.5-18 ▪ AIR
165. Peat JK, Dickerson J, Li J. 2007. Effects of damp and
mould in the home on respiratory health: a review of the lit-
erature. Allergy 53(2):120–128.
166. Gent JF, Kezik JM, Hill ME, Tsai E, Li DW Leaderer BP.
2012. Household mold and dust allergens: Exposure, sensi-
tization and childhood asthma morbidity. Environ Res 118-
(Oct):86–93.
167. Nielsen KF, Huttunen K, Hyvärinen A, Andersen B,
Jarvis BB, Hirvonen MR. 2002. Metabolite profiles of
Stachybotrys isolates from water-damaged buildings and
their induction of inflammatory mediators and cytotoxicity
in macrophages. Mycopathologia 154(4):201–206.
168. Nikulin M. 1999. Toxigenic Stachybotrys chartarum in
water-damaged building materials: ecology and biology.
University of Helsinki, Helsinki, Finland.
169. Taskinen T, Laitinen S, Nevalainen A, Vepsäläinen A,
Meklin T, Reiman M, Korppi M, Husman T. 2002. Immu-
noglobulin G antibodies to moulds in school-children from
moisture problem schools. Allergy 57(1):9–16.
170. Jaakkola JJK, Hwang BF, Jaakkola N. 2005. Home damp-
ness and molds, parental atopy, and asthma in childhood: a
six-year population-based cohort study. Environ Health Per-
spect 113(3):357.
171. Ehrenstein V, Mutius V, Kries V. 2001. Reduced risk of
hay fever and asthma among children of farmers. Clin Exp
Allergy 30(2):187–193.
172. Jaakkola MS, Nordman H, Piipari R, Uitti J, Laitinen J,
Karjalainen A, Hahtola P, Jaakkola JJK. 2002. Indoor
dampness and molds and development of adult-onset
asthma: a population-based incident case-control study.
Environ Health Perspect 110(5):543.
173. Kilpeläinen M, Terho E, Helenius H, Koskenvuo M.
2002. Childhood farm environment and asthma and
sensitization in young adulthood. Allergy 57(12):1130–
1135.
174. Karvala K, Toskala E, Luukkonen R, Uitti J, Lappalainen
S, Nordman H. 2011. Prolonged exposure to damp and
moldy workplaces and new-onset asthma. Int Arch Occup
Environ Health 84(7):713–721.
175. Murr AH, Goldberg AN, Pletcher SD, Dillehay K,
Wymer LJ, Vesper SJ. 2012. Some chronic rhinosinusitis
patients have elevated populations of fungi in their sinuses.
Laryngoscope 122:1438–1445.
176. Vesper S, McKinstry C, Haugland R, Wymer L, Bradham
K, Ashley P, Cox D, Dewalt G, Friedman W. 2007. Devel-
opment of an environmental relative moldiness index for
US homes. J Occup Environ Med 49(8):829–833.
177. Pieckova E, Jesenska Z, Wilikins K. 1999. Microscopic
fungi and their metabolites in dwellings—a bioassay study,
p. 351–354. In Johanning E (ed), Bioaerosols, Fungi and
Mycotoxins: Health Effects, Assessment, Prevention and cOn-
trol. Eastern New York Occupational and Environmental
Health Center, Albany, NY.
178. Dillon HK, Miller JD, Sorenson W, Douwes J, Jacobs
RR. 1999. Review of methods applicable to the assessment
of mold exposure to children. Environ Health Perspect 107
(Suppl 3):473–480.
179. Miller JD, Dales R, White J. 1999. Exposure measures for
studies of mold and dampness and respiratory health,
p. 298–305. In Johanning E (ed), Bioaerosols, Fungi and
Mycotoxins: Health Effects, Assessment, Prevention and Con-
trol. Eastern New York Occupational and Environmental
Health Center, Albany, NY.
180. Ammann HM. 1999. ACGIH TLV Statement on bioaer-
osols; American Council of Government Industrial
Hygienists, p. 520–521. In Johanning E (ed), Bioaerosols,
Fungi and Mycotoxins: Health Effects, Assessment, Prevention
and Control. Eastern New York Occupational and Environ-
mental Health Center, Albany, NY.
181. Anonymous. 1994. Recommendations, p. 531–538. In
Samson RA, Flannigan B, Flannigan ME, VerHoeff AP,
Adan OCG, Hoekstra ES (eds), Health Implications of Fungi
in Indoor Environments. Elsevier, Amsterdam.
182. Barnes C, Buckley S, Pacheco F, Portnoy J. 2002.
IgE-reactive proteins from Stachybotrys chartarum. Ann
Allergy Asthma Immunol 89(1):29–33.
183. Chung YJ, Copeland LB, Doerfler DL, Ward MDW.
2010. The relative allergenicity of Stachybotrys chartarum
compared to house dust mite extracts in a mouse model.
Inhal Toxicol 22(6):460–468.
184. Nagayoshi M, Tada Y, West J, Ochiai E, Watanabe
A, Toyotome T, Tanabe N, Takiguchi Y, Shigeta A,
Yasuda T. 2011. Inhalation of Stachybotrys chartarum
evokes pulmonary arterial remodeling in mice, attenuated
by rho-kinase inhibitor. Mycopathologia 1–11.
185. Yike I, Dearborn D. 2011. Guest editorial—novel insights
into the pathology of Stachybotrys chartarum. Mycopatholo-
gia 1–3.
186. Rakkestad KE, Skaar I, Ansteinsson VE, Solhaug A,
Holme JA, Pestka JJ, Samuelsen JT, Dahlman HJ, Hon-
gslo JK, Becher R. 2010. DNA damage and DNA damage
responses in THP-1 monocytes after exposure to spores of
either Stachybotrys chartarum or Aspergillus versicolor or to
T-2 toxin. Toxicol Sci 115(1):140–155.
187. Henson PM, Murphy RC. 1989. Mediator of the Inflamma-
tory Process. Elsevier, New York.
188. Rylander R, Jacobs RR. 1994. Organic Dusts Exposure,
Effects, and Prevention. CRC Press, Boca Raton, FL.
189. Rylander R. 1994. Office and domestic environments,
p. 247–255. In Rylander R Jacobs RR (eds), Organic Dusts:
Exposure, Effects, and Prevention. Lewis Publishers, Boca
Raton, FL.
190. Rylander R, Etzel R. 1999. Introduction and summary:
workshop on children’s health and indoor mold exposure.
Environ Health Perspect 107(Suppl 3):465–468.
191. Johanning E. 1999. Overview of waste management
in the United States and recent research activities
about composting related occupational health risks,
p. 127–140. In Eikmann T, Hofmann R (eds), Stand von
Wissenschaft, Forschung und Technik zu siedlungshygienischen
Aspekten der Abfallentsorgung und -verwertung. (Status
assessment of science, research and techniques related to hygiene
in waste-management and -use) Berlin, Kommission Reinhal-
tung der Luft im VDI und DIN. Band 30. Schriftenreihedes
Vereins für Wasser-. Boden- und Lufthygiene, Berlin,
Germany.
192. National Institute for Occupational Safety and Health.
1994. Request for assistance in preventing organic dust toxic
syndrome. DHHS, NIOSH, Cincinnati, OH.
193. Bünger J, Muller M, Stalder K, Hallier E. 1999. Immuno-
logical biomonitoring in the assessment of exposure to air-
borne fungi from waste handling, p. 112–118. In Johanning
E (ed), Bioaerosols, Fungi and Mycotoxins: Health Effects,
Assessment, Prevention and Control. Eastern New York
Occupational and Environmental Health Center, Albany,
NY.
194. Chiba S, Okada S, Suzuki Y, Watanuki Z, Mitsuishi Y,
Igusa R, Sekii T, Uchiyama B. 2009. Cladosporium
species-related hypersensitivity pneumonitis in household
environments. Intern Med 48(5):363–367.
195. Jacobs RL, Thorner RE, Holcomb JR, Schwietz LA,
Jacobs FO. 1986. Hypersensitivity pneumonitis caused by
Cladosporium in an enclosed hot-tub area. Ann Intern
Med 105(2):204–206.
196. Lee SK, Kim SS, Nahm DH, Park HS, Oh YJ, Park KJ,
Kim SO, Kim SJ. 2000. Hypersensitivity pneumonitis
caused by Fusarium napiforme in a home environment.
Allergy 55(12):1190–1193.

197. Selman M, Lacasse Y, Pardo A, Cormier Y. 2010. Hyper-
sensitivity pneumonitis caused by fungi. Proc Am Thorac
Soc 7(3):229–236.
198. Bhan U, Newstead MJ, Zeng X, Ballinger MN, Standi-
ford LR, Standiford TJ. 2011. Stachybotrys chartarum-
Induced Hypersensitivity Pneumonitis Is TLR9 Depend-
ent. Am J Pathol 179(6):2779–2787.
199. Lincoff G, Mitchel DH. 1977. Toxic and Hallucinogenic
Mushroom Poisoning. A Handbook for Physicians and Mush-
room Hunters. Van Nostrand Reinhold, New York.
200. Gonmori K, Fujita H, Yokoyama K, Watanabe K, Suzuki
O. 2011. Mushroom toxins: a forensic toxicological review.
Forens Toxicol 29(2):85–94.
201. Marasas W, Nelson PE. 1987. Mycotoxicology: Introduction
to the Mycology, Plant Pathology, Chemistry, Toxicology, and
Pathology of Naturally Occuring Mycotoxicoses in Animals and
Man. Pennsylvania State University Press, State College.
202. Krogh P. 2008. The role of mycotoxins in disease of ani-
mals and man. J Appl Microbiol 67(s18):99s–104s.
203. Croft WA, Jarvis BB, Yatawara C. 1986. Airborne out-
break of trichothecene toxicosis. Atmos Environ (1967)
20(3):549–552.
204. Smoragiewicz W, Cossette B, Boutard A, Krzystyniak K.
1993. Trichothecene mycotoxins in the dust of ventilation
systems in office buildings. Int Arch Occup Environ Health
65(2):113–117.
205. Dill I, Trautmann C, Szewzyk R. 1997. Mass development
of Stachybotrys chartarum on compostable plant pots made
from recycled paper. Mycoses 40:110–114.
206. Schiefer HB. 1990. Mycotoxins in Indoor Air: A Critical
Toxicological Viewpoint, Vol. 1, Indoor Air 0 90, Proceedings
of the 5th International Conference on Indoor Air Quality
and Climate, Toronto. 167–172.
207. Richard J, Payne GA, Desjardins A, Maragos C, Norred
W, Pestka J. 2003. Mycotoxins: Risks in Plant, Animal and
Human Systems. Council for Agricultural Science and
Technology, Ames, IA.
208. Peitzsch M, Sulyok M, Täubel M, Vishwanath V,
Seegers-Krop E, Borràs-Santos A, Hyvärinen A, Neva-
lainen A, Krska R, Larsson LP. 2012. Microbial secondary
metabolites in school buildings inspected for moisture dam-
age in Finland, The Netherlands and Spain. J Environ
Monit 14(8):2044–2053.
209. Larsen TO, Smedsgaard J, Nielsen KF, Hansen ME, Fris-
vad JC. 2005. Phenotypic taxonomy and metabolite profil-
ing in microbial drug discovery. Nat Prod Rep 22(6):
672–695.
210. De Costa C. 2002. St Anthony’s fire and living ligatures:
a short history of ergometrine. Lancet 359(9319):
1768–1770.
211. Peraica M, Radic B, Lucic A, Pavlovic M. 1999. Toxic
effects of mycotoxins in humans. Bull World Health Organ
77(9):754–766.
212. Matossian MK. 1991. Poisons of the Past: Molds, Epidemics,
and History. Yale University Press, New Haven, CT.
213. Ammann HM. 2005. Mold toxicity: risk assessment for
humans exposed indoors, p. 52–58. In Johanning E (ed),
Bioaerosols, Fungi, Bacteria, Mycotoxins and Human Health:
Patho-physiology, Clinical Effects, Exposure Assessment, Pre-
vention and Control in Indoor Environments and Work. Fungal
Research Group Foundation, Albany, NY.
214. Miller VR, Ammann HM. 2005. Risk assessment of
fungal bioaerosols in indoor environments: Current
paradigms and evolving concepts, p. 474–481. In Johan-
ning E (ed), Bioaerosols, Fungi, Bacteria, Mycotoxins and
Human Health: Patho-physiology, Clinical Effects, Exposure
Assessment, Prevention and Control in Indoor Environments
and Work. Fungal Research Group Foundation, Albany,
NY.
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-19
215. Auger PL, Pépin P, Miller JD, Gareis M, Doyon J, Bou-
chard R, Pinard M-F, Mainville C. 1999. Chronic toxic
encephalopathies apparently related to exposure to toxi-
genic fungi, p. 131–138. In Johanning E (ed), Bioaerosols,
Fungi and Mycotoxins: Health Effects, Assessment, Prevention
and Control. Eastern New York Occupational and Environ-
mental Health Center, Albany, NY.
216. Ciegler A, Burmeister HR, Vesonder RF, Hesseltine
CW. 1981. Mycotoxins: occurrence in the environment,
In Shank RC (ed), Mycotoxins and N-nitroso Compound
Environmental Risk. CRC Press, Boca Raton, FL.
217. IARC. 1993. IARC Monographs in the evaluation of carcino-
genic risks to humans. World Health Organization.
218. Sorenson W. 1999. Fungal spores: hazardous to health?
Environ Health Perspect 107(Suppl 3):469.
219. Ueno Y. 1983. Trichothecenes. Chemical, Biological and
Toxicological Aspects. Elsevier Science, Amsterdam, The
Netherlands.
220. Ueno Y. 1985. The toxicology of mycotoxins. Crit Rev Tox-
icol 14(2):99–132.
221. Quintela S, Villarán MC, Armentia IL, Elejalde E. 2013.
Ochratoxin A removal in wine: a review. Food Control
30(2):439–445.
222. Saxena N, Ansari KM, Kumar R, Chaudhuri BP, Dwi-
vedi PD, Das M. 2011. Role of mitogen activated protein
kinases in skin tumorigenicity of Patulin. Toxicol Appl Phar-
macol 257(2):264–271.
223. Zain ME. 2011. Impact of mycotoxins on humans and ani-
mals. J Saudi Chem Soc 15(2):129–144.
224. Ferrante M, Sciacca S, Conti GO. 2012. Carcinogen role
of food by mycotoxins and knowledge gap, p. 133–162.
In Pesheva M, Dimitrov M, Stoycheva TS (eds), Carcino-
gen, InTech, Rijeka, Croatia.
225. Dearborn DG, Yike I, Sorenson W, Miller MJ, Etzel RA.
1999. Overview of investigations into pulmonary hemor-
rhage among infants in Cleveland, Ohio. Environ Health
Perspect 107(Suppl 3):495.
226. Hodgson MJ, Morey P, Leung WY, Morrow L, Miller D,
Jarvis BB, Robbins H, Halsey JF, Storey E. 1998.
Building-Associated Pulmonary Disease From Exposure
to Stachybotrys chartarum and Aspergillus versicolor.
J Occup Environ Med 40(3):241–249.
227. Gordon WA, Cantor JB, Johanning E, Charatz HJ, Ash-
man TA, Breeze JL, Haddad L, Abramowitz S. 2004. Cog-
nitive impairment associated with toxigenic fungal
exposure: a replication and extension of previous findings.
Appl Neuropsychol 11(2):65–74.
228. Gordon WA, Cantor JB. 2004. The diagnosis of cognitive
impairment associated with exposure to mold. Adv Appl
Microbiol 55:361–374.
229. Ammann HM. 2003. Is indoor mold contamination a
threat to health? Part two. J Environ Health 66(2):47–9.
230. Sorenson W. 1990. Mycotoxins as potential occupational
hazards. Dev Ind Microbiol 31:205–211.
231. Lanier C, André V, Séguin V, Heutte N, El Kaddoumi A,
Bouchart V, Picquet R, Garon D. 2012. Recurrence of
Stachybotrys chartarum during mycological and toxicologi-
cal study of bioaerosols collected in a dairy cattle shed.
Ann Agric Environ Med 19(1):61.
232. Mayer S, Curtui V, Usleber E, Gareis M. 2007. Airborne
mycotoxins in dust from grain elevators. Mycotoxin Res 23
(2):94–100.
233. Johanning E, Gareis M, Nielsen K, Dietrich R,
Märtlebauer E. 2002. Airborne mycotoxin sampling and
screening analysis. Indoor Air, 1–6.
234. Brasel TL, Douglas DR, Wilson SC, Straus DC. 2005.
Detection of airborne Stachybotrys chartarum macrocyclic
trichothecene mycotoxins on particulates smaller than
conidia. Appl Environ Microbiol 71(1):114–122.
3.2.5-20 ▪ AIR
235. Bloom E, Nyman E, Must A, Pehrson C, Larsson L. 2009.
Molds and mycotoxins in indoor environments—a
survey in water-damaged buildings. J Occup Environ Hyg
6(11):671–678.
236. Täubel M, Sulyok M, Vishwanath V, Bloom E, Turunen
M, Järvi K, Kauhanen E, Krska R, Hyvärinen A, Larsson
L. 2011. Co-occurrence of toxic bacterial and fungal secon-
dary metabolites in moisture-damaged indoor environ-
ments. Indoor Air 21(5):368–375.
237. Peitzsch M, Bloom E, Haase R, Must A, Larsson L. 2012.
Remediation of mould damaged building materials—effi-
ciency of a broad spectrum of treatments. J Environ Monit
14(3):908–915.
238. WHO. 1990. Environmental Health Criteria 105, Selected
Mycotoxins: Ochratoxins, Trichothecenes, Ergot. World
Health Organization, Geneva.
239. Jarvis BB, Salemme J, Morals A. 2006. Stachybotrys toxins.
1. Nat Toxins 3(1):10–16.
240. Jarvis BB, Zhou Y, Jiang J, Wang S, Sorenson W, Hin-
tikka EL, Nikulin M, Parikka P, Etzel RA, Dearborn
DG. 1996. Toxigenic molds in water-damaged buildings:
dechlorogriseofulvins from Memnoniella echinata. J Nat
Prod 59(6):553–554.
241. Tuomi T, Reijula K, Johnsson T, Hemminki K, Hintikka
EL, Lindroos O, Kalso S, Koukila-Kähkölä P, Mussalo-
Rauhamaa H, Haahtela T. 2000. Mycotoxins in crude
building materials from water-damaged buildings. Appl
Environ Microbiol 66(5):1899–1904.
242. Smith JE, Moss MO. 1985. Mycotoxins. Formation, Analy-
sis and Significance. John Wiley & Sons, New York.
243. Hendry KM, Cole EC. 1993. A review of mycotoxins in
indoor air. J Toxicol Environ Health A 38(2):183–198.
244. Kusak V, Jelinek S, Sula J. 1970. Possible role of Aspergil-
lus flavus in the pathogenesis of Schneeberg and Jachymov
disease. Neoplasma 17(5):441–449.
245. Andersson M, Nikulin M, Köljalg U, Andersson M,
Rainey F, Reijula K, Hintikka E, Salkinoja-Salonen
M. 1997. Bacteria, molds, and toxins in water-damaged
building materials. Appl Environ Microbiol 63(2):387–
393.
246. Gareis M, Johanning E, Dietrich R. 1999. Mycotoxin
cytotoxicity screening of field samples, p. 202–213. In
Johanning E (ed), Bioaerosols, Fungi and Mycotoxins: Health
Effects, Assessment, Prevention and Control. Eastern
New York Occupational and Environmental Health Cen-
ter, Albany, NY.
247. Creasia D, Thurman J, Wannemacher R, Bunner D.
1990. Acute inhalation toxicity of T-2 mycotoxin in the
rat and guinea pig. Fundam Appl Toxicol 14(1):54–59.
248. Betina V. 1989. Mycotoxins. Chemical, Biological and Envi-
ronmental Aspects. Elsevier, Amsterdam.
249. Hintikka EL. 1999. The effects of inhaled spores of myco-
toxin producing fungi on animals, p. 214–220. In Johan-
ning E (ed), Bioaerosols, Fungi and Mycotoxins: Health
Effects, Assessment, Prevention and Control. Eastern
New York Occupational and Environmental Health Cen-
ter, Albany, NY.
250. Okumurai H, Yoshino N, Sugiura Y, Sugamata M. 1999.
Tichothecenes as a potent inducer of apoptosis, p.
221–231. In Johanning E (ed), Bioaerosols, Fungi and Myco-
toxins: Health Effects, Assessment, Prevention and Control.
Eastern New York Occupational and Environmental
Health Center, Albany.
251. Forgacs J. 1972. Mycotoxin, p. 95–128. In Kadis S, Agl SJ
(eds), Microbial Toxins. Academic Press, New York.
252. Jarvis BB. 1990. Mycotoxins and indoor air quality, p.
201–214. In Morey P, Feeley J, Otten J (eds), Biological Con-
taminants in Indoor Environments. American Society for
Testing and Materials, Philadelphia.
253. Johanning E. 1995. Health problems related to fungal
exposure—primarily of Stachybotrys atra, p. 169–182. In
Johanning E, Yang CS (eds), Fungi and Bacteria in Indoor
Air Environments: Health Effects, Detection and Remediation.
Eastern New York Occupational Health Program, Albany,
NY.
254. Etzel RA, Montana E, Sorenson W, Kullman GJ, Allan
TM, Dearborn DG. 1998. Acute pulmonary hemorrhage
in infants associated with exposure to Stachybotrys atra
and other fungi. Arch Pediatr Adolesc Med 152(8):757.
255. Flappan SM, Portnoy J, Jones P, Barnes C. 1999. Infant
pulmonary hemorrhage in a suburban home with water
damage and mold (Stachybotrys atra). Environ Health Per-
spect 107(11):927.
256. CDC. 2000. Update: pulmonary hemorrhage/hemosiderosis
among infants—Cleveland, Ohio, 1993–1996. Morbid. Mor-
tal. Wkly. Rep. 49(9):180–184.
257. Dearborn D. 2013. Mold. In Landrigan P, Etzel R (eds),
Children’s Environmental Health. Oxford University Press,
New York.
258. Nikulin M, Reijula K, Jarvis BB, Hintikka E-L. 1996.
Experimental lung mycotoxicosis in mice induced by Sta-
chybotrys atra. Int J Exp Pathol 77(5):213–218.
259. Elidemir O, Colasurdo GN, Rossmann SN, Fan LL. 1999.
Isolation of Stachybotrys from the lung of a child with pul-
monary hemosiderosis. Pediatrics 104(4):964–966.
260. Tripi PA, Modlin S, Sorenson W, Dearborn DG. 2000.
Acute pulmonary haemorrhage in an infant during
induction of general anaesthesia. Pediatr Anesth 10(1):
92–94.
261. Sorenson W, Frazer DG, Jarvis BB, Simpson J, Robinson
V. 1987. Trichothecene mycotoxins in aerosolized conidia
of Stachybotrys atra. Appl Environ Microbiol 53(6):
1370–1375.
262. Pasanen AL, Nikulin M, Tuomainen M, Berg S, Parikka
P, Hintikka EL. 1993. Laboratory experiments on mem-
brane filter sampling of airborne mycotoxins produced by
Stachybotrys atra corda. Atmos Environ A Gen Top 27(1):
9–13.
263. Wicklow D, Shotwell O. 1983. Intrafungal distribution of
aflatoxins among conidia and sclerotia of Aspergillus flavus
and Aspergillus parasiticus. Can J Microbiol 29(1):1–5.
264. Corrier D. 1991. Mycotoxicosis: mechanisms of immuno-
suppression. Vet Immunol Immunopathol 30(1):73–87.
265. Jarvis BB, Hinkley SF. 1999. Analysis for Stachybotrys tox-
ins, p. 232–239. In Johanning E (ed), Bioaerosols, Fungi and
Mycotoxins: Health Effects, Assessment, Prevention and Con-
trol. Eastern New York Occupational and Environmental
Health Center, Albany, NY.
266. Gordon K, Masotti R, Waddell W. 1993. Tremorgenic
encephalopathy: a role of mycotoxins in the production
of CNS disease in humans? Can J Neurol Sci 20(3):237.
267. Gordon W, Johanning E, Haddad L. 1999. Cognitive
impairment associated with exposure to toxigenic fungi,
p. 94–98. In Johanning E (ed), Bioaerosols, Fungi and Myco-
toxins: Health Effects, Assessment, Prevention and Control.
Eastern New York Occupational and Environmental
Health Center, Albany, NY.
268. Jarmai K. 1929. Viskosusseptikamien bei alteren Fohlen
und erwachsenen Pferden (Viscosussepsis in older colts
and grown-up horses). Deutsche Tier Wochens 33:517–
519.
269. Ozegovic L. 1971. Straw related disease in farm workers (in
Russian). Veterinaria (Sarajevo) 20:263–267.
270. McLaughlin CS, Vaughan MH, Campbell IM, Wei CM,
Stafford ME, Hansen BS. 1977. Inhibition of protein
synthesis by trichothecenes, p. 263–273. In Rodrick JV,
Hesseltine CM, Mehlman MA (eds), Mycotoxins in Human
and Animal Health. Pathotox, Park Forest South, IL.

271. Goodwin W, Hass C, Fabian C, Heller-Bettinger I, Hoog-
straten B. 1978. Phase I evaluation of anguidine (diacetox-
yscirpenol, NSC–141537). Cancer 4(1):23–26.
272. Hintikka E. 1978. Human stachybotryotoxicosis, p. 87–89.
In Wyllie TD, Morehouse LG (eds), Mycotoxic Fungi, Myco-
toxins and Mycotoxicosis. Marcel Dekker, New York.
273. Tutelyan VA, Dadiani KR, Voitko NE. 1992. Changes in
aerological indicators of immune status of workers in con-
tact with mycotoxin-contaminated foodstuffs (in Russian).
Gigiena Truda i Professional’nye Zabolevaniya (Occup Safety
Health) 8:18–20.
274. Pestka J, Moorman M, Warner R. 1989. Dysregulation of
IgA production and IgA nephropathy induced by the tri-
chothecene vomitoxin. Food Chem Toxicol 27(6):
361–368.
275. Di Paolo N, Guarnieri A, Loi F, Sacchi G, Mangiarotti A,
Di Paolo M. 1993. Acute renal failure from inhalation of
mycotoxins. Nephron 64(4):621–625.
276. Lewis C, Smith J, Anderson J, Murad Y. 1994. The pres-
ence of mycotoxin-associated fungal spores isolated from
the indoor air of the damp domestic environment and
cytotoxic to human cell lines. Indoor Built Environ 3(6):
323–330.
277. Wilkins K, Larsen K, Simkus M. 2003. Volatile metabo-
lites from indoor molds grown on media containing wood
constituents. Environ Sci Pollut Res 10(4):206–208.
278. Fischer G, Müller T, Schwalbe R, Ostrowski R, Dott W.
2000. Exposure to airborne fungi, MVOC and mycotoxins
3.2.5. Airborne Fungi and Mycotoxins ▪ 3.2.5-21
in biowaste-handling facilities. Int J Hyg Environ Health
203(2):97–104.
279. Korpi A, Kasanen JP, Alarie Y, Kosma VM, Pasanen AL.
1999. Sensory irritating potency of some microbial volatile
organic compounds (MVOCs) and a mixture of five
MVOCs. Arch Environ Health 54(5):347–352.
280. Keller R, Senkpiel K, Ohgke H. 1999. Use of MVOC
measurements and odor perception as indicator of
mould in indoor areas, p. 532–537. In Johanning E (ed),
Bioaerosols, Fungi and Mycotoxins: Health Effects, Assess-
ment, Prevention and Control. Eastern New York Occupa-
tional and Environmental Health Center, Albany, NY.
281. Hänninen OO. 2011. WHO guidelines for indoor air qual-
ity: dampness and mold, p. 277–302. In Adan OC, Samson
RA (eds), Fundamentals of Mold Growth in Indoor Environ-
ments and Strategies for Healthy Living. Wageningen Aca-
demic Press, Wageningen.
282. Wady L, Bunte A, Pehrson C, Larsson L. 2003. Use of gas
chromatography-mass spectrometry/solid phase microex-
traction for the identification of MVOCs from moldy build-
ing materials. J Microbiol Meth 52(3):325–332.
283. Larsen TO, Frisvad JC. 1995. Characterization of volatile
metabolites from 47 Penicillium taxa. Mycol Res 99(10):
1153–1166.
284. Betancourt DA, Dean TR, Menetrez MY, Moore SA.
2006. Characterization of Microbial Volatile Organic Com-
pounds (MVOC) Emitted by Stachybotrys chartarum. Air &
Waste Management Association, Durham, NC.
 Airborne Bacteria, Archaea, and Endotoxin
PETER S. THORNE, CAROLINE DUCHAINE, AND PASCALE BLAIS LECOURS
3.2.6
Environments can be divided into four main categories
according to the concentrations of bacteria and endotoxin
contamination present (Table 1). The first category consists
of highly contaminated environments, including livestock
barns, row crop operations, agricultural transfer and process-
ing settings, and industrial operations that involve handling
or processing of organic matter supporting active microbial
growth. The second category includes occupational environ-
ments with moderate contamination such as food processing
facilities, industrial facilities not handling organic waste
materials, and domestic environments with significant con-
tamination sources. The third group encompasses the rela-
tively low contamination environments such as medical
and dental clinics, office environments, homes, and schools.
Finally, the fourth group includes environments that are
deliberately very low in bioaerosol concentration, such as
industrial clean rooms, pharmaceutical manufacturing areas,
surgical suites and some research laboratories. This fourth
group is not discussed in this chapter.
ENVIRONMENTS WITH HIGH
CONCENTRATIONS OF BACTERIA AND
ENDOTOXIN
Livestock production and concentrated animal feeding oper-
ations (CAFOs) are highly contaminated environments.
These include swine, poultry, dairy, and beef cattle facilities
where Staphylococcus, Pseudomonas, Bacillus, Listeria, Entero-
coccus, Nocardia, and Lactobacillus are commonly present
(1). Bacteria, endotoxin, and organic dust are the major air-
borne contaminants in swine CAFOs, and culturable air-
borne bacterial levels have been reported as high as 105
CFU/m3 (2, 3). The concentration and the species composi-
tion of airborne Gram-negative bacteria have been studied
in poultry, swine, and cattle barns (4). Between 0.02% and
5.2% of the total number of culturable aerobic bacteria
were Gram-negative with Enterobacteriaceae, Pseudomonada-
ceae, and Neisseriaceae dominant. In animal houses using
straw as bedding material Enterobacter agglomerans (Pantoea
agglomerans) was predominant, whereas when no bedding
material was used, E. coli was the most commonly isolated
bacterium (4).
Bioaerosol sampling in 24 swine barns yielded culturable
bacterial concentrations of 6.1 × 105 CFU/m3 and total
microorganisms at 13 × 107 cells/m3 (5). Representative
doi:10.1128/9781555818821.ch3.2.6
3.2.6-1
viable bacteria included Aerococcus, E. coli, Klebsiella, Pseu-
domonas, Rhodococcus, Staphylococcus, Streptococcus, Xantha-
monas, and Yersinia. Health problems observed in swine
building workers include local and systemic inflammatory
responses, cross-shift change in respiratory function, expec-
toration, and symptoms of cough and dyspnea. Cases of
occupational asthma and asthma-like syndrome are common
(see below) (6).
A major concern regarding agricultural airborne bacte-
ria is the propagation of antibiotic-resistant strains within
the environment or to people working or surrounding the
CAFOs because the nontherapeutic use of antibiotics in
swine production selects for antibiotic resistance in normal
flora and pathogenic bacteria. Enterococcus strains collected
from a CAFO were tested and analyzed for resistance to
erythromycin, clindamycin, virginiamycin, tetracycline, and
vancomycin (7). Regardless of bacterial species, 98% of the
isolates expressed high-level resistance to at least two antibi-
otics commonly used in swine production, suggesting that
the inhalation of air from these facilities may serve as an
exposure pathway for the transfer of multidrug-resistant bac-
terial pathogens from swine to humans (7). A separate study
detected antibiotic-resistant bacteria inside and downwind
of the CAFOs, including strains with resistance to ampicil-
lin, erythromycin, oxytetracycline, penicillin, tetracycline,
and tylosin within airborne Staphylococcus aureus, Salmonella
spp., and fecal coliforms. (8). Antibiotic resistance genes
from bacteria have also been found in swine barn bioaerosols
(7, 9, 10) and dairy farms. In cattle barns, S. aureus show-
ing resistance to penicillin, ampicillin, and cefaclor were
grown from aerosol samples collected using an impaction
method (11). In cage- and floor-raise poultry operations,
Enterococcus spp., E. coli, Staphylococcus spp., Campylobacter
spp., and Clostridium perfringens may be present. Zinc baci-
tracin resistance gene, erythromycin resistance gene (ermA),
and tetracycline resistance gene (tetA(C)) were also detected
(12). Farm animals are often carriers of methicillin-resistant
S. aureus (MRSA). It was shown that 40% of pigs in the
Netherlands carried MRSA (9), and other studies confirmed
the MRSA carrier status of pigs is widespread in Canada, the
United States, and Germany. A recent study (13) reported
a concentration of up to 4,000 CFU/m3 of culturable
MRSA in farms. It also suggested that the airborne bacteria
could then spread to humans and animals through the air-
borne route.
3.2.6-2 ▪ AIR
TABLE 1 Categories of environmental contamination for
bacteria and endotoxin
Environments with high concentrations of bacteria
and endotoxin
Concentrated animal feeding operations (swine, poultry, cattle)
Traditional livestock barns
Row crop harvesting
Grain elevators
Animal feed mills
Agricultural transfer and processing settings
Industrial composting operations
Solid waste transport and handling
Wastewater treatment plants
Other industries handling or processing of organic matter
Environments with moderate concentrations of bacteria and
endotoxin
Cotton textile industry
Vegetable- and seed-processing facilities
Metals machining with metal-working fluids
Forest products industry
Vegetable and flower greenhouse operations
Slaughterhouses (abattoirs)
Rodent vivaria
Domestic environments with significant contamination sources
Industrial operations not handling organic waste materials
Outdoor air (in vicinity of bioaerosol sources)
Environments with low concentrations of bacteria and endotoxin
Office environments
Domiciles
Day care centers and schools
Hospitals and medical clinics
Dental clinics
Podiatry clinics
Outdoor air (away from bioaerosol sources)
Environments with very low concentrations of bacteria and
endotoxin
Pharmaceutical manufacturing facilities
Industrial clean rooms
Surgical suites
Specialized research laboratories
Most of the exposure control efforts for livestock facilities
have focused on personal respiratory protection. Environ-
mental exposure control methods are not common. In swine
buildings vegetable oil sprinkling has been tested to decrease
the overall airborne dust levels and endotoxin with success
(14, 15). However, there was little effect on airborne bacterial
concentrations. This method was shown to reduce airway
inflammation in exposed workers (16). Added oil in animal
feed was tested to reduce airborne dust with modest improve-
ment (17).
Dairy barns can also be highly contaminated with bacte-
ria. Hay and straw used for bedding are very important bio-
aerosol sources. More recently, shredded newspaper or
sawdust has replaced straw in some regions. Hypersensitivity
pneumonitis, or farmer’s lung disease, can develop in dairy
farmers and is caused by, among other substances, a subset
of four thermophilic mycelial bacteria, Saccharopolyspora rec-
tivirgula, Saccharomonospora viridis, Thermoactinomyces sac-
chari, and Thermoactinomyces vulgaris, growing in moldy hay
and straw (18). Culturable S. rectivirgula can be found at con-
centrations in excess of 103 CFU/m3, and up to 107 16S cop-
ies/m3 were detected for S. rectivirgula using a molecular
approach for quantification (19). Factors such as hay packing
techniques (20) and hay preservatives (21) have been studied
for their effects on the colonization of hay by thermophilic
actinomycetes. Large cylindrical bales support the growth of
these bacteria very well, and bacterial hay preservatives are
not efficient in reducing the airborne concentration of bacte-
rial aerosols. Barn ventilation is inadequate for reducing the
airborne load of actinomycetes spores (20). Reports from Fin-
land and France proposed that molds play a major role in
farmer’s lung disease but that S. rectivirgula, which is classi-
cally incriminated, does not (22, 23). These findings have
not been confirmed in other countries where S. rectivirgula
is still considered the main etiologic agent for farmer’s lung
disease. Pathogens carried by cattle can be found in the
barn environment, including in the air. Environmental con-
tamination with Mycobacterium avium subspecies paratubercu-
losis (MAP) was detected in a barn sheltering cows shedding
this pathogen (24, 25). The persistence of MAP was studied
after carrier animals left the barn and a complete cleaning was
undertaken. This study showed the difficulty of eliminating
this bacterium.
Manure spreading can produce bacterial aerosols, which
can be propagated to the surrounding areas. Bacterial aerosols
were quantified 2 m and 20 m downwind of manure spreading
and culturable bacteria were found in air samples collected
during the biosolid application process (26, 27). However,
most of the bacterial species present in the biosolid were
not detected in air samples. Die-off of the bacterial cells could
be responsible for those observations.
Workers in grain elevators and animal feed production are
exposed to very high levels of bioaerosols and bacterial con-
centrations up to 109 CFU/m3 have been reported in silos
(28, 29). Bacterial species in these environments include
Bacillus spp. and Gram-negative organisms (Pseudomonas
spp., Alcaligenes spp., Citrobacter spp., and Klebsiella spp.).
Thermophilic actinomycetes have been isolated from grain
samples but were not quantified in the aerosols (28). Hyper-
sensitivity cases among corn workers have been described, but
mold seems to be the more important etiological agent (30).
There are many opportunities for exposure reduction in this
industry with improved application of local ventilation sys-
tems. This approach is most effective when the source of bac-
terial aerosols is truck or rail car unloading, grain conveyance
systems, or feed bagging operations.
Residents and workers at sites where composting is prac-
ticed frequently express concern about bioaerosol emissions
and exposures. Exposure monitoring has demonstrated that
when the compost pile was being turned, shredded, or loaded,
bioaerosols were generated with concentrations approxi-
mately 2 logs higher than background levels and the organ-
isms were detected 40 m downwind of the facility (31).
Release of bioaerosols from compost piles was evaluated at a
green waste composting facility. Experiments conducted on
the surface of static piles generated specific bioaerosol emis-
sion rates at ground level, between 13 and 22 × 103 CFU/
m2 of mesophilic actinomycetes (32). The bioaerosols emit-
ted from a suburban yard waste composting facility were meas-
ured, and high bacterial concentrations were detected (up to

7 × 104 CFU/m3). Concentrations of total bacteria, Gram-
positive bacteria, Gram-negative bacteria, and actinomycetes
demonstrated decreasing concentration with distance from
the compost pile and higher downwind versus upwind con-
centrations (33). However, the results suggest that despite
the high exposure of workers, there was not an elevated risk
for residents living in the vicinity of composting facilities
because the concentrations recorded downwind of the opera-
tional area were not significantly different from background
levels (31).
Exposure to compost and potting soils poses an increased
risk of infectious diseases to horticultural workers. An out-
break of Pontiac fever was described in a nursery in New
Zealand (34). Cultural Legionella ssp. were detected in 61 of
88 compost samples from four organic waste facilities in
Switzerland (35). The majority of these were identified as
L. pneumophila. Bioaerosol sampling 5 m from compost piles
were positive for L. pneumophila in 11% of samples. Thus,
composting operators and horticultural workers should be
considered at-risk populations when tracking Legionella
infections.
Workers in the solid waste industry are exposed to bioaer-
osols (36, 37), leading to gastrointestinal and respiratory ill-
nesses (38–41). Studies have indicated a high prevalence
(14% of workers with cough and 17% with nasal irritation)
of respiratory illnesses and inflammation, and these work-
related illnesses are related to bacterial exposure with up to
105 CFU/m3 (42). Adverse health effects including organic
dust toxic syndrome were described in workers in the waste
recycling industry during manual sorting of industrial and
household waste (43). Although bacterial aerosols have
been described in this type of environment, emphasis has
been put on airborne and dust-borne endotoxin and glucan
exposure (36, 44, 45). A Canadian study described bacterial
aerosols in household waste recycling plants (46). This study
showed that in summer, the average concentrations of total
bacteria were greater than 104 CFU/m3 in the receiving, sort-
ing, and shipping areas of some plants.
In municipal wastewater treatment plants the concentra-
tions of airborne bacteria typically range from 102 to 105
CFU/m3 depending on the area of plant, with sludge dewater-
ing areas reaching the highest levels. Airborne bacterial
concentrations of 102 CFU/m3 have been reported for Gram-
positive bacteria, Gram-negative bacteria, and thermophilic
actinomycetes (47–50). Commonly detected taxa include
Enterobacter cloacae, Acinetobacter calcoaceticus, Pseudomonas
spp., Stenotrophomonas maltophilia, Thermoactinomyces thalpo-
philus, and Thermoactinomyces vulgaris. Various water treat-
ment methods were compared with regard to bacterial aerosol
releases and demonstrated that mechanical aeration of sludge
generates substantially higher bacterial aerosols than diffused
aeration systems, which pose a lesser hazard for human health
(51). Pathogenic bacteria were among airborne bacteria
found in Polish wastewater treatment plants. (50). Gram-
positive bacteria prevailed with corynebacteria accounting
for over half the culturable mesophilic bacteria. Staphylococ-
cus spp., Micrococcus spp., and Bacillus spp. were also preva-
lent. Among the Gram-negative genera, Enterobacter,
Acinetobacter, Pseudomonas, and Stenotrophomonas domi-
nated. Antibiotic-resistant bacteria have also been described
in wastewater treatment plants (52, 53), and airborne Legion-
ella pneumophila has also been reported (54). Reuse of partially
treated domestic wastewater for irrigation is a common prac-
tice in arid regions (55). Exposure to potentially pathogenic
bacteria is a concern as bacteria from the wastewater become
airborne (56).
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-3
ENVIRONMENTS WITH MODERATE
CONCENTRATIONS OF BACTERIA AND
ENDOTOXIN
The cotton textile, vegetable and seed processing, machin-
ing, forest products industries, and slaughterhouse facilities
are environments with moderate concentrations of bacteria
and endotoxin. Outdoor environments may also contain
high airborne concentrations of bacterial contaminants
resulting from their proximity to bacterial-laden industries
or activities. Cotton and wool workers are exposed to airborne
Gram-negative bacteria and endotoxin (57), but dust control
measures are effective for reducing exposures to bacteria and
endotoxin as evidenced by improved worker health and com-
pliance with exposure guidelines.
The relationship between bioaerosol exposure in the
potato starch industry and work-related respiratory symptoms
has been described (58), and it seems to be linked to dust
exposure. However, bacterial exposures have rarely been
described. Airborne microorganisms have been investigated
in the sugar beet processing industry, and workers may be
exposed to high culturable microorganisms, mostly Gram-
positive bacteria (59). Cases of organic dust toxic syndrome
were reported in agricultural seed processing and exposures
to bacteria, mold spores and endotoxin were documented
(60). High risk tasks include dumping, mixing, and sieving.
Local exhaust ventilation systems can improve the air quality
in those plants if properly designed and operated.
Soluble oil metal-working fluids (MWFs) used in machin-
ing plants are susceptible to bacterial colonization since water
used for the emulsion creates an optimal environment for
growth. Bacterial contamination is a problem for the industry
because it changes the biophysical properties of the MWF
reducing the system efficiency during metal boring, hobbing,
tapping, grinding, and honing operations (61). Aerosols are
generated during metal machining from systems that contain
MWFs, tramp oils, additives, emulsifiers, and bacteria that
colonize the fluids. Although biocides are used in MWFs,
the bacterial diversity of contaminants is complex and Pseu-
domonas spp., Micrococcus spp., and Mycobacterium spp. are
important representatives (62, 63). Other taxa have also
been described in MWF bioaerosols including Enterobacter
agglomerans, Alcaligenes faecalis, Streptococcus spp., Staphylo-
coccus spp., Corynebacterium spp., Streptomyces spp., other
Gram-positive bacteria, and molds and yeasts (62, 64, 65).
The presence of mycobacteria in MWF has been associated
with hypersensitivity pneumonitis outbreaks (66, 67), and
recent animal studies have clearly shown that Mycobacterium
immunogenum in MWF can induce the lung pathology and T-
lymphocyte profiles consistent with hypersensitivity pneu-
monitis (68, 69).
The forest products industry is often confronted with
health and safety issues surrounding the control of bioaerosol
exposures. These bioaerosols are mainly comprised of wood-
decaying molds, although wood dust, bacteria, respiratory
allergens, and carcinogens from the wood are also found
(70). Western red cedar asthma, sequoiosis, and maple bark
disease are respiratory illnesses associated with wood industry
bioaerosols. The link between airborne bacteria and diseases
in the wood industry is less clear, and molds are the primary
microbial agents of concern, but microbiological air sampling
performed in sawmills demonstrated bacterial levels up to 105
CFU/m3 (71–74). Debarking operations were the most highly
contaminated sites with molds, bacteria, and endotoxins (71,
75). More than 34 species or genera of bacteria were identified
in the Polish coniferous wood industry (76). Bacillus spp. and
3.2.6-4 ▪ AIR
the Pseudomonadacea family are the predominant bacteria
found in sawmills (74).
Paper production processes involve handling and transfor-
mation of wood products such as wood chips, timber logs, or
recycled paper fiber. Those products may have microbial con-
tamination. When dispersed into the air during production,
bacteria and molds may be inhaled by workers and induce
airway inflammation (77), and endotoxins are important
contaminants of this workplace (78). Bacterial bioaerosol
concentrations of 310 CFU/m3 have been reported with spe-
cies from the Enterobacteriaceae family the most predomi-
nant taxa.
As with livestock production facilities, slaughterhouses
have significant bacterial aerosol contamination, some of
which may be pathogenic. The main concerns in this setting
are minimizing contamination of the product and protecting
the health of the workers. In a recent study, the highest counts
of airborne bacteria were found in the receiving/hanging
and defeathering stations of poultry slaughtering facilities.
Concentrations were lower in the processing sites with chilled
air (79). Areas with intense animal movement generated the
highest bacteria concentrations. S. aureus (79) and antibiotic-
resistant E. coli (80) were among the pathogens present in
the air of poultry slaughtering plants. In beef processing
plants, E. coli O157:H7 and Salmonella enterica were commonly
detected in air samples (81). A study performed in turkey,
chicken, and duck houses as well as duck slaughterhouses
revealed that a remarkable proportion, up to 18%, of the 16S
sequences were affiliated with Jeotgalicoccus spp. in bioaerosol
samples (82). Microbial cross-contamination occurs by aerosol
or large airborne droplets and airborne particulates and feathers
(4). Levels of contamination in the back-splitting areas (1,800
CFU/m3) were generally higher than that in the weighing areas
(1,070 CFU/m3) and showed that airborne bacteria have an
important role in carcass contamination (83). Jericho et al.
evaluated the bacterial aerosols generated at the high-line
speed cattle abattoir operations and studied the importance
of air sampling in the Hazard Analysis and Critical Control
Point plan. They concluded that the control of aerosols should
be treated as a critical control point (84). Airborne bacteria
in chilling facilities of poultry abattoirs have also been studied
(85). Bioaerosols were extremely high and Micrococcaceae and
Gram-positive irregular rods were detected in addition to strep-
tococci and members of the Enterobacteriaceae and Micrococ-
caceae. The authors concluded that cross-contamination may
occur in evaporative chilling process.
Poultry slaughterhouse workers were studied for respiratory
health and exposure to bioaerosols including airborne bacte-
ria. Up to 106 culturable airborne bacteria were detected,
mainly from coagulase-negative staphylococcal strains. An
overshift increase in respiratory symptoms was found but did
not correlate with the individual exposure to bioaerosols
(86). The presence of aerosolized Salmonella spp. and Escher-
ichia coli inside and outside commercial layer facilities was
studied, and the results showed that those bacteria are present
inside such facilities and outside up to 13 m from the ventila-
tion fans (87). The role of live birds on bioaerosols generation
has been demonstrated, and up to 105 CFU/m3 were detected
where fecal matter was removed using a disposal system featur-
ing a central opening in the floor through which the matter
automatically dropped to an open-air lower level. Marginally
lower bioaerosol concentrations were detected when con-
veyor belts were used (88).
In outdoor settings, highly contaminated substrates such
as pig manure, wastewater, and compost facilities may intro-
duce vast quantities of microorganisms into the air, which
can then spread over a wide area. It has been established
that the air in rural areas where swine CAFOs are present
may have high concentrations of odors and vapors from
microbial degradation of manure and airborne microbes
(89). Other factors such as the presence of various contami-
nation sources and the distance from them, relative humidity,
wind, and human activities also contribute to the dispersion
of contaminants and may influence the outdoor airborne
microbial burden (89).
Cooling towers have been shown to be responsible for
Legionnaires’ disease outbreaks in France in 2003–2004,
and several people who were infected by the responsible cool-
ing tower were 1 to 3 km away (90). In Quebec, Canada, in
the summer of 2012, 180 cases, including 13 fatalities, of
Legionnaires’ disease were diagnosed and linked to a contami-
nated cooling tower. Legionella cells may be transmitted
through the air over a long distance. The normal microbial
content of outdoor air also varies in nature and concentration
depending on the time of the day, season, the temperature,
and other factors. A background range, within which varia-
tions in microbial content would be considered normal,
may be established in a given geographical area. This outdoor
contamination by bacteria often serves as a comparison point
in any indoor airborne bacteria analysis.
ENVIRONMENTS WITH LOW
CONCENTRATIONS OF BACTERIA AND
ENDOTOXIN
Day cares and schools are grouped with low contamination
environments, but they are important settings associated
with children’s exposure to bioaerosols because they spend
large amounts of time within those facilities and children
often harbor more bacteria than adults. Several studies have
looked at mold exposure in schools after water damage but
limited information is available regarding bacterial exposure
in schools. Generally school classrooms are contaminated
with airborne bacteria, but their concentration rarely exceeds
a few hundred CFU/m3 (91). One study using molecular
approaches to assess bacterial bioaerosols in occupied univer-
sity classrooms showed that levels may reach 105 16S copies/
m3 (92). Fox et al. demonstrated that occupied classrooms
have a higher level of bacterial markers (lipopolysaccharides
and muramic acid) and dust than unoccupied rooms, and that
the release of bacteria from children could be responsible for
this phenomenon (93). It has also been shown that naturally
ventilated classrooms generally have higher airborne bacte-
ria than mechanically ventilated rooms (94). In this study
of 39 schools, staphylococci, micrococci, Bacillus spp., and
corynebacteria were detected. Personal exposure of school-
teachers to bioaerosols was assessed, and it was shown that
when the workplace is heavily crowded, airborne bacterial
load is higher (95). Ventilation and air intake problems
have been studied in public schools, demonstrating that
occlusion of air intakes negatively pressurized the building
and caused sump vents to become air intakes (96). This
poor air handling system maintenance resulted in contamina-
tion of classrooms by Bacillus spp. with up to 760 CFU/m3
detected.
Office environments have been studied to assess the influ-
ence of indoor environmental quality on physical and psy-
chological health impacts. Bioaerosols in office buildings
have been evaluated in commercial buildings and the factors
influencing indoor environmental quality assessed (97),
including culturable and total bioaerosols. Another study pre-
senting the bacterial aerosols in air-conditioned office

buildings in a cold climate found low (<103 CFU/m3) bacte-
rial concentrations (98). Total bioaerosols (including bacte-
ria) reached up to 104 organisms/m3 and little difference was
reported for culturable bacteria between indoor and outdoor
concentrations.
Air of dental clinics may be contaminated by airborne
microorganisms (99–101) of various sources, such as from
the mouth of patients (102) and dental unit waterlines
(103, 104). The use of high-speed hand pieces (e.g., ultra-
sonic scaler, rotating devices) generate aerosols and dental
unit water lines are susceptible to the formation of biofilms
(105–107). Several pathogens and opportunistic pathogens
are often detected such as Legionella spp., Mycobacterium
spp., and Pseudomonas spp. (106, 108–110).
Airborne bacterial concentrations in homes have been
described in a few studies (67, 111–114), and Sessa et al.
reported that airborne bacterial concentrations were two-
to fivefold higher in the presence of people than in empty
homes (67). They suggested consideration of occupancy as
a step to identify acceptable levels for bioaerosols in common
indoor environments. Actinomycetes and their indoor/out-
door ratio were evaluated across seasons in a study by Lee
et al. (113). Actinomycetes were detected indoors with con-
centrations that differed only slightly from outdoor levels.
Streptomyces spp. have been proposed as an indicator of water
damage in houses (114) and basements have been studied in
which 41 compliant and noncompliant houses were com-
pared for their bioaerosol content (111). Culturable bacteria
were detected at 970 CFU/m3 on the main floor and 1,100
CFU/m3 in the basements, suggesting that the basement
represented an important amplification site for compliant
homes. Homes flooded in Cedar Rapids, Iowa, undergoing
remediation had a geometric mean living room airborne bac-
teria concentration of 710 CFU/m3 while the concentra-
tion for homes still in the process of remediation was 1,100
CFU/m3 (112).
Humidifiers that hold stagnant water quickly become con-
taminated by bacterial biofilms. Although, the infection risk
is associated with exposure to Legionellaea-contaminated
bioaerosols (115, 116), actinomycetes are also sometimes
incriminated. Water droplets released by humidifiers can
contain bacteria and the size of the droplets favors lung dep-
osition. In an experimental study, it was shown that humidi-
fiers release bacteria when the humidifier reservoir was
contaminated with bacteria (117). Although several papers
report cases of humidifier lung, very few draw conclusions
about the etiological agents.
ARCHAEA
Archaea are microorganisms devoid of a nucleus that make up
one of the three domains of life. The archaeal domain is
composed of two main groups, the Euryarchaeota and the
Crenarchaeota. The Euryarchaeota group contains the meth-
anogenic and halophilic (highly salt tolerant) species, as well
as several thermophilic organisms. The Crenarchaeota phy-
lum includes thermophilic and hyperthermophilic Archaea
(118–120). Archaea possess characteristics from both
bacteria and eukaryotes. Their genomic processing systems
(replication, transcription, translation, and DNA repair)
resemble that of eukaryotes, whereas their metabolism is
more like that of bacteria (118, 120, 121). Archaea have dis-
tinct characteristics such as the capacity of several to survive
extreme conditions, their resistance to some antibiotics, a cell
wall devoid of peptidoglycan, and unique membrane lipids
(122–125).
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-5
Until recently, Archaea were considered as extremophile
microorganisms only living in hostile environments such as
hot springs, acid water sources, or salt lakes. Their culture
conditions are often very stringent and therefore Archaea
were rarely characterized in human environments. However,
through molecular biology techniques, these microorganisms
have now been detected in soil (126, 127), water (128), and
the gastrointestinal tract of ruminant animals and humans
(129) and are now considered ubiquitous organisms.
Human Exposures to Archaea
Archaea were identified in bioaerosols sampled from swine,
dairy, and poultry barns (19, 130, 131) as well as in industrial
facilities such as wastewater treatment plants. Archaea are
usually found in environments with high concentrations of
airborne bacteria and endotoxin, but their concentrations
are generally two to three orders of magnitude lower than
for bacteria. Their airborne concentrations range from 108
Archaea/m3 in swine barns to 106 Archaea/m3 in dairy and
poultry barns. Airborne archaeal numbers were lower in
wastewater treatment plants where 104 Archaea/m3 were
detected. Because bioaerosol-related Archaea are mainly
from fecal sources in work environments, most archaeal spe-
cies found in bioaerosols are methanogens, producing meth-
ane in anoxic conditions, such as the ruminant gut. Species
of Methanobrevibacter sp. (M. smithii, M. ruminantium,
M. wosei) and Methanosarcina sp. (M. mazei, M. siciliae), as
well as Methanosphaera stadtmanae, were found in the air of
agricultural settings (19, 130, 131). Archaeal species detected
in bioaerosols from wastewater treatment plants were Metha-
nospirillum hungatei, Methanocorpusculum labreanum, and Hal-
oarcula marismortui.
Humans are exposed to Archaea via their own intestinal
microflora (132, 133) but also via inhalation of airborne
Archaea from bioaerosols of various work facilities. These
Archaea are mainly methanogenic species from fecal sources
and are thus likely killed by the aerosolization process due to
their high oxygen sensitivity. Therefore, they have little
infectious potential but could be involved in diseases arising
from sensitization or innate immune modulation. Since bur-
dens of bioaerosols from livestock facilities are very high and
other microbial components of these bioaerosols have been
shown to induce respiratory diseases, Archaea could play a
similar role. For example, Archaea could be associated with
phlegm symptoms reported by cage-poultry barn workers
(134). These workers are exposed to lower concentrations
of total bacteria and endotoxins than floor-poultry barn
workers (135), but to a much higher archaeal burden (130).
Since Archaea were only recently recognized as an important
component of bioaerosols in animal facilities, more data are
needed to establish their impact on respiratory health in
agriculture.
Murine Studies for Archaea Toxicity
The immunogenic potential of Archaea in the lungs was
studied in a chronic airway-exposure mouse model (136).
This model was based on an established murine model for
hypersensitivity pneumonitis (68, 137, 138). Mice were
exposed for 3 weeks to either one of two archaeal species
found in bioaerosols of agricultural settings, M. smithii or
M. stadtmanae. Although both Archaea showed pulmonary
immunogenic properties, only M. stadtmanae induced inflam-
mation. Archaea found in bioaerosols must be further studied
to determine if the immune responses they induce are tolero-
genic or proinflammatory.
3.2.6-6 ▪ AIR
Molecular Biology Methods for Analysis of
Airborne Archaea
Molecular biological methods are very specific and efficient
and can be used to detect and quantify microorganisms inde-
pendent of their culturability (139, 140). They are the only
efficient methods for airborne archaeal detection, since
most Archaea found thus far in bioaerosols are sensitive to
oxygen and are unlikely to grow on culture media after aero-
solization. Quantitative PCR (qPCR) allows real-time quan-
tification of total bacteria, Archaea, and fungi and has
been applied in agricultural and industrial bioaerosol charac-
terization studies using universal primers and organism-
specific probes (19, 131, 141–143). qPCR was also used for
bacterial and fungal genus- or species-specific detection
(144, 145) and for other genes such as antimicrobial resist-
ance genes (146). The most studied genes for bacterial and
archaeal detection are the 16S ribosomal RNA fragments.
These genes are composed of highly conserved regions of
the bacterial and archaeal domains and can thus be used for
species-level identification. Apart from quantification,
molecular fingerprinting methods such as denaturing gradient
gel electrophoresis, temperature gradient gel electrophoresis,
single-strand conformational polymorphism, terminal restric-
tion fragment length polymorphism, and automated riboso-
mal intergenic spacer analysis are used to characterize
bacterial and archaeal biodiversity of bioaerosols (19, 130,
131, 135, 142). These methods have limited sensitivity and
poor detection limits, which can limit species detection.
Cloning the 16S PCR amplicons and sequencing the inserts
after building a 16S gene clone library is also useful for
biodiversity characterization, but this method is time-
consuming. Finally, pyrosequencing allows identification
of all targeted microorganisms in a sample without bias
but is very expensive (147). Rapid innovation holds prom-
ise for further improving quantitation of Archaea in the
airborne environment.
ENDOTOXIN
Endotoxins are a family of surface glycolipids of Gram-
negative bacteria that are ubiquitous in the environment.
They are regularly found in indoor and outdoor air, water,
soil, and food. They are amphipathic molecules bearing a
structurally unique hydrophobic lipid A moiety embedded
in the outer membrane connected by keto-deoxy-manno-
octanoic acid to a polysaccharide moiety (O antigen) that
protrudes from the microbial membrane into the extracellular
space. The lipid A component of the molecule is conserved
across bacterial species and carries the toxic potency.
Throughout evolution, a wide range of vertebrate and non-
vertebrate species have developed highly sensitive responses
to endotoxin that promote host defense against invading
Gram-negative bacteria. Recognition of endotoxin leads to
the induction of inflammation, targeting of Gram-negative
bacteria for elimination, and the subsequent clearance of
the endotoxin (148–150). While these responses guard
against colonization and infection, they can also induce path-
ology when exposure or subsequent responses to endotoxin
are inadequately controlled.
Exposure to endotoxin occurs on a daily basis from nonin-
fectious and infectious Gram-negative bacteria in both occu-
pational and nonoccupational environments. Endotoxin is a
component of particulate matter in ambient air (151–154), in
air and reservoir dust in homes (155–159), office buildings
(97), and in a variety of industrial and agricultural settings
(6, 160). Inhalation of endotoxin has been linked to the
induction and exacerbation of asthma and other airway dis-
eases (157, 159, 161). High levels of exposure found in a vari-
ety of occupational environments lead to a high prevalence of
respiratory and systemic symptoms (43, 162).
The hazards of inhaled endotoxin first became apparent
from studies of lung disease among workers exposed to
vegetable and cotton dust (163, 164) and were later refined
with measured endotoxin exposures (57, 165, 166). Studies
in the 1990s characterized pulmonary responses, including
lung inflammation and cytokine upregulation, on inhalation
of endotoxin and organic dusts containing endotoxin in
humans (167–171) and experimental animals (167, 172–
174). In the past decade there has been a focus on endotoxin
exposures in the domestic environment with large studies in
the United States and Europe (155, 156, 159, 175–178).
Endotoxin Structure and Recognition
Each Gram-negative bacterium contains approximately 106
endotoxin molecules in its outer membrane (179). When
released from bacteria through membrane fragmentation or
blebbing, the lipid A regions are sequestered within aggre-
gates. The lipid A regions are covalently bound to a carbohy-
drate chain of variable length and composition. Endotoxins
exist as lipopolysaccharides (LPS) or lipooligosaccharides
(LOS) depending on the genus of bacteria from which they
arise. A variety of mammalian proteins confer innate immune
recognition of Gram-negative bacteria by binding to sites
within the conserved inner core region and lipid A regions.
The lipid A portion of endotoxin consists of a diphosphory-
lated β-1,6-D-glucosamine disaccharide linked to 3-hydroxy
fatty acids that are further substituted by nonhydroxylated
fatty acids. The fatty acids of the lipid A region are the key
determinants of the potency of endotoxin (148). Hexa-
acylated endotoxins including LPS of E. coli and LOS of
Neisseria meningitides are capable of inducing cytokine upre-
gulation at picomolar concentrations.
Specific endotoxin recognition and signaling requires
ordered interactions involving several extracellular and
cell surface host proteins, including LPS-binding protein
(LBP), CD14, MD-2, and Toll-like receptor (TLR) 4
(180). Murine exposure studies demonstrate that TLR4-
dependent responses to endotoxin in the resting airway are
mediated primarily by alveolar macrophages with less
involvement of airway epithelia. Human and mouse studies
have shown that aggregates of endotoxin bind to LBP and
are disaggregated and then attach to CD14 as individual
molecules. The CD14:endotoxin complex then interacts
with MD-2, sending a signal via TLR4 leading to the upregu-
lation of NF-kB via either a MyD88-dependent or a MyD88-
independent pathway (181, 182). This signal then induces
production of proinflammatory cytokines. Thus, the action
of LBP, CD14, MD-2, and TLR4 convert each environmen-
tal endotoxin particle to many activating protein:endotoxin
complexes. MD-2 is likely limited in the resting airway
and only arises when myeloid cells are recruited to the lung
and airway epithelia are upregulated, leading to amplified
responses to endotoxin. This organization of unresponsive-
ness to low levels of endotoxin in the resting airway may be
advantageous, promoting host responses to invading bacteria
while minimizing inflammatory reactions to more incidental
exposure to endotoxin-bearing environmental pollutants
such as ambient air pollution. Tissue sites that routinely
come into contact with free endotoxin aggregates, such as
the gut, are hyporesponsive to endotoxin apparently due to
low expression of MD-2 and TLR4 (183, 184).

Endotoxin Inhalation and Lung Inflammation
Acute inhalation exposure to purified endotoxin or to partic-
ulate matter containing endotoxin can induce airflow
obstruction, airway hyperreactivity, inflammation, and sys-
temic symptoms including fever, chills, myalgia, and malaise.
The inflammation is marked by the appearance of neutrophils
and pro-inflammatory cytokines in lung lavage (174, 180)
and includes clinical conditions such as asthma, asthma-like
syndrome, organic dust toxic syndrome, and byssinosis (43).
Chronic exposure may induce changes resembling allergic
asthma and may also contribute to chronic obstructive pul-
monary disease (157, 185). Endotoxin inhalation in mice
induces similar host responses and changes (172, 180, 186–
188) and thus provides a suitable animal model for character-
ization of the molecular and cellular determinants of lung
responses to endotoxin. Studies have provided time course
(167, 180) and dose-response relationships (174) for induc-
tion of inflammation after endotoxin inhalation and have
demonstrated the roles of LBP, CD14 and TLR4 in the patho-
genesis of endotoxin-induced disease (180, 189, 190).
Variability of Endotoxin Responsiveness
Strain variability to inhaled endotoxin in mice and reprodu-
cible differential responsiveness between humans provide
evidence of genetic determinants of individual susceptibility
to inflammation induced by endotoxin. Thorne et al. devel-
oped dose-response curves for lung inflammation induced by
endotoxin inhalation in a responsive strain of mice (C3HeB-
FeJ) and a hyporesponsive strain (C3H/HeJ [Tlr4Lps-d]) and
showed that the hyporesponsive mice had 1,000-fold lower
neutrophil recruitment to the lung for the same exposure
(174). These C3H/HeJ (Tlr4Lps-d) mice have a spontaneous
mutation in the LPS response locus (the toll-like receptor 4
gene, Tlr4Lps-d) that effects binding of endotoxin:MD2 to
TLR4. Hadina et al. showed that endotoxin-induced respon-
siveness was absent in MD-2 null and CD-14 null mice and
that penta-acylated endotoxin produced diminished inflam-
matory responses (180). In human studies, Kline et al. dem-
onstrated that some normal subjects exposed to increasing
inhaled doses of endotoxin had marked hyper- or hypores-
ponsiveness (169) and that this was associated with a poly-
morphism in TLR4 (191). A study of 24 candidate single
nucleotide polymorphisms in 408 agricultural workers
showed modification of the relationship between endotoxin
exposure and pulmonary function by genetic variants in
CD14 and MD2 (192). Polymorphisms in control pathways
for other host molecules likely exist that also effect endotoxin
responsiveness.
Diseases Associated with Endotoxin Exposure
Exposure to inhaled endotoxin is associated with several dis-
tinct clinical syndromes, including occupational and nonoc-
cupational asthma, asthma-like syndrome, organic dust toxic
syndrome, chronic bronchitis, and increased severity of
hypersensitivity pneumonitis.
Asthma
Asthma is a chronic pulmonary disease often punctuated by
episodic exacerbations. It is characterized by eosinophilic or
neutrophilic inflammation and hyperresponsive airways. Dra-
matic increases in the prevalence and morbidity of asthma
were observed in the 1980s and 1990s, especially in devel-
oped nations with higher standards of living (193). It is clear
that exposures to common allergens are an important factor in
asthma induction. However, exposure to endotoxin is also
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-7
important, although its effects in the induction and progres-
sion of asthma appear to be complex and may depend on the
dose and timing of exposure as well as the particular health
outcome examined (159, 161, 194–199).
Asthma-like Syndrome
Asthma-like syndrome is a nonallergic respiratory disease that
occurs acutely from inhalation of organic dust containing
endotoxin. This syndrome occurs among livestock confine-
ment workers, grain and cotton textile workers, sewage treat-
ment plant workers, and compost workers. The syndrome has
been called byssinosis in cotton workers, bagassosis in sugar-
cane workers, and porcinosis among swine workers. It
includes neutrophilic lung infiltrates and production of
inflammatory cytokines leading to cough, chest tightness,
wheeze, and shortness of breath (6). Lung function declines
across a work shift are common. Exposures to endotoxin
above 200 endotoxin units (EU)/m3 are generally needed
to induce this disease (200). Byssinosis is a similar syndrome
associated with exposures in the cotton textiles industry (57,
201). Its features include acute bronchoconstriction, chest
tightness, and dyspnea that are worse on returning to work
after a weekend. Long-term exposure leads to chronic bron-
chitis and airflow limitations. Much of the effect is attribut-
able to endotoxin exposure, although other microbial cell
wall and plant components likely play a role (202, 203).
Organic Dust Toxic Syndrome
Organic dust toxic syndrome (ODTS) is an acute inflamma-
tory condition of the small bronchioles and alveoli associated
with high exposure to endotoxin (>1,000 EU/m3), fungi, or
both. Symptoms include fever, malaise, dyspnea, and fatigue
that typically peak 4 to 8 h after exposure and resolve within
48 h (6). A European study of pig farmers found a lifetime
prevalence of ODTS of 23% (204). This was lower than an
earlier U.S. study of swine farmers that reported 34% lifetime
prevalence (205). Self-reported ODTS in the past 12 months
was assessed in 877 Dutch farmers and agricultural industry
workers at 18.6% (206). This was considerably higher than
reported for veterinary students (1.9%) or Danish farm
apprentices (0.6%). However biofuel workers also experi-
enced a high incidence with 14.9% reporting ODTS in the
past 12 months.
Hypersensitivity Pneumonitis
Hypersensitivity pneumonitis (HP), also known as extrinsic
allergic alveolitis, is an allergic disease associated with
thermophilic bacteria, fungi, and bird antigens (207). Cases
of HP have arisen in recent years among metal machin-
ists exposed to water-based MWFs treated with certain bio-
cides. Investigation of these outbreaks suggested a role for
a particular organism, Mycobacteria immunogenum, found
growing in the MWFs. Laboratory study has shown that
M. immunogenum in MWFs can induce HP and that endo-
toxin coexposure with M. immunogenum augments the sever-
ity of the HP (68).
Occupational Exposures to Endotoxin
Occupational exposures to endotoxin have been evaluated
extensively beginning with the cotton industry and the live-
stock industry in the early 1980s (2, 57). Table 2 provides a
summary of endotoxin exposure levels measured in a variety
of occupational environments with a focus on larger studies
that used current methodology. The table demonstrates that
concentrations of airborne endotoxin are typically highest
for livestock confinement facilities, reaching geometric
3.2.6-8 ▪ AIR

TABLE 2 Endotoxin concentrations in occupational environments

Geometric mean endotoxin
Setting Size fraction N Reference
concentration, EU/m3
Swine operations Area CFCa filter 81 8,290 (244)
Swine operations Area inhalable filter (154)
Hoop barns 30 3,250
Conventional confinement barns 40 3,100
Swine operations Area impinger 21 4,385 (75)
Area CFC filter 21 3,927
Swine operations Personal inhalable filter 350 920 (256)
Poultry operations Area CFC filter 81 1,340 (244)
Dairy operations Personal inhalable filter 194 647 (250)
Area respirable filter 216 16.8
Soybean harvesting (closed tractor cab) Personal CFC filter 32 56 (257)
Agricultural seed processing industry Personal inhalable filter 100 1,800 (60)
Grain elevators Personal CFC filter 410 2,860 (29)
Personal respirable filter 410 83.2
Animal feed manufacturing Personal inhalable filter 530 12–285 (258)
Area inhalable filter 79 19
Potato processing Personal inhalable filter 195 9–102 (200)
Area inhalable filter 68 1–4,000
Sawmills Area impinger 59 740 (75)
Area CFC filter 62 188
Domestic organic waste composting facilities Personal inhalable filter 182 205 (45)
Domestic waste collection Personal inhalable filter 176 40 (45)
Wastewater treatment plants Personal inhalable filter 460 27 (259)
Wastewater treatment plants Area CFC filter 104 82.6 (49)
Engine machining plant Area CFC filters 48 54 (63)
Podiatry clinics Personal inhalable filter 12 9.6 (260)
Vegetable greenhouses Personal inhalable filter 69 320 (median) (261)
Animal facilities Area inhalable filter (233)
Swine 39 7,220
Chicken 44 4,080
Turkey 36 14,100
Dairy 44 102
Academic veterinary clinic stables Personal inhalable filter (262, 263)
Poultry 12 1,470
Ruminants 24 228
Horse 42 608
a
CFC: closed-face cassette.

mean levels ranging from 920 to 8,290 EU/m3 for swine,
1,340 EU/m3 for poultry, and 650 EU/m3 for dairy. Personal
inhalable fraction endotoxin exposures among U.S. grain ele-
vator workers had a geometric mean concentration of 2,860
EU/m3 and this was 10-fold higher than the mean level in
Dutch animal feed manufacturing facilities. A study of the
agricultural seed industry initiated by cases of ODTS meas-
ured a geometric mean of 1,800 EU/m3 with some measure-
ments above 100,000 EU/m3. Soybean harvesting, potato
processing, organic waste composting, and sawmills ranged
from about 100 to 1,000 EU/m3. Wastewater treatment plants
and metal machining plants using water-based MWFs fell
in the range of 27–83 EU/m3. Stables associated with a
large veterinary clinic had geometric mean concentrations
ranging from 228 to 1470 EU/m3. A substantial proportion
of workers in agriculture continue to face daily exposures
exceeding 200 EU/m3.
Exposure to Endotoxin Indoors
Although it has been over 50 years since the first recognition
that house dust contains endotoxin (208), a greater appreci-
ation that this source of endotoxin carried a health risk came
decades later with the observation of Michel et al. (157) that
the severity of asthma and the use of asthma medication in
clinic patients was associated with mattress and floor dust
endotoxin concentrations. A number of epidemiologic stud-
ies have drawn attention to endotoxin exposure in domiciles
and its role in asthma. Table 3 lists mean values of endotoxin
assessed through active air sampling, reservoir dust sampling,
and passive air sampling using an electrostatic dust collector
in homes and outdoor air. Michel et al. (157) studied endo-
toxin in reservoir dust in the homes of patients with stable
chronic rhinitis or asthma and noted a significant association
of mattress and floor dust endotoxin concentration, but not

mite allergen concentration, with the severity of asthma and
asthma medication use. Studies from Bavaria (156), Canada
(209), and Sweden (210) provided evidence that growing up
on a family farm leaves one less likely to develop allergies and
allergic asthma. This protective effect (the hygiene hypothe-
sis) has received support from studies demonstrating a shift
away from an allergic phenotype with early life exposure to
endotoxin (195). However, studies from Australia (211),
Europe (212), New Zealand (213), and the United States
(214, 215) found the same or higher asthma rates (i.e., no pro-
tective effect) between farm and nonfarm children in one or
more study groups. The role of endotoxin as a risk factor for
asthma and wheezing has been investigated in urban cohort
studies (155, 157, 159, 216–219) with the finding that higher
endotoxin in house dust was associated with higher preva-
lence of wheeze but lower allergy.
Further insights have come from studies of the agricultural
workforce, which experiences a wider range of endotoxin
exposures. Smit et al. found that endotoxin exposure was pos-
itively associated with bronchial hyperresponsiveness and
wheeze (220). However, endotoxin exposure was also associ-
ated with diminished atopy and serum levels of specific IgE to
grass pollen. A study of agricultural workers exposed to a wide
range of airborne endotoxin concentrations demonstrated
that exposure exerted a significant protective effect for self-
reported allergy but significantly increased risk for developing
chronic bronchitis, ODTS, and, at the highest quartile of
exposure (>1,000 EU/m3), wheezing. Importantly, stratifying
subjects by whether they grew up on a farm did not alter the
risk for chronic bronchitis, ODTS, asthma, or self-reported
allergy (206, 221).
Intense study of immune regulation in early life has
attempted to explain the protective role of exposures to
microbial agents and endotoxin in the development of aller-
gies and allergic asthma. New insights into the role of pattern
recognition receptors demonstrate closer interplay between
innate and adaptive immune functioning (222). The obser-
vation of functional and structural mimicry between the
house dust mite allergen, Der p2, and the endotoxin binding
protein MD-2 shows that environmental allergens may stim-
ulate innate immunity through the Toll-like receptor signal-
ing pathway (223). T-regulatory cells have been identified as
key to the development of tolerance to environmental aller-
gens early in life. Failure to develop an appropriate level of
tolerance leading to a Th2 phenotype may be attributable
to an uncoupling of T-regulatory cells from Th2 effector cells
(224, 225).
In the National Survey of Endotoxin in Housing, Thorne
et al. analyzed endotoxin in 2,542 house dust samples from
five locations within homes that were systematically selected
to represent the demographics and housing stock of the entire
United States (159). Information on health and living condi-
tions was assessed through questionnaire and on-site evalua-
tion of 2,456 residents of 831 homes. Significant relationships
were observed between increasing endotoxin levels and diag-
nosed asthma, asthma symptoms in the past year, current use
of asthma medications, and wheezing among residents of the
homes. Those with elevated bedding and bedroom floor
endotoxin were nearly three times as likely to have experi-
enced recent asthma symptoms. Allergic subjects with higher
endotoxin exposure were no more likely to have diagnosed
asthma or asthma symptoms than were nonallergic subjects.
These authors then analyzed 37 candidate variables in a
repeated-measures analysis of variance to identify independ-
ent predictors of endotoxin (176). The major predictors of
household endotoxin were the sampling location within
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-9
the home, geographic region of the United States, educa-
tional attainment of the residents, presence of children, dog
ownership, problems with cockroaches, evidence of food
debris, and evidence of smoking in the home.
Flood waters deliver bacteria into submerged homes and
offer ideal conditions for microbial growth and endotoxin
production. When the flood waters recede, the endotoxin
remains. Samples taken in flood-damaged homes in New
Orleans after Hurricane Katrina demonstrated endotoxin lev-
els 10- to 50-fold higher than those normally observed in
homes (226). During remediation of homes, exposure levels
rose a further 10-fold. Seventy-three homes in Cedar Rapids,
Iowa, that were partially submerged in the flood of 2008 were
studied as they were being remediated or when remediation
was complete (112). Endotoxin contamination levels were
lower in these homes than in Katrina-flooded homes (GM
= 22.3 EU/m3) (227) and homes with remediation com-
pleted had lower levels of airborne inhalable endotoxin
(GM = 0.3 EU/m3) than homes where remediation was
ongoing (GM = 1.55 EU/m3). Remediation successfully
reduced exposures to acceptable levels, however, residents
were left with increased prevalence of wheeze and prescrip-
tion medication use for breathing problems compared to
before the flood (112).
Methods for the Assessment of Endotoxin
Limulus Amoebocyte Lysate Assays
Assessment of exposures to endotoxin entails sample collec-
tion, extraction from the matrix, and analysis of endotoxin
(228). Airborne concentrations of endotoxin are usually
determined from samples collected onto filter media. Reser-
voir dust or surface samples are most often collected using vac-
uum sampling onto filter media or electrostatic wipes (229).
Extraction is performed in an aqueous medium with shaking
or ultrasonifying, with or without Tween 20, usually without
added buffers. Methods comparisons for endotoxin extraction
protocols have been extensively studied (230–232).
Analysis for endotoxin most commonly employs the Limu-
lus amoebocyte lysate (LAL) assay (233). The LAL assay is
based on the extreme sensitivity of an enzymatic cascade in
amoebocytes taken from the hemolymph of horseshoe crabs
(Limulus polyphemus, Tachypleus tridentatus) and originates
from studies of bacterial infections in horseshoe crabs
(234). The LAL contains a mixture of proenzymes that are
activated by endotoxin in a cascading set of reactions via Fac-
tors C and B (235). Kinetic assays based on the LAL measure
spectral changes in microplate wells based on endotoxin
activation of enzymatic cleavage of a substrate (measured
at 405 nm) or the development of turbidity (measured at
340 nm) at a controlled temperature (228). A commercial
assay based on recombinant Factor C (rFc) has been intro-
duced that yields a fluorescent product of the reaction (exci-
tation 380 nm, detection 440 nm). Analysis of 402 field and
510 laboratory inhalable air sample pairs using the rFc assay
and the kinetic chromogenic LAL assay yielded correlation
coefficients of 0.93 and 0.86, respectively (P < 0.0001 for
both) (233).
In our laboratory, environmental samples are generally
extracted in sterile, pyrogen-free ( pf ) water or pf water with
0.05% Tween 20 for 1 h at 22°C with continuous shaking
(228, 233). Extracts are centrifuged and supernatants are
transferred into borosilicate glass tubes that have been heated
for 4 h at 200°C to remove endotoxin activity. They are then
analyzed using the kinetic chromogenic LAL assay. Twofold
serial dilutions of endotoxin standards and sample extracts are
3.2.6-10 ▪ AIR
prepared in sterile pf water. A 12-point calibration curve and
4-point endotoxin determination in duplicate is typically
performed. The standard curve ranges from 0.0244 to 50.0
EU/ml of standard endotoxin from E. coli 0111:B4. Lyo-
philized standard endotoxin, chromogenic substrate, and
LAL preparations are reconstituted using pf water. Aliquots
(100 μl) of the serial dilutions of endotoxin standards
and extracts are pipetted into a pf polystyrene microplate
and assayed via the addition of the LAL reagent and
substrate. The absorbance in each well is measured at
405 nm every 30 s for 90 min. Endotoxin determinations
are based on the maximum slope of the absorbance versus
time plot for each well. Four assay reagent blank wells serve
as reference and control for the pf status of the reagent
water, centrifuge tubes, pipette tips, and microplates. The
endotoxin value for a sample is typically calculated from
the arithmetic mean of those dilutions that fall in the
middle two-thirds of the standard curve. The detection
limit for airborne endotoxin measurements is approximately
0.02 EU/m3 (2 pg/m3) for a 1 m3 air sample eluted into a
10 ml extraction volume. Thus, the LAL assay is a highly
sensitive bioassay with physiologic relevance.
There have been international interlaboratory compari-
sons of sample extraction and endotoxin analysis protocols
for agricultural dust (236, 237), and cotton dust (238–240).
An additional study compared five different methods for
MWF air samples within one laboratory (241). The overrid-
ing conclusion of these is that intralaboratory variability is
generally low (<30%) and to minimize interlaboratory varia-
tion it is necessary to harmonize the sample extraction meth-
ods, harmonize the analysis methods, and match lysate
supplier and individual lysate lot. If these conditions are
met, laboratories experienced in the LAL will generally agree
within half a log and results are well correlated.
Endotoxin exposure assessment is influenced by the
method of sampling, sample storage conditions, the extrac-
tion method employed, the assay reagents, and methodology
(75, 242–245). For air sampling in occupational or domestic
settings, filter sampling of endotoxin-laden particulate matter
is typically used. Binder-free glass fiber filters are optimal in
terms of extractability of the endotoxin and the capacity for
dust collection but have less stable gravimetric properties
due to loss of glass fibers on O-rings within the sampler. Sam-
ple extraction has been evaluated extensively with conflicting
views expressed as to whether it is best to extract samples with
pf water, pf water with 0.05% Tween 20, or with added buffers
(230, 231). Three recent studies have systematically eval-
uated the performance of the LAL assay. Spaan et al. simulta-
neously collected 400 air samples in two occupational
environments to investigate the effects on airborne endo-
toxin concentration of filter type (glass fiber or Teflon), trans-
port conditions (with/without desiccant), sample storage
temperature (20°C or 4°C), extraction solution ( pf water or
pf water plus 0.05% Tween 20), extract storage temperature
(20°C or 4°C), and assay solution ( pf water or pf water plus
0.05% Tween 20). No differences were observed for transport
conditions and storage temperature of extracts. However, use
of glass fiber filters, storage of samples in the freezer, and
extraction in pf water plus 0.05% Tween 20 resulted in signif-
icantly higher estimated endotoxin concentrations (232).
Further work by these investigators led to the recommenda-
tion that airborne endotoxin samples should be extracted in
pf water plus 0.05% Tween 20 to obtain optimal endotoxin
yields as long as airborne and bulk dust samples can be diluted
50-fold or more in pf water to avoid interference of Tween 20
with the Limulus lysate (231).
In a recent study from our laboratory, we measured
14
C-labeled endotoxin in aggregates, membrane blebs, and
intact bacteria and compared these results to the endotoxin
activity in the LAL assay and in a murine in vivo lung inflam-
mation model (230). These experiments showed that neither
pf water, pf water with Tween 20, nor previously employed
extraction buffers effectively recovered endotoxin from
Gram-negative bacterial membranes even though endotoxin
presented in this form was fully able to induce lung inflamma-
tion in vivo. The inability of the LAL assay to fully detect bio-
active endotoxin in whole bacteria has troubling implications
for exposure assessment in environmental, epidemiological,
and laboratory studies where endotoxin often appears in dif-
ferent forms (i.e., aggregates, membrane blebs, or whole bac-
teria) (230).
Air sampling has been the mainstay of endotoxin exposure
assessment in occupational settings and because endotoxin
exposures are most strongly associated with airway diseases,
the inhalable fraction of the particulate matter is most
often sampled. The IOM, Button, and PAS-6 inhalable
dust samplers are suitable in most environments (246–249).
Several studies have performed both inhalable and respirable
endotoxin sampling (Table 2) and observed respirable con-
centrations 35- to 40-fold lower than inhalable endotoxin
(29, 250).
Reservoir dust sampling for endotoxin assessment has
been performed extensively in epidemiological studies of
asthma (see Table 3). While short-term air sampling may
be sufficient for occupational health studies, it is argued
that measurement of endotoxin in reservoir dust may better
represent long-term exposure than a single, short-term air
sample (159). In addition, surface sampling can be performed
quickly in a single home visit or by the residents themselves.
This approach generally consists of using a commercial vac-
uum cleaner fitted with a HEPA sock, extraction thimble,
or filter canister and vacuuming a defined area for a set
amount of time (e.g., 2 m2 for 5 min). The bulk dust is usually
sieved to remove hair, fur, sand, and other noninhalable
material, weighed, and then stored frozen to await extraction
and analysis. However, vacuum sampling of carpeted floors
and mattresses collects material that originated there and
was never airborne and therefore may lead to significant expo-
sure misclassification. Other surface sampling techniques
include wipe sampling of smooth surfaces with electrostatic
wiping cloths (229).
Two recent studies reported on the variability of reservoir
dust endotoxin within homes and lack of agreement with air
samples. Reports from the U.S. National Survey of Endotoxin
in Housing (159, 176) documented correlation coefficients
between sites within homes ranging from 0.22 to 0.42, dem-
onstrating that endotoxin levels at one location in the home
are not informative about another. Endotoxin concentrations
in 24-h air samples collected in family rooms were poorly cor-
related with family room floor endotoxin in a study of 85 Iowa
households (r = 0.10) (251). A study of 15 Boston-area
homes found no association of airborne and floor dust endo-
toxin concentrations in the bedrooms (158). These authors
reported higher variance within homes than between homes
for floor dust and airborne endotoxin.
Chemical Methods and Immunoassays
Chemical methods for quantification of endotoxin have been
developed employing gas chromatography-mass spectrometry
(GC-MS) (252, 253). These methods require special extrac-
tion procedures and assess the amounts of 3-hydroxy fatty
acids typically bearing 12–16 carbons. This method has
 3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-11

TABLE 3 Endotoxin concentrations in ambient, domestic and office environments

Mean
Setting Size fraction n Reference
concentration
Ambient air
Southern California (13 locations) PM10 filter 99 0.44 EU/m3 (153)
Munich, Germany PM2.5 filter 158 0.015 EU/m3 (151)
Hettstedt and Zerbst, Germany PM10 filter 42 9.0 EU/m3 (152)
Outside Iowa swine barns (all seasons) Area inhalable filter 7 hog barns (154)
33 m upwind 20 4.9 EU/m3
33 m downwind 20 136 EU/m3
165 m downwind 20 30 EU/m3
Indoor environments—air sampling
Office Buildings Area inhalable filter 24 1.1 EU/m3 (97)
Rural iowa households (LEIP Study) Play area inhalable filter 85 homes 5.82 EU/m3 (251)
326 samples

Houses of university faculty, staff, Bedroom filter 15 homes 0.64 EU/m3 (158)
students in Boston area 142 samples
Huts burning biomass fuel Inhalable filter (264)
Nepal – wood 16 43 EU/m3
Nepal – dung 15 365 EU/m3
Malawi – wood 4 113 EU/m3
Malawi - corn stover 2 1609 EU/m3
Indoor environments—reservoir dust sampling
U.S. National Survey of Endotoxin in Housing Reservoir dust 831 homes (159, 176)
Bedding 470 samples 20.6 EU/mg
Bedroom floor 588 samples 37.7 EU/mg
Living room floor 489 samples 71.1 EU/mg
Living room sofa 468 samples 45.3 EU/mg
Kitchen 454 samples 84.4 EU/mg
New York City Reservoir dust (175)
Bedroom floor 301 samples 75.9 EU/mg
Boston Reservoir dust (265)
Family room floor 404 samples 79 EU/mg
Bedroom floor 323 samples 63 EU/mg
Kitchen floor 245 samples 100 EU/mg
East Baltimore Reservoir dust (266)
Bedroom floors 85 samples 64.6 EU/mg
Boston Reservoir dust (267)
Homes – bedding & floor 118 samples 7.0 EU/mg
Schools – floors (median)
13.4 EU/mg
(median)
U.S. National Health and Nutrition Reservoir dust P.S. Thorne
Examination Survey Bed + bedroom floors 6963 samples 18.96 EU/mg unpublished
data
Munich, Leipzig households (LISA Study) Reservoir dust (155)
Bedding 1,884 samples 2.9 EU/mg
Bavarian, Austrian, Swiss farmhouses (ALEX Study) Reservoir dust (156)
Bedding 319 samples 37.8 EU/mg
Dutch households Reservoir dust (44)
Living room floor 99 samples 6.0 EU/mg
Kitchen floor 97 samples 13.5 EU/mg
Indoor environments—electrostatic dust collector sampling
Dutch households Passive air samples 10 samples 986 EU/m2 (268)
(Continued on next page)
3.2.6-12 ▪ AIR

TABLE 3 (Continued )
Mean
Setting Size fraction n Reference
concentration
Companion animal hospital Passive air samples 1 hospital 2280 EU/m2 (269)
14-day deployment 30 samples
Iowa homes Passive air samples (270)
Apartment buildings 14-day deployment 36 samples 205 EU/m2
Farm homes 7-day deployment 30 samples 934 EU/m2
14-day deployment 30 samples 2,010 EU/m2
28-day deployment 30 samples 2,770 EU/m2

been coupled with the LAL method to investigate the
impact of sample extraction procedures on endotoxin deter-
minations in the LAL (236). Analytical improvements
have been made in the analysis of 3-hydroxy fatty acids as sur-
rogates for the detection of lipid A. Alwis et al. (254) inves-
tigated the use of GC ion-trap MS for detection of
trimethylsilyl methyl ester derivatives of 3-hydroxy fatty
acids. This requires a multistep reaction and purification
methodology that is less amenable to high throughput than
the LAL kinetic chromogenic method. It has the advantage
that it can measure both free and extractable cell-bound
endotoxin (extraction in 2 M HCl in methanol). It has the
disadvantage that the equipment and labor costs are high
and mammalian body fluids also contain these 3-hydroxy
fatty acids. Immunoassays have also been developed but
have not been widely adopted. Although the biological activ-
ity rests with the lipid A portion of the molecule, the antibod-
ies produced against endotoxin recognize the polysaccharide
moiety (255). Thus, at this time the LAL assay remains the
most widely employed method.
Quality Assurance Measures for Endotoxin
Exposure Assessment
Adherence to rigorous quality assurance (QA) protocols is
essential for reliable endotoxin exposure assessment. Mini-
mum QA measures should include the use of sample blanks,
high and low endotoxin bench control samples, blind random
repeat samples, careful tracking of lots of Limulus lysate and
other reagents, a multipoint standard curve covering several
orders of magnitude, and consistent laboratory personnel
throughout the study. Inclusion of internal and external
QA audit is also advised. The multipoint standard curve
and the high and low bench control samples assayed at multi-
ple dilutions should be included on each microplate analyzed.
Rules for judging if assays are out of compliance should be
incorporated into large studies. One example is to use the first
20 assays of the bench control samples to establish mean and
standard deviations from which to judge subsequent assays
using control rules such as those proposed by Westgard
(www.westgard.com).
In support of a recent epidemiologic study, we performed
endotoxin measurements for 6,963 reservoir house dust sam-
ples. As a component of the quality assurance protocol, we
created high and low bench control samples by blending
and sieving 50 distinct house dust samples taken from the
family rooms, bedrooms, and kitchens of 41 homes to gener-
ate 1,100 cryovials each containing 50 mg of control house
dust. This approach was used to identify assays for reanalysis.
In this study, all field samples and 722 bench control samples
were run using a single lot of Limulus lysate. Data showed that
361 assays of the high bench control dust yielded one sample
above and none below the 3 SD control point. There were 12
samples above and none below the 2 SD control point. For
the low bench control dust, two exceeded the 3 SD point
range with none below and five were above the 2 SD control
point with none below. The geometric mean for the high
bench control samples were 83.2 EU/mg (n = 165) for phase
1 of the analysis and 81.2 EU/mg (n = 196) for phase 2 (geo-
metric GSD = 1.27 for both). Low bench controls were 16.2
EU/mg (geometric SD = 1.30, n = 165) in phase 1 and 16.1
EU/mg (geometric SD = 1.29, n = 196) in phase 2. This dem-
onstrates the high degree of precision attainable for house
dust analyses using the kinetic chromogenic LAL assay.
CONCLUSIONS
Airborne bacteria and endotoxin appear at elevated concen-
trations in a variety of occupational and nonoccupational
environments. Especially high levels are found in agriculture
and in industrial operations handling wet organic matter. Ele-
vated exposures in domestic environments are often associ-
ated with defective ventilation systems or water intrusion
and can also be problematic. Principal adverse health effects
include asthma exacerbation, organic dust toxic syndrome,
asthma-like syndrome, and hypersensitivity pneumonitis.
Endotoxin exerts its inflammatory potency through an innate
immune signaling pathway via TLR4. Exposure to endotoxin
at key life stages appears to reduce the induction of an allergic
phenotype. Susceptibilities differ between individuals
dependent on genetic differences in the recognition and
response pathways. Endotoxin can be measured with great
sensitivity using the kinetic chromogenic LAL assay. How-
ever, care must be taken to ensure that sample collection
and extraction methods are optimized and QA protocols are
strictly followed.
REFERENCES
1. Predicala BZ, Urban JE, Maghirang RG, Jerez SB, Good-
band RD. 2002. Assessment of bioaerosols in swine barns
by filtration and impaction. Curr Microbiol 44:136–140.
2. Donham KJ. 1991. Association of environmental air con-
taminants with disease and productivity in swine. Am J Vet
Res 52:1723–1730.
3. Duchaine C, Grimard Y, Cormier Y. 2000. Influence of
building maintenance, environmental factors, and seasons
on airborne contaminants of swine confinement buildings.
J Sci Occup Environ Health Saf 61:56–63.
4. Zucker BA, Muller W. 2000. Investigation on airborne
microorganisms in animals stables 3. Report: relationship
between inhalable endotoxin, inhalable dust and airborne
bacteria in a hen house. Berl Munch Tierarztl 113:
279–283.
5. Kiekhaefer MS, Donham KJ, Whitten P, Thorne PS.
1995. Cross seasonal studies of airborne microbial

populations and envirnoment in swine buildings: implica-
tions for worker and animal health. Ann Agric Environ Med
2:37–44.
6. Schenker MB, Christiani D, Cormier Y, Dimich-Ward
H, Doekes G, Doman J, Douwes J, Dowling K, Enarson
D, Green F, Heederik D, Husman K, Kennedy S, Kull-
man G, Lacasse Y, Lawson B, Malmberg P, May J,
McCurdy S, Merchant J, Myers J, Nieuwenhuijsen M,
Olenchock S, Saiki C, Schwartz D, Seiber J, Thorne P,
Wagner G, White N, Xu X, Chan-Yeung M. 1998. Respi-
ratory health hazards in agriculture. Am J Respir Crit Care
Med 158:S1–S76.
7. Chapin A, Rule A, Gibson K, Buckley T, Schwab K.
2005. Airborne multidrug-resistant bacteria isolated from
a concentrated swine feeding operation. Environ Health Per-
spect 113:137–142.
8. Gibbs SG, Green CF, Tarwater PM, Scarpino PV. 2004.
Airborne antibiotic resistant and nonresistant bacteria and
fungi recovered from two swine herd confined animal feed-
ing operations. J Occup Environ Hyg 1:699–706.
9. de Neeling AJ, van den Broek MJ, Spalburg EC, van
Santen-Verheuvel MG, Dam-Deisz WD, Boshuizen
HC, van de Giessen AW, van Duijkeren E, Huijsdens
XW. 2007. High prevalence of methicillin resistant Staph-
ylococcus aureus in pigs. Vet Microbiol 122:366–372.
10. Ling AL, Pace NR, Hernandez MT, LaPara TM. 2013.
Tetracycline resistance and class 1 integron genes associ-
ated with indoor and outdoor aerosols. Environ Sci Technol
47:4046–4052.
11. Alvarado CS, Gibbs SG, Gandara A, Flores C, Hurd
WW, Green CF. 2012. The potential for community expo-
sures to pathogens from an urban dairy. J Environ Health 74:
22–28.
12. Just NA, Letourneau V, Kirychuk SP, Singh B, Duchaine
C. 2012. Potentially pathogenic bacteria and antimicro-
bial resistance in bioaerosols from cage-housed and floor-
housed poultry operations. Ann Occup Hyg 56:440–449.
13. Masclaux FG, Sakwinska O, Charriere N, Semaani E,
Oppliger A. 2013. Concentration of airborne Staphylococ-
cus aureus (MRSA and MSSA), total bacteria, and endo-
toxins in pig farms. Ann Occup Hyg 57:550–557.
14. Nonnenmann MW, Donham KJ, Rautiainen RH, O’Sh-
aughnessy PT, Burmeister LF, Reynolds SJ. 2004. Vege-
table oil sprinkling as a dust reduction method in swine
confinement. J Agric Saf Health 10:7–15.
15. Siggers JL, Kirychuk SP, Lemay SP, Willson PJ. 2011.
Size distribution of particulate and associated endotoxin
and bacteria in traditional swine barn rooms and rooms
sprinkled with oil. J Agromed 16:271–279.
16. Senthilselvan A, Zhang Y, Dosman JA, Barber EM, Hol-
feld LE, Kirychuk SP, Cormier Y, Hurst TS, Rhodes CS.
1997. Positive human health effects of dust suppression
with canola oil in swine barns. Am J Respir Crit Care Med
156:410–417.
17. Mankell KO, Janni KA, Walker RD, Wilson ME, Petti-
grew JE, Jacobson LD, Wilcke WF. 1995. Dust suppression
in swine feed using soybean oil. J Anim Sci 73:981–985.
18. Pepys J, Jenkins PA, Festenstein GN, Gregory PH, Lacey
ME, Skinner FA. 1963. Farmer’s lung. Thermophilic acti-
nomycetes as a source of “Farmer’s lung hay” antigen. Lan-
cet 2:607–611.
19. Blais Lecours P, Veillette M, Marsolais D, Duchaine C.
2012. Characterization of bioaerosols from dairy barns:
reconstructing the puzzle of occupational respiratory dis-
eases by using molecular approaches. Appl Environ Microbiol
78:3242–3248.
20. Ranalli G, Grazia L, Roggeri A. 1999. The influence of
hay-packing techniques on the presence of saccharopoly-
spora rectivirgula. J Appl Microbiol 87:359–365.
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-13
21. Duchaine C, Meriaux A, Brochu G, Cormier Y. 1999.
Airborne microflora in Quebec dairy farms: lack of effect
of bacterial hay preservatives. Am Ind Hyg Assoc J 60:
89–95.
22. Erkinjuntti-Pekkanen R, Reiman M, Kokkarinen JI,
Tukiainen HO, Terho EO. 1999. IgG antibodies, chronic
bronchitis, and pulmonary function values in farmer’s lung
patients and matched controls. Allergy 54:1181–1187.
23. Roussel S, Reboux G, Dalphin JC, Bardonnet K, Millon
L, Piarroux R. 2004. Microbiological evolution of hay and
relapse in patients with farmer’s lung. Occup Environ Med
61:e3.
24. Eisenberg S, Nielen M, Hoeboer J, Bouman M, Heederik
D, Koets A. 2011. Mycobacterium avium subspecies paratu-
berculosis in bioaerosols after depopulation and cleaning of
two cattle barns. Vet Rec 168:587.
25. Eisenberg SW, Nielen M, Hoeboer J, Rutten V, Heederik
D, Koets AP. 2012. Environmental contamination with
Mycobacterium avium subspecies paratuberculosis within
and around a dairy barn under experimental conditions.
J Dairy Sci 95:6477–6482.
26. Brooks JP, Tanner BD, Gerba CP, Pepper IL. 2006. The
measurement of aerosolized endotoxin from land applica-
tion of Class B biosolids in Southeast Arizona. Can J Micro-
biol 52:150–156.
27. Brooks JP, Tanner BD, Josephson KL, Gerba CP, Pep-
per IL. 2004. Bioaerosols from the land application of
biosolids in the desert southwest USA. Water Sci Technol
50:7–12.
28. Dutkiewicz J, Olenchock SA, Sorenson WG, Gerencser
VF, May JJ, Pratt DS, Robinson VA. 1989. Levels of bac-
teria, fungi, and endotoxin in bulk and aerosolized corn
silage. Appl Environ Microbiol 55:1093–1099.
29. Schwartz DA, Thorne PS, Yagla SJ, Burmeister LF,
Olenchock SA, Watt JL, Quinn TJ. 1995. The role of
endotoxin in grain dust-induced lung disease. Am J Respir
Crit Care Med 152:603–608.
30. Moreno-Ancillol A, Dominguez-Noche C, Gil-Adrados
AC, Cosmes PM. 2004. Hypersensitivity pneumonitis
due to occupational inhalation of fungi-contaminated
corn dust. J Investig Allergol Clin Immunol 14:165–167.
31. Sanchez-Monedero MA, Stentiford EI, Urpilainen ST.
2005. Bioaerosol generation at large-scale green waste com-
posting plants. J Air Waste Manag Assoc 55:612–618.
32. Taha MP, Pollard SJ, Sarkar U, Longhurst P. 2005. Esti-
mating fugitive bioaerosol releases from static compost
windrows: feasibility of a portable wind tunnel approach.
Waste Manag 25:445–450.
33. Hryhorczuk D, Curtis L, Scheff P, Chung J, Rizzo M,
Lewis C, Keys N, Moomey M. 2001. Bioaerosol emissions
from a suburban yard waste composting facility. Ann Agric
Environ Med 8:177–185.
34. Cramp GJ, Harte D, Douglas NM, Graham F, Schousboe
M, Sykes K. 2010. An outbreak of Pontiac fever due to
Legionella longbeachae serogroup 2 found in potting mix in
a horticultural nursery in New Zealand. Epidemiol Infect
138:15–20.
35. Conza L, Pagani SC, Gaia V. 2013. Presence of Legionella
and free-living amoebae in composts and bioaerosols from
composting facilities. PLoS One 8:e68244.
36. Douwes J, Wouters I, Dubbeld H, van Zwieten L, Steer-
enberg P, Doekes G, Heederik D. 2000. Upper airway
inflammation assessed by nasal lavage in compost workers:
a relation with bio-aerosol exposure. Am J Ind Med 37:
459–468.
37. Van Tongeren M, Van Amelsvoort L, Heederik D. 1997.
Exposure to organic dusts, endotoxins, and microorganisms
in the municipal waste industry. Int J Occup Environ Health
3:30–36.
3.2.6-14 ▪ AIR
38. Sigsgaard T, Abel A, Donbaek L, Malmros P. 1994. Lung
function changes among recycling workers exposed to
organic dust. Am J Ind Med 25:69–72.
39. Sigsgaard T, Bach B, Malmros P. 1990. Respiratory
impairment among workers in a garbage-handling plant.
Am J Ind Med 17:92–93.
40. Thorn J, Beijer L, Rylander R. 1998. Airways inflamma-
tion and glucan exposure among household waste collec-
tors. Am J Ind Med 33:463–470.
41. Wouters IM, Hilhorst SK, Kleppe P, Doekes G, Douwes
J, Peretz C, Heederik D. 2002. Upper airway inflammation
and respiratory symptoms in domestic waste collectors.
Occup Environ Med 59:106–112.
42. Heldal KK, Eduard W. 2004. Associations between acute
symptoms and bioaerosol exposure during the collection of
household waste. Am J Ind Med 46:253–260.
43. Douwes J, Thorne P, Pearce N, Heederik D. 2003. Bio-
aerosol health effects and exposure assessment: progress
and prospects. Ann Occup Hyg 47:187–200.
44. Wouters IM, Douwes J, Doekes G, Thorne PS, Brunek-
reef B, Heederik DJ. 2000. Increased levels of markers of
microbial exposure in homes with indoor storage of organic
household waste. Appl Environ Microbiol 66:627–631.
45. Wouters IM, Spaan S, Douwes J, Doekes G, Heederik D.
2006. Overview of personal occupational exposure levels to
inhalable dust, endotoxin, beta(1 → 3)-glucan and fungal
extracellular polysaccharides in the waste management
chain. Ann Occup Hyg 50:39–53.
46. Lavoie J, Guertin S. 2001. Evaluation of health and safety
risks in municipal solid waste recycling plants. J Air Waste
Manag Assoc 51:352–360.
47. Bauer H, Fuerhacker M, Zibuschka F, Schmid H, Pux-
baum H. 2002. Bacteria and fungi in aerosols generated
by two different types of wastewater treatment plants. Water
Res 36:3965–3970.
48. Laitinen S, Kangas J, Kotimaa M, Liesivuori J, Martikai-
nen PJ, Nevalainen A, Sarantila R, Husman K. 1994.
Workers’ exposure to airborne bacteria and endotoxins at
industrial wastewater treatment plants. Am Ind Hyg Assoc
J 55:1055–1060.
49. Lee JA, Johnson JC, Reynolds SJ, Thorne PS, O’Shaugh-
nessy PT. 2006. Indoor and outdoor air quality assessment
of four wastewater treatment plants. J Occup Environ Hyg 3:
36–43.
50. Prazmo Z, Krysinska-Traczyk E, Skorska C, Sitkowska J,
Cholewa G, Dutkiewicz J. 2003. Exposure to bioaerosols
in a municipal sewage treatment plant. Ann Agric Environ
Med10:241–248.
51. Brandi G, Sisti M, Amagliani G. 2000. Evaluation of the
environmental impact of microbial aerosols generated by
wastewater treatment plants utilizing different aeration sys-
tems. J Appl Microbiol 88:845–852.
52. Korzeniewska E, Harnisz M. 2013. Extended-spectrum
beta-lactamase (ESBL)-positive Enterobacteriaceae in
municipal sewage and their emission to the environment.
J Environ Manag 128:904–911.
53. Korzeniewska E, Korzeniewska A, Harnisz M. 2013.
Antibiotic resistant Escherichia coli in hospital and munici-
pal sewage and their emission to the environment. Ecotox-
icol Environ Saf 91:96–102.
54. Fykse EM, Aarskaug T, Thrane I, Blatny JM. 2013.
Legionella and non-Legionella bacteria in a biological treat-
ment plant. Can J Microbiol 59:102–109.
55. O’Hara RE, Rubin R. 2005. Reducing bioaerosol disper-
sion from wastewater treatment and its land application: a
review and analysis. J Environ Health 68:24–29.
56. Paez-Rubio T, Viau E, Romero-Hernandez S, Peccia J.
2005. Source bioaerosol concentration and rRNA gene-
based identification of microorganisms aerosolized at a
flood irrigation wastewater reuse site. Appl Environ Microbiol
71:804–810.
57. Rylander R. 1987. The role of endotoxin for reactions after
exposure to cotton dust. Am J Ind Med 12:687–697.
58. Hollander A, Heederik D, Kauffman H. 1994. Acute res-
piratory effects in the potato processing industry due to a
bioaerosol exposure. Occup Environ Med 51:73–78.
59. Forster HW, Crook B, Platts BW, Lacey J, Topping MD.
1989. Investigation of organic aerosols generated during
sugar beet slicing. Am Ind Hyg Assoc J 50:44–50.
60. Smit LA, Wouters IM, Hobo MM, Eduard W, Doekes G,
Heederik D. 2006. Agricultural seed dust as a potential
cause of organic dust toxic syndrome. Occup Environ Med
63:59–67.
61. Thorne PS, Sprince NL. 2004. Metal working fluids,
p. 1043–1054. In: Rosenstock L, Cullen MR, Brodkin
CA, Redlich CA (eds), Textbook of Occupational and Envi-
ronmental Medicine, 2nd ed. Elsevier, Philadelphia, PA.
62. Mattsby-Baltzer I, Sandin M, Ahlstrom B, Allenmark S,
Edebo M, Falsen E, Pedersen K, Rodin N, Thompson
RA, Edebo L. 1989. Microbial growth and accumulation
in industrial metal-working fluids. Appl Environ Microbiol
55:2681–2689.
63. Thorne PS, DeKoster JA, Subramanian P. 1996. Environ-
mental assessment of aerosols, bioaerosols, and airborne
endotoxins in a machining plant. Am Ind Hyg Assoc J
57:1163–1167.
64. Gorny RL, Szponar B, Larsson L, Pehrson C, Prazmo Z,
Dutkiewicz J. 2004. Metalworking fluid bioaerosols at
selected workplaces in a steelworks. Am J Ind Med 46:
400–403.
65. Perkins SD, Angenent LT. 2010. Potential pathogenic
bacteria in metalworking fluids and aerosols from a machin-
ing facility. FEMS Microbiol Ecol 74:643–654.
66. Centers for Disease C, Prevention. 1996. Biopsy-
confirmed hypersensitivity pneumonitis in automobile
production workers exposed to metalworking fluids—
Michigan, 1994–1995. Morb Mortal Wkly Rep 45:
606–610.
67. Sessa R, Di PM, Schiavoni G, Santino I, Altieri A, Pinelli
S, Del PM. 2002. Microbiological indoor air quality in
healthy buildings. New Microbiol 25:51–56.
68. Thorne PS, Adamcakova-Dodd A, Kelly KM, O’Neill M
E, Duchaine C. 2006. Metalworking fluid with mycobacte-
ria and endotoxin induces hypersensitivity pneumonitis in
mice. Am J Respir Crit Care Med 173:759–768.
69. Gordon T, Nadziejko C, Galdanes K, Lewis D, Donnelly
K. 2006. Mycobacterium immunogenum causes hypersensi-
tivity pneumonitis-like pathology in mice. Inhal Toxicol
18:449–456.
70. Demers PA, Teschke K, Kennedy SM. 1997. What to do
about softwood? A review of respiratory effects and recom-
mendations regarding exposure limits. Am J Ind Med 31:
385–398.
71. Duchaine C, Meriaux A, Thorne PS, Cormier Y. 2000.
Assessment of particulates and bioaerosols in eastern
Canadian sawmills. J Sci Occup Environ Health Saf 61:
727–732.
72. Eduard W, Sandven P, Levy F. 1992. Relationships
between exposure to spores from Rhizopus microsporus and
Paecilomyces variotii and serum IgG antibodies in wood
trimmers. Int Arch Allergy Immunol 97:274–282.
73. Eduard W, Sandven P, Levy F. 1993. Serum IgG antibod-
ies to mold spores in two Norwegian sawmill populations:
relationship to respiratory and other work-related symp-
toms. Am J Ind Med 24:207–222.
74. Oppliger A, Rusca S, Charriere N, Vu Duc T, Droz PO.
2005. Assessment of bioaerosols and inhalable dust expo-
sure in Swiss sawmills. Ann Occup Hyg 49:385–391.

75. Duchaine C, Thorne PS, Meriaux A, Grimard Y, Whit-
ten P, Cormier Y. 2001. Comparison of endotoxin expo-
sure assessment by bioaerosol impinger and filter-sampling
methods. Appl Environ Microbiol 67:2775–2780.
76. Dutkiewicz J, Krysinska-Traczyk E, Prazmo Z, Skorska
C, Sitkowska J. 2001. Exposure to airborne microorgan-
isms in Polish sawmills. Ann Agric Environ Med 8:71–80.
77. Prazmo Z, Dutkiewicz J, Skorska C, Sitkowska J, Chol-
ewa G. 2003. Exposure to airborne Gram-negative bacteria,
dust and endotoxin in paper factories. Ann Agric Environ
Med 10:93–100.
78. Rylander R, Thorn J, Attefors R. 1999. Airways inflam-
mation among workers in a paper industry. Eur Respir J
13:1151–1157.
79. Liang R, Tian J, She R, Meng H, Xiao P, Chang L. 2013.
Airborne microbial composition in a high-throughput
poultry slaughtering facility. J Food Prot 76:413–419.
80. Gregova G, Kmetova M, Kmet V, Venglovsky J, Feher A.
2012. Antibiotic resistance of Escherichia coli isolated from
a poultry slaughterhouse. Ann Agric Environ Med 19:
75–77.
81. Schmidt JW, Arthur TM, Bosilevac JM, Kalchayanand
N, Wheeler TL. 2012. Detection of Escherichia coli
O157:H7 and Salmonella enterica in air and droplets at three
U.S. commercial beef processing plants. J Food Prot 75:
2213–2218.
82. Martin E, Fallschissel K, Kampfer P, Jackel U. 2010.
Detection of Jeotgalicoccus spp. in poultry house air. Syst
Appl Microbiol 33:188–192.
83. Rahkio TM, Korkeala HJ. 1997. Airborne bacteria and
carcass contamination in slaughterhouses. J Food Prot 60:
38–42.
84. Jericho KW, Kozub GC, Gannon VP, Taylor CM. 2000.
Microbiological testing of raw, boxed beef in the context of
hazard analysis critical control point at a high-line-speed
abattoir. J Food Prot 63:1681–1686.
85. Fries R, Graw C. 1999. Water and air in two poultry proc-
essing plants’ chilling facilities—a bacteriological survey.
Br Poult Sci 40:52–58.
86. Hagmar L, Schutz A, Sjoholm A. 1990. Over-shift
decrease in lung function in poultry slaughterhouse work-
ers. Am J Ind Med 17:77–78.
87. Davis M, Morishita TY. 2005. Relative ammonia concen-
trations, dust concentrations, and presence of Salmoneua
species and Escherichia coli inside and outside commercial
layer facilities. Avian Dis 49:30–35.
88. Venter P, Lues JF, Theron H. 2004. Quantification of bio-
aerosols in automated chicken egg production plants. Poult
Sci 83:1226–1231.
89. Iowa State University and The University of Iowa
Study Group. 2002. Iowa Concentrated Animal Feed-
ing Operations Air Quality Study. http://www.public-
health.uiowa.edu/ehsrc/CAFOstudy/CAFO_final2-14.pdf
[accessed: 16 July 2015].
90. Nguyen TM, Ilef D, Jarraud S, Rouil L, Campese C, Che
D, Haeghebaert S, Ganiayre F, Marcel F, Etienne J,
Desenclos JC. 2006. A community-wide outbreak of
legionnaires disease linked to industrial cooling towers—
how far can contaminated aerosols spread? J Infect Dis
193:102–111.
91. Jo WK, Seo YJ. 2005. Indoor and outdoor bioaerosol levels
at recreation facilities, elementary schools, and homes.
Chemosphere 61:1570–1579.
92. Hospodsky D, Qian J, Nazaroff WW, Yamamoto N,
Bibby K, Rismani-Yazdi H, Peccia J. 2012. Human occu-
pancy as a source of indoor airborne bacteria. PLoS One 7:
e34867.
93. Fox A, Harley W, Feigley C, Salzberg D, Toole C, Sebas-
tian A, Larsson L. 2005. Large particles are responsible for
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-15
elevated bacterial marker levels in school air upon occupa-
tion. J Environ Monitor 7:450–456.
94. Bartlett KH, Kennedy SM, Brauer M, van Netten C, Dill
B. 2004. Evaluation and determinants of airborne bacterial
concentrations in school classrooms. J Occup Environ Hyg
1:639–647.
95. Toivola M, Alm S, Reponen T, Kolari S, Nevalainen A.
2002. Personal exposures and microenvironmental con-
centrations of particles and bioaerosols. J Environ Monit
4:166–174.
96. Thorne PS. 1993. Sump additives as a source of bioaerosols
in a school building. Vet Hum Toxicol 35:141–143.
97. Reynolds SJ, Black DW, Borin SS, Breuer G, Burmeister
LF, Fuortes LJ, Smith TF, Stein MA, Subramanian P,
Thorne PS, Whitten P. 2001. Indoor environmental qual-
ity in six commercial office buildings in the midwest United
States. Appl Occup Environ Hyg 16:1065–1077.
98. Golofit-Szymczak M, Gorny RL. 2010. Bacterial and fun-
gal aerosols in air-conditioned office buildings in Warsaw,
Poland—the winter season. Int J Occup Saf Ergon 16:
465–476.
99. Clark A. 1974. Bacterial colonization of dental units and
the nasal flora of dental personnel. Proc R Soc Med 67:
1269–1270.
100. Ghiabi N. 1998. Air contamination during use of air abra-
sion instrumentation. J Clin Pediatr Dent 23:37–43.
101. Kedjarune U, Kukiattrakoon B, Yapong B, Chowanadisai
S, Leggat P. 2000. Bacterial aerosols in the dental clinic:
effect of time, position and type of treatment. Int Dent J
50:103–107.
102. Petti S, Tarsitani G. 1998. Intra-individual variations of
salivary microbial levels in young adults. Eur J Oral Sci
106:616–622.
103. Barbeau J, Gauthier C, Payment P. 1998. Biofilms, infec-
tious agents, and dental unit waterlines: a review. Can
J Microbiol 44:1019–1028.
104. Bennett AM, Fulford MR, Walker JT, Bradshaw DJ,
Martin MV, Marsh PD. 2000. Microbial aerosols in gen-
eral dental practice. Br Dent J 189:664–667.
105. Barbeau J, Buhler T. 2001. Biofilms augment the number
of free-living amoebae in dental unit waterlines. Res Micro-
biol 152:753–760.
106. Barbeau J, Nadeau C. 1997. Dental unit waterline
microbiology: a cautionary tale. J Can Dent Assoc 63:
775–779.
107. Noce L, Di Giovanni D, Putnins EE. 2000. An evaluation
of sampling and laboratory procedures for determination of
heterotrophic plate counts in dental unit waterlines. J Can
Dent Assoc 66:262.
108. Dutil S, Tessier S, Veillette M, Laflamme C, Meriaux A,
Leduc A, Barbeau J, Duchaine C. 2006. Detection of
Legionella spp. by fluorescent in situ hybridization in dental
unit waterlines. J Appl Microbiol 100:955–963.
109. Pagniano RP, Scheid RC, Rosen S, Beck FM. 1985. Air-
borne microorganisms collected in a preclinical dental lab-
oratory. J Dent Educ 49:653–655.
110. Reinthaler FF, Mascher F, Stunzner D. 1988. Serological
examinations for antibodies against Legionella species in
dental personnel. J Dent Res 67:942–943.
111. Dekoster JA, Thorne PS. 1995. Bioaerosol concentrations
in noncompliant, compliant, and intervention homes in
the Midwest. Am Ind Hyg Assoc J 56:573–580.
112. Hoppe KA, Metwali N, Perry SS, Hart T, Kostle PA,
Thorne PS. 2012. Assessment of airborne exposures and
health in flooded homes undergoing renovation. Indoor
Air 22:446–456.
113. Lee T, Grinshpun SA, Martuzevicius D, Adhikari A,
Crawford CM, Luo J, Reponen T. 2006. Relationship
between indoor and outdoor bio-aerosols collected with a
3.2.6-16 ▪ AIR
button inhalable aerosol sampler in urban homes. Indoor
Air 16:37–47.
114. Rintala H, Hyvarinen A, Paulin L, Nevalainen A. 2004.
Detection of streptomycetes in house dust—comparison of
culture and PCR methods. Indoor Air 14:112–119.
115. Kaan JA, Simoons-Smit AM, MacLaren DM. 1985.
Another source of aerosol causing nosocomial Legion-
naires’ disease. J Infect 11:145–148.
116. Zuravleff JJ, Yu VL, Shonnard JW, Rihs JD, Best M.
1983. Legionella pneumophila contamination of a hospital
humidifier. Demonstration of aerosol transmission and sub-
sequent subclinical infection in exposed guinea pigs. Am
Rev Respir Dis 128:657–661.
117. Rhame FS, Streifel A, McComb C, Boyle M. 1986. Bub-
bling humidifiers produce microaerosols which can carry
bacteria. Infect Control 7:403–407.
118. Allers T, Mevarech M. 2005. Archaeal genetics—the
third way. Nat Rev Genet 6:58–73.
119. Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P.
2008. Mesophilic Crenarchaeota: proposal for a third arch-
aeal phylum, the Thaumarchaeota. Nat Rev Microbiol 6:
245–252.
120. Schleper C, Jurgens G, Jonuscheit M. 2005. Genomic
studies of uncultivated archaea. Nat Rev Microbiol 3:
479–488.
121. Forterre P, Brochier C, Philippe H. 2002. Evolution of the
archaea. Theor Popul Biol 61:409–422.
122. Albers SV, Meyer BH. 2011. The archaeal cell envelope.
Nat Rev Microbiol 9:414–426.
123. Cavicchioli R. 2011. Archaea—timeline of the third
domain. Nat Rev Microbiol 9:51–61.
124. Gill EE, Brinkman FS. 2011. The proportional lack of
archaeal pathogens: do viruses/phages hold the key? Bioes-
says 33:248–254.
125. Kates M. 1992. Archaebacterial lipids: structure, biosyn-
thesis and function. Biochem Soc Symp 58:51–72.
126. Angel R, Claus P, Conrad R. 2012. Methanogenic archaea
are globally ubiquitous in aerated soils and become active
under wet anoxic conditions. ISME J 6:847–862.
127. Bates ST, Berg-Lyons D, Caporaso JG, Walters WA,
Knight R, Fierer N. 2011. Examining the global distribu-
tion of dominant archaeal populations in soil. ISME J 5:
908–917.
128. DeLong EF. 1992. Archaea in coastal marine environ-
ments. Proc Natl Acad Sci USA 89:5685–5689.
129. Saengkerdsub S, Ricke SC. 2013. Ecology and character-
istics of methanogenic archaea in animals and humans. Crit
Rev Microbiol 40:97–116.
130. Just N, Blais Lecours P, Marcoux-Voiselle M, Kirychuk
S, Veillette M, Singh B, Duchaine C. 2013. Archaeal
characterization of bioaerosols from cage-housed and floor-
housed poultry operations. Can J Microbiol 59:46–50.
131. Nehme B, Gilbert Y, Letourneau V, Forster RJ, Veillette
M, Villemur R, Duchaine C. 2009. Culture-independent
characterization of archaeal biodiversity in swine confine-
ment building bioaerosols. Appl Environ Microbiol 75:
5445–5450.
132. Dridi B, Raoult D, Drancourt M. 2011. Archaea as emerg-
ing organisms in complex human microbiomes. Anaerobe
17:56–63.
133. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlef-
sen L, Sargent M, Gill SR, Nelson KE, Relman DA. 2005.
Diversity of the human intestinal microbial flora. Science
308:1635–1638.
134. Kirychuk SP, Dosman JA, Reynolds SJ, Willson P, Sen-
thilselvan A, Feddes JJ, Classen HL, Guenter W. 2006.
Total dust and endotoxin in poultry operations: compari-
son between cage and floor housing and respiratory effects
in workers. J Occup Environ Med 48:741–748.
135. Just N, Kirychuk S, Gilbert Y, Letourneau V, Veillette
M, Singh B, Duchaine C. 2011. Bacterial diversity charac-
terization of bioaerosols from cage-housed and floor-housed
poultry operations. Environ Res 111:492–498.
136. Blais Lecours P, Duchaine C, Taillefer M, Tremblay C,
Veillette M, Cormier Y, Marsolais D. 2011. Immunogenic
properties of archaeal species found in bioaerosols. PLoS
One 6:e23326.
137. Blanchet MR, Israel-Assayag E, Cormier Y. 2004. Inhib-
itory effect of nicotine on experimental hypersensitivity
pneumonitis in vivo and in vitro. Am J Respir Crit Care
Med 169:903–909.
138. Gudmundsson G, Hunninghake GW. 1997. Interferon-
gamma is necessary for the expression of hypersensitivity
pneumonitis. J Clin Invest 99:2386–2390.
139. Eduard W, Heederik D, Duchaine C, Green BJ. 2012.
Bioaerosol exposure assessment in the workplace: the
past, present and recent advances. J Environ Monit 14:
334–339.
140. Lange JL, Thorne PS, Lynch N. 1997. Application of flow
cytometry and fluorescent in situ hybridization for assess-
ment of exposures to airborne bacteria. Appl Environ Microb
63:1557–1563.
141. Letourneau V, Meriaux A, Goyer N, Chakir J, Cormier
Y, Duchaine C. 2010. Biological activities of respirable
dust from Eastern Canadian peat moss factories. Toxicol
In Vitro 24:1273–1278.
142. Nehme B, Letourneau V, Forster RJ, Veillette M, Duch-
aine C. 2008. Culture-independent approach of the bacte-
rial bioaerosol diversity in the standard swine confinement
buildings, and assessment of the seasonal effect. Environ
Microbiol 10:665–675.
143. Oppliger A, Charriere N, Droz PO, Rinsoz T. 2008.
Exposure to bioaerosols in poultry houses at different stages
of fattening; use of real-time PCR for airborne bacterial
quantification. Ann Occup Hyg 52:405–412.
144. Fallschissel K, Kampfer P, Jackel U. 2009. Direct detec-
tion of salmonella cells in the air of livestock stables by real-
time PCR. Ann Occup Hyg 53:859–868.
145. Schafer J, Kampfer P, Jackel U. 2011. Detection of Saccha-
ropolyspora rectivirgula by quantitative real-time PCR. Ann
Occup Hyg 55:612–619.
146. Letourneau V, Nehme B, Meriaux A, Masse D, Cormier
Y, Duchaine C. 2010. Human pathogens and tetracycline-
resistant bacteria in bioaerosols of swine confinement
buildings and in nasal flora of hog producers. Int J Hyg Envi-
ron Health 213:444–449.
147. Nonnenmann MW, Bextine B, Dowd SE, Gilmore K,
Levin JL. 2010. Culture-independent characterization of
bacteria and fungi in a poultry bioaerosol using pyrose-
quencing: a new approach. J Occup Environ Hyg 7:
693–699.
148. Alexander C, Rietschel ET. 2001. Bacterial lipopolysac-
charides and innate immunity. J Endotoxin Res 7:
167–202.
149. Beutler B, Rietschel ET. 2003. Innate immune sensing
and its roots: the story of endotoxin. Nat Rev Immunol 3:
169–176.
150. Weiss J. 2003. Bactericidal/permeability-increasing pro-
tein (BPI) and lipopolysaccharide-binding protein (LBP):
structure, function and regulation in host defence against
Gram-negative bacteria. Biochem Soc Trans 31:785–790.
151. Carty CL, Gehring U, Cyrys J, Bischof W, Heinrich J.
2003. Seasonal variability of endotoxin in ambient fine
particulate matter. J Environ Monit 5:953–958.
152. Heinrich J, Pitz M, Bischof W, Krug N, Borm PJA. 2003.
Endotoxin in fine (PM2.5) and coarse (PM2.5–10) particle
mass of ambient aerosols. A temporo-spatial analysis. Atmos
Environ 37:3659–3667.

153. Mueller-Anneling L, Avol E, Peters JM, Thorne PS.
2004. Ambient endotoxin concentrations in PM10 from
Southern California. Environ Health Perspect 112:
583–588.
154. Thorne PS, Ansley AC, Perry SS. 2009. Concentrations
of bioaerosols, odors, and hydrogen sulfide inside and
downwind from two types of swine livestock operations.
J Occup Environ Hyg 6:211–220.
155. Bolte G, Bischof W, Borte M, Lehmann I, Wichmann
HE, Heinrich J, LISA Study Group. 2003. Early endo-
toxin exposure and atopy development in infants: results
of a birth cohort study. Clin Exp Allergy 33:770–776.
156. Braun-Fahrlander C, Riedler J, Herz U, Eder W, Waser
M, Grize L, Maisch S, Carr D, Gerlach F, Bufe A, Lau-
ener RP, Schierl R, Renz H, Nowak D, von Mutius E,
Allergy and Endotoxin Study Team. 2002. Environmental
exposure to endotoxin and its relation to asthma in school-
age children. N Engl J Med 347:869–877.
157. Michel O, Kips J, Duchateau J, Vertongen F, Robert L,
Collet H, Pauwels R, Sergysels R. 1996. Severity of
asthma is related to endotoxin in house dust. Am J Respir
Crit Care Med 154:1641–1646.
158. Park JH, Spiegelman DL, Burge HA, Gold DR, Chew
GL, Milton DK. 2000. Longitudinal study of dust and air-
borne endotoxin in the home. Environ Health Perspect 108:
1023–1028.
159. Thorne PS, Kulhankova K, Yin M, Cohn R, Arbes SJ,
Jr., Zeldin DC. 2005. Endotoxin exposure is a risk factor
for asthma: the national survey of endotoxin in United
States housing. Am J Respir Crit Care Med 172:
1371–1377.
160. Duquenne P, Marchand G, Duchaine C. 2013. Measure-
ment of endotoxins in bioaerosols at workplace: a critical
review of literature and a standardization issue. Ann Occup
Hyg 57:137–172.
161. Singh J, Schwartz DA. 2005. Endotoxin and the lung:
Insight into the host-environment interaction. J Allergy
Clin Immunol 115:330–333.
162. Douwes J, Thorne PS, Pearce N, Heederik D. 2003.
Biological agents—recognition, p. 219–292. In: Perkins
JL (ed), Modern Industrial Hygiene, vol. 2. ACGIH, Cincin-
nati, OH.
163. Cavagna G, Foa V, Vigliani EC. 1969. Effects in man and
rabbits of inhalation of cotton dust or extracts and purified
endotoxins. Br J Ind Med 26:314–321.
164. Pernis B, Vigliani EC, Cavagna G. 1960. The role of endo-
toxins in the pathogenesis of diseases caused by the inhala-
tion of vegetable dusts. Med Lav 51:780–794.
165. Castellan RM, Olenchock SA, Hankinson JL, Millner
PD, Cocke JB, Bragg CK, Perkins HH, Jr., Jacobs RR.
1984. Acute bronchoconstriction induced by cotton dust:
dose-related responses to endotoxin and other dust factors.
Ann Intern Med 101:157–163.
166. Clapp WD, Becker S, Quay J, Watt JL, Thorne PS, Frees
KL, Zhang X, Koren HS, Lux CR, Schwartz DA. 1994.
Grain dust-induced airflow obstruction and inflammation
of the lower respiratory tract. Am J Respir Crit Care Med
150:611–617.
167. Deetz DC, Jagielo PJ, Quinn TJ, Thorne PS, Bleuer SA,
Schwartz DA. 1997. The kinetics of grain dust-induced
inflammation of the lower respiratory tract. Am J Respir
Crit Care Med 155:254–259.
168. Jagielo PJ, Thorne PS, Watt JL, Frees KL, Quinn TJ,
Schwartz DA. 1996. Grain dust and endotoxin inhalation
challenges produce similar inflammatory responses in nor-
mal subjects. Chest 110:263–270.
169. Kline JN, Cowden JD, Hunninghake GW, Schutte BC,
Watt JL, Wohlford-Lenane CL, Powers LS, Jones MP,
Schwartz DA. 1999. Variable airway responsiveness to
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-17
inhaled lipopolysaccharide. Am J Respir Crit Care Med
160:297–303.
170. Michel O, Ginanni R, Le Bon B, Content J, Duchateau J,
Sergysels R. 1992. Inflammatory response to acute inhala-
tion of endotoxin in asthmatic patients. Am Rev Respir Dis
146:352–357.
171. Michel O, Nagy AM, Schroeven M, Duchateau J, Neve J,
Fondu P, Sergysels R. 1997. Dose-response relationship to
inhaled endotoxin in normal subjects. Am J Respir Crit Care
Med 156:1157–1164.
172. Jagielo PJ, Thorne PS, Kern JA, Quinn TJ, Schwartz
DA. 1996. Role of endotoxin in grain dust-induced lung
inflammation in mice. Am J Physiol 270:L1052–1059.
173. Schwartz DA, Thorne PS, Jagielo PJ, White GE, Bleuer
SA, Frees KL. 1994. Endotoxin responsiveness and grain
dust-induced inflammation in the lower respiratory tract.
Am J Physiol 267:L609–617.
174. Thorne PS, McCray PB, Howe TS, O’Neill MA. 1999.
Early-onset inflammatory responses in vivo to adenoviral
vectors in the presence or absence of lipopolysaccharide-
induced inflammation. Am J Respir Cell and Molecular
Biol 20:1155–1164.
175. Perzanowski MS, Miller RL, Thorne PS, Barr RG, Div-
jan A, Sheares BJ, Garfinkel RS, Perera FP, Goldstein IF,
Chew GL. 2006. Endotoxin in inner-city homes: associa-
tions with wheeze and eczema in early childhood.
J Allergy Clin Immunol 117:1082–1089.
176. Thorne PS, Cohn RD, Mav D, Arbes SJ, Zeldin DC.
2009. Predictors of endotoxin levels in U.S. housing. Envi-
ron Health Perspect 117:763–771.
177. Celedon JC, Milton DK, Ramsey CD, Litonjua AA,
Ryan L, Platts-Mills TA, Gold DR. 2007. Exposure to
dust mite allergen and endotoxin in early life and asthma
and atopy in childhood. J Allergy Clin Immunol 120:144–
149.
178. Blanc PD, Eisner MD, Katz PP, Yen IH, Archea C, Ear-
nest G, Janson S, Masharani UB, Quinlan PJ, Hammond
SK, Thorne PS, Balmes JR, Trupin L, Yelin EH. 2005.
Impact of the home indoor environment on adult asthma
and rhinitis. J Occup Environ Med 47:362–372.
179. Post DM, Zhang D, Eastvold JS, Teghanemt A, Gibson
BW, Weiss JP. 2005. Biochemical and functional charac-
terization of membrane blebs purified from Neisseria menin-
gitidis serogroup B. J Biol Chem 280:38383–38394.
180. Hadina S, Weiss JP, McCray PB, Jr., Kulhankova K,
Thorne PS. 2008. MD-2-dependent pulmonary immune
responses to inhaled lipooligosaccharides: effect of acyla-
tion state. Am J Respir Cell Mol Biol 38:647–654.
181. Gioannini TL, Teghanemt A, Zhang D, Coussens NP,
Dockstader W, Ramaswamy S, Weiss JP. 2004. Isolation
of an endotoxin-MD-2 complex that produces Toll-like
receptor 4-dependent cell activation at picomolar concen-
trations. Proc Natl Acad Sci USA 101:4186–4191.
182. Gioannini TL, Weiss JP. 2007. Regulation of interactions
of Gram-negative bacterial endotoxins with mammalian
cells. Immunol Res 39:249–260.
183. Abreu MT, Vora P, Faure E, Thomas LS, Arnold ET,
Arditi M. 2001. Decreased expression of Toll-like
receptor-4 and MD-2 correlates with intestinal epithelial
cell protection against dysregulated proinflammatory gene
expression in response to bacterial lipopolysaccharide.
J Immunol 167:1609–1616.
184. Otte JM, Cario E, Podolsky DK. 2004. Mechanisms of
cross hyporesponsiveness to Toll-like receptor bacterial
ligands in intestinal epithelial cells. Gastroenterology 126:
1054–1070.
185. Schwartz DA, Donham KJ, Olenchock SA, Popendorf
WJ, Van Fossen DS, Burmeister LF, Merchant JA.
1995. Determinants of longitudinal changes in spirometric
3.2.6-18 ▪ AIR
function among swine confinement operators and farmers.
Am J Respir Crit Care Med 151:47–53.
186. George CL, White ML, O’Neill ME, Thorne PS,
Schwartz DA, Snyder JM. 2003. Altered surfactant pro-
tein A gene expression and protein metabolism associated
with repeat exposure to inhaled endotoxin. Am J Physiol
Lung Cell Mol Physiol 285:L1337–L1344.
187. Lorenz E, Jones M, Wohlford-Lenane C, Meyer N, Frees
KL, Arbour NC, Schwartz DA. 2001. Genes other than
TLR4 are involved in the response to inhaled LPS.
Am J Physiol Lung Cell Mol Physiol 281:L1106–1114.
188. Thorne PS, Yeske CP, Karol MH. 1987. Monitoring
guinea pig core temperature by telemetry during inhalation
exposures. Fundam Appl Toxicol 9:398–408.
189. Brass DM, Savov JD, Whitehead GS, Maxwell AB,
Schwartz DA. 2004. LPS binding protein is important in
the airway response to inhaled endotoxin. J Allergy Clin
Immunol 114:586–592.
190. Jeyaseelan S, Chu HW, Young SK, Freeman MW,
Worthen GS. 2005. Distinct roles of pattern recognition
receptors CD14 and Toll-like receptor 4 in acute lung
injury. Infect Immun 73:1754–1763.
191. Arbour NC, Lorenz E, Schutte BC, Zabner J, Kline JN,
Jones M, Frees K, Watt JL, Schwartz DA. 2000. TLR4
mutations are associated with endotoxin hyporesponsive-
ness in humans. Nat Genet 25:187–191.
192. Smit LA, Heederik D, Doekes G, Koppelman GH, Bot-
tema RW, Postma DS, Wouters IM. 2011. Endotoxin
exposure, CD14 and wheeze among farmers: a gene—envi-
ronment interaction. Occup Environ Med 68:826–831.
193. Kleeberger SR, Peden D. 2005. Gene-environment inter-
actions in asthma and other respiratory diseases. Annu Rev
Med 56:383–400.
194. Gereda JE, Leung DY, Liu AH. 2000. Levels of environ-
mental endotoxin and prevalence of atopic disease. J Am
Med Assoc 284:1652–1653.
195. Gereda JE, Leung DY, Thatayatikom A, Streib JE, Price
MR, Klinnert MD, Liu AH. 2000. Relation between
house-dust endotoxin exposure, type 1 T-cell develop-
ment, and allergen sensitisation in infants at high risk of
asthma. Lancet 355:1680–1683.
196. Liu AH. 2002. Endotoxin exposure in allergy and asthma:
reconciling a paradox. J Allergy Clin Immunol 109:
379–392.
197. Michel O, Ginanni R, Duchateau J, Vertongen F, Le Bon
B, Sergysels R. 1991. Domestic endotoxin exposure and
clinical severity of asthma. Clin Exp Allergy 21:441–448.
198. Tulic MK, Wale JL, Holt PG, Sly PD. 2000. Modification
of the inflammatory response to allergen challenge after
exposure to bacterial lipopolysaccharide. Am J Respir Cell
Mol Biol 22:604–612.
199. Wan GH, Li CS, Lin RH. 2000. Airborne endotoxin
exposure and the development of airway antigen-specific
allergic responses. Clin Exp Allergy 30:426–432.
200. Zock JP, Hollander A, Heederik D, Douwes J. 1998.
Acute lung function changes and low endotoxin exposures
in the potato processing industry. Am J Ind Med 33:
384–391.
201. Kennedy SM, Christiani DC, Eisen EA, Wegman DH,
Greaves IA, Olenchock SA, Ye TT, Lu PL. 1987. Cotton
dust and endotoxin exposure-response relationships in cot-
ton textile workers. Am Rev Respir Dis 135:194–200.
202. Lai PS, Christiani DC. 2013. Long-term respiratory health
effects in textile workers. Curr Opin Pulm Med 19:
152–157.
203. Wang XR, Zhang HX, Sun BX, Dai HL, Hang JQ, Eisen
EA, Wegman DH, Olenchock SA, Christiani DC. 2005.
A 20-year follow-up study on chronic respiratory effects of
exposure to cotton dust. Eur Respir J 26:881–886.
204. Radon K, Danuser B, Iversen M, Jorres R, Monso E,
Opravil U, Weber C, Donham KJ, Nowak D. 2001. Res-
piratory symptoms in European animal farmers. Eur Respir J
17:747–754.
205. Donham KJ, Merchant JA, Lassise D, Popendorf WJ,
Burmeister LF. 1990. Preventing respiratory disease in
swine confinement workers: intervention through applied
epidemiology, education, and consultation. Am J Ind Med
18:241–261.
206. Basinas I, Schlunssen V, Heederik D, Sigsgaard T, Smit
LA, Samadi S, Omland O, Hjort C, Madsen AM, Skov S,
Wouters IM. 2012. Sensitisation to common allergens and
respiratory symptoms in endotoxin exposed workers: a
pooled analysis. Occup Environ Med 69:99–106.
207. Fink JN, Ortega HG, Reynolds HY, Cormier YF, Fan LL,
Franks TJ, Kreiss K, Kunkel S, Lynch D, Quirce S, Rose
C, Schleimer RP, Schuyler MR, Selman M, Trout D,
Yoshizawa Y. 2005. Needs and opportunities for research
in hypersensitivity pneumonitis. Am J Respir Crit Care
Med 171:792–798.
208. Peterson RD, Wicklunds PE, Good RA. 1964. Endotoxin
activity of a house dust extract. J Allergy 35:134–142.
209. Ernst P, Cormier Y. 2000. Relative scarcity of asthma and
atopy among rural adolescents raised on a farm. Am J Respir
Crit Care Med 161:1563–1566.
210. Klintberg B, Berglund N, Lilja G, Wickman M, van
Hage-Hamsten M. 2001. Fewer allergic respiratory disor-
ders among farmers’ children in a closed birth cohort
from Sweden. Eur Respir J 17:1151–1157.
211. Downs SH, Marks GB, Mitakakis TZ, Leuppi JD, Car
NG, Peat JK. 2001. Having lived on a farm and protection
against allergic diseases in Australia. Clin Exp Allergy 31:
570–575.
212. Leynaert B, Neukirch C, Jarvis D, Chinn S, Burney P,
Neukirch F, European Community Respiratory Health
Survey. 2001. Does living on a farm during childhood pro-
tect against asthma, allergic rhinitis, and atopy in adult-
hood? Am J Respir Crit Care Med 164:1829–1834.
213. Wickens K, Lane JM, Fitzharris P, Siebers R, Riley G,
Douwes J, Smith T, Crane J. 2002. Farm residence and
exposures and the risk of allergic diseases in New Zealand
children. Allergy 57:1171–1179.
214. Chrischilles E, Ahrens R, Kuehl A, Kelly K, Thorne P,
Burmeister L, Merchant J. 2004. Asthma prevalence
and morbidity among rural Iowa schoolchildren. J Allergy
Clin Immunol 113:66–71.
215. Merchant JA, Naleway AL, Svendsen ER, Kelly KM,
Burmeister LF, Stromquist AM, Taylor CD, Thorne
PS, Reynolds SJ, Sanderson WT, Chrischilles EA.
2005. Asthma and farm exposures in a cohort of rural
Iowa children. Environ Health Perspect 113:350–356.
216. Litonjua AA, Milton DK, Celedon JC, Ryan L, Weiss ST,
Gold DR. 2002. A longitudinal analysis of wheezing in
young children: the independent effects of early life expo-
sure to house dust endotoxin, allergens, and pets.
J Allergy Clin Immunol 110:736–742.
217. Horick N, Weller E, Milton DK, Gold DR, Li R, Spiegel-
man D. 2006. Home endotoxin exposure and wheeze in
infants: correction for bias due to exposure measurement
error. Environ Health Perspect 114:135–140.
218. Rennie DC, Lawson JA, Kirychuk SP, Paterson C, Will-
son PJ, Senthilselvan A, Cockcroft DW. 2008. Assess-
ment of endotoxin levels in the home and current
asthma and wheeze in school-age children. Indoor Air 18:
447–453.
219. Zhao Z, Sebastian A, Larsson L, Wang Z, Zhang Z, Nor-
back D. 2008. Asthmatic symptoms among pupils in rela-
tion to microbial dust exposure in schools in Taiyuan,
China. Ped Allergy Immunol 19:455–465.

220. Smit LA, Heederik D, Doekes G, Lammers JW, Wouters
IM. 2010. Occupational endotoxin exposure reduces the
risk of atopic sensitization but increases the risk of bronchial
hyperresponsiveness. Int Arch Allergy Immunol 152:
151–158.
221. Basinas I, Sigsgaard T, Heederik D, Takai H, Omland O,
Andersen NT, Wouters IM, Bonlokke JH, Kromhout H,
Schlunssen V. 2012. Exposure to inhalable dust and endo-
toxin among Danish livestock farmers: results from the
SUS cohort study. J Environ Monit 14:604–614.
222. Sigsgaard T, Hoffmann HJ, Thorne PS. 2008. The role of
innate immunity in occupational allergy: recent findings.
Curr Opin Allergy Clin Immunol 8:120–125.
223. Trompette A, Divanovic S, Visintin A, Blanchard C,
Hegde RS, Madan R, Thorne PS, Wills-Karp M, Gioan-
nini TL, Weiss JP, Karp CL. 2009. Allergenicity resulting
from functional mimicry of a Toll-like receptor complex
protein. Nature 457:585–588.
224. Robinson DS. 2010. The role of the T cell in asthma.
J Allergy Clin Immunol 126:1081–1091.
225. Strickland DH, Holt PG. 2011. T regulatory cells in child-
hood asthma. Trends Immunol 32:420–427.
226. Chew GL, Wilson J, Rabito FA, Grimsley F, Iqbal S,
Reponen T, Muilenberg ML, Thorne PS, Dearborn
DG, Morley RL. 2006. Mold and endotoxin levels in the
aftermath of Hurricane Katrina: a pilot project of homes
in New Orleans undergoing renovation. Environ Health Per-
spect 114:1883–1889.
227. Rao CY, Riggs MA, Chew GL, Muilenberg ML, Thorne
PS, Van Sickle D, Dunn KH, Brown C. 2007. Character-
ization of airborne molds, endotoxins, and glucans in
homes in New Orleans after Hurricanes Katrina and Rita.
Appl Environ Microbiol 73:1630–1634.
228. Thorne PS. 2000. Inhalation toxicology models of
endotoxin- and bioaerosol-induced inflammation. Toxicol-
ogy 152:13–23.
229. Thorne PS, Metwali N, Avol E, McConnell RS. 2005.
Surface sampling for endotoxin assessment using electro-
static wiping cloths. Ann Occup Hyg 49:401–406.
230. Thorne PS, Hoppe KA, Hadina S, Adamcakova-Dodd A,
Weiss JP. 2011. Bioactivity of 14C-labeled endotoxin in
aggregates, shed blebs and whole bacteria: potency compar-
ison in the LAL assay, cell cultures, and mouse bioassays.
Eur Respir J 38:4182.
231. Spaan S, Doekes G, Heederik D, Thorne PS, Wouters
IA. 2008. Effect of extraction and assay media on analysis
of airborne endotoxin. Appl Environ Microb 74:
3804–3811.
232. Spaan S, Heederik DJJ, Thorne PS, Wouters IM. 2007.
Optimization of airborne endotoxin exposure assessment:
Effects of filter type, transport conditions, extraction solu-
tions, and storage of samples and extracts. Appl Environ
Microb 73:6134–6143.
233. Thorne PS, Perry SS, Saito R, O’Shaughnessy PT,
Mehaffy J, Metwali N, Keefe T, Donham KJ, Reynolds
SJ. 2010. Evaluation of the Limulus amebocyte lysate and
recombinant factor C assays for assessment of airborne
endotoxin. Appl Environ Microbiol 76:4988–4995.
234. Bang FB. 1956. A bacterial disease of Limulus polyphemus.
Bull Johns Hopkins Hosp 98:325–351.
235. Iwanaga S. 1993. The Limulus clotting reaction. Curr Opin
Immunol 5:74–82.
236. Reynolds SJ, Milton DK, Heederik D, Thorne PS, Don-
ham KJ, Croteau EA, Kelly KM, Douwes J, Lewis D,
Whitmer M, Connaughton I, Koch S, Malmberg P,
Larsson BM, Deddens J, Saraf A, Larsson L. 2005. Inter-
laboratory evaluation of endotoxin analyses in agricultural
dusts—comparison of LAL assay and mass spectrometry.
J Environ Monit 7:1371–1377.
3.2.6. Airborne Bacteria and Endotoxin ▪ 3.2.6-19
237. Reynolds SJ, Thorne PS, Donham KJ, Croteau EA, Kelly
KM, Lewis D, Whitmer M, Heederik DJ, Douwes J, Con-
naughton I, Koch S, Malmberg P, Larsson BM, Milton
DK. 2002. Comparison of endotoxin assays using agricul-
tural dusts. AIHA J 63:430–438.
238. Chun DT, Bartlett K, Gordon T, Jacobs RR, Larsson
BM, Larsson L, Lewis DM, Liesivuori J, Michel O, Mil-
ton DK, Rylander R, Thorne PS, White EM, Brown
ME, Gunn VS, Wurtz H. 2006. History and results of
the two inter-laboratory round robin endotoxin assay stud-
ies on cotton dust. Am J Ind Med 49:301–306.
239. Chun DT, Chew V, Bartlett K, Gordon T, Jacobs R,
Larsson BM, Lewis DM, Liesivuori J, Michel O,
Rylander R, Thorne PS, White EM, Gunn VC, Wurtz
H. 2002. Second inter-laboratory study comparing endo-
toxin assay results from cotton dust. Ann Agric Environ
Med 9:49–53.
240. Chun DT, Chew V, Bartlett K, Gordon T, Jacobs RR,
Larsson BM, Larsson L, Lewis DM, Liesivuori J,
Michel O, Milton DK, Rylander R, Thorne PS, White
EM, Brown ME. 2000. Preliminary report on the results
of the second phase of a round- robin endotoxin assay
study using cotton dust. Appl Occup Environ Hyg 15:
152–157.
241. Thorne PS, Bartlett KH, Phipps J, Kulhankova K. 2003.
Evaluation of five extraction protocols for quantification of
endotoxin in metalworking fluid aerosol. Ann Occup Hyg
47:31–36.
242. Douwes J, Versloot P, Hollander A, Heederik D, Doekes
G. 1995. Influence of various dust sampling and extraction
methods on the measurement of airborne endotoxin. Appl
Environ Microbiol 61:1763–1769.
243. Milton DK, Johnson DK, Park JH. 1997. Environmental
endotoxin measurement: interference and sources of varia-
tion in the Limulus assay of house dust. Am Ind Hyg Assoc J
58:861–867.
244. Thorne PS, Reynolds SJ, Milton DK, Bloebaum PD,
Zhang X, Whitten P, Burmeister LF. 1997. Field evalua-
tion of endotoxin air sampling assay methods. Am Ind
Hyg Assoc J 58:792–799.
245. Zock JP, Hollander A, Doekes G, Heederik D. 1996. The
influence of different filter elution methods on the meas-
urement of airborne potato antigens. Am Ind Hyg Assoc J
57:567–570.
246. Aizenberg V, Grinshpun SA, Willeke K, Smith J, Baron
PA. 2000. Performance characteristics of the button per-
sonal inhalable aerosol sampler. J Sci Occup Environ Health
Saf 61:398–404.
247. de Vocht F, Huizer D, Prause M, Jakobsson K, Peplonska
B, Straif K, Kromhout H. 2006. Field comparison of inhal-
able aerosol samplers applied in the european rubber man-
ufacturing industry. Int Arch Occup Environ Health 79:
621–629.
248. Kenny LC, Aitken R, Chalmers C, Fabries JF, Gonzalez-
Fernandez E, Kromhout H, Liden G, Mark D, Riediger
G, Prodi V. 1997. A collaborative European study of per-
sonal inhalable aerosol sampler performance. Ann Occup
Hyg 41:135–153.
249. Li SN, Lundgren DA, Rovell-Rixx D. 2000. Evaluation of
six inhalable aerosol samplers. J Sci Occup Environ Health
Saf 61:506–516.
250. Kullman GJ, Thorne PS, Waldron PF, Marx JJ, Ault B,
Lewis DM, Siegel PD, Olenchock SA, Merchant JA.
1998. Organic dust exposures from work in dairy barns.
Am Ind Hyg Assoc J 59:403–413.
251. Thorne PS, Chrischilles EA, Kuehl AKW, Kelly KM,
Metwali N, Harris LM, O’Neill ME, Quella AK, Walker
RE. 2006. Reduction of endotoxin and glucan exposures in
the Louisa/Keokuk Environmental Intervention Project
3.2.6-20 ▪ AIR
(LEIP) for rural childhood asthma. J Allergy Clin Immun
117:S215–S215.
252. Sonesson A, Larsson L, Fox A, Westerdahl G, Odham
G. 1988. Determination of environmental levels of
peptidoglycan and lipopolysaccharide using gas chro-
matography with negative-ion chemical-ionization mass
spectrometry utilizing bacterial amino acids and hydroxy
fatty acids as biomarkers. J Chromatogr 431:1–15.
253. Sonesson A, Larsson L, Schutz A, Hagmar L, Hallberg T.
1990. Comparison of the limulus amebocyte lysate test and
gas chromatography-mass spectrometry for measuring lipo-
polysaccharides (endotoxins) in airborne dust from
poultry-processing industries. Appl Environ Microbiol 56:
1271–1278.
254. Alwis KU, Larsson L, Milton DK. 2006. Suppression of
ionization and optimization of assay for 3-hydroxy fatty
acids in house dust using ion-trap mass spectrometry.
Am J Ind Med 49:286–295.
255. Rietschel ET, Brade H, Holst O, Brade L, Muller-
Loennies S, Mamat U, Zahringer U, Beckmann F, Seydel
U, Brandenburg K, Ulmer AJ, Mattern T, Heine H,
Schletter J, Loppnow H, Schonbeck U, Flad HD,
Hauschildt S, Schade UF, Di Padova F, Kusumoto S,
Schumann RR. 1996. Bacterial endotoxin: chemical con-
stitution, biological recognition, host response, and immu-
nological detoxification. Curr Top Microbiol Immunol 216:
39–81.
256. Preller L, Heederik D, Kromhout H, Boleij JS, Tielen
MJ. 1995. Determinants of dust and endotoxin exposure
of pig farmers: development of a control strategy using
empirical modelling. Ann Occup Hyg 39:545–557.
257. Roy CJ, Thorne PS. 2003. Exposure to particulates, micro-
organisms, beta(1–3)-glucans, and endotoxins during soy-
bean harvesting. AIHA J 64:487–495.
258. Smid T, Heederik D, Mensink G, Houba R, Boleij JS.
1992. Exposure to dust, endotoxins, and fungi in the ani-
mal feed industry. Am Ind Hyg Assoc J 53:362–368.
259. Smit LA, Spaan S, Heederik D. 2005. Endotoxin exposure
and symptoms in wastewater treatment workers. Am J Ind
Med 48:30–39.
260. Coggins MA, Hogan VJ, Kelly M, Fleming GT, Roberts
N, Tynan T, Thorne PS. 2012. Workplace exposure to
bioaerosols in podiatry clinics. Ann Occup Hyg 56:
746–753.
261. Madsen AM, Hansen VM, Nielsen SH, Olsen TT.
2009. Exposure to dust and endotoxin of employees in
cucumber and tomato nurseries. Ann Occup Hyg 53:
129–138.
262. Samadi S, Rietbroek NN, Dwars RM, Jamshidifard AR,
Heederik DJ, Wouters IM. 2011. Endotoxin and beta-
(1 → 3)-glucan exposure in poultry and ruminant clinics.
J Environ Monit 13:3254–3261.
263. Samadi S, Wouters IM, Houben R, Jamshidifard AR,
Van Eerdenburg F, Heederik DJ. 2009. Exposure to inhal-
able dust, endotoxins, beta(1 → 3)-glucans, and airborne
microorganisms in horse stables. Ann Occup Hyg 53:
595–603.
264. Semple S, Devakumar D, Fullerton DG, Thorne PS,
Metwali N, Costello A, Gordon SB, Manandhar DS,
Ayres JG. 2010. Airborne endotoxin concentrations in
homes burning biomass fuel. Environ Health Perspect 118:
988–991.
265. Park JH, Spiegelman DL, Gold DR, Burge HA, Milton
DK. 2001. Predictors of airborne endotoxin in the home.
Environ Health Perspect 109:859–864.
266. Mazique D, Diette GB, Breysse PN, Matsui EC, McCor-
mack MC, Curtin-Brosnan J, Williams DL, Peng RD,
Hansel NN. 2011. Predictors of airborne endotoxin
concentrations in inner city homes. Environ Res 111:
614–617.
267. Sheehan WJ, Hoffman EB, Fu C, Baxi SN, Bailey A,
King EM, Chapman MD, Lane JP, Gaffin JM, Permaul
P, Gold DR, Phipatanakul W. 2012. Endotoxin exposure
in inner-city schools and homes of children with asthma.
Ann Allergy AsthmaImmunol 108:418–422.
268. Noss I, Doekes G, Sander I, Heederik DJ, Thorne PS,
Wouters IM. 2010. Passive airborne dust sampling with
the electrostatic dustfall collector: optimization of storage
and extraction procedures for endotoxin and glucan meas-
urement. Ann Occup Hyg 54:651–658.
269. Samadi S, Heederik DJ, Krop EJ, Jamshidifard AR, Wil-
lemse T, Wouters IM. 2010. Allergen and endotoxin
exposure in a companion animal hospital. Occup Environ
Med 67:486–492.
270. Kilburg-Basnyat B, Metwali N, Thorne PS. 2015. Effect
of deployment timeon endotoxin and allergne exposure
assessment using electrostatic dust collectors. Ann Occup
Hyg 59:104–115.
 Airborne Viruses
SYED A. SATTAR, NITIN BHARDWAJ, AND M. KHALID IJAZ
3.2.7
Viruses are well known for their ability to survive in air (1, 2),
and mathematical models can predict such survival (3).
Viruses may become airborne through the expulsion of con-
taminated liquids and bodily fluids of animals and humans;
dust and dried virus-containing material may also enter air
when it is disturbed. Regardless of the means of aerosolization,
viruses can be carried over considerable distances by air cur-
rents indoors (4–6), on cruise ships (7, 8), and outdoors (9,
10). Wind-blown carriage of animal pathogenic viruses has
long been shown to cause outbreaks of disease far away from
the source (9–12). For example, an epizootic of pseudorabies
in swineherds arose from airborne spread of virus across
an area of nearly 150 km2 (13), and retrospective studies of
similar outbreaks indicated airborne spread of the virus up
to 17 km (11). Intercontinental transport of human viruses
through atmospheric dispersion of airborne particles has
also been postulated. Hammond et al. (14) suggest that
such long-distance transport of airborne viruses may explain
the pandemics of influenza while Gloster et al. (15) devel-
oped a model based on a 1967 outbreak of foot-and-mouth
disease virus (FMDV) in the United Kingdom to estimate
the risks associated with long-distance airborne virus spread.
The respiratory tract can retain viruses when contami-
nated air is inhaled, and many human (1, 16, 17) and animal
(1) pathogenic viruses can spread by this route; subsequent
shedding of virus may then occur and increase further the
virus load in air. While direct and indirect exposure to viruses
in air can occur by other means, infections from the inhala-
tion and retention, including translocation and ingestion fol-
lowing inhalation of droplet nuclei, are generally regarded as
true airborne spread (Fig. 1).
Infectious viruses (1) or their nucleic acids (18) have been
detected in both experimental (19) and field (20, 21) samples
of air. Further, experimentally contaminated air has been
shown to spread infectious agents (19, 22, 23). Naturally
occurring airborne outbreaks of viral disease target mainly
the respiratory tract. However, several viruses associated
with gastroenteritis (18, 24, 25), cancer (19), or skin infec-
tions (26) may result from the translocation and ingestion
of particles retained in the upper respiratory tract (27).
Recent years have witnessed a revival of interest in the
potential of air to spread microbial pathogens, including
viruses (27). Among the reasons for this are the devastating
outbreaks of newly discovered viral infections, such as severe
doi:10.1128/9781555818821.ch3.2.7
3.2.7-1
acute respiratory syndrome (SARS) and avian influenza (H5:
N1) and their likely spread by air (20, 28–31). Outbreak
investigations also strongly suggest air as a vehicle in the trans-
mission of noroviral infections (24, 25) while continued sur-
veillance of airborne adenovirus (32) suggests that infections
in the hospital and the community may occur through this
route. Viral bioagents released deliberately or accidentally
(28, 33) continue to gain much attention, including a recent
study performed on the potentially pandemic H5:N1 influ-
enza virus, which focused on demonstrating airborne trans-
mission in ferrets (34). These concerns have generated
much activity in the development of equipment and proce-
dures to test (35–38) and decontaminate indoor air (39)
and protect individuals against airborne contagion (40–42).
In general, however, airborne spread of viruses is rapid, effec-
tive, and difficult to prevent and control (43).
METHODS FOR THE STUDY
OF AIRBORNE VIRUSES
The methodologies for generating, storing, and collecting
viral aerosols have already been reviewed (44–51). The sam-
pling techniques for airborne microorganisms are discussed in
chapter 3.2.2. This chapter critically reviews the available
information on the role of air in the spread of vertebrate
viruses under controlled laboratory conditions as well as viral
decontamination of indoor air.
Proper study of the aerobiology of viruses requires special-
ized and custom-built equipment (52). Further, virus sus-
pensions with relatively high titers (>107 infective units/
ml) are essential for such studies because virus infectivity
can be lost during generation, storage, and collection of aero-
sols and there is also dilution of the nebulized material in air.
Work with airborne viruses also requires stringent safety pre-
cautions (53). Basically, the following equipment and proce-
dures are used for the study of airborne viruses and other types
of microorganisms.
Aerosol Generation
To study the aerobiology of viruses, it is necessary to gener-
ate particles that are small enough (<5 µm diameter) to
remain suspended in air, thus permitting their retention in
the respiratory tract on inhalation (42). Several nebulizers
are commercially available for this purpose (46, 47, 51, 54).
3.2.7-2 ▪ AIR

FIGURE 1 Direct or indirect exposure of susceptible hosts to aerosolized viruses. doi:10.1128/9781555818821.ch3.2.7.f1

Nebulizers for the aerosol delivery of therapeutic drugs also subject to the influence of various environmental factors
produce particles in the respirable range (55) and can be (46, 51). Figure 2 is a micrograph of a virus-containing par-
adapted for generating microbial aerosols. The nature and ticle captured on a spider thread (59). In spite of the limita-
composition of the virus-suspending medium determine the tions (51), ultrafine threads are the best means to study the
size distribution of the particles aerosolized and protect the
viability of the virus during nebulization and subsequent
aging (1).
A physical tracer in the virus suspension to be sprayed is
needed to differentiate between physical loss of the infectious
virus due to the settling out of the aerosol and its actual bio-
logical decay. Uranine, a fluorescent dye, added to the virus
suspension can be as effective as using a radiolabeled virus
(56). At the levels needed, the dye is generally harmless to
viruses and cell cultures and is cheaper and safer to use than
radioisotopes. However, dyes may affect the photosensitivity
of viruses (46). Bacterial spores are also suitable tracers (46,
51). Nebulization fluids with mixtures of viruses have been
used to directly compare their airborne stabilities (57). Anti-
foam agents in the virus suspension may also be needed to
reduce frothing during nebulization (56).

Retention and Aging of Aerosols

Whereas the nebulized virus can be held as a static aerosol in
any type of closed container, a rotating stainless steel drum
(56, 58) is used to store virus aerosols in a dynamic state for
studying the influence of various environmental factors.
The drum is housed in a larger chamber which is vented
through HEPA filters to prevent exposure to accidental virus
leakage. The drum is continuously rotated mechanically
along its axis at a predetermined rate (e.g., 4 rpm) to keep
the aerosolized material from settling out. This allows the
virus to stay suspended in air for weeks. The air inside the
drum can be preconditioned to the desired relative humidity
(RH) level and temperature. Thorough air flushing of the
drum is required between experiments. However, such drums FIGURE 2 Scanning electron micrograph of an aerosol particle
or similar devices are not available commercially. attached to a strand of spider web. Bar = 0.5 µm. (Reproduced with
Ultrafine (<1 µm in diameter) natural (spider web) and permission from Dr. B. Kournikakis of the Defence Research Estab-
artificial (tungsten) threads can be used as supports for small lishment, Suffield, Alberta, Canada.).
virus-containing particles. Such anchored particles are still doi:10.1128/9781555818821.ch3.2.7.f2

influence of atmospheric chemicals (47) and light and irradi-
ation (59, 60) on airborne viruses. Handley and Roe (61)
have used an artificial fiber (2–3 µm diameter) of ethylene
vinyl acetate as an alternative to spider threads; tests with bac-
terial aerosols have found them to perform as well as spider
threads. Their suitability for viruses remains to be evaluated.
Collection and Sizing of Aerosols
Most of the available aerosol collection devices (1, 46, 49–
51) take in relatively small volumes (about 10 liters/min) of
air. In an all-glass impinger (AGI), the air drawn in passes
through a tube with a limiting orifice and impacts on the
surface of the collecting fluid, usually Earl balanced salt solu-
tion. An antifoaming agent in the collection fluid is usually
needed to reduce frothing during the aerosol sampling process
(56). The volume of air sampled depends on the sampling
rate of the impinger and the run-time. If the air is being
sampled from an aerosol-holding device, its volume and the
total number of samples to be collected for a given experiment
will determine the volume of individual air samples. Pro-
longed operation of AGI for air sampling can also lead to
the evaporation of the sampling fluid, thereby affecting the
concentration of solutes in it. Use of preimpingers with
AGI improves the collection of airborne viruses (1). The vol-
ume of air sampled must be replaced with fresh air to avoid
creating a vacuum in the aerosol-holding device.
In the field, commercial large-volume air samplers, which
can easily process 10 m3 of air per min, can be used to recover
airborne viruses (1, 46, 51) by impacting the air on a thin layer
of a collecting fluid, such as Earl balanced salt solution. The
collection efficiencies of various large-volume air samplers-
vary considerably depending on the type of virus, the type
and volume of air sampled, the nature of the collection fluid,
and the rate of sampling. Wallis et al. (61) reported the use of
acidic buffer–moistened cartridge filters to recover poliovi-
ruses aerosolized during the flushing of toilets; the air could
be sampled with flow rates as high as 100 liters/min, but the
virus eluted from the filters required further concentration.
Determination of the size distribution of airborne particles
containing infectious viruses has required the adaptation (1)
of the Andersen six-stage sampler (35). Instead of a bacterio-
logical agar medium, tryptose phosphate broth containing
3% gelatin is placed in the sampler’s petri plates. After sam-
pling, the collection medium is liquefied by holding the plates
at 37°C for 1 h and poured out for virus titration and measure-
ments of the physical tracer levels. In a May three-stage glass
impinger (62), virus retention corresponds to that in the
human upper respiratory tract, bronchioles, and alveoli. A
comparison of the three-stage Andersen sampler and the
three-stage May AGI gave similar results in the sampling
and sizing of artificially generated biological aerosols (63).
Several other methods have been designed based on the
Andersen sampler model, including compact cascade impac-
tors (64); these can test pressurized flow as well as the develop-
ment of filter cartridges including those made of Teflon and
gelatin (65). In all cases, the adsorption of virus particles to
media and/or surfaces is the critical requirement for the enu-
meration of viruses.
Exposing Experimental Animals or Human
Subjects to Airborne Viruses
Table 1 summarizes information from selected reports pub-
lished in the past 42 years. Studies on the susceptibility of ani-
mals to airborne viruses, or experiments to determine the
minimal infective dose of a given virus by the respiratory
3.2.7. Airborne Viruses ▪ 3.2.7-3
route, require exposing animals to standardized clouds of
the infectious agent under controlled conditions. Tradition-
ally, the Henderson apparatus (134) has been used for this
purpose when working with small animals such as mice and
rats. In this device, individual animals are placed in cylindri-
cal holders with only the tip of the animal’s face (including
the nostrils) protruding from one end of the holder. The hold-
ers are then attached to ports on a larger tube connected to
the aerosol source, and the only means of exposure of the ani-
mals to the aerosols is by inhalation of the air in the tube.
Devices for working with cattle (101) or pigs (135) have
also been described. It is important to note that the aerosol-
ized virus must not be allowed to deposit on areas where the
test animals could be exposed to it in ways other than through
inhalation. Chung et al. (136) reported the development of a
low-cost wind tunnel for exposing human subjects to aerosols
of >2 µm diameter. Particles generated by a Collison-type
nebulizer and introduced into the tunnel by using an aerosol
injector are uniformly distributed to expose human subjects.
However, no reports have been published on the use of the
device with airborne viruses.
SOURCES OF VIRAL AEROSOLS
Activities such as sneezing, coughing, flushing toilets, and
changing diapers as well as shaking, homogenization, and
sonication of virus-containing materials can generate infec-
tious aerosols (1). For some gastrointestinal viruses, aerosol-
ized feces and vomitus may also transmit infection (24).
Inhalation of infectious virus particles released into air during
certain surgical procedures can also pose risks to human
health. Under experimental conditions, human papillomavi-
rus was isolated from aerosols and shown to transfer infection
from one tissue to another (19).
Preventing the generation of and avoiding exposure to
such aerosols is particularly important in laboratory settings
where infectious material is handled (137). According to
Pike (138), 27% of the cases of laboratory-acquired infec-
tions were due to airborne viruses; cases in research settings
accounted for more than 67% of such infections. Whereas
improvements in the design and construction of biohazard
containment equipment (44) and better enforcement of bio-
safety procedures (139) have considerably reduced the risk of
aerosol exposure, many laboratory workers do not appear to
be fully aware of the dangers of infectious aerosols. Improved
training of laboratory workers, continued vigilance, and fur-
ther improvements in equipment design and procedures are
needed to make these workplaces safer still.
AIRBORNE SURVIVAL OF HUMAN
AND ANIMAL PATHOGENIC VIRUSES
How long a given virus can remain infectious in air depends
on the nature of the virus and the medium in which it was sus-
pended before becoming airborne, ambient temperature, RH,
atmospheric gases, lighting, and irradiation (1). In 1971, Ben-
bough attempted to understand the survival of airborne
viruses including polio- and Langat viruses (140). Forty years
later, Posada (3) developed mathematical models to estimate
the survival of these two viruses as well as the influenza virus.
More recently, Yang and Marr (141) have hypothesized that
pH changes in airborne particles due to RH-dependent evap-
oration of water affect the survival of aerosolized enveloped
viruses by inducing conformational changes in their surface
glycoproteins. Despite the interest in certain viruses for their
airborne potential, there is an absence of standardization of
3.2.7-4 ▪ AIR

TABLE 1 Experimental challenge of animals or humans to viral aerosols: chronological list of selected studies published between 1970
and 2012

Virus(es) Host(s) Remarks Ref.

Rauscher murine Mice Nearly 40% of exposed mice developed leukemia within 25 months. (66)
leukemia virus
Maloney murine sarcoma Mice When tumor extracts were aerosolized, both viruses survived for at least 2 h, (67)
and leukemia virus but mice exposed to the aerosols did not develop an infection.
complex
Adenovirus 12 Hamsters Airborne virus shown to be pathogenic to newborn animals. (68)
Adenovirus 4, Human volunteers Determined infectivity of these viruses by aerosols. Small particle aerosol (69, 70)
coxsackievirus A21 inoculation led to infection with only 0.5 TCID50. Human infection
occurred in 50% of cases with 9 TCID50
Parainfluenzavirus type 3 Calves First report of extensive purulent pneumonia in calves after exposure to (70)
aerosols of the virus and Pasteurella haemolytica.
Parainfluenzavirus type 1 Mice Transmissibility rates did not increase after serial airborne challenge with the (71)
virus.
Marek’s disease virus Chickens Exposure to effluent air from “donor cages” housing infected animals resulted (72)
in a high incidence of the infection in test chicks; passage of contaminated
air through certain filters partially or completely prevented such infection.
Chickens Establishment of an aerosol-based Marek’s disease virus infection model. (73)
Newcastle disease virus Chickens Antibody response to aerosol challenge much better than that from (74)
(vaccine) administration in drinking water; no clinical disease; virus recovered from
lungs for 10 days.
Newcastle disease virus Chickens Unvaccinated birds shed much higher levels of virus than those previously (75)
vaccinated.
Chickens Occurrence and transmission of Newcastle disease virus aerosol originating (76)
from infected chickens under experimental conditions
Vesicular stomatitis virus Mice Exposure of mice for 1 h to ozone resulted in a 70% increase in respiratory (77)
deposition of the virus.
Influenza virus Mice Under conditions of aerosol inhalation, mice were found to be a suitable model (54)
for studies on pathogenesis.
Mice Extrapulmonary virus was in direct quantitative relationship to the extent of (78)
lung involvement.
Mice Mouse resistance to viral pneumonia affected in presence of manganese (79)
dioxide. Studied the dynamics of B lymphocytes in lungs after aerosol
challenge;
Mice IgA- and IgM-containing cells appeared first, followed by IgG-containing (80)
ones.
Mice Intranasal immunization was found to be superior to aerosol exposure in (81)
immunization of mice.
Horses Challenge of animals to aerosolized influenza virus was found to be a reliable (82)
means of assessing immunization against the disease.
Humans Administered the vaccine by either nose drops or large-particle aerosols; (83)
aerosol route was more effective and well tolerated.
Horses Vaccinated and unvaccinated ponies were exposed to aerosols of a field isolate (84)
of influenza virus; vaccinated animals showed complete protection even
after 15 months of immunization.
Mice Mouse nose only inhalation model provides the means to evaluate the efficacy (85)
of drug and vaccine candidates against the relevant route of challenge,
thereby providing data that may better predict clinical outcome.
Ferrets Several studies demonstrate genetically modified A/H5N1 virus acquired (34, 86,
mutations during passage in ferrets to ultimately become air-transmissible 87)
in ferrets.
Bovine rhinotracheitis Cattle Compared clinical and immunological responses after aerosol exposure or (88)
virus (bovine intramuscular inoculation; in both cases, the virus generally elicited
herpesvirus 1) comparable levels of serum antibody but not measurable nasal antibody;
aerosol-exposed cattle shed virus from the nose while the others did not.
Bovine herpesvirus Calves Strain 1.3 produced severe encephalitis with minimal respiratory lesions, (89)
strains 1.1 and 1.3 whereas strain 1.1 gave respiratory lesions but no neurological disease;
colostrum-fed animals became subclinically infected.
(Continued on next page)
 3.2.7. Airborne Viruses ▪ 3.2.7-5

TABLE 1 (Continued )
Virus(es) Host(s) Remarks Ref.
Bovine herpesvirus 1 Cattle Susceptible calves exposed to air from infected animals became (90)
infected.
Calves Five-minute exposure of vaccinated and control animals to experimentally (70)
generated viral aerosols through a face mask.
Calves Administration of the combination modified-live BHV-1 vaccine yielded (89)
significant disease-sparing effects in calves experimentally infected with
aerosolized virulent field strains of BHV-1
Bovine viral diarrhea Cattle Susceptible calves exposed to air from infected animals became infected. (90)
virus
Bovine respiratory Cattle Susceptible calves exposed to air from infected animals became infected. (90)
syncytial virus
Bovine respiratory Holstein calves Animals exposed to aerosols of the virus manifested moderate to severe signs of (91)
syncytial virus respiratory disease.
Calves RSV activates the vitamin D intracrine pathway in the infected calf lung, and (92)
vitamin D status of the host does have an immunomodulatory role during
RSV infection
Feline caliciviruses Cats Concluded that aerosol transmission probably plays little part in the spread of (93)
these viruses.
Feline caliciviruses Cats Animals could be infected by aerosol exposure or direct intranasal instillation. (94)
African swine fever virus Pigs Concluded that the primary route of infection in pigs was through the lower (95)
(KWH/12) respiratory tract.
Pigs Animals became infected after challenge with aerosolized virus. (96)
Rinderpest virus Cattle Low or high RH were shown to increase the probability of disease transmission (97)
by the respiratory route, but any aerial spread across distances greater than a
few meters was believed to occur principally at night.
Japanese B encephalitis Mice, rats, Mice and hamsters highly susceptible to aerosol challenge; guinea pigs and rats (98)
virus hamsters, guinea seroconverted but survived the infection; squirrel monkeys died after a high
pigs, squirrel dose of infectious virus.
monkeys
Rauscher murine Mice Infection of BALB/c mice through aerosols of leukemogenic virus was (99)
leukemia virus possible.
Foot-and-mouth disease Cattle Cattle infected with aerosolized virus. (100, 101)
virus
Calves In situ hybridization method showed presence of virus in target organs as early (102)
as 6 h after aerosol challenge.
Cattle Sampling exhaled air for as little as 1 min allowed the detection of FMDV in (100)
cattle infected experimentally.
Aujeszky’s disease virus Pigs Virus transmitted to seronegative pigs exposed to air from boxes containing (103)
(Herpesvirus) infected pigs.
Pigs Naive pigs exposed to the breath of intranasally infected ones through a mask- (104)
to-mask means became infected as evidenced by seroconversion. A good
model for transmission and vaccination studies.
Pigs After a 15-min exposure of 6 animals to aerosolized virus (4.5 log TCID50) all (105)
became infected. A proportion of sentinel pigs housed with the infected
animals also became infected.
Infectious bronchitis Chickens Birds exposed to aerosols had earlier and slightly more severe respiratory (106)
virus (Australian symptoms.
T strain)
Lassa fever virus Guinea pigs and Both species susceptible when exposed to small-particle aerosols. (107)
monkeys
Mouse rotavirus virus Mice Neonatal mice developed acute gastroenteritis within 48 h of exposure to viral (108)
aerosols.
Rift Valley fever virus Rats When exposed to viral aerosols, 97% of the unvaccinated and 32% of the (109)
vaccinated animals died; levels of serum neutralizing antibody were
predictive of the protective effect.
Rhesus macaques The mutagen-attenuated RVF MP-12 vaccine was protective against (110)
intravenous and aerosol challenge with virulent RVFV in macaques.
Mice DNA vaccine protects against aerosolized RVFV challenge. (111)
(Continued on next page)
3.2.7-6 ▪ AIR

TABLE 1 Experimental challenge of animals or humans to viral aerosols: chronological list of selected studies published between 1970
and 2012 (Continued )

Virus(es) Host(s) Remarks Ref.

Hantaan Seoul and Rats Female Wistar rats (12–16 weeks old) exposed to aerosols of the viruses (112)
Puumala viruses became infected.
Andes virus (hantavirus) Cynomolgus Only hantavirus known to spread from person to person. Cynomolgus (113)
macaques monkeys exposed to aerosolized virus; no clinical disease but antibody
production.
Fowl pox virus Chickens Day-old chicks were vaccinated by the aerosol route or by injection in the wing (114)
web; both methods induced immunity but the aerosol method more suitable
for mass vaccinations.
Marburg virus Guinea pigs and Susceptible animals exposed to air from infected ones became infected. (115)
rhesus monkeys
Monkeys Monkeys vaccinated with a VSV vector expressing the glycoprotein of Zaire (116)
Ebolavirus (ZEBOV) or Marburg (MARV) were completely protected
against an aerosol exposure of ZEBOV and MARV, respectively.
Monkeys Angola strain of Marburg virus causes lethal hemorrhagic fever in monkeys (117)
exposed via aerosol route.
Mice (MARV) and (ZEBOV) are more virulent when administered via the IP route (118)
rather than by aerosol infection in experimental mice.
Junin virus Rhesus monkeys All animals exposed to aerosolized virus became sick within 3 weeks and died; (119)
clinical picture similar to that seen in humans.
Equine herpesvirus 1 Horses Exposure of pregnant mares to viral aerosols resulted in abortions, but virus (120)
could not be demonstrated in the aborted fetuses.
Horses Protection induced by the modified live virus vaccine is superior to that (118)
induced by the inactivated combination vaccine against an aerosol
challenge
Respiratory coronavirus Pigs Aerosol challenge of young pigs induced a strong response in bronchial lymph (121)
nodes but not in the mesenteric lymph nodes.
Venezuelan equine Monkeys The protective effect of vaccines against aerosolized challenge with VEEV was (122)
encephalitis virus tested in cynomolgus monkeys implanted with temperature-monitoring
(VEEV) radiotelemetry devices.
Mice Animals first exposed to the recombinant virus intranasally and (123)
challenged with aerosolized VEEV. Potential for vaccination of
humans against VEEV.
Swine fever virus Swine Naive pigs were housed in an isolation unit with three compartments (23)
(Asfivirus) with separate ventilation systems. An infected pig in a separate
compartment was able to infect only those animals with a connected
ventilation system.
Ovine herpesvirus 2 Sheep Causes malignant catarrhal fever in ruminants. Virus refractory to in vitro (124)
cultivation. Uninfected sheep exposed to aerosols of various dilutions of
nasal secretions from infected sheep. Minimal infective dose determined for
further studies on viral pathogenesis and vaccine protection.
Sheep Clinical signs and lesions resembling MCF can develop in sheep when they are (125)
exposed to a high dose of aerosolized OvHV-2.
Cattle Nasal secretions collected from sheep experiencing OvHV-2 shedding (126)
episodes were infectious for cattle via nebulization and is capable of
inducing MCF.
Rabbit Infection of rabbits by intranasal nebulization is a potentially useful model that (127)
mimics the natural route of infection and may be used to study viral
replication and pathogenesis.
Sheep Lung is the primary replication site for OvHV-2 during initial infection in (128)
sheep.
Pig MCF can be experimentally induced in pigs by aerosol challenge using sheep (129)
nasal secretions containing OvHV-2.
Smallpox virus (variola) Cynomolgus Development of a subhuman primate model for the study of aerosolized (130)
monkeys smallpox pathogenesis and transmission.
Rabbitpox virus Rabbit The aerosol infection model in rabbits may serve as an appropriate model for (30)
evaluation of antivirals under development for the therapeutic treatment of
human smallpox.
(Continued on next page)

TABLE 1 (Continued )
Virus(es) Host(s)
Rabbit Smallpox-like disease can be produced in rabbits in a controlled fashion
through exposure to a small-particle rabbitpox virus aerosol, and
rabbitpox spreads from animal to animal by the airborne route in a
laboratory setting.
Porcine reproductive and Pigs Naive pigs were exposed for three hours to virus aerosolized up to 150 m
respiratory disease away; three of the animals became infected. Virus also recovered from
virus (Arteriviridae) the air.
Pigs PRRSV isolate pathogenicity may influence aerosol transmission of the virus.
the experimental protocols and wide variations in the system
of reporting the results make direct comparisons of the find-
ings from different studies extremely difficult.
The literature on this topic has been reviewed (1); how-
ever, some new information has been developed in light of
the more recent outbreaks of FMDV, SARS, and pandemic
influenza over the past decade. The following is an update
of the information from experimental and field studies on
selected human and animal pathogenic viruses. Some of
the viruses selected are known to spread by the airborne route,
whereas for the others there appears to be a real or perceived
risk for airborne transmission. It should be noted that many of
the viruses spread through aerosols are not normally associ-
ated with infections of the respiratory tract.
VZV
Investigations of chickenpox outbreaks clearly show airborne
spread of varicella-zoster virus (VZV). Sawyer et al. (26)
found VZV DNA in air samples from rooms housing patients
with zoster or chickenpox; and the air remained positive up to
24 h after patient discharge. Whereas none of the VZV
DNA-positive samples had infectious virus, integrating cell
culture and reverse transcription quantitative PCR (ICC-
RT-qPCR) assay-based technology may prove to be very use-
ful in studying the airborne transport of viruses in indoor and
outdoor settings.
Hepburn et al. (142) have described an outbreak of chick-
enpox in a military field hospital. Several patients and staff
members in an isolation ward for cases of gastroenteritis
were affected, and virus spread was most likely through air.
This incident highlights the weaknesses in field hospital
design and illustrates how easily biological warfare agents
could spread in such settings (143).
A recent study to identify the spread of VZV in patients
treated with oral acyclovir (144) showed a significant air-
borne presence of the virus. By day 3 of acyclovir treatment,
29% of air purifiers had detectable traces of VZV DNA.
This partially correlated with the detection of VZV in throat
swabs of between 17% and 32% of patients.
Viruses Causing Acute Gastroenteritis
Caliciviruses
That norovirus can spread by air is suggested from outbreak
investigations (24, 25, 145, 146). The first such report was
based on an outbreak associated with a hospital emergency
room (18): individuals who simply walked through the con-
taminated area also became infected. Contamination of air
is believed to occur mainly by the aerosolization of the virus
during vomiting (147). However, proper studies on the
3.2.7. Airborne Viruses ▪ 3.2.7-7
Remarks Ref.
(131)
(132)
(133)
airborne spread of these viruses are difficult because they can-
not be cultured in the laboratory.
Rotaviruses
Rotaviruses are among the major causes of acute gastroenter-
itis in humans and animals (21, 148). Apart from their spread
by the fecal-oral route, epidemiological and experimental
studies suggest airborne transmission of rotaviral infections
(148, 149). The behavior of murine and other strains of
bovine rotaviruses in air appears to parallel that of previously
studied human and animal rotaviruses (56, 150).
Viruses Causing Infections
of the Respiratory Tract
The behavior of aerosolized influenza and parainfluenza
viruses of humans and animals has been studied in some detail
(1). More recent studies in this regard have focused on chal-
lenging vaccinated animals with viral aerosols to study the
protective effect of immunization (Table 1).
The relative importance of air and other vehicles in the
spread of rhinovirus colds continues to be debated (151,
152). Airborne human rhinovirus type 14 (RV-14) has
shown behavior in ways typical of other picornaviruses
(153); its half-life was nearly 14 h when it was aerosolized
from tryptose phosphate broth and the aerosols were held at
20°C with 80% RH. This finding suggests that rhinoviruses
can remain infectious in air long enough to permit aerial
spread. Recent improvements in the recovery and detection
of rhinoviruses from indoor air (5) and their application in
field investigations of rhinovirus colds (6) provide additional
evidence in support of this method of transmission.
Papillomaviruses
Reports of warts in the respiratory tract of laser therapists
(154, 155) suggest exposure to papillomaviruses in the smoke
plumes from vaporized verrucae. Garden et al. (156) first
reported using DNA probes for the detection of papillomavi-
rus DNA in such plumes from CO2 laser vaporization of
bovine fibropapilloma and human verrucae. Subsequently,
PCR was used to confirm that the DNA in the laser plumes
corresponded to the type in the lesion (157). In fact, infec-
tious papillomavirus particles have also been detected in
the vapors produced during CO2 laser treatment as well as
electrocoagulation of bovine warts (142). The same study
also found that the amounts of DNA generated during the
laser treatment of human and bovine warts were higher
when compared to those released during electrocoagulation;
surgical masks were shown to be effective in filtering out
the viruses in the vapors from both types of treatment.
3.2.7-8 ▪ AIR
Inoculation of cattle with debris from CO2 laser plumes gen-
erated during the vaporization of bovine warts failed to infect
the cattle (158). Another PCR-based study (159) reported
widespread contamination of the facial area of the therapists
and the operating room environment with papillomavirus
DNA released during CO2 laser treatment and electrocoagu-
lation of human warts and neoplasia.
The experiments of Garden et al. (19, 156) give further
proof that laser plumes generate infectious aerosols and that
viruses in them have the potential to cause disease (160).
Bovine cutaneous fibropapillomas were subjected to simu-
lated excision surgery using a CO2 laser and the material aero-
solized as the laser plume was collected. It was not only
positive for papillomavirus DNA, it also induced tumors in
the skin of inoculated calves.
Retroviruses
Little is known about airborne spread of retroviruses (1). The
advent of AIDS, however, has led to some recent studies on
HIV-1. Although airborne spread of AIDS is not known, sur-
vival of HIV-1 in air was tested to assess the risk of exposure of
health care personnel during orthopedic surgery. Infectious
HIV-1 was detected in aerosols from certain types of surgical
power tools (161), and the blood-containing aerosols were
in the respirable range (162). Baggish et al. (163) reported
detection of HIV-1 in CO2 laser smoke when pellets of exper-
imentally infected cells were vaporized.
Airborne spread of HIV-1 and other retroviruses could
occur where high-titered suspensions of such viruses are
handled, and strict adherence to safety precautions is neces-
sary to eliminate the risk. It would be useful to know if cell-
associated HIV-1 (in both infectious and proviral forms)
behaves differently in air when compared to cell-free virus.
Viruses Causing Hemorrhagic Fevers
The ability of viruses causing hemorrhagic fevers to be trans-
mitted through air remains unclear (164). Their handling
requires the highest level of biohazard containment (53).
However, laboratories with such facilities are limited, and
those that exist may not be equipped for aerobiological stud-
ies. In many cases, the airborne spread of these viruses has
been discovered through laboratory accidents or patterns of
disease transmission in hospitals (165, 166) and animal hold-
ing facilities.
The survival of Lassa fever virus in artificially generated
aerosols was favored at low (30%) RH levels, and even at
32°C the virus survived long enough to permit its dispersal
by air (107). Experimental aerosol exposure of monkeys and
guinea pigs could infect and kill them, with the median
50% infectious dose for guinea pigs being as low as 15 PFUs.
Laboratory-acquired infections in those handling hantavi-
ruses or animals carrying these viruses are well documented
(167–171). The most likely means of exposure in these
studies was by inhalation of infectious aerosols. Inhalation
of artificially generated aerosols of hantaviruses can infect
rats (112).
Rhesus monkeys became acutely ill and died when
exposed to aerosols of Junin virus (119), the agent of Argen-
tine hemorrhagic fever. The symptoms and pathology of the
disease were very similar to those after parenteral exposure
and also mimicked the clinical syndrome in humans.
The lab-acquired case of Sabia virus infection was most
likely due to aerosol exposure (159).
SARS Coronavirus
In 2003, an outbreak of a novel coronavirus (CoV) that
caused an atypical pneumonia (SARS) startled the world
and was poised to be the next pandemic. However, to date,
the virus has caused just over 8,400 cases with fewer than
1,000 deaths and studies into disease management have all
but curtailed the potential for a pandemic (172). Recently,
a novel virus, named the Middle East respiratory syndrome
coronavirus (MERS-CoV), which produces a clinical picture
similar to that of SARS, was isolated from infected humans in
Saudi Arabia (63).
The available epidemiological evidence strongly suggests
that SARS spreads through droplets (173), and this spread
is much easier to control than that through aerosols due to
the inability of droplets to travel over longer distances. How-
ever, the pattern of spread of SARS-CoV in at least two
instances is highly suggestive of airborne spread.
A cluster of 329 SARS cases was recorded in one apart-
ment complex in Hong Kong, with the majority of them
occurring on several floors in one wing (173). This pattern
of spread is highly suggestive of virus dissemination by air,
and the aerosolization of the virus was speculated to have
occurred from malfunctioning sewers in the building (173).
Rodent pests in the building have been hypothesized as pos-
sible amplifiers and disseminators of the virus (173); the virus,
most likely acquired from infected residents, may have multi-
plied in the rodents, released in their excreta and then possi-
bly becoming airborne. While hantaviruses, for example, can
infect humans from aerosolized rodent excreta (112), the role
of air in the transmission of SARS-CoV in this outbreak
remains speculative at this stage. However, a more detailed
analysis of the outbreak of SARS at that apartment complex
suggests that the virus released by the patients themselves may
have spread through air (20).
Limited airborne spread of SARS may have occurred on
board commercial aircraft. In one such instance, a sympto-
matic index case infected at least 22 of 120 (18.3%) passen-
gers and crew during a 3-h flight (20). Some of those
infected were seated over 2 m away from the index case, a dis-
tance much longer than the 0.9 m generally believed to be
the limit for droplet transmission. In such retrospective inves-
tigations, it is virtually impossible to rule out the role of other
possible means of virus spread. Also, the inside of an aircraft
combines features that may be more conducive to airborne
spread of pathogens.
A recent report from Canada suggests that oxygen delivery
masks with open vents could promote the dispersal of respira-
tory pathogens such as SARS-CoV through their enhanced
release in mists of exhaled pulmonary gases (174). The
exhaled moist air ejected from such oxygen masks is believed
to carry pathogen-laden droplets over longer distances and
possibly contribute to an increased risk of spread of respiratory
infections in nosocomial settings. Additional investigations
are needed to first prove that viruses such as SARS-CoV
can retain their infectivity better in the warm, moisture-laden
air exhaled from oxygen masks. The findings on the influence
of RH and air temperature on the airborne survival of corona-
virus 229E would tend to suggest otherwise (175).
The findings of earlier studies on the effect of RH and air
temperature on experimentally aerosolized 229E, another res-
piratory coronavirus of humans, further reinforce the poten-
tial of SARS-CoV to spread by air (175). As is true for
enveloped viruses in general, 229E survived better at 30–
50% RH than at 80% RH when the air temperature was about
20°C. Under these conditions, the half-lives of the virus at
30%, 50%, and 80% RH were 27, 67, and 3 h, respectively.

Lowering the air temperature to 6°C increased the half-lives
of the virus at 30% and 50% RH to 34 and 103 h, respec-
tively. But the lower air temperature produced the most dra-
matic effect on virus survival at 80% RH and changed its
half-life from 3 to over 86 h.
Avian Influenza
In 1997, an H5:N1 variant of influenza was detected in
humans (176). Since then, the presence of avian influenza
virus has been rigorously monitored and actions to prevent
its spread are quickly instituted. Although there is ample evi-
dence of transmission of virus from animals to humans, there
has not yet been any demonstration of human-to-human
transfer of infection without close contact and transmission
of high levels of virus from one individual to another. Several
groups have developed strains of the H5:N1 virus that could
spread through the airborne route in ferrets (34, 86, 87).
These studies demonstrate the capability of the virus to
develop an airborne state, yet the rigorous laboratory work
required to achieve this outcome suggest there is little likeli-
hood of the natural evolution of an airborne route for this
virus. Yet the use of airborne precautions including a pro-
posed Containment Level 4 requirement when working
with these mammalian transmissible viruses (177) is strongly
recommended.
Viruses of Domestic Animals
FMDV
Foot-and-mouth disease virus (FMDV) has been studied
extensively for its capacity to spread through air (1, 104,
105), and several airborne outbreaks of the disease have
been documented (12). The 2001 outbreak in the United
Kingdom resulted in the slaughter of more than 6 million cat-
tle. A prediction model was proposed by Gloster et al. (11) in
which 7 of 12 farms were infected through airborne transmis-
sion of the virus from nearby farms. They used meteorological
models to defend this theory. Other models to forecast and
analyze outbreaks of FMDV are now available and continue
to be postulated (10, 178). More recently FMDV outbreaks
have been reported from Republic of Korea and Egypt (179,
180). Airborne FMDV has a relatively small minimal in-
fective dose; when sheep were experimentally exposed in cab-
inets to virus-infected pigs, as little as a dose of 10 cell culture
infective units was found to be sufficient to infect 50% of the
animals exposed (181).
NDV
There is convincing evidence for the airborne spread of New-
castle disease virus (NDV). Recently, Hietala et al (21)
showed that exotic NDV was detected in air samplers within
2 h of being introduced into a flock of commercial poultry
flocks. Thus, air filters in poultry houses are highly effective
in preventing outbreaks of the disease (182). Kournikakis
et al. (59, 183) have used a vaccine (LaSota) strain of NDV
as a model to study the airborne survival and behavior of
enveloped viruses as well as field test protective equipment
and methodologies for the rapid collection and identification
of viruses in air. The vaccine, available in the lyophilized
form, is relatively easy to work with and safe. Each vaccine
vial can yield nearly 108 PFU/ml. The influence of air tem-
perature and RH on the airborne survival of this virus was
very similar to that of other enveloped viruses (1, 184). A
combination of low levels of RH (20–30%) and a low temper-
ature (about 10°C) is optimal for its survival in air in the dark
3.2.7. Airborne Viruses ▪ 3.2.7-9
(59); under such conditions, 56% of the aerosolized virus
remained infectious even after 6 h, while the corresponding
figure for 20°C was 39%. There was nearly a 99% drop in virus
infectivity when NDV-containing aerosols were captured on
spider threads and held under daylight conditions with or
without a cloud cover (Fig. 3).
Negative air ions may reduce the spread of NDV in
chicken farms (185). Infected animals were placed upwind
from the susceptible ones and under controlled conditions
of air temperature (26.7°C), RH (50%), and ventilation rates
(0.34–1.36 m3/min); there was a 28% reduction in virus
transmission when negative ion generators were used.
Pseudorabies Virus
Even though the herpesvirus that causes pseudorabies is rel-
atively fragile, its capacity to spread through air parallels
(13) that of FMDV, a picornavirus. The virus in air sur-
vived the best when RH was 55% and the air temperature
was 4°C (186); the half-life of the virus under these optimal
conditions was less than 1 h. Although pigs are the natural
host species, this virus is capable of causing fatal encepha-
litis in other species as well. Recently pseudorabies virus
infection was reported in hunting dogs in the United States
(187).
AEROSOL CHALLENGE STUDIES
Several studies have challenged human or animal hosts with
artificially generated viral aerosols (1). Table 1 summarizes
the information from selected studies published on this topic
since 1970. The following conclusions are based on a critical
analysis of these studies. (1) In most cases, susceptible hosts
became infected on exposure to the test virus, but often the
amount of virus inhaled was either unknown or may have
been unrealistically high. (2) In certain cases, the experimen-
tal design and set-up did not exclude virus exposure of test
subjects by means other than through inhalation. (3) in spite
of the importance of reporting RH and air temperature in
such experiments, many of the investigators failed to mention
these parameters. (4) Of necessity, nearly all of these studies
have used laboratory-adapted strains of the test virus(es);
the extent to which these data apply to viruses circulating
in the field remains undetermined and perhaps undetermin-
able. (5) Any meaningful comparison of the data from these
studies is difficult because there are no standard procedures
for aerosol challenge studies. Therefore, much develop-
mental work is needed before accurate and reliable informa-
tion on the behavior of airborne viruses can be generated
and such data can be applied to designing mathematical mod-
els and strategies to control and prevent disease spread
through air.
COLLECTION OF NATURALLY OCCURRING
VIRAL AEROSOLS
Collecting infectious viruses from naturally occurring aerosols
continues to be difficult because the devices currently avail-
able for the purpose are generally noisy, bulky, expensive,
and somewhat inefficient. As is true for many other types of
environmental sampling, it is often too late to look for the
suspected virus in air in an outbreak investigation. Even if
it were feasible, regular monitoring of air for viruses is not rec-
ommended because of the limited significance of the findings.
The data from molecular biological techniques (e.g., PCR/
ICC-RT-qPCR) for viruses in air sample concentrates will
be much more meaningful if the nucleic acid detected could
3.2.7-10 ▪ AIR

FIGURE 3 Biological decay of NDV captured on a spider web and held under sunlight (a) and under cloudy conditions
(b). Control samples were held in the dark. (Reproduced with permission from Dr. B. Kournikakis of the Defence
Research Establishment, Suffield, Alberta, Canada.) Expon. = The rate of loss in the viability of the virus.
doi:10.1128/9781555818821.ch3.2.7.f3

reliably signal presence or absence of virus infectivity. How-
ever, there are certain situations where sampling of air for nat-
urally occurring viral aerosols can be extremely valuable. This
is exemplified by the studies of bat caves for airborne rabies
virus (188, 189). In this regard, not only was infectious rabies
virus recovered from the air in the caves, but experimental
animals exposed (in insect-proof cages) to such air died of
rabies (188, 189).
Airborne spread of naturally occurring viral infections of
humans and animals is known. In some of these cases (e.g.,
measles virus, influenza virus), air may be the chief mode of
virus transmission. Others may be isolated instances of air-
borne spread of a virus, which is normally transmitted by
direct contact or by other vehicles. But it must be noted
that any virus that can survive aerosolization has the potential
for airborne transmission.
TECHNOLOGIES TO RID INDOOR
AIR OF VIRUSES
The recent revival of interest in the potential of air as a
vehicle for human and animal pathogens has spawned
renewed interest in the disinfection of air to inactivate air-
borne pathogens and prevent their spread. The principles of
disinfection have been extensively reviewed (190, 191),
although the majority of literature has focused on animate
and inanimate surfaces as well as water. Many technologies
aimed at removal/inactivation of viruses in indoor air have
been developed, and Table 2 provides a summary of such
reports. While some of the experimental work was not pub-
lished in peer-reviewed journals, the efficacy of these technol-
ogies requires that they be mentioned here. Moreover, the
means by which these technologies were tested demonstrates
 3.2.7. Airborne Viruses ▪ 3.2.7-11

the inherent limitations of the respective test methods, while that UV254 inactivation of four bacteriophages (MS-2, phi
reemphasizing the need for standardization of such testing X174, phi 6, T-7) decreased as relative humidity increased
methods. At present, there are only two strategies that have at temperatures of 25°C to 28°C and speculated that de-
proven their effectiveness at reducing airborne viral load, creased UV254 susceptibility under higher RH conditions
and they are briefly described next. resulted from attenuation of UV254 by water sorption onto
the viral surface. In contrast, in a study involving bacterio-
Oxygen-Based Methods phage MS-2, respiratory adenovirus serotype 2, and mouse
hepatitis virus (coronavirus), Walker and Ko (197) reported
While oxygen is a necessity of life, several species of the ele-
that UV254 inactivation increased as RH increased (temper-
ment are highly reactive and lethal. The use of reactive oxy-
ature conditions not reported). This same general trend was
gen represents a highly rapid and effective means to reduce
reported by McDevitt et al. (196), who stated there was an
the microbial load on surfaces and water (201) and in air
increase in UV254 susceptibility of the vaccinia virus with
(202). Currently, several reactive oxygen-generating technol-
an increase in relative humidity (temperature conditions
ogies have been developed and tested with three specific
not reported). Given the conflicting data reported, clarity
active agents showing the highest potential. They include
on the mechanism of UV inactivation of aerosolized viruses
ozone, vaporized hydrogen peroxide (VHP), and cold gas
awaits outcome of further studies (213).
plasma. In all three cases, a water vapor containing the active
The use of UV to inactivate microbes in air has relied on
agent is developed and then expelled into the target environ-
the stable placement of a UV-generating device that contin-
ment. The subsequent mist allows for widespread contact
ually flows air in an enclosed space. In theory, as time pro-
with any particles including viruses in the air.
gresses, the total volume of air will be disinfected, leaving
Ozone (O3) has been used for water disinfection for well
the environment safe. For potential application of UV to
over a century and continues to be an excellent means to ren-
decontaminate airborne pathogens, the challenge remains
der water safe. Studies have demonstrated that airborne ozone
to optimize UV for large volume of air including determina-
can reduce the levels of several bacteria (203, 204) and
tion of the UV inactivation constant for airborne pathogens
viruses, including bacteriophages (205, 206) and influenza
under different environmental conditions (temperature and
viruses (207). However, ozone is toxic and concentrations
RH). Another concern of the use of UV254 in the context
required for inactivation are magnitudes higher than the per-
of airborne viruses was raised by Mattle and Kohn (214)
missible exposure levels stated by the U.S. National Institute
who observed a tailing effect in which viruses appeared to sur-
for Occupational Safety and Health. The use of ozone has
vive the UV process and remain viable and infectious. The
been limited to specific environments where full enclosure
work was conducted in water as opposed to air, yet the find-
and posttreatment ventilation is possible.
ings clearly demonstrated that viruses can be shielded from
In contrast to ozone, hydrogen peroxide is nontoxic at
UV radiation through aggregation and then rescue other
effective levels of microbial inactivation. This trait has
viruses through recombination during high multiplicity of
enabled the use of VHP in several health care environ-
infection. The potential for reduced activity due to aggre-
ments. There has been some concern that catalase-producing
gates—either through clumping of viruses or adsorption to
microorganisms, such as Staphylococcus aureus, may resist
airborne particles such as bodily fluids—remains to be
VHP (208). While this may not a concern with viruses, the
investigated.
use of VHP against human pathogenic viruses remains unex-
plored. VHP has demonstrated efficacy against prions (209),
which are known to be the hardest pathogens to inactivate. CONCLUDING REMARKS
Oxygen gas plasma is formed when oxygen is exposed to
Increasing use of recycled air will further enhance the risk of
an electric field. Oxygen gas plasma consists of a high number
exposure of susceptible individuals to airborne viruses. Such
of highly reactive elemental species including ozone and oxy-
spread of viruses is also believed to exacerbate asthma and
gen radicals. The use of plasma for disinfection and steril-
other ailments of the respiratory tract (215, 216). Therefore,
ization began in the 1960s and focused on oxygen-based
there is an urgent need for improvements in the design
measures by Nelson and Berger (210), who used oxygen gas
and operation of air-handling/recycling systems, particularly
plasma to inactivate spores of Bacillus subtilis and Clostridium
in hospitals, clinics, offices, residential complexes, sports
sporogenes. Terrier et al. (200) showed that oxygen gas plasma
facilities, airports (http://www.securityinfowatch.com/news/
could inactivate several airborne respiratory viruses, includ-
10594478/indianapolis-airport-plans-airborne-infection-iso
ing human parainfluenzavirus-3, respiratory syncytial virus,
lation-room), and aircraft (217, 218). The continuing in-
and influenza virus H5:N2. Moreover, the levels of ozone
crease in the numbers of immunosuppressed and transplant
were significantly lower than the permissible exposure level.
patients, including the aging population, also underscores
this need. A U.S. survey on the association of outbreaks
Ultraviolet Irradiation of infections with competitive sports has shown the infec-
All living organisms are threatened by exposure to UV light. tious agents involved to be predominantly viruses, with sev-
Cutler and Zimmerman reviewed the principles of UV irradi- eral instances of airborne transmission (219). Innovative
ation as it relates to inactivation of infectious agents (211) approaches are required to combine energy and resource con-
and conclude the wavelengths between 200 and 280 nm (ger- servation to reduce the risk of infection spread through con-
micidal UV) affect the double-bond stability of adjacent car- ditioned and recycled air. Additional studies are also needed
bon atoms in molecules including pyrimidines, purines, and to establish better methods for disinfecting airborne viruses in
flavin, leading to inactivation of microbial agents by forma- recycled air.
tion of dimers in RNA (uracil and cytosine) and DNA (thy- Airborne dissemination of noninfectious viruses or their
mine and cytosine). A wavelength of 253.7 nm (UV254) has components can lead to problems in laboratories using PCR
become the standard wavelength for microbial inactivation. or other such techniques, and presently available contain-
The mechanism of UV inactivation of aerosolized viruses ment facilities may not be quite adequate to address this issue
varies depending on the study. Tseng and Li (212) reported (220). Infectious virus–containing aerosol generation by
 3.2.7-12 ▪ AIR
TABLE 2 Existing and emerging technologies for removal/inactivation of viruses in indoor air in institutional and domestic settings (since 2005)
Test methodology used for
Active/technology Virus(es) tested Field trials Remarks Reference
evaluation
The Microgenix air Aerosolized MS-2 phage as a Simulated HVAC system where NA Sensors needed to warn of (192)
purification system surrogate for viruses. coarse-filtered air was drawn at 60– filter blockage and/or UV
(MAPS) is composed of a 1,500 m3/h. The aerosolized phage lamp failure. Using the
chemical-coated filter and a passed through the duct-mounted phage obviated any
ultraviolet (UV) source test system and sampled biosafety concerns.
operating at a flow rate of down-stream using cyclone
500 m3/h designed to samplers operating at 700 liters/
reduce airborne microbes in min. The inactivation efficiency
HVAC systems. was 97.34% with UV on and
61.46% with UV off.
Hydroxl/Odorox Product Claims to inactivate, bacteria, NA NA Hydroxyl generating device http://eairsolutions.com/
Technology viruses, mold, and fungal claiming broad-spectrum pbenefits.htm
spores both on surfaces and microbicidal activity.
in air. Unable to find paper on
how this technology has
been evaluated supporting
microbicidal claims
including neutralization of
hazardous gasses such as
H2S, converting odors
(volatile organic
compounds) into harmless
by-products and clean
every surface and crevice
penetrating into and
through hockey/football
equipment, cushions,
carpet and other difficult
materials eliminating smell
and bacteria.
Nonthermal plasma reactors H5N2 avian flu strain (A/ Viral aerosols were generated by a Systematic airborne For H5N2, a 4- to 5-log (193)
unit: the mobile Finch/England/2051/91 six-jet Collison nebulizer. The decontamination studies in single-pass reduction was
air-decontamination unit H5N2), a strain not known sampler design was based on the an airborne-infections seen when the system was
used in this study (Plasmair to be pathogenic to humans principles of a traditional AGI 30 isolation (AII) room were activated, compared with a
T2006; AirInSpace, Paris, and used serves as a liquid impinger, the only difference used to test the 1-log reduction when the
France) is an surrogate for the lethal being that 50 ml sterile plastic effectiveness of the system was turned off. Tests
FDA-registered medical H5N1 avian and H1N1 disposable vials were used as Plasmair air conducted at different
device (510k K070722) swine influenza viruses. collection vessels. decontamination unit temperatures and relative
intended as a room air under different room humidity (RH) showed
purifier/recirculating air ventilation scenarios. The similar performance at
cleaner and used for test room 32.8 m3 (3.6 × temperatures ranging from
filtering out and 3.25 × 3 2.8 m) was located 10°C to 40°C and at
inactivating airborne in a closed wing of the varying RH up to
 particles from the air for
medical purposes. The
unit’s physical dimensions
are 1.94 m high, 30.9 m
wide, 30.6 m deep, and its
nominal operating range is
500–2,000 m3/h. Air is
drawn from the base of the
unit near the floor, treated
with nonthermal plasma
reactors, and exhausted
from the top.
Phocatox Technologies: this Viruses, bacteria such as
multiphase process uses
HEPA filtration, hydroxyl
radical production, purified
O3 and vaporized gas-phase
hydrogen peroxide (H2O2)
production, singlet oxygen
and oxyradical plasma
production, and sterilizing
germicidal UVC radiation,
all generated internally in
one portable unit to
produce unparalleled
decontamination
capabilities on site.
TriAir T250: hydroxyl Gram-positive,
radicals
Inov8 Air Disinfection unit: Gram-positive,
hydroxyl radicals.
hospital, and all joints in condensing conditions (e.
the wall, ceiling, and g., RH 98%), which is
windows were hermetically characteristic of
sealed with nonthermal plasma
aluminum-backed tape. decontamination systems.
NA NA Claims air and surface http://phocatox.com/
methicillin-resistant decontamination covering
Staphylococcus aureus wide range of pathogens,
(MRSA) and Clostridium including airborne viruses,
difficile. bacteria such as MRSA
and C. difficile, fungi, fine
particulate matter such as
aerosols, smoke, fumes,
dust, ash and pollen, and
volatile organic
compounds.
NA NA Claims to kill broad range of http://www.
Gram-negative bacteria, microorganisms, including tri-airdevelopments.com/
and enveloped and viruses. products_intro.cfm
non-enveloped viruses.
The aerobiology chambers The units were installed in In vitro virucidal efficacy of http://www.inov8.com/
Gram-negative bacteria, approximately the size of a three “Nightingale” wards the units against

3.2.7. Airborne Viruses ▪ 3.2.7-13

fungi, mycobacteria
including MS2 phage as
surrogate for nonenveloped
viruses.
Sharp Air Purifier, which Bacteria and viruses.
combines plasmacluster ion
technology with multiple
layers of filtration.
single-bedded isolation room
located at Leeds University, UK,
and Health Protection Agency,
UK, were used (experimental
details not available).
NA
of Hereford County aerosolized MS2 phage was
Hospital substantiated by absence of
norovirus spread in the
wards expected to spread to
other patients in the
absence of units.
NA Make broad claims http://www.sharp.ca/
inactivating allergens, en-CA/ForHome/
odors including bacteria HomeEnvironment/
and viruses, particularly AirPurifier.aspx?
influenza and rhinovirus. PCI=HowItWorks
(Continued on next page)

TABLE 2 Existing and emerging technologies for removal/inactivation of viruses in indoor air in institutional and domestic settings (since 2005) (Continued )
Active/technology
Upper-room 254 nm (UVC)
light (UV254)
Upper-room 254 nm (UVC)
light (UV254)
Upper-room 254 nm (UVC)
light (UV254)
Virus(es) tested
Porcine reproductive and
respiratory syndrome
(PRRS) virus.
Influenza A virus (H1N1,
PR-8).
Vaccinia virus (as a surrogate
for the smallpox virus).
3.2.7-14 ▪ AIR
Test methodology used for
Field trials Remarks Reference
evaluation
PRRS virus continuously flowed from NA PRRS virus was more (193)
Reservoir 1 to Reservoir 2 and then susceptible to UV254 as
across a UV254 exposure field temperature decreased;
separated by a manifold (each most susceptible to UV254
quartz tube represented a different inactivation at RH
level of UV254 treatment) in between 25–79%, less
Reservoir 2 to equally distribute susceptible at RH ≤24%,
aerosolized PRRS virus into four and least susceptible at
quartz tubes. Airborne PRRS virus ≥80% RH.
received 4 levels of UV254
treatment at each combination of
temperature and RH.
The effect of UV254 (virus ranging NA Influenza virus susceptibility (194)
from 4 to 12 J/m2) was evaluated increases with decreasing
against influenza at three relative relative humidity.
humidity levels (25%, 50%, and
75%) using a bench top
aerobiology chamber. The virus
was aerosolized by six-jet Collison
nebulizer and air was drawn
through the chamber by a pump at
25 liters/min through a manifold
attached to 2 SKC Biosamplers,
each operating at 12.5 liters/min.
Under simulated real-world NA The upper-room UV fixtures (195)
conditions including the effects of produced decreases in
convection, mechanical mixing, airborne virus
temperature, and RH, the concentrations that would
effectiveness of UV254 air require additional
disinfection was determined. ventilation of more than
Aerosolized virus was delivered 87 air changes per hour.
into the testing chamber (1.5 m The susceptibility of
above the floor in the center of a vaccinia virus to UV
climate controlled 4.60 m × increased at low RH.
62.97 m × 63.05 m high room)
equipped with a ceiling fan and two
black boxes containing 100 W
light bulbs (simulate body heat of
two people). The UV light source
was five wall-mounted Hygeaire®
 UVC fixtures, each using one 25 W
low-pressure mercury discharge
lamp with a UVC output of
5. Fixtures were mounted 2.3 m
from the floor.
Upper-room 254 nm Vaccinia virus (as surrogate for The aerobiology chamber consisted of NA Vaccinia virus susceptibility (196)
(UVC) light (UV254) smallpox). three sections, (a) aerosol increased with decreasing
generation (six-jet Collison RH. They concluded that
nebulizer) and RH control virus susceptibility did not
(measured via an Omega RH32 appear to be a function of
temperature and RH meter), (b) virus-containing aerosol
UVC exposure, and (c) sampling. particle size.
Air was drawn through the
chamber by a pump at 28.3 liters/
min through a manifold attached to
the sampler. A HEPA filter was
connected after the samplers to
remove all of viral aerosols before
the airstream entered the pump.
When sampling was not in
progress, the aerosol-laden
airstream running through the
chamber was bypassed around the
samplers and the flow was directed
to the HEPA filter. The entire
apparatus was set up inside a 6-foot
Class II biosafety cabinet and
maintained under negative pressure
with respect to the cabinet interior.
The UV doses tested ranged from
0.1 to 3.2 J/m(2), at three RH
levels (20%, 60%, and 80%).

3.2.7. Airborne Viruses ▪ 3.2.7-15

Upper-room 254 nm (UVC) Respiratory adenovirus
light (UV254) (serotype 2), murine
hepatitis virus (MHV)-
coronavirus (a surrogate for
the SARS coronavirus),
and the bacteriophage
MS-2.
The testing chamber consisted of
three parts: the intake plenum
(165 mm high × 165 mm wide ×
244 mm long), the main body (50
mm high × 260 mm wide × 455
mm long), and the exhaust plenum
(49 mm high × 49 mm wide × 100
mm long). A 25 4 nm UV fixture
containing six 36 W UV lamps,
was installed above a fused quartz
UV exposure window (260 mm
NA Both the MS-2 and (197)
adenovirus aerosols were
very resistant to UV air
disinfection, with a
reduction of <1 log of
aerosolized infectious virus
at a UV dose of 2608 íW s/
cm2. The susceptibility of
coronavirus aerosols was 7–
10 times that of the MS-2
and adenovirus aerosols.
(Continued on next page)
 3.2.7-16 ▪ AIR
TABLE 2 Existing and emerging technologies for removal/inactivation of viruses in indoor air in institutional and domestic settings (since 2005) (Continued )
Test methodology used for
Active/technology Virus(es) tested Field trials Remarks Reference
evaluation
wide × 260 mm long) in the main Unlike bacterial aerosols,
body. Airflow through the system there was no significant
was maintained at 12.5 liters/min protective effect of high
throughout the experiments. The RH on UV susceptibility of
airflow pattern was checked using the tested viral aerosols.
smoke tubes. The aerosols were
generated by a six-jet Collison
nebulizer and samples of viral
aerosols were collected during the
experiments using an AGI-30
liquid impinger.
Ultraviolet germicidal Four different bacteriophages: A Collison three-jet nebulizer (BGI, NA Airborne viruses with (198)
irradiation (UVGI) – eight ssRNA (MS-2, ATCC Waltham, MA) was used to single-stranded nucleic
UVGI lamps (Philips 15597-B1), ssDNA ( phi nebulize the bacteriophages. acid (ssRNA and ssDNA)
Germicidal Lamp, X174, ATCC 13706-B1), Exposure of aerosolized virus to a were more susceptible to
TUV8W/G8 T-5, Holland) dsRNA ( phi 6 with given intensity of UV was carried UV inactivation than were
that emitted at peak 253.7 envelope lipid, ATCC out by passing the airborne viral those with
nm (UV-C) for germicidal 21781-B1), and dsDNA particles through a cylinder (5 cm double-stranded ones
action. (T7, ATCC 11303-B1). diameter, 28 cm length, made of (dsRNA and dsDNA). For
quartz) at a distance from 0 to 30 all tested viruses at the
cm from the UV source (with a same UVGI dose,
radiation peak at 254 nm). The inactivation at 85% RH
UV irradiance intensity was was higher than that at
measured using a radiometer 55% RH, which was
(P-97503-00, Cole-Parmer). For thought to be due to water
sampling Andersen one-stage sorption onto a virus
viable impactor was used. surface that provided
protection against
UV-induced DNA/RNA
damage at higher RH.
Ozone generator Four different bacteriophages: The exposure chamber used was ∼23 NA Using Anderson six-stage (199)
(OZ1PCS-V/SW, Ozotech, ssRNA (MS-2, ATCC liters (ID 14 cm and height 38 cm). sampler they found more
Yreka, CA) with pure 15597-B1), ssDNA ( phi The testing conditions were: RH, than 95% of
 oxygen at 3 l/min. Ozone X174, ATCC 13706-B1),
levels were measured by an dsRNA ( phi 6 with
ozone analyzer (model 401, envelope lipid, ATCC
Advanced Pollution 21781-B1), and dsDNA
Instruments, San Diego, (T-7, ATCC 11303-B1).
CA)
Cold oxygen plasma Human parainfluenzavirus-3;
respiratory syncytial Virus;
influenza virus H5N2.
medium (RH 55%) or high (85%)
at 25–28°C temperature. The
gaseous ozone was continuously
generated through a Teflon tube
into the chamber at a flow rate of 3
liters/min. In addition, the
virus-containing aerosols were also
generated continuously into this
chamber from a Collison three-jet
nebulizer at a flow rate of 3 liters/
min. The test system was located in
a chemical hood so that the
exhausted gas was vented outside.
For virus-containing aerosol
samples collection, Andersen
one-stage viable impactor was used
before and after ozone treatment.
This stage has 400 0.25-mm holes
and has a sampling flow rate of
28.3 liters/min (corresponding to a
velocity of 24 m/s) when 20 ml
Luria-Bertani broth is used with 3%
gelatin plates.
Viruses were nebulized into a flow
tunnel with sampling ports
containing 3 ml of phosphate
buffer saline) both upstream and
downstream of the plasma. Air
speed was fixed at 0.9 m/s.
virus-containing aerosols
to be less than 2.1 μm in
diameter. They found the
order of resistance to ozone
as phi6 < phi X174 < MS-2
<T-7. Considering viral
ultrastructure, in
comparison with simple
capsid architecture ( phi
X174), more complex virus
capsid (MS-2 and T-7) was
observed to be less
susceptible to ozone
(P<0.05). They found T-7
with more complex capsid
architecture to be less
susceptible to ozone than
MS-2 and based on this
observation conclude,
MS-2 could not be a
suitable indicator for virus
inactivation by ozone in
air. They also found the
susceptibility of viruses to
ozone was higher at 85%
RH than that at 55% RH.
This might be related to
the generation of more
radicals from ozone which
reacted with more water
vapor at higher RH.
NA There is potential for the use (200)
of oxygen plasma in
airborne disinfection on
3.2.7. Airborne Viruses ▪ 3.2.7-17
the condition that the
ozone levels remain safe.
3.2.7-18 ▪ AIR
surgical tools (142, 159, 221) and laboratory equipment
(222) suggest that the potential of any such technology
should be assessed at the design stage and corrective measures
taken. This will require improved communication between
design engineers and end-users of the devices, including
infection control practitioners. More information is also
needed on effective means of protecting animals and humans
in case of deliberate or accidental release of viral aerosols.
Recent attempts at developing predictive models of exposure
to continuous or temporary sources of aerosolized virus show
that sheltering indoors may have a protective effect (43).
The increased awareness of airborne infection, particularly
in hospitals, has escalated the use of face masks, but the pro-
tective ability of these masks has been questioned in studies
using respirable but nonmicrobial aerosols (222). There are
no standardized methods to test the ability of face masks to
withstand challenge with airborne infectious agents.
Recently, alternation of the lung airway surface properties
by spraying normal saline into the respiratory tract has been
found to diminish the number of exhaled bioaerosols by
human subjects by over 70% (217). This appears to be due
to changes in the surface tension of the fluid lining the respi-
ratory tract. While corroboration of these findings is needed,
the use of a readily available and nontoxic fluid such as nor-
mal saline may prove to be an effective and broad-spectrum
means of reducing the aerial loading of infectious agents
from patients with respiratory viral and other infections.
Immunization of humans and animals by exposure to arti-
ficially aerosolized virus is a very attractive alternative to cur-
rent practices of parenteral or oral vaccination. Apart from
being rapid and practical, such antigen administration more
closely mimics natural exposure to many viruses and gives
longer-lasting immunity (123, 223). However, there are
attendant risks in introducing live, albeit attenuated virus
directly into the respiratory tract of susceptible hosts. Inhala-
tion of inactivated viruses or even their purified antigens and
other components of the vaccine may also cause allergic
reactions. Therefore, much initial testing will be necessary
before any mass vaccinations through aerosolization could
be attempted.
Rapid inactivation (59) and higher levels of dilution of
vertebrate viruses in open air suggest that they may be less
likely to spread under outdoor conditions, but there is now
enough evidence that certain types of animal pathogenic
viruses can be carried by air over several kilometers, leading
to outbreaks at distant locations (9, 13, 224). While some sug-
gest that this is true for human pathogens as well (14), much
work is needed to substantiate this possibility. Perhaps long-
distance transport of viruses may explain the peculiar trans-
continental movement of outbreaks of rotaviral infections
(225).
Additional work is required to develop safe and effective
means of inactivating viruses in air, and standardized methods
are also needed to test as well as validate the claims of devices
marketed for this purpose. Recent studies have shown that
ozone has the potential to inactivate viruses in aerosolized
blood products (205). Further work on such an approach
could provide alternative means of ridding heat-sensitive
injectables of harmful viruses.
Airborne viruses continue to be a threat to human and
animal health in spite of the sophisticated design and efficient
protective functioning of the respiratory system (226). Much
of the research thus far has documented outbreaks of airborne
viral infections and identified to some degree the factors that
influence virus survival in air. Now greater emphasis is needed
on reducing generation of infectious aerosols in indoor as well
as outdoor settings and on enhancing the removal and/or
inactivation of viruses (33, 45) and other infectious agents
in air. Many human and animal pathogenic viruses, some of
which are relatively obscure and out of the mainstream, are
considered potential bioweapons; recent accounts (227–
229) of their production and stockpiling attest to this. The
true extent of past and possibly ongoing work in this regard
will remain shrouded in mystery because the details may never
become available to the general scientific community.
We are grateful to Bill Kournikakis and Canada’s Defence Research
Establishment for the permission to reproduce the scanning electron micro-
graph and the graphs. The assistance of Jason Tetro of CREM in the prep-
aration of this manuscript is gratefully acknowledged.
REFERENCES
1. Sattar SA, Ijaz MK. 1987. Spread of viral infections by
aerosols. CRC Crit Rev Environ Control 17:89–131.
2. Whon TW, Kim MS, Roh SW, Shin NR, Lee HW, Bae
JW. 2012. Metagenomic characterization of airborne viral
DNA diversity in the near-surface atmosphere. J Virol 86:
8221–8231.
3. Posada JA, Redrow J, Celik I. 2010. A mathematical
model for predicting the viability of airborne viruses.
J Virol Methods 164:88–95.
4. Goldmann DA. 2000. Transmission of viral respiratory
infections in the home. Pediatr Infect Dis J 19:S97–102.
5. Myatt TA, Johnston SL, Rudnick S, Milton DK. 2003.
Airborne rhinovirus detection and effect of ultraviolet irra-
diation on detection by a semi-nested RT-PCR assay. BMC
Pub Hlth 3:5.
6. Myatt TA, Johnston SL, Zuo Z, Wand W, Kebadze T,
Rudnick S, Milton DK. 2004. Detection of airborne rhi-
novirus and its relation to outdoor air supply in office envi-
ronments. Am J Respir Crit Care Med 169:1187–1190.
7. Carling PC, Bruno-Murtha LA, Griffiths JK. 2009.
Cruise ship environmental hygiene and the risk of norovi-
rus infection outbreaks: an objective assessment of 56 ves-
sels over 3 years. Clin Infect Dis 49:1312–1317.
8. Gupta S. 2002. Cruising for trouble. Don’t blame the ocean
liners. You can get this stomach bug almost anywhere. Time
160:86.
9. Christensen LS, Mortensen S, Botner A, Strandbygaard
BS, Ronsholt L, Henriksen CA, Andersen JB. 1993. Fur-
ther evidence of long-distance airborne transmission of
Aujeszky disease (pseudorabies) virus. Vet Rec 132:
317–321.
10. Donaldson AI, Alexandersen S. 2002. Predicting the
spread of foot and mouth disease by airborne virus. Rev
Sci Tech 21:569–575.
11. Gloster J, Champion HJ, Mansley LM, Romero P,
Brough T, Ramirez A. 2005. The 2001 epidemic of foot-
and-mouth disease in the United Kingdom: epidemiologi-
cal and meteorological case studies. Vet Rec 156:793–803.
12. Maragon S, Facchin E, Moutou F, Massirio I, Vincenzi G,
Davies G. 1994. The 1993 Italian foot-and-mouth-disease
epidemic—epidemiologic features of the 4 outbreaks iden-
tified in Verona Province (Veneto Region). Vet Rec 135:
53–57.
13. Grant RH, Scheidt AB, Rueff LR. 1994. Aerosol trans-
mission of a viable virus affecting swine: explanation of
an epizootic of pseudorabies. Int J Biometeorol 38:33–39.
14. Hammond GW, Raddatz RL, Gelskey DE. 1989. Impact of
atmospheric dispersion and transport of viral aerosols on the
epidemiology of influenza. Rev Infect Dis 11:494–497.
15. Gloster J, Jones A, Redington A, Burgin L, Sorensen JH,
Turner R, Dillon M, Hullinger P, Simpson M, Astrup P,

Garner G, Stewart P, D’Amours R, Sellers R, Paton P.
2010. Airborne spread of foot-and-mouth disease—model
intercomparison. Vet J 183:278–286.
16. Echavarria M, Kolavic SA, Cersovsky S, Mitchell F, San-
chez JL, Polyak C, Innis BL, Binn LN. 2000. Detection of
adenoviruses (AdV) in culture-negative environmental
samples by PCR during an AdV-associated respiratory dis-
ease outbreak. J Clin Microbiol 38:2982–2984.
17. Eickhoff TC. 1994. Airborne nosocomial infection—a
contemporary perspective. Infect Contr Hosp Epidemiol
15:663–672.
18. Sawyer MH, Chamberlin CJ, Wu YN, Aintablian N,
Wallace MR. 1994. Detection of varicella-zoster virus
DNA in air samples from hospital rooms. J Infect Dis
169:91–94.
19. Garden JM, O’Banion MK, Bakus DA, Olson C. 2002.
Viral disease transmitted by laser-generated plume (aero-
sol). Arch Dermatol 138:1303–1307.
20. Booth TF, Kournikakis B, Bastien N, Ho J, Kobasa D,
Stadnyk L, Li Y, Spence M, Paton S, Henry B, Mederski
B, White D, Low DE, McGeer A, Simor A, Vearncombe
M, Downey J, Jamieson FB, Tang P, Plummer F. 2005.
Detection of airborne severe acute respiratory syndrome
(SARS) coronavirus and environmental contamination
in SARS outbreak units. J Infect Dis 191:1472–1477.
21. Hietala SK, Hullinger PJ, Crossley BM, Kinde H, Ardans
AA. 2005. Environmental air sampling to detect exotic
Newcastle disease virus in two California commercial poul-
try flocks. J Vet Diagn Invest 17:198–200.
22. Barry M, Bia F, Cullen M, Dembry L, Fischer S, Geller G,
Hierholzer W, Mcphedran P, Rainey P, Russi M, Snyder
E, Wrone E, Gonzalez JP, Ricohesse R, Tesh R, Ryder R,
Shope R, Quinn WP, Galbraith PD, Cartter ML, Hadler
JL, Demaria A. 1994. Arenavirus infection—Connecti-
cut, 1994. J Am Med Assoc 272:998–999.
23. Dewulf J, Laevens H, Koenen F, Mintiens K, de Kruif A.
2000. Airborne transmission of classical swine fever virus
under experimental conditions. Vet Rec 147:735–738.
24. Marks PJ, Vipond IB, Carlisle D, Deakin D, Fey RE,
Caul EO. 2000. Evidence for airborne transmission of Nor-
walk-like virus (NLV) in a hotel restaurant. Epidemiol Infect
124:481–487.
25. Marks PJ, Vipond IB, Regan FM, Wedgwood K, Fey RE,
Caul EO. 2003. A school outbreak of Norwalk-like virus:
evidence for airborne transmission. Epidemiol Infect 131:
727–736.
26. Suzuki K, Yoshikawa T, Tomitaka A, Matsunaga K,
Asano Y. 2204. Detection of aerosolized varicella-zoster
virus DNA in patients with localized herpes zoster.
J Infect Dis 189:1009–1012.
27. Slote L. 1976. Viral aerosols—a potential occupationally
related health threat in aerated wastewater treatment sys-
tems. J Environ Hlth 38:310–314.
28. National Institute of Standards and Technology Chal-
lenges in microbial sampling in the indoor environment work-
shop summary. Report TN 1737. 2012. U.S. Dept.
Commerce. http://www.nist.gov/mml/bbd/biomolecular/
upload/Sloan-Roport-Final-TN-1737.pdf.
29. Li KS, Guan Y, Wang J, Smith GJD, Xu KM, Duan L,
Rahardjo PA, Puthavathana P, Buranathai C, Nguyen
TD, Estoepangestie ATS, Chaisingh A, Auewarakul P,
Long HT, Hanh NTH, Webby RJ, Poon LLM, Chen
H, Shortridge KF, Yuen KY, Webster RG, Peiris JSM.
2004. Genesis of a highly pathogenic and potentially pan-
demic H5N1 influenza virus in eastern Asia. Nature 430:
209–213.
30. Roy CJ, Milton DK. 2004. Airborne transmission of
communicable infection—the elusive pathway. N Engl J
Med 350:1710–1712.
3.2.7. Airborne Viruses ▪ 3.2.7-19
31. Yu IT, Wong TW, Chiu YL, Lee N, Li Y. 2005. Temporal-
spatial analysis of severe acute respiratory syndrome among
hospital inpatients. Clin Infect Dis 40:1237–1243.
32. Wan GH, Huang CG, Huang YC, Huang JP, Yang SL,
Lin TY, Tsao KC. 2012. Surveillance of airborne adenovi-
rus and Mycoplasma pneumoniae in a hospital pediatric
department. PLoS One 7:e33974. doi:10.1371/journal.
pone.0033974.
33. Gooze LL, Hughes EC. 2003. Smallpox. Semin Respir
Infect 18:196–205.
34. Herfst S, Schrauwen EJ, Linster M, Chutinimitkul S, de
Wit E, Munster VJ, Sorrell EM, Bestebroer TM, Burke
DF, Smith DJ, Rimmelzwaan GF, Osterhaus AD, Fouch-
ier RA. 2012. Airborne transmission of influenza A/H5N1
virus between ferrets. Science 336:1534–1541.
35. Andersen AA. 1958. New sampler for the collection, siz-
ing, and enumeration of viable airborne particles.
J Bacteriol 76:471–484.
36. Crook B. 1995. Inertial samplers: biological perspective,
p 247–267. In Cox CS, Wathes CM (eds), Bioaerosols hand-
book. CRC Press, Boca Raton, FL.
37. Crook B. 1995. Non-inertial samplers: biological perspec-
tive, p 269–283. In Cox CS, Wathes CM (eds), Bioaerosols
handbook. CRC Press Inc, Boca Raton, FL.
38. Zimmerman J, Reist PC, Turner AG. 1987. Comparison
of two biological aerosol sampling methods. Appl Environ
Microbiol 53:99–104.
39. Utrup LJ, Frey AH. 2004. Fate of bioterrorism-relevant
viruses and bacteria, including spores, aerosolized into
an indoor air environment. Experiment Biol Med 229:
345–350.
40. Agranovski IE, Safatov AS, Borodulin AI, Pyankov OV,
Petrishchenko VA, Sergeev AN, Agafonov AP, Ignatiev
GM, Sergeev AA, Agranovski V. 2004. Inactivation of
viruses in bubbling processes utilized for personal bioaero-
sol monitoring. Appl Environ Microbiol 70:6963–6967.
41. Franz DR, Jahrling PB, Friedlander AM, McClain DJ,
Hoover DL, Bryne WR, Pavlin JA, Christopher GW, Eit-
zen EM, Jr. 1997. Clinical recognition and management of
patients exposed to biological warfare agents. J Am Med
Assoc 278:399–411.
42. Verreault D, Moineau S, Duchaine C. 2008. Methods for
sampling of airborne viruses. Microbiol Mol Biol Rev 72:
413–444.
43. Casal J, Planas-Cuchi E, Casal J. 1999. Sheltering as a pro-
tective measure against airborne virus spread. J Haz Mat 68:
179–189.
44. Clark RP. 1995. Safety cabinets, fume cupboards and other
containment systems, p 473–504. In Cox CS, Wathes CM
(eds), Bioaerosols handbook. CRC Press, Boca Raton, CL.
45. Clark S, Scarpino P. 1996. Disinfection of microbial aero-
sols, p 215–223. In Hurst CJ (ed), Modeling disease transmis-
sion and its prevention by disinfection. Cambridge University
Press, Cambridge, UK.
46. Cox CS. 1987. The aerobiological pathway of microorganisms.
John Wiley & Sons, Chichester, UK.
47. Cox CS. 1989. Airborne bacteria and viruses. Sci Prog
73:469–499.
48. Ijaz MK, Sattar SA, Alkarmi T, Dar FK, Bhatti AR,
Elhag KM. 1994. Studies on the survival of aerosolized
bovine rotavirus (Uk) and a murine rotavirus. Compr
Immunol Microbiol Infect Dis 17:91–98.
49. Mitchell JP. 1995. Aerosol generation for instrument cali-
bration, p 101–175. In Cox CS, Wathes CM (eds), Bioaer-
osols handbook. CRC Press, Boca Raton, FL.
50. Mitchell JP. 1995. Particle size analyzers: practical proce-
dures and laboratory techniques, p 177–246. In Cox CS,
Wathes CM (eds), Bioaerosols handbook. CRC Press, Boca
Raton, FL.
3.2.7-20 ▪ AIR
51. Spendlove JC. 1957. Production of bacterial aerosols in a
rendering plant process. Pub Hlth Rep 72:176–180.
52. Wathes CM, Johnson HE. 1991. Physical protection
against airborne pathogens and pollutants by a novel ani-
mal isolator in a level 3 containment laboratory. Epidemiol
Infect 107:157–170.
53. Fleming DO, Richardson JH, Tulis JJ, Vesley D. 2000.
Laboratory safety: principles and practices. ASM Press, Wash-
ington, DC.
54. Frankova V. 1975. Inhalatory infection of mice with
influenza A0/PR8 virus. I. The site of primary virus replica-
tion and its spread in the respiratory tract. Acta Virol 19:
29–34.
55. Knoch M, Wunderlich E, Geldner S. 1994. A nebulizer
system for highly reproducible aerosol delivery. J Aerosol
Med Dep Clear Eff Lung 7:229–237.
56. Ijaz MK, Karim YG, Sattar SA, Johnson-Lussenburg
CM. 1987. Development of methods to study the survival
of airborne viruses. J Virol Meth 18:87–106.
57. Ijaz MK, Alharbi S, Uduman S, Cheema Y, Sheek-Hus-
sen MM, Alkhair AR, Shalabi AG, Ijaz MK, Bin-Othman
SA, Sattar SA, Liddle LF. 1994. Seasonality and preva-
lence of rotavirus in Al-Ain, United Arab Emirates. Clin
Diagn Virol 2:323–329.
58. Goldberg LJ, Watkins HM, Boerke EE, Chatigny MA.
1958. The use of a rotating drum for the study of aerosols
over extended periods of time. Am J Hyg 68:85–93.
59. Kournikakis B, Simpson M, Netolitzky D. 1988. Photo-
inactivation of Newcastle disease virus in aerosol and in sol-
ution. Defence Research Establishment Suffield, Ralston,
AB.
60. May KR, Druett HA. 1968. A microthread technique for
studying the viability of microbes in a simulated airborne
state. J Gen Microbiol 51:353–366.
61. Wallis C, Melnick JL, Rao VC, Sox TE. 1985. Method for
detecting viruses in aerosols. Appl Environ Microbiol 50:
1181–1186.
62. May KR. 1966. Multistage liquid impinger. Bacteriol Rev
30:559–570.
63. Zaki AM, van Boheemen S, Bestebroer TM, Osterhaus
AD, Fouchier RA. 2012. Isolation of a novel coronavirus
from a man with pneumonia in Saudi Arabia. N Engl J
Med 367:1814–1820.
64. Lippmann M. 1961. A compact cascade impactor for field
survey sampling. Am Ind Hyg Assoc J 22:348–353.
65. Fabian P, McDevitt JJ, Houseman EA, Milton DK. 2009.
Airborne influenza virus detection with four aerosol sam-
plers using molecular and infectivity assays: considerations
for a new infectious virus aerosol sampler. Indoor Air 19:
433–441.
66. McKissick GE, Griesemer RA, Farrell RL. 1970. Aerosol
transmission of Rauscher murine leukemia virus. J Natl
Cancer Inst 45:625–636.
67. Hinshaw VS, Schaffer FL, Chatigny MA. 1976. Evalua-
tion of Moloney murine sarcoma and leukemia-virus com-
plex as a model for airborne oncogenic virus biohazards—
survival of airborne virus and exposure of mice. J Natl Can-
cer Inst 57:775–778.
68. Davis GW, Grieseme RA, Shadduck JA, Farrell RL.
1971. Effect of relative humidity on dynamic aerosols of
adenovirus-12. Appl Microbiol 21:676–679.
69. Couch RB, Cate TR, Douglas RG Jr., Gerone PJ, Knight
V. 1966. Effect of route of inoculation on experimental res-
piratory viral disease in volunteers and evidence for air-
borne transmission. Bacteriol Rev 30:517–529.
70. Jericho KWF, Kozub GC. 2004. Experimental infectious
respiratory disease in groups of calves: lobar distribution,
variance, and sample-size requirements for vaccine evalua-
tion. Can J Vet Res 68:118–127.
71. van der Veen J, Poort Y, Birchfield DJ. 1972. Effect
of relative humidity on experimental transmission of
Sendai virus in mice. 1. Proc Soc Exp Biol Med 140:
1437–1440.
72. Calnek BW, Hitchner SB. 1973. Survival and disinfection
of Mareks-disease virus and effectiveness of filters in pre-
venting airborne dissemination. Poult Sci 52:35–43.
73. Abdul-Careem MF, Javaheri-Vayeghan A, Shanmugana-
than S, Haghighi HR, Read LR, Haq K, Hunter DB,
Schat KA, Heidari M, Sharif S. 2009. Establishment of
an aerosol-based Marek’s disease virus infection model.
Avian Dis 53:387–391.
74. Kim SJ, Spradbrow PB. 1978. Administration of a vac-
cine prepared from the Australian V4 strain of Newcastle
disease virus by aerosol and drinking water. Aust Vet J 54:
486–489.
75. Hugh-Jones M, Allan WH, Dark FA, Harper GJ. 1973.
The evidence for the airborne spread of Newcastle disease.
J Hyg (Lond) 71:325–339.
76. Li X, Chai T, Wang Z, Song C, Cao H, Liu J, Zhang X,
Wang W, Yao M, Miao Z. 2009. Occurrence and transmis-
sion of Newcastle disease virus aerosol originating from
infected chickens under experimental conditions. Vet
Microbiol 136:226–232.
77. Fairchild GA. 1974. Ozone effect on respiratory deposition
of vesicular stomatitis virus aerosols. Am Rev Respir Dis
109:446–451.
78. Frankova V, Rychterova V. 1975. Inhalatory infection of
mice with influenza A0/PR8 virus. II. Detection of the virus
in the blood and extrapulmonary organs. Acta Virol 19:
35–40.
79. Maigetter RZ, Ehrlich R, Fenters JD, Gardner DE. 1976.
Potentiating effects of manganese-dioxide on experimental
respiratory infections. Environ Res 11:386–391.
80. Owens SL, Osebold JW, Zee YC. 1981. Dynamics of
B-lymphocytes in the lungs of mice exposed to aerosolized
influenza virus. Infect Immun 33:231–238.
81. Johansson BE, Kilbourne ED. 1991. Comparison of intra-
nasal and aerosol infection of mice in assessment of im-
munity to influenza-virus infection. J Virol Methods 35:
109–114.
82. Mumford JA, Jessett DM, Rollinson EA, Hannant D,
Draper ME. 1994. Duration of protective efficacy of equine
influenza immunostimulating complex/tetanus vaccines.
Vet Rec 134:158–162.
83. Gruber WC, Hinson HP, Holland KL, Thompson JM,
Reed GW, Wright PF. 1993. Comparative trial of large-
particle aerosol and nose drop administration of live attenu-
ated influenza vaccines. J Infect Dis 168:1282–1285.
84. Paillot R, Hannant D, Kydd JH, Daly JM. 2006. Vaccina-
tion against equine influenza: quid novi? Vaccine 24:4047–
4061.
85. Bowen LE, Rivers K, Trombley JE, Bohannon JK, Li SX,
Boydston JA, Eichelberger MC. 2012. Development of a
murine nose-only inhalation model of influenza: compari-
son of disease caused by instilled and inhaled A/PR/8/34.
Front Cell Infect Microbiol 2:74.
86. Lednicky JA, Hamilton SB, Tuttle RS, Sosna WA, Dan-
iels DE, Swayne DE. 2010. Ferrets develop fatal influenza
after inhaling small particle aerosols of highly pathogenic
avian influenza virus A/Vietnam/1203/2004 (H5N1). Virol
J 7:231.
87. Tuttle RS, Sosna WA, Daniels DE, Hamilton SB,
Lednicky JA. 2010. Design, assembly, and validation of
a nose-only inhalation exposure system for studies of
aerosolized viable influenza H5N1 virus in ferrets. Virol J
7:135.
88. Frank GH, Marshall RG, Smith PC. 1977. Clinical and
immunologic responses of cattle to infectious bovine

rhinotracheitis virus after infection by viral aerosol or intra-
muscular inoculation. Am J Vet Res 38:1497–1502.
89. Belknap EB, Collins JK, Ayers VK, Schultheiss PC.
1994. Experimental-infection of neonatal calves with neu-
rovirulent bovine herpesvirus type-1.3. Vet Path 31:
358–365.
90. Mars MH, Bruschke CJ, van Oirschot TJ. 1999. Airborne
transmission of BHV1, BRSV, and BVDV among cattle is
possible under experimental conditions. Vet Microbiol 66:
197–207.
91. Elazhary MA, Galina M, Roy RS, Fontaine M, Lamothe
P. 1980. Experimental infection of calves with bovine res-
piratory syncytial virus (Quebec strain). Can J Comp Med
44:390–395.
92. Sacco RE, Nonnecke BJ, Palmer MV, Waters WR, Lipp-
olis JD, Reinhardt TA. 2012. Differential expression of
cytokines in response to respiratory syncytial virus infec-
tion of calves with high or low circulating 25-hydroxyvita-
min D3. PLoS One 7:e33074. doi:10.1371/journal.
pone.0033074.
93. Wardley RC, Povey RC. 1977. Aerosol transmission of
feline calciciviruses. An assessment of its epidemiological
importance. Brit Vet J 133:504–508.
94. Omerod E, McClandlish IA, Jarrett O. 1979. Disease pro-
duced by feline caliciviruses. Vet Rec 104:65–69.
95. Wilkinson PJ, Donaldson AI. 1977. Transmission studies
with African swine fever virus. The early distribution of
virus in pigs infected by airborne virus. J Comp Pathol 87:
497–501.
96. Wilkinson PJ, Donaldson AI, Greig A, Bruce W.
1977. Transmission studies with African swine fever virus.
Infections of pigs by airborne virus. J Comp Pathol 87:487–
495.
97. Hyslop NS. 1979. Observations on the survival and infec-
tivity of airborne Rinderpest virus. Int J Biometeorol 23:1–7.
98. Larson EW, Dominik JW, Slone TW. 1980. Aerosol
stability and respiratory infectivity of Japanese B-encepha-
litis virus. Infect Immun 30:397–401.
99. Merekalova ZI, Orekhova NM, Zharona EI. 1981. Possi-
ble ways of excreting murine leukemia virus into the envi-
ronment and the so-called input gate for infection. Exper
Oncol 3:59–62.
100. Arzt J, Pacheco JM, Rodriguez LL. 2010. The early
pathogenesis of foot-and-mouth disease in cattle after aero-
sol inoculation. Identification of the nasopharynx as the
primary site of infection. Vet Pathol 47:1048–1063.
101. McVicar JW, Eisner RJ. 1983. Aerosol exposure of cat-
tle to foot-and-mouth disease virus. J Hyg (Lond) 91:
319–328.
102. Brown CC, Meye RF, Olander HJ, House C, Mebus CA.
1992. A pathogenesis study of foot-and-mouth-disease
in cattle, using in situ hybridization. Can J Vet Res 56:
189–193.
103. Donaldson AI, Wardley RC, Martin S, Ferris NP. 1983.
Experimental Aujeszky’s disease in pigs: excretion, survival
and transmission of the virus. Vet Rec 113:490–494.
104. Gillespie RR, Hill MA, Kanitz CL, Knox KE, Clark LK,
Robinson JP. 2000. Infection of pigs by Aujeszky’s disease
virus via the breath of intranasally inoculated pigs. Res Vet
Sci 68:217–222.
105. Gillespie RR, Hill MA, Kanitz CL. 1996. Infection of pigs
by aerosols of Aujeszky’s disease virus and their shedding of
the virus. Res Vet Sci 60:228–233.
106. Ratanasethakul C, Cumming RB. 1983. The effect of
route of infection and strain of virus on the pathology of
Australian infectious bronchitis. Aust Vet J 60:209–213.
107. Stephenson EH, Larson EW, Dominik JW. 1984. Effect of
environmental factors on aerosol-induced Lassa virus infec-
tion. J Med Virol 14:295–303.
3.2.7. Airborne Viruses ▪ 3.2.7-21
108. Prince DS, Astry C, Vonderfecht S, Jakab G, Shen FM,
Yolken RH. 1986. Aerosol transmission of experimental
rotavirus infection. Pediatr Infect Dis 5:218–222.
109. Anderson GW, Lee JO, Anderson AO, Powell N, Mangi-
afico JA, Meadors G. 1991. Efficacy of a Rift-Valley fever
virus-vaccine against an aerosol infection in rats. Vaccine 9:
710–714.
110. Morrill JC, Peters CJ. 2011. Protection of MP-12-vacci-
nated rhesus macaques against parenteral and aerosol chal-
lenge with virulent rift valley fever virus. J Infect Dis 204:
229–236.
111. Bhardwaj N, Pierce BR, Ross TM. 2012. Immunization
with DNA vaccine expressing Gn coupled to C3d prevents
clinical symptoms of infection and protects mice against an
aerosol Rift Valley fever virus infection. J Bioterr Biodefense
2:S3:006.
112. Nuzum EO, Rossi CA, Stephenson EH, Leduc JW. 1988.
Aerosol transmission of Hantaan and related viruses to lab-
oratory rats. Am J Trop Med Hyg 38:636–640.
113. McElroy AK, Smith JM, Hooper JW, Schmaljohn CS.
2004. Andes virus M genome segment is not sufficient to
confer the virulence associated with Andes virus in Syrian
hamsters. Virol 326:130–139.
114. Deuter A, Southee DJ, Mockett APA. 1991. Fowlpox
virus—pathogenicity and vaccination of day-old chickens
via the aerosol route. Res Vet Sci 50:362–364.
115. Pokhodiaev VA, Gonchar NI, Pshenichnov VA. 1991.
An experimental study of the contact transmission of the
Marburg virus. Vopr Virusol 36:506–508.
116. Geisbert TW, Daddario-Dicaprio KM, Geisbert JB, Reed
DS, Feldmann F, Grolla A, Stroher U, Fritz EA, Hensley
LE, Jones SM, Feldmann H. 2008. Vesicular stomatitis
virus-based vaccines protect nonhuman primates against
aerosol challenge with Ebola and Marburg viruses. Vaccine
26:6894–6900.
117. Alves DA, Glynn AR, Steele KE, Lackemeyer MG,
Garza NL, Buck JG, Mech C, Reed DS. 2010. Aerosol
exposure to the Angola strain of Marburg virus causes lethal
viral hemorrhagic fever in cynomolgus macaques. Vet
Pathol 47:831–851.
118. Goodman LB, Wagner B, Flaminio MJ, Sussman KH,
Metzger SM, Holland R, Osterrieder N. 2006. Compari-
son of the efficacy of inactivated combination and modi-
fied-live virus vaccines against challenge infection with
neuropathogenic equine herpesvirus type 1 (EHV-1). Vac-
cine 24:3636–3645.
119. Kenyon RH, Mckee KT, Zack PM, Rippy MK, Vogel AP,
York C, Meegan J, Crabbs C, Peters CJ. 1992. Aerosol
infection of rhesus macaques with Junin virus. Intervirol
33:23–31.
120. Smith KC, Whitwell KE, Binns MM, Dolby CA, Han-
nant D, Mumford JA. 1992. Abortion of virologically neg-
ative foetuses following experimental challenge of pregnant
pony mares with equid herpesvirus 1. Equine Vet J 24:
256–259.
121. Brim TA, Vancott JL, Lunney JK, Saif LJ. 1994. Lympho-
cyte-proliferation responses of pigs inoculated with trans-
missible gastroenteritis virus or porcine respiratory
coronavirus. Am J Vet Res 55:494–501.
122. Pratt WD, Gibbs P, Pitt ML, Schmaljohn AL. 1998. Use
of telemetry to assess vaccine-induced protection against
parenteral and aerosol infections of Venezuelan equine
encephalitis virus in non-human primates. Vaccine 16:
1056–1064.
123. Phillpotts RJ, O’brien L, Appleton RE, Carr S, Bennett
A. 2005. Intranasal immunisation with defective adeno-
virus serotype 5 expressing the Venezuelan equine ence-
phalitis virus E2 glycoprotein protects against airborne
challenge with virulent virus. Vaccine 23:1615–1623.
3.2.7-22 ▪ AIR
124. Taus NS, Traul DL, Oaks JL, Crawford TB, Lewis GS,
Li H. 2005. Experimental infection of sheep with ovine
herpesvirus 2 via aerosolization of nasal secretions. J Gen
Virol 86:575–579.
125. Li H, O’Toole D, Ki O, Oaks JL, Crawford TB. 2005.
Malignant catarrhal fever-like disease in sheep after intra-
nasal inoculation with ovine herpesvirus-2. J Vet Diagn
Invest 17:171–175.
126. Taus NS, Oaks JL, Gailbreath K, Traul DL, O’Toole D,
Li H. 2006. Experimental aerosol infection of cattle (Bos
taurus) with ovine herpesvirus 2 using nasal secretions
from infected sheep. Vet Microbiol 116:29–36.
127. Gailbreath KL, Taus NS, Cunha CW, Knowles DP, Li H.
2008. Experimental infection of rabbits with ovine herpes-
virus 2 from sheep nasal secretions. Vet Microbiol 132:
65–73.
128. Li H, Cunha CW, Davies CJ, Gailbreath KL, Knowles
DP, Oaks JL, Taus NS. 2008. Ovine herpesvirus 2 repli-
cates initially in the lung of experimentally infected sheep.
J Gen Virol 89:1699–1708.
129. Li H, Brooking A, Cunha CW, Highland MA, O’Toole
D, Knowles DP, Taus NS. 2012. Experimental induc-
tion of malignant catarrhal fever in pigs with ovine herpes-
virus 2 by intranasal nebulization. Vet Microbiol 159:
485–489.
130. Jahrling PB, Hensley LE, Martinez MJ, Leduc JW,
Rubins KH, Relman DA, Huggins JW. 2004. Exploring
the potential of variola virus infection of cynomolgus mac-
aques as a model for human smallpox. Proc Nat Acad Sci
101:15196–15200.
131. Nalca A, Nichols DK. 2011. Rabbitpox: a model of air-
borne transmission of smallpox. J Gen Virol 92:31–35.
132. Dee SA, Deen J, Jacobson L, Rossow KD, Mahlum C,
Pijoan C. 2005. Laboratory model to evaluate the role of
aerosols in the transport of porcine reproductive and respi-
ratory syndrome virus. Vet Rec 156:501–504.
133. Cho JG, Deen J, Dee SA. 2007. Influence of isolate path-
ogenicity on the aerosol transmission of porcine reproduc-
tive and respiratory syndrome virus. Can J Vet Res 71:
23–27.
134. Henderson DW. 1952. An apparatus for the study of air-
borne infection. J Hyg (Lond) 50:53–68.
135. Osborne AD, Saunders JR, Sebunya TK, Willson P,
Green GH. 1985. A simple aerosol chamber for experi-
mental reproduction of respiratory disease in pigs and other
species. Can J Comp Med 49:434–435.
136. Chung IP, Dunnrankin D, Phalen RF, Oldham MJ. 1992.
Low-cost wind-tunnel for aerosol inhalation studies. Am
Industl Hyg Assoc J 53:232–236.
137. Cole EC, Cook CE. 1998. Characterization of infectious
aerosols in health care facilities: an aid to effective engi-
neering controls and preventive strategies. Am J Infect
Contrl 26:453–464.
138. Pike RM. 1979. Laboratory-associated infections: in-
cidence, fatalities, causes, and prevention. Annu Rev Micro-
biol 33:41–66.
139. Biosafety in Microbiological and Biomedical Laboratories
2009. 5th ed. U. S. Government Printing Office, Wash-
ington, DC. http://www.cdc.gov/biosafety/publications/
bmbl5/.
140. Benbough JE. 1971. Some factors affecting the survival of
airborne viruses. J Gen Virol 10:209–220.
141. Yang W, Marr CL. 2012. Mechanisms by which ambient
humidity may affect viruses in aerosols. Appl Environ Micro-
biol 78:6781–6788.
142. Sawchuk WS, Weber PJ, Lowy DR, Dzubow LM. 1989.
Infectious papillomavirus in the vapor of warts treated with
carbon dioxide laser or electrocoagulation: detection and
protection. J Am Acad Dermatol 21:41–49.
143. Hepburn NC, Brooks TJG. 1991. An outbreak of chick-
enpox in a military field hospital—the implications for bio-
logical warfare. J Roy Soc Med 84:721–722.
144. Suzuki K, Yoshikawa T, Ihira M, Ohashi M, Suga S,
Asano Y. 2003. Spread of varicella-zoster virus DNA to
the environment from varicella patients who were treated
with oral acyclovir. Pediatr Int 45:458–460.
145. Caul EO. 1994. Small round-structured airuses—airborne
transmission and hospital control. Lancet 343:1240–1242.
146. Gellert GA, Waterman SH, Ewert D, Oshiro L, Giles
MP, Monroe SS, Gorelkin L, Glass RI. 1990. An outbreak
of acute gastroenteritis caused by a small round structured
virus in a geriatric convalescent facility. Infect Contrl
Hosp Epidemiol 11:459–464.
147. Chadwick PR, Walker M, Rees AE. 1994. Airborne
transmission of a small round-structured virus. Lancet
343:171.
148. Saif LJ, Theil KW. 1990. Viral diarrheas of man and animals.
CRC Press, Boca Raton, FL.
149. Ijaz MK, Sattar SA, Johnson-Llussenburg CM, Spring-
thorpe VS. 1985. Comparison of the airborne survival of
calf rotavirus and poliovirus type-1 (Sabin) aerosolized as
a mixture. Appl Environ Microbiol 49:289–293.
150. Patel MM, Pitzer VE, Alonso WJ, Vera D, Lopman B,
Tate J, Viboud C, Parashar UD. 2013. Global seasonality
of rotavirus disease. Pediatr Infect Dis J 32:e134–e147. doi:
10.1097/INF.0b013e31827d3b68.
151. Dick EC, Jennings LC, Mink KA, Wartgow CD, Inhorn
SL. 1987. Aerosol transmission of rhinovirus colds. J Infect
Dis 156:442–448.
152. Gwaltney JM. 1997. Rhinoviruses, p 815–838. In Evans
AS, Kaslow RA (eds), Viral infections of humans: epidemiol-
ogy and control, 4th ed. Plenum Medical, New York.
153. Karim YG, Ijaz MK, Sattar SA, Johnson-Lussenburg
CM. 1985. Effect of relative humidity on the airborne
survival of rhinovirus-14. Can J Microbiol 31:1058–
1061.
154. Calero L, Brusis T. 2003. Laryngeal papillomatosis—first
recognition in Germany as an occupational disease in
an operating room nurse. Laryngo-Rhino-Otolo 82:790–
793.
155. Hallmo P, Naess O. 1991. Laryngeal papillomatosis with
human papillomavirus DNA contracted by a laser surgeon.
Eur Arch Otorhinolaryngol 248:425–427.
156. Manson LT, Damrose EJ. 2013. Does exposure to laser
plume place the surgeon at high risk for acquiring clinical
human papillomavirus infection? Laryngoscope 123:
1319–1320.
157. Kashima HK, Kessis T, Mounts P, Shah K. 1991.
Polymerase chain-reaction identification of human papil-
lomavirus DNA in CO2-laser plume from recurrent respira-
tory papillomatosis. Otolaryngol Head Neck Surg 104:
191–195.
158. Wisniewski PM, Warhol MJ, Rando RF, Sedlacek TV,
Kemp JE, Fisher JC. 1990. Studies on the transmission
of viral disease via the CO2-laser plume and ejecta.
J Reprod Med 35:1117–1123.
159. Bergbrant IM, Samuelsson L, Olofsson S, Jonassen F,
Ricksten A. 1994. Polymerase chain-reaction for monitor-
ing human papillomavirus contamination of medical per-
sonnel during treatment of genital warts with CO2-laser
and electrocoagulation. Acta Dermato-Venereolog 74:
393–395.
160. Barrett WL, Garber SM. 2003. Surgical smoke—a review
of the literature—is this just a lot of hot air? Surg Endosc
Other Intervent Tech 17:979–987.
161. Johnson GK, Robinson WS. 1991. Human immunodefi-
ciency virus-1 (HIV-1) in the vapors of surgical power
instruments. J Med Virol 33:47–50.

162. Jewett DL, Heinsohn P, Bennett C, Rosen A, Neuilly C.
1992. Blood-containing aerosols generated by surgical
techniques—a possible infectious hazard. Am Indust Hyg
Assoc J 53:228–231.
163. Baggish MS, Poiesz BJ, Joret D, Williamson P, Refai A.
1991. Presence of human-immunodeficiency-virus DNA
in laser smoke. Lasers Surg Med 11:197–203.
164. Leffel EK, Reed DS. 2004. Marburg and Ebola viruses as
aerosol threats. Biosec Bioterr Biodef Strat Pract Sci 2:
186–191.
165. Dalgard DW, Hardy RJ, Pearson SL, Pucak GJ, Quander
RV, Zack PM, Peters CJ, Jahrling PB. 1992. Combined
simian hemorrhagic-fever and Ebola virus-infection in cyn-
omolgus monkeys. Lab Animal Sci 42:152–157.
166. Leduc JW. 1989. Epidemiology of hemorrhagic fever
viruses. Rev Infect Dis 11:S730–S735.
167. Alexeyev OA, Baranov BA. 1993. Puumala virus infection
without signs of renal involvement. Scand J Infect Dis 25:
525–527.
168. Lee HW, Johnson KM. 1982. Laboratory-acquired infec-
tions with Hantaan virus, the etiologic agent of Korean
hemorrhagic fever. J Infect Dis 146:645–651.
169. Niklasson BS. 1992. Haemorrhagic fever with renal syn-
drome, virological and epidemiological aspects. Pediatr
Nephrol 6:201–204.
170. Peters CJ, Khan AS. 2002. Hantavirus pulmonary syn-
drome: the new American hemorrhagic fever. Clin Infect
Dis 34:1224–1231.
171. Wong TW, Chan YC, Yap EH, Joo YG, Lee HW, Lee
PW, Yanagihara R, Gibbs CJ Jr., Gajdusek DC. 1988.
Serological evidence of hantavirus infection in laboratory
rats and personnel. Int J Epidemiol 17:887–890.
172. Keeler N, Lingappa L. 2004. Severe acute respiratory
syndrome: public health response and clinical practice
update for an emerging disease. Curr Opin Pediatr 16:
61–69.
173. Ng SK. 2003. Possible role of an animal vector in the
SARS outbreak at Amoy Gardens. Lancet 362:570–572.
174. Somogyi R, Vesely AE, Azami T, Preiss D, Fisher J, Cor-
reia J, Fowler RA. 2004. Dispersal of respiratory droplets
with open vs closed oxygen delivery masks: implications
for the transmission of severe acute respiratory syndrome.
Chest 125:1155–1157.
175. Ijaz MK, Brunner AH, Sattar SA, Nair RC, Johnson-
Lussenburg CM. 1985. Survival characteristics of airborne
human coronavirus-229E. J Gen Virol 66:2743–2748.
176. Saw TA, Lim W, Shortridge K, Tam J, Liu KK, Mak KH,
Tsang T, Ho YY, Lee FY, Kwong H. 1998. Isolation of
avian influenza A (H5N1) viruses from humans—Hong
Kong, May–December 1997. J Am Med Assoc 279:
263–264.
177. Imperiale MJ, Hanna MGIII 2012. Biosafety considera-
tions of mammalian-transmissible H5N1 influenza. MBio
3:e00043–12. doi:10.1128/mBio.00043-12.
178. Moutou F, Durand B. 1994. Modelling the spread of foot-
and-mouth disease virus. Vet Res 25:279–285.
179. Kandeil A, El-Shesheny R, Kayali G, Moatasim Y, Bagato
O, Darwish M, Gaffar A, Younes A, Farag T, Kutkat MA,
Ali MA. 2013. Characterization of the recent outbreak of
foot-and-mouth disease virus serotype SAT2 in Egypt.
Arch Virol 158:619–627.
180. Park JH, Lee KN, Ko YJ, Kim SM, Lee HS, Park JY, Yeh
JY, Kim MJ, Lee YH, Sohn HJ, Cho IS, Kim B. 2013.
Diagnosis and control measures of the 2010 outbreak of
foot-and-mouth disease A type in the Republic of Korea.
Transbound Emerg Dis 60:188–192.
181. Gibson CF, Donaldson AI. 1986. Exposure of sheep to
natural aerosols of foot-and-mouth disease virus. Res Vet
Sci 41:45–49.
3.2.7. Airborne Viruses ▪ 3.2.7-23
182. Hopkins SR, Drury LN. 1971. Efficacy of air filters in pre-
venting transmission of Newcastle Disease. Avian Dis 15:
596–603.
183. Kournikakis B, Netolitzky D, Fildes J. 1987. Effects of
temperature and relative humidity on the survival of
Newcastle Disease Virus aerosols in the rotating drum
(U). (1261). Defence Research Establishment Suffield,
Ralston, AB.
184. Burge HA. 1995. Airborne contagious disease, p 25–47. In
Burge HA (ed), Bioaerosols. CRC Press, Boca Raton, FL.
185. Mitchell BW, King DJ. 1994. Effect of negative air ioniza-
tion on airborne transmission of Newcastle disease virus.
Avian Dis 38:725–732.
186. Schoenbaum MA, Zimmerman JJ, Beran GW, Murphy
DP. 1990. Survival of pseudorabies virus in aerosol.
Am J Vet Res 51:331–333.
187. Cramer SD, Campbell GA, Njaa BL, Morgan SE, Smith
SK, McLin WR, Brodersen BW, Wise AG, Scherba G,
ILangohr IM, Maes RK. 2011. Pseudorabies virus infec-
tion in Oklahoma hunting dogs. J Vet Diagn Invest 23:
915–923.
188. Winkler WG. 1968. Airborne rabies virus isolation. Wildl
Dis 4:37–40.
189. Winkler WG, Fashinell TR, Leffingwell L, Howard P,
Conomy P. 1973. Airborne rabies transmission in a labora-
tory worker. J Am Med Assoc 226:1219–1221.
190. Sattar SA. 2004. Microbicides and the environmental con-
trol of nosocomial viral infections. J Hosp Infect 56(Suppl
2):S64–S69.
191. Sattar SA. 2007. Hierarchy of susceptibility of viruses to
environmental surface disinfectants: a predictor of activity
against new and emerging viral pathogens. J AOAC Int 90:
1655–1658.
192. Griffiths WD, Bennett A, Speight S, Parks S. 2005.
Determining the performance of a commercial air purifica-
tion system for reducing airborne contamination using
model micro-organisms: a new test methodology. J Hosp
Infect 61:242–247.
193. Cutler TD, Wang C, Hoff SJ, Zimmerman JJ. 2012.
Effect of temperature and relative humidity on ultraviolet
(UV 254) inactivation of airborne porcine respiratory
and reproductive syndrome virus. Vet Microbiol 159:47–52.
194. McDevitt JJ, Rudnick SN, Radonovich LJ. 2012. Aerosol
susceptibility of influenza virus to UV-C light. Appl Environ
Microbiol 78:1666–1669.
195. McDevitt JJ, Milton DK, Rudnick SN, First MW. 2008.
Inactivation of poxviruses by upper-room UVC light in a
simulated hospital room environment. PLoS One 3:
e3186. doi:10.1371/journal.pone.0003186.
196. McDevitt JJ, Lai KM, Rudnick SN, Houseman EA, First
MW, Milton DK. 2007. Characterization of UVC light
sensitivity of vaccinia virus. Appl Environ Microbiol 73:
5760–5766.
197. Walker CM, Ko G. 2007. Effect of ultraviolet germicidal
irradiation on viral aerosols. Environ Sci Tech 41:5460–
5465.
198. Tseng CC, Li CS. 2005. Inactivation of virus-containing
aerosols by ultraviolet germicidal irradiation. Aerosol Sci
Tech 39:1136–1142.
199. Tseng CC, Li CS. 2006. Ozone for inactivation of aerosol-
ized bacteriophages. Aerosol Sci Tech 40:683–689.
200. Terrier O, Essere B, Yver M, Barthelemy M, Bouscam-
bert-Duchamp M, Kurtz P, VanMechelen D, Morfin F,
Billaud G, Ferraris O, Lina B, Rosa-Calatrava M, Moules
V. 2009. Cold oxygen plasma technology efficiency against
different airborne respiratory viruses. J Clin Virol 45:
119–124.
201. Wigginton KR, Pecson BM, Sigstam T, Bosshard F, Kohn
T. 2012. Virus inactivation mechanisms: impact of
3.2.7-24 ▪ AIR
disinfectants on virus function and structural integrity.
Environ Sci Technol 46:12069–12078.
202. Fan L, Song J, Hildebrand PD, Forney CF. 2002. Interac-
tion of ozone and negative air ions to control micro-organ-
isms. J Appl Microbiol 93:144–148.
203. Ishizaki K, Shinriki N, Matsuyama H. 1986. Inactivation
of Bacillus spores by gaseous ozone. J Appl Bacteriol 60:
67–72.
204. Kowalski WJ, Bahnfleth WP, Striebig BA, Whittam TS.
2003. Demonstration of a hermetic airborne ozone disin-
fection system: studies on E. coli. Am Indust Hyg Assoc J
64:222–227.
205. Kekez MM, Sattar SA. 1997. A new ozone-based method
for virus inactivation: preliminary study. Phys Med Biol 42:
2027–2039.
206. Tseng C, Li C. 2008. Inactivation of surface viruses by gas-
eous ozone. J Environ Hlth 70:56–62.
207. Tanaka H, Sakurai M, Ishii K, Matsuzawa Y. 2009. Inac-
tivation of influenza virus by ozone gas. IHI Engr Rev 42:
108–111.
208. Pottage T, Macken S, Walker JT, Bennett AM. 2012.
Meticillin-resistant Staphylococcus aureus is more resistant
to vaporized hydrogen peroxide than commercial Geobacil-
lus stearothermophilus biological indicators. J Hosp Infect 80:
41–45.
209. Fichet G, Comoy E, Duval C, Antloga K, Dehen C, Char-
bonnier A, McDonnell G, Brown P, Lasmezas CI, Deslys
JP. 2004. Novel methods for disinfection of prion-contami-
nated medical devices. Lancet 364:521–526.
210. Nelson CL, Berger JT. 1989. Inactivation of micro-
organisms by oxygen gas plasma. Curr Microbiol 18:
275–276.
211. Cutler TD, Zimmerman JJ. 2011. Ultraviolet irradiation
and the mechanisms underlying its inactivation of infec-
tious agents. Anim Hlth Res Rev 12:15–23.
212. Tseng CC, Li CS. 2007. Inactivation of viruses on surfaces
by ultraviolet germicidal irradiation. J Occup Environ Hyg
4:400–405.
213. Sagripanti JL, Voss L, Marschall HJ, Lytle CD. 2013.
Inactivation of vaccinia virus by natural sunlight and by
artificial UVB radiation. Photochem Photobiol 89:132–138.
214. Mattle MJ, Kohn T. 2012. Inactivation and tailing during
UV254 disinfection of viruses: contributions of viral aggre-
gation, light shielding within viral aggregates, and recombi-
nation. Environ Sci Technol 46:10022–10030.
215. Korppi M. 1988. Viruses and airborne allergens as precip-
itants of obstructive respiratory difficulties in children.
Ann Clin Res 20:417–422.
216. Quintiliani R. 1994. The epidemiology of mild, moderate
and severe asthma. Adv Therapy 11:70–77.
217. Edwards DA, Man JC, Brand P, Katstra JP, Sommerer K,
Stone HA, Nardell E, Scheuch G. 2004. Inhaling to mit-
igate exhaled bioaerosols. Proc Natl Acad Sci 101:17383–
17388.
218. Wenzel RP. 1996. Airline travel and infection. N Engl J
Med 334:981–982.
219. Goodman LB, Wimer C, Dubovi EJ, Gold C, Wagner B.
2012. Immunological correlates of vaccination and infec-
tion for equine herpesvirus. Clin Vaccine Immunol 19:
235–241.
220. Saksena NK, Dwyer D, Barré-Sinoussi F. 1991. A “senti-
nel” technique for monitoring viral aerosol contamination.
J Infect Dis 164:1021–1022.
221. Cukier J, Price MF, Gentry LO. 1989. Suction lipo-
plasty—biohazardous aerosols and exhaust mist—the
clouded issue. Plastic Reconst Surg 83:494–497.
222. Weber A, Willeke K, Marchioni R, Myojo T, McKay R,
Donnelly J, Liebhaber F. 1993. Aerosol penetration and
leakage characteristics of masks used in the health care
industry. Am J Infect Contrl 21:167–173.
223. Roth Y, Chapnik JS, Cole P. 2003. Feasibility of aerosol
vaccination in humans. Ann Otol Rhinol Laryngol 112:
264–270.
224. Banks M. 1993. DNA restriction-fragment-length-poly-
morphism among British isolates of Aujeszkys disease
virus—use of the polymerase chain-reaction to discrimi-
nate among strains. Brit Vet J 102:155–163.
225. Ho MS, Glass RI, Pinsky PF, Anderson LJ. 1988. Rota-
virus as a cause of diarrheal morbidity and mortality in
the United States. J Infect Dis 158:1112–1116.
226. Skerrett SJ. 1994. Host defenses against respiratory infec-
tion. Med Clin N Am 78:941–966.
227. Alibek K, Handleman S. 1999. Biohazard. Random House,
New York.
228. Madjid M, Lillibridge S, Mirhaji P, Casscells W.
2003. Influenza as a bioweapon. J Roy Soc Med 96:
345–346.
229. Sidwell RW, Smee DF. 2003. Viruses of the Bunya- and
Togaviridae families: potential as bioterrorism agents and
means of control. Antiviral Res 57:101–111.
 Aerobiology of Agricultural Pathogens
ESTELLE LEVETIN
3.2.8
All crop plants are subject to diseases caused by viruses, bac-
teria, oomycetes, and fungi. Although plant diseases had
been recognized since ancient times, they were not connected
with pathogenic microorganisms until the 19th century. In
1861, during the investigations of the Irish potato blight,
the German botanist Anton de Bary proved experimentally
that Phytophthora infestans, previously described as present
on the infected potatoes, was actually the cause of the disease.
This discovery gave birth to the science of plant pathology.
Soon various plant pathogenic fungi were described; how-
ever, pathogenic bacteria and viruses were not identified until
later in the 19th century. Even with the many advances in
agriculture during the 20th century, plant diseases still pose
a major threat to the world’s crops in the 21st century, making
the world’s food supply highly vulnerable.
Studying the spread of plant disease within a crop and from
field to field within a region is often referred to as “botanical
epidemiology.” Dispersal of the pathogen is the necessary step
for repeated cycles of infection and multiplication and, there-
fore, the spread of the epidemic. A thorough understanding of
dispersal is necessary for predicting the onset and severity of
the disease. Implicit within the scope of dispersal are the proc-
esses of release, transport, and deposition with each of these
steps influenced by the environment (1–4). Although prop-
agules of plant pathogens are dispersed by wind, rain, soil
water, insects, and even humans, this chapter is focused
on the discussion of the airborne spread of those pathogens.
Airborne transport may be the least understood vehicle of
dispersal.
RANGE OF AIRBORNE TRANSPORT
Plant pathogens dispersed by airborne transport generally
produce enormous numbers of propagules that are passively
carried through the atmosphere. The spread of plant dis-
ease through this aerobiological pathway can proceed over
short distances through focal spread as well as over long dis-
tances (5).
An infection focus is an area of a crop with a contagious
disease. Foci are often circular; however, if they are strongly
affected by wind, they may become comet- or V-shaped.
Foci generally have a constant radial expansion, with the
rate varying with the scale of the infection from a few centi-
meters per day for a localized infection to hundreds of
doi:10.1128/9781555818821.ch3.2.8
3.2.8-1
kilometers per year for a pandemic (5, 6). The spread of the
disease can be described as microscale, mesoscale, or synoptic
scale using meteorological terms that describe the horizontal
movement of atmospheric particles.
Disease Spread
Microscale spread is usually limited to less than a few hundred
meters within one field and occurs within one growing sea-
son. This roughly corresponds to a zero-order epidemic (5,
7). The focus begins with a single successful propagule caus-
ing an infection and creating a lesion. After several genera-
tions of localized pathogen spread for polycyclic diseases,
which produce many generations of inocula and many cycles
of infection during a single growing season, the focus may
reach a detectable size. In an annual crop, these are often
about 1 m in diameter around the initial source of infection
but may be larger.
If growing conditions are unsuitable for the pathogen, the
focus may stop expanding or even disappear with new growth
in the canopy. However, when conditions are favorable for
the pathogen, the disease will spread. As the primary focus
continues to expand, secondary foci and later tertiary foci
appear. This continued focal spread over a larger area is con-
sidered mesoscale spread and corresponds to a first-order
epidemic. This may be restricted to one field but may spread
over many fields or up to an area several hundred square kilo-
meters or even over part of a continent during a single grow-
ing season (5).
Synoptic or macroscale spread occurs when the epidemic
progresses for several years and spreads over an area of several
thousand square kilometers. This is also referred to as a
second-order epidemic. This pandemic may cover a whole
continent after a certain number of years (5). These concepts
that describe disease spread, from microscale to synoptic, have
been incorporated into various spore dispersal models and
simulations (8–11).
Long-Distance Transport
Long-range transport often explains the introduction of a
pathogen to a new area or its reintroduction to areas where
overwintering cannot occur (12). Because fungi are the
most abundant plant pathogens, most of the well-studied
examples of long-distance transport involve fungal spores.
The transport of Puccinia graminis uredospores from the
3.2.8-2 ▪ AIR
southern United States and Mexico to the wheat belt in the
northern United States and Canada has been well docu-
mented, as have other instances of long-distance dispersal
of rust spores (12, 13). Pedgley (3) and Brown and Hovmøller
(12) reviewed several cases of long-distance transport and
provided descriptions of the atmospheric conditions that pro-
mote these events. Griffin and colleagues (14, 15) described
the airborne microorganisms collected in the Caribbean
atmosphere during African dust events, and Mims and
Mims (16) described the long-distance transport of fungal
spores in smoke from the Yucatán. Long-distance transport
also explains the introduction of soybean rust into the United
States during the September 2004 hurricanes (17).
Backward trajectory analysis is frequently used to trace the
previous movement of a spore-laden parcel of air and locate
the inoculum source. Once the source is identified, forward
trajectories are used to indicate further potential areas of fall-
out. Various dispersion models have been used to trace the
movement of spores from dissemination at a source to deposi-
tion at a sink by calculating trajectories based on upper air
winds, temperature, and other parameters. If the source
strength is known, these models can be also used to calcu-
late the amount of spores deposited. The HYSPLIT (hybrid
single-particle Lagrangian integrated trajectory) model has
been widely and successfully used to track the movement of
several pathogens and predict the occurrence of disease out-
breaks (18–21). Most of the literature on long-distance trans-
port suggests that disease introduction through a single
transport episode or repeated incursions from great distances
are rare events that may have dramatic effects on plant popu-
lations (12). By contrast, Aylor (22) stressed the low probabil-
ity of these events and developed a model to explain the
continental spread of plant disease through short successive
steps that parallel the seasonal availability of the host plants.
Movement in the Atmosphere
Fungal spores are a normal component of the turbulent layer
of the atmosphere, and airborne transport normally occurs
there, with the dispersive power related to the intensity of tur-
bulence (23, 24). Airborne spores are present in large num-
bers whenever the ground is not covered with ice and snow.
Spores are discharged from fungi growing as parasites or sap-
robes, and atmospheric concentrations may exceed 200 000
spores/m3 of air (25, 26). Wind is the major factor, with gusts
and lulls affecting takeoff, transport, and deposition (1, 13).
In a crop, spores have to pass from the laminar layer close
to the leaf surface into the turbulent layer within the crop.
Gusts and turbulence enhance spore removal from leaves by
sweeping away the layer of slow-moving air next to the leaf
surface (1). Before spores reach the free air above the crop,
they must also pass through the boundary layer surrounding
the crop. Possibly as many as 90% of the spores are deposited
within the crop itself (27). The percentage that escapes from
the canopy depends on the balance between deposition and
turbulence with greater escape during more turbulent winds.
The position of the spores within the canopy also affects
escape. Spores produced lower in the canopy will have lower
rates of escape because they are exposed to slower winds and
less turbulence. In the early stages of a disease, when the infec-
tion may be confined to the bottom or middle of the canopy,
the spread of the disease may be limited. This may change as
the disease progresses. After the infection reaches the top of
the canopy, the increased rate of spore escape may permit
the disease to spread rapidly across the field (1).
Carried by wind, spores are transported both horizontal-
ly and vertically. Horizontally, spores can be carried for
thousands of kilometers, dispersing pathogens into new areas
(13, 28). Vertically, they may be carried upward by convec-
tive activity or thermals and have been recovered from alti-
tudes higher than 5,000 m. In spring and summer, when
the ground is heated to a greater degree, vertical movements
of air masses are more common, as is greater turbulence. The
atmosphere during this period is often characterized by the
maximum concentrations of airborne spores, including
many pathogenic spores from fungi of genera such as Erysiphe,
Drechslera, and Venturia.
Dispersal Mechanisms
Dispersal of spores into the atmosphere is dependent on the
method of discharge by individual taxa as well as environmen-
tal factors, such as temperature, humidity, and wind speed.
Spores are released into the air passively or by active discharge
mechanisms. A review of dispersal mechanisms for plant
pathogens has been published by McCartney and Fitt (29).
Passive Mechanisms
Many pathogenic fungi, such as the conidia of the powdery
mildews, Alternaria spp., and Drechslera spp. as well as rust ure-
dospores and smut teliospores, are passively dispersed in dry
weather. For the majority of these dry-dispersed spores, the
airborne concentration is dependent on the ease with which
spores are detached from the parent mycelium (or fruiting
structure) and on atmospheric conditions such as wind speed
and turbulence. Wind gusts facilitate their removal, as does
the movement of the leaf itself in the wind. The role of
wind gusts in the dispersal of spores has been reviewed by
Aylor (1). Strong gusts are able to disperse spores from leaf
surfaces even when the average wind speeds are too low to
accomplish this (3). Because wind dispersal for both patho-
genic and saprobic fungi is promoted by warm, dry weather,
the entrained propagules are referred to as the “dry air spora.”
They include the most abundant fungi in the environment
and usually peak during the afternoon hours, when humidity
is low and wind speeds are increased (24). Although there are
major similarities in the dry air spora worldwide, at any one
time the air spora may be dominated by nearby sources of
spores (26). During crop harvesting or mowing, incredible
numbers of spores (up to 109 spores/m3) may be dispersed
into the atmosphere, with Alternaria, Cladosporium, and
Epicoccum being the most abundant taxa (26, 30, 31). Pas-
sively dispersed spores may also be removed from the leaf
surface by raindrops as they strike the surface, causing a puff-
ing action that may remove the conidia (31, 32). This typi-
cally occurs when the first raindrops fall, and it may result
in increased atmospheric concentrations of conidia of Clado-
sporium, Alternaria, rust uredospores, or other fungi. Release of
puffball spores has a similar puffing action as raindrops strike
the mature fruiting bodies (24, 33). At maturity, puffballs
have a thin flexible wall that covers the gleba, or spore
mass. When raindrops strike this thin outer wall, spores are
discharged in puffs or plumes; strong gusts of wind or contact
with animals can also result in spore puffs.
Rain splash can also propel spores into the atmosphere,
and it is second to wind in importance as a means of pathogen
dispersal (29, 34–36). While rain is the most important
vehicle for splash dispersal, overhead irrigation can also dis-
perse plant pathogens. Splash-borne spores or bacteria are
usually produced in mucilage (a sticky or slimy polysaccharide
layer), which inhibits their direct removal by wind. However,
the mucilage also protects them from desiccation during dry
weather. The first raindrops dissolve the mucilage and leave
a spore suspension available for splash dispersal by additional

raindrops. Rajasab and Chawda (37) studied the dispersal of
Colletotrichum gloeosporioides conidia by rain splash and found
that the number of conidia varied from drop to drop. Splash
droplets from the first water drop to strike the leaf released
only a few conidia and some mucilage, with the highest num-
ber of conidia found in the fifth water drop. The size of the
droplet may also affect the number of conidia, with the small-
est droplets unable to carry spores. Most spores are carried by
droplets greater than 100 µm diameter (3, 31).
Rain splash confines dispersal to periods when the wet
conditions on the new host are also favorable for germination
and even reproduction (34, 38). The severity of bacterial
brown spot of snap bean was found to increase after rain
due to the rapid multiplication of the pathogen (39).
Many splash-dispersed spores have adhesive properties
that enable them to stick to the new host surface (40). There
is no evidence of a common attachment mechanism or
chemical component; these properties are varied and include
glycoproteins, lipids, and polysaccharides. In some species,
preformed adhesives are passively released from the spore on
hydration,while in others, adhesion is a metabolically active
process requiring the synthesis of new compounds. Addition-
ally, the extracellular matrix that surrounds some spores is
known to play a role in adhesion.
Splash-dispersed spores, such as conidia of Fusarium,
Cercospora, and Pseudocercosporella, frequently have smooth,
thin, hyaline walls and an elongated shape. Although they
lack the protective feature of thick pigmented walls com-
monly found on members of the dry air spora, they are pro-
tected from desiccation by wet conditions during dispersal
and deposition (34).
The effect of wind on dispersal of splash-borne spores
depends on the size of the raindrops. Very large droplets car-
rying spores are unaffected by wind, and small droplets may
evaporate and allow the spores to be dispersed through the
air as a true aerosol (34, 41). Most splash droplets are too large
to form aerosols. As a result, rain splash appears to be most
important for local dispersal, with gradients much shorter
than those of wind-dispersed spores (29, 35). Various studies
have examined the maximum height and distance that inoc-
ulum can be spread by rain splash for various plant pathogens
(42–44).
Bock et al. (42) examined the downwind dispersal dis-
tance of Xanthomonas axonopodis pv. Citri from canker-
infected citrus trees. Using methods that simulated both
rain splash and wind, they found that wind-driven rain was
able to disperse the bacterium up to 12 m from the host plant.
Lovell et al. (43) looked at the risk of inoculum transfer due to
rain splash for Septoria tritici conidia in winter wheat. Spore
dispersal from lesions on the bottom leaves to newly emergent
leaves occurs through rain splash. The authors showed that in
some cultivars the risk of infection decreased with height,
with a 90% risk for leaves that were 10 cm above the infected
leaves decreasing to a 10% risk for leaves that were 30 cm
above the infection. They also found that the distance
between diseased and healthy leaves varied among cultivars
tested. Paul et al. (44) examined the dispersal of Gibberella
zeae (anamorph Fusarium graminearum) spores in wheat fields
over three growing seasons. During every rain event, they
found that macroconidia were consistenly splash dispersed,
collected, and identified in all spore collectors positioned at
0, 30, and 100 cm above the ground. The results indicated
rain splash is able to spread pathogen spores to spike heights
within the wheat canopy.
In addition to dispersal, raindrops are very efficient spore
collectors, and the majority of airborne spores in the 20–30
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-3
µm diameter range are washed out by raindrops. Even a very
light rain can effectively wash spores from the atmosphere;
however, the significance of this rain scrubbing is difficult
to quantify (24).
Active Mechanisms
Active discharge mechanisms propel spores into the turbu-
lent layer, independent of wind. These mechanisms are wide-
spread among fungi, with the ballistics of some species being
quite spectacular. Many ascospores and basidiospores are
actively discharged by mechanisms that require moisture or
high humidity. The atmospheric concentration of these
spores often peaks at night or during early morning hours
when the humidity is high (1, 5). Aerobiological studies
have shown that ascospores are often abundant in the atmos-
phere during and after rainfall (34, 45–47). The explosive dis-
charge of ascospores from the ascus during periods of high
atmospheric moisture is a widespread characteristic of many
ascomycetes. High osmotic pressure develops within the ascus
either by the direct absorption of rainwater or by the swelling
of hygroscopic material within the ascus (48, 49). The result-
ing pressure causes the ascus tip to rupture, forcing the spores
out explosively into the turbulent layer. Discharge of basidio-
spores from mushrooms and bracket fungi also requires atmos-
pheric moisture, and the mechanism has been described as a
surface tension catapult (50, 51). High concentrations of
atmospheric basidiospores have been reported from various
locations with the release occurring during times of high
humidity (24, 45, 46, 52). Although basidiospore dispersal
is not directly tied to rainfall, fruiting bodies of mushrooms
and other basidiomycetes frequently develop following rain.
Release of basidiospores from these fruiting bodies, therefore,
results in high atmospheric concentrations during seasons
when rainfall is frequent (52).
Survival in the Atmosphere
Fungal spores that are adapted to airborne dispersal are often
much more resistant to environmental stress than are the
parent hyphae. However, they are still vulnerable to certain
types of environmental damage while airborne. Exposure to
harmful radiation and extremes of temperature and humidity
can decrease the viability or infectivity of pathogenic species.
Changes in relative humidity, often caused by changing
wind speeds, may affect survival, especially for thin-walled
spores, which may easily plasmolyze. The risks of desiccation
are usually greatest during the daytime and close to the
ground. At night and at high altitudes, conditions are less
stressful; spores have even been reported to germinate in
the clouds (25). Spores may also serve as condensation nuclei
for rain (13).
Radiation in the upper atmosphere, especially UV radia-
tion, may also affect the survival of spores carried into the
upper air. The effects of radiation may outweigh the effects
of temperature and humidity. Again, thin-walled colorless
spores may be more vulnerable because they lack the protec-
tion provided by melanin, which is present in the cell walls of
pigmented spores. Low temperatures in the upper atmosphere
may be preservative, however, and protect spores from UV
damage (24). Because of the erosion of the ozone layer in
the upper atmosphere, the effects of UV radiation on fungal
spores have been the focus of several studies. Scientists are
interested in possible decreases in spore germination, myce-
lium development, and spore formation caused by the radia-
tion (53, 54).
Despite environmental hazards, many spores are able to
survive long-range transport, but the percentage of viable
3.2.8-4 ▪ AIR
spores that actually reach a target and cause infection is low.
Spores can be deposited on the crop surface by sedimentation,
impaction, boundary layer exchange, turbulence, or electro-
static deposition and through raindrops (13).
PATHOGENS SPREAD BY THE
AEROBIOLOGICAL PATHWAY
A wide variety of plant pathogens, including viruses, bacteria,
oomycetes, and fungi, are dispersed through the atmosphere.
When they are deposited on a susceptible host, infection can
occur, and when environmental conditions are favorable, the
resulting disease spread may lead to widespread crop loss. A
thorough understanding of the role of the aerobiological
pathway in pathogen dispersal is necessary for the manage-
ment and control of disease.
Viruses
Most plant viruses are spread by means of insect vectors,
although relatively few insect groups transmit viruses (55–
57). Insects with sucking mouth parts (such as members of
the order Homoptera) transmit viruses by their stylet. Aphids
are the most dangerous insect vectors and are known to trans-
mit the majority of stylet-borne viruses. Whiteflies, another
important group of vectors in the same order, transmit a large
number of viruses belonging to several different families
including the begomoviruses (family Geminiviridae), which
are pathogens that severely affect a range of host species (58).
Barley yellow dwarf virus (BYDV) is often considered the
most widespread and economically important viral disease of
cereal crops; it affects more than 150 species, including wheat,
rice, barley, and oats. There are little data on the exact mon-
etary losses due to BYDV, but an estimated 5% loss of crops
infected by this virus would be greater than $1 billion in
the United States alone. BYDV can be an important limiting
factor for grain production throughout the world (59, 60).
Many wild grasses are also infected and thereby serve as reser-
voirs of infection. Barley yellow dwarf disease is caused by a
large complex of luteoviruses which share common aphid
vectors. This topic has been reviewed in several publications
(59, 61, 62), and Miller et al. (63) examined the phylogenetic
relationships of viruses in this group. BYDV is controlled by
the use of resistant plant lines (64) or pesticides. Recently,
researchers have been using genetic engineering to develop
crops resistant to BYDV (65). However, in many developing
areas where disease occurrence is high, farmers must live with
the losses caused by these viruses (59). The aerobiological
focus of BYDV research concerns the atmospheric movement
of the aphid vectors and the importance of winds in local and
long-distance aphid migration (61).
Plant viruses can become airborne when infected plants
are damaged by high winds, but in general evidence on the
aerosol dispersal of plant viruses is limited (41). However, air-
sampling methods to detect airborne viruses have improved in
recent years, along with the use of molecular techniques for
identification (66). A recent study by Whon et al. (67) used
confocal microscopy, transmission electron microscopy, and
metegenomic analysis to characterize the diversity of viral
DNA from air samples above three different land-use areas
in Korea. The authors found that the viral abundance in the
air samples ranged from 1.7 × 106 to 4.0 × 107 viruses/m3.
Although the majority of sequences were uncharacterized,
the metagenomic analysis showed that the identified sequen-
ces belonged to 12 virus families, including a variety of plant
viruses. An additional aerobiological mechanism for virus
transmission is through pollen. The transport of virus particles
by airborne pollen has been established for more than 15 plant
viruses with reviews by Cooper et al. (68) and Mink (69).
Plant-pathogenic viruses have been known for more than
100 years; however, the number of recognized virus diseases
has increased dramatically recently. Several theories have
been suggested to account for the recent increases: develop-
ment of resistance to commonly used pesticides among insect
vectors, migration and explosion of aphid and whitefly popu-
lations in more temperate regions, changes in cultivation
practices; and the introduction of exotic transplants, which
may enhance the long-distance movement of both virus
and vector (58, 70, 71).
Bacteria
Approximately 100 species of bacteria are serious phytopath-
ogens causing devastating crop losses each year. Bacterial dis-
eases are especially common in humid tropical areas but can
be destructive wherever warm, moist conditions are present
(55). Bacterial plant pathogens are generally Gram-negative
bacilli with species of Erwinia, Pseudomonas, Xanthomonas,
and Agrobacterium being well-studied phytopathogens; how-
ever, nonpathogenic species in these genera also occur (39,
41, 72–75). Clavibacter, Rhodococcus, and Streptomyces species
are familiar Gram-positive plant pathogens (76, 77), and
Conn et al. (78) identified the Gram-positive coccoid bacte-
rium Leuconostoc mesenteroides as the cause of a postharvest
decay of tomato.
Dispersal
Dispersal of pathogenic bacteria from plants is generally pas-
sive via water, wind-blown water, and animals. Agricultural
practices and agricultural workers, their machinery, and their
tools also play a major role in the spread of pathogenic bacte-
ria, with humans typically responsible for most long-distance
dispersal. In warm, humid climates, where dew and rain are
common, dispersal by rain splash is the major means of disease
spread (29, 34, 35, 41, 42, 79).
Phytopathogenic bacteria have no special mechanism for
producing airborne propagules; however, they become air-
borne during rain or irrigation as well as when crops are har-
vested (41). Bacteria can be passively carried from surfaces of
plants or soil by air currents and may even become airborne as
single cells, but they are usually carried on rafts of plant mate-
rial or in droplet nuclei following splash dispersal. In a review
of air-sampling techniques, Lacey and Venette (41) cite many
studies that document the airborne dispersal of bacterial
pathogens.
None of the phytopathogenic bacteria are capable of form-
ing endospores; thus, they remain susceptible to solar radia-
tion and desiccation and therefore limited in their ability to
survive long-distance aerial dispersal. Survival requires rapid
transport to an environment with adequate moisture, nutri-
tion, and temperature. Loss of water causing dehydration or
desiccation represents the major stress to airborne bacteria.
Dehydration is generally caused by the evaporation of water
from droplets carrying bacteria (80). This supports the sugges-
tion by Hirano and Upper (39) that splash dispersal is signifi-
cant only for local disease spread. However, it is also known
that some bacteria can serve as cloud condensation nuclei.
Franc (81) showed that viable Erwinia carotovora cells were
recovered from precipitation in Colorado. The storm system
that deposited the precipitation had originated in the Pacific
Ocean off the coast of California. The ability of E. carotovora
cells to function as cloud condensation nuclei and be pro-
tected from desiccation and radiation may explain their abil-
ity to survive this long-distance transport.

In recent years, another aspect of plant-associated bacteria
has been the focus of intense research and numerous reviews
(82–87). Contaminated spinach, lettuce, tomatoes, sprouts,
and cantaloupes have all made headlines when human ill-
ness, including numerous hospitalizations and deaths, have
been traced to Salmonella enterica, Escherichia coli O157:H7,
or Listeria monocytogenes growing on vegetables and fruits in
the field. This has led to many studies to determine the origin
of the pathogens as well as the ecological role of these bacteria
on plants. In several instances, the original source of contam-
ination has been traced to animal feces in the field. Animal
sources included cattle, feral pigs, rodents, opossum, birds,
sheep, deer, and coyotes. Other sources of enteric pathogens
include contaminated irrigation water or flood waters where
fecal material was deposited directly into the water or entered
through runoff from animal pastures following rain (82). Sub-
sequent spread to vegetables and fruits can occur through root
colonization from the soil or onto leaf surfaces from splash dis-
persal during rain or irrigation. Entry of pathogens into the
plant occurs through natural openings, such as stomata or
hydathodes, or through accidental wounds. Progress is also
being made in determining the methods of proliferation
within the plant. Over the past two decades, this issue has
led to the realization that plants can play a central role in
the life cycle of enteric pathogens. Further research in this
area is needed to fully understand the attachment, external
colonization, and internal colonization of these microbes as
well as identify the environmental factors that promote
pathogen proliferation.
Biocontrol Agents
The airborne dispersal of bacteria that are not pathogens has
been the focus of other agricultural research. Biocontrol
agents include bacteria, fungi, and viruses that can be used
to control plant pests and pathogens. The use of bacteria as
biocontrol agents has expanded greatly in recent decades,
and several species are currently used in various agricultural
applications including frost protection, control of plant dis-
eases, and control of insect pests (88–92). In some instances,
the organisms are naturally occurring microbial pathogens of
insects that are being used for pest control, or they are epi-
phytic competitors of foliar pathogens. In other cases, the
organisms are genetically engineered microorganisms, which
have been designed to perform specific actions.
Much research has focused on Bacillus thuringiensis, a ubiq-
uitous bacterium that naturally produces crystalline proteins
with insecticidal properties. Several reviews on this topic
have been published (93–95). This species exists as a soil
microbe and insect pathogen as well as a leaf-surface saprobe.
Dried inoculum containing endospores and crystals of insec-
ticidal proteins has been marketed since the 1960s as a safe
alternative to chemical pesticides. These biocontrol agents
enter the aerobiological pathway because they are applied as
sprays or dusts on plants to control a variety of insects, espe-
cially lepidopteran pests. It is the most widely used biopesti-
cide and accounts for 90% of the market (93).
There are many subspecies of B. thuringiensis, which differ
in the number and type of plasmids they contain. The genes
for the insecticidal proteins are borne on the plasmids (96).
More than 1,000 strains of B. thuringiensis have been isolated,
and more than 100 insecticidal proteins have been identified
and sequenced (95). Different strains of B. thuringiensis have
different insecticidal activities toward specific insect pests.
There are more than 400 B. thuringiensis preparations re-
gistered in the United States for control of various insect
pests. Although B. thuringiensis toxins degrade within a few
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-5
days, the endospores can survive for several years after spray
applications.
B. thuringiensis genes have been transferred into other
bacteria, including several species that are plant epiphytes
and endophytes (94). Pseudomonas fluorescens expressing B.
thuringiensis genes lengthens the life of toxins in the field
because P. fluorescens can propagate on the plant host and
continue producing crystal proteins. Clavibacter xyli, a bacte-
rial endophyte of corn with B. thuringiensis genes, can be ino-
culated into a corn plant to protect against the European corn
borer. Since the mid-1990s the genes for these insecticidal
proteins have also been genetically engineered into various
crops, with corn and cotton the most widely grown.
Researchers are also focusing on methods using other bac-
teria as biocontrol agents. These include using several other
bacteria with insecticidal toxins, such as Xenorhabdus nemato-
philus and Photorhabdus luminescens (93). These two species
are mutualistic symbionts within nematodes. The nematodes
act as vectors transporting the bacteria into the body of insect
larvae. Toxin genes from these bacteria have been cloned,
and some have been expressed in Escherichia coli. It has
been suggested that Xenorhabdus and Photorhabdus will be
the next generation of biopesticides. Other scientists are
studying P. fluorescens, P. syringae, Bacillus cereus, and other
Bacillus species to control various bacterial or fungal patho-
gens (88, 92, 96–99). P. fluorescens A506 is used to control
fire blight on apples and pears and is available commercially
as BlightBan A506. Fire blight, caused by the Erwinia amylo-
vora, is the most damaging bacterial disease of apples and
pears (92). Control of fire blight can often be achieved by
spray application of A506 to the flowers. Growth of the P.
fluorescens restricts the growth of E. amylovora on the stigma
surface and reduces the incidence of blossom blight. The
use of BlightBan A506 has shown increased use by growers
in recent years because many E. amylovora populations have
become resistant to streptomycin, which had been the stand-
ard control measure. Experiments to determine optimum tim-
ing and weather for applications of BlightBan, either alone or
in combination with other biocontrol agents, are still ongoing
(92, 100). P. fluorescens A506 is also registered commercially
to control frost injury of pear, cherry, apple, almond, peach,
tomato, potato, and strawberry (88).
Oomycetes
Oomycetes are fungal-like organisms in the Phylum Oomy-
cota of the Kingdom Stramenopila (Chromista). These
organisms were previously considered simple fungi, but bio-
chemical, ultrastructural, and phylogenetic evidence indicate
that the oomycetes are not in the main path of fungal evolu-
tion. Molecular evidence aligns the oomycetes with the
golden-brown and brown algae in the Kingdom Stramenopila
(101). Like true fungi, oomycetes have a hyphal body and
reproduce by spores. Also, like the true fungi, many important
and destructive plant pathogens are included in this group.
Plant diseases caused by oomycetes include downy mildews,
white rusts, damping-off diseases, and root rots. The most
important genus in this phylum is Phytophthora, which causes
serious diseases in a variety of crop plants. Some of the patho-
genic oomycetes are soil-borne fungi and others are foliar
pathogens with an airborne dispersal phase. This section
highlights two oomycetes with airborne propagules (Table 1).
Late Blight of Potato
The potato ranks as the world’s fourth most common food
crop, but the continued increase of this crop in developing
nations may be limited by the cost of disease control measures.
3.2.8-6 ▪ AIR

TABLE 1 Major oomycete and fungal pathogens with aerobiological dispersal

Disease Organism References
Late blight of potato Phytophthora infestans (101–111)
Blue mold Peronospora tabacina (1, 2, 21, 112–115)
Apple scab Venturia inaequalis (34, 104, 116–124)
Early blight, leaf spots Alternaria spp. (25, 125–130)
Fusarium head blight Fusarium spp. (131–142)
Coffee rust Hemelia vestatrix (12, 13)
Wheat stem rust Puccinia graminis (12, 13, 143–152)
Wheat leaf rust Puccinia triticina (syn. P. recondita) (13, 147, 148)
Soybean rust Phakopsora pachyrhizi (17, 18, 153–164)
Smut fungi Ustilago spp., Tilletia spp. (19, 125, 165–168)

Potatoes are attacked by many pathogens and pests, and the
total amount of agricultural chemicals applied to the potato
crop is larger than for any other food crop (102). Late blight
of potato has been the most important potato disease since
the 1840s, when it caused the destruction of the potato
crop in Ireland and the resulting widespread famine. Today,
more than 170 years later, late blight ranks as the most
destructive agricultural disease (103).
Phytophthora infestans is the pathogen responsible for late
blight of potato (and tomato). This oomycete occurs wher-
ever potatoes are grown, and all potato cultivars are suscepti-
ble. Populations of the pathogen are generally short-lived,
and a field with infected potatoes during a particular year
may or may not show diseased plants the following year
(104). The oomycete invades host tissue, resulting in the
rapid death of infected parts. Productivity of the potato plants
is greatly reduced, and tuber destruction frequently occurs as
well. Sporangia are produced on aerial hyphae, which grow
out from the stomata and are dispersed by wind, possibly car-
ried for tens of kilometers (104). At low temperatures (10–
15°C) and high humidity, sporangia germinate by producing
numerous zoospores, whereas at higher temperatures, each
sporangium gives rise to a single germ tube that develops
into hyphae. As a result during cool, wet periods, the produc-
tion of zoospores leads to an astonishingly rapid spread of the
disease. Without fungicidal protection, a blighted field can be
destroyed within a couple of days. Similar disease progression
also occurs in infected tomato fields (105). Multibillion-
dollar annual losses are caused by this pathogen for both
potato and tomato growers (106, 107).
The economic impact of fungicide application to control
P. infestans (or other pathogens) must not be overlooked.
Often the cost of repeated applications may actually outweigh
the value of the crop at harvest time. Other techniques, such
as sanitation, crop rotation, and use of genetically resistant
cultivars can reduce the need for fungicides (107, 108).
This multifaceted approach is referred to as integrated pest
management and has been widely recommended for control-
ling plant disease. To reduce fungicide use, detailed informa-
tion is needed about a pathogen’s life cycle, especially
the dispersal phase. For pathogens with airborne dispersal,
aerobiological studies and knowledge of the environmental
conditions that promote dispersal can supply information
on the optimum timing of fungicide applications to protect
the crop.
Potato late blight forecasting has a 45-year history, and in
many potato growing regions, when meteorological condi-
tions indicate that Phytophthora spread is likely to occur,
warnings are issued to apply fungicides (108). Today, many
of the forecasts and warnings are readily available online.
Forecasting models, such as Blitecast, have been successful
in reducing the number of fungicide applications. Other mod-
els have successfully incorporated air-sampling data for blight
forecasting. A recent Lagrangian stochastic simulation model
(109) may be useful for predicting releases rates for sporangia
and the spread of disease between fields. The aerobiological
models could possibly lead to a 50% reduction in fungicide
applications (105).
P. infestans overwinters as mycelia in infected tubers, but
in areas where both mating types occur, it can also overwinter
as oospores in the soil. When the pathogen was introduced to
Europe and the United States in the 1840s, only one mating
type (A1) was introduced. In 1950, the second mating type
(A2) was identified in central Mexico, the native home of
the oomycete. The A2 mating type was confined to Mexico
until 1980, when it spread through Europe. It is speculated
that this mating type was introduced in a large shipment of
potatoes from Mexico to Europe in the late 1970s. Outbreaks
of the A2 type began occurring in Europe in 1980; this type
has subsequently spread throughout the world (102, 110,
111). The introduction of the A2 mating type has increased
concerns about the possibility of sexual reproduction and
genetic recombination occurring in the field with new strains
developing. This has broad implications for potato breeders
searching for blight resistance. These concerns were rein-
forced by the outbreaks of late blight in the United States
and Canada during the 1990s. The strains of P. infestans
involved were nearly all resistant to metalaxyl, the fungicide
most widely used to control late blight (110, 111). Currently
fungicide-resistant strains have been found worldwide (103).
Two types of long-distance dispersal of P. infestans appa-
rently have taken place. Intercontinental migration has
been associated with the transport of infected plants or tubers
by humans. This occurred in the 1840s and again before the
1980 outbreak of the A2 mating type. Long-distance dispersal
over tens of kilometers is attributed to wind-blown sporangia.
Aylor et al. (109) showed that wind speeds of 1–2 m/s were
sufficient to disperse sporangia from an infected potato can-
opy a distance of 10–20 km in less than 3 h. Sporangia can sur-
vive for some hours at the reduced humidity encountered
during transport (104). Maps showing the rapid progress of
blight epidemics in the 1840s suggest that a second-order epi-
demic of late blight could occur during a single growing
season.
Much of the current focus of late blight research is cen-
tered on developing potato varieties resistant to the pathogen

using traditional methods as well as genetic engineering (103,
106). A wild potato species, Solanum bulbocastanum, is resist-
ant to all known races of P. infestans, and the resistance genes
from this species have been identified and cloned (106, 112).
Hybrids between white potato and S. bulbocastanum may
someday produce blight-resistant cultivars. Until these culti-
vars are a reality, growers must rely on forecasting models and
fungicides to keep late blight in check.
Blue Mold
Tobacco blue mold caused by the oomycete Peronospora taba-
cina (syn. P. hyoscyami f. sp. tabacina) is an unpredictable dis-
ease of both wild and cultivated tobacco, causing devastation
some years and not appearing at all in others. Blue mold was
first described in Australia during the 19th century, and the
organism was identified by Baily in 1890 (113). In North
America, the disease was confined to seedbeds until 1979,
when the first serious epidemic occurred. The infection rate
was especially severe in both 1979 and 1980, with a second-
order epidemic advancing at rates of 10–32 km/day north-
ward through the eastern United States and into southern
Canada. Crop losses in the United States and Canada during
these two years were estimated at approximately $350 million.
Both host plants and pathogen exist year-round in tropical
and subtropical areas such as the Mediterranean and Carib-
bean basins. In temperate regions, tobacco is grown as an
annual; in addition, P. tabacina is not able to overwinter in
temperate zones. As a result, the long-distance transport of
inocula from tropical regions, in the form of asexual sporan-
gia, must recur each year.
Infection can occur within 4 h after a sporangium lands on
the leaf surface. A symptom-free incubation period, typically
5–7 days, ends with the appearance of yellow lesions and the
development of new sporangia. Unlike Phytophthora, Perono-
spora produces no zoospores, and the sporangia themselves
(often referred to as sporangiospores or conidia) are the
only asexual propagules. At times sporangia can be extremely
abundant, producing up to 1 million spores/cm2 of lesion area
(2). Sporangia are released from the sporangiophore by the
twisting movements of the sporangiophore that occurs
when the leaf dries in the morning as humidity decreases
and temperature increases (114). During the spring in Con-
necticut, the greatest number of sporangia was found in the
air in the early morning; in summer, the highest concentra-
tions occurred later in the morning when turbulence may
have been greater (1). Cool, wet, overcast weather favors
the rapid advance of the pathogen; clear, hot, dry weather
stops disease spread (114).
Each spring in the eastern United States weather condi-
tions are favorable for the northward transport of Peronospora
sporangia from southern sources. Case studies of epidemics
occurring from 1979 to 1986 suggest at least two likely path-
ways of disease spread (114). In some years, the pathogen
spread northward from Florida and Georgia. In other years,
the disease occurred first in Kentucky and North Carolina
without first occurring in states farther southeast. The source
of the inoculum in this second pathway was believed to
be Nicotiana repanda in south-central Texas or cultivated
tobacco in Mexico. These cases have shown that long-
distance transport of sporangia can lead to severe epidemics
of blue mold, and forecasting systems can potentially provide
time for tobacco farmers to apply fungicides. A numerical
model was developed that encompasses the release, transport,
and deposition of blue mold sporangia (21).
Wiglesworth et al. (115) reported a forecasting system
using molecular probes. They described the amplification of
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-7
an oligonucleotide fragment of DNA from P. tabacina that
could be used to detect the fungus in infected parts of tobacco
plants. Spore traps along with the amplification of this DNA
fragment were used to predict a local blue mold outbreak.
Other predictive models make use of extant disease outbreaks,
weather fronts, and weather forecasts. From 1995 to 2011 the
North American Plant Disease Forecasting Center at the
Department of Plant Pathology, North Carolina State Uni-
versity, successfully used the HYSPLIT trajectory model to
predict outbreaks of blue mold in the eastern United States
(169). The disease forecast employed documented reports
of blue mold along with meteorological models from the
National Oceanic and Atmospheric Administration, to plot
trajectories of inoculum-laden parcels of air. Forecasters
used existing and continuing sources of blue mold to develop
daily trajectories during the growing season. The daily fore-
cast produced with HYSPLIT trajectories described current
and future weather conditions at the source and along the
anticipated pathway, with emphasis given to those atmos-
pheric conditions that favored sporulation at the source, sur-
vival during transport, and deposition. An overall outlook
described the likelihood of blue mold spread over the subse-
quent 48 h. These forecasts were available on the Internet,
and the forecasters believed that this model provided valuable
information for controlling blue mold epidemics.
Fungi
Over 70% of all major crop diseases are caused by fungi with
thousands of fungal species recognized as plant pathogens. It
is estimated that fungal diseases cost more than $3.5 billion
annually to U.S. farmers alone (33). In general spores of
most fungal pathogens are adapted for airborne transport;
however, much of the aerobiological research on agricultural
pathogens has focused on a limited number of fungi that cause
economically important diseases (41). Air sampling can be a
valuable tool because the management of plant disease
requires an understanding of the airborne dispersal of inocula.
Routine examination of air samples from Tulsa, Oklahoma,
analyzed in the author’s laboratory provides an understanding
of the diversity of the spore types in the atmosphere (45, 116,
125). These data are representative of airborne spores found
in temperate climates throughout the world. The dry air spora
is dominated by Cladosporium as it is in most areas; in Tulsa
hourly concentrations of these spores have sometimes
exceeded 100,000 spores/m3 (25). In addition, spores of Alter-
naria, Bipolaris, Curvularia, Drechslera, Epicoccum, Helmintho-
sporium, Nigrospora, Pithomyces, Stemphylium, and Torula as
well as smut teliospores typically occur in daily air samples
during dry weather. In wet weather, ascospores predominate
with Leptosphaeria often abundant along with Venturia, Didy-
mella, and spores from members of the family Diatrypaceae
during certain seasons. Asexual conidia of Cercospora, Fusa-
rium, and various basidiospores are also abundant in moist
weather. The following discussion examines a few of the sig-
nificant fungal pathogens that are dispersed by aerobiological
pathway (Table 1).
Apple Scab
Scab is the most important disease of apples with all commer-
cial varieties susceptible to attack by the ascomycete Venturia
inaequalis (33). The disease occurs in every country where
apple trees are grown, and similar scab diseases affect pear
and hawthorn. Scab lesions occur on both leaves and fruit;
the disease can also cause premature defoliation (117). Con-
trol of apple scab requires repeated application of fungicides
3.2.8-8 ▪ AIR
(up to 18–25 times per season) to protect the crop. Without
them, 70–100% of the crop would be unsalable (33).
The fungus overwinters as immature fruiting bodies on
dead leaves on the ground in an orchard. The ascocarps
mature in the spring, and ascospores are actively discharged
from ascocarps during rainfall (34). Stensvand et al. (118)
reported the capture of ascospores during dew on several occa-
sions; however, dew is not generally accepted as a spore release
mechanism (119). Although discharged when it is raining,
the spores are dry, airborne spores. The rain causes the asci
to swell and release the spores (34). In one study, it was
observed that the airborne concentration of V. inaequalis
ascospores decreased rapidly at greater heights above ground.
The concentrations at 3 m were only 6% of the values at 15
cm. A spore dispersal model suggests that this decrease was
due to the rapid increase of wind speed and eddies above
ground level (120). The spores become a major component
of the air spora in orchards and are carried by wind to leaves,
where they cause a primary infection (105, 121). During the
spring there is usually a constant source of ascospores. Ponti
and Cavanni (105) showed that ascospores typically could
be found in the air of Italian apple orchards during periods
of rainfall from mid-March to mid-June. In the northeastern
United States, ascospores were found in orchards from mid-
April to early June (122). Once the primary infection is estab-
lished, conidia of the anamorphic stage (Spilocaea pomi) begin
developing. The more primary inoculum present, the more
rapidly the disease will build up and the more serious the epi-
demic will ultimately be.
Fungicide applications should be timed to coincide with
periods of rainfall in the spring when ascospores are released.
Air sampling has been used to provide information about the
duration of the ascospore season in a particular area and there-
fore help limit the use of fungicides to the periods when it is
absolutely necessary (105). Much research has been focused
on modeling the aerial dispersal of ascospores to help evaluate
the risk of infection (119, 121, 123, 124, 126).
Alternaria
Alternaria is a genus of asexual or anamorphic fungi assigned
to the form class Hyphomycetes. Fungi in this genus are ana-
morphs of the ascomycete genus Lewia. Alternaria species are
characterized by distinctive large multicellular dictyospores
that have a beak and are produced in chains. Species of Alter-
naria occur as parasites on a number of crop plants, causing
early blight or leaf spot diseases, or as saprobes on a wide vari-
ety of organic substrates. Crop losses caused by Alternaria
pathogens are less serious than those caused by rusts or downy
mildews; however, this genus is prominent in aerobiological
literature because it is recognized as an important aeroaller-
gen. Aerobiological surveys including those conducted by
the American Academy of Allergy, Asthma, and Immunol-
ogy (127) routinely report the presence of Alternaria conidia
in the atmosphere. This ubiquitous genus can frequently be
found in atmospheric concentrations of several hundred to
1,000 or more spores per cubic meter of air with peak concen-
trations typically occurring during late summer or fall (25,
116, 128).
Many pathogenic species of Alternaria have worldwide dis-
tribution, including Alternaria solani on potato and tomato,
A. brassicae on members of the Brassicaceae, A. porri on
onions, and A. alternata on a variety of host species (129).
In general, Alternaria pathogens often attack plants under
stress, especially those affected by drought, insect infestation,
or senescence. Young seedlings are also susceptible to Alter-
naria, but established to middle-aged plants may be less
vulnerable. However, Vloutoglou and Kalogerakis (130)
found that two tomato varieties were susceptible to A. solani
infection at all ages. The plants became increasingly vulner-
able as they aged, with the greatest susceptibility at the repro-
ductive stage.
Alternaria spores are passively dispersed from infected
leaves by moderate to strong gusty winds with velocities of
2–3 m/s required for spore release. As a component of the
dry air spora, dispersal typically occurs during midday when
conditions are warm and dry with high wind speeds. Studies
have shown significant correlations between airborne Alter-
naria concentrations and temperature (47, 116, 131). The
greatest dispersal occurs during dry weather that immediately
follows periods of rain or heavy dew. Prolonged dry windy
periods deplete spore reserves on the leaves and inhibit spor-
ulation, which requires moist conditions. Spores can also dis-
perse when washed or splashed from leaf surfaces by rain and
by irrigation; however, for Alternaria, this method of dispersal
is less important. By contrast, high humidity inhibits the
release of spores from wet leaves, and airborne spores are
washed from the atmosphere by rain and irrigation. Several
predictive models based on meteorology have been developed
to aid in timing fungicide application (128, 129, 131).
At the present time there is no nationwide network of
plant pathologists conducting air sampling to detect airborne
pathogens. However, the American Academy of Allergy,
Asthma, and Immunology sponsors a network that consists
of approximately 84 certified air sampling and reporting sta-
tions in the United States. In the 1998 network report, air-
borne fungal spore concentrations were reported from 46
sampling stations. Alternaria conidia were among the top 10
spore types (in term of atmospheric concentrations) reported
from 45 stations, and at 36 stations Alternaria spores were
among the top 5 (127). Although some of the airborne
Alternaria spores may be from saprobic species of Alternaria,
these spores still represent a significant component of the
air spora in the United States and may indicate a potential
threat to crops.
Fusarium Head Blight
Fusarium head blight (FHB) is an important disease of wheat
causing small, shriveled, bleached grains, which reduce grain
quality and yield. FHB occurs throughout the wheat-growing
regions of the world and is caused by several species of Fusa-
rium and Microdochium. Literature on FHB is extensive with
several recent reviews (132–136).
In most areas Fusarium graminearum (teleomorph Gibber-
ella zeae) is the most prevalent species causing FHB. While
wheat is the most important host, FHB also attacks barley
and other small grains. In addition to head blight in these
crops, F. graminearum causes ear rot of corn. Besides affecting
crop yield, F. graminearum and other Fusarium species have
the ability to synthesize a variety of trichothecene mycotox-
ins, including deoxynivalenol (DON), nivalenol, and zeara-
lenone. The type of toxin varies with the FHB species, the
pathogen race, and environmental conditions. The toxins
cause an additional economic loss due to rejection or down-
grading of the harvested grains (134). The presence of toxins
in the grain is of major concern because they remain after
processing and cooking and pose a significant health risk to
humans and domesticated animals (132). For example,
DON inhibits protein synthesis. When this toxin is ingested
in sufficient quantities, it primarily affects the digestive system
although other organs can also be affected. For this reason,
DON levels in grain are regulated in most developed coun-
tries. In the United States the limit is 1 ppm for human

consumption, whereas in the European Union the limit is 0.5
ppm. Zearalenone causes estrogenic effects in livestock and
humans and can interfere with ovulation, conception, and
fetal development. At the present time there are no regulatory
limits for zearalenone in the United States or European
Union (135).
F. graminearum is a homothallic species, producing both
asexual macroconidia and sexual ascospores. It is believed
that this may give F. graminearum an epidemiological advant-
age over other FHB pathogens, which can only produce con-
idia (133). The fungus overwinters on crop residue in the field
or in soil and gives rise to macroconidia as well as perithecia
that produce asci and ascospores. Ascospores are forcibly
ejected from the perithecium during and after rainfall (137).
Although macroconidia are dispersed by wind and rain-
splash for all Fusarium species, ascospores are considered the
primary inoculum for F. graminearum (G. zeae) (137).
Numerous studies have shown that high humidity and rainfall
promote spore production on crop residue and the resulting
spores are then blown to developing wheat. Wheat is most
susceptible during flowering, but some infection can occur
later during early grain development. Current models for dis-
ease forecasting use local weather conditions during the week
before flowering to predict risk for FHB development and
make decisions for fungicide application (138). These models
assume all spores are produced locally and that no long-
distance transport occurs; however, several studies suggest
otherwise.
Maldonado-Ramirez et al. (139) collected viable F. grami-
nearum (G. zeae) spores in the planetary boundary layer of the
atmosphere at 60 m above agricultural fields using a remotely
piloted vehicle equipped with spore sampling equipment.
Sampling took place in late May and June over four years
using Fusarium-selective media. More than 12,000 F. grami-
nearum colonies developed from spores collected during
158 flights. The mean number of colonies collected during
cloudy weather was significantly greater than during sunny
or rainy conditions. The authors conclude that recovery at
60 m indicate that long-distance transport of FHB pathogens
should be considered in disease forecasting. Schmale et al.
(140) used autonomous (self-controlled) unmanned aerial
vehicles to collect air samples onto Fusarium-selective
medium. Isolates of F. graminearum were collected during
11 flights at 40–320 m above an agricultural field in Virginia.
All 11 isolates were sequenced to confirm identification, and
mycotoxin potential was determined for each isolate. The iso-
lates were also inoculated onto spring wheat to verify FHB
development. Several isolates were collected from flights dur-
ing fall and winter when no local host crops were present. The
authors suggest that spores of these isolates were likely trans-
ported to the area from states to the south, highlighting the
potential long-distance transport of FHB pathogens.
Other studies (138, 141–143) have focused on the in-
fluence of environmental conditions on FHB disease devel-
opment, fungal biomass, and mycotoxin production in
attempts to improve forecasting for epidemics. These studies
all show that environmental factors are important during spe-
cific windows of time; however, no single factor or time slot
could predict all disease occurrences. Continued research
into the environmental risk factors is essential to improve pre-
dictive forecasts for disease management.
Rusts
There are about 6,000 species of rust fungi, which attack
a wide range of hosts among seed plants and cause some
of the most destructive plant diseases. Millennia before
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-9
scientists understood that pathogenic fungi could cause plant
diseases, rust epidemics were studied, and the reddish lesions
on plants were noted. These epidemics were recognized in
ancient Greece and described in the writings of Aristotle
and Theophrastus, who even noted that different plants var-
ied in their susceptibility to these diseases. Today, rust fungi
remain among the most serious agricultural pathogens.
Some of the most important diseases of cereal crops are caused
by these basidiomycetes, which have produced serious epi-
demics throughout history. In addition, coffee, apple, soy-
bean, and pine trees as well as the economies dependent on
these crops have been devastated by rust fungi.
The long-distance transport of rust fungi has been studied
more extensively than that of any other fungal pathogen. Cof-
fee rust, caused by Hemileia vastatrix, destroyed the coffee
plantations in Ceylon in the 1870s and 1880s and today
threatens production wherever coffee is grown. In 1966, an
outbreak of coffee rust in Angola produced spores that were
apparently carried by favorable wind currents across the
Atlantic Ocean. The spores were washed out by rainfall
over coffee plantations in Brazil approximately five to seven
days later (13). A similar example of long-distance dispersal
has been described for sugarcane rust caused by Puccinia mel-
anocephala. Meteorological data indicate that sugarcane rust
was introduced into the Dominican Republic from Cameroon
via uredospores transported by cyclonic winds across the
Atlantic Ocean (12). The remaining focus of this section is
on the pathogens causing wheat rust and soybean rust.
Wheat Rust
Three devastation rust diseases threaten wheat: stem rust
of wheat, yellow rust (also know as stripe rust), and leaf
rust. Historically stem rust has been the most important wheat
rust pathogen; in recent years yellow rust and leaf rust epidem-
ics have been especially frequent and severe in many parts of
the world (144, 145). The focus here will be on stem rust and
the growing threat caused by new races of this pathogen.
Puccinia graminis f. sp. tritici is the fungus responsible for
stem rust of wheat (170). The organism has a complex life
cycle that involves five spore stages on two separate host
plants (wheat and barberry). Basidiospores which develop
from overwintering teliospores are capable of infecting a
young barberry leaf, eventually giving rise to spermatia and
receptive hyphae and then aeciospores, also on the barberry
leaf. The aeciospores transfer the infection to wheat where
the spores germinate, and hyphae enter the plant through sto-
mata. Once the mycelium is established in wheat, the uredial
stage develops within two weeks, with uredia appearing on
the stem as long narrow lesions that produce dark red powdery
masses of uredospores (also called urediniospores). Uredo-
spores become airborne and reinfect new wheat plants to pro-
duce repeated generations of uredia. Near the end of the
growing season, the uredia turn black when two-celled over-
wintering teliospores replace the uredospores (33, 55).
The major vehicles for disease spread are the uredospores,
which are easily carried by wind from one plant to another,
giving rise to epidemics. In fact, they can be carried by pre-
vailing winds for hundreds or thousands of kilometers (13,
146). In the mild climate of northern Mexico and southern
Texas, this stage can continue all winter and give rise to spring
infections in northern states, with spores being carried by pre-
vailing southerly winds. Uredospores can also be carried back
to the south in late summer and fall. Initially demonstrated in
1923 by Stakman, the movement of rust uredospores along
the “Puccinia pathway” in North America is one of the best-
known examples of long-distance dispersal (147). Puccinia
3.2.8-10 ▪ AIR
graminis, P. triticina (formerly P. recondita), P. coronata, and
Bipolaris maydis as well as insect vectors of viral diseases may
all be connected with dispersal along this pathway (13).
Wheat plants infected by P. graminis are severely weak-
ened but not destroyed, and the grain yield is significantly
reduced. Wheat plants throughout the world are threatened,
and it is estimated that worldwide more than a million metric
tons of wheat are lost annually due to stem rust. Evidence of
the long-distance transport of P. graminis uredospores in other
parts of the world was included in a review by Nagarajan and
Singh (13). In Europe, two pathways have been studied: an
eastern European pathway originating in Turkey and Roma-
nia and a western pathway from Morocco and Spain. Both
pathways converge in the Scandinavian countries. In India,
the uredospores of P. graminis and P. triticina survive through-
out the year in the Nilgiri Hills in southern India. Dispersal to
central and northern India occurs during November under
the influence of tropical cyclones (13).
The long-distance dispersal of P. graminis uredospores
between the eastern and western wheat-growing regions of
Australia has been documented, and the overseas transport
of stem rust has also been investigated. Two strains discovered
in Australia were shown to be identical to strains found in
South Africa in terms of pathogenicity and isozyme patterns.
In addition, upper air wind currents studied by use of a
weather balloon support the possibility of this transport (13).
Successful long-distance dispersal of wheat rust uredo-
spores is also dependent on source strength and viability. A
mature uredium is capable of producing about 10,000 uredo-
spores/day over a period of several weeks (55). With a 5% dis-
ease severity of 50 uredia, a single plant would produce
500,000 uredospores/day. It has been estimated that a field
of wheat with a moderate infection of stem rust would produce
4 × 1012 uredospores/day/ha (13). Nagarajan and Singh (13)
reported that following long-distance transport, the spores had
to be deposited within 120 h after takeoff to be infectious.
Eversmeyer and Kramer (148) studied the survival of P. grami-
nis and P. triticina uredospores under a variety of temperature
conditions in the field. At subfreezing temperatures during
winter, no spores were viable after 96 h, but during spring
10–20% of the inoculum was viable after 120 h and a fraction
of 1% survived for 456 h. In environmental chamber ex-
periments, spores remained viable for up to 864 h at constant
temperatures between 10°C and 30°C. However, the viability
of spores exposed to freezing and subfreezing temperatures
declined rapidly within a few hours of exposure (149).
Although conditions in the atmosphere may be similarly
harsh, the enormous numbers of spores produced will ensure
that at least a small number of spores will reach a suitable host.
Since the mid-1950s, P. graminis has been controlled in
most countries by planting rust-resistant varieties of wheat.
Approximately 60 stem rust resistance genes have been iden-
tified, although not all of these genes are equally effective
(144, 145, 150). Several resistant genes originated in wheat
species, while others were identified in closely related grasses
and integrated into wheat using conventional breeding tech-
niques. One of the most effective resistance genes had been
Sr31, which was originally identified in rye and had been
incorporated into most wheat varieties around the world.
Since the development of wheat varieties with the Sr31
gene in the 1980s, stem rust had been under control in
most countries. Another broadly incorporated resistance
gene has been Sr24, originally identified in tall wheatgrass,
a wild relative of bread wheat (150, 151). The last major
stem rust epidemic during the 20th century was in Ethiopia
during 1993–94 (145).
Ug99 Lineage
A severe infection of stem rust of wheat was detected in
1998 in Uganda. The fungus was characterized early in
1999 as a new race of P. graminis and was named Ug99 (tech-
nically, race TTKSK). Ug99 was the first race of P. graminis to
show virulence to the Sr31 resistant gene. Ug99 soon spread
to Kenya (2001) and Ethiopia (2003) where major epidemics
developed (145). Subsequently, seven variants of Ug99 have
been identified. In addition to Sr31, the Ug99 lineage (Ug99
and variants) shows virulence toward many other Sr genes;
however, each variant differs in virulence patterns to specific
Sr genes. At least three of the variants show resistance to
Sr24, and combined virulence to both Sr31 and Sr24 has
been found in variants identified in seven African countries
where they are often the predominant race. The Ug99 lineage
has now been identified in 11 countries, 9 in Africa (Uganda,
Kenya, Ethiopia, Sudan, Tanzania, Eritrea, South Africa,
Mozambique, and Zimbabwe) and 2 in Asia (Yemen and
Iran) (145, 152). The rapid movement of the pathogen and
the high virulence suggest that commercial wheat cultivars
worldwide are vulnerable.
The original Ug99 is the most widely distributed race in
this lineage and, through 2012, has been the only race occur-
ring outside of Africa. However, the other variants are
increasing their distribution, and proliferation out of Africa
is likely to occur (145, 152). The spread of the Ug99 lineage
in east Africa is believed to be occurring along existing airflow
pathways (144); however, the dispersal of Ug99 variants from
east Africa to southern Africa is still uncertain (145). The
pathogen could have spread along wind patterns. Trajectories
from Kenya show the regular movement of air masses into
Tanzania from January to March. Rust spores from infected
plants in Tanzania could have moved on to cause infections
in Mozambique, Zimbabwe, and South Africa. The move-
ment of soybean rust followed a similar path in Africa from
1996 to 2001. Alternatively, the southern movement could
be explained by the accidental introduction due to the exten-
sive trade and travel between Kenya and South Africa (145).
The presence of Ug99 in Iran is a major threat to the
wheat-growing areas of south Asia. The previous spread of
yellow rust from east Africa into the Middle East and then
into Pakistan and India during the 1990s illustrates the real
possibility of this dispersal pathway (153). HYSPLIT trajecto-
ries from infected areas of Iran show the potential spread to
the southeast into Pakistan. Trajectories from Iran also
show the potential spread to the northeast into the Caucasus
and Central Asia (145). A more direct introduction into
south Asia from Yemen and Eritrea is also considered a possi-
bility. HYSPLIT trajectories in 2011 from infected areas in
these countries showed that airflow was able to reach south
Asia within 72 h (144).
The occurrence of Ug99 lineage in South Africa, espe-
cially those races with combined Sr31 and Sr24 virulence,
is an additional cause for concern due to the possible intro-
duction to Australia. Modeled trajectories from source loca-
tions in South Africa during October and November 2010
reached Australia in six to eight days (144). This possibility
is supported by an earlier introduction of stem rust into Aus-
tralia in 1969, which is believed to have originated in south-
ern Africa (12).
The global wheat community is actively involved in
developing strains of wheat resistant to pathogens in the
Ug99 lineage. The use of resistant varieties is the most feasible
option for many poor farmers in Africa and Asia, since the
cost for chemical control would be prohibitive. Several strains
are promising and initial plantings of many of these in several

countries have shown good yield. Continued developing of
resistant varieties is necessary to ensure genetic diversity
and protect against new Ug99 variants (145).
Extensive monitoring efforts have been established
through various international agencies to provide early warn-
ings of wheat rust epidemics. In 2010 the UN Food and Agri-
culture Organization initiated Rust SPORE (http://www.fao.
org/agriculture/crops/rust/stem/en/) a site to track the progress
of Ug99. Similar efforts are also under way with the Global
Cereal Rust Monitoring System (GCRMS), which was devel-
oped by an international team under the auspices of the
Borlaug Global Rust Initiative. Modern technology is a com-
ponent of the GCRMS monitoring system utilizing smart-
phones and tablet computers. Apps have been developed
that record and upload GPS data and photos of infected
plants and also allow for automatic data entry into a rust sur-
vey form and central database (152). These efforts need to be
continued to prevent further spread of the Ug99 lineage.
Soybean Rust
Another rust disease that has been attracting considerable
attention is Asian soybean rust, which has become a major
concern in all soybean-growing countries around the world.
The disease is caused by Phakopsora pachyrhizi. Although an
additional species of Phakopsora, P. meibomiae, also attacks
soybean, P. pachyrhizi is far more virulent and is the focus of
this discussion. Asian soybean rust is the most destructive
foliar pathogen on soybean and reports of loss range have
reached 80% in some areas. Yield losses over 50% are com-
mon when meteorological conditions favor disease develop-
ment (55, 154). The fungus causes numerous uredial lesions
on the leaves, reducing photosynthetic capacity of the host
and subsequently reducing the yield of soybeans. Severe
infection causes defoliation resulting in even greater loss.
The life cycle of soybean rust is far simpler than that described
FIGURE 1 Global spread of soybean rust caused by Phakopsora pachyrhizi. Map supplied by Annalisa Ariatti, Dept. of Plant
Pathology, Pennsylvania State University, University Park, PA 16802. doi:10.1128/9781555818821.ch3.2.8.f1
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-11
for stem rust of wheat. P. pachyrhizi can produce both uredia
and telia on the same host plants. Uredia give rise to abun-
dant uredospores that readily become airborne and spread
the infection. Telia with teliospores have been reported late
in the growing season on infected plants and in greenhouses,
but this stage is rare and teliospores are not important for dis-
ease spread (55, 155). Unlike many rust fungi that have a nar-
row host range, P. pachyrhizi can infect more than 150 species
in 53 genera of the Fabaceae, including several other edible
legumes (154, 156). Kudzu, a widespread weed in many parts
of the world, can also be infected by this fungus and serve as
an overwintering host in warm climates (157). In the south-
eastern United States, kudzu is an aggressive, invasive species,
which can quickly grow over trees and shrubs.
P. pachyrhizi was first described in Japan in 1902 (158). By
the mid-1930s the pathogen was reported from several other
countries in Asia and in Australia and was identified in India
in the early 1950s. Reports in the late 1990s showed that the
pathogen had spread to several countries in Africa, and in
2001 it was reported from Paraguay and Brazil in South Amer-
ica (159). Over the next few years, the pathogen spread
through much of the soybean-growing areas of South Amer-
ica (Fig. 1) causing significant yield losses (155, 159, 160).
The fungus was first detected in the continental United
States in November 2004 (17). Researchers tracked the
movement of the P. pachyrhizi spores in South and Central
America using various dispersal models. Evidence from these
models indicated that the pathogen was introduced during
the 2004 hurricane season, specifically during Hurricane
Ivan. The movement of the pathogen through South Amer-
ica and into the United States shows the importance of the
aerobiological pathway for long-distance dispersal (18, 157).
Since the introduction of P. pachyrhizi into North Amer-
ica, Asian soybean rust has been detected in all major
soybean-growing areas (157, 161). Airborne transport of
3.2.8-12 ▪ AIR
uredospores explains the yearly movement of the disease in
the United States by short hops or long-distant transport.
The pathogen overwinters on kudzu in Florida and other
states bordering the Gulf of Mexico (156). In these areas,
kudzu leaves remain green during winter, and in the spring,
uredospores from infected kudzu spread the disease to soybean
plants in southern states. During the growing season the
pathogen spreads north, infecting both soybean and kudzu
in other states. Kudzu can serve as a bridge host in these areas
since kudzu leaves emerge before soybean. Reports from the
field indicate that these are the predominant hosts, although
other legumes are also infected (156, 161). By 2009, soybean
rust infections had been identified in 20 U.S. states; Ontario,
Canada; and 9 states in Mexico (157, 162). However, the
extent of pathogen distribution in the United States varies
year by year based on meteorological conditions, which influ-
ence disease severity and dispersal (161). Phakopsora uredo-
spores are easily airborne, and washout during rain is
especially important because wet conditions favor spore ger-
mination and disease development. Optimum temperatures
are from 17°C to 27°C, while temperatures above 30°C and
sunny skies repress disease development (157).
Control of soybean rust is through the use of fungicides
and resistant varieties of soybean; however, commercial culti-
vars grown in the United States are not resistant to
P. pachyrhizi (154, 162). Although several genes have been
identified that confer high levels of resistance, they are not
effective against all races of the pathogen. Cultivars with
single-gene resistance have been used in some countries
where the pathogen has become endemic; however, failures
of these cultivars have occurred (162). Durable resistance
will require a combination of resistance genes since single-
gene resistance is quickly overcome by variability in the
pathogen. Ongoing screening programs in the United States
and other countries continue the search for additional resist-
ant genes.
Until resistant cultivars are available, management of the
pathogen will depend on the use of fungicides timed to peri-
ods when environmental conditions are optimal for pathogen
dispersal and infection. Consequently, monitoring and fore-
casting are essential for successful control of Asian soybean
rust. The U.S. Department of Agriculture has sponsored a
multifaceted soybean rust–monitoring program in North
America (157, 161, 163). A central website, Integrated Pest
Management—Pest Information Platform for Extension
and Education, (http://sbr.ipmpipe.org) provides real-time
information of the status of Phakopsora in North America
on a daily basis. Sentinel plots, spore trapping, and disease
forecasts based on aerobiological models are all part of this
program. Forecasts incorporate known source areas, meteoro-
logical data from the National Weather Service, and outputs
from several trajectory models (19, 163, 164) to predict the
release, dispersal, and deposition of airborne uredospores.
The program has proven successful and has resulted in high
yields in some areas when timely fungicide application has
saved a crop. In addition, it has also resulted in significant
financial savings when forecasts prevented unnecessary fungi-
cide application (161, 163). Numerous other websites offer
additional information on the pathogen along with recom-
mendations for fungicide use.
Smuts
There are more than 1,500 species of smut fungi classified in
77 genera with Tilletia and Ustilago the most commonly occur-
ring genera. About 800 species of smut fungi attack members
of the Poaceae, including important crops such as corn,
wheat, barley, sugarcane, and sorghum, causing millions of
dollars in damage each year. Many native plants are also
affected, with a total of over 4,100 host species (165,166).
The visible spore stage of smut fungi is the teliospore stage.
Enormous numbers of teliospores develop from the mycelium,
usually within galls, and are dispersed by wind. There is no
repeating stage; each infected plant produces only one gener-
ation of teliospores. Many smuts overwinter as teliospores;
however, other smut species overwinter as mycelia in infected
grain. Under suitable environmental conditions, each telio-
spore gives rise to a basidium, which produces haploid basi-
diospores. The basidiospores may germinate giving rise to
fine hyphae, or they may reproduce by budding, thereby form-
ing a yeast-like stage. Fusion of compatible basidiospores or
yeast cells or hyphae results in the establishment of dikaryotic
hyphae. Dikaryon formation is necessary before the fungus
can invade host tissue (55, 165, 166).
Corn smut is a widespread disease occurring wherever corn
is grown; however, it is more prevalent on sweet corn than on
other varieties. It is generally controlled through resistant
cultivars. The fungus Ustilago maydis can form galls on any
aboveground plant part but is most conspicuous when the
galls develop on the ear. The size and location of the galls
reflect the degree of crop loss. Galls on the ear result in total
loss, whereas those in other places may reduce yield or cause
stunted growth (55). In recent years, U. maydis has become
a major research tool in elucidating the molecular aspects of
disease development (166).
The fungus overwinters as teliospores on plant debris and
in the soil. In spring, the teliospores germinate; each produces
a basidium that forms basidiospores. These are carried by wind
or splashed by rain to developing tissues of corn plants; gen-
erally young seedlings or the growing tissues of older plants
are vulnerable. Once the dikaryon is established, the resulting
mycelium grows within the developing plant and eventually
stimulates host cells to divide and form galls, which may reach
a size of 15 cm in diameter. As the galls mature, the interior
darkens as the mycelium is converted into teliospores. The
galls are initially covered by a membrane, which later rup-
tures, releasing masses of dry teliospores (55).
Various studies have also documented the presence of
smut spores in the atmosphere. Halwagy (167) reported that
Ustilago spores were the second most common spore type
identified in the Kuwait atmosphere. Hasnain et al. (168)
examined the concentration of airborne smut spores from
three cities in Saudi Arabia and found maximum monthly
concentration reached 4,000 spores/m3 in Jizan during the
fall with the lower concentration in the other locations.
Although all smuts were considered as a single category, the
authors suggest that the category included several Ustilago
species, Tilletia caries, and Sphacelotheca occidentalis.
Crotzer and Levetin (171) examined the airborne concen-
tration of smut teliospores in the Tulsa, Oklahoma, atmos-
phere during 1991 and 1992 and attempted to identify the
most abundant smut spores. This study showed that telio-
spores occurred in the atmosphere on 100% of the days
from May through October in 1991 and 1992. The average
daily concentration generally ranged from 100 to 1,000 telio-
spores/m3 during both seasons with the mean for the 1991
season at 291 spores/m3 and the peak at 1,874 spores/m3,
which occurred on 5 July. The mean for the 1992 season
was 356 spores/m3, with the peak of 5,906 spores/m3 on 12
May. Although consistently present in the atmosphere during
the period studied, the concentration of these spores had
many fluctuations due to daily variations in wind speed, pre-
cipitation, and relative humidity as well as phenologies of the

host species and pathogens. Different species of smuts were
observed at different times of the year. During May and
June, prevalent teliospores included those species that infect
Bermuda grass (U. cynodontis), Johnson grass (Sphacelotheca
occidentalis), oat (U. kolleri), and wheat (U. tritici). Smuts
identified during September and October included U. maydis,
which infects corn, and U. brumivora and U. bullata, which
are pathogenic to several native Oklahoma grasses (31).
Other data show that airborne smut spores in Tulsa are not
limited to the May–October period but are also prevalent ear-
lier in spring and later in the fall (116). Although some of the
smuts identified in Tulsa were pathogens of native grasses,
many others were important crop pathogens.
In addition to the significance of smut fungi as plant
pathogens, it is recognized that smut teliospores could serve
as potential aeroallergens because the atmosphere is often
saturated with these spores for extended periods of time. In
the 1998 report from the air sampling network of the Amer-
ican Academy of Allergy, Asthma, and Immunology, smut
spores were among the top 10 spore types, in terms of atmos-
pheric concentration, reported from 44 out of 46 sampling
stations that provide counts for fungal spores. At 28 of these
stations, smut spores were in the top five (127). A few smut
extracts are routinely used for diagnosis and desensitization;
however, the full extent of allergenicity and the clinical sig-
nificance of various smut species are not known (171). At
the 44 monitoring stations, which are scattered across the
United States, smut spores represent a significant component
of the air spora and may represent a significant source of
pathogen inoculum (127).
Emerging Plant Pathogens and Biological Warfare
Historically, long-distance transport has introduced patho-
gens into new areas with devastating results (12, 172). Late
blight of potato, Dutch elm disease, and chestnut blight are
well-known examples from the 19th and early 20th centuries.
In recent decades other newly evolved pathogens, new and
virulent races of known pathogens, or pathogens with
expanded distributions have been the focus of much research
and reviews (12, 172–174). The Ug99 lineage of P. graminis
and P. pachyrhizi, which have been described already, are
two examples of emerging fungal pathogens; however, other
diseases caused by invasive viral, bacterial, and oomycete
pathogens are also worthy of mention.
A number of viral diseases have emerged in the past two
decades that have limited vegetable production in temperate,
subtropical, and tropical areas (58). Many of these are diseases
caused by begomoviruses and transmitted by white flies. Gen-
erally, these are existing diseases that have spread to new areas
or have developed the ability to infect new hosts. Tomato yel-
low leaf curl disease is an example of a viral disease that is now
widespread and is one of the most damaging viral diseases
affecting tomatoes. It is caused by a complex of 15 viral spe-
cies, which also affect other crops, ornamentals, and weeds.
This disease was first described in the Jordan Valley in 1939
and has affected tomato growing in the Middle East since
the 1970s. Today one of the viruses in the complex, tomato
yellow leaf curl virus has been report from the Mediterranean
region, the Far East, Australia, North America, and Central
America. It is believed that the global spread of this
virus began in the Mediterranean region in the 1990s (58).
The virus entered the United States in south Florida in
1997 and spread rapidly. It is believed that infected plant
material was the source, although the origin was never iden-
tified (175).
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-13
Citrus canker is a bacterial disease of citrus trees caused by
Xanthomonas citri subspecies citri (syn. Xanthomonas axonopo-
dis pv. citri). It is a devastating disease of citrus plants through-
out the tropical and subtropical growing areas with significant
economic impact (172, 174, 176). The bacteria are easily dis-
persed by rain splash and wind, and the presence of the disease
in citrus growing areas results in immediate quarantine follow-
ing national and international regulations (176). The disease
is thought to have originated in Southeast Asia and now
occurs in more than 30 countries (172). The first introduc-
tion into the United States was in 1910 in Florida on citrus
plants from Japan. It spread through Florida and other south-
eastern states and was finally eradicated in the 1930s (176).
Subsequent introductions to Florida occurred in the late
1980s and early 1990s. Eradication efforts destroyed millions
of citrus trees; however, additional outbreaks in 1994 and
1995 led to opposition of eradication measures and legal chal-
lenges. Meanwhile, tropical storms and hurricanes resulted in
disease spread through the citrus-growing areas of the state.
The disease was declared endemic in 2006 and eradication
efforts halted. Today citrus canker is controlled through the
use of voluntary eradication, resistant varieties, and frequent
bactericidal spraying (174).
Phytophthora ramorum is an oomycete pathogen that causes
sudden oak death (177). The disease was first described in
mid-1990s on ornamental Rhododendron species in Germany
and on oak and tanoak in California. By 2012 the known host
range was 100 species of woody plants in 40 genera. This has
resulted in destruction of forest ecosystems in California and
southern Oregon and the British Isles, along with economic
devastation in plant nurseries in the United States and
Europe. Dispersal of the P. ramorum sporangia by rain splash
or wind-driven rain is responsible for disease spread within a
forest. Long-distance dispersal of the pathogen has been
through the shipment of contaminated plants from nurseries.
Continued spread of the pathogen by both mechanisms is
likely (177).
With present-day international travel and trade, contami-
nated plants or contaminated materials can be accidentally
introduced from one country to another even when inspec-
tions occur at borders. As a result invasive plant diseases
will continue to plague crops throughout the world. Early
detection, rapid response, and education are necessary steps
to manage emerging pathogens (174).
In addition to the accidental introduction of pathogens
into new areas through natural processes or the human-
assisted movement of contaminated materials, the deliberate
introduction of pathogens has also been an area of concern for
many years. Fungal spores that attack crop plants have a long
history as components of the biological warfare arsenal in the
United States and other countries. The safety of our food
crops has been the focus of a number of articles suggesting
that even modest outbreaks of diseases targeting any major
food crop could severely hurt the economy in the United
States or other countries (174, 178, 179).
In 1969 the United States renounced the use of weapons
of biological warfare. This helped lay the groundwork for the
1972 Biological and Toxin Weapons Convention, which was
signed by 141 countries. The signatories were required to stop
work on biological and toxic weapons and destroy their stock-
piles. Before the unilateral termination of the U.S. offensive
biological weapons program, a substantial effort to develop
these weapons, including anticrop weapons, existed (179).
Between 1951 and 1969 the U.S. program stockpiled more
than 30,000 kg of stem rust spores (P. graminis f. sp. tritici),
which infect wheat, and nearly a ton of Pyricularia oryzae
3.2.8-14 ▪ AIR

spores to be used as an antirice weapon. Also included in the
arsenal were propagules of pathogens causing late blight of
potato and sclerotium rot, which attacks soybeans, sugar
beets, sweet potatoes, and cotton. In addition to these stock-
piles, weapon systems were developed for delivery (179, 180).
Awareness of anticrop weaponry has increased, possibly
fostered by revelations that Iraq actively pursued a biological
weapons program in the late 1980s. Iraqi scientists had exam-
ined the potential use of five bacterial strains, five viruses, one
fungal plant pathogen, and four toxins, two of which were
mycotoxins (181). The plant pathogen on which they focused
was a Tilletia sp., the causal agent of wheat covered smut (178).
That research was assumed to be aimed at destroying the wheat
crop of neighboring Iran (179, 181). The Iraqi program pro-
duced Tilletia spores, infected fields near the town of Mosul,
then harvested the wheat as a stockpile of spores. No attempt
to recover the Tilletia sp. was made, and the infected crop was
destroyed in 1990 (181); however, the susceptibility of eco-
nomic resources remains a significant threat (179).
This chapter has discussed many plant pathogens that can
disperse to hosts at various distances and can cause extensive
damage and economic loss; thus, the focus on plant pathogens
as anticrop weapons is understandable. Dispersal by the aero-
biological pathway has been depicted in several articles

TABLE 2 Examples of online plant disease forecasting systems

Forecast system Pathogen URL
Fusarium Head Blight Fusarium http://www.wheatscab.psu.edu/
Prediction Center graminearum
ipmPIPE (Integrated Pest http://www.ipmpipe.org/
Management—Pest
Information Platform for
Extension and Education)
Soybean rust Phakopsora
pachyrhizi
Cucurbit downy mildew Pseudoperonospora cubensis
Michigan State University
Potato late blight Phytophthora infestans http://www.lateblight.org/forecasting.php
Weather monitoring
North Dakota State University
Fusarium head blight Fusarium graminearum http://www.ag.ndsu.nodak.edu/cropdisease
Tan spot Pyrenophora triticirepentis http://www.ag.ndsu.nodak.edu/cropdisease
Stagonospora blotch Septoria tritici http://www.ag.ndsu.nodak.edu/cropdisease
Wheat leaf rust Puccinia triticina http://www.ag.ndsu.nodak.edu/cropdisease
Potato late blight Phytophthora infestans http://ndawn.ndsu.nodak.edu/potatolb-form.html
Sclerotinia stem rot of canola Sclerotinia sclerotiorum http://www.ag.ndsu.edu/sclerotinia/
Norway Potato Late Blight Forecasting Phytophthora infestans http://www.vips-landbruk.no/
Oklahoma State University
Pecan scab model Cladosporium caryigenum http://agweather.mesonet.ou.edu/models/newpecanscab/
default.php
Peanut leaf spot model Cercospora arachidícola http://www.mesonet.org/index.php/agriculture/category/
crop/peanut/leaf_spot_advisor
Spinach white rust Albugo occidentalis http://www.mesonet.org/index.php/agriculture/category/
horticulture/spinach/white_rust_advisor
Oregon State University
Tomato-potato Smith late blight risk Phytophthora infestans http://uspest.org/risk/tom_pot_map
UN Food and Agriculture Organization
Rust SPORE Pucinnia graminis Ug99 lineage http://www.fao.org/agriculture/crops/rust/stem/en/
University of California, Davis Phytophthora infestans http://www.ipm.ucdavis.edu/DISEASE/DATABASE/
potatolateblight.html
Disease Model Database
Potato late blight
University of Idaho
Potato late blight Phytophthora infestans http://www.uiweb.uidaho.edu/ag/plantdisease/lbhome.htm
Western Australia http://www.agric.wa.gov.au/PC_92989.html
Department of Agriculture
Canola blackleg Leptosphaeria maculans
Cereal rust disease Puccinia graminis Puccinia triticina
Puccinia striiformis

describing scenarios of bioterrorist attacks (179, 182), and
long-distance transport acts as a natural delivery system for bio-
logical particles that severely impact plant or human health.
One focus of modern phytopathological research is
increasingly on the development and use of dispersal models
to track pathogen outbreaks and meteorological conditions
conducive to the spread of a pathogen. These efforts must
continue. Understanding the mechanisms that transport
pathogens into a given area and then establishing prophylac-
tic measures would enhance efforts to protect crops against
emerging pathogens and terrorist attacks.
Disease Forecasting and Information Technology
A variety of epidemiological models have been developed to
predict spore dispersal and disease gradients (6, 8–11, 21, 22,
31, 55, 109, 121, 174, 183–186). While many of these models
provide useful information to plant pathologists and epidemi-
ologists, they are of little direct use to the average farmer. The
growers need practical applications of these models in the
form of a forecast. Disease forecasting for various agricultural
pathogens has been ongoing for a number of years. The ulti-
mate goal of forecasting is to warn growers when disease out-
break is likely to occur so that control measures can be used. A
further goal is to reduce costly and environmentally hazardous
pesticide application during periods when disease occurrence
is not likely.
In recent years, the widespread availability of personal
computers and the growth of the Internet have created a rev-
olution in information technology that has spread to all areas,
including plant pathology and epidemiology (187). The
innovations in technology have made plant disease forecasts
more widely available online. As a result, information is often
in the hands of the grower within minutes. Some forecasting
systems are supported by aerobiological data, some are sup-
ported by simulations and experimental data, and others
rely solely on meteorological data. These websites translate
many of the disease forecasting models into “user-friendly”
applications including direct access on smartphones. Many
sites use a multimedia approach to provide forecasts as well
as education about the pathogen and symptoms on infected
plants. Table 2 provides a small sampling of websites that pro-
vide disease forecasts in various parts of the world. (It should
be noted that websites change frequently. The URLs listed in
Table 2 are those currently available; however, it may be nec-
essary to use a search engine if the listed URL does not con-
nect to the site.)
Social media is also being utilized for distributing informa-
tion on pathogens. The PNW (Pacific North West) Plant
Disease Management Facebook page provides updated infor-
mation on disease occurrence in the northwest United States.
The Facebook page is a supplement to the Plant Disease Man-
agement Handbook that is published by Oregon State Univer-
sity, Washington State University, and the University of
Idaho. Comments with questions and photos from growers
and homeowners are posted with responses from the PNW
staff. The Texas Plant Disease Diagnostic Lab from the
Department of Plant Pathology and Microbiology at Texas
A&M University provides similar information for the south
central states.
SUMMARY
Airborne fungal spores and other plant disease propagules are
always present in the natural environment and pose a con-
stant threat to food crops. With the rapidly expanding world
population, diseases that threaten the food supply could result
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-15
in widespread starvation. World food production must con-
tinue to increase in the coming decades to keep pace with
the projected population increases. For food production to
increase, control of plant disease is essential. Aerobiological
studies must be part of any effort to understand the distribu-
tion and epidemiology of agricultural pathogens that rely on
air currents for dispersal.
Many aerobiological studies have been conducted on the
local or long-distance spread of individual plant pathogens.
Some of these studies are performed to develop mathematical
or computer models for forecasting disease epidemics and
predicting their onset and severity. The online availability
of disease forecasts is a major step in quickly putting the infor-
mation in the hands of the users.
For other pathogens, the lack of accurate spore dispersal
data has often limited the development of dispersal models.
These models require an understanding of all stages of disper-
sal, from spore takeoff to transport to deposition within the
crop as well as knowledge of the percentage of the spores
that escape local deposition and the percentage that survive
long-distance transport. The sensitivity of molecular techni-
ques is improving pathogen detection and measurement and
may help answer some of the questions about long-distance
transport.
Air sampling in the plant pathology community has gen-
erally been limited to individual studies for individual patho-
gens. There is no general monitoring for diverse pathogens.
Although the importance of the Puccinia Pathway for the dis-
persal of P. graminis uredospores is well known, there is no net-
work of monitoring stations that could provide information
on the advance of the pathogen or which other organisms
use the same aerobiological pathway. Such networks exist
in the aeroallergen research community. Because many spores
from phytopathogens are also well-known allergens, some of
these data are available. An important objective for the future
should be greater coordination among scientists interested in
atmospheric transport.
REFERENCES
1. Aylor DE. 1990. The role of intermittent wind in the disper-
sal of fungal pathogens. Ann Rev Phytopathol 28:73–92.
2. Campbell CL. 1999. The importance of dispersal mecha-
nisms in the epidemiology of Phytophthora blights and
downy mildews on crop plants. Ecosyst Health 5:146–157.
3. Pedgley DE. 1991. Aerobiology: the atmosphere as a source
and sink for microbes, p 43–59. In Andrews JH, Hirano SS
(eds), Microbial ecology of leaves. Springer, New York.
4. Waggoner PE, Aylor DE. 2000. Epidemiology: a science of
patterns. Ann Rev Phytopathol 38:71–94.
5. Zadoks J, Van den Bosch F. 1994. On the spread of plant
disease: a theory on foci. Ann Rev Phytopathol 32:503–521.
6. Van den Bosch F, Metz JAJ, Zadoks JC. 1999. Pandemics
of focal plant disease, a model. Phytopathology 89:495–505.
7. Heesterbeek JAP, Zadoks JC. 1987. Modelling pandemics
of quarantine pests and diseases: problems and perspectives.
Crop Prot 6:211–221.
8. Xu X-M, Ridout M. 1998. Effects of initial epidemic con-
ditions, sporulation rate, and spore dispersal gradient on the
spatio-temporal dynamics of plant disease epidemics. Phy-
topathology 88:1000–1012.
9. Aylor DE. 1999. Biophysical scaling and the passive dis-
persal of fungus spores: relationship to integrated pest man-
agement strategies. Agr For Meteorol 97:275–292.
10. Aylor DE. 2002. Aerobiology of fungi in relation to cap-
ture and release by plants, p 341–361. In Lindow SE,
3.2.8-16 ▪ AIR
Hecht-Poinar EI, Elliott VJ (eds), Phyllosphere microbiology.
APS Press, St. Paul, MN.
11. de Jong MD, Bourdot GW, Powell J, Goudriaan J. 2002.
A model of the escape of Sclerotinia sclerotiorum ascospores
from pasture. Ecol Model 150:83–105.
12. Brown JKM, Hovmoller MS. 2002. Aerial dispersal of
pathogens on the global and continental scales and its
impact on plant disease. Science 297:537–541.
13. Nagarajan S, Singh DV. 1990. Long-distance dispersion of
rust pathogens. Annu Rev Phytopathol 28:139–153.
14. Griffin DW, Garrison VH, Herman JR, Shinn EA. 2001.
African desert dust in the Caribbean atmosphere: microbi-
ology and public health. Aerobiologia 17:203–213.
15. Griffin DW, Kellogg CA, Garrison VH, Lisle JT, Borden
TC, Shinn EA. 2003. Atmospheric microbiology in the
northern Caribbean during African dust events. Aerobiolo-
gia 19:143–157.
16. Mims SA, Mims FM. 2004. Fungal spores are transported
long distances in smoke from biomass fires. Atmos Environ
38:651–655.
17. Schneider RW, Hollier CA, Whitam HK, Palm ME,
McKemy JM, Hernandez JR, Levy L, DeVries-Paterson
R. 2005. First report of soybean rust caused by Phakopsora
pachyrhizi in the continental United States. Plant Dis 89:
774.
18. Isard SA, Gage SH, Comtois P, Russo JM. 2005. Princi-
ples of the atmospheric pathway for invasive species applied
to soybean rust. BioScience 55:851–861.
19. Pan Z, Yang XB, Pivonia S, Xue L, Pasken R, Roads J.
2006. Long-term prediction of soybean rust entry into the
continental United States. Plant Dis 90:840–846.
20. Holmes G, Main C, Keever ZIII, 2004. Cucurbit downy mil-
dew: a unique pathosystem for disease forecasting, p 69–80. In
Spencer-Phillips P, Jeger M (eds), Advances in downy mildew
research. Kluwer Academic, Dordrecht, The Netherlands.
21. Yao C, Pal Arya S, Davis J, Main CE. 1997. A numerical
model of the transport and diffusion of Peronospora tabacina
spores in the evolving atmospheric boundary layer. Atmos
Environ 31:1709–1714.
22. Aylor DE. 2003. Spread of plant disease on a continental
scale: role of aerial dispersal of pathogens. Ecology 84:
1989–1997.
23. Cox CS. 1987. The aerobiological pathway of microorganisms.
John Wiley & Sons, New York.
24. Gregory PH. 1973. The microbiology of the atmosphere, 2nd
ed. Halstead Press, New York.
25. Burch M, Levetin E. 2002. Effects of meteorological con-
ditions on spore plumes. Int J Biometeorol 46:107–117.
26. Lacey J. 1991. Aerobiology and health: the role of airborne
fungal spores in respiratory disease, p 157–185. In Hawks-
worth DL (ed), Frontiers in mycology. CAB International,
Kew, UK.
27. Frinking HD. 1993. A historical perspective on aerobiol-
ogy as a discipline. In Isard SA (ed), Alliance for aerobiology
research workshop report. Illinois Natural History Survey,
Champaign, IL.
28. Govi G. 1992. Aerial diffusion of phytopathogenic fungi.
Aerobiologia 8:84–93.
29. McCartney HA, Fitt BDL. 1998. Dispersal of foliar fungal
plant pathogens: mechanisms, gradients and spatial pat-
terns. In Jones DG (ed), The epidemiology of plant disease.
Kluwer, Dordrecht, The Netherlands.
30. Burge H, Muilenberg M, Chapman J. 1991. Crop plants as
a source of fungus spores of medical importance, p 223–235.
In Andrews J, Hirano S (eds), Microbial ecology of leaves.
Springer, New York.
31. Lacey J. 1996. Spore dispersal: its role in ecology and dis-
ease: the British contribution to fungal aerobiology. Mycol
Res 100:641–660.
32. Geagea L, Huber L, Sacbe I. 1999. Dry-dispersal and rain-
splash of brown (Puccinia recondita f. sp. tritici) and yellow
(P. striiformis) rust spores from infected wheat leaves
exposed to simulated raindrops. Plant Pathol 48:472–482.
33. Kendrick B. 2000. The fifth kingdom, 3rd ed. A Focus Text,
Newburyport, MA.
34. Fitt BDL, McCartney HA, Walklate PJ. 1989. The role of
rain in dispersal of pathogen inoculum. Ann Rev Phytopathol
27:241–270.
35. Huber L, Madden LV, Fitt BDL. 1998. Rain-splash and
spore dispersal: a physical perspective, p 348–370. In Jones
DG (ed), The epidemiology of plant diseases. Kluwer Aca-
demic, Dordrecht, The Netherlands.
36. Pielaat A, Madden L, Gort G. 1998. Spores splashing
under different environmental conditions: a modeling
approach. Phytopathology 88:1131–1140.
37. Rajasab AH, Chawda HT. 1994. Dispersal of the conidia
of Colletotrichum gloeosporioides by rain and the develop-
ment of anthracnose on onion. Grana 33:162–165.
38. Huber L, Gillespie TJ. 1992. Modeling leaf wetness in
relation to plant disease epidemiology. Ann Rev Phytopathol
30:553–577.
39. Hirano SS, Upper CD. 1990. Population biology and epi-
demiology of Pseudomonas syringae. Ann Rev Phytopathol
28:155–177.
40. Tucker SL, Talbot NJ. 2001. Surface attachment and pre-
penetration stage development by plant pathogenic fungi.
Ann Rev Phytopathol 39:385–417.
41. Lacey J, Venette J. 1995. Outdoor air sampling techniques,
p 407–471. In Cox CS, Wathes CM (eds), Bioaerosols hand-
book. Lewis, Boca Raton, FL.
42. Bock C, Parker P, Gottwald T. 2005. Effect of simulated
wind-driven rain on duration and distance of dispersal of
Xanthomonas axonopodis pv. Citri from canker-infected cit-
rus trees. Plant Dis 89:71–80.
43. Lovell D, Parker S, Hunter T, Welham S, Nichols A.
2003. Position of inoculum in the canopy affects the risk
of septoria tritici blotch epidemics in winter wheat. Plant
Pathol 53:11–21.
44. Paul P, El-Allaf S, Lipps P, Madden L. 2004. Rain splash
dispersal of Gibberella zeae within wheat canopies in Ohio.
Phytopathology 94:1342–1349.
45. Levetin E. 1995. Fungi, p 87–120. In Burge H (ed), Bioaer-
osols. Lewis, Boca Raton, FL.
46. Levetin E, Horner WE. 2002. Fungal aerobiology: expo-
sure and measurement, p 10–27. In Breitenbach M, Cra-
meri R, Lehrer SB (eds), Fungal allergy and pathogenicity.
Chem. Immunol., vol. 81. Krager, Basel.
47. Troutt C, Levetin E. 2001. Correlation of spring spore con-
centrations and meteorological conditions in Tulsa, Okla-
homa. Int J Biometeorol 45:64–74.
48. Deacon JW. 1997. Modern mycology, 3rd ed. Blackwell Sci-
ence, Oxford, UK.
49. Moore-Landecker E. 1990. Fundamentals of the fungi. Pren-
tice Hall, Englewood Cliffs, NJ.
50. Money NP. 1998. More g’s than the Space Shuttle: ballis-
tospore discharge. Mycologia 90:547–558.
51. Webster J, Davey RA, Turner JCR. 1989. Vapour as the
source of water in Buller’s drop. Mycol Res 93:297–302.
52. Levetin E. 1991. Identification and concentration of air-
borne basidiospores. Grana 30:123–128.
53. Moody SA, Newsham KK, Ayres PG, Paul ND. 1999.
Variation in the responses of litter and phylloplane fungi
to UV-B radiation (290–315 nm). Mycol Res 103:1469–
1477.
54. Sundin G. 2002. Ultraviolet radiation on leaves: its influ-
ence on microbial communities and their adaptations, p
27–42. In Lindow S, Hecht-Poinar E, Elliott V (eds), Phyl-
losphere microbiology. APS Press, St. Paul, MN.

55. Agrios G. 2005. Plant pathology. Elsevier Academic Press,
Boston, MA.
56. Ng JC, Perry KL. 2004. Transmission of plant viruses by
aphid vectors. Mol Plant Pathol 5:505–511.
57. Gutiérrez S, Michalakis Y, Munster M, Blanc S. 2013.
Plant feeding by insect vectors can affect life cycle, popula-
tion genetics and evolution of plant viruses. Funct Ecol 27:
610–622. doi:10.1111/1365-2435.12070.
58. Navas-Castillo J, Fiallo-Olivé E, Sánchez-Campos S.
2011. Emerging virus diseases transmitted by whiteflies.
Ann Rev Phytopathol 49:219–248.
59. Miller WA, Rasochova L. 1997. Barley yellow dwarf
viruses. Ann Rev Phytopathol 35:167–190.
60. Strange RN, Scott PR. 2005. Plant disease: a threat to
global food security. Ann Rev Phytopathol 43:83–116.
61. Irwin ME, Thresh JM. 1990. Epidemiology of barley yel-
low dwarf: a study in ecological complexity. Ann Rev Phy-
topathol 28:393–424.
62. Hall G, Peters J, Little D, Power A. 2010. Plant commun-
ity diversity influences vector behaviour and barley yellow
dwarf virus population structure. Plant Pathol 59:1152–
1158.
63. Miller WA, Liu S, Beckett R. 2002. Barley yellow dwarf
virus: Luteoviridae or Tombusviridae? Mol Plant Pathol 3:
177–183.
64. Plumb RT. 2002. Viruses of Poaceae: a case history in plant
pathology. Plant Pathol 51:673–682.
65. Fokunang CN, Beynon JL, Watson KA, Battey NH,
Dunwell JM, Tembe-Fokunang EA. 2004. Advancement
in genetic modification technologies towards disease resist-
ance and food crop production. Biotechnology 3:1–20.
66. Verreault D, Moineau S, Duchaine C. 2008. Methods for
sampling of airborne viruses. Microbiol Mol Biol Rev 72:
413–444.
67. Whon TW, Kim M-S, Roh SW, Shin N-R, Lee H-W, Bae
J-W. 2012. Metagenomic characterization of airborne viral
DNA diversity in the near-surface atmosphere. J Virol 86:
8221–8231.
68. Cooper JI, Kelley SE, Massalski PR. 1988. Virus-pollen
interactions. Adv Dis Vector Res 5:221–249.
69. Mink GI. 1993. Pollen and seed-transmitted viruses and
viroids. Ann Rev Phytopathol 31:375–402.
70. Gottwald TR, Gibson GJ, Garnsey SM, Irey M. 1999.
Examination of the effect of aphid vector population com-
position on the spatial dynamics of citrus tristeza virus
spread by stochastic modeling. Phytopathology 89:603–608.
71. Hibino H. 1996. Biology and epidemiology of rice viruses.
Annu Rev Phytopathol 34:249–274.
72. Barras F, van Gijsegem F, Chatterjee AK. 1994. Extracel-
lular enzymes and pathogenesis of soft-rot Erwinia. Ann Rev
Phytopathol 32:201–234.
73. Romantschuk M. 1992. Attachment of plant patho-
genic bacteria to plant surfaces. Ann Rev Phytopathol 30:
225–243.
74. Salmond GP. 1994. Secretion of extracellular virulence
factors by plant pathogenic bacteria. Ann Rev Phytopathol
32:181–200.
75. Mansfield J, Genin S, Magori S, Citovsky V, Sriariyanum
M, Ronald P, Dow M, Verdier V, Beer SV, Machado MA,
Toth I, Salmond G, Foster GD. 2012. Top 10 plant patho-
genic bacteria in molecular plant pathology. Mol Plant
Pathol 13:614–629.
76. Eichenlaub R, Gartemann K-H. 2011. The Clavibacter
michiganensis subspecies: molecular investigation of Gram-
positive bacterial plant pathogens. Ann Rev Phytopathol 49:
445–464.
77. Francis I, Holsters M, Vereecke D. 2010. The Gram-
positive side of plant-microbe interactions. Environ Micro-
biol 12:1–12.
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-17
78. Conn KE, Ogawa JM, Manji BT, Adaskaveg JE. 1995.
Leuconostoc mesenteroides subsp. mesenteroides, the first
report of a coccoid bacterium causing a plant disease. Phy-
topathology 85:593–599.
79. Bock CH, Cook AZ, Parker PE, Gottwald TR, Graham
JH. 2012. Short distance dispersal of splashed bacteria of
Xanthomonas citri subsp. citri from canker-infected grape-
fruit tree canopies in turbulent wind. Plant Pathol 61:
829–836.
80. Marthi B. 1994. Resuscitation of microbial bioaerosols, p
192–225. In Lighthart B, Mohr AJ (eds), Atmospheric
microbial aerosols: theory and applications. Chapman and
Hall, New York.
81. Franc GD. 1994. Atmospheric transport of Erwinia caroto-
vora. Presented at Proc. 21st Conf. on Agricultural and For-
est Meteorology and 11th Conf. on Biometeorology and
Aerobiology, March 1994, San Diego, CA.
82. Barak JD, Schroeder BK. 2012. Interrelationships of food
safety and plant pathology: the life cycle of human patho-
gens on plants. Ann Rev Phytopathol 50:241–266.
83. Berger CN, Sodha SV, Shaw RK, Griffin PM, Pink D,
Hand P, Frankel G. 2010. Fresh fruit and vegetables as
vehicles for the transmission of human pathogens. Environ
Microbiol 12:2385–2397.
84. Brandl MT. 2006. Fitness of human enteric pathogens on
plants and implications for food safety. Ann Rev Phytopathol
44:367–392.
85. Critzer FJ, Doyle MP. 2010. Microbial ecology of food-
borne pathogens associated with produce. Curr Opin Bio-
tech 21:125–130.
86. Lynch MF, Tauxe RV, Hedberg CW. 2009. The growing
burden of foodborne outbreaks due to contaminated fresh
produce: risks and opportunities. Epidemiol Infect 137:
307–315.
87. Teplitski M, Barak JD, Schneider KR. 2009. Human
enteric pathogens in produce: un-answered ecological
questions with direct implications for food safety. Curr
Opin Biotech 20:166–171.
88. Fravel DR. 2005. Commercialization and implementation
of biocontrol. Ann Rev Phytopathol 43:337–359.
89. Haas D, Keel C. 2003. Regulation of antibiotic production
in root-colonizing Pseudomonas spp. and relevance for bio-
logical control of plant disease. Ann Rev Phytopathol 41:
117–153.
90. Janisiewicz WJ, Korsten L. 2002. Biological control of
postharvest diseases of fruits. Ann Rev Phytopathol 40:
411–441.
91. Paulitz TC, Belanger RR. 2001. Biological control in
greenhouse systems. Ann Rev Phytopathol 39:103–133.
92. Stockwell V, Johnson K, Loper J. 2002. Biological control
of fire blight: understanding interactions among introduced
and indigenous microbial communities, p 225–239. In
Lindow S, Hecht-Poinar E, Elliott V (eds), Phyllosphere
microbiology. APS Press, St. Paul, MN.
93. Chattopadhyay A, Bhatnagar N, Bhatnagar R. 2004. Bac-
terial insecticidal toxins. Crit Rev Microbiol 30:33–54.
94. de Maagd RA, Bravo A, Berry C, Crickmore N, Schnepf
HE. 2003. Structure, diversity, and evolution of protein
toxins from spore-forming entomopathogenic bacteria.
Ann Rev Genet 37:409–433.
95. Schnepf E, Crickmore N, Van Rie J, Lereclus D, Baum J,
Feitelson J, Zeigler DR, Dean DH. 1998. Bacillus thurin-
giensis and its pesticidal crystal proteins. Microbiol Mol
Biol Rev 62:775–806.
96. Atlas RM, Bartha R. 1998. Microbial ecology: fundamentals
and applications. Benjamin Cummings Science Publishing,
Menlo Park, CA.
97. Chen J, Jacobson L, Handelsman J, Goodman R. 1996.
Compatibility of systemic acquired resistance and microbial
3.2.8-18 ▪ AIR
biocontrol for suppression of plant disease in a laboratory
assay. Mol Ecol 5:73–80.
98. Paul B, Chereyathmanjiyil A, Masih I, Chapuis L, Benoit
A. 1998. Biological control of Botrytis cinerea causing grey
mould disease of grapevine and elicitation of stilbene phy-
toalexin (resveratrol) by a soil bacterium. FEMS Microbiol
Lett 165:65–70.
99. Umesha S, Dharmesh SM, Shetty SA, Krishnappa M,
Shetty HS. 1998. Biocontrol of downy mildew disease of
pearl millet using Pseudomonas fluorescens. Crop Prot 17:
387–392.
100. Lindow SE, Suslow TV. 2003. Temporal dynamics of the
biocontrol agent Pseudomonas fluorescens strain A506 in
flowers in inoculated pear trees. Phytopathology 93:727–
737.
101. Levesque CA. 2011. Fifty years of oomycetes—from con-
solidation to evolutionary and genomic exploration. Fungal
Divers 50:35–46.
102. Niederhauser JS. 1993. International cooperation in
potato research and development. Ann Rev Phytopathol
31:1–25.
103. Garelik G. 2002. Taking the bite out of potato blight. Sci-
ence 298:1702–1704.
104. Fry WE, Goodwin SB, Matuszak JM, Spielman LJ, Mil-
groom MG, Drenth A. 1992. Population genetics and
intercontinental migrations of Phytophthora infestans. Ann
Rev Phytopathol 30:107–130.
105. Ponti I, Cavanni P. 1992. Aerobiology in plant protection.
Aerobiologia 8:94–101.
106. Song J, Bradeen JM, Naess SK, Raasch JA, Wielgus SM,
Haberlach GT, Liu J, Kuang H, Austin-Phillips S, Buell
CR, Helgeson JP, Jiang J. 2003. Gene RB cloned from
Solanum bulbocastanum confers broad spectrum resistance
to potato late blight. Proc. Nat. Acad. Sci 100:9128–9133.
107. Nowicki M, Foolad MR, Nowakowska M, Kozik EU.
2012. Potato and tomato late blight caused by Phytophthora
infestans: an overview of pathology and resistance breeding.
Plant Dis 96:4–17.
108. Johnson DA, Cummings TF, Abi Ghanem R, Alldredge
JR. 2009. Association of solar irradiance and days of precip-
itation with incidence of potato late blight in the semiarid
environment of the Columbia Basin. Plant Dis 93:
272–280.
109. Aylor DE, Fry WE, Mayton H, Andrade-Piedra JL. 2001.
Quantifying the rate of release and escape of Phytophthora
infestans sporangia from a potato canopy. Phytopathology
91:1189–1196.
110. Fry WE, Goodwin SB. 1997. Re-emergence of potato and
tomato late blight in the United States. Plant Dis 1:
1349–1357.
111. Fry WE, Mizubuti ES. 1998. Potato late blight, p 371–388.
In Jones DG (ed), The epidemiology of plant disease. Kluwer
Academic, Dordrecht, The Netherlands.
112. Lokossou AA, Rietman H, Wang M, Krenek P, van der
Schoot H, Henken B, Hoekstra R, Vleeshouwers VG,
van der Vossen EA, Visser RG. 2010. Diversity, dis-
tribution, and evolution of Solanum bulbocastanum late
blight resistance genes. Mol Plant Microbe Interact 23:
1206–1216.
113. Johnson GI. 1989. Peronospora hyoscyami de Bary: taxo-
nomic history, strains and host range, p 1–8. In McKeen
WE (ed), Blue mold of tobacco. APS Press, St. Paul, MN.
114. Main C, Davis J. 1989. Epidemiology and biometeorology
of tobacco blue mold, p 201–215. In McKeen WE (ed), The
blue mold of tobacco. APS Press, St. Paul, MN.
115. Wiglesworth MD, Nesmith WC, Schardl CL, Li D, Siegel
MR. 1994. Use of specific repetitive sequences in Perono-
spora tabacina for the early detection of the tobacco blue
mold pathogen. Phytopathology 84:425–430.
116. Levetin E, Dorsey K. 2006. Contribution of leaf surface
fungi to the air spora. Aerobiologia 22:3–12.
117. Bus VG, Rikkerink EH, Caffier V, Durel C-E, Plummer
KM. 2011. Revision of the nomenclature of the differential
host-pathogen interactions of Venturia inaequalis and
Malus. Ann Rev Phytopathol 49:391–413.
118. Stensvand A, Amundsen T, Semb L, Gadoury DM,
Seem RC. 1998. Discharge and dissemination of asco-
spores by Venturia inaequalis during dew. Plant Dis 82:
761–764.
119. Rossi V, Ponti I, Marinelli M, Giosue S, Bugiani R. 2001.
Environmental factors influencing the dispersal of Venturia
inaequalis ascospores in the orchard air. J Phytopathol 149:
11–19.
120. Aylor DE. 1995. Vertical variation of aerial concentration
of Venturia inaequalis ascospores in an apple orchard. Phyto-
pathology 85:175–180.
121. Aylor DE. 1998. The aerobiology of apple scab. Plant Dis
82:838–849.
122. Aylor DE. 1996. Comparison of the seasonal pattern of air-
borne Venturia inaequalis ascospores with the release poten-
tial of V. inaequalis ascospores from a source. Phytopathology
86:769–776.
123. Gadoury DM, Stensvand A, Seem RC. 1998. Influence of
light, relative humidity, and maturity of populations on dis-
charge of ascospores of Venturia inaequalis. Phytopathology
88:902–909.
124. Jones AL. 1998. Apple scab: role of environment in patho-
gen and epidemic development, p 389–403. In Jones DG
(ed), The epidemiology of plant diseases. Kluwer Academic,
Dordrecht, The Netherlands.
125. Sterling M, Rogers C, Levetin E. 1999. An evaluation of
two methods used for microscopic analysis of airborne fun-
gal spore concentrations from the Burkard spore trap. Aero-
biologia 15:9–18.
126. Rossi V, Ponti I, Marinelli M, Giosue S, Bugiani R. 1999.
Field evaluation of some models estimating the seasonal
pattern of airborne ascospores of Venturia inaequalis.
J Phytopathol 147:567–575.
127. American Academy of Allergy A, and Immunology
(AAAAI) 1999. 1998 Pollen and spore report. American
Academy of Allergy, Asthma, and Immunology, Milwau-
kee, WI.
128. Iglesias I, Rodríguez-Rajo F, Méndez J. 2007. Evaluation
of the different Alternaria prediction models on a potato
crop in A Limia (NW of Spain). Aerobiologia 23:27–34.
129. Rotem J. 1994. The genus Alternaria: biology, epidemiology,
and pathogenicity. APS Press, St. Paul, MN.
130. Vloutoglou I, Kalogerakis S. 2000. Effects of inoculum
concentration, wetness duration and plant age on develop-
ment of early blight (Alternaria solani) and on shedding of
leaves in tomato plants. Plant Pathol 49:339–345.
131. Escuredo O, Seijo MC, Fernández-González M, Iglesias
I. 2011. Effects of meteorological factors on the levels of
Alternaria spores on a potato crop. Int J Biometeorol 55:
243–252.
132. Leplat J, Friberg H, Abid M, Steinberg C. 2013. Survival
of Fusarium graminearum, the causal agent of Fusarium head
blight. A review. Agron Sust Devel 33:97–111.
133. Xu X, Nicholson P. 2009. Community ecology of fungal
pathogens causing wheat head blight. Annu Rev Phytopathol
47:83–103.
134. Osborne LE, Stein JM. 2007. Epidemiology of Fusarium
head blight on small-grain cereals. Int J Food Microbiol
119:103–108.
135. West JS, Holdgate S, Townsend JA, Edwards SG, Jen-
nings P, Fitt BD. 2012. Impacts of changing climate and
agronomic factors on fusarium ear blight of wheat in the
UK. Fungal Ecol 5:53–61.

136. Trail F. 2009. For blighted waves of grain: Fusarium grami-
nearum in the postgenomics era. Plant Physiol 149:103–
110.
137. Trail F, Xu H, Loranger R, Gadoury D. 2002. Physiolog-
ical and environmental aspects of ascospore discharge in
Gibberella zeae (anamorph Fusarium graminearum). Mycolo-
gia 94:181–189.
138. Kriss AB, Paul PA, Madden LV. 2010. Relationship
between yearly fluctuations in Fusarium head blight inten-
sity and environmental variables: a window-pane analysis.
Phytopathology 00:784–797.
139. Maldonado-Ramirez SL, Schmale DG, Shields EJ,
Bergstrom GC. 2005. The relative abundance of viable
spores of Gibberella zeae in the planetary boundary layer
suggests the role of long-distance transport in regional
epidemics of Fusarium head blight. Agr Forest Meteorol
32:20–27.
140. Schmale DG, Ross SD, Fetters TL, Tallapragada P,
Wood-Jones AK, Dingus B. 2012. Isolates of Fusarium gra-
minearum collected 40–320 meters above ground level
cause Fusarium head blight in wheat and produce tricho-
thecene mycotoxins. Aerobiologia 28:1–11.
141. Kriss AB, Paul PA, Xu X, Nicholson P, Doohan FM,
Hornok L, Rietini A, Edwards SG, Madden LV. 2012.
Quantification of the relationship between the environ-
ment and Fusarium head blight, Fusarium pathogen den-
sity, and mycotoxins in winter wheat in Europe. Eur
J Plant Pathol 133:975–993.
142. Schaafsma AW, Tamburic-Ilincic L, Hooker DC. 2005.
Effect of previous crop, tillage, field size, adjacent crop,
and sampling direction on airborne propagules of Gibberella
zeae/Fusarium graminearum, fusarium head blight severity,
and deoxynivalenol accumulation in winter wheat. Can
J Plant Pathol 27:217–224.
143. Brennan JM, Egan D, Cooke BM, Doohan FM. 2005.
Effect of temperature on head blight of wheat caused by
Fusarium culmorum and F. graminearum. Plant Pathol 54:
156–160.
144. Hodson D. 2011. Shifting boundaries: challenges for rust
monitoring. Euphytica 179:93–104.
145. Singh RP, Hodson DP, Huerta-Espino J, Jin Y, Bhavani
S, Njau P, Herrera-Foessel S, Singh PK, Singh S, Govin-
dan V. 2011. The emergence of Ug99 races of the stem rust
fungus is a threat to world wheat production. Ann Rev Phy-
topathol 49:465–481.
146. Hau B, de Vallavieille-Pope C. 1998. Wind-dispersed dis-
eases, p 323–347. In Jones DG (ed), The epidemiology of plant
diseases. Kluwer Academic, Dordrecht, The Netherlands.
147. Eversmeyer MG, Kramer CL. 2000. Epidemiology of
wheat leaf and stem rust in the central Great Plains of the
USA. Ann Rev Phytopathol 38:491–513.
148. Eversmeyer MG, Kramer CL. 1994. Survival of Puccinia
recondita and P. graminis urediniospores as affected by expo-
sure to weather conditions at one meter. Phytopathology 84:
332–334.
149. Eversmeyer MG, Kramer CL. 1995. Survival of Puccinia
recondita and P. graminis urediniospores exposed to tem-
peratures from subfreezing to 35 C. Phytopathology 85:161.
150. Visser B, Herselman L, Park RF, Karaoglu H, Bender
CM, Pretorius ZA. 2011. Characterization of two new
Puccinia graminis f. sp. tritici races within the Ug99 lineage
in South Africa. Euphytica 179:119–127.
151. Singh RP, Hodson DP, Huerta-Espino J, Jin Y, Njau P,
Wanyera R, Herrera-Foessel SA, Ward RW. 2008. Will
stem rust destroy the world’s wheat crop? Adv Agron 98:
271–309.
152. Hodson D, Grønbech-Hansen J, Lassen P, Alemayehu Y,
Arista J, Sonder K, Kosina P, Moncada P, Nazari K, Park
R. 2012. Tracking the wheat rust pathogens, p 11–22. In
3.2.8. Aerobiology of Agricultural Pathogens ▪ 3.2.8-19
McIntosh R (ed), Borlaug Global Rust Initiative 2012 techni-
cal workshop. Beijing, China.
153. Singh RP, Hodson DP, Jin Y, Huerta-Espino J, Kinyua
MG, Wanyera R, Njau P, Ward RW. 2006. Current status,
likely migration and strategies to mitigate the threat to
wheat production from race Ug99 (TTKS) of stem rust
pathogen. In CAB reviews: perspectives in agriculture, veteri-
nary science, nutrition and natural resources. doi:10.1079/
PAVSNNR20061054.
154. Hartman GL, Miles MR, Frederick RD. 2005. Breeding
for resistance to soybean rust. Plant Dis 89:664–666.
155. Carmona M, Gally M, Lopez S. 2005. Asian soybean rust:
incidence, severity, and morphological characterization of
Phakopsora pachyrhizi (Uredinia and Telia) in Argentina.
Plant Dis 89:109.2.
156. Hartman G, Hillz C, Twizeyimanaz M, Miles M, Bandyo-
padhyay R. 2012. Interaction of soybean and Phakopsora
pachyrhizi, the cause of soybean rust, p 59–72. In Hemmings
D (ed), Plant sciences reviews 2011. CAB International,
Cambridge, MA.
157. Li X, Esker PD, Pan Z, Dias AP, Xue L, Yang XB. 2010.
The uniqueness of the soybean rust pathosystem: an
improved understanding of the risk in different regions of
the world. Plant Dis 94:796–806.
158. Goellner K, Loehrer M, Langenbach C, Conrath U,
Koch E, Schaffrath U. 2010. Phakopsora pachyrhizi, the
causal agent of Asian soybean rust. Mol Plant Pathol 11:
169–177.
159. Yorinori J, Paiva W, Frederick R, Costamilan L, Bertag-
nolli P, Hartman G, Godoy C, Nunes Jr J. 2005. Epidem-
ics of soybean rust (Phakopsora pachyrhizi) in Brazil and
Paraguay from 2001 to 2003. Plant Dis 89:675–677.
160. Ploper LD, Gonzalez V, Galvez MR, de Ramallo NV,
Zamorano MA, Garcia G, Castagnaro AP. 2005. Detec-
tion of soybean rust caused by Phakopsora pachyrhizi in
Northwestern Argentina. Plant Dis 89:774.2.
161. Isard SA, Barnes CW, Hambleton S, Ariatti A, Russo
JM, Tenuta A, Gay DA, Szabo LJ. 2011. Predicting soy-
bean rust incursions into the North American continental
interior using crop monitoring, spore trapping, and aerobio-
logical modeling. Plant Dis 95:1346–1357.
162. Kim K-S, Unfried JR, Hyten DL, Frederick RD, Hart-
man GL, Nelson RL, Song Q, Diers BW. 2012. Molecular
mapping of soybean rust resistance in soybean accession PI
561356 and SNP haplotype analysis of the Rpp1 region in
diverse germplasm. Theor Appl Genet 125:1339–1352.
163. Hershman D, Sikora E, Giesler L. 2011. Soybean rust PIPE:
past, present, and future. J Integr Pest Manage 2:D1–D7.
164. Isard SA, Russo JM, Ariatti A. 2007. The Integrated
Aerobiology Modeling System applied to the spread of soy-
bean rust into the Ohio River valley during September
2006. Aerobiologia 23:271–282.
165. Vanky K. 2002. Illustrated genera of smut fungi. APS Press,
St. Paul, MN.
166. Brefort T, Doehlemann G, Mendoza-Mendoza A, Reiss-
mann S, Djamei A, Kahmann R. 2009. Ustilago maydis as
a pathogen. Annu Rev Phytopathol 47:423–445.
167. Halwagy MH. 1989. Seasonal airspora at three sites in
Kuwait 1977–1982. Mycol Res 93:208–213.
168. Hasnain SM, Fatima K, Al-Frayh A, Al-Sedairy ST.
2005. Prevalence of airborne basidiospores in three coastal
cities of Saudi Arabia. Aerobiologia 21:139–145.
169. LaMondia J. 2010. January temperatures predict tobacco
blue mold severity: evidence for local source and long-
distance transport of inoculum in Connecticut. Plant Dis
94:119–124.
170. Leonard KJ, Szabo LJ. 2005. Stem rust of small grains and
grasses caused by Puccinia graminis. Mol Plant Pathol 6:
99–111.
3.2.8-20 ▪ AIR
171. Crotzer V, Levetin E. 1996. The aerobiological signifi-
cance of smut spores in Tulsa, Oklahoma. Aerobiologia
12:177–184.
172. Anderson PK, Cunningham AA, Patel NG, Morales FJ,
Epstein PR, Daszak P. 2004. Emerging infectious diseases
of plants: pathogen pollution, climate change and agro-
technology drivers. Trends Ecol Evol 19:535–544.
173. Fisher MC, Henk DA, Briggs CJ, Brownstein JS, Madoff
LC, McCraw SL, Gurr SJ. 2012. Emerging fungal threats
to animal, plant and ecosystem health. Nature 484:
186–194.
174. Fletcher J, Luster D, Bostock R, Burans J, Cardwell K,
Gottwald T, McDaniel L, Royer M, Smith K. 2010.
Emerging infectious plant diseases, p 337–366. In Scheid
WM, Grayson ML, Hughes JM (eds), Emerging infectious
diseases. ASM Press, Washington, DC.
175. Polston JE, McGovern RJ, Brown LG. 1999. Introduction
of tomato yellow leaf curl virus in Florida and implications
for the spread of this and other geminiviruses of tomato.
Plant Dis 83:984–988.
176. Gottwald TR, Hughes G, Graham JH, Sun XS, Riley T.
2001. The citrus canker epidemic in Florida: the scientific
basis of regulatory eradication policy for an invasive species.
Phytopathology 91:30–34.
177. Grunwald NJ, Garbelotto M, Goss EM, Heungens K,
Prospero S. 2012. Emergence of the sudden oak death
pathogen Phytophthora ramorum. Trends Microbiol 20:
131–138.
178. Madden L, Wheelis M. 2003. The threat of plant patho-
gens as weapons against US crops. Ann Rev Phytopathol
41:155–176.
179. Rogers P, Whitby S, Dando M. 1999. Biological warfare
against crops. Sci Am 280:70–75.
180. Christopher GW, Cieslak TJ, Pavlin JA, Eitzen EM Jr.,
1997. Biological warfare. J Am Med Assoc 278:412–417.
181. Zilinskas RA. 1997. Iraq’s biological weapons: the past as
future? J Am Med Assoc 278:418–424.
182. Inglesby TV. 1999. Anthrax: a possible case history. Emerg
Infect Dis 5:556–560.
183. Hardwick NV. 1998. Disease forecasting. In Jones DG
(ed), The epidemiology of plant disease. Kluwer Academic,
Dordrecht, The Netherlands.
184. Paulitz T, Dutilleul P, Yamasaki S, Fernando W, Seaman
W. 1999. A generalized two-dimensional Gaussian model
of disease foci of head blight of wheat caused by Gibberella
zeae. Phytopathology 89:74–83.
185. Rosa M, Gozzini B, Orlandini S, Seghi L. 1995. A com-
puter program to improve the control of grapevine downy
mildew. Comput Electron Agr 12:311–322.
186. De Wolf ED, Isard SA. 2007. Disease cycle approach
to plant disease prediction. Ann Rev Phytopathol 45:
203–220.
187. Newton AC, Gaunt RE. 1998. Information technology
in epidemiology, p 278. In Jones DG (ed), The epidemiol-
ogy of plant diseases. Kluwer Academic, Dordrecht, The
Netherlands.
 Legionellae and Legionnaires’ Disease
CLARESSA E. LUCAS AND BARRY S. FIELDS
3.2.9
Bacteria of the family Legionellaceae are representatives
of a group of gammaproteobacteria that survive as intra-
cellular parasites or endosymbionts of free-living protozoa.
The Legionellaceae maintained their anonymity until 1977,
primarily because of their unique ecology. Initially, the bac-
terium Legionella pneumophila was labeled the Legionnaires’
disease bacterium. This nomenclature is somewhat mislead-
ing since colonizing a human to cause disease is an ecolog-
ical dead-end for the bacterium. Humans may have
coexisted with these bacteria for a very long time, but it is
only recent industrial technology that has provided these
organisms a means for causing infection in humans.
LEGIONNAIRES’ DISEASE
Legionnaires’ disease is a consequence of altering the envi-
ronment for human benefit. The illness continues to make
up a significant proportion of cases of pneumonia that occur
in the developed world. Legionnaires’ disease is the most
common etiologic agent of waterborne disease in drinking
water in the United States, where the incidence of disease
has more than doubled in the past decade (1, 2). Although
the disease can be effectively treated with appropriate antimi-
crobial agents, measures aimed at preventing transmission of
the bacteria from environmental sources to susceptible hosts
are varied and of questionable efficacy.
Legionnaires’ disease occurs when sufficient numbers of
Legionellae are aerosolized and subsequently inhaled by a suscep-
tible host (3). Legionnaires’ disease and Legionella have been
included in the aerobiology section of this manual because
the bacteria are transmitted to the host via aerosols. However,
because these bacteria are primarily found in freshwater envi-
ronments, many of the methods discussed are associated with
microbiological examination of water. Legionellae may be diffi-
cult to detect in the environment or clinically without special-
ized materials and methods. The fastidious nature of these
bacteria is primarily due to the fact that they replicate intracell-
ularly within single-cell protozoans in the environment or
alveolar macrophages in humans. Thus, the methods used to
detect Legionellae have evolved from research in both clinical
and environmental microbiology.
Legionellosis
Legionellae were first isolated and identified due to investiga-
tion of an explosive, common-source outbreak of pneumonia
doi:10.1128/9781555818821.ch3.2.9
3.2.9-1
among people attending an American Legion convention in
Philadelphia in 1976 (4, 5). The highly publicized original
reports of the epidemic identified 182 cases of disease, 29 of
which were fatal. Legionellae are now associated with two forms
of respiratory illness, collectively referred to as legionellosis (3)
Legionnaires’ disease is the pneumonic and more severe form
of illness (Table 1). The other form of respiratory illness is
named Pontiac fever after the first documented outbreak,
which occurred at a health department in Pontiac, Michigan
(6). Possible explanations for the manifestation of two disease
syndromes caused by the same bacteria include the inability of
some Legionellae to multiply in human tissue for reasons of vir-
ulence, host range, bacterial viability, or host susceptibility (6–
8). A few limited studies suggest Pontiac fever may be associ-
ated with exposure to bacterial endotoxin and not true intra-
cellular replication (9, 10).
Community-acquired pneumonia (CAP) incidence stud-
ies have estimated that there are between 8,000 and 18,000
cases of legionellosis annually in the United States, which
is approximately 25-fold higher than the number of cases
reported to the Centers for Disease Control and Prevention
(CDC) (6). In fact, the majority of cases of legionellosis are
sporadic, with less than 5% outbreak-related (1, 11). The
sources of CAP are often difficult to identify, because of
both the ubiquitous nature of the bacterium and the reluc-
tance of physicians to obtain respiratory specimens to match
with environmental isolates (12, 13). Although Legionellae
are extremely common in fresh water and building system
environments worldwide, they are only infrequently the
cause of disease because several conditions must be met to
infect a susceptible host (14–16). For disease to occur, Legion-
ellae must be present in an environmental reservoir, multiply
from low to high concentrations, and become disseminated in
a form that can be inhaled by susceptible individuals (3, 6).
This chain of causality serves as the basis for understanding
the epidemiology of legionellosis as well as development of
prevention strategies (17–19).
LEGIONELLACEAE AND THEIR ECOLOGY
Bacteria of the genus Legionella are Gram-negative, aerobic,
non–spore forming, and rod shaped. Cells are 0.3 to 0.9 by
1 to 20 μm and motile, with one or more polar flagella.
Legionellae use amino acids as their carbon and energy sources
and do not oxidize or ferment carbohydrates (20). As of this
3.2.9-2 ▪ AIR
TABLE 1 Characteristic of Legionnaires’ disease and Pontiac
fever
Disease
Characteristic Legionnaires’
Pontiac fever
disease
Illness Progressive Influenza-like illness
pneumonia, (nonpneumonic)
sometimes fatal
Incubation 2–10 days 36 h (mean)
period
Attack rate <5% >95%
Risk factors Cigarette smoking None known
Diabetes mellitus
Cancer
COPD
End-stage renal
disease
AIDS
writing there are at least 53 named species comprising approx-
imately 73 serogroups in genus Legionella, with 20 species
known to be associated with disease (21). However, environ-
mental surveillance studies frequently recover Legionellae that
cannot be conclusively identified by either serological or
molecular methods, so the number of named species in genus
Legionella is likely to increase as more novel isolates are
characterized. L. pneumophila serogroup 1 is responsible for
approximately 80% of all cases of legionellosis in most of
the developed world except for Australia and New Zealand,
where L. longbeachae may be the etiological agent for up to
50% of cases (1, 22). However, all species and serogroups
are considered potentially pathogenic, though infections
from Legionellae other than L. pneumophila serogroup 1 are
infrequently reported because they are rare and lack diagnos-
tic reagents (23). There also appear to be a number of uniden-
tified Legionellae which cannot be grown on routine Legionella
media. These bacteria have been given the acronym name
Legionella-like amoebal pathogens because they have been
detected through their ability to grow intracellularly in proto-
zoan cells (24).
The Ecology of Legionella
Water is the major reservoir for Legionellae. The bacteria are
found in freshwater environments worldwide, even in habi-
tats as inimical as Antarctic lakes and the Mir space station
(25, 26). Legionellae do not survive in dry environments, so
early outbreaks associated with construction were most likely
the results of massive descalement of the plumbing systems
due to changes in water pressure and not direct transmission
through dry soils (27, 28). However, transmission of
L. longbeachae from moist potting soils has recently been con-
firmed in some cases of sporadic Legionnaires’ disease but not
in epidemic proportions and only after direct contact with the
compost (29). Where conditions are consistently moist and
within permissive temperature ranges, such as building water
systems, Legionellae may persist for decades, occupying a reser-
voir in biofilms conducive to growth and resistant to attempts
at disinfection (30–32).
Initially, it was difficult to explain the pervasiveness of
Legionellae in aquatic environments. The bacteria are fastidi-
ous and require an unusual combination of nutrients in
bacteriologic medium (33). These levels of nutrients would
rarely be found in aquatic environments and, if present, would
amplify faster growing bacteria that would compete with the
Legionellae. However, the nutrients required by Legionellae
represent the need for an intracellular environment, parasitiz-
ing free-living protozoa, and not soluble nutrients commonly
found in fresh water as first described by Rowbotham (34).
Figure 1 shows L. pneumophila multiplying within the ciliated
protozoan Tetrahymena pyriformis. Legionellae have been
reported to multiply in 14 species of amoebas, 2 species of cili-
ated protozoa, and 1 species of slime mold while growth in the
absence of protozoa has only been documented on laboratory
media (35, 36). The host range of a given species of Legionella
may be determined by genetic factors that evolve quickly due
to horizontal transfer of foreign DNA (37, 38). Several studies
have described the relationship between Legionellae and pro-
tozoa in aquatic environments as well as the accidental nature
of Legionellae extending their host range to alveolar macro-
phages to cause disease in humans (3, 6, 39).
Infection of both protozoan and human phagocytic cells is
accomplished utilizing a novel type IVb secretion system
composed of 25 genes, designated the Dot/Icm system. The
Dot/Icm system delivers substrates that allow the bacteria to
subvert host cell processes for their own propagation. Over
200 different substrates are known or suspected to be translo-
cated by the Dot/Icm system, which vary considerably
between species and between strains of the same species
(40, 41). Deletion of any single or a few substrates does not
result in complete inability to replicate intracellularly but
does sometimes constrict the effective host range of the
mutant compared to its isogenic wild type (42, 43). Thus,
understanding the roles that protozoa play in the ecology of
Legionellae is critical to the development of successful preven-
tion strategies. To understand the ecology of Legionellae, the
bacteria must be considered in the context of their natural
microbial community and not as independent inhabitants
of freshwater environments.
Aerosol Transmission from Aquatic Environments
Inhalation of Legionellae in aerosolized droplets is the primary
means of transmission for legionellosis (3, 11, 17, 44). The
aerosolized droplets must be of a respirable size (between 1
and 5 μm). No person-to-person transmission of Legion-
naires’ disease has been documented. A number of devices
have been implicated as the most common sources of aerosol
transmission of Legionellae (Table 2). Such sources are of two
general types: those producing aerosols of contaminated pot-
able water, such as showers and faucets; and those from non-
potable water, such as cooling towers and whirlpool spas.
However, the source of transmission may not be in the imme-
diate vicinity of affected persons, generate only small amounts
of aerosols, or otherwise not fit the canonical sources most
often linked with legionellosis outbreaks (45, 46). Thus, effi-
cient identification of sources of transmission requires a
multidisciplinary approach that includes clinical epidemiol-
ogy, traditional microbiology, and molecular epidemiology
techniques.
DETECTION OF LEGIONELLAE IN THE
ENVIRONMENT
It is generally accepted that culturing environmental samples
for the presence of Legionella is necessary to establish the
source of an outbreak of Legionnaires’ disease, to determine
the efficacy of a disinfection program, or to evaluate the

FIGURE 1 Gimenez stain of the ciliated protozoan T. pyriformis infected with L. pneumophila. Chains of multiplying Legionella cells are
contained within vesicles, as observed in human phagocytic cells infected with L. pneumophila. Magnification, ×1,650.
doi:10.1128/9781555818821.ch3.2.9.f1
potential of a device to transmit the disease. Routine envi-
ronmental sampling can also be beneficial in institutions
housing persons at extremely high risk for acquiring Legion-
naires’ disease, such as solid organ transplant wards, with
the understanding that the objective is no detectable Legion-
ellae in the water system (18). However, recommendations
concerning the routine culturing of environmental samples
in the absence of documented cases of legionellosis represent
an area of considerable controversy for a variety of reasons.
Detection of Legionellae in an environmental source is not
necessarily evidence of the potential for disease because there
are little data to inform appropriate sampling methodology or
interpretation of results (47). In addition, the type of sample
collected, transport procedures, and variable laboratory sam-
ple handling processes can have a large effect on both the
number of positive sites identified and the number of Legion-
ellae recovered (14, 48–53). In fact variability approaching
three logs is not uncommon in even “standardized” samples
TABLE 2 Sources known to transmit Legionellae via aerosols
Type of water Transmitting device
Potable Showers
Tap water faucets
Respiratory therapy equipment
Nonpotable Cooling towers and evaporative condensers
Whirlpool spas
Decorative fountains
Ultrasonic mist machines
Humidifiers
Source: Modified from Bopp et al. (1981) Journal of Clinical Microbiology
13:714–719, with permission.
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-3
spiked with live bacteria or reconstituted from lyophilized
stock for proficiency testing purposes (54, 55).
Determining a scientifically based number of Legionella
that are acceptable in water systems would be valuable infor-
mation that could assist in control of the disease. Several
attempts have been made to address this issue by modeling
exposures after a review of scientific literature (56, 57). How-
ever, assessing the validity of such models faces considerable
ethical hurdles because test populations should not be know-
ingly exposed to various levels of Legionella to establish what is
“safe.” Differences in building design and siting, occupant sus-
ceptibility, and inherent virulence of various Legionella spe-
cies and strains would also add confounding variables to
any standard model. Nevertheless, action plans that propose
either a percent positivity of sites sampled or number of viable
Legionellae detected by culture as indicators of disease poten-
tial have been adopted and promulgated by some public
health authorities. To date, all numerical action plans are
derived from a table published by a commercial testing lab,
but there are no studies in the literature that directly measure
the efficacy this model (58). In contrast, percent positivity
action plans have accumulated more data since the original
study was performed in 1983, but the sensitivity (59%) and
specificity (74%) of the metric remain too low to be a useful
predictor of disease (47, 59). Continued scientific analysis is
needed to establish when and where testing for the presence
of Legionella may be beneficial to public health disease pre-
vention efforts.
Collection of Bulk Water and Biofilm Swab Samples
The number and types of sites that should be sampled to
detect Legionellae must be determined on an individual basis
due to the diversity of plumbing, heating, ventilation, and air
3.2.9-4 ▪ AIR
conditioning (HVAC) systems in use. Institutions that may be
sampled can include industrial facilities, hospitals, hotels,
retirement homes, public facilities, or domestic residences.
The CDC offers a downloadable environmental assessment
tool that may be useful for identifying sources to sample in
a variety of facilities (http://www.cdc.gov/legionella/files/
EnvironmentalAssessmentInstrument.pdf ). In general, any
water source that may be aerosolized should be considered a
potential source for the transmission of Legionellae. The bac-
teria may be present at extremely low concentrations in
municipal water supplies, but rarely detected. Instead, Legion-
ellae tend to colonize plumbing systems and point-of-use
(POU) devices where they multiply to detectable numbers.
Amplification of Legionellae requires temperatures above 25°
C, though lower temperatures tend to be bacteriostatic but
not bactericidal (20, 60). Legionellae do not survive drying,
so condensate from air-conditioning equipment that fre-
quently evaporates is not a likely source (27).
The two primary sample types collected are bulk water and
swabs of POU devices or system surfaces that contain biofilm
(61, 62). Legionella bacteria in potable water are not homoge-
neous, but may exist in clumps or dispersed single cells. Thus,
detection of Legionella in potable water often requires that
samples are concentrated to reduce sampling error. Testing
simulated potable water volumes between 0.25 and 1 liters
yielded similar results during proficiency testing, so collecting
volumes in this range is preferred (55, 63). Smaller volumes
of nonpotable water that may have higher organic loads,
such as water from evaporative condensers or whirlpool
spas, may be acceptable if sample concentration is not antici-
pated, but it is generally preferable to collect the largest vol-
ume possible.
It is recommended that swab samples of biofilms should
be taken in conjunction with water samples from accessible
sites such as the basins of cooling towers, spa water lines, fau-
cet aerators, or showerheads. Discordance in both positivity
and species presentation between bulk water and swab sam-
ples has been observed to range between 19% and 50%, so
a comprehensive survey of a plumbing or HVAC system
should include both types of samples (14, 48, 49). Cotton-
tipped swabs may inhibit recovery of Legionellae and should
not be used (64). Swabs should be taken from interior pipe
and device surfaces with the aerator or showerhead removed.
Swabs should be submerged in a small volume of water taken
from the same source to prevent drying during transport to the
laboratory.
Air Sampling
Examination of water samples is the most efficient microbio-
logic method for identifying sources of Legionellae. Air sam-
pling is an insensitive means of detecting these bacteria and
therefore is of limited value. In certain instances, it may be
helpful to demonstrate the presence of Legionellae in aerosol
droplets associated with suspected reservoirs of the bacterium.
Air sampling has been used to better define the roles of cer-
tain devices such as showers, faucets, and evaporative con-
densers in disease transmission or to document the distance
droplets may travel from source devices. It may also be used
to quantitate or determine the size of particles containing
Legionellae, but this approach requires stringent controls and
calibration (61, 65, 66). Samplers should be placed in loca-
tions representative of human exposure and investigators
should wear an Occupation Safety and Health Administra-
tion–approved respirator since sampling involves exposure
to potentially infectious aerosols.
The basic methods for sampling airborne bacteria include
impingement in liquids, impaction onto solid surfaces,
filtration, and precipitation. See Chapter 3.2.2 for a detailed
discussion of sampling techniques for airborne micro-
organisms. Methods that have been used to sample air for
Legionellae include impingement in liquid media using an
all-glass impinge (AGI), impaction of solid medium using
Andersen samplers, and the use of settle plates (61, 65, 66).
Except for settle plates, a nonquantitative sampling method,
these methods require a vacuum source and a means of
controlling air flow. Several configurations of air sampling
equipment can be used, that usually incorporate a device
for controlling airflow (flowmeter-manometer) connected
in the vacuum line between sampler and vacuum source. A
variety of hand-held solid impaction devices for use in the
field are commercially available.
AGI (AGI-30, Ace Glass, Vineland, NJ) use the principle
of impingement and washing of air, in which organisms are
entrapped in a liquid medium and have been successfully
used to sample for Legionellae (66). Because of the velocity
at which samples are collected, clumps tend to be fragmented,
leading to a more accurate count of bacteria present in the air.
The disadvantages of this method are that this velocity tends
to destroy some vegetative cells, it does not differentiate par-
ticle sizes, and the apparatus is easily broken in the field. Yeast
extract broth or buffered yeast extract broth are recommended
as the liquid medium for AGI-30 sampling of Legionellae and
may be processed by methods used for the culture of water
samples (61, 65, 66).
Six-stage Andersen samplers are viable particle samplers
in which particles pass through jet orifices of decreasing
size in a cascade fashion until they hit an agar surface. The
agar plates are then removed and incubated to culture any
Legionellae present. The stage distribution of the Legionellae
should indicate the extent to which the bacteria would
have penetrated the respiratory system. The advantages of
this sampling method are that the equipment is more durable,
the sampler can determine the number and size of droplets
containing Legionellae, and agar plates can be placed directly
in an incubator with no further manipulations. Disadvan-
tages include drying of artificial media and efficient collec-
tion of fungal spores, both of which can have a negative
impact on the recovery of viable Legionellae plates (65, 66).
If samples must be shipped to a laboratory, the plates should
be packed and shipped without refrigeration as soon as
possible.
Atypical Environmental Samples
Occasionally, samples other than bulk water, biofilm swabs,
or air are collected during outbreak investigations. Legion-
naires’ disease outbreaks have been linked to sources through
sampling such atypical sources as whirlpool spa filters, indus-
trial fluids, and potting soil (29, 67, 68). Direct transmission
from industrial fluids and moist soil have been documented,
whereas filter samples provide a reservoir from which public
health officials may recover Legionellae during investigations
when the source transmission device has been drained and
cleaned. When taking atypical environmental samples, the
goal should be to collect portions that remain moist and at
ambient temperatures conducive to bacterial growth or bac-
teriostasis. Liquid samples may be processed according to
the same methods as bulk water. Solid samples, such as spa fil-
ters, should be covered with water from the source or sterile
water during transport to prevent drying and agitated (i.e.,
vortexed or sonicated) to release trapped bacteria into the
liquid milieu for plating. A method for Legionellae recovery

from colloidal soil was described by Travis et al. (69). It
should be noted that solid and colloidal samples often con-
tain large numbers of competitive heterotrophic bacteria
and fungi, so sample processing methods that enrich or select
for Legionellae are encouraged.
Sample Transport and Storage
All liquid samples should be treated with an appropriate vol-
ume of 0.1 N sodium thiosulfate (0.5 ml per 1 liter) to remove
residual halide disinfectant. Continued contact with even
low levels of residual disinfectant may result in the inactiva-
tion of Legionellae during transport, yielding false negative cul-
ture results (70). Samples should be transported to the
laboratory at typical ambient room temperature (25–28°C)
in insulated containers for receipt within 24 h from the
time of collection. If longer holding times are anticipated,
samples should be stored at 4–6°C, but care must be taken
to avoid freezing. Ideally, samples should be processed and
plated within 72 h of collection. However, reprocessing sam-
ples is often necessary due to overgrowth of heterotrophic bac-
teria and fungi. Legionellae have been recovered from
reprocessed samples stored at 4–6°C for up to a few months
(61).
Sample Processing
A schematic representation of the methods for processing
water samples for culture is shown in Figure 2. The procedure
chosen depends on the expected degree of total bacterial con-
tamination in a particular sample. Potable waters generally
have low bacterial concentrations and are either cultured
directly or concentrated to detect Legionellae. Nonpotable
waters, such as those from cooling towers, generally do not
require concentration because of their high bacterial loads.
Swabs, filters, and other colloidal or solid samples should be
agitated to remove bacteria from the surface to a liquid milieu.
In practice, this can be accomplished by vortexing for 1 min
FIGURE 2 Overview of procedures for the culture of water samples to detect Legionellae. doi:10.1128/9781555818821.ch3.2.9.f2
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-5
on high setting or three 30-s bursts of sonication. The liquid
into which the bacteria are released is assumed to contain a
high bacterial load and plated directly, enriched for Legionel-
lae, or subjected to selective treatments.
Sample Concentration
Samples may be concentrated 10-fold or more by using either
filtration or centrifugation, though filtration is used more fre-
quently because the method is more conducive to high
throughput. Water is passed through a 0.2 μm pore size poly-
carbonate filter by applying vacuum to the apparatus mani-
fold. Filters containing nitrocellulose should not be used
as bacteria may adhere to them more closely and be difficult
to remove by agitation. Polycarbonate membranes allow
suspended particles to collect on the filter surface without
being trapped as with matrix-type filters (64). The filter
membrane is then aseptically removed from the filtration
apparatus, resuspended in 5–10 ml sterile water, and vortexed
for 1 min or sonicated in an ultrasound tank for 2–10 min
(61, 62, 64). Samples may be concentrated by centrifugation
at 6,000 g for 10 min. All but 10 ml of the supernatant is
then removed and the pellet resuspended by vortexing (71).
With either method of sample concentration, the starting
and ending volumes should be documented so that the con-
centration factor may be determined.
Acid or Heat Pretreatment
A selective procedure is required to reduce the number of
heterotrophic bacteria before culturing some water samples
with high total bacterial concentrations. NonLegionellae
bacteria can be selectively killed by either acid pretreatment
or brief exposure to higher temperatures, though in some stud-
ies acid treatment seems to yield slightly greater recovery
when nonselective media are used (72). Legionellae are
more resistant to lower pH and brief exposures to higher
temperatures than many other freshwater bacteria. For acid
3.2.9-6 ▪ AIR
pretreatment, the sample is mixed with an acid ( pH 2.2)
buffer and incubated at room temperature 10–60 min
(61, 62, 64). The sample is neutralized by the buffer within
the buffered charcoal yeast extract (BCYE) agar and so
must be spread on the agar plate immediately after the period
of incubation with the acid buffer ends. Alternatively, the
acid may be neutralized by an equal volume of 0.1 N potas-
sium hydroxide, but this dilutes the sample and is not a
recommended practice. Heat pretreatment is accomplished
by incubating 10 ml of sample in a 50°C water bath for
30 min (62).
Heat Enrichment
Heat enrichment should not be confused with heat pretreat-
ment, because it depends on the amplification of Legionellae
within autochthonous protozoan hosts and not the resistance
of Legionellae to higher temperatures. Instead, heat enrich-
ment may be considered a form of coculture. Samples are
incubated at 35°C for 2–6 weeks before aliquots are cultured.
Such enrichment can improve the recovery of Legionellae
by up to 30%, but the considerable length of time between
sampling event and results makes it impractical in many situa-
tions (73).
Culture Media
Legionellae were isolated by procedures used for the recovery of
rickettsiae as early as 1943 (TATLOCK strain, now
L. micdadei) although the bacteria were not identified as
Legionellae until 1979 (20, 74, 75). The bacterium was first
isolated on bacteriologic agar, using Mueller-Hinton agar
supplemented with hemoglobin and IsoVitaleX (MH-IH)
(76). The essential component in hemoglobin was found to
be a soluble form of iron while L-cysteine was the essential
amino acid provided by the IsoVitaleX. These refinements
led to the development of Feeley-Gorman agar, which pro-
vided better recovery of the organism from tissue. Later, starch
was replaced with charcoal to detoxify the medium and the
amino acid source was changed to yeast extract, resulting in
charcoal yeast extract agar (33). Charcoal yeast extract agar
is the base form for most media used to grow Legionellae.
The medium used for the culture of Legionellae has been
improved several times, eventually resulting in the medium
TABLE 3 Components and supplements of BCYE agar for culturing Legionellae from the environment
Component Concentration
Charcoal 2.0 g/liter
Yeast extract 10.0 g/liter
ACESa buffer 10.0 g/liter
Ferric pyrophosphate 0.25 g/liter
L-cysteine 0.4 g/liter
Potassium α-ketoglutarate 1.0 g/liter
Agar 17.0 g/liter
Glycine 3.0 g/liter
Polymyxin B 50–100 U/ml
Vancomycin or cefamandole 1–5 g or 4 mg/liter
Anisomycin or cycoheximide 80 µg/ml (for either)
Bromocresol blue 10 mg/liter
Bromocresol purple 10 mg/liter
Bovine serum albumin 10 g/liter
a
N-(2-acetamido)-2-aminothanesulfonic acid.
currently used, BCYE agar (33, 77, 78). The most widely
used form of BCYE agar is also supplemented with
α-ketoglutarate (51, 77, 78). Table 3 lists the primary compo-
nents of BCYE agar and the supplements added for various
purposes (33, 79–81).
Culture of environmental samples requires the use of selec-
tive and nonselective media in conjunction with the previ-
ously described pretreatment procedures. Most laboratories
use multiple plates for each sample, including a BCYE agar
plate containing the three antimicrobial agents plus glycine
(Table 3) (55, 61, 62). Media prepared with indicator dyes,
which impart a color specific for certain species of Legionella,
has fallen out of general use since most laboratories identify
Legionellae with serological or molecular methods (64, 82).
Although the majority of Legionella spp. grow readily on
BCYE agar, some require supplementation with bovine serum
albumin to enhance growth (79). Additionally, heat may be a
preferable pretreatment than acid when media containing
glycine is used (83, 84). All agar plates are inoculated with
0.1 ml of the sample by the spread plate method and incu-
bated at 35–37°C for 4–10 days. Incubation in a humidified
2.5% CO2 may increase the growth rate of isolates from clin-
ical samples but does not seem to significantly affect recovery
from environmental samples (55, 64).
Identification of Legionellae and Subtyping Isolates
Macroscopic colonies of Legionellae typically appear on BCYE
after approximately 72 h of incubation but may require a total
incubation of 7 days or longer. Plates should be examined
after 4 days’ incubation and again at 7–10 days before being
discarded. Plates should be examined with a dissecting micro-
scope and a light source to detect bacterial colonies resem-
bling Legionellae. After approximately 4 days of incubation,
these colonies are 2 to 4 mm in diameter, convex, and round
with entire edges (Figure 3). The center of the colony is usu-
ally a bright white with a textured appearance that has been
described as “cut glass–like” or speckled. The white center
of the colony is often bordered with blue, purple, green, or
red iridescence. Some species of Legionellae produce colonies
that exhibit blue-white or red autofluorescence (85, 86). Pri-
mary isolation plates can be examined with long-wave UV
light to detect autofluorescent colonies.
Purpose
Base component
Base component
Base component
Base component
Base component
Base component
Base component
Selective agent
Selective agent (gram negative)
Selective agent (gram positive)
Selective agent (fungal)
Indicator dye
Indicator dye
Supplement for some fastidious Legionellae

FIGURE 3 Two Legionella colonies and a non-Legionella bacterial
colony as seen through a dissecting microscope upon primary isola-
tion (4 days of incubation). Note the white “cut-glass” appearance
of the center of the colony and the purple iridescence which borders
it. The iridescence can be one of several colors; the significance of
the colors is unknown. doi:10.1128/9781555818821.ch3.2.9.f3
Colonies resembling Legionellae can be presumptively
identified on the basis of their requirement for L-cysteine
by subculture on blood agar or BCYE without L-cysteine.
Subcultured colonies that grow on BCYE agar but not on
blood agar or BCYE without L-cysteine are presumed to
belong to genus Legionella. The exception to this rule is
L. oakridgensis, which may grow in the absence of L-cysteine,
especially after serial passage on artificial media (87). Legion-
ellae are relatively inert in many biochemical test media, so
these tests are of limited value in identification. Historically,
definitive identification of Legionellae has been accomplished
by serological methods, using a direct fluorescent antibody
(DFA) or slide agglutination test with specific antisera.
Legionella species may also be determined by fatty acid ana-
lysis or DNA hybridization (61, 86, 88). Fatty acid analysis
is useful for discriminating between numerous isolates, but
can confound speciation and subtyping when used on blue-
white autofluorescent strains (89). DNA hybridization is
probably the most precise of the methods historically used
to identify Legionella, and recent advances with oligonucleo-
tide arrays show promise, but to date such methods are tech-
nically challenging to accomplish (82, 90). Serological tests
such as DFA and slide agglutination are limited in that
antisera for many species and serogroups are not available
commercially and those that are can have confounding cross-
reactions between nonpneumophila species (6, 91–93). The
use of molecular methods for both speciation and subtyping
has recently increased due to greater discriminatory power,
reproducibility, and ease of interpretation compared to sero-
logical methods (82). Comparing the sequence of the mip
gene to a database such as NCBI (http://www.ncbi.nlm.nih.
gov/pubmed/) is a reliable method for identifying nonpneu-
mophila species of genus Legionella (94, 95).
Associating an environmental isolate of Legionella with
a clinical isolate from a patient with legionellosis usually
requires a subtyping procedure. L. pneumophila serogroup
1 (Lp1) accounts for most cases of disease but Lp1 is a fairly
heterogeneous serogroup that may be divided into a number
of subtypes using various techniques (1, 96). Typing Legionel-
lae to 1 of 17 serogroups employs polyvalent or monoclonal
antisera and may be adequate for identifying reservoirs of
some of the more uncommon L. pneumophila. However, the
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-7
variety of strains and distribution of Lp1 typically necessitate
more elaborate subtyping procedures to discriminate within
this serogroup. Thus, identification of L. pneumophila to the
serogroup level is not generally sufficient to implicate an envi-
ronmental isolate as the source of disease. Several groups of
monoclonal antibodies have been developed for Lp1 subtyp-
ing purposes that are able to categorize Lp1 into 10–12 type
strains (92, 97). Although this has been extremely useful dur-
ing certain outbreak investigations, the drawback to this
method is that the monoclonal antibodies are not available
commercially and may only be supplied in limited quantities
by the developers (82).
Subtyping strains of L. pneumophila using sequenced-
based typing (SBT) of seven gene fragments has recently
increased in use and proven to be a quick, reliable, and power-
ful tool for epidemiologically linking environmental and clin-
ical isolates during investigations of disease or environmental
surveillance studies (98, 99). Sequence types submitted by
volunteers worldwide may be compared using the database
of sequence types maintained by the European Study Group
for Legionella Infections (http://www.hpa-bioinformatics.org.
uk/legionella/legionella_sbt/php/sbt_homepage.php). How-
ever, SBT may only be used to subtype the species L. pneumo-
phila, and the current database is heavily weighted toward
isolates of serogroup 1 (100–102). Thus, fragment length
molecular methods such as pulsed-field gel electrophoresis
(PFGE), amplified-fragment length polymorphism, or arbi-
trarily primed PCR are still preferred for subtyping non-
pneumophila species (82).
Figure 4 provides an algorithm that combines serological
and molecular methods, which can be used by environmental
laboratories for speciation and typing of isolates that are
confirmed as genus Legionella. The serological tests most
frequently employed are those used to identify the species
L. pneumophila and distinguish between serogroups 1 and 2–
14, respectively, because there are many commercial identi-
fication kits available. In practice, commercial environ-
mental laboratories in the United States rarely provide
information beyond the first tier of identification: L. pneumo-
phila serogroup 1 (Lp1), L. pneumophila of undetermined
serogroup 2–14 (Lp2–14), or Legionella spp. nonpneumophila
(Legionella spp.) (55). The more discriminatory typing
methods, such as monoclonal antibody, SBT, or PFGE, are
usually reserved for research and public health laboratories.
It should be noted that serogroup, sequence type, and PFGE
(or other fragment) patterns are complementary subtyping
methods that have no direct relation to each other. Thus,
employing a combination of some or all of these methods
to an isolate allows for the most precise definition of
subtype (82).
Nonculture Methods for Detection of Legionellae
Several nonculture methods have been developed to detect
Legionellae in environmental samples. These methods offer
the potential of greatly increased sensitivity. However, cul-
ture remains the method of choice for detecting Legionellae
in environmental samples primarily because nonculture
methods cannot provide information regarding the viability
of the bacteria. Nonculture methods include detection of
the organisms with specific antisera by DFA staining and
procedures to detect nucleic acids of Legionellae using PCR.
DFA use for detection is limited by the number of specific
antisera available since there is no antiserum that specifically
reacts with all Legionella species. Reports on the sensitivity
and specificity of DFA testing of environmental specimens
3.2.9-8 ▪ AIR

Cysteine Auxotrophy
Confirmed Genus Legionellaa
DFA or slide agglutination
to Lp1 and Lp2-14

Legionella
Lp1 spp.
Sequence Based Typing
Lp2-14
International MAb Serologyb
PFGE (or other fragment analysis) DFA or slide agglutination
with species-specific
(OPTIONAL: Serotype with antisera
serogroup specific antisera by or
Sequence Type DFA or slide agglutination) mip sequence
or Sequence Based Typing
Type Strain PFGE (or other fragment
analysis)
or Species
Fragment Pattern Identification
(Serogroup)
(and) PFGE (or other
fragment analysis)
Sequence Type,
or
Fragment Pattern Fragment Pattern

FIGURE 4 Algorithm for speciating and subtyping genus-confirmed Legionella isolates. There are numerous commercial vendors of reagents
to perform DFA or slide agglutination to quickly determine whether an isolate is species L. pneumophila and, if so, whether it belongs to
serogroup 1 (Lp1) or serogroups 2–14 (Lp2–14). Lp1 isolates may be further subtyped by SBT, serology with a panel of monoclonal antibodiesb,
and/or fragment analyses such as PFGE. Commercial antisera also exist to specifically serotype Lp2–14 isolates. Subtyping Lp2–14 isolates may
be achieved by SBT and/or fragment analyses. Speciating nonpneumophila Legionella spp. isolates is accomplished by employing species-specific
antisera in DFA or slide agglutination tests or by sequencing the mip gene. Genes for all seven of the genes used in SBT are generally not present
in nonpneumophila species, so the only way to reliably subtype these isolates is by fragment analyses.aL. oakridgensis may not display cysteine
auxotrophy. When performing comprehensive environmental surveys or where epidemiological evidence suggests the involvement of this spe-
cies in cases of disease, it may be advantageous to subject suspect colonies to the nonpneumophila Legionella spp. branch (far right) of the algo-
rithm.bThere are no commercially available sources of the international panel for monoclonal antibody subtyping. However, limited supplies
may be obtained for research or public health surveillance purposes from the developers (97). doi:10.1128/9781555818821.ch3.2.9.f4

vary greatly, with most indicating that the test is relatively
insensitive and nonspecific (64, 103).
The use of PCR for detecting nucleic acids of Legionellae
in the environment can be a valuable tool in some inves-
tigations of outbreaks of legionellosis and environmental
surveys when used in conjunction with culture (104,
105). Legionella gene targets that have proven to be the
most useful include 5S rRNA, 16S rRNA, and mip, but
advances in whole genome sequencing and comparative
genomics may yield targets with greater specificity and,
therefore, suitability for use in environmental samples
(90, 106, 107). Most investigations of epidemic legionello-
sis have used culture to detect Legionellae in the environ-
ment. As a result, most of our epidemiologically relevant
information concerning legionellosis is based on direct cul-
ture data. Until a better understanding of the diversity and
distribution of Legionellae in the environment is attained,
results from nonculture-based methods should be inter-
preted cautiously (13).
ENVIRONMENTAL APPROACHES TO
CONTROLLING LEGIONELLOSIS
Practical information concerning treatment processes that
effectively control Legionellae is limited. Various biocides
and alternative disinfection methods, such as heat
eradication, UV irradiation, ionization, and ozonation have
been tested to determine their abilities to kill Legionellae
(108, 109). However, results obtained in laboratory studies
do not always translate into effective prevention protocols,
and all methods of Legionella control applied to water systems
have failed in some instances (110–113). Several countries
have produced guidelines or codes of practice relating to the
control of Legionellae. However, research to substantiate these
practices is scarce and the prevailing rationale for these rec-
ommendations is often empirical (111, 114). Currently, there
are 13 US guidelines that address the prevention and control
of legionellosis: five federal, two state, one county, and four
from professional societies (115). These guidelines address
maintenance temperatures, biocide levels, and emergency
procedures for a number of building water systems including
potable systems, cooling towers, and heated spas. If one uti-
lizes the chain of causation leading to an outbreak of legion-
ellosis (as described earlier in this chapter) a control strategy
would only need to interrupt the chain of events to be suc-
cessful. The following discussion of control strategies is
arranged according to this scheme.
Environmental Reservoirs and Water Supplies
It is highly unlikely that Legionellae could be eradicated
from the aquatic environment since they are integral mem-
bers of freshwater ecosystems, where they are ubiquitous but
present in relatively low concentrations (60, 116–119). The

difficulty in detecting Legionellae in water treatment plants
and municipal water supplies is probably due to the lower
temperatures of these waters. Legionellae are more frequently
detected and present in higher concentrations in warm or
thermally altered environments, such as building water sys-
tems. A community-based approach to control that may
reduce Legionella colonization of building potable water sys-
tems is municipal use of monochloramine as a disinfectant
(49, 120, 121). However, there is at least one report that
describes an alteration of POU biofilm to contain more myco-
bacteria so the benefit in Legionellae control must be weighed
against possible negative effects from other pathogenic water-
borne microorganisms (122).
Amplifying Reservoirs
As previously mentioned, temperature is a critical factor in
the ability of Legionellae to colonize reservoirs in which they
are amplified. Other microorganisms and factors critical to
the growth of Legionellae are almost universally present in
freshwater environments, and it is temperature that governs
the numbers of these bacteria. The ideal temperature range
for most species of Legionellae to amplify is between 25°C
and 42°C (6, 64). At lower temperatures the bacteria may
be metabolically inactive but are still viable while at temper-
atures above 45°C the bacteria are rapidly killed. Therefore, it
has been suggested that potable hot water systems be main-
tained at temperatures between 55°C and 60°C to prevent
growth of bacteria. However, this may be problematic for
some institutions, especially hospitals, where the potential
for scalding patients exists or where state regulations are pro-
hibitive (123).
Practices to control Legionellae in amplifying reservoirs
can be divided into two categories: routine maintenance
and emergency decontamination. Examples of routine main-
tenance practices include the requirement that respiratory
therapy equipment is not rinsed with potable water before
patient use or cleaning and startup protocols for cooling
towers, evaporative condensers, and whirlpool spas (115,
124). These documents suggest following manufacturer’s
recommendations regarding cleaning and biocide treatment
of these devices. Well-maintained equipment does not per-
mit excessive growth of Legionellae. Emergency decontami-
nation protocols describing hyperchlorination and cleaning
procedures for cooling towers and hot tubs implicated in
the transmission of legionellosis have been developed
(125, 126). Additionally, guidelines for the control of noso-
comial legionellosis are presented in documents addres-
sing nosocomial pneumonia, the hospital environment,
and facilities that perform organ or hematopoietic stem
cell transplants (19, 115, 127). The principal approaches
to emergency disinfection are heat flushing with water main-
tained at ≥60°C, hyperchlorination, and physical cleaning
of water storage tanks. However, potable systems are easily
recolonized and may require continuous intervention such
as raising hot water temperatures or continuous addition of
disinfectant to maintain a measurable residual concentration
(111, 128).
Aerosol Transmission
Some control strategies are intended to prevent exposure
of susceptible individuals to aerosols that may contain Legion-
ellae. Cooling tower manufacturers have improved the per-
formance of drift eliminators, and these are now a common
fixture on cooling towers (123). Modern eliminator modules
can reduce the water loss from cooling towers by orders of
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-9
magnitude, which is beneficial for not only disease pre-
vention but also energy conservation. Early investigations
of Legionnaires’ disease associated with cooling towers also
resulted in adopting the recommendation for relocation of
cooling towers or air intake vents so that cooling tower
exhaust would not be carried directly into the HVAC systems
of buildings. Emergency control measures may also involve
a physical barrier between aerosol sources by installing
POU filters on showerheads and faucets. However, the
requirement that these filters be changed frequently is often
cost-prohibitive for routine usage (129–132).
Health care professionals, engineers, and industry associa-
tions continue to increase their awareness and understanding
of Legionellae and legionellosis. Unfortunately, it can be diffi-
cult to determine whether increased understanding has led to
measures that have reduced the incidence of legionellosis.
Quick empirical antibiotic treatment of pneumonias within
4 h of admission, while providing better outcomes for the
patient, inhibits public health surveillance on the incidence
of disease. Moreover, fear of litigation may have caused the
adoption of some prevention strategies that lack sufficient sci-
entific support. Effective strategies can only be measured by
approaches that both understand the incidence of disease
and measure the efficacy of preventative actions. Although
Legionella cannot be eradicated from the environment, dis-
ease is preventable by controlling transmission to humans
from environmental sources. Reducing amplification within
manmade structures and controlling the release of contami-
nated aerosols minimizes the risk of exposure to susceptible
individuals.
REFERENCES
1. Hicks LA, Garrison LE, Nelson GE, Hampton LM. 2012.
Legionellosis—United States, 2000–2009. Am J Trans 12:
250–253.
2. Brunkard JM, Ailes E, Roberts VA, Hill V, Hilborn ED,
Craun GF, Rajasingham A, Kahler A, Garrison L, Hicks
L, Carpenter J, Wade TJ, Beach MJ, YoderJS, CDC.
2011. Surveillance for waterborne disease outbreaks associ-
ated with drinking water—United States, 2007–2008. Mor-
bid Mortal Weekly Rep Surveill Summ 60:38–68.
3. Palusinska-Szysz M, Cendrowska-Pinkosz M. 2009.
Pathogenicity of the family Legionellaceae. Arch Immunol
Ther Exp 57:279–290.
4. McDade JE, Shepard CC, Fraser DW, Tsai TR, Redus
MA, Dowdle WR. 1977. Legionnaires’ disease: isolation
of a bacterium and demonstration of its role in other respi-
ratory disease. N Eng J Med 297:1197–1203.
5. Fraser DW, Tsai TR, Orenstein W, Parkin WE, Beecham
HJ, Sharrar RG, Harris J, Mallison GF, Martin SM,
McDade JE, Shepard CC, Brachman PS. 1977. Legion-
naires’ disease: description of an epidemic of pneumonia.
N Eng J Med 297:1189–1197.
6. Fields BS, Benson RF, Besser RE. 2002. Legionella and
Legionnaires’ disease: 25 years of investigation. Clin Micro-
biol Rev 15:506–526.
7. Molmeret M, Bitar DM, Han L, Kwaik YA. 2004. Cell
biology of the intracellular infection by Legionella pneumo-
phila. Micr Infect 6:129–139.
8. Friedman H, Yamamoto Y, Klein TW. 2002. Legionella
pneumophila pathogenesis and immunity. Sem Ped Infect
Dis 13:273–279.
9. Castor ML, Wagstrom EA, Danila RN, Smith KE, Naimi
TS, Besser JM, Peacock KA, Juni BA, Hunt JM, Bartkus
JM, Kirkhorn SR, Lynfield R. 2005. An outbreak of
Pontiac fever with respiratory distress among workers
3.2.9-10 ▪ AIR
performing high-pressure cleaning at a sugar-beet process-
ing plant. J Infect Dis 191:1530–1537.
10. Fields BS, Haupt T, Davis JP, Arduino MJ, Miller PH,
Butler JC. 2001. Pontiac fever due to Legionella micdadei
from a whirlpool spa: possible role of bacterial endotoxin.
J Infect Dis 184:1289–1292.
11. Hadler SC, Castro KG, Dowdle W, Hicks L, Noble G,
Ridzon R. 2011. Epidemic Intelligence Service investiga-
tions of respiratory illness, 1946–2005. Am J Epidemiol
174:S36–S46.
12. Hollenbeck B, Dupont I, Mermel LA. 2011. How often is
a work-up for Legionella pursued in patients with pneumo-
nia? a retrospective study. BMC Infect Dis 11:237.
13. Tronel H, Hartemann P. 2009. Overview of diagnostic
and detection methods for legionellosis and Legionella
spp. Lett Appl Microbiol 48:653–656.
14. Lim YH, Relus Kek YL, Lim PY, Yap HM, Vivien Goh
TL, Ng LC. 2011. Environmental surveillance and molec-
ular characterization of Legionella in tropical Singapore.
Trop Biomed 28:149–159.
15. Lee HK, Shim JI, Kim HE, Yu JY, Kang YH. 2010. Dis-
tribution of Legionella species from environmental water
sources of public facilities and genetic diversity of L. pneu-
mophila serogroup 1 in South Korea. Appl Environ Microbiol
76:6547–6554.
16. Kozak NA, Benson RF, Brown E, Alexander NT,
Taylor TH Jr., Shelton BG, Fields BS. 2009. Distribution
of lag-1 alleles and sequence-based types among Legionella
pneumophila serogroup 1 clinical and environmental
isolates in the United States. J Clin Microbiol 47:
2525–2535.
17. Den Boer JW, Verhoef L, Bencini MA, Bruin JP, Jansen
R, Yzerman EP. 2007. Outbreak detection and secondary
prevention of Legionnaires’ disease: a national approach.
Int J Hyg Envir Heal 210:1–7.
18. Sehulster L, Chinn RY, CDC, HICPAC. 2003. Guide-
lines for environmental infection control in health-care
facilities. Recommendations of CDC and the Healthcare
Infection Control Practices Advisory Committee (HIC-
PAC). MMWR Recomm Rep 52:1–42.
19. Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R,
CDC, HICPAC. 2004. Guidelines for preventing health-
care–associated pneumonia, 2003: recommendations of
CDC and the Healthcare Infection Control Practices
Advisory Committee. MMWR Recomm Rep 53:1–36.
20. Brenner DJ, Steigerwalt AG, McDade JE. 1979. Classifi-
cation of the Legionnaires’ disease bacterium: Legionella
pneumophila, genus novum, species nova, of the family
Legionellaceae, familia nova. Ann Int Med 90:656–658.
21. Edelstein PH, Edelstein MA, Shephard LJ, Ward KW,
Ratcliff RM. 2012. Legionella steelei sp. nov., isolated
from human respiratory specimens in California, USA,
and South Australia. Int. J. Syst. Evol. Micr 62:1766–1771.
22. Graham FF, White PS, Harte DJ, Kingham SP. 2012.
Changing epidemiological trends of legionellosis in
New Zealand, 1979–2009. Epidemiol Infect 140:1481–1496.
23. Alli OA, Zink S, Von Lackum NK, Abu-Kwaik Y. 2003.
Comparative assessment of virulence traits in Legionella spp.
Microbiology 149(Pt 3):631–641.
24. Lamoth F, Greub G. 2010. Amoebal pathogens as emerg-
ing causal agents of pneumonia. FEMS Microbiol Rev 34:
260–280.
25. Carvalho FR, Nastasi FR, Gamba RC, Foronda AS, Pel-
lizari VH. 2008. Occurrence and diversity of Legionella-
ceae in polar lakes of the Antarctic peninsula. Curr
Microbiol 57:294–300.
26. Ott CM, Bruce RJ, Pierson DL. 2004. Microbial charac-
terization of free floating condensate aboard the Mir space
station. Microbial Ecol 47:133–136.
27. Katz SM, Hammel JM. 1987. The effect of drying, heat,
and pH on the survival of Legionella pneumophila. Ann
Clin Lab Sci 17:150–156.
28. Mermel LA, Josephson SL, Giorgio CH, Dempsey J,
Parenteau S. 1995. Association of Legionnaires’ disease
with construction: contamination of potable water? Infect
Control Hosp Epidemiol 16:76–81.
29. Whiley H, Bentham R. 2011. Legionella longbeachae and
legionellosis. Emerg Infect Dis 17:579–583.
30. Silk BJ, Moore MR, Bergtholdt M, Gorwitz RJ, Kozak
NA, Tha MM, Brown EW, Winchester JL, Labus BJ,
Rowley P, Middaugh JP, Fields BS, Hicks LA. 2012. Eight
years of Legionnaires’ disease transmission in travellers to a
condominium complex in Las Vegas, Nevada. Epidemiol
Infect 140:1993–2002.
31. Rangel-Frausto MS, Rhomberg P, Hollis RJ, Pfaller MA,
Wenzel RP, Helms CM, Herwaldt LA. 1999. Persistence
of Legionella pneumophila in a hospital’s water system: a
13-year survey. Infect Control Hosp Epidemiol 20:793–797.
32. Cowgill KD, Lucas CE, Benson RF, Chamany S, Brown
EW, Fields BS, Feikin DR. 2005. Recurrence of Legion-
naire’s disease at a hotel in the United States Virgin
Islands over a 20-year period. Clin Infect Dis 40:1205–1207.
33. Feeley JC, Gibson RJ, Gorman GW, Langford NC,
Rasheed JK, Mackel DC, Baine WB. 1979. Charcoal-
yeast extract agar: primary isolation medium for Legionella
pneumophila. J Clini Microbiol 10:437–441.
34. Rowbotham TJ. 1980. Preliminary report on the pathoge-
nicity of Legionella pneumophila for freshwater and soil
amoebae. J Clin Pathol 33:1179–1183.
35. Fields BS. 1996. The molecular ecology of legionellae.
Trends Microbiol 4:286–290.
36. Solomon JM, Isberg RR. 2000. Growth of Legionella pneu-
mophila in Dictyostelium discoideum: a novel system for
genetic analysis of host-pathogen interactions. Trends
Microbiol 8:478–480.
37. Ensminger AW, Yassin Y, Miron A, Isberg RR. 2012.
Experimental evolution of Legionella pneumophila in mouse
macrophages leads to strains with altered determinants of
environmental survival. PLoS Pathogens 8:e1002731.
38. O’Connor TJ, Adepoju Y, Boyd D, Isberg RR. 2011.
Minimization of the Legionella pneumophila genome reveals
chromosomal regions involved in host range expansion.
Proc Natl Acad Sci USA 108:14733–14740.
39. Swanson MS, Hammer BK. 2000. Legionella pneumophila
pathogesesis: a fateful journey from amoebae to macro-
phages. Ann Rev Microbiol 54:567–613.
40. Nagai H, Kubori T. 2011. Type IVB secretion systems of
Legionella and other Gram-negative bacteria. Front Micro-
biol 2:136.
41. Gomez-Valero L, Rusniok C, Cazalet C, Buchrieser C.
2011. Comparative and functional genomics of Legionella
identified eukaryotic like proteins as key players in host-
pathogen interactions. Front Microbiol 2:208.
42. Rolando M, Buchrieser C. 2012. Post-translational
modifications of host proteins by Legionella pneumophila: a
sophisticated survival strategy. Fut Microbiol 7:369–381.
43. Luo ZQ. 2012. Legionella secreted effectors and innate
immune responses. Cell Microbiol 14:19–27.
44. Yoder J, Roberts V, Craun GF, Hill V, Hicks LA,
Alexander NT, Radke V, Calderon RL, Hlavsa MC,
Beach MJ, Roy SL, CDC. 2008. Surveillance for water-
borne disease and outbreaks associated with drinking water
and water not intended for drinking—United States,
2005–2006. Morbid Mortal Weekly Rep Surveill Summ 57:
39–62.
45. Phares CR, Russell E, Thigpen MC, Service W, Crist
MB, Salyers M, Engel J, Benson RF, Fields B, Moore
MR. 2007. Legionnaires’ disease among residents of a long-

term care facility: the sentinel event in a community out-
break. Am J Infect Cont 35:319–323.
46. O’Loughlin RE, Kightlinger L, Werpy MC, Brown E,
Stevens V, Hepper C, Keane T, Benson RF, Fields
BS, Moore MR. 2007. Restaurant outbreak of Legion-
naires’ disease associated with a decorative fountain:
an environmental and case-control study. BMC Infect Dis
7:93.
47. Allen JG, Myatt TA, Macintosh DL, Ludwig JF, Mine-
gishi T, Stewart JH, Connors BF, Grant MP, McCarthy
JF. 2012. Assessing risk of health care-acquired Legion-
naires’ disease from environmental sampling: the limits of
using a strict percent positivity approach. Am J Infect Contr
40:917–921.
48. Ditommaso S, Giacomuzzi M, Gentile M, Moiraghi AR,
Zotti CM. 2010. Effective environmental sampling
strategies for monitoring Legionella spp contamination in
hot water systems. Am J Infect Cont 38:344–349.
49. Flannery B, Gelling LB, Vugia DJ, Weintraub JM, Sale-
rno JJ, Conroy MJ, Stevens VA, Rose CE, Moore MR,
Fields BS, Besser RE. 2006. Reducing Legionella coloniza-
tion in water systems with monochloramine. Emerg Infect
Dis 12:588–596.
50. Ohno A, Kato N, Sakamoto R, Kimura S, Yamaguchi
K. 2008. Temperature-dependent parasitic relationship
between Legionella pneumophila and a free-living amoeba
(Acanthamoeba castellanii). Appl Environ Microbiol 74:
4585–4588.
51. Luck PC, Igel L, Helbig JH, Kuhlisch E, Jatzwauk L.
2004. Comparison of commercially available media for
the recovery of Legionella species. Int J Hyg Envir Health
207:589–593.
52. Leoni E, Legnani PP. 2001. Comparison of selective proce-
dures for isolation and enumeration of Legionella species
from hot water systems. J Appl Microbiol 90:27–33.
53. McCoy WF, Downes EL, Leonidas LF, Cain MF,
Sherman DL, Chen K, Devender S, Neville MJ. 2012.
Inaccuracy in Legionella tests of building water systems
due to sample holding time. Water Res 46:3497–3506.
54. Boulanger CA, Edelstein PH. 1995. Precision and accu-
racy of recovery of Legionella pneumophila from seeded tap
water by filtration and centrifugation. Appl Enbiron Micro-
biol 61:1805–1809.
55. Lucas CE, Taylor TH Jr., Fields BS. 2011. Accuracy and
precision of Legionella isolation by US laboratories in the
ELITE program pilot study. Water Res 45:4428–4436.
56. Buse HY, Schoen ME, Ashbolt NJ. 2012. Legionellae
in engineered systems and use of quantitative microbial
risk assessment to predict exposure. Water Res 46:921–933.
57. Schoen ME, Ashbolt NJ. 2011. An in-premise model for
Legionella exposure during showering events. Water Res
45:5826–5836.
58. Morris GK, Shelton BG. 1991. Legionella in environmen-
tal samples: hazard analysis and suggested remedial actions.
Pathcon Technical Bulletin 13.
59. Best M, Yu VL, Stout J, Goetz A, Muder RR, Taylor F.
1983. Legionellaceae in the hospital water-supply. Epide-
miological link with disease and evaluation of a method
for control of nosocomial legionnaires’ disease and Pitts-
burgh pneumonia. Lancet 2:307–310.
60. Borella P, Guerrieri E, Marchesi I, Bondi M, Messi P.
2005. Water ecology of Legionella and protozoan: environ-
mental and public health perspectives. Biotechnol Annu Rev
11:355–80.
61. Fields BS. 2005. Procedures for the recovery of Legionella
from the environment, pp. 1–13. In: Services HaH, (ed),
2nd ed. Public Health Service, Atlanta, GA.
62. ISO. 1998. Water quality—detection and enumeration of
Legionella. ISO, Geneva, Switzerland,.
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-11
63. ISO. 2004. Water quality—detection and enumeration of
Legionella—Part 2: Direct membrane filtration method for
waters with low bacteria counts. ISO, Geneva, Switzerland.
64. Kozak NA, Lucas CE, Winchell JM. 2013. Identification
of Legionella in the environment. Meth Molecular Biology
954:3–25.
65. Ishimatsu S, Miyamoto H, Hori H, Tanaka I, Yoshida S.
2001. Sampling and detection of Legionella pneumophila
aerosols generated from an industrial cooling tower. Ann
Occup Hyg 45:421–427.
66. Breiman RF, Cozen W, Fields BS, Mastro TD, Carr SJ,
Spika JS, Mascola L. 1990. Role of air sampling in investi-
gation of an outbreak of legionnaires’ disease associated
with exposure to aerosols from an evaporative condenser.
J Infect Dis 161:1257–1261.
67. Fry AM, Rutman M, Allan T, Scaife H, Salehi E, Benson
R, Fields B, Nowicki S, Parrish MK, Carpenter J, Brown
E, Lucas C, Horgan T, Koch E, Besser RE. 2003. Legion-
naires’ disease outbreak in an automobile engine manufac-
turing plant. J Infectious Diseases 187:1015–1018.
68. Jernigan DB, Hofmann J, Cetron MS, Genese CA,
Nuorti JP, Fields BS, Benson RF, Carter RJ, Edelstein
PH, Guerrero IC, Paul SM, Lipman HB, Breiman R.
1996. Outbreak of Legionnaires’ disease among cruise
ship passengers exposed to a contaminated whirlpool spa.
Lancet 347:494–499.
69. Travis TC, Brown EW, Peruski LF, Siludjai D, Jorakate
P, Salika P, Yang G, Kozak NA, Kodani M, Warner AK,
Lucas CE, Thurman KA, Winchell JM, Thamthitiwat S,
Fields BS. 2012. Survey of Legionella species found in Thai
soil. Int J Microbiol 2012:218791.
70. Wiedenmann A, Langhammer W, Botzenhart K. 2001. A
case report of false negative Legionella test results in a chlori-
nated public hot water distribution system due to the lack of
sodium thiosulfate in sampling bottles. Int J Hyg Envir Heal
204:245–249.
71. Brindle RJ, Stannett PJ, Cunliffe RN. 1987. Legionella
pneumophila: comparison of isolation from water specimens
by centrifugation and filtration. Epidemiol Infect 99:
241–247.
72. Reinthaler FF, Sattler J, Schaffler-Dullnig K, Weinmayr
B, Marth E. 1993. Comparative study of procedures for iso-
lation and cultivation of Legionella pneumophila from tap
water in hospitals. J Clin Microbiol 31:1213–1216.
73. Sanden GN, Morrill WE, Fields BS, Breiman RF, Bar-
baree JM. 1992. Incubation of water samples containing
amoebae improves detection of legionellae by the culture
method. Appl Environ Microbiol 58:2001–2004.
74. Bozeman FM, Humphries JW, Campbell JM. 1968. A new
group of rickettsia-like agents recovered from guinea pigs.
Acta Virol 12:87–93.
75. Hebert GA, Moss CW, McDougal LK, Bozeman FM,
McKinney RM, Brenner DJ. 1980. The rickettsia-like
organisms TATLOCK (1943) and HEBA (1959): bacteria
phenotypically similar to but genetically distinct from
Legionella pneumophila and the WIGA bacterium. Ann
Intern Med 92:45–52.
76. Feeley JC, Gorman GW, Weaver RE, Mackel DC, Smith
HW. 1978. Primary isolation media for Legionnaires dis-
ease bacterium. J Clin Microbiol 8:320–325.
77. Edelstein PH. 1981. Improved semiselective medium for
isolation of Legionella pneumophila from contaminated clin-
ical and environmental specimens. J Clin Microbiol 14:
298–303.
78. Pasculle AW, Feeley JC, Gibson RJ, Cordes LG, Myero-
witz RL, Patton CM, Gorman GW, Carmack CL, Ezzell
JW, Dowling JN. 1980. Pittsburgh pneumonia agent:
direct isolation from human lung tissue. J Infect Dis 141:
727–732.
3.2.9-12 ▪ AIR
79. Morrill WE, Barbaree JM, Fields BS, Sanden GN, Martin
WT. 1990. Increased recovery of Legionella micdadei and
Legionella bozemanii on buffered charcoal yeast extract agar
supplemented with albumin. J Clini Microbiol 28:616–68.
80. Vickers RM, Stout JE, Yu VL, Rihs JD. 1987. Manual of
culture methodology for Legionella. Sem Resp Infect 2:
274–279.
81. Vickers RM, Brown A, Garrity GM. 1981. Dye-
containing buffered charcoal-yeast extract medium for
differentiation of members of the family Legionellaceae.
J Clini Microbiol 13:380–382.
82. Luck C, Fry NK, Helbig JH, Jarraud S, Harrison TG.
2013. Typing methods for Legionella. Meth Mol Biol 954:
119–148.
83. DeLuca G, Stampi S, Lezzi L, Zanetti F. 1999. Effect of
heat and acid decontamination treatments on the recovery
of Legionella pneumophila from drinking water using two
selective media. New Microbiol 22:203–208.
84. Ta AC, Stout JE, Yu VL, Wagener MM. 1995. Compari-
son of culture methods for monitoring Legionella species in
hospital potable water systems and recommendations for
standardization of such methods. J Clin Microbiol 33:
2118–2123.
85. Amemura-Maekawa J, Hayakawa Y, Sugie H, Moribaya-
shi A, Kura F, Chang B, Wada A, Watanabe H. 2004.
Legioliulin, a new isocoumarin compound responsible for
blue-white autofluorescence in Legionella (Fluoribacter)
dumoffii under long-wavelength UV light. Biochem Biophys
Res Commun 323:954–959.
86. Wilkinson IJ, Sangster N, Ratcliff RM, Mugg PA, Davos
DE, Lanser JA. 1990. Problems associated with identifica-
tion of Legionella species from the environment and isola-
tion of six possible new species. Appl Environ Microbiol
56:796–802.
87. Orrison LH, Cherry WB, Tyndall RL, Fliermans CB,
Gough SB, Lambert MA, McDougal LK, Bibb WF, Bren-
ner DJ. 1983. Legionella oakridgensis: unusual new species
isolated from cooling tower water. Appl Environ Microbiol
45:536–545.
88. Thacker WL, Wilkinson HW, Benson RF. 1983. Com-
parison of slide agglutination test and direct immunofluor-
escence assay for identification of Legionella isolates. J Clin
Microbiol 18:1113–1118.
89. Diogo A, Verissimo A, Nobre MF, da Costa MS. 1999.
Usefulness of fatty acid composition for differentiation of
Legionella species. J Clin Microbiol 37:2248–2254.
90. Yong SF, Tan SH, Wee J, Tee JJ, Sansom FM, Newton
HJ, Hartland EL. 2010. Molecular detection of Legionella:
moving on from mip. Front Microbiol 1:123.
91. Benson RF, Fields BS. 1998. Classification of the genus
Legionella. Sem Resp Infect 13:90–99.
92. Joly JR, McKinney RM, Tobin JO, Bibb WF, Watkins
ID, Ramsay D. 1986. Development of a standardized
subgrouping scheme for Legionella pneumophila sero-
group 1 using monoclonal antibodies. J Clin Microbiol 23:
768–771.
93. Boissinot M, Ramsay D, Barthe C, Joly JR. 1987. Anti-
genic variability of the outer membrane antigens of Legion-
ella pneumophila serogroups 1 to 8. Can J Microbiol 33:
607–613.
94. Ratcliff RM, Lanser JA, Manning PA, Heuzenroeder
MW. 1998. Sequence-based classification scheme for the
genus Legionella targeting the mip gene. J Clin Microbiol
36:1560–1567.
95. Ratcliff RM. 2013. Sequence-based identification of
Legionella. Meth Mol Biol 954:57–72.
96. Guyard C, Low DE. 2011. Legionella infections and
travel associated legionellosis. Travel Med Infect Dis 9:
176–186.
97. Luck PC, Helbig JH, Ehret W, Marre R, Witzleb W.
1992. Subtyping of Legionella pneumophila serogroup 1
strains isolated in Germany using monoclonal antibodies.
Zentral Bakteriol 277:179–187.
98. Scaturro M, Losardo M, De Ponte G, Ricci ML. 2005.
Comparison of three molecular methods used for subtyping
of Legionella pneumophila strains isolated during an epi-
demic of legionellosis in Rome. J Clin Microbiol 43:
5348–5350.
99. Jonas D, Meyer HG, Matthes P, Hartung D, Jahn B,
Daschner FD, Jansen B. 2000. Comparative evaluation
of three different genotyping methods for investigation of
nosocomial outbreaks of Legionnaires’ disease in hospitals.
J Clin Microbiol 38:2284–2291.
100. Farhat C, Mentasti M, Jacobs E, Fry NK, Luck C. 2011.
The N-acylneuraminate cytidyltransferase gene, neuA, is
heterogenous in Legionella pneumophila strains but can be
used as a marker for epidemiological typing in the consen-
sus sequence-based typing scheme. J Clin Microbiol 49:
4052–4058.
101. Ratzow S, Gaia V, Helbig JH, Fry NK, Luck PC. 2007.
Addition of neuA, the gene encoding N-acylneuraminate
cytidylyl transferase, increases the discriminatory ability
of the consensus sequence-based scheme for typing Legion-
ella pneumophila serogroup 1 strains. J Clin Microbiol 45:
1965–1968.
102. Fry NK, Bangsborg JM, Bergmans A, Bernander S, Eti-
enne J, Franzin L, Gaia V, Hasenberger P, Baladron Jime-
nez B, Jonas D, Lindsay D, Mentula S, Papoutsi A,
Struelens M, Uldum SA, Visca P, Wannet W, Harrison
TG. 2002. Designation of the European Working Group
on Legionella Infection (EWGLI) amplified fragment
length polymorphism types of Legionella pneumophila
serogroup 1 and results of intercentre proficiency testing
Using a standard protocol. Eur J Clin Microbiol Infect Dis
21:722–728.
103. Bates MN, Maas E, Martin T, Harte D, Grubner M, Mar-
golin T. 2000. Investigation of the prevalence of Legionella
species in domestic hot water systems. N Z Med J 113:
218–220.
104. Nazarian EJ, Bopp DJ, Saylors A, Limberger RJ, Musser
KA. 2008. Design and implementation of a protocol for the
detection of Legionella in clinical and environmental sam-
ples. Diagn Microbiol Infect Dis 62:125–132.
105. Miller LA, Beebe JL, Butler JC, Martin W, Benson R,
Hoffman RE, Fields BS. 1993. Use of polymerase chain
reaction in an epidemiologic investigation of Pontiac fever.
J Infect Dise 168:769–772.
106. Chang CW, Chou FC. 2011. Methodologies for quantify-
ing culturable, viable, and total Legionella pneumophila in
indoor air. Indoor Air 21:291–299.
107. Joly P, Falconnet PA, Andre J, Weill N, Reyrolle M, Van-
denesch F, Maurin M, Etienne J, Jarraud S. 2006. Quan-
titative real-time Legionella PCR for environmental water
samples: data interpretation. Appl Environ Microbiol 72:
2801–2808.
108. Dupuy M, Mazoua S, Berne F, Bodet C, Garrec N, Her-
belin P, Menard-Szczebara F, Oberti S, Rodier MH, Sor-
eau S, Wallet F, Hechard Y. 2011. Efficiency of water
disinfectants against Legionella pneumophila and Acantha-
moeba. Water Res 45:1087–1094.
109. Loret JF, Robert S, Thomas V, Cooper AJ, McCoy
WF, Levi Y. 2005. Comparison of disinfectants for
biofilm, protozoa and Legionella control. J Water Health 3:
423–433.
110. Krojgaard LH, Krogfelt KA, Albrechtsen HJ, Uldum SA.
2011. Cluster of Legionnaires disease in a newly built
block of flats, Denmark, December 2008–January 2009.
Euro Surveill 16:19759.

111. Mouchtouri VA, Goutziana G, Kremastinou J, Hadji-
christodoulou C. 2010. Legionella species colonization in
cooling towers: risk factors and assessment of control meas-
ures. Am J Infec Cont 38:50–55.
112. Hambidge A. 2001. Reviewing efficacy of alternative water
treatment techniques—Part 2. Health Estate 55:24–26.
113. Hambidge A. 2001. Reviewing efficacy of alternative water
treatment techniques. Health Estate 55:23–25.
114. Ricketts KD, Joseph CA, EWGLI. 2006. The impact of
new guidelines in Europe for the control and prevention
of travel-associated Legionnaires’ disease. Int J Hyg Envir
Health 209:547–552.
115. Fields BS, Moore MR. 2006. Control of legionellae in
the environment: a guide to the US guidelines. ASHRAE
Trans 112:691–699.
116. Wingender J, Flemming HC. 2011. Biofilms in drinking
water and their role as reservoir for pathogens. Int J Hyg
Envir Health 214:417–423.
117. Declerck P. 2010. Biofilms: the environmental playground
of Legionella pneumophila. Environ Microbiol 12:557–566.
118. Pagnier I, Merchat M, La Scola B. 2009. Potentially
pathogenic amoeba-associated microorganisms in cooling
towers and their control. Fut Microbiol 4:615–629.
119. Berry D, Xi C, Raskin L. 2006. Microbial ecology of
drinking water distribution systems. Curr Opin Biotech 17:
297–302.
120. Marchesi I, Cencetti S, Marchegiano P, Frezza G, Borella
P, Bargellini A. 2012. Control of Legionella contamination
in a hospital water distribution system by monochloramine.
Am J Infect Cont 40:279–281.
121. Moore MR, Pryor M, Fields B, Lucas C, Phelan M, Besser
RE. 2006. Introduction of monochloramine into a muni-
cipal water system: impact on colonization of buildings by
Legionella spp. Appl Environ Microbiol 72:378–383.
122. Pryor M, Springthorpe S, Riffard S, Brooks T, Huo Y,
Davis G, Sattar SA. 2004. Investigation of opportunistic
pathogens in municipal drinking water under different
supply and treatment regimes. Water Sci Technol 50:83–90.
123. ASHRAE. 2000. ASHRAE Guideline 12–2000—Minimiz-
ing the Risk of Legionellosis Associated with Building Water
3.2.9. Legionellae and Legionnaires’ Disease ▪ 3.2.9-13
Systems. Report No. 12-2000. American Society for
Heating, Refrigeration, and Air Conditioning Engineers,
Atlanta, GA.
124. Mastro TD, Fields BS, Breiman RF, Campbell J,
Plikaytis BD, Spika JS. 1991. Nosocomial Legionnaires’
disease and use of medication nebulizers. J Infect Dis 163:
667–671.
125. CTI. 2012. Legionellosis Guideline: Best Practices for
Control of Legionella (WTP-148) (08). Cooling Technol-
ogy Institute. Available from: http://www.cti.org/cgi-bin/
download.pl.
126. CDC. 2012. Disinfection of hot tubs contaminated with
Legionella. Available from: http://www.cdc.gov/legionella/
files/hot-tub-disinfection.pdf.
127. CDC. 2007. Surveillance for travel-associated legionnaires
disease—United States, 2005–2006. Morbid Mortal Weekly
Rep 56:1261–123.
128. Perola O, Kauppinen J, Kusnetsov J, Karkkainen UM,
Luck PC, Katila ML. 2005. Persistent Legionella pneumo-
phila colonization of a hospital water supply: efficacy of con-
trol methods and a molecular epidemiological analysis.
APMIS 113:45–53.
129. Marchesi I, Marchegiano P, Bargellini A, Cencetti S,
Frezza G, Miselli M, Borella P. 2011. Effectiveness of
different methods to control Legionella in the water supply:
ten-year experience in an Italian university hospital. J Hosp
Infect 77:47–51.
130. Molloy SL, Ives R, Hoyt A, Taylor R, Rose JB. 2008. The
use of copper and silver in carbon point-of-use filters for
the suppression of Legionella throughput in domestic water
systems. J Appl Microbiol 104:998–1007.
131. Daeschlein G, Kruger WH, Selepko C, Rochow M,
Dolken G, Kramer A. 2007. Hygienic safety of reusable
tap water filters (Germlyser) with an operating time of 4
or 8 weeks in a haematological oncology transplantation
unit. BMC Infect Dis 7:45.
132. Sheffer PJ, Stout JE, Wagener MM, Muder RR. 2005.
Efficacy of new point-of-use water filter for preventing
exposure to Legionella and waterborne bacteria. Am J Infect
Cont 33(5 Suppl 1):S20–S25.
 Pathogenic Viruses and Protozoa Transmitted by Soil
PASCAL DELAQUIS, JULIE BRASSARD, AND ALVIN GAJADHAR
3.3.1
HUMAN PATHOGENS ASSOCIATED WITH
SOIL AND THEIR PUBLIC HEALTH
SIGNIFICANCE
Soil ecosystems support diverse and complex communities of
prokaryotic and eukaryotic species. While the vast majority
are harmless to humans, some have significant roles in the eti-
ology of noninfectious, infectious, or parasitic human diseases
with a range of acute, chronic, or recurrent symptoms and out-
comes. Eukaryotic helminths including hookworms are likely
the most common global cause of infections transmitted
by soil (1). Improvements in sanitary infrastructure and
enhanced access to anthelminthic drugs have greatly reduced
rates of infection, but their impact on public health in impov-
erished or developing areas remains considerable (2).
Endemic diseases caused by soilborne fungi are also rare in
the developed world, but it should be noted that infections
that were uncommon until immunosuppressive agents and
drugs became widely used in the second half of the 20th cen-
tury are growing in importance (3). In contrast, diseases
caused by soilborne pathogenic viruses and protozoa, which
are the focus of this chapter, are reported worldwide and
include growing threats to public health. Although some of
the associated species appear to be ubiquitous in soils, epide-
miological data reveal spatial and temporal differences in
both global and regional disease burdens. Historical variabil-
ity in the risk of transmission and the incidence of disease is
shaped by complex socioeconomic factors that include the
scope of public health measures, availability of medical
care, economic development, and cultural practices, among
others. In addition, environmental pressures associated with
rapid development and population growth, notably the dis-
posal of industrial and sanitary wastes generated by humans
or livestock, are known to affect the incidence of infectious
diseases (4). There is increasing concern that the purposeful
application of manure or biosolids from wastewater treatment
plants and declining irrigation water quality are likely increas-
ing the role of soil as a reservoir for human pathogens (5). The
public health burden associated with specific soilborne
human pathogens may be difficult to measure accurately
due to historic or geographic divergence in the prevalence
of illnesses and inconsistencies in reporting between jurisdic-
tions. Furthermore, innate difference in the pathogenicity of
individual species and variable clinical manifestations may
conceal the role of soil in disease transmission.
doi:10.1128/9781555818821.ch3.3.1
3.3.1-1
Microorganisms that normally inhabit soil habitats may be
called edaphic (from the Greek for ground or soil). While no
strictly edaphic viral pathogens have been identified to date,
protozoa with known or suspected edaphic lifestyles are
clearly capable of causing disease in humans. In addition,
soil can act as a reservoir for a diversity of transient pathogenic
viruses and protozoa, notably where soil quality has been
affected by animal or human activity.
Viruses
Viruses are primarily transmitted by the fecal–oral route of
transmission. In some cases very low infectious doses are
needed to cause infection in the human host. Viruses behave
as inert particles outside their host due to their inability to
multiply. However, replication within the cells of a host’s gas-
trointestinal system can lead to the shedding of very large
titers in feces.
Enteric species of noroviruses, rotaviruses, adenoviruses,
astroviruses, enteroviruses, and the hepatitis A and E viruses
cause a range of human illnesses, including gastroenteritis,
diarrhea, hepatitis, conjunctivitis, meningitis, encephalitis,
and diseases of the nervous system. Once underdiagnosed,
enteric viruses are now recognized as a major cause of food
poisonings and are suspected to cause 50 million infections
in the United States annually (6, 7). Although shellfish
and ready-to-eat foods were traditionally considered most
susceptible to viral contamination, there has been an upswing
in poisonings associated with unprocessed foods, notably
fresh fruits and vegetables (8, 9). Epidemiological studies
indicate that fresh foods can be contaminated during pro-
duction, harvest, postharvest handling, or transportation
(10). Irrigation water, soil, and water used in postharvest
processing have been suggested as primary sources of contam-
ination (11, 12). Enteric viruses can persist in soil due to
their inherent capacity to withstand extremes of pH, heat,
or freezing (13, 14).
Noroviruses
Noroviruses (NoVs) are nonenveloped, positive-sense,
single-stranded RNA viruses with a very stable icosahedral
capsid that protects a genome of 7.5–7.7 kb. NoVs belong
to the Norovirus genus of the Caliciviridae family, which is
subdivided into five genogroups (GI, GII, GIII, GIV, and
GV). The majority of strains pathogenic to humans belong
3.3.1-2 ▪ SOIL
to genogroup GII, although a few from genogroups GI and
GIV have been reported. Many different strains and geno-
types are known, but GII genotype 4 (GII.4) is currently
the most prevalent. In the past decade, strains of genotype
GII.4 alone have been caused three global gastroenteritis
pandemics (15). NoVs are the leading cause of foodborne
outbreaks and sporadic acute gastroenteritis (16). The symp-
toms of NoV infection generally include diarrhea, dehydra-
tion, vomiting, nausea, headache, abdominal pain, and
sometimes fever. The symptoms last for two to three days, dur-
ing which an individual sheds up to 100 billion virus particles
per gram of fecal matter (17). Infection occurs primarily by
ingestion of contaminated food or water (18, 19), although
airborne transmission of NoV through vomitus particles has
been demonstrated (20, 21). A very small number of infec-
tious particles (fewer than 10) may cause infection (22). In
the United States, 33% of epidemics linked to the consump-
tion of contaminated fresh foods are associated with NoV
(23), which are also responsible for 90% of cases of nonbac-
terial gastroenteritis (18, 24). Long-term immunity to NoV
infection does not appear possible, and individuals may be
infected repeatedly.
Rotaviruses
Rotaviruses belong to the Rotavirus genus of the Reoviridae
family. They are nonenveloped viruses with a double-
walled capsid containing segmented double-stranded
RNA and a complex genome of about 18.5 kb. Rotaviruses
are subdivided into five distinct groups (A–E) (25). Group
A rotaviruses (GARVs) are a primary cause of severe gas-
troenteritis in infants and young children around the world
(26, 27). GARVs are estimated to cause 600,000 deaths
annually in developing areas and thousands of hospitaliza-
tions in industrialized countries (28). Symptoms of rotavi-
rus infection include vomiting and watery diarrheas that
are often accompanied by fever and abdominal pain lasting
for three to eight days. The virus can remain active on sur-
faces for extended periods of time, thereby promoting the
spread of infection (29). Stability in the environment
also favours transmission through food and water (30).
Zoonotic transmission of GARVs has been suggested given
widespread incidence of infection in wild or domesticated
animals and that genetic reassortment between human
and animal strains has been observed (31–33). The recent
introduction of vaccination against GARV in humans
could potentially promote the emergence of new strains
from animal reservoirs or the reemergence of old human
strains (34).
Hepatitis A Virus
The hepatitis A virus (HAV) belongs to the Hepatovirus genus
of the Picornaviridae family. This virus has a nonenveloped
icosahedral capsid with a positive-sense, single-stranded
RNA genome of 7.5 kb. The incidence of hepatitis A varies
considerably around the world and is much higher in develop-
ing countries where sanitary conditions and wastewater
treatment are poor (35). HAV infections acquired by inges-
tion of contaminated food or water may result in severe
illnesses. The symptoms include fever, malaise, nausea, vom-
iting, dark urine, and jaundice. Infection is generally asymp-
tomatic in children but may lead to complications in older
individuals, where it is associated with chronic liver patholo-
gies. The virus is able to survive in different environments,
such as water, food, and wastewater treatment plant dis-
charges (36).
Hepatitis E Virus
The hepatitis E virus (HEV) belongs to the Hepevirus genus
of the Hepeviridae family and is likely responsible for the
vast majority of cases of non–A, non–B hepatitis world-
wide. It is a nonenveloped, positive-sense, single-stranded
RNA virus of 7.2 kb that is subdivided into four distinct
genogroups (1–4) (37). Genogroups 1 and 2 are found
mainly in humans in Asia and certain developing countries
(38, 39). Animals such as swine appear to be reservoirs for
genogroups 3 and 4, and cases of direct transmission to
humans have been reported (38, 40–42). It has also been
suggested that these genogroups could be responsible for
increasing rates of infection with HEV in industrialized
countries (43–46). Contaminated water and food appear
to be the primary and most common modes of transmission,
given that person-to-person transmission accounts for only
0.7–0.8% of infection (47, 48). The symptoms of HEV
infection include nausea, general malaise, fever, dark urine,
and jaundice. Young children appear to be asymptomatic
(38). However, the mortality rate in pregnant women
approaches 20% versus 1% in the general population
(49). Immunosuppressed individuals are more susceptible
to HEV infection and chronic liver sequelae (50–52). Fre-
quent contamination of drinking water supplies in endemic
regions is indicative of HEV stability in the environment.
The recent demonstration of HEV on strawberries irrigated
with surface water provides evidence that this likely con-
tributes to indirect routes of infection via foods produced
or washed in contaminated water (53).
Other Enteric Viruses
Many other enteric viruses are known to cause gastroenteritis,
encephalitis, meningitis, conjunctivitis, and respiratory
illnesses in animals and humans, including astroviruses,
adenoviruses, caliciviruses, enteroviruses, toroviruses, coro-
naviruses, kobuviruses, and picobirnaviruses. In most cases
little is known about their prevalence, distribution, and
impact on public health. Most are primarily transmitted by
person-to-person contact, through fecal matter and respira-
tory secretions. Because they are shed in large quantities in
fecal matter, these viruses may be found in surface waters or
wastewater used intensively for irrigation and in biosolids or
animal manures applied as fertilizers (54–56). Consequently,
a role for soil in the transmission of diseases caused by these
pathogens cannot be discounted.
Protozoa
The protozoa include a large group of eukaryotic unicellular
free-living, commensal, mutualistic, or parasitic heterotro-
phic microorganisms with variable, potentially complex,
and often highly specific life cycles. Many protozoan species
are ubiquitous in soils where they contribute immensely to
soil diversity and fertility. Parasitic species that are consid-
ered pathogenic for animals or humans are distributed across
many genera and are released with feces from hosts to enter
the environment. Some are geographically endemic or con-
sidered as emerging, particularly in the developed world,
where the globalization of travel and food distribution sys-
tems, changing agricultural practices, and climate change,
among other factors, are probably contributing to the
exploitation of native hosts and the evolution of alternative
transmission routes. Protozoa employ a variety of means for
locomotion, including movement by means of cytoplasmic
extensions called pseudopodia in amebae, propulsion by
the oar-like motion of numerous cilia which protrude
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-3
TABLE 1 Selected pathogenic protozoa
Protozoan group Species
Amoebae (Sarcodina) Acanthamoeba spp.
Entamoeba histolytica
Naegleria fowleri
Flagellate (Mastrigophora) Giardia spp.
Ciliate (Ciliophora) Balantidium coli
Sporozoa (Apicomplexa) Toxoplasma gondii
Cyclospora cayetanensis
Cryptosporidium parvum
from the cell in ciliates, and wave-like motion of long flag-
ella in flagellates. Selected pathogenic protozoan species
considered in this chapter are provided in Table 1.
Amebae
Acanthamoeba spp.
Amebae of the genus Acanthamoeba spp. have been iso-
lated worldwide from a range of environments but prevalence
data suggest that soil is the primary habitat (57). The ubiqui-
tous nature of Acanthamoeba and evidence of interaction
with humans was confirmed through a random survey of
114 individuals from 37 countries which revealed the pres-
ence of anti-Acanthamoeba antibodies in 85% of the subjects
tested (58). Acanthamoeba spp. undergo two stages during
their life cycle. During the vegetative trophozoite stage,
spine-like structures known as acanthopodia protrude from
the cell to assist in adhesion to surfaces, cellular movement,
or capturing prey. Adverse conditions (lack of food, extremes
in temperatures or pH) induce the formation of the second
stage characterized by the formation of a double-walled resist-
ant cyst that can remain viable for several years in distilled
water and up to 20 years in the dry state (59, 60).
Ten of the 25 known species of Acanthamoeba are known
to cause infections in humans. These include infections of the
skin, nasal passages, brain, and lung in immunocompromised
individuals, and rarely but usually fatal granulomatous amebic
encephalitis (61, 62). Acanthamebic keratitis, the most com-
mon disease caused by Acanthamoeba spp., can lead to corneal
ulcers or blindness. The species A. castellanii, A. polyphaga,
A. hatchetti, A. culbertsoni, A. rhysodes, A. griffini, A. quina,
and A. lugdunensis have been clinically associated with the
disease, which is reported with increasing frequency in con-
tact lens wearers (63). Transmission can occur by contami-
nated water or lens solution, although dust is suspected in
cases affecting non–contact lens wearers. The ability of Acan-
thamoeba to harbor a wide range of viral, bacterial, and protist
endosymbionts led one author to describe the species as the
“Trojan horse of the microbial world” (61). Analysis of soil
and sand by coculture with A. polyphaga by Evstigneeva
et al. (64) yielded 33 bacterial species with the ability to
grow in amebae, including 20 potentially pathogenic species.
According to one estimate, 102 of 539 known human or ani-
mal bacterial pathogens are able to resist and potentially pro-
liferate within amebae (65). These potential endosymbionts
include several soil-associated pathogens such as E. coli
O157:H7, Campylobacter spp., and Legionella spp. Hence,
the role of free-living amebae such as Acanthamoeba spp. in
the transmission of other microbial pathogens may be under-
estimated or unrecognized.
Entamoeba hystolitica
The genus Entamoeba contains numerous morphologically
similar amebae differentiated on the basis of cyst size, number
of cyst nuclei, appearance of chromatoid bars, and host spe-
cies. At least six species are found in man including Enta-
moeba histolytica, Entamoeba dispar, Entamoeba moshkovskii,
and Entamoeba coli. Four appear to be harmless, but E. histo-
lytica causes amebiasis (amebic dysentery) and a wide range
of other invasive diseases in humans, including amebic liver
abscess, respiratory tract infections, cerebral infections, and
genitourinary infections. In parallel with Acanthamoeba,
E. histolytica has a simple lifestyle consisting of a vegetative
trophozoite stage and an infectious cyst stage. The latter
enters the host through the mouth, survives the acid in the
stomach, travels through the small intestine, and excysts in
the terminal ileum or colon to form the trophozoite stage
(66). Trophozoites can adhere to colonic epithelial cells,
inducing the development of multiple ulcers, inflammation
of the mucosa, necrosis, and perforation of the intestinal
wall, leading to frequent episodes of mucousy or bloody diar-
rhea characteristic of amebiasis. Spread of the highly motile
trophozoites from the intestine via the bloodstream can
lead to amebic liver abscess and other extraintestinal lesions.
Infected individuals shed large numbers of mature cysts.
Amebic dysentery contributes significantly to global mor-
bidity and mortality associated with diarrheal diseases, and
according to some estimates E. histolytica is the third leading
etiologic agent of parasitic disease after Plasmodium and Schis-
tosoma (67). E. histolytica is found in every region of the world,
but infection rates are highest in socioeconomically deprived
populations in tropical or subtropical regions where sanita-
tion is poor. Infection is believed to occur primarily by the
ingestion of contaminated food or water. Cysts are readily
detected in water and soil from endemic regions, where
they are believed to survive for long periods, although sur-
vival in soil has not been examined in detail. Furthermore,
analysis of risk factors for infection occasionally leads to con-
tradictory conclusions. For example, a study in Brazil identi-
fied place of residence, age, ingestion of raw vegetables, and
drinking water quality as important risk factors (68). In con-
trast, socioeconomic and personal hygiene factors (hand
washing) rather than exposure to human and animal excreta
or agricultural activities determined the likelihood of infec-
tion with E. histolytica in a study carried out in Vietnam
(69). Consistent recovery of E. histolytica from fresh produce
grown in endemic regions suggests that food crops could serve
as a vector for transmission (70). However, it remains uncer-
tain whether contamination occurs primarily by transfer of
the pathogen from contaminated water/soil to the crop during
cultivation or during unsanitary handling of the crop at or
after harvest.
Naegleria fowleri
The free-living ameba Naegleria fowleri is sometimes
referred to as the “brain-eating ameba.” This evocative desig-
nation is derived from the symptoms associated with primary
amebic meningoencephalitis, a rare and usually fatal acute
hemorrhagic disease caused by N. foeleri. Early symptoms
include the sudden onset of headaches, high fever, stiff
neck, nausea, vomiting, irritability, and restlessness. Later
symptoms may include photophobia, lethargy, confusion,
bizarre behavior, and seizures, followed by coma and death.
N. fowleri amebae infect the host through the nose and
migrate to the brain via the olfactory nerve. Using sucker-like
appendages or amebostomes extending from the cell surface,
the amebae “feed” on infected tissues, resulting in significant
3.3.1-4 ▪ SOIL
necrosis and hemorrhaging, particularly in the olfactory
bulbs. The cysts can conceivably convert to trophozoites
inside the host if they are inhaled. Primary amebic menin-
goencephalitis has occurred in several countries around the
world, and N. fowleri appears to be widely distributed in
warm waters, ponds, lakes, rivers, hot springs, and near warm-
water discharges of industrial plants (71). This suggests a pri-
marily aquatic habitat, but the species is also found in soils
associated with contaminated water bodies (72).
The life cycle of N. fowleri includes three stages. The free-
feeding trophozoite stage is presumably infective. Tropho-
zoites can undergo transformation to a flagellate stage under
conditions of nutrient deprivation in water and encyst under
adverse environmental conditions. Little is known about the
prevalence of each stage in the natural environment. The
majority of illnesses reported to date appear to be waterborne,
although acquisition by inhalation of soil or mud has been
proposed in cases where infection could not be associated
with direct exposure to contaminated water. Hence available
evidence hints that the species may survive in moist or wet
soils and that the latter should be considered a vector for
the transmission of the pathogen.
Flagellate: Giardia spp.
The taxonomy of the flagellated protozoan Giardia is some-
what unclear. Six species are generally recognized: G. agilis,
G. ardeae, G. muris, G. microti, and G. psittaci, which infect
various animals, and G. duodenalis, which infects mammals
including humans. The latter is sometimes referred to as
G. intestinalis or G. lamblia. G. duodenalis is now considered
a complex consisting of eight host specific assemblages (A
to H). G. duodenalis from assemblages A and B can infect
humans and other mammals.
Infection in humans can lead to giardiasis (sometimes
referred to as beaver fever). The disease is characterized by
the onset of fatty, yellowish, foul-smelling diarrhea, weakness,
weight loss, abdominal cramps, flatulence, abdominal disten-
sion, and occasionally nausea, vomiting, and fever. The ill-
ness is rarely fatal and usually lasts from two to four weeks.
A proportion of infected individuals (estimated at 30–50%)
develop a long-term illness for which treatment is ineffectual,
continuing to cause the patient discomfort or pain, sequelae
that seriously affect quality of life. Children may suffer more
serious consequences, including retarded growth and devel-
opment and poor cognitive function (73). Infection of
human populations with G. duodenalis is common and global
in occurrence. Infection rates rarely exceed 10% in developed
countries but may reach 30% in some countries in Asia
(Bangladesh, Cambodia, China, India, Indonesia, Laos,
Malaysia, Nepal, the Philippines, Thailand, Turkey, Saudi
Arabia, Vietnam), Central America (Cuba, Mexico, Nicara-
gua), South America (Argentina, Brazil, Colombia, Peru),
and Africa (northern Africa, west Africa, South Africa)
(74). The simple life cycle of Giardia includes a dividing
trophozoite stage and an environmentally resistant infective
cyst stage. Ingested cysts pass into the duodenum, where
excystation occurs, releasing two trophozoites that multiply
rapidly via binary fission. Clinical symptoms occur as a result
of damage to the mucous membrane.
Ciliate: Balantidium coli
Balantidium coli is the only ciliated protozoan known to infect
humans. Infection can lead to balantidiasis, an uncommon
and rarely fatal disease with a range of clinical presentations,
including asymptomatic carriage, chronic infection that
results in nonbloody diarrhea, cramping, halitosis, and
abdominal pain, to fulminating balantidiosis with mucoid,
bloody stools (75). Intestinal hemorrhaging, ulceration, and
perforation appear to be mediated by the production of pro-
teolytic enzymes (76). Balantiasis is reported sporadically
worldwide, and is endemic in rural regions of Latin America,
the Philippines, Papua New Guinea, and West Irian. Epide-
miological data from endemic regions has shown that proxim-
ity to pigs and pig feces are significant risk factors for
balantiasis in humans. The species is widely distributed in
wild and domesticated pigs, which are asymptomatic hosts.
For example, B. coli was detected in 55.1% of swine surveyed
in Oklahoma (77). B. coli has a relatively simple life cycle.
Cysts with thickened walls that offer protection from desicca-
tion and other environmental stress are shed in the feces of
swine or infected humans. Cysts ingested with food or water
survive the stomach and excyst in the colon where the motile
trophozoite penetrates into the mucosa layer of the large
intestine. Encystation of the trophozoite also occurs in the
colon and the rectum, resulting in the release of infectious
cysts to complete the cycle (76). B. coli oocysts are known
to survive well in feces and have been isolated from water
and soil in pig production areas (78). These findings hint at
the possibility of persistence in the natural environment but
little is known about factors that favor long-term survival of
oocysts in water or soil.
Sporozoa
Cryptosporidium spp.
Within the oocyst-forming coccidian protozoa the genus
Cryptosporidium currently includes 16 species and an addi-
tional 33 distinct genotypes of oocyst-forming coccidian pro-
tozoa of which a few are zoonotic parasites. C. parvum,
C. muris, C. hominis, C. felis, C. andersoni, and C. wrairi are
known to infect mammals; C. baileyi and C. meleagridis birds;
C. serpentis reptiles; and C. nasorum tropical fish. C. parvum
and C. hominis are most commonly associated with human
infection. Unlike C. hominis, which is almost exclusively a
parasite of humans, C. parvum has a broad host range and
causes infection in humans, farm animals (notably cattle),
companion animals, and rodents. Although isolates from
individual Cryptosporidium species are morphologically and
developmentally similar, many are host-specific. The com-
plex life cycle of C. parvum and C. hominis is completed
within a single host, culminating in the shedding of mature,
immediately infective sporulated oocysts in the feces. Oocysts
contain four crescent-shaped infective sporozoites that are
released on excystation in the small intestine. The sporozoites
penetrate the gut epithelium, initiating an infection and
progress to multiple cycles of asexual replication followed by
sexual multiplication to yield oocysts which sporulate in
situ. Infection in humans can lead to acute short-term disease
which has the potential to become severe and resist treatment
in children and immunocompromised individuals. Profuse
diarrhea is the most common symptom of infection, followed
by abdominal pain and vomiting. Cryptosporidiosis typically
has a longer symptomatic period than bacterial gastrointesti-
nal infections, lasting from one to two weeks (79). Severe and
persistent infections are common in immunocompromised
individuals, with invasion of other organ systems including
the lungs and the bile duct (80). Cryptosporidiosis is consid-
ered an emergent disease in industrialized countries where the
rate of reporting has increased over the past 20 years, but chil-
dren in the developing world are disproportionately affected.
Infection with Cryptosporidium is one of the leading causes of
child mortality associated with diarrheal diseases (67).
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-5
Lozano et al. (67) reported a decline in death rates between
1990 and 2010, but the lack of rapid diagnostic methods,
effective vaccines, or treatments for immunocompromised
patients suggest that the burden of disease is likely to remain
significant for the near future.
Cryptosporidiosis occurs in every region of the world (81,
82). However, differences in the global, regional, and sea-
sonal distribution of Cryptosporidium species have emerged
through macroepidemiological analysis of clinical specimens
using molecular tools that differentiate at the species/
genotype and subtype levels. C. hominis is more prevalent
in North and South America, Australia, and Africa, whereas
C. parvum causes more human infections in Europe and the
Middle East. C. hominis is usually the predominant species
in humans in developing countries (83). Recurrent outbreaks
associated with poor water quality sustained the long-held
view that cryptosporidiosis is a waterborne illness. Increasing
reports of transmission by other routes have shaped a broader
perspective that acknowledges the importance of other
pathways, including food and soil. Several outbreaks of cryp-
tosporidiosis have been linked to the consumption of conta-
minated food, notably fresh fruits and vegetables. Isolation of
infective oocysts from produce has been reported in many
countries (84–86). Oocysts resist environmental stresses in
soil and retain infectivity better than many other pathogens.
Differences in physiochemical and biological soil properties
influence stability and experimental outcomes tend to vary,
but available evidence indicates that they can survive for
months (87). Consequently, irrigation with water of poor
microbiological quality, runoff due to rain or floods, and the
accidental or purposeful application of fecal material may
contribute high concentrations of infective Cryptosporidium
oocysts to soil. Contamination of vegetables by viral and
bacterial pathogens in soil has been demonstrated experimen-
tally (88, 89). Although direct evidence is lacking, there is
a strong probability that Cryprosporidium oocysts are also
transferred from soil to growing crops.
Cyclospora cayetanensis
The genus Cyclospora includes 19 cyst-forming coccidian
species, including Cyclospora cayatenensis, which only infects
humans (90). That humans are host to C. cayetanensis and
that infection can lead to disease are comparatively recent
findings derived from cohort studies carried out in Peru in
the late 1980s (91). It is now recognized as the etiological
agent responsible for cyclosporiasis, an intestinal illness char-
acterized by prolonged watery diarrhea, abdominal cramping,
weight loss, anorexia, myalgia, and occasionally vomiting
and/or fever that can last for several weeks, with bouts of
remittance and relapse. The severity of illness is variable
depending on the age of the host and size of the infective
dose. Cyclosporiasis has been reported in the Americas, the
Caribbean, Europe, Australia, Asia, and Africa. In industrial-
ized countries the disease has a strong association with travel
to or the consumption of foods imported from endemic
regions (92, 93). In contrast, infection with C. cayetanensis
is widespread in some developing countries, notably Haiti,
Guatemala, Peru, and Nepal (94). Little is known about
the life cycle of C. cayetanensis or the pathogenic mechanisms
responsible for the clinical manifestations of the disease.
Unsporulated oocysts secreted in the feces of infected individ-
uals sporulate in the environment. Ingested sporulated
oocysts release sporozoites that invade and replicate in upper
intestinal tract enterocytes in a similar manner to other coc-
cidians. The formation of coccidian oocysts allow for their
high resistance to many environmental conditions (95).
Cyclospora oocysts appear to retain infectivity for long periods
of time in the environment and are resistant to chlorine treat-
ments used in water purification (96). Direct contact with
contaminated soil has been identified as a risk factor for cyclo-
sporiasis (97). An indirect link with contaminated soil was
suspected in an outbreak associated with consumption of
leafy vegetables (98). However, survival of C. cayetanensis
oocysts in soil has not been examined in detail.
Toxoplasma gondii
The genus Toxoplasma currently contains a single
species, Toxoplasma gondii, a ubiquitous obligate intracellular
protozoan that is estimated to infect one third of the world
population (99). The majority of human infections are
asymptomatic, but severe disease can occur in immunocom-
promised individuals and newborns who acquire the disease
in utero from infected mothers. In addition to abortion, still-
birth, or early mortality, clinical symptoms in surviving
infants range from jaundice, rash, and hepatosplenomegaly
(enlargement of the liver and spleen) to developmental
abnormalities, decreased visual acuity, cerebral calcifications,
hydrocephalus or microcephaly, and psychomotor retarda-
tion. In immunocompromised individuals central nervous
system toxoplasmosis can lead to intracranial lesions or ence-
phalitis, and ocular toxoplasmosis leading to retinal lesions
that can result in poor or permanent loss of vision (100).
Wild or domesticated cats are the definitive host for
T. gondii, and warm-blooded wild animals, livestock, and
birds can serve as intermediate hosts. Cats are usually infected
by ingestion of encysted bradyzoites in the tissues of an
infected intermediate host. The infectious bradyzoites are
released into the intestinal lumen, where they penetrate the
intestinal wall and replicate throughout the body as rapidly
dividing tachyzoites. Simultaneous invasion and replication
within the intestinal epithelial cells results in sexual repro-
duction and the formation of oocysts which are excreted in
the feces of cats. Sporulation (formation of infectious sporo-
zoites within the oocyst) may take 1–5 days depending
on environmental conditions. In contrast with infection in
cats, oocyst production does not occur in intermediate hosts.
However, ingestion of encysted bradyzoites, in insufficiently
cooked meat, for example, can trigger infection. The cocci-
dian oocysts of T. gondii are very resistant and may persist in
the environment for extended periods of time. A study
designed to establish rates of survival in wet and dry soil
showed that oocysts remained infective for more than 400
days in wet soil (101). Contamination of soil by cat feces
can be highly localized and represent an enduring source of
infectious oocysts (102).
CONVENTIONAL METHODS FOR THE
DETECTION OF SOILBORNE PATHOGENS
Elution, Concentration, and Enumeration of
Soilborne Viruses
Efficient recovery is a key step in the analysis of soil for the
presence of viruses. The goal of the analysis is the separation
of the viral particles from the sample matrix and their concen-
tration in an aqueous solution free of contaminants to permit
detection in laboratory-grown cell lines. Viruses naturally
carry a negative surface charge that promotes adsorption to
soil components through electrostatic interactions. Adsorp-
tion to soil particles is a reversible process influenced by soil
composition (organic matter, clay content) and pH (103–
105). Consequently, elution buffers of defined composition
3.3.1-6 ▪ SOIL
with respect to pH, ionic strength, and organic content are
designed to promote desorption of the viruses from soil par-
ticles. Beef extract is widely used as a source of organic matter.
Other eluent formulations may contain chaotropic mole-
cules, such as glycine, or detergents, such as Tween 80,
SDS, and EDTA. Most are adjusted to high pH to reduce
the electrostatic forces between viruses and soil. A concen-
tration step may be combined with elution for environmen-
tal samples that contain very low concentrations of virus
particles. The concentration step can be done through a
combination of organic flocculation, phase separation, pre-
cipitation, and ultrafiltration.
Various protocols combining elution buffers and concen-
tration methods have been proposed. A widely used reference
protocol was proposed by Hurst et al. (106). Briefly, a 50 g soil
sample is resuspended in 50 ml buffer containing 10% beef
extract, Na2HPO4, and citric acid. The suspension is agitated
for 30 min, and the pH is adjusted to 7.0 ± 0.1 with 1 M HCl
or NaCl. The suspension is then centrifuged at 2,500 × g for
30 min. The supernatant is recovered by successive filtration
through sterile membranes of type AP40, AP25, AP20, and
AP15 (Millipore) to remove soil particles and other microor-
ganisms from the matrix. The resulting supernatant can be
used directly for a virus count in cell culture and may be fur-
ther concentrated or frozen at −80°C. An additional concen-
tration step generally increases the sensitivity of detection.
This can be achieved by adjustment of the beef extract con-
centration to 3% by the addition of approximately 2.3 vol-
umes of sterile water. The suspension is stirred and the pH
is adjusted to 3.5 ± 0.1 with 1 M HCl until a flocculent pre-
cipitate forms, a process that takes approximately 30 min.
The solution is then centrifuged at 1,000 × g for 5 min at 4°
C, and the supernatant is carefully removed. The precipitate
is resuspended in 8–10 ml 0.05 M glycine buffer ( pH 11) to
maintain pH at a minimum of 9.5. Once the precipitate has
dissolved, a final centrifugation is carried out at 1,000 × g
for 10 min at 4°C, and the supernatant is collected.
The pH is adjusted to 7.5, antibiotics (200 U penicillin/ml,
200 µg streptomycin sulfate, 2.5 µg amphotericin B/ml) and
0.5 ml of a 3 M NaCl solution are added, and the solution
is used for the virus count or stored at −80°C until use (106).
The detection and quantification of infectious virus
particles are carried out in laboratory-grown mammalian
cell cultures that support replication of the particles. Quantal
methods of quantification (determination of the 50% tissue
culture infective dose) and plaque assays (determination of
the number of plaque-forming units) are the two conven-
tional methods used to detect and quantify viruses recovered
in the elution and concentration steps. Virus counts are
derived from the number of plaques produced during cultiva-
tion on monolayer cells (1 plaque = 1 viral infection) (107).
The plaques are made visible by the addition of a layer of
semisolid agar–based nutrient medium to the surface of the
cells. In contrast, the quantal method assesses the amount
of virus required to kill 50% of the infected host cells or to
produce a cytopathic effect in 50% of the inoculated cell
monolayer (107). A number of cell lines, such as BGMK (buf-
falo green monkey kidney cells), MA-104 (rhesus monkey
kidney cells), RD (human rhabdomyosarcoma cells),
MRC-5 (human lung fibroblast cells), and A549 (human
lung adenocarcinoma epithelial cells), are used for the quan-
tification of enteroviruses (108). However, not all enteric
viruses can be detected and quantified by these methods. A
number of viruses, including noroviruses, astroviruses, sapovi-
ruses, rotaviruses, and the hepatitis A and E viruses, are very
difficult or impossible to grow in the laboratory because of the
lack of cell lines that support in vitro replication. In addition,
some culturable viruses such as adenoviruses do not produce a
cytopathic effect and cannot be quantified using these
procedures.
Detection of Soilborne Protozoa
Traditional methods for the detection of protozoa in environ-
mental samples are adapted from diagnostic procedures used
in the analysis of clinical specimens (feces, tissues, fluids)
where the target species are often present in sufficient num-
bers to allow detection and identification with a microscope.
Protozoa are usually present in low numbers and are unevenly
distributed in comparatively dry soil matrixes. Consequently,
their detection against the large background microflora and
extraneous components in soil may require a series of prelimi-
nary steps, including dispersion in an aqueous phase, elution,
concentration, and purification by variable methods deter-
mined by the properties of both sample and target protozoa.
Concentration is normally achieved by sedimentation, floc-
culation, filtration, or centrifugation, followed by purification
by density gradient separation, flotation, or immunomagnetic
separation. Table 2 provides methods that have been applied
in the detection of pathogenic protozoan species in soil.
MOLECULAR METHODS FOR THE DETECTION
OF SOILBORNE PATHOGENIC VIRUSES AND
PROTOZOA
Coextraction of PCR Inhibitors from the Matrix
Soil is without question one of the most complex environ-
mental matrixes. The soil environment contains variable
amounts of decaying organic matter, minerals, metals, and
complex microbiota. As a result of advances in molecular
biology, it is now possible to detect, quantify, and study micro-
bial diversity using various methods, such as PCR, conven-
tional reverse-transcription PCR (RT-PCR), real-time PCR
and RT-PCR, and sequencing (109). The direct extraction
of microbial DNA or RNA from soil remains the key step
in the application of molecular methods for the examination
of microbial diversity or the detection of specific species
(110). The extraction and purification of DNA and RNA
of sufficient quality to permit sensitive analysis can be chal-
lenging. Soil constituents can compromise molecular detec-
tion either by inhibiting the activity of Taq polymerase
during DNA amplification by PCR or by binding nucleic
acids during extraction. Phenolic compounds, notably humic
and fulvic acids, are well-known inhibitors of PCR that bind
to DNA and interfere with Taq polymerase (109). Humic and
fulvic acids are frequently coextracted with nucleic acids from
environmental samples (111) and cause false-negative
results. Various methods that can be used alone or in combi-
nation have been proposed for the purification of DNA/RNA
in the presence of humic acid, including agarose gel electro-
phoresis, use of polyvinylpolypyrrolidone (PVPP), size-exclu-
sion chromatography, silica-based column chromatography,
isopropanol precipitation, and polyethylene glycol precipita-
tion (112–118). Dilution of the extracted nucleic acids can
also improve detection by PCR (119).
Commercial kits for the extraction of nucleic acids from
environmental samples that employ various means to remove
potential inhibitors are available. Most are designed for use
with very small samples (0.2–5 g), and their performance
can be limited where target nucleic acid concentrations are
low. Here an indirect approach requiring separation of the
virus particles from the matrix by elution, followed by lysis
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-7

TABLE 2 Methods applicable to the detection of pathogenic protozoa in soil

Target and references Dispersion, concentration, purification
Acanthamoeba spp. Suspend 2 g soil in 20 ml distilled H2O
Entamoeba histolytica 100 g soil in conical bottom cups with 1 liter
detergent solution (0.001% q.s.p.), filtration
through gauze and sedimentation by gravity
(144)
Naegleria fowleri Dispersion in water, sedimentation followed by
application of methods used for water analysis
Swab samples collected by passing sterile gauze
over rocks and soil, resuspension, concentration
by centrifugation (72)
Giardia spp. Dispersion in water, centrifugation and flotation in
sucrose (Sheather’s sugar solution) (146, 147).
Note: after a few minutes in Sheather’s solution
Giardia stages may collapse. Here a zinc sulfate
solution with g.s.p. of 1.18 is may be used instead
of Sheather’s
Balantidium coli Flotation in sucrose solution specific gravity 1.18
Cryptosporidium below (78)
Toxoplasma gondii Dispersion of 5 g soil in 10 ml deionized water,
flotation in 20 ml cold sucrose solution (specific
gravity 1.2), centrifugation (1,500× g, 20 min),
resuspension and centrifugation at 1,500× g,
20 min) (101, 148)
Cyclospora cayetanensis None reported to date but methods used for
T. gondii are likely applicable
Cryptosporidium parvum Dispersion in distilled H2O or phosphate-buffered
saline, flotation in ZnSO4.7H2O, cold sucrose,
or NaCl (best) (149)
Dispersion in 10 volumes of 0.01% Tween 20,
sedimentation, flotation with sucrose solution
(specific gravity 1.40), 350 × g for 10 min,
resuspension in distilled water, centrifugation at
1,540 × g for 10 min, analysis of pellet (150)
Detection
Nonnutrient agar (1% (w/v) Oxoid no. 1 agar in Page’s
amoeba saline (PAS) (2.5 mM NaCl, 1 mM KH2PO4, 0.5
mM Na2HPO4, 40 mM CaCl2.6H2O, and 20 mM
MgSO4.7H2O) supplemented with 4% (w/v) malt extract
and 4% (w/v) yeast extract, pH adjusted to 6.9 with KOH.
Approximately 5 ml late log phase culture Escherichia coli or
Enterobacter aerogenes are poured onto nonnutrient agar
plates and left for 5 min, after which excess culture fluid is
removed and plates are left to dry before their inoculation
with soil slurry. Incubate at 30°C and observe daily for the
presence of Acanthamoeba trophozoites (142, 143)
Trophozoites detected by ELISA with Entamoeba Ridascreen
(144). Note: a number of commercial antigen-based kits
are available
Nested PCR (72)
Real-time PCR, melt-curve analysis (145)
Oocyst detection by light microscopy
Oocyst detection by light microscopy
Oocyst detection by UV microscopy; real-time PCR (101)
Oocyst detection and identification by real-time PCR and
melt-curve analysis (125)
Detection of oocysts by epifluorescence microscopy, direct
immunofluorescence antibody method (149).
Immunomagnetic separation (Dynabeads
anti-Cryptosporidium Kit, Invitrogen, IDEXX
Laboratories), immunofluorescence microscopy (150)

of the particles and extraction of nucleic acids may be
required. The indirect approach permits elution and con-
centration from larger samples followed by concentration
by flocculation using methods already described (106). The
selection of a method for extraction is guided by the nature
of the virus of interest. For RNA viruses extraction using a
commercial solution (TRIzol, TRI Reagent) containing
organic solvents and guanidine thiocyanate can provide
good nucleic acid yield, removal of the phenolic compounds
and inhibition of RNase activity. Briefly, 1 ml of TRI Reagent
is added to 150 µl of the previously eluted and concentrated
viral suspension followed by incubation at room temperature
for 5 min. Chloroform (250 µl) is added to the suspension,
which is shaken vigorously by hand for 15 s and incubated
at room temperature for 3 min. The sample is then centri-
fuged at 12,000 × g for 15 min at 4°C. The colorless
(aqueous) phase is separated from the colored (organic) phase
and 500 µl isopropanol is added to precipitate the RNA. The
suspension is mixed, incubated at room temperature for 10
min, and centrifuged at 12,000 × g for 10 min at 4°C. The
supernatant is carefully removed, and 1 ml ethanol 75% is
added to the RNA-containing precipitate. The suspension
is mixed and then centrifuged once more at 7,500 × g for 5
min. The resulting pellet is air-dried for 10 min before being
resuspended in 50 µl DNase- and RNase-free water. The
extracted RNA may be stored at −80°C until use. Other
extraction methods based on the use of magnetic beads (Invi-
trogen Dynabeads Silane Viral NA Kit, MagNA Pure Com-
pact Nucleic Acid Isolation Kits) can also be considered for
the extraction of total viral nucleic acid (DNA and RNA).
An additional purification step can be included to remove
all traces of humic acid by passing the suspension through
3.3.1-8 ▪ SOIL
commercial PVPP mini-columns (Zymo Research), followed
by centrifugation at 14,000 × g for 1 min.
Molecular Methods for Viruses
Where target viruses replicate with great difficulty or cannot
replicate in the laboratory molecular methods must be used
for detection. Methods based on the amplification of nucleic
acids are now the preferred approach for the detection of most
viruses because they are fast, easily transferable between labo-
ratories, relatively affordable, very sensitive, and usable for the
majority of target species, whether DNA or RNA viruses.
Conventional PCR based on in vitro amplification of target
DNA produces a large number of copies of the target DNA.
The PCR reaction is conducted in three stages: denatura-
tion of the DNA by heating to 94°C to separate the DNA
strands, hybridization of the primers to the target sequence
at 50–60°C, and elongation at 72°C by DNA polymerase
(Taq polymerase). The cycle can be repeated 30–40 times
to obtain large amounts of the target DNA sequence. The
products of PCR can be detected and analyzed quickly by
agarose gel or capillary electrophoresis with a variety of stain-
ing and detection options available. For RNA viruses, reverse
transcription must be carried out before PCR to convert the
RNA into cDNA through the action of reverse transcriptase.
The design of primers is critical, particularly when the sample
contains a small amount of the virus in soil. The primers must
target a conserved region of the genome, have low degener-
acy, and have high melting temperatures to avoid nonspecific
matches. Sequence libraries can be useful for designing
specific primers. Many RT-PCR and PCR systems are avail-
able for detecting enteric viruses in various types of matrixes
(foods, human specimens, water, wastewater), and a recently
published review by Rodriguez-Lazaro et al. (120) provides a
good summary of methods for HEV, HAV, NoVs, rotaviruses,
enteroviruses, and adenoviruses, including detection limits
for each assay.
The use of positive, negative, and internal controls is
extremely important in molecular detection methods, partic-
ularly where nucleic acids are extracted from environmental
samples. Compounds that inhibit the PCR reaction can per-
sist despite rigorous purification procedures. To overcome this
problem, known DNA or RNA can be added to nucleic acid
extracts before amplification. Amplification of an internal
control (the known DNA or RNA) provides evidence of a
successful reaction and confirms negative results. Conversely,
failure of internal control amplification indicates inhibition
of PCR by coextracted material and puts the outcome in
doubt. It is recommended that all amplicons from positive
reactions obtained with RT-PCR and PCR undergo sequenc-
ing to confirm the amplified sequence.
Real-time PCR and RT-PCR are based on constant mon-
itoring of DNA during the reaction. Fluorescent probes in the
reaction mixture bind to either double-stranded DNA (SYBR
chemistry) or a target DNA sequence (TaqMan and beacon
chemistries). The probes fluoresce only when bound to
DNA, and the increase in the fluorescent signal is directly
proportional to the number of amplicons generated by the
PCR reaction. Measurement of fluorescence emitted in
each cycle makes it possible to monitor the PCR during the
exponential phase, where the first significant increase in
the amount of amplicons is directly correlated with the initial
quantity in the original target matrix (template). Real-time
technology has several advantages. In addition to allowing
progress to be viewed as the reaction actually occurs, this
technology provides direct confirmation of the sequence by
hybridization with a probe specific to the target DNA and
makes quantification possible, a very useful feature for virus
detection. As with conventional PCR, however, amplifica-
tion can be inhibited by compounds from the matrix, and
controls must always be run in parallel. The various reference
systems for real-time PCR and RT-PCR for the main enteric
viruses have been inventoried by Rodriguez-Lazaro et al.
(120). Despite their undeniable advantages for the detection
of viruses in soil and other environmental samples, molecular
detection methods currently do not distinguish between
infectious and noninfectious virus particles.
Molecular Methods for Protozoa
Although the ability to detect protozoa in clinical and envi-
ronmental samples has improved over the last decades, the
approaches and methods used for detection have remained
largely unchanged. Microscopic examination for the identifi-
cation of amebae, trophozoites, cysts, and oocysts is still con-
sidered the “gold standard” despite important disadvantages.
Microscopy is slow, tedious, and subject to human error
because morphological similarities between closely related
protozoan species make differentiation difficult. Antigen-
based detection has improved sensitivity, notably since the
advent of direct immunofluorescence assays or enzyme-linked
immunosorbent assays (ELISAs). For example, Giardia lam-
blia can be reliably detected using an ELISA developed by
Mank et al. (121). In contrast, antigen-based detection of
Cryptosporidium is far less reliable (122). Flow cytometry has
been used to improve the sensitivity, specificity, speed, and
throughput of immunofluorescence-based detection, but
the cost of analysis is high. Consequently, considerable efforts
have been aimed at developing methods for the identification
of protozoa that are based on the sensitive detection of nucleic
acids. The direct extraction of protozoan DNA from soil is
subject to the limitations described earlier. In addition,
nucleic acid yields from clinical or environmental samples
tend to be low because cysts and oocysts resist physical,
mechanical, or enzymatic lysis during extraction (including
extraction with commercial extraction kits). Recovery can
be improved by freezing in liquid N2 followed by thawing at
room temperature, although several cycles may be needed
before resistant cysts or oocysts are lysed.
There are few reports of direct and successful extraction of
protozoan nucleic acid from soil samples. An exception is
found in Liang and Keely (123), which describes the recovery
of mRNA from Cryptosporidium parvum oocysts for the assess-
ment of viability by reverse transcription RT-PCR. These
authors compared the efficiency of several commercial extrac-
tion kits in combination with chemical and physical treat-
ments. The most efficient total RNA extraction was
obtained with the PowerSoil total RNA isolation kit (Mo
Bio, Carlsbad, CA) followed by bead beating, although detec-
tion limits were variable in sandy, loamy, and clay soils. In the
absence of reliable methods for the direct extraction of
nucleic acids from other protozoan species in soil, elution
and concentration prior to DNA/RNA extraction remains
essential. A number of sensitive nucleic acid–based proce-
dures can then be applied for identification. Lalonde and
Gajadhar (124) developed a sensitive real-time quantitative
PCR assay that targets highly conserved regions of ribosomal
DNA for the detection of Cyclospora cayetanensis. Real-time
quantitative PCR combined with fluorescent melt-curve
analysis allows the detection and differentiation of coccidians
oocysts including Cryptosporidium parvum, Toxoplasma gondii,
and Cyclospora cayetanensis from a single DNA extract (125).
A multiplex real-time TaqMan PCR assay that can simulta-
neously detect Acanthamoeba spp. and Naegleria fowleri in
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-9
DNA extracts obtained by digestion with proteinase K in a
50 mM Tris-HCl ( pH 8.5), 1% aureth-12, and 1 mM
EDTA buffer that is widely used for the recovery of DNA
from protozoa (126, 127). Balantidium coli can be detected
by conventional PCR using primers given in Nilles-Bije
and Rivera (128) from DNA obtained by extraction with
Chelex 100 as described by Walsh et al. (129).
TRANSMISSION OF SOILBORNE HUMAN
PATHOGENIC VIRUSES AND PROTOZOA
While direct person-to-person contact remains the most com-
mon route of infection with enteric viruses, adventitious
transmission leading to sporadic, endemic, or epidemic dis-
ease is well documented in the medical literature. Pathogenic
protozoa are occasionally transmitted horizontally, but persis-
tent rates of infection in endemic regions and the random
emergence of new disease foci are also indicative of alterna-
tive modes of transmission. The enteric viruses and protozoa
discussed in the present work are shed in the feces of infected
animals or humans. Consequently, transfer from infected to
uninfected hosts is likely to occur along the fecal–oral route
of disease transmission. Figure 1 illustrates possible direct
and indirect pathways of infection for viral and protozoan
pathogens along the fecal–oral route. The risk of infection
associated with specific pathways is influenced by complex
biotic or abiotic factors that are clearly and perhaps increas-
ingly influenced by human behaviors.
Humans who interact daily with soil, notably in underde-
veloped areas with large agrarian populations, are susceptible
FIGURE 1 The role of soil in the transmission of human pathogens. doi: 10.1128/9781555818821.ch3.3.1.f1
to several well-known fungal and bacterial diseases acquired
by direct contamination of open wounds with soil or though
the inhalation of dust. Most pathogenic viruses and protozoa
transmitted by soil are less opportunistic and do not cause
infection unless they are ingested by the host. Purposeful con-
sumption of soil is rare, but inadvertent consumption in food
has been estimated to range between 9 and 96 mg per day in
Western countries (130). Outbreaks of bacterial gastroenter-
itis associated with exposure to contaminated soil during out-
door sporting activities suggest that accidental consumption
may also lead to infection (131). Although direct transmis-
sion from soil is clearly possible, the contribution of infections
acquired by this route to the overall burden of illness caused
by viral or protozoan human pathogens is likely small.
Contamination of soil with feces is unavoidable. Many
wildlife species serve as reservoir hosts for viral or protozoan
pathogens. In addition, human activity contributes to the
fecal contamination of soil through land application of
human waste, animal manures, slurries, biosolids, and sludges
from wastewater plants. In the United States alone, approxi-
mately 5.6 million dry tons of biosolids are generated annually
and 60% is applied on land (132). Recognition of the micro-
biological risks associated with the land application of fecal
materials has led to the development of practices meant to
promote the decay of pathogens, reduce human contact
with contaminated soil, and prevent dispersion of pathogens
in the natural or agricultural environments. The management
of fecal wastes and applications in agriculture are generally
subject to regulatory oversight in the developed world.
Despite these measures, infections caused by pathogenic
3.3.1-10 ▪ SOIL
viruses or protozoa and occasional public health crises persist.
Most viruses and protozoa adsorb to soil particles at variable
rates depending on pH, cation concentration, and organic
content. However, unadsorbed or free microorganisms perco-
late through soil and are transported over distances that vary
with water flow, soil texture, and composition (56, 133).
Transport is generally delayed in soils with a high organic
matter and clay content but is favored in sandy or very satu-
rated soils. Without barriers to restrict further movement,
potentially infectious microorganisms can contaminate
ground and surface waters, including those used to irrigate
agricultural crops destined for feeding livestock and for
human consumption (134, 135). Transfer of pathogenic
viruses and protozoa to fresh produce by contaminated irriga-
tion water is well established (53, 136, 137). In addition,
there is increasing evidence that contact with soil or absorp-
tion via the root system of the plant can lead to similar out-
comes (138). The consumption of contaminated crops is
clearly an indirect pathway for the transmission of viral and
protozoan to new hosts.
The enteric viruses and protozoa described in the present
chapter are often described as waterborne pathogens due to
strong evidence linking outbreaks of disease to water used
for drinking or bathing. Water clearly provides an important
means for the bulk movement of infectious microorganisms in
the natural or agricultural environments. With the exception
of the truly edaphic Acanthamoeba spp. pathogenic viruses
and protozoa are generally considered to be transient in soil.
However, the repeated application of contaminated irrigation
water and sources of fecal matter may create stable reservoirs
of pathogenic viruses and protozoa in soil. There is strong
anecdotal evidence that pathogens are transported from soil
reservoirs to surface and ground waters in epidemiologically
relevant numbers under favorable hydrogeological condi-
tions. For example, outbreaks of cryptosporidiosis have fre-
quently followed periods of sudden or heavy precipitation
(139–141). Pathogen transport from point sources on land
is suspected in many such incidents, but definitive evidence
is often lacking. Field outbreak investigations are normally
initiated well after the onset of a public health emergency.
Changes in conditions after a contamination event, incom-
plete understanding of complex environment–pathogen–
host interactions and the difficulties inherent to the recovery
and detection of viruses and protozoa in environmental sam-
ples, notably soil, undoubtedly contribute to uncertainty
about the ultimate origin of pathogens in many outbreaks.
Given the anticipated increase in land use intensification,
urbanization, and climate variability in both developed and
developing countries, it is clear that there is a need to improve
methods for the detection of such pathogens in soil with a
view to clarify the role of soil as a reservoir for infection.
REFERENCES
1. Brooker S, Clements ACA, Bundy DAP. 2006. Global
epidemiology, ecology and control of soil-transmitted hel-
minth infections. Adv Parasitol 62:221–261.
2. de Silva NR, Brooke S, Hotez PJ, Montresor A, Engels D,
Savioli L. 2003. Soil-transmitted helminth infections:
updating the global picture. Trrends Parasitol 19:547–551.
3. Baumgartner DJ. 2012. Soil-related bacterial and fungal
infections. J Am Board Fam Med 5:734–744.
4. Butler CD. 2012. Infectious disease emergence and global
change: thinking systemically in a shrinking world. Infect
Dis Pov 1:5.
5. Santamaria J, Toranzos GA. 2003. Enteric pathogens and
soils: a short review. Int Microbiol 6:5–9.
6. Palacios G, Oberste MS. 2005. Enteroviruses as
agents of emerging infectious diseases. J Neurovirol 11:
424–433.
7. Newell DG, Koopmans M, Verhoef L, Duizer E, Aidara-
Kane A, Sprong H, Opsteegh M, Langelaar M, Threfall J,
Scheutz F, van der Giessen J, Kruse H. 2010. Food-borne
diseases—the challenges of 20 years ago still persist while
new ones continue to emerge. Int J Food Microbiol 139
(Suppl 1):S3–S15.
8. Berger CN, Sodha SV, Shaw RK, Griffin PM, Pink D,
Hand P, Frankel G. 2010. Fresh fruit and vegetables as
vehicles for the transmission of human pathogens. Environ
Microbiol 12:2385–2397.
9. Lynch MF, Tauxe RV, Hedberg CW. 2009. The growing
burden of foodborne outbreaks due to contaminated fresh
produce: risks and opportunities. Epidemiol Infect 137:
307–315.
10. Dentinger CM, Bower WA, Nainan OV, Cotter SM,
Myers G, Dubusky LM, Fowler S, Salehi ED, Bell BP.
2001. An outbreak of hepatitis A associated with green
onions. J Infect Dis 183:1273–1276.
11. Beuchat LR. 1996. Pathogenic microorganisms associated
with fresh produce. J Food Prot 59:204–216.
12. Fallahi S, Mattison K. 2011. Evaluation of murine norovi-
rus persistence in environments relevant to food produc-
tion and processing. J Food Prot 74:1847–1851.
13. Bagdasaryan GA. 1964. Survival of viruses of the Entero-
virus group (poliomyelitis, echo, coxsackie) in soil and on
vegetables. J Hyg Epidemiol Microbiol Immunol 120:
497–505.
14. Yeager JG, O’Brien RT. 1979. Enterovirus inactivation in
soil. Appl Environ Microbiol 38:694–701.
15. Patel MM, Hall AJ, Vinje J, Parashar UD. 2009. Norovi-
ruses: a comprehensive review. J Clin Virol 44:1–8.
16. Patel MM, Widdowson MA, Glass RI, Akazawa K, Vinje
J, Parashar UD. 2008. Systematic literature review of role
of noroviruses in sporadic gastroenteritis. Emerg Infect Dis
14:1224–1231.
17. Koo HL, Ajami N, Atmar RL, DuPont HL. 2010. Noro-
viruses: the leading cause of gastroenteritis worldwide. Dis-
cov Med 10:61–70.
18. Atmar RL, Estes MK. 2001. Diagnosis of noncultivatable
gastroenteritis viruses, the human caliciviruses. Clin Micro-
biol Rev 14:15–37.
19. Koopmans M, Duizer E. 2004. Foodborne viruses: an
emerging problem. Int J Food Microbiol 90:23–41.
20. Marks PJ, Vipond IB, Carlisle D, Deakin D, Fey RE,
Caul EO. 2000. Evidence for airborne transmission of
Norwalk-like virus (NLV) in a hotel restaurant. Epidemiol
Infect 124:481–487.
21. Marks PJ, Vipond IB, Regan FM, Wedgwood K, Fey RE,
Caul EO. 2003. A school outbreak of Norwalk-like virus:
evidence for airborne transmission. Epidemiol Infect 131:
727–736.
22. Teunis PF, Moe CL, Liu P, Miller SE, Lindesmith L,
Baric RS, Le Pendu J, Calderon RL. 2008. Norwalk virus:
how infectious is it? J Med Virol 80:1468–1476.
23. Hall AJ, Eisenbart VG, Etingue AL, Gould LH, Lopman
BA, Parashar UD. 2012. Epidemiology of foodborne nor-
ovirus outbreaks, United States, 2001–2008. Emerg Infect
Dis 18:1566–1573.
24. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS,
Shapiro C, Griffin PM, Tauxe RV. 1999. Food-related ill-
ness and death in the United States. Emerg Infect Dis 5:
607–625.
25. Estes MK, Kapikian AZ. 2007. Rotaviruses, 5th ed. Lippin-
cott Williams & Wilkins, Philadelphia, PA.
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-11
26. Parashar UD, Gibson CJ, Bresee JS, Glass RI. 2006.
Rotavirus and severe childhood diarrhea. Emerg Infect Dis
12:304–306.
27. Sanchez-Padilla E, Grais RF, Guerin PJ, Steele AD,
Burny ME, Luquero FJ. 2009. Burden of disease and circu-
lating serotypes of rotavirus infection in sub-Saharan
Africa: systematic review and meta-analysis. Lancet Infect
Dis 9:567–576.
28. Kapikian A Z. 2001. A rotavirus vaccine for prevention of
severe diarrhoea of infants and young children: develop-
ment, utilization and withdrawal. Novartis Found Symp
238:153–171; discussion 171–159.
29. Sattar SA, Tetro JA (eds). 2001. Rotavirus, 2nd ed, vol 2.
Marcel Dekker, New York.
30. Hurst CJ, Gerba CP, Cech I. 1980. Effects of environmen-
tal variables and soil characteristics on virus survival in soil.
Appl Environ Microbiol 40:1067–1079.
31. Banyai K, Esona MD, Kerin TK, Hull JJ, Mijatovic S,
Vasconez N, Torres C, de Filippis AM, Foytich KR,
Gentsch JR. 2009. Molecular characterization of a rare,
human-porcine reassortant rotavirus strain, G11P[6], from
Ecuador. Arch Virol 154:1823–1829.
32. Banyai K, Esona MD, Mijatovic S, Kerin TK, Pedreira C,
Mercado J, Balmaseda A, Perez MC, Patel MM, Gentsch
JR. 2009. Zoonotic bovine rotavirus strain in a diarrheic
child, Nicaragua. J Clin Virol 46:391–393.
33. Banyai K, Martella V, Molnar P, Mihaly I, Van Ranst M,
Matthijnssens J. 2009. Genetic heterogeneity in human
G6P[14] rotavirus strains detected in Hungary suggests
independent zoonotic origin. J Infect 59:213–215.
34. Matthijnssens J, Bilcke J, Ciarlet M, Martella V, Banyai
K, Rahman M, Zeller M, Beutels P, Van Damme P, Van
Ranst M. 2009. Rotavirus disease and vaccination: impact
on genotype diversity. Future Microbiol 4:1303–1316.
35. Fiore AE. 2004. Hepatitis A transmitted by food. Clin
Infect Dis 38:705–715.
36. Rzezutka A, Cook N. 2004. Survival of human enteric
viruses in the environment and food. FEMS Microbiol Rev
28:441–453.
37. Lu L, Li C, Hagedorn CH. 2006. Phylogenetic analysis of
global hepatitis E virus sequences: genetic diversity, sub-
types and zoonosis. Rev Med Virol 16:5–36.
38. Aggarwal R, Jameel S. 2011. Hepatitis E. Hepatology 54:
2218–2226.
39. Okamoto H. 2007. Genetic variability and evolution of
hepatitis E virus. Virus Res 127:216–228.
40. Li TC, Chijiwa K, Sera N, Ishibashi T, Etoh Y, Shinohara
Y, Kurata Y, Ishida M, Sakamoto S, Takeda N, Miyamura
T. 2005. Hepatitis E virus transmission from wild boar
meat. Emerg Infect Dis 11:1958–1960.
41. Masuda J, Yano K, Tamada Y, Takii Y, Ito M, Omagari K,
Kohno S. 2005. Acute hepatitis E of a man who consumed
wild boar meat prior to the onset of illness in Nagasaki,
Japan. Hepatol Res 31:178–183.
42. Meng XJ. 2011. From barnyard to food table: the omnipre-
sence of hepatitis E virus and risk for zoonotic infection and
food safety. Virus Res 161:23–30.
43. Dalton HR, Bendall R, Ijaz S, Banks M. 2008a. Hepatitis
E: an emerging infection in developed countries. Lancet
Infect Dis 8:698–709.
44. Dalton H., Stableforth W, Thurairajah P, Hazeldine S,
Remnarace R, Usama W, Farrington L, Hamad N,
Sieberhagen C, Ellis V, Mitchell J, Hussaini SH, Banks
M, Ijaz S, Bendall RP. 2008. Autochthonous hepatitis
E in Southwest England: natural history, complications
and seasonal variation, and hepatitis E virus IgG sero-
prevalence in blood donors, the elderly and patients with
chronic liver disease. Eur J Gastroenterol Hepatol 20:
784–790.
45. Kwo PY, Schlauder GG, Carpenter HA, Murphy PJ,
Rosenblatt JE, Dawson GJ, Mast EE, Krawczynski K,
Balan V. 1997. Acute hepatitis E by a new isolate acquired
in the United States. Mayo Clin Proc 72:1133–1136.
46. Lewis HC, Wichmann O, Duizer E. 2010. Transmission
routes and risk factors for autochthonous hepatitis E virus
infection in Europe: a systematic review. Epidemiol Infect
138:145–166.
47. Aggarwal R, Naik SR. 1994. Hepatitis E: intrafamilial
transmission versus waterborne spread. J Hepatol 21:
718–723.
48. Somani SK, Aggarwal R, Naik SR, Srivastava S, Naik S.
2003. A serological study of intrafamilial spread from
patients with sporadic hepatitis E virus infection. J Virol
10:446–449.
49. Emerson S, Purcell RH. 2003. Hepatitis E virus. Rev Med
Virol 13:145–154.
50. Halac U, Béland K, Lapierre P, Patey N, Ward P, Brassard
J, Houde A, Alvarez F. 2012. Chronic hepatitis E infection
in children with liver transplantation. Gut 61:597–603.
51. Halac U, Béland K, Lapierre P, Patey N, Ward P, Bras-
sard J, Houde A, Alvarez F. 2012. Cirrhosis due to chronic
hepatitis E infection in a child post-bone marrow trans-
plant. J Pediatr 160:871–874.
52. Kamar N, Selves J, Mansuy JM, Ouezzani L, Peron JM,
Guitard J, Cointault O, Esposito L, Abravanel F, Danjoux
M, Durand D, Vinel JP, Izopet J, Rostaing L. 2008. Hep-
atitis E virus and chronic hepatitis in organ-transplant
recipients. N Engl J Med 358:811–817.
53. Brassard J, Gagné MJ, Généreux M, Côté C. 2012. Detec-
tion of human food-borne and zoonotic viruses on irrigated,
field-grown strawberries. Appl Environ Microbiol 78:
3763–3766.
54. Simmons FJ, Xagoraraki I. 2011. Release of infectious
human enteric viruses by full-scale wastewater utilities.
Water Res 45:3590–3598.
55. Wong K, Onan BM, Xagoraraki I. 2010. Quantification
of enteric viruses, pathogen indicators, and Salmonella bac-
teria in class B anaerobically digested biosolids by culture
and molecular methods. Appl Environ Microbiol 76:
6441–6448.
56. Ziemer CJ, Bonner JM, Cole D, Vinje J, Constantini V,
Goyal S, Gramer M, Mackie R, Meng XJ, Myers G, Saif
LJ. 2010. Fate and transport of zoonotic, bacterial, viral,
and parasitic pathogens during swine manure treatment,
storage, and land application. J Anim Sci 88:E84–E94.
57. Mergeryan H. 1991. The prevalence of Acanthamoeba in
the human environment. Rev Infect Dis 13(Suppl. 5):
S390–S391.
58. Brindley N, Matin A, Khan NA. 2009. Acanthamoeba cas-
tellanii: high antibody prevalence in racially and ethnically
diverse populations. Exp Parasitol 121:254–256.
59. Mazur T, Hadas E, Iwanicka I. 1995. The duration of the
cyst stage and the viability and virulence of Acanthamoeba
isolates. Trop Med Parasitol 46:106–108.
60. Sriram R, Shoff M, Booton G, Fuerst P, Visvesvara GS.
2008. Survival of Acanthamoeba cysts after desiccation for
more than 20 years. J Clin Microbiol 46:4045–4048.
61. Siddiqui R, Khan NA. 2012. Biology and pathogenesis of
Acanthamoeba. Parasit Vectors 5:6–19.
62. Marciano-Cabral F, Cabral G. 2003. Acanthamoeba spp. as
agents of disease in humans. Clin Microbiol Rev 16:
273–307.
63. Verani JR, Lorick SA, Yoder JS, Beach MJ, Braden CR,
Roberts JM, Conover CS, Chen S, McConnell K, Chang
DC, Park BJ, Jones DB, Visvesvara GS, Roy SL. 2009.
National outbreak of Acanthamoeba keratitis associated
with use of a contact lens solution, United States. Emerg
Infect Dis 15:1236–1242.
3.3.1-12 ▪ SOIL
64. Evstigneeva A, Raoult D, Karpachevskiy L, La Scola B.
2009. Amoeba co-culture of soil specimens recovered 33
different bacteria, including four new species and Strepto-
coccus pneumoniae. Microbiol 155:657–664.
65. Thomas V, McDonnell G, Denyer SP, Maillard JY. 2010.
Free-living amoebae and their intracellular pathogenic
microorganisms: risks for water quality. FEMS Microbiol
Rev 34:231–259.
66. Stanley SL Jr. 2003. Amoebobiasis. Lancet 361:1025–
1034.
67. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K,
Aboyans V, Abraham J, Adair T, Aggarwal R, Ahn SY,
Alvarado M, Anderson HR, Anderson LM, Andrews
KG, Atkinson C, Baddour LM, Barker-Collo S, Bartels
DH, Bell ML, Benjamin EJ, Bennett D, Bhalla K, Bikbov
B, Bin Abdulhak A, Birbeck G, Blyth F, Bolliger I,
Boufous S, Bucello C, Burch M, Burney P, Carapetis J,
Chen H, Chou D, Chugh SS, Coffeng LE, Colan SD,
Colquhoun S, Colson KE, Condon J, Connor MD,
Cooper LT, Corriere M, Cortinovis M, de Vaccaro KC,
Couser W, Cowie BC, Criqui MH, Cross M, Dabhadkar
KC. et al. 2012. Global and regional mortality from 235
causes of death for 20 age groups in 1990 and 2010: a system-
atic analysis for the Global Burden of Disease Study 2010.
Lancet 380:2095–2128.
68. Benetton ML, Goncalves AV, Meneghini ME, Silva EF,
Carneiro M. 2005. Risk factors for infection by the Enta-
moeba histolytica/E. dispar complex: an epidemiological
study conducted in outpatient clinics in the city of Manaus,
Amazon Region, Brazil. Trans R Soc Trop Med Hyg 99:
532–540.
69. Pham Duc P, Nguyen-Viet H, Hattendorf J, Zinsstag J,
Dan Cam P, Odermat P. 2011. Risk factors for Entamoeba
histolytica infection in an agricultural community in Hanam
province, Vietnam. Parasit Vectors 4:102–111.
70. Shahnazi M, Jafari-Sabet M. 2010. Prevalence of parasitic
contamination of raw vegetables in village of Qazvin prov-
ince, Iran. Foodborne Pathog Dis 7:1025–1030.
71. Marciano-Cabral F, Cabral G. 2007. The immune response
to Naegleria fowleri amoebae and pathogenesis of infection.
FEMS Immunol Med Microbiol 51:243–259.
72. Maclean RC, Richardson DJ, LePardo R, Marciano-
Cabral F. 2004. The identification of Naegleria fowleri
from water and soil samples by nested PCR. Parasitol Res
93:211–217.
73. Robertson LJ, Hanevik K, Escobedo AA, Mørch K, Lan-
geland N. 2010. Giardiasis—why do the symptoms some-
times never stop? Trends Parasitol 26:75–82.
74. Feng Y, Xiao L. 2011. Zoonotic potential and molecular
epidemiology of Giardia species and giardiasis. Clin Micro-
biol Rev 24:110–140.
75. Yazar S, Altuntas F, Sahin I, Atambay M. 2004. Dysentery
caused by Balantidium coli in a patient with non-Hodgkin’s
lymphoma from Turkey. World J Gastroenterol 10:
458–459.
76. Schuster FL, Ramirez-Avila L. 2009. Current
world status of Balantidium coli. Clin Microbiol Rev 21:
626–638.
77. Morris RG, Jordan HE, Luce WG, Coburn TC, Maxwell
CV. 1984. Prevalence of gastrointestinal parasitism in
Oklahoma swine. Am J Vet Res 45:2421–2423.
78. Yatswako S, Faleke OO, Gulumbe ML, Daneji AI. 2007.
Cryptosporidium oocysts and Balantidium coli cysts in pigs
reared semi-intensively in Zuru, Nigeria. Pak J Biol Sci
10:3435–3439.
79. Dawson D. 2005. Foodborne protozoan parasites. Int J Food
Microbiol 103:207–227.
80. Farthing MJG. 2000. Clinical aspects of human crypto-
sporidiosis, p 50–74. In Petry F. (ed), Cryptosporidiosis and
Microsporidiosis. Contributions to Microbiology, vol 6. Karger,
Basel.
81. Mor Siobhan M, Tzipor S. 2008. Cryptosporidiosis in chil-
dren in sub-Saharan Africa: a lingering challenge. Clin
Infect Dis 47:915–921.
82. Fletcher SM, Stark D, Harkness J, Ellis J. 2012. Enteric
protozoa in the developed world: a public health perspec-
tive. Clin Microbiol Rev 25:420–449.
83. Xiao L. 2010. Molecular epidemiology of cryptosporidiosis:
an update. Exper Parasitol 124:80–89.
84. Fayer R, Morgan U, Upton SJ. 2000. Epidemiology of
Cryptosporidium: transmission, detection and identifica-
tion. Int J Parasitol 30:1305–1322.
85. Robertson LJ, Gjerde B. 2001. Occurrence of parasites
on fruits and vegetables in Norway. J Food Prot 64:
1793–1798.
86. Maikai BV, Baba-Onaja EBT, Elisha IA. 2013. Contam-
ination of raw vegetables with Cryptosporidium oocysts in
markets within Zaria metropolis, Kaduna State, Nigeria.
Food Cont 31:45–48.
87. Peng X, Murphy T, Holden NM. 2008. Evaluation of the
effect of temperature on the die-off rate for Cryptosporidium
parvum oocysts in water, soils, and feces. Appl Environ
Microbiol 74:7101–7107.
88. Mootian G, Wu WH, Matthews K. 2009. Transfer of
Escherichia coli O157:H7 from soil, water, and manure con-
taminated with low numbers of the pathogen to lettuce
plants. J Food Prot 72:2308–2312.
89. Wei J, Kniel KE. 2010. Pre-harvest viral contamination of
crops originating from fecal matter. Food Environ Virol 2:
195–206.
90. Lainson R. 2005. The genus Cyclospora (Apicomplexa:
Eimeriidae), with a description of Cyclospora schneideri n.
sp. in the snake Anilius scytale scytale (Aniliidae) from
Amazonian Brazil—a review. Mem Inst Oswaldo Cruz
100:103–110.
91. Ortega YR, Giman RH, Sterling CR. 1994. A new cocci-
dian parasite (Apicomplexa: Eimeriidae) from humans.
J Parasitol 80:625–629.
92. Puente S, Morente A, García-Benayas T, Subirats M,
Gascon J, González-Lahozm JM. 2006. Cyclosporiasis: a
point source outbreak acquired in Guatemala. J Travel
Med 13:334–337.
93. Mansfield LS, Gajadhar AA. 2004. Cyclospora cayetanensis
to food and waterborne coccidian parasite. Vet Parasitol
126:73–90.
94. Ortega YR, Sanchez R. 2010. Update on Cyclospora caye-
tanensis, a food-borne and waterborne parasite. Clin Micro-
biol Rev 23:218–234.
95. Gajadhar AA, Allen JR. 2004. Factors contributing to the
public health and economic importance of waterborne zoo-
notic parasites. Vet Parasitol 126:3–14.
96. Rabold G, Hoge CW, Shlim DR, Kefford C, Rajah R,
Echeverria P. 1994. Cyclospora outbreak associated with
chlorinated drinking water. Lancet 344:1360–1361.
97. Chacin-Bonilla L. 2010. Epidemiology of Cyclospora caye-
tanensis: a review focusing in endemic areas. Acta Tropica
115:181–193.
98. Doller PC, Dietrich K, Filipp N, Brockmann S, Dreweck
C, Vonthein R, Wagner-Wiening C, Wiedenmann A.
2002. Cyclosporiasis outbreak in Germany associated with
the consumption of salad. Emerg Infect Dis 8:992–994.
99. Kim K, Weis LM 2008. Toxoplasma: the next 100 years.
Microbes Infect 10:978–984.
100. Holland GN. 2004. Ocular toxoplasmosis: a global reas-
sessment. Part II: disease manifestations and management.
Am J Ophthalmol 137:1–17.
101. Lélu M, Villena I, Dardé ML, Aubert D, Geers R, Dupuis
E, Marnef F, Poullem ML, Gotteland C, Dumètre A,
 3.3.1. Pathogenic Viruses and Protozoa Transmitted by Soil ▪ 3.3.1-13
Gilot-Fromont E. 2012. Quantitative estimation of the
viability of Toxoplasma gondii oocysts in soil. Appl Environ
Microbiol 78:5127–5132.
102. Alfonso E, Lemoine M, Poulle ML, Ravat MC, Romand S,
Thulliez P, Villena I, Aubert D, Rabilloud M, Riche B,
Gilot-Fromont E. 2008. Spatial distribution of soil
contamination by Toxoplasma gondii in relation to cat
defecation behavior in an urban area. Int J Parasitol 38:
1017–1023.
103. Goyal SM, Gerba CP. 1979. Comparative adsorption
of human enteroviruses, simian rotavirus, and selected
bacteriophages to soils. Appl Environ Microbiol 38:241–
247.
104. Sobsey MD, Dean CH, Knuckles ME, Wagner RA. 1980.
Interactions and survival of enteric viruses in soil materials.
Appl Environ Microbiol 40:92–101.
105. Straub TM, Pepper IL Gerba CP. 1992. Persistence of
viruses in desert soils amended with anaerobically digested
sewage sludge. Appl Environ Microbiol 58:636–641.
106. Hurst CJ, Schaub SA, Sobsey MD, Farrah SR, Gerba CP,
Rose JB, Goyal SM, Larkin EP, Sullivan R, Tierney JT,
Tierney JT, O'Brien RT, Safferman RS, Morris ME,
Wellings FM, Lewis AL, Berg G, Britton PW, Winter
JA. 1991. Multilaboratory evaluation of methods for
detecting enteric viruses in soils. Appl Environ Microbiol
57:395–401.
107. Payment P. 2007. Cultivation and assay and animal
viruses, p. 93–100. In Hurst CJ (ed), Manual of Envi-
ronmental Microbiology, 3rd ed. ASM Press, Washington,
DC.
108. Prim N, Rodriguez G, Margall N, Del Cuerpo M,
Trallero G, Rabella N. 2013. Combining cell lines to
optimize isolation of human enterovirus from clinical
specimens: report of 25 years of experience. J Med Virol
85:116–120.
109. Rajendhran J, Gunasekaran P. 2008. Strategies for access-
ing soil metagenome for desired applications. Biotechnol
Adv 26:576–590.
110. McIlroy SJ, Porter K, Seviour RJ, Tillett D. 2009.
Extracting nucleic acids from activated sludge which reflect
community population diversity. Antonie Van Leeuwenhoek
96:593–605.
111. Dong D, Yan A, Liu H, Zhang X, Xu Y. 2006. Removal of
humic substances from soil DNA using aluminium sulfate.
J Microbiol Meth 66:217–222.
112. Berthelet M, Whyte LG, Greer CW. 1996. Rapid, direct
extraction of DNA from soils for PCR analysis using poly-
vinylpolypyrrolidone spin columns. FEMS Microbiol Lett
138:17–22.
113. Jackson CR, Harper JP, Willoughby D, Roden EE,
Churchill PF. 1997. A simple, efficient method for the sep-
aration of humic substances and DNA from environmental
samples. Appl Environ Microbiol 63:4993–4995.
114. Martin-Laurent F, Philippot L, Hallet S, Chaussod R,
Germon JC, Soulas G, Catroux G. 2001. DNA extraction
from soils: old bias for new microbial diversity analysis
methods. Appl Environ Microbiol 67:2354–2359.
115. Miller DN, Bryant JE, Madsen EL, Ghiorse WC. 1999.
Evaluation and optimization of DNA extraction and puri-
fication procedures for soil and sediment samples. Appl
Environ Microbiol 65:4715–4724.
116. Miller DN. 2001. Evaluation of gel filtration resins for the
removal of PCR-inhibitory substances from soils and sedi-
ments. J Microbiol Meth 44:49–58.
117. Rondon MR, August PR, Bettermann AD, Brady SF,
Grossman TH, Liles MR, Loiacono KA, Lynch BA,
MacNeil IA, Minor C, Tiong CL, Gilman M, Osburne
MS, Clardy J, Handelsman J, Goodman RM. 2000.
Cloning the soil metagenome: a strategy for accessing the
genetic and functional diversity of uncultured microorgan-
isms. Appl Environ Microbiol 66:2541–2547.
118. Zhou J, Bruns MA, Tiedje JM. 1996. DNA recovery from
soils of diverse composition. Appl Environ Microbiol 62:
316–322.
119. Burgmann H, Pesaro M, Widmer F, Zeyer J. 2001. A strat-
egy for optimizing quality and quantity of DNA extracted
from soil. J Microbiol Meth 45:7–20.
120. Rodriguez-Lazaro D, Cook N, Ruggeri FM, Sellwood
Nasser JA, Nascimento MS, D’Agostino M, Santos R,
Saiz JC, Rzeżutka A, Bosch A, Gironés R, Carducci A,
Muscillo M, KovačK, Diez-Valcarce M, Vantarakis A,
von Bonsdorff CH, de Roda Husman AM, Hernández
M, van der Poel WH. 2012. Virus hazards from food, water
and other contaminated environments. FEMS Microbiol
Rev 36:786–814.
121. Mank TG, Zaat JO, Deelder AM, van Eijk JT, Polderman
AM. 1997. Sensitivity of microscopy versus enzyme immu-
noassay in the laboratory diagnosis of giardiasis. Eur J Clin
Microbiol Infect Dis 16:615–619.
122. Weitzel T, Dittrich S, Mohl I, Adusu E, Jelinek T. 2006.
Evaluation of seven commercial antigen detection tests for
Giardia and Cryptosporidium in stool samples. Clin Microbiol
Infect 12:656–659.
123. Liang Z, Keely A. 2011. Detection of viable Cryptospori-
dium parvum in soil by reverse transcription–real-time
PCR targeting hsp70 mRNA. Appl Environ Microbiol 77:
6476–6485.
124. Lalonde LF, Gajadhar AA. 2008. Highly sensitive and
specific PCR assay for reliable detection of Cyclospora
cayetanensis oocysts. Appl Environ Microbiol 74:4354–
4358.
125. Lalonde LF, Gajadhar AA. 2011. Detection and differen-
tiation of coccidian oocysts by real-time PCR and melting
curve analysis. J Parasitol 97:725–730.
126. Qvarnstrom Y, Visvesvara GS, Sriram R, da Silva AJ.
2006. Multiplex real-time PCR assay for simultaneous
detection of Acanthamoeba spp., Balamuthia mandrillaris,
and Naegleria fowleri. J Clin Microbiol 44:3589–3595.
127. da Silva AJ, Schwartz DA, Visvesvara GS, de Moura H,
Slemenda SB, Pieniazek NJ. 1996. Sensitive PCR diagno-
sis of infections by Enterocytozoon bieneusi (microsporidia)
using primers based on the region coding for small-subunit
rRNA. J Clin Microbiol 34:986–987.
128. Nilles-Bije ML, Rivera WL. 2010. Ultrastructural and
molecular characterization of Balantidium coli isolated in
the Philippines. Parasitol Res 106:387–394.
129. Walsh PS, Metzger DA, Higuchi R. 1991. Chelex 100 as a
medium for simple extraction of DNA for PCR-based typ-
ing from forensic material. Bio Tech 10:506–513.
130. Abrahams PW. 2002. Soils: their implications to human
health. Sci Total Environ 291:1–32.
131. Stuart TL, Sandhu J, Stirling R, Corder J, Ellis A, Misa P,
Goh S, Wong B, Martiquet P, Hoang L, Galanis E. 2010.
Campylobacteriosis outbreak associated with ingestion of
mud during a mountain bike race. Epidemiol Infect 138:
1695–1703.
132. Land NR. 2002. Biosolids Applied to Land: Advancing Stand-
ards and Practices. National Academies Press, Washington,
DC.
133. Mawdsley JL, Brooks AE, Merry RJ. 1996. Movement of
the protozoan pathogen Cryptosporidium parvum through
three contrasting soil types. Biol Fertil Soils 21:30–36.
134. Tyrrel SF, Quinton JN. 2003. Overland flow transport of
pathogens from agricultural land receiving faecal wastes.
J Appl Microbiol 94(Suppl):87S–93S.
135. Borchardt MA, Spencer SK, Kieke BA, Lambertini E,
Loge FJ. 2012. Viruses in nondisinfected drinking water
from municipal wells and community incidence of acute
3.3.1-14 ▪ SOIL
gastrointestinal illness. Environ Health Perspect 120:
1272–1279.
136. Steele M, Odumeru J. 2004. Irrigation water as source of
foodborne pathogens on fruit and vegetables. J Food Prot
67:2839–2849.
137. Chaidez C, Soto M, Gortares P, Mena K. 2005. Occur-
rence of Cryptosporidium and Giardia in irrigation water
and its impact on the fresh produce industry. Int J Environ
Health Res 15:339–345.
138. Hirneisen KA, Sharma M, Kniel KE. 2012. Human
enteric pathogen internalization by root uptake into food
crops. Foodborne Pathog Dis 9:396–405.
139. Atherton F, Newman CP, Casemore DP. 1995. An out-
break of waterborne cryptosporidiosis associated with a
public water supply in the UK. Epidemiol Infect 115:
123–131.
140. MacKenzie WR, Hoxie NJ, Proctor ME, Gradus MS,
Blair KA, Peterson JJ, Kazmierczak DG, Addiss KR,
Fox JB, et al. 1994. A massive outbreak in Milwaukee of
Cryptosporidium infection transmitted through the water
supply. N Engl J Med 331:161–167.
141. Lake IR, Bentham G, Kovats RS, Nichols GL. 2005.
Effects of weather and river flow on cryptosporidiosis.
J Water Health 3:469–474.
142. Khan NA, Jarroll EL, Paget TA. 2001. Acanthamoeba can
be differentiated by the polymerase chain reaction and sim-
ple plating assays. Curr Microbiol 43:204–208.
143. Kilic A, Tanyuksel M, Sissons J, Jayasekera S, Khan NA.
2004. Isolation of Acanthamoeba isolates belonging to T2,
T3, T4 and T7 genotypes from environmental samples in
Ankara, Turkey. Acta Parasitol 49:246–252.
144. Da Silva Barbosa A, Antunnes Uchôa CM, Laurentino
da Silva V, Nascimento Duarte A, Bendeira Vianna M,
Monteiro Fonseca AB, Pereira Bastos OM. 2012. Compa-
rative study of parasitological techniques and ELISA for
analysis of environmental samples. Brazil Rev Ibero-
Latinoam Parasitol 71:90–96.
145. Robinson BS, Monis PT, Dobson PJ. 2006. Rapid, sensi-
tive, and discriminating identification of Naegleria spp. by
real-time PCR and melting-curve analysis. Appl Environ
Microbiol 72:5857–5863.
146. Barwick RS, Mohammed HO, White ME, Bryant RB.
2000. Detection of Cryptosporidium muris and Cryptospori-
dium parvum in soil samples. Biol Fert Soils 31:385–390.
147. Barwick R, Mohammed H, White M, Bryant R. 2003.
Prevalence of Giardia spp. and Cryptosporidium spp. on
dairy farms in southeastern New York state. Prev Vet Med
59:1–11.
148. Lélu M, Gilot-Fromont E, Aubert D, Afonso E, Dupuis
E, Gotteland C, Dardé ML, Marnef F, Poulle ML,
Richaume-Jolion A, Villena I. 2011. Development of a
sensitive method for Toxoplasma gondii oocyst extraction
in soil samples. Vet Parasitol 183:59–67.
149. Kuckzinska E, Shelton DR. 1999. Method for detection
and enumeration of Cryptosporidium parvum oocysts in
feces, manures, and soils. Appl Environ Microbiol 99:
2820–2826.
150. Petersen HH, Enemark HL, Olsenm A, Amin MG, Dals-
gaard A. 2012. Transport of Cryptosporidium parvum
oocysts in soil columns following applications of raw and
separated liquid slurries. Appl Environ Microbiol 78:
5994–6000.
 Natural Soil Reservoirs for Human Pathogenic and
Fecal Indicator Bacteria
MARIA LAURA BOSCHIROLI, JOSEPH FALKINHAM, SABINE FAVRE-BONTÉ,
SYLVIE NAZARET, PASCAL PIVETEAU, MICHAEL SADOWSKY,
MURULEE BYAPPANAHALLI, PASCAL DELAQUIS, AND ALAIN HARTMANN
3.3.2
INTRODUCTION AND BACKGROUND
Context
Soils are potential reservoirs for human pathogens, and con-
taminated soils put humans and grazing animals at risk of con-
tamination through direct contact or consumption of crops.
Adjacent water resources may be contaminated through
leaching and runoff. Recreational activities in contaminated
environments may also lead to health risks. Alternatively,
soils may also act as a barrier (or a filter) against human patho-
gens through competition with endogenous soil microbial
communities. Less information is available concerning
soil reservoirs compared to aquatic reservoirs. This chapter
focuses on several bacterial pathogens that are significant
for human health and that are known to persist in soil.
True pathogens like Listeria monocytogenes, Clostridium botuli-
num, C. tetanii, C. perfringens, Bacillus cereus, Mycobacterium
bovis, Mycobacterium avium, and enterotoxic E. coli are used
as models to illustrate available methods. Methods concern-
ing the detection and characterization of opportunistic path-
ogens that occur naturally in soils are presented using
Stenotrophomonas maltophilia as an example. Plant and animal
hosts; plant roots; rhizosphere; soil microfauna, mesofauna,
and macrofauna; wild animals; and livestock that may serve
as reservoirs for human bacterial pathogens are described
as well as specific methods. The following is a review of
approaches to detection, isolation, enumeration, and typing
of pathogenic bacteria in soils. Molecular detection is used
to describe the enumeration without colony isolation, usually
involving DNA amplification through PCR. PCR-based
detection methods have the disadvantage of not yielding via-
ble organisms for subsequent characterization, however, has
the advantage of detecting viable but not culturable
(VBNC) cells. Alternatively, colony formation, though miss-
ing VBNC cells, is suitable for subsequent typing and source
tracking.
Key Soilborne Bacterial Pathogens of Humans
Bacteria that favor soil as a habitat may be called edaphic, from
the Greek word for ground or soil. A number of edaphic bac-
terial species are capable of causing disease in humans. In
addition, soil can act as a reservoir for a diversity of patho-
genic bacteria, notably accompanying contamination by ani-
mal or human fecal material.
doi:10.1128/9781555818821.ch3.3.2
3.3.2-1
Clostridium
The genus Clostridium includes over 200 species of Gram-
positive anaerobic spore-forming bacteria of which approxi-
mately 50 have been associated with clinical conditions in
animals and humans. The most significant species from a his-
torical perspective is Clostridium tetani, the etiologic agent
responsible for tetanus (also known as trismus or lockjaw),
which may manifest as generalized, localized, or neonatal
forms. C. tetani is a noninvasive pathogen that enters affected
tissues primarily in the form of spores. The name Clostridium
botulinum defines a heterogeneous complex of clostridia capa-
ble of producing protein neurotoxins termed botulinum neu-
rotoxins that are among the most potent toxins known to
humans. C. botulinum strains belonging to types A, B, E,
and F are associated with human botulism. Four clinical forms
of botulism are recognized: foodborne, infant, wound, and
adult intestinal colonization, with the most common being
an intoxication that results from ingestion of food containing
the preformed toxins. C. perfringens is responsible for both
infections and intoxications in humans. The infection of
deep tissue wounds or trauma leading to gas gangrene, clostri-
dial myonecrosis was a common occurrence in wounded sol-
diers before the advent of modern battlefield care.
Bacillus
The genus Bacillus encompasses more than 250 diverse spe-
cies of aerobic to facultative anaerobic Gram-positive
endospore-forming rod-shaped bacteria. Six soilborne species
including B. cereus, B. anthracis, B. mycoides, B. pseudomy-
coides, B. thuringiensis, and B. weihenstephanensis are consid-
ered taxonomically distinct in most classification schemes
and are referred to collectively as the Bacillus cereus or patho-
genic Bacillus group. However, recent phylogenetic analysis
has revealed a high degree of relatedness among these species,
suggesting that this group represents a single species within
the genus (1). There is no convincing clinical evidence
that B. thuringiensis, B. mycoides, or B. pseudomycoides are
pathogenic to humans. In contrast, B. cereus and the compa-
ratively rare B. weihenstephanensis can cause emetic vomiting
or diarrheal foodborne illnesses. Illnesses caused by B. cereus
or B. weihenstephanensis are not uncommon but generally do
not require medical attention. It should be noted that occa-
sional reports of severe complications or mortality associated
with emetic strains point to the risk that more virulent forms
3.3.2-2 ▪ SOIL
could emerge in the future (2). B. anthracis has long been rec-
ognized as a major threat to the well-being of both humans
and livestock. Diseases caused by B. anthracis are referred to
collectively as anthrax.
Listeria
The genus Listeria comprises seven species of Gram-positive
short rods capable of growth under aerobic to microaerophilic
conditions. L. monocytogenes and L. ivanovii are pathogenic
to humans, although reports of infection with the latter
are scarce (3). L. monocytogenes is a facultative intracellular
pathogen capable of causing severe infections in humans
and animals, including wild and domesticated ruminants.
L. monocytogenes is an opportunistic pathogen. Infection
can lead to meningitis, septicemia, and organ failure in sus-
ceptible individuals including the immunodeficient, the eld-
erly, and neonates, as well as miscarriages in pregnant women.
Mortality rates can approach 50% in these populations (4).
Long considered primarily a veterinary pathogen, a surge in
human cases since the late 1970s has been associated with
the consumption of contaminated food, and numerous out-
breaks have been reported, primarily in developed nations.
Food contamination by L. monocytogenes is probably due to
its capacity to grow in low-temperature food processing
plants.
E. coli and Enterococci
Both E. coli and enterococci have historically been used
worldwide as fecal indicator bacteria in environmental matri-
ces. However, research over the past three decades has signifi-
cantly expanded our understanding of the ecology of E. coli,
enterococci, and other enteric bacteria in extraintestinal hab-
itats. The widespread occurrence of these bacteria in a range
of environments, encompassing terrestrial and aquatic habi-
tats, has challenged the long-held paradigm that enteric bac-
teria are primarily homoeothermic (i.e., restricted to the
gastrointestinal tract of humans and warm-blooded animals)
and that their outside occurrence is incidental or transient.
Watershed streams and rivers, riparian soils, associated drain-
age areas, beach aquatic vegetation, sand, and swimming
water ecosystems have received more attention as these bacte-
ria are known to influence water quality and human health
(5–7).
Over several decades, the occurrence of enteric bacteria
in soil and other environments was considered as a result of
incidental fecal contamination from anthropogenic sewage,
domesticated animals, and wildlife sources. Thus, many of
the studies on enteric bacterial occurrence in soil and sedi-
ment environments have typically focused on watersheds
affected by agriculture, including croplands, rangelands, and
other animal husbandry operations (8, 9). More recently,
other studies have shown that E. coli and enterococci are
equally common in soils across a range of climatic and geo-
graphic conditions (5).
Mycobacterium
The genus Mycobacterium contains about 90 named species
distributed in three groups: slow growers, rapid growers, and
strains that are not culturable in vitro (namely, M. leprae).
Mycobacterium are acid-fast bacilli, are aerobic and nonmo-
bile, and belong to the Acidobacteria order. Species belong-
ing to the M. tuberculosis group (M. tuberculosis, M. bovis,
M. microti, and M. africanum) cause tuberculosis, which is a
chronic granulomatous lung disease, affecting humans, many
other mammals, birds, fish, and reptiles. Tuberculosis may
spread to other organs or proceed to a generalized infection
(“military tuberculosis”) (10). Environmental opportunistic
mycobacteria are not obligate human pathogens. They include
a large number of nontuberculous mycobacterial (NTM) spe-
cies, including M. avium, M. avium subspecies paratuberculosis
(responsible for Johne’s disease in cattle) and M. intracellulare.
Environmental mycobacteria may have health impacts like
severe prosthetic joint infections and may be involved in
chronic bowel disease (M. avium) (11).
S. maltophilia
S. maltophilia has been described in the past decades as an
global environmental emerging Gram-negative multidrug-
resistant organism that is commonly associated with respira-
tory infections in humans (12). It represents the third most
frequent nonfermenting Gram-negative bacterium impli-
cated in nosocomial infections (13) and is increasingly
isolated from cystic fibrosis patients (14). S. maltophilia is
characterized by a high intrinsic resistance to a wide range
of antimicrobics. Its intrinsic resistance is due to the presence
of broad-spectrum efflux pumps such as SmeABC and Sme-
DEF, the L1 and L2 metallo-β-lactamases, AAC60 -Iz and
APH30 -IIa aminoglycoside-modifying enzymes, and a Qnr
peptide that confers resistance to quinolones (15).
DISTRIBUTION IN SOILS AND HABITATS
Clostridium
Surveys of soils in various parts of the world invariably reveal
the presence of C. tetani spores, although recovery rates are
variable and the relationship between occurrence, soil type,
or use is poorly understood (16). Spores are also shed in the
feces of humans and animal species, including horses, cattle,
sheep, dogs, rats, and chickens, and have been recovered from
human skin and the hides of livestock (17). Hence transmis-
sion may occur by direct exposure to soil or indirectly by con-
tact with surfaces contaminated via animal feces. The
prevalence and distribution of specific types of C. botulinum
spores varies with geographic region, habitat, and land use.
Clusters of outbreaks implicating fish products and preva-
lence data for fresh and marine water and sediments indicate
that type E may be aquatic (18). In contrast type B, and to a
lesser extent type A, are isolated from soils with high fre-
quency, suggesting a soil habitat (16, 19). However it should
be noted that C. botulinum type A, B, and E have been
detected in all environments, including the gastrointestinal
tract or feces of birds, mammals, and fish. Fecal shedding
undoubtedly contributes to the spread of C. botulinum spores
to soil. Notermans et al. (20) presented direct evidence that
spreading manure from cows infected with type B increased
the frequency of isolation of the same type from pasture soil
and grass. C. perfringens is edaphic; the species is readily iso-
lated from soils or sediments, however, it is also common in
the intestines and feces of many animal species (21). Further-
more, genotypic characterization of isolates associated with
food poisoning have cast doubt on the assumption that soil
is the main reservoir and points to additional ecological
niches, although soil is likely implicated as an indirect route
of transmission (22, 23).
Bacillus
Members of the B. cereus group are the most common Bacillus
species isolated from soil and are global in geographic distribu-
tion. Recovery using culture-based methods alone does not
provide proof of endemism because such methods do not
always differentiate between vegetative cells and endospores.
 3.3.2. Natural Soil Reservoirs for Human Pathogenic and Fecal Indicator Bacteria ▪ 3.3.2-3
The latter may be dispersed and accumulate in environments
that are not normally colonized by Bacillus spp. and additional
evidence is needed to establish the ecological niches that are
occupied by individual species. B. cereus and B. weihenstepha-
nensis have been recovered from a variety of soils and sedi-
ments, dust, and plants (24). Germination, growth, and
sporulation in soil extracts or soil microcosms suggest they
are soil saprophytes, although both can clearly occupy addi-
tional niches (25). Limited germination and growth of
B. anthracis in a model rhizosphere system has been demon-
strated (26), but evidence of a true saprophytic lifestyle is
lacking. Soil is still considered the main reservoir, although
much remains to be learned about the ecology of the species
outside of the animal host.
L. monocytogenes
Early work on the ecology of L. monocytogenes in the natural
environment revealed that the species is readily isolated from
agricultural or urban soils. Extensive survival or growth in soil
have been documented in experimental systems although soil
type, chemistry, biological activity, moisture content, motil-
ity and water activity influence experimental outcomes
(26). While soil is commonly considered to be a reservoir
where L. monocytogenes leads a saprophytic lifestyle, it is clear
that much remains to be learned about the ecology of the spe-
cies in the soil environment. The mechanisms underlying
resistance to stresses and the shift from soil-dwelling sapro-
phyte to lethal pathogen are also poorly understood (27,
28). L. monocytogenes is a very versatile pathogen able to
respond to a variety of suboptimal growth conditions. This
feature ensures that the pathogen can persist in a range of
environments, and it may be regarded as a ubiquitous envi-
ronmental bacterium. It has been isolated from soil, plants
and vegetables, decaying vegetation, surface water, ground-
water, biowastes, and composts (27, 29–37). Transfer of the
pathogen to soil after land-spreading of low-quality organic
fertilizers such as manure collected from contaminated live-
stock (38) or low-quality composts has been documented
(39). The use of contaminated wastewater for crop irrigation
is likely to introduce pathogenic bacteria into soil (40).
E. coli and Other Enteric Bacteria
Whether E. coli and enterococci are part of the indigenous
soil microbiota or transient populations from recent contam-
inations continues to be debated; several lines of evidence,
however, tend to support the resident population argument.
First, bacterial ubiquity: populations of indicator bacteria
E. coli, fecal coliforms, enterococci/fecal streptococci, and
some pathogens have been recovered in a range of soils, water,
and associated watersheds in subtropical islands Oahu,
Hawaii; Guam; Puerto Rico (41–46); subtropical continental
United States, south Florida (47, 48); and the Great Lakes
riparian soils and Irish maritime soils, both temperate (49–
51). E. coli and other fecal bacteria in many soils examined
also display patchiness, a common trait of many other indig-
enous bacteria (46, 50–52). Nevertheless, the findings that
E. coli is able to survive and persist in soils across seasons
(49, 51, 53, 54) renders support to the argument that these
bacteria have become resident or adapted to the soil
environment.
E. coli have also been shown to respond to changing sea-
sons in northern temperate soils of Lake Superior watersheds
(United States), with higher densities in summer to early fall,
June to October, relative to the winter and spring months,
February to May (51). Population fluxes have similarly been
shown to be influenced by other environmental variables,
such as moisture (48, 54–57) and nutrients (51, 54, 55).
While E. coli and some members of the Enterobacteriaceae
appear to be more susceptible to desiccation than enterococci
(55), residual populations can recover following rehydration
(55, 56). These findings collectively suggest that fecal bacte-
rial populations not only respond to changing soil conditions
but are also able to survive and persist in this environment
like other indigenous bacteria. E. coli growth occurs in indig-
enous soils in response to added nutrients and/or in the pres-
ence of reduced competition for nutrients (51, 54, 55). The
findings that metabolically diverse strains of E. coli and enter-
ococci are found in soils (58), the recovery of clonal E. coli
populations from soil (41, 51), and that certain E. coli strains
grow better in soil than others (51) support the growth
hypothesis. It has been postulated that the most likely habitat
for growth of E. coli in the soil environment is in the
nutrient-rich rhizosphere region of plants (59).
Mycobacterium
Soil is the natural habitat for many NTMs. A variety of Myco-
bacterium species have been recovered from soils (60, 61).
Although soil is not a natural habitat for the obligate patho-
gen M. bovis, it has been isolated from soils (62–64). In con-
trast, soil is a natural habitat of NTMs that include such
opportunistic pathogens as M. avium and M. intracellulare:
the M. avium complex (65–67). Long-term persistence of
mycobacteria in soil is likely due to their relative resistance
to environmental stressors, slow growth, and dormancy. In a
number of instances (68), the presence of mycobacteria in
soil has been linked to disease in animals or humans, under-
scoring the importance of developing methods for their
detection and/or isolation from soil.
S. maltophilia
Before being recognized as an emerging opportunistic patho-
gen, S. maltophilia is primarily known as a ubiquitous envi-
ronmental microorganism described in a variety of natural
environments such as soil (69, 70) or sediments (71). Rhizo-
sphere is a niche for this terrestrial species, and their role as
plant growth promoters or biological control agents of plant
pathogens has been the subject of numerous studies (72,
73). Their presence has also been reported in extreme ecosys-
tems such as deep sea or high altitudes (74, 75) as well as pol-
luted sites (76).
METHODS FOR DETECTING
AND QUANTIFYING SOIL
BACTERIAL PATHOGENS
Recovery of Bacteria from Soil
Detection or isolation of mycobacteria from soil presents dif-
ficulties as mycobacteria adhere tightly to soil particulates, are
difficult to lyse, and are capable of dormancy. Fortunately
methods are available for recovery, detection and enumera-
tion, and typing of soilborne mycobacteria. Separation of bac-
teria from soil can be challenging and may require rigorous
procedures. The hydrophobicity of mycobacteria cells (77)
promotes strong attachment to soil particles, and the use of
extractants containing detergents to disrupt hydrophobic
bonds is called for. In addition, incorporation of protein
hydrolysates, such as those included in microbial media (e.g.,
casamino acids or yeast extract), increases separation of
mycobacterial cells from soil particulates (78). Care must be
taken when attempting to separate mycobacterial cells from
soils in estuaries, as those soils may have high concentrations
3.3.2-4 ▪ SOIL
of salt. High salt concentrations increase the strength of
hydrophobic bonds. Increased recovery of mycobacteria
from soils can be attained by digestion of soil slurries with
enzymes able to hydrolyze the extracellular matrix that sur-
rounds cells in biofilms on soil particulates (79).
Mycobacteria are readily phagocytized and grow in proto-
zoa and amebae (80, 81). Most of the methods for preparing
mycobacterial cell suspensions result in lysis of the fragile cells
of both protozoa and amebae, so any intracellular mycobacte-
ria will be released to the suspension. Immunomagnetic beads
functionalized with antimycobacterial antibodies to bind and
concentrate mycobacterial cells from soil samples (82, 83) are
useful.
Culture-Dependent Detection and Enumeration
Culture-based methods for pathogen identification are the
gold standard and can be highly sensitive. They require
only basic equipment and yield viable organisms for further
characterization. These methods usually involve morpholog-
ical, biochemical, and/or genotypic typing after isolation and
culture of single colonies grown on selective media to further
confirm isolate identification. These steps might make this
approach lengthy and are not ideal for surveys requiring the
investigation of a very large number of samples.
Traditional methods for the detection or enumeration of
pathogens in soil have long relied on the use of semi-selective
differential solid media and incubation conditions. Methods
used for the analysis of soil are generally adapted from meth-
ods developed for clinical purposes or food analysis. In the
absence of standardized methods, the selection of an
approach for detection or enumeration of bacterial pathogens
should be guided by several considerations. Physiological
injuries derived from exposure to temperature extremes,
osmotic shock, competition, dehydration or the presence of
inhibitors such as minerals, humic substances, and man-made
contaminants among others, can severely compromise growth
on selective media and the accuracy of quantitative analysis.
Recovery methods on solid media using agar overlay/under-
lays or filtration methods may be used to improve the perform-
ance of selective media. Enrichment and pre-enrichment are
routinely applied for the recovery of physiologically compro-
mised cells, low populations, or for strictly qualitative assess-
ments of soil quality. Examples of approaches and media used
with specific bacterial pathogens are given in Table 1.
L. monocytogenes
Methods readily available for environmental and soil samples
are adapted from the food industry. The U.S. Department of
Agriculture, FDA, and ISO have developed three main
TABLE 1 Examples of methods used for the recovery of human pathogens from soils
Species Sample preparation
Clostridium perfringens Recovery from soil by filtration
followed by decimal dilution
Clostridium botulinum Dilution in saline
Clostridium tetani Heat shock in primary enrichment
medium
Bacillus anthracis Flotation extraction and ethanol
purification of spores
Listeria monocytogenes Direct enrichment
reference methods. These methods are culture-based and
require enrichment in selective liquid media followed by
isolation on selective agar media. The FDA Bacteriological
and Analytical Method uses a one-step enrichment in Listeria
enrichment broth (89), whereas the ISO 11290 method
requires a pre-enrichment step in half-strength Fraser broth
followed by subculturing in full-strength Fraser. In the USDA
method 993.12 (90), the two-step enrichment proceeds in
modified University of Vermont medium. In these reference
methods, selective enrichments are plated on Oxford and/
or PALCAM agar media. Several chromogenic selective agars
such as ALOA, Rapid’L, mono, or CHROMagar Listeria test
are available commercially for isolation of Listeria sp. and
L. monocytogenes. Because of the diversity of the soil microor-
ganisms, nonspecific growth may be observed and selectivity
of the media may be unacceptably low. Addition of cyclohex-
imide is useful to avoid overgrowth of agar by fungi. L. mono-
cytogenes grows well in enrichment at 4°C, giving it a
competitive advantage, but long incubation periods of several
weeks are required. A major consideration for the use of these
methods is that they are labor-intensive and require up to 7
days prior to definite identification of the pathogen.
E. coli and Other Enteric Bacterial Species
The isolation of fecal bacteria from environmental sources
is fairly straightforward due to the availability of selective
and differential media for many of these microorganisms
(91), including toxigenic E. coli (92). Many pathogens reside
in soil from application of manures, septage, and sewage (93)
or addition of municipal wastes. For example, thermal toler-
ant E. coli and enterococci can be directly isolated from soil,
water, sediment, and sand environments by plating onto
modified mTEC and mEI agar media, respectively. This can
be done by direct plating or drop plating or by using the mem-
brane filtration method (94). There are also selective and dif-
ferential media for many enteric pathogens, but since the
numbers of these microbes in soils may be low, they may
require prior enrichment or resuscitation prior to enumera-
tion (95).
Culture-Dependent Isolation and Enumeration of
Mycobacterium spp.
There are no truly selective media for the very slow-growing
mycobacteria, and overgrowth with bacteria and fungi is a
serious problem. The inclusion of malachite green in the
standard Middlebrook broth and agar media will help sup-
press fungi. Cetylpyridinium chloride 0.005% for 30 min to
24 h exposure will kill most nonmycobacteria. A variety of
other decontaminating agents have been used (e.g., 1%
Media Reference
Lactose sulfite broth 84
Cooked meat medium; confirmation by 85
intraperitoneal injection of spent medium in mice
Enrichment in blood broth, selective plating on 86
Zeissler’s agar
Culture on PLET polymyxin, lysozyme, EDTA, 87
thallium acetate medium
Selective enrichment in Listeria enrichment broth, 88
plating on Oxford medium
 3.3.2. Natural Soil Reservoirs for Human Pathogenic and Fecal Indicator Bacteria ▪ 3.3.2-5
NaCl for 30 min, 1% oxalic acid for 30 min), but they kill a
substantial number of mycobacterial cells and should be
avoided (78). Furthermore, it is important to ensure that
any decontaminating agent be inactivated before suspensions
are spread on agar medium (e.g., M7H10 agar or inoculated
into broth medium M7H9 or M. avium subsp. paratuberculosis
[MAP]).
Following the suspension and separation of mycobacterial
cells from soil slurries, colonies can be identified and enumer-
ated on media encouraging the growth of mycobacteria. The
best medium is Middlebrook 7H10 agar containing 1% glyc-
erol and 10% oleic acid–albumin. The medium supports
growth of mycobacteria colonies within 1–4 weeks and is
clear, allowing for the detection of the small and faint trans-
parent colony variants. Transparent colony variants are more
virulent and the form most often recovered from infected
individuals. Transparent colonies cannot be seen on the egg-
based Lowenstein-Jensen solid medium. Direct plating has an
advantage over plating from broth cultures as mycobacteria
will grow in broth. Mycobacterial-like colonies (e.g., 1 mm
diameter, yellow or unpigmented) appearing as soon as 7
days and as long as 3–4 weeks are picked, restreaked, and acid-
fast stained. Viable counts of mycobacteria undoubtedly
underestimate total populations because dormant cells are
not recovered (96).
Acid-fast colonies are best identified by PCR-based tech-
niques, rather than cultural, biochemical, and enzymatic tests
that lack the necessary discriminatory power. Some of the
PCR primers can be employed for direct detection and enu-
meration qPCR of mycobacteria in suspensions as they
amplify species-specific sequences; for example IS900 for
M. avium subspecies paratuberculosis.
Culture of S. maltophilia
Considering the opportunistic pathogen S. maltophilia, isola-
tion on agar plates has been initially performed on Xanthomo-
nas maltophilia selective medium (XMSM) as described for
isolation of Xanthomonas maltophilia from bulk soil and plant
rhizosphere (69). This culture medium contains at least six
antibacterial agents, two antifungal agents, maltose, and bro-
mothymol blue, allowing the screening of orange colonies
with a yellow halo after 2 days of incubation. XMSM prepara-
tion is complex, time-consuming, expensive, and not highly
selective or specific enough for the isolation of S. maltophilia
recovered from the environment (97). The VIA medium was
then proposed for the isolation of S. maltophilia from clinical
and environmental samples (97). VIA contains imipenem as
a selective agent for S. maltophilia. Furthermore, vancomycin
can be included to inhibit the growth of Enterococcus faecium,
and amphotericin B to inhibit fungi. Mannitol and bromo-
thymol blue are added to VIA agar to distinguish between
mannitol-fermenting and -nonfermenting microorganisms.
S. maltophilia does not produce acid from mannitol, and
thus appears as green colonies with a blue halo. VIA medium
was demonstrated to be less inhibitory to S. maltophilia and
more selective than XMSM in preventing the growth of other
bacteria like maltose fermenters. The combination of VIA
medium isolation and PCR screening is effective for the iso-
lation and confirmation of S. maltophilia (98).
Enrichment Methods Using Immunomagnetic
Capture of Surface-Exposed Antigens
Immunomagnetic capture (IMC) is a procedure useful for the
specific detection of slow-growing bacteria because these are
difficult to cultivate from environmental matrixes. Indeed,
IMC was used to physically concentrate M. bovis from soils,
sediments, or feces to improve molecular detection sensitiv-
ity and optimize culture-based detection methods. The prin-
ciple of this method involves microscopic magnetic beads
coated with monoclonal or polyclonal antibodies, peptides,
or other macromolecules that target bacteria of interest.
Two options are available: direct IMC, where magnetic
beads coated with the antibody or other ligands are added
to samples, and indirect IMC, where a primary antibody
or ligand is added to samples, then magnetic beads coated
with a secondary antibody are added to the samples. Mag-
netic separation of the complex bead-target bacteria allows
selective recovering of the target-enriched sample fractions.
Sweeney et al. (63, 83) successfully employed either direct
or indirect IMC to selectively recover M. bovis from badger
feces and soils. Subsequent cultivation of M. bovis from
IMC-enriched samples was performed on semi-selective or
selective media (83).
More recently, direct IMC of MAP from milk samples was
achieved using two phage display-derived peptides coating
tosyl-activated Dynabeads (99). This method of IMC, fol-
lowed by optimized phage assay, allowed detection of MAP
in milk samples within 48 h. Finally, Stewart et al. (100)
used a combination of a peptide ligand with a monoclonal
antibody to capture M. bovis from contaminated animal tis-
sues. These most recent methods need to be validated for
soils.
Molecular Detection and Quantification of
Bacterial Pathogens from Soils
Culture-dependent approaches are time-consuming and
therefore limit the number of samples that can be tested at
any one time. They are not always very specific, they involve
extensive biosafety measures, and some organisms simply
resist cultivation. Molecular methods for the detection of
pathogens are more rapid and specific. Molecular detection
and enumeration methods are developed in the following sec-
tions for several bacterial species or genera.
DNA Extraction from Soil and Soil Particulates
Numerous commercial kits for extracting DNA from soil
microbial communities are available and widely used. One
drawback of these methods is that DNA extracted from soil
matrixes may contain a number of compounds, including
humic and fulvic acids, that can inhibit PCR (62, 101).
Thus, suitable methods to evaluate PCR inhibition by inhib-
itors coextracted in soil DNA must be included in protocols
(102, 103). Alternative soil DNA extraction and purification
methods, which give higher quality DNA, are also available
(104, 105). An ISO method for soil DNA extraction was
also recently validated (106).
Molecular Detection of Soil Mycobacterium
Molecular methods are advisable for the detection of myco-
bacteria in soil because they are very slow growing on isolation
media. The lipid-rich outer membrane of mycobacterial cells
(77) makes them recalcitrant to DNA extraction. A thorough
inventory of mycobacterial populations in a particular sample
requires the use of either a collection of species-specific pri-
mers or sequencing PCR-generated 16S rRNA amplicons
(107). Systematic studies of factors influencing PCR detec-
tion of mycobacteria in soils have been recently published
(62, 101). Nested PCR primers used for the simultaneous
3.3.2-6 ▪ SOIL
identification of Mycobacterium, M. avium, and M. intracellu-
lare (108) can be used for direct enumeration of Myco-
bacterium and M. avium by qPCR (109). For any of the
PCR-based techniques, it is important to ensure the sensitiv-
ity, limit of detection, and specificity. Many of the available
primers for Mycobacterium species identification were devel-
oped for use with clinical samples, which are inherently less
diverse than soils and therefore less likely to yield false posi-
tives. Finding primers that are truly specific for different
Mycobacterium species is a problem (110). As more Mycobac-
terium species are discovered, detection methods will contin-
ually require reassessment (111).
Coextraction of inhibitors with DNA poses an acute
problem for the direct PCR detection or enumeration of
Mycobacterium spp. as one of their preferred habitats is fulvic-
and humic-rich peat-rich pine forest and estuarine soils (112,
113). Thus, any method for PCR detection or enumeration
of mycobacteria in soils should include a step to reduce
humic and fulvic acid concentrations. One approach is to
employ a commercial kit specifically designed for soil
DNA isolation.
Molecular Detection of L. monocytogenes
Several methods have been described for the direct molecular
detection of L. monocytogenes (114). Relying on either con-
ventional or real-time PCR, these methods have been devel-
oped for foodstuffs. Recently, a Taqman method was designed
for the direct detection of the pathogen from soil (103). Pri-
mers and the Taqman probe target an internal fragment of prs,
the gene coding a phosphoribosylpyrophosphate synthetase.
This assay is specific to L. monocytogenes and quantitative
down to 104 cells/g of soil.
Deep Sequencing of Metagenomic Soil DNA to
Detect Pathogens
Metagenomic characterization of bacterial communities
provides an assessment of microbial phylogenetic structure
and diversity and has recently emerged as a powerful tool
in the study of a variety of matrixes including soils (115),
marine (116), and freshwater (117, 118) environments.
While earlier approaches to metagenomic analyses relied
on the direct cloning of DNA or PCR-amplified 16S
rDNA from the environmental into E. coli (119), they suf-
fered from bias associated with cloning and sequencing a
relatively small number of clones (120). In contrast, next-
generation sequencing has overcome many of these limita-
tions by using massive parallel sequencing of selective and
highly variable region of the 16S rDNA (e.g., the V4,
V5, or V6 hypervariable regions) from any matrix, without
the need for cloning. This has allowed the identification of
rare populations in low abundance in bacterial communities
and the characterization of microbial communities in
marine waters (121), soils (122), and freshwater (118) envi-
ronments. As expected, the microbial community structure
in freshwater differ significantly from those found in soils or
in marine environments (123). Using primers specific for
16S rDNA, such as V4, V5, V6, or V4–V6, only limited
taxonomic information can determined accurately. Conse-
quently, information concerning pathogens can be only
inferred from higher level taxonomic assignments made to
family or genera containing pathogens. More recently,
genomic-based microfluidic platforms have been used in
conjunction with sequence based analyses to better define
multiple pathogens in waterways (124).
METHODS FOR STRAIN PHENOTYPING/
GENOTYPING: COMPARISON OF
PHENOTYPES OR GENOTYPES OF
ENVIRONMENTAL VERSUS CLINICAL ORIGIN
Fecal Indicators (E. coli, Enterococci) and Other
Enteric Bacteria
In a recent study Luo et al. (125) showed that environmental
E. coli strains share traits that are clearly distinct from their
enteric counterparts (125). However, these findings were
based on a limited number of environmental E. coli strains
(7), representing freshwater beach, soil, and sediment habi-
tats; thus, more environmental E. coli strains from geograph-
ically different locations, habitats, and substrates are needed
to better understand and characterize the populations that
are often referred to autochthonous, environmental strains.
While the original source of E. coli or other fecal indicator
bacteria in soil remains speculative, isolates from fecal contri-
butions, especially from wildlife, can be identified (51, 53). In
an earlier study, Fujioka and Byappanahalli (58) proposed that
the soil environment does not necessarily select for one or two
dominant strains of fecal indicator bacteria (E. coli, entero-
cocci); instead, numerous strains with a range of metabolic
diversity can potentially inhabit this environment. To further
address this hypothesis, Byappanahalli et al. (41) examined
the population structure and genetic relatedness of E. coli iso-
lated from subtropical Hawaii and temperate Great Lakes
watershed soils. A few significant findings of this study were
that E. coli isolates were genetically diverse in all soils exam-
ined, unique E. coli strains were found in both tropical and
temperate soils, and E. coli isolates were largely clustered by
location. These authors concluded that multiple E. coli gen-
otypes can adapt to become part of the soil microflora. E. coli
genotypic diversity in soil has also been examined in other
studies, including temperate Irish soils (49), pristine forest
soils of Puerto Rico (46), and a recreational meadow in the
United States that was affected by animal fecal inputs (52).
Collectively, these studies support the persistence of geneti-
cally diverse E. coli strains, as widely reported in the studies
by Byappanahalli et al. (41) and Ishii et al. (126).
Aside from soil, a number of studies have examined the
population structure and genetic relatedness of other fecal
bacteria, including some pathogens, in other habitats/sub-
strates, including beach sand (53, 127), the macrophytic
green algae Cladophora (50, 126, 128, 129), forage crops
(130, 131), and other aquatic vegetation (129). Some signifi-
cant findings of these studies include a recovery of genotypi-
cally diverse of E. coli (53, 127) and different Enterococcus
spp. in beach sand from recreational beaches (53). It is
thought that the abundance of these bacteria in beach sand
is due to in situ growth under certain conditions (53, 127,
132, 133). E. coli O157:H7, Shigella, Salmonella, and Campy-
lobacter have been recovered in the green algae Cladophora in
the Great Lakes (126, 134) and elsewhere (135).
Plant-adapted, epiphytic Enterococcus spp. have been
observed in a number of forage crops (130, 131) and sub-
merged aquatic vegetation (129). Whether these bacteria
have any ecological roles in these habitats remain purely spec-
ulative, however, their distribution and abundance in the
environment can have potential impact on water quality
and human health.
Identification and Characterization of Mycobacteria
Although much of what is known about mycobacterial
ecology has come from isolates identified by cultural,
 3.3.2. Natural Soil Reservoirs for Human Pathogenic and Fecal Indicator Bacteria ▪ 3.3.2-7
TABLE 2 Summary of DNA-based methods for identification/
typing of Mycobacteria
Mycobacterium
Target Reference
species
Mycobacterium spp. 16S rRNA PCR 108
Mycobacterium spp. hsp65 RFAa 137, 138
Mycobacterium spp. LRF-PFGE 139
M. bovis IS1081 RFLP 144
M. avium subsp. IS900 RFLP 145
paratuberculosis
M. avium IS1245/1311 RFLP 140
M. avium rep-PCR 141
M. intracellulare MIRU-RFLP 143
a
RFA: Restriction Fragment Analysis.
biochemical, and enzymatic tests, current practice employs
DNA sequence analysis for identification. A motivator for
such a shift was knowledge that M. avium and M. intracellulare
species that only share 70% DNA similarity (136) and could
not be distinguished by cultural, biochemical, or enzymatic
tests. One widely used approach is analysis of restriction frag-
ments generated by restriction endonuclease HaeIII and
BstEII of the mycobacterial 65 kDa heat shock protein gene
hsp65 (137, 138). As a result of the difficulty in cultivating
M. avium subsp. paratuberculosis, positive identification of
that species is almost always made by amplification of the
species-specific IS900 (67). As the list of mycobacterial spe-
cies continues to grow, it has become apparent that many spe-
cies share similar if not identical 16S rRNA sequences as is
the case for M. intracellulare and M. chimaera (110). In addi-
tion to examining the sequence of other genes, for example,
rpoB and the internal transcribed sequence of the rRNA
genes, the field is being increasingly driven toward requiring
more and longer sequence data.
Typing tools for the major mycobacterial pathogens are
presented in Table 2. Fortunately, a variety of methods are
available including those using large restriction fragment
(LRF) polymorphisms employing pulse-field gel electro-
phoresis (PFGE) (139) and insertion sequences (ISs) (140),
repetitive sequences rep-PCR (141) or tandem repeat units
such as the mycobacterial interspersed repeat unit (MIRU)
and variable number tandem repeat (VNTR) analysis (142,
143). Although many consider LRF polymorphisms the
gold standard, the preparation of cells for restriction endonu-
clease digestion followed by PFGE is simply too time-
consuming and laborious. Furthermore, the data generated
by LRF-PFGE and IS polymorphisms is not amenable to
easily transferable, numerical types, unlike MIRU-VNTR.
In that MIRU-VNTR analysis provides the number of repeats
for each tandem repeat amplified by PCR, the data are trans-
missible throughout the world.
Typing of L. monocytogenes
The phylogeny of the species L. monocytogenes is character-
ized by the grouping of isolates in four lineages (146, 147)
and major clonal complexes distributed worldwide are recog-
nized (148).
PFGE, multilocus sequence typing, and ribotyping are
widely used for genotyping L. monocytogenes isolates. For a
comprehensive review of the genotyping methods available,
see Jadhav et al. (114). Some reports suggest that distribution
of clones and serotypes differ among environmental versus
clinical and food isolates (148–151).
Among farm animal isolates, EcoRI ribotyping identified
several subtypes (152, 153). In a case study, Nightingale
et al. (153) showed that the incidence of one of the ribotypes
was higher in soil, whereas others were preferentially associ-
ated with fecal samples and animals listeriosis. In other studies
focusing on the genotypic diversity of Listeria isolates from
soil, water, and vegetation from natural areas of the state of
New York and from urban soil, water, and vegetation samples
found in cities (88, 154), higher detection of specific subtypes
in specific sample sites was also reported. Whether niche-spe-
cific genotypes exist is still a debate, but environmental iso-
lates are usually potentially virulent strains when tested in
in vitro cell invasion assays (36, 37).
CONCLUSION
Microbiologists have studied soil bacteria for well over 150
years, yet there remain major challenges in assessing soilborne
pathogens of human concern. Devising specific and effective
culturing methods is a challenge, particularly in light of the
enormous diversity in soil microbial communities. Many
pathogens, for example, Mycobacterium spp., are very fastidi-
ous and slow-growing, or require complex nutritional supple-
ments that support the growth of soil heterotrophs. Variation
in soil properties, for example, clay content or the presence of
humic materials, causes great variation in the extractability of
bacteria or nucleic acid. Coextraction of PCR inhibitors must
be considered and controlled for. There is growing evidence
that several major human bacterial pathogens may inhabit
soils, some harbored in protozoa or other hosts. Developments
of molecular tools for the detection, quantitation, and char-
acterization of pathogens, largely flowing from work in the
medical and food quality sectors, are being deployed by soil
microbiologists. Altogether, methodology is very much in a
state of flux, and there is great potential for exciting new
methods available in the future.
REFERENCES
1. Alcaraz LD, Moreno-Hagelsieb G, Eguiarte LE, Souza V,
Herrera-Estrella L, Olmedo G. 2010. Understanding the
evolutionary relationships and major traits of Bacillus
through comparative genomics. BMC Genomics 11:332.
2. Dierick K, Van Coillie E, Swiecicka I, Meyfroidt G, Dev-
lieger H, Meulemans A, Hoedemaekers G, Fourie L,
Heyndrickx M, Mahillon J. 2005. Fatal family outbreak
of Bacillus cereus-associated food poisoning. J Clin Microbiol
43:4277–4279.
3. Guillet C, Join-Lambert O, Le Monnier A, Leclercq A,
Mechaï F, Mamzer-Bruneel MF, Bielecka MK, Scortti
M, Disson O, Berche P, Vazquez-Boland J, Lortholary
O, Lecuit M. 2010. Human listeriosis caused by Listeria iva-
novii. Emerg Infect Dis 16:136–138.
4. Donnelly CW. 2001. Listeria monocytogenes: a continuing
challenge. Nutr Rev 59:183–194.
5. Byappanahalli MN, Ishii S. 2011. Environmental sources
of fecal bacteria, p 93–110. In Sadowsky MJ, Whitman RL
(eds), The Fecal Bacteria. ASM Press, Washington, DC.
6. Ferguson D, Signoretto C. 2011. Environmental persis-
tence and naturalization of fecal indicator organisms,
p 379–397. In Hagedorn C, Blanch AR, Harwood VJ
(eds), Microbial Source Tracking: Methods, Applications,
and Case Studies. Springer, New York.
7. Whitman R, Nevers M, Przybyla-Kelly K, Byappanahalli
MN. 2011. Physical and biological factors influencing
3.3.2-8 ▪ SOIL
environmental sources of fecal indicator bacteria in surface
water, p 111–134. In Sadowsky MJ, Whitman RL (eds),
The Fecal Bacteria. ASM Press, Washington, DC.
8. Howell JM, Coyne MS, Cornelius P. 1995. Fecal bacteria
in agricultural waters of the Bluegrass Region of Kentucky.
J Environ Qual 24:411–419.
9. Stephenson GR, Street LV. 1978. Bacterial variations in
streams from a Southwest Idaho rangeland watershed.
J Environ Qual 7:150–156.
10. Grange JM. 1996. The biology of the genus Mycobacterium.
J Appl Bacteriol 81:S1–S9.
11. Primm TP, Lucero CA, Falkinham JO. 2004. Health
impacts of environmental mycobacteria. Clin Microbiol
Rev 17:98–106.
12. Brooke JS. 2012. Stenotrophomonas maltophilia: an emerg-
ing global opportunistic pathogen. Clin Microbiol Rev 25:
2–41.
13. Jones RN, Sader HS, Beach ML. 2003. Contemporary
in vitro spectrum of activity summary for antimicrobial
agents tested against 18569 strains non-fermentative
Gram-negative bacilli isolated in the SENTRY Antimicro-
bial Surveillance Program (1997–2001). Int J Antimicrob
Agents 22:551–556.
14. Razvi S, Quittell L, Sewall A, Quinton H, Marshall B,
Saiman L. 2009. Respiratory microbiology of patients
with cystic fibrosis in the United States, 1995 to 2005.
Chest 136:1554–1560.
15. Sanchez MB, Hernandez A, Martinez JL. 2009. Stenotro-
phomonas maltophilia drug resistance. Future Microbiol 4:
655–660.
16. Smith LD. 1978. Occurrence of Clostridium botulinum and
Clostridium tetani in soil of United-States. Health Lab Sci
15:74–80.
17. Cook TM, Protheroe RT, Handel JM. 2001. Tetanus: a
review of the literature. Br J Anaesth 87:477–487.
18. Huss HH. 1980. Distribution of Clostridium botulinum.
Appl Environ Microbiol 39:764–769.
19. Smith GR, Young AM. 1980. Clostridium botulinum in
British soil. J Hyg (Lond) 85:271–274.
20. Notermans S, Dufrenne J, Oosterom J. 1981. Persistence
of Clostridium botulinum Type-B on a cattle farm after
an outbreak of botulism. Appl Environ Microbiol 41:
179–183.
21. Matches JR, Liston J, Curran D. 1974. Clostridium perfrin-
gens in the environment. Appl Microbiol 28:655–660.
22. Li JH, Sayeed S, McClane BA. 2007. Prevalence of enter-
otoxigenic Clostridium perfringens isolates in Pittsburgh
(Pennsylvania) area soils and home kitchens. Appl Environ
Microbiol 73:7218–7224.
23. Lahti P, Lindstrom M, Somervuo P, Heikinheimo A,
Korkeala H. 2012. Comparative genomic hybridization
analysis shows different epidemiology of chromosomal
and plasmid-borne cpe-carrying Clostridium perfringens
Type A. PLoS One 7:e46162.
24. Arnesen LPS, Fagerlund A, Granum PE. 2008. From soil
to gut: Bacillus cereus and its food poisoning toxins. FEMS
Microbiol Rev 32:579–606.
25. Vilain S, Luo Y, Hildreth MB, Brozel VS. 2006. Analysis
of the life cycle of the soil saprophyte Bacillus cereus in
liquid soil extract and in soil. Appl Environ Microbiol 72:
4970–4977.
26. Saunders BD, Wideman M. 2007. Ecology of Listeria spe-
cies and L. monocytogenes in the natural environment.
Food Sci Technol 161:21–54.
27. Freitag NE, Port GC, Miner MD. 2009. Listeria monocyto-
genes—from saprophyte to intracellular pathogen. Nat Rev
Microbiol 7:623–628.
28. Gorski L, Duhe JM, Flaherty D. 2011. The Sigma
B operon is a determinant of fitness for a Listeria
monocytogenes serotype 4b strain in soil. Foodborne Pathog
Dis 8:699–704.
29. Welshimer HJ. 1960. Survival of Listeria monocytogenes in
soil. J Bacteriol 80:316–320.
30. Weis J, Seeliger HPR. 1975. Incidence of Listeria monocy-
togenes in nature. Appl Microbiol 30:29–32.
31. Dowe MJ, Jackson ED, Mori JG, Bell CR. 1997. Listeria
monocytogenes survival in soil and incidence in agricultural
soils. J Food Prot 60:1201–1207.
32. Fenlon DR, Wilson J, Donachie W. 1996. The incidence
and level of Listeria monocytogenes contamination of food
sources at primary production and initial processing.
J Appl Bacteriol 81:641–650.
33. Garrec N, Picard Bonnaud F, Pourcher AM. 2003.
Occurrence of Listeria sp. and L. monocytogenes in sewage
sludge used for land application: effect of dewatering, lim-
ing and storage in tank on survival of Listeria species.
FEMS Immunol Med Microbiol 35:275–283.
34. Welshimer HJ. 1968. Isolation of Listeria monocytogenes
from vegetation. J Bacteriol 95:300–303.
35. Welshimer HJ, Donker-Voet J. 1971. Listeria monocyto-
genes in nature. Appl Microbiol 21:516–519.
36. Lyautey E, Hartmann A, Pagotto F, Tyler K, Lapen DR,
Wilkes G, Piveteau P, Rieu A, Robertson WJ, Medeiros
DT, Edge TA, Gannon V, Topp E. 2007. Characteristics
and frequency of detection of fecal Listeria monocytogenes
shed by livestock, wildlife, and humans. Can J Microbiol
53:1158–1167.
37. Lyautey E, Lapen DR, Wilkes G, McCleary K, Pagotto F,
Tyler K, Hartmann A, Piveteau P, Rieu A, Robertson
WJ, Medeiros DT, Edge TA, Gannon V, Topp E. 2007.
Distribution and characteristics of Listeria monocytogenes
isolates from surface waters of the South Nation River
watershed, Ontario, Canada. Appl Environ Microbiol 73:
5401–5410.
38. Guan TY, Holley RA. 2003. Pathogen survival in swine
manure environments and transmission of human enteric
illness: a review. J Environ Qual 32:383–392.
39. Al-Ghazali MR, Al-Azawi SK. 1990. Listeria monocyto-
genes contamination of crops grown on soil treated with
sewage sludge cake. J Appl Bacteriol 69:642–647.
40. Paillard D, Dubois V, Thiebaut R, Nathier F,
Hoogland E, Caumette P, Quentin C. 2005. Occur-
rence of Listeria spp. in effluents of French urban waste-
water treatment plants. Appl Environ Microbiol 71:
7562–7566.
41. Byappanahalli MN, Roll BM, Fujioka RS. 2012. Evidence
for occurrence, persistence, and growth potential of Escher-
ichia coli and enterococci in Hawaii’s soil environments.
Microbes Environ 27:164–170.
42. Goto DK, Yan T. 2011. Effects of land uses on fecal indi-
cator bacteria in the water and soil of a tropical watershed.
Microbes Environ 26:254–260.
43. Fujioka R, Sian-Denton C, Borja M, Castro J, Morphew
K. 1999. Soil: the environmental source of Escherichia coli
and enterococci in Guam’s streams. J Appl Microbiol 85:
83S–89S.
44. Hardina CM, Fujioka RS. 1991. Soil—the environmental
source of Escherichia coli and enterococci in Hawaii streams.
Environ Toxicol Water Qual 6:185–195.
45. Hazen TC. 1988. Fecal coliforms as indicators in tropical
waters—a review. Tox Assess 3:461–477.
46. Lasalde C, Rodriguez R, Toranzos GA, Smith HH. 2005.
Heterogeneity of uidA gene in environmental Escherichia
coli populations. J Water Health 3:297–304.
47. Desmarais TR, Solo-Gabriele HM, Palmer CJ. 2002.
Influence of soil on fecal indicator organisms in a tidally
influenced subtropical environment. Appl Environ Micro-
biol 68:1165–1172.
 3.3.2. Natural Soil Reservoirs for Human Pathogenic and Fecal Indicator Bacteria ▪ 3.3.2-9
48. Solo-Gabriele HM, Wolfert MA, Desmarais TR, Palmer
CJ. 2000. Sources of Escherichia coli in a coastal subtropical
environment. Appl Environ Microbiol 66:230–237.
49. Brennan FP, Abram F, Chinalia FA, Richards KG,
O’Flaherty V. 2010. Characterization of environmentally
persistent Escherichia coli isolates leached from an Irish
soil. Appl Environ Microbiol 76:2175–2180.
50. Byappanahalli MN, Shively DA, Nevers MB, Sadowsky
MJ, Whitman RL. 2003. Growth and survival of Escheri-
chia coli and enterococci populations in the macro-alga
Cladophora (Chlorophyta). FEMS Microbiol Ecol 46:
203–211.
51. Ishii S, Yan T, Shively DA, Byappanahalli MN, Whit-
man RL, Sadowsky MJ. 2006. Cladophora (Chlorophyta)
spp. harbor human bacterial pathogens in nearshore
water of Lake Michigan. Appl Environ Microbiol 72:
4545–4553.
52. Bergholz PW, Noar JD, Buckley DH. 2011. Environmen-
tal patterns are imposed on the population structure of
Escherichia coli after fecal deposition. Appl Environ Microbiol
77:211–219.
53. Byappanahalli MN, Whitman RL, Shively DA, Sadow-
sky MJ, Ishii S. 2006. Population structure, persistence,
and seasonality of autochthonous Escherichia coli in temper-
ate, coastal forest soil from a Great Lakes watershed. Environ
Microbiol 8:504–513.
54. Whitman RL, Nevers MB, Byappanahalli MN. 2006.
Examination of the watershed-wide distribution of Escher-
ichia coli along Southern Lake Michigan: an integrated
approach. Appl Environ Microbiol 72:7301–7310.
55. Byappanahalli MN, Fujioka RS. 2004. Indigenous soil
bacteria and low moisture may limit but allow faecal bacte-
ria to multiply and become a minor population in tropical
soils. Water Sci Technol 50:27–32.
56. Chandler DS, Craven JA. 1980. Relationship of soil-
moisture to survival of Escherichia coli and Salmonella typhi-
murium in soils. Aust J Agric Res 31:547–555.
57. Hunter C, McDonald A. 1991. The occurrence of coliform
bacteria in the surface soils of 2 catchment areas in the
Yorkshire Dales. J Inst Water Environ Manag 5:534–538.
58. Fujioka RS, Byappanahalli MN 2001. Microbial ecology
controls the establishment of fecal bacteria in tropical soil
environment, p. 273–283. In Matsuo TH, Takizawa K,
Satoh SH (eds), Advances in Water and Wastewater Treat-
ment Technology: Molecular Technology, Nutrient Removal,
Sludge Reduction and Environmental Health. Elsevier,
Amsterdam.
59. Walker TS, Bais HP, Grotewold E, Vivanco JM. 2003.
Root exudation and rhizosphere biology. Plant Physiol
132:44–51.
60. Biet F, Boschiroli ML, Thorel MF, Guilloteau LA.
2005. Zoonotic aspects of Mycobacterium bovis and Myco-
bacterium avium-intracellulare complex (MAC). Vet Res
36:411–436.
61. Kazda J, Pavlik I, Falkinham JO, Hruska K. 2009. The
Ecology of Mycobacteria: Impact on Animal and Human
Health. Springer, Rotterdam.
62. Young JS, Gormley E, Wellington EM. 2005. Molecular
detection of Mycobacterium bovis and Mycobacterium bovis
BCG (Pasteur) in soil. Appl Environ Microbiol 71:1946–
1952.
63. Sweeney FP, Courtenay O, Ul-Hassan A, Hibberd V,
Reilly LA, Wellington EMH. 2006. Immunomagnetic
recovery of Mycobacterium bovis from naturally infected
environmental samples. Lett Appl Microbiol 43:364–369.
64. Travis ER, Gaze WH, Pontiroli A, Sweeney FP, Porter D,
Mason S, Keeling MJC, Jones RM, Sawyer J, Aranaz A,
Rizaldos EC, Cork J, Delahay RJ, Wilson GJ, Hewinson
RG, Courtenay O, Wellington EMH. 2011. An inter-
laboratory validation of a real time PCR assay to measure
host excretion of bacterial pathogens, particularly of Myco-
bacterium bovis. PLoS One 6:e27369.
65. Brooks RW, Parker BC, Gruft H, Falkinham JO. 1984.
Epidemiology of infection by nontuberculous mycobacte-
ria.5. Numbers in eastern-United-States soils and correla-
tion with soil characteristics. Am Rev Respir Dis 130:
630–633.
66. Radomski N, Thibault VC, Karoui C, de Cruz K,
Cochard T, Gutiérrez C, Supply P, Biet F, Boschiroli
ML. 2010. Determination of genotypic diversity of Myco-
bacterium avium subspecies from human and animal origins
by mycobacterial interspersed repetitive-unit-variable-
number tandem-repeat and IS1311 restriction fragment
length polymorphism typing methods. J Clin Microbiol
48:1026–1034.
67. Pickup RW, Rhodes G, Arnott S, Sidi-Boumedine K,
Bull TJ, Weightman A, Hurley M, Hermon-Taylor J.
2005. Mycobacterium avium subsp paratuberculosis in the
catchment area and water of the river Taff in South Wales,
United Kingdom, and its potential relationship to cluster-
ing of Crohn’s disease cases in the city of Cardiff. Appl Envi-
ron Microbiol 71:2130–2139.
68. De Groote MA, Pace NR, Fulton K, Falkinham JO.
2006. Relationships between Mycobacterium isolates from
patients with pulmonary mycobacterial infection and pot-
ting soils. Appl Environ Microbiol 72:7602–7606.
69. Juhnke ME, des Jardin E. 1989. Selective medium for iso-
lation of Xanthomonas maltophilia from soil and rhizosphere
environments. Appl Environ Microbiol 55:747–750.
70. Romano G, Stampi S, Zanetti F, De Luca G, Tonelli E.
1997. Occurrence of gram-negative bacteria in drinking
water undergoing softening treatment. Int J Hyg Environ
Health 200:152–162.
71. Dungan RS, Yates SR, Frankenberger WT. 2003. Trans-
formations of selenate and selenite by Stenotrophomonas
maltophilia isolated from a seleniferous agricultural drainage
pond sediment. Environ Microbiol 5:287–295.
72. Berg G, Eberl L, Hartmann A. 2005. The rhizosphere as a
reservoir for opportunistic human pathogenic bacteria.
Environ Microbiol 7:1673–1685.
73. Ryan RP, Monchy S, Cardinale M, Taghavi S, Crossman
L, Avison MB, Berg G, van der Lelie D, Dow JM. 2009.
The versatility and adaptation of bacteria from the genus
Stenotrophomonas. Nat Rev Microbiol 7:514–525.
74. Romanenko LA, Uchino M, Tanaka N, Frolova GM,
Slinkina NN, Mikhailov VV. 2008. Occurrence and
antagonistic potential of Stenotrophomonas strains iso-
lated from deep-sea invertebrates. Arch Microbiol 189:
337–344.
75. Flores MR, Ordoñez OF, Maldonado MJ, Farías ME.
2009. Isolation of UV-B resistant bacteria from two high
altitude Andean lakes (4,400 m) with saline and non saline
conditions. J Gen Appl Microbiol 55:447–458.
76. Matyar F, Kaya A, Dinçer S. 2008. Antibacterial agents
and heavy metal resistance in Gram-negative bacteria iso-
lated from seawater, shrimp and sediment in Iskenderun
Bay, Turkey. Sci Total Environ 407:279–285.
77. Brennan PJ, Nikaido H. 1995. The envelope of mycobac-
teria. Ann Rev Biochem 64:29–63.
78. Brooks RW, George KL, Parker BC, Falkinham JO,
Gruft H. 1984. Recovery and survival of nontuberculous
mycobacteria under various growth and decontamination
conditions. Can J Microbiol 30:1112–1117.
79. Thorel MF, Moreau R, Charvin M, Ebiou D. 1991.
Release of mycobacteria from natural environment by
enzymes. C R Séances Soc Biol Fil 185:331–337.
80. Strahl ED, Gillaspy GE, Falkinham JO. 2001. Fluores-
cent acid-fast microscopy for measuring phagocytosis of
3.3.2-10 ▪ SOIL
Mycobacterium avium, Mycobacterium intracellulare, and
Mycobacterium scrofulaceum by Tetrahymena pyriformis and
their intracellular growth. Appl Environ Microbiol 67:
4432–4439.
81. Cirillo JD, Falkow S, Tompkins LS, Bermudez LE. 1997.
Interaction of Mycobacterium avium with environmen-
tal amoebae enhances virulence. Infect Immun 65:
3759–3767.
82. Mazurek GH, Reddy V, Murphy D, Ansari T. 1996.
Detection of Mycobacterium tuberculosis in cerebrospinal
fluid following immunomagnetic enrichment. J Clin Micro-
biol 34:450–453.
83. Sweeney FP, Courtenay O, Hibberd V, Hewinson RG,
Reilly LA, Gaze WH, Wellington EMH. 2007. Environ-
mental monitoring of Mycobacterium bovis in badger feces
and badger sett soil by real-time PCR, as confirmed by
immunofluorescence , immunocapture, and cultivation.
Appl Environ Microbiol 73:7471–7473.
84. Voidarou C, Bezirtzoglou E, Alexopoulos A, Plessas S,
Stefanis C, Papadopoulos I, Vavias S, Stavropoulou E,
Fotou K, Tzora A, Skoufos I. 2011. Occurrence of Clostri-
dium perfringens from different cultivated soils. Anaerobe
17:320–324.
85. Lúquez C, Bianco MI, de Jong LIT, Sagua MD, Arenas
GN, Ciccarelli AS, Fernández RA. 2005. Distribution of
botulinum toxin-producing clostridia in soils of Argentina.
Appl Environ Microbiol 71:4137–4139.
86. Sanada I, Nishida S. 1965. Isolation of Clostridium tetani
from soil. J Bacteriol 89:626–629.
87. Dragon DC, Rennie RP, Elkin BT. 2001. Detection of
anthrax spores in endemic regions of northern Canada.
J Appl Microbiol 91:435–441.
88. Sauders BD, Overdevest J, Fortes E, Windham K,
Schukken Y, Lembo A, Wiedmann M. 2012. Diversity of
Listeria species in urban and natural environments. Appl
Environ Microbiol 78:4420–4433.
89. Hitchins AD. 2003. Detection and enumeration of Listeria
monocytogenes in foods. In U.S. Food and Drug Administra-
tion, Bacteriological Analytical Manual. Washington, DC.
90. USDA, Service FSaI. 2013. Isolation and identification of
Listeria monocytogenes from red meat, poultry, egg and envi-
ronmental samples. In USDA Food and Safety and Inspec-
tion Service, Microbiology Laboratory Guidebook, http://
www.fsis.usda.gov/wps/portal/fsis/topics/science/laboratories-
and-procedures/guidebooks-and-methods/microbiology-
laboratory-guidebook/microbiology-laboratory-guidebook.
91. Atlas R. 2010. Handbook of Microbiological Media, 4th ed.
ASM Press and CRC Press, Washington, DC.
92. Park SH, Ryu S, Kang DH. 2011. Improved selective and
differential medium for isolation of Escherichia coli O157:
H7. J Clin Microbiol 49:405–408.
93. Santamaria J, Toranzos GA. 2003. Enteric pathogens and
soil: a short review. Int Microbiol 6:5–9.
94. Zuberer DA. 1994. Recovery and enumeration of viable
bacteria, p 119–144. In Weaver RW (ed), Methods of Soil
Analysis, Part 2. Microbiological and Biochemical Properties.
Soil Society of America, Madison, WI.
95. Pepper IL, Gerba CP. 2009. Cultural methods, p 173–189.
In Maier RM, Pepper IL, Gerba CP (eds), Environmental
Microbiology, 2nd ed. Academic Press, San Diego, CA.
96. Archuleta RJ, Hoppes PY, Primm TP. 2005. Mycobacte-
rium avium, enters a state of metabolic dormancy in
response to starvation. Tuberculosis 85:147–158.
97. Kerr KG, Denton M, Todd N, Corps CM, Kumari P,
Hawkey PM. 1996. A new selective differential medium
for isolation of Stenotrophomonas maltophilia. Eur J Clin
Microbiol Infect Dis 15:607–610.
98. Pinot C, Deredjian A, Nazaret S, Brothier E, Cournoyer
B, Segonds C, Favre-Bonté S. 2011. Identification of
Stenotrophomonas maltophilia strains isolated from environ-
mental and clinical samples: a rapid and efficient proce-
dure. J Appl Microbiol 111:1185–1193.
99. Foddai A, Elliott CT, Grant IR. 2010. Maximizing capture
efficiency and specificity of magnetic separation for Myco-
bacterium avium subsp. paratuberculosis cells. Appl Environ
Microbiol 76:7550–7558.
100. Stewart LD, McNair J, McCallan L, Thompson S, Kula-
kov LA, Grant IR. 2012. Production and evaluation
of antibodies and phage display-derived peptide ligands
for immunomagnetic separation of Mycobacterium bovis.
J Clin Microbiol 50:1598–1605.
101. Radomski N, Lucas FS, Moilleron R, Cambau E, Haenn
S, Moulin L. 2010. Development of a real-time qPCR
method for detection and enumeration of Mycobacterium
spp. in surface water. Appl Environ Microbiol 76:
7348–7351.
102. Colinon C, Deredjian A, Hien E, Brothier E, Bouziri L,
Cournoyer B, Hartmann A, Henry S, Jolivet C, Ranjard
L, Nazaret S. 2013. Detection and enumeration of Pseudo-
monas aeruginosa in soil and manure assessed by an ecfX
qPCR assay. J Appl Microbiol 114:1734–1749.
103. Locatelli A, Depret G, Jolivet C, Henry S, Dequiedt S,
Piveteau P, Hartmann A. 2013. Nation-wide study of
the occurrence of Listeria monocytogenes in French soils
using culture-based and molecular detection methods.
J Microbiol Meth 93:242–250.
104. Ranjard L, Lejon DPH, Mougel C, Schehrer L, Merdino-
glu D, Chaussod R. 2003. Sampling strategy in molecular
microbial ecology: influence of soil sample size on DNA
fingerprinting analysis of fungal and bacterial communities.
Environ Microbiol 5:1111–1120.
105. Martin-Laurent F, Philippot L, Hallet S, Chaussod R,
Germon JC, Soulas G, Catroux G. 2001. DNA extraction
from soils: old bias for new microbial diversity analysis
methods. Appl Environ Microbiol 67:4397–4397.
106. Petric I, Philippot L, Abbate C, Bispo A, Chesnot T, Hal-
lin S, Laval K, Lebeau T, Lemanceau P, Leyval C, Lind-
strom K, Pandard P, Romero E, Sarr A, Schloter M,
Simonet P, Smalla K, Wilke BM, Martin-Laurent F.
2011. Inter-laboratory evaluation of the ISO standard
11063 “Soil quality—method to directly extract DNA
from soil samples.” J Microbiol Meth 84:454–460.
107. Feazel LM, Baumgartner LK, Peterson KL, Frank DN,
Harris JK, Pace NR. 2009. Opportunistic pathogens
enriched in showerhead biofilms. Proc Natl, Acad Sci
USA 106:16393–16398.
108. Wilton S, Cousins D. 1992. Detection and identification
of multiple mycobacterial pathogens by DNA amplifica-
tion in a single tube. PCR Meth Appl 1:269–273.
109. Wang H, Edwards M, Falkinham JO, Pruden A. 2012.
Molecular survey of the occurrence of Legionella spp., Myco-
bacterium spp., Pseudomonas aeruginosa, and amoeba hosts
in two chloraminated drinking water distribution systems.
Appl Environ Microbiol 78:6285–6294.
110. Tortoli E, Rindi L, Garcia MJ, Chiaradonna P, Dei R,
Garzelli C, Kroppenstedt RM, Lari N, Mattei R, Mario-
ttini A, Mazzarelli G, Murcia MI, Nanetti A, Piccoli P,
Scarparo C. 2004. Proposal to elevate the genetic variant
MAC-A, included in the Mycobacterium avium complex,
to species rank as Mycobacterium chimaera sp. nov.
Int J Syst Evol Microbiol 54:1277–1285.
111. Tortoli E. 2003. Impact of genotypic studies on mycobac-
terial taxonomy: the new mycobacteria of the 1990s. Clin
Microbiol Rev 16:319–354.
112. Kirschner RA, Parker BC, Falkinham JO. 1992. Epidemi-
ology of infection by nontuberculous mycobacteria—
Mycobacterium avium, Mycobacterium intracellulare, and
Mycobacterium scrofulaceum in acid, brown-water swamps
 3.3.2. Natural Soil Reservoirs for Human Pathogenic and Fecal Indicator Bacteria ▪ 3.3.2-11
of the Southeastern United-States and their association
with environmental variables. Am Rev Respir Dis 145:
271–275.
113. Iivanainen EK, Martikainen PJ, Raisanen ML, Katila
ML. 1997. Mycobacteria in boreal coniferous forest soils.
FEMS Microbiol Ecol 23:325–332.
114. Jadhav S, Bhave M, Palombo EA. 2012. Methods used for
the detection and subtyping of Listeria monocytogenes.
J Microbiol Meth 88:327–341.
115. Liles MR, Manske BF, Bintrim SB, Handelsman J, Good-
man RM. 2003. A census of rRNA genes and linked
genomic sequences within a soil metagenomic library.
Appl Environ Microbiol 69:2684–2691.
116. Venter JC, Remington K, Heidelberg JF, Halpern AL,
Rusch D, Eisen JA, Wu DY, Paulsen I, Nelson KE,
Nelson W, Fouts DE, Levy S, Knap AH, Lomas MW,
Nealson K, White O, Peterson J, Hoffman J, Parsons
R, Baden-Tillson H, Pfannkoch C, Rogers YH, Smith
HO. 2004. Environmental genome shotgun sequencing
of the Sargasso Sea. Science 304:66–74.
117. Lindstrom ES, Kamst-Van Agterveld MP, Zwart G. 2005.
Distribution of typical freshwater bacterial groups is associ-
ated with pH, temperature, and lake water retention time.
Appl Environ Microbiol 71:8201–8206.
118. Mueller-Spitz SR, Goetz GW, McLellan SL. 2009. Tem-
poral and spatial variability in nearshore bacterioplankton
communities of Lake Michigan. FEMS Microbiol Ecol 67:
511–522.
119. Schmidt TM, Delong EF, Pace NR. 1991. Analysis of
a marine picoplankton community by 16 s ribosomal-
RNA gene cloning and sequencing. J Bacteriol 173:
4371–4378.
120. von Wintzingerode F, Gobel UB, Stackebrandt E. 1997.
Determination of microbial diversity in environmental
samples: pitfalls of PCR-based rRNA analysis. FEMS
Microbiol Rev 21:213–229.
121. Brown MV, Philip GK, Bunge JA, Smith MC, Bissett A,
Lauro FM, Fuhrman JA, Donachie SP. 2009. Microbial
community structure in the North Pacific ocean. ISME J
3:1374–1386.
122. Buee M, Reich M, Murat C, Morin E, Nilsson RH, Uroz
S, Martin F. 2009. 454 Pyrosequencing analyses of forest
soils reveal an unexpectedly high fungal diversity. New
Phytol 184:449–456.
123. Zwart G, Crump BC, Agterveld MPKV, Hagen F, Han
SK. 2002. Typical freshwater bacteria: an analysis of avail-
able 16S rRNA gene sequences from plankton of lakes and
rivers. Aquat Microbiol Ecol 28:141–155.
124. Ishii S, Segawa T, Okabe S. 2013. Simultaneous quantifi-
cation of multiple food- and waterborne pathogens by use
of microfluidic quantitative PCR. Appl Environ Microbiol
79:2891–2898.
125. Luo CW, Walk ST, Gordon DM, Feldgarden M, Tiedje
JM, Konstantinidis KT. 2011. Genome sequencing of
environmental Escherichia coli expands understanding of
the ecology and speciation of the model bacterial species.
Proc Natl Acad Sci USA 108:7200–7205.
126. Ishii S, Ksoll WB, Hicks RE, Sadowsky MJ. 2006. Pres-
ence and growth of naturalized Escherichia coli in temperate
soils from lake superior watersheds. Appl Environ Microbiol
72:612–621.
127. Ishii S, Hansen DL, Hicks RE, Sadowsky MJ. 2007.
Beach sand and sediments are temporal sinks and sources
of Escherichia coli in lake superior. Environ Sci Technol 41:
2203–2209.
128. Byappanahalli MN, Whitman RL, Shively DA, Ferguson
J, Ishii S, Sadowsky MJ. 2007. Population structure of
Cladophora-borne Escherichia coli in nearshore water of
Lake Michigan. Water Res 41:3649–3654.
129. Badgley BD, Ferguson J, Vanden Heuvel A, Kleinheinz
GT, McDermott CM, Sandrin TR, Kinzelman J, Junion
EA, Byappanahalli MN, Whitman RL, Sadowsky MJ.
2011. Multi-scale temporal and spatial variation in geno-
typic composition of Cladophora-borne Escherichia coli pop-
ulations in Lake Michigan. Water Res 45:721–731.
130. Muller T, Ulrich A, Ott EM, Muller M. 2001. Identifica-
tion of plant-associated enterococci. J Appl Microbiol 91:
268–278.
131. Ott EM, Muller T, Muller M, Franz CMAP, Ulrich A,
Gabel M, Seyfarth W. 2001. Population dynamics and
antagonistic potential of enterococci colonizing the phyl-
losphere of grasses. J Appl Microbiol 91:54–66.
132. Hartz A, Cuvelier M, Nowosielski K, Bonilla TD, Green
M, Esiobu N, McCorquodale DS, Rogerson A. 2008.
Survival potential of Escherichia coli and enterococci in
subtropical beach sand: implications for water quality man-
agers. J Environ Qual 37:898–905.
133. Yamahara KM, Walters SP, Boehm AB. 2009. Growth of
enterococci in unaltered, unseeded beach sands subjected
to tidal wetting. Appl Environ Microbiol 75:1517–1524.
134. Byappanahalli MN, Sawdey R, Ishii S, Shively DA,
Ferguson JA, Whitman RL, Sadowsky MJ. 2009. Sea-
sonal stability of Cladophora-associated Salmonella in Lake
Michigan watersheds. Water Res 43:806–814.
135. Gubelit YI, Vainshtein MB. 2011. Growth of enterobacte-
ria on algal mats in the eastern part of the Gulf of Finland.
Inland Water Biol 4:132–136.
136. Baess I. 1974. Isolation and purification of desoxyribonu-
cleic acid from mycobacteria. Acta Pathol Microbiol Scand
B Microbiol Immunol 82:780–784.
137. Telenti A, Marchesi F, Balz M, Bally F, Bottger EC, Bod-
mer T. 1993. Rapid identification of mycobacteria to the
species level by polymerase chain reaction and restriction
enzyme analysis. J Clin Microbiol 31:175–178.
138. Steingrube VA, Gibson JL, Brown BA, Zhang YS,
Wilson RW, Rajagopalan M, Wallace RJ. 1995. PCR
amplification and restriction endonuclease analysis of a
65-kilodalton heat shock protein gene sequence for
taxonomic separation of rapidly growing mycobacteria.
J Clin Microbiol 33:1686–1686.
139. Arbeit RD, Slutsky A, Barber TW, Maslow JN, Niemc-
zyk S, Falkinham JO, Oconnor GT, Vonreyn CF. 1993.
Genetic diversity among strains of Mycobacterium avium
causing monoclonal and polyclonal bacteremia in patients
with AIDS. J Infect Dis 167:1384–1390.
140. van Soolingen D, Bauer J, Ritacco V, Leão SC, Pavlik I,
Vincent V, Rastogi N, Gori A, Bodmer T, Garzelli C,
Garcia MJ. 1998. IS1245 restriction fragment length poly-
morphism typing of Mycobacterium avium isolates: proposal
for standardization. J Clin Microbiol 36:3051–3054.
141. Cangelosi GA, Freeman RJ, Lewis KN, Livingston-
Rosanoff D, Shah KS, Milan SJ, Goldberg SV. 2004.
Evaluation of a high-throughput repetitive-sequence-based
PCR system for DNA fingerprinting of Mycobacterium
tuberculosis and Mycobacterium avium complex strains.
J Clin Microbiol 42:2685–2693.
142. Dauchy FA, Degrange S, Charron A, Dupon M, Xin Y,
Bebear C, Maugein J. 2010. Variable-number tandem-
repeat markers for typing Mycobacterium intracellulare
strains isolated in humans. BMC Microbiol 10:93.
143. Ichikawa K, Yagi T, Inagaki T, Moriyama M, Nakagawa
T, Uchiya K, Nikai T, Ogawa K. 2010. Molecular typing
of Mycobacterium intracellulare using multilocus variable-
number of tandem-repeat analysis: identification of loci and
analysis of clinical isolates. Microbiology 156:496–504.
144. Skuce RA, Brittain D, Hughes MS, Neill SD. 1996. Dif-
ferentiation of Mycobacterium bovis isolates from animals by
DNA typing. J Clin Microbiol 34:2469–2474.
3.3.2-12 ▪ SOIL
145. Collins DM, Cavaignac S, deLisle GW. 1997. Use of four
DNA insertion sequences to characterize strains of the
Mycobacterium avium complex isolated from animals. Mol
Cell Probes 11:373–380.
146. Nadon CA, Woodward DL, Young C, Rodgers EG,
Wiedmann M. 2001. Correlations between molecular sub-
typing and serotyping of Listeria monocytogenes. J Clin
Microbiol 39:2704–2707.
147. Ragon M, Wirth T, Hollandt F, Lavenir R, Lecuit M, Le
Monnier A, Brisse S. 2008. A new perspective on Listeria
monocytogenes evolution. PLoS Pathog 4:e1000146.
148. Chenal-Francisque V, Lopez J, Cantinelli T, Caro V,
Tran C, Leclercq A, Lecuit M, Brisse S. 2011. Worldwide
distribution of major clones of Listeria monocytogenes. Emerg
Infect Dis 17:1110–1112.
149. Ward TJ, Gorski L, Borucki MK, Mandrell RE,
Hutchins J, Pupedis K. 2004. Intraspecific phylogeny
and lineage group identification based on the prfA viru-
lence gene cluster of Listeria monocytogenes. J Bacteriol
186:4994–5002.
150. Wiedmann M, Bruce JL, Keating C, Johnson AE, McDo-
nough PL, Batt CA. 1997. Ribotypes and virulence gene
polymorphisms suggest three distinct Listeria monocytogenes
lineages with differences in pathogenic potential. Infect
Immun 65:2707–2716.
151. Gray MJ, Zadoks RN, Fortes ED, Dogan B, Cai S, Chen
YH, Scott VN, Gombas DE, Boor KJ, Wiedmann M.
2004. Listeria monocytogenes isolates from foods and
humans form distinct but overlapping populations. Appl
Environ Microbiol 70:5833–5841.
152. Gudmundsdottir KB, Aalbaek B, Sigurdarson S, Gun-
narsson E. 2004. The diversity of Listeria monocytogenes
strains from 10 Icelandic sheep farms. J Appl Microbiol 96:
913–921.
153. Nightingale KK, Schukken YH, Nightingale CR, Fortes
ED, Ho AJ, Her Z, Grohn YT, McDonough PL,
Wiedmann M. 2004. Ecology and transmission of Listeria
monocytogenes infecting ruminants and in the farm envi-
ronment. Appl Environ Microbiol 70:4458–4467.
154. Sauders BD, Durak MZ, Fortes E, Windham K,
Schukken Y, Lembo AJ Jr., Akey B, Nightingale KK,
Wiedmann M. 2006. Molecular characterization of Listeria
monocytogenes from natural and urban environments.
J Food Prot 69:93–105.
 The Evolving Science of Microbial Source Tracking
VALERIE J. HARWOOD, CHARLES HAGEDORN, AND MICHAEL SADOWSKY
3.4.1
A review of monitoring data compiled annually across the
United States clearly shows that levels of fecal indicator bac-
teria (FIB) in many water bodies exceed regulatory standards
(1). In eight of the past nine years, closings and advisories for
U.S. beaches have exceeded 20,000 annually, and over 80%
of these were triggered by elevated FIB (2). The central
hypothesis of the regulatory paradigm is that FIB levels that
exceed regulatory standards indicate elevated human health
risk (3), which can lead to substantial negative economic
and social consequences to local communities, for example,
local economic losses of up to $100,000 per day (4). Swim-
ming advisories also negatively impact public perceptions of
water safety and quality and those local, state, and federal
agencies responsible for protecting public health (5, 6).
In 2007, the U.S. Environmental Protection Agency
(EPA) sponsored a workshop on critical research needs
to inform development of new recreational water quality cri-
teria (7). An overarching short-term goal identified by the
experts was to “conduct research so that monitoring using
indicators can help to distinguish human from nonhuman
sources of fecal contamination.” Eight years later, microbial
source tracking (MST), or methods and approaches designed
to determine the dominant source(s) of fecal pollution in
water bodies, is an increasingly active research area in which
new methodologies are proposed, sometimes with bewilder-
ing frequency. Fig. 1 depicts the essential objective of MST.
An important use of MST methods is to improve estimates
of human health risk resulting from recreational water expo-
sure (8). Exposure to human fecal matter during recreational
water use carries known risk of contracting gastroenteritis
and other, potentially more severe diseases such as hepatitis
(7, 9). The EPA’s recreational water quality standards are
based on human health risk calculated from swimmers’ expo-
sure to fecal contamination that has some contribution from
domestic sewage (3), but are based on FIB levels, which pro-
vide no information about contamination source or the type
(s) of pathogens that may be present.
The use of FIB alone to assess water quality has several
important limitations (reviewed in [10]), including (a) the
nearly ubiquitous presence in feces prevents source identifi-
cation, which in turn (b) precludes accurate risk assessment
and (3) prevents effective remediation of contaminated
waters. Routine monitoring for waterborne human patho-
gens is currently impractical due to their low concentration,
doi:10.1128/9781555818821.ch3.4.1
3.4.1-1
the wide variety of pathogens that may be present, and the
high cost of laboratory analysis (10, 11). Possible sources of
fecal contamination and FIB include leaking sewer lines
and septic systems (12, 13), wild animals (14), animals
raised for food (15–18), and pets (19). Environmental res-
ervoirs include sand, soil, sediments, and aquatic vegetation
(20–22), and decaying algae and shoreline wrack (23, 24).
Urban development and impervious surface area (25) is
generally linked to elevated FIB levels in surface waters,
which can be contributed from stormwater (26) and agri-
cultural runoff (18) or dry-weather drains; however, the rel-
ative contributions of human sewage, animal fecal material,
and environmental reservoirs are largely “unknown” (1, 27,
28). Clearly, with the many different potential FIB sources
in coastal waters, effective water quality management
requires more information than FIB monitoring data alone
can provide.
Substantial efforts have been made over the past decades
to identify human sources of fecal contamination in recrea-
tional waters and reduce such sources of contamination, par-
ticularly in urban coastal environments (29). However,
numerous waters still experience health advisories due to
high FIB levels (2, 30). The fecal waste of animals other
than humans can contribute FIB and pathogens to water;
therefore, nonhuman fecal sources cannot be ignored as
potential contributors to human health risk (31, 32). MST
methods to detect fecal waste from human and other animal
sources have been developed, and many others are being
explored (reviewed in [10]).
The MST section of the Manual of Environmental Micro-
biology is intended to acquaint the reader with the various
types of MST methods and practices in current use, as well
as providing a means for judging the potential usefulness of
a given method for a practical application. Chapter 3.4.2
describes performance metrics such as sensitivity, specific-
ity, and limit of detection and how they can be used to
assess the strengths and weaknesses of specific methods.
Chapters 3.4.3 and 3.4.4 discuss MST methods that iden-
tify human and animal sources of pollution, respectively.
In Chapter 3.4.5, planning and implementation of field
studies is discussed, as well as considerations for the poten-
tial use of MST as a regulatory tool. Chapter 3.4.6 discusses
the ways modeling approaches can inform water quality and
MST investigations.
3.4.1-2 ▪ MICROBIAL SOURCE TRACKING
FIGURE 1 The essential objective of microbial source tracking
doi: 10.1128/9781555818821.ch3.4.1
MST METHODOLOGY
Regardless of the analytical method(s) chosen for an MST
study, an initial investigatory step should be to determine
the most probable contributors to contamination in a water-
shed (catchment). Many exploratory studies of a watershed
begin with a sanitary-survey type of approach, in which the
applicable watershed(s) is explored on foot and via aerial
mapping. Possible contamination sources can also be identi-
fied with input from local citizens and agencies (33). Subse-
quent analytical steps frequently use PCR or quantitative
PCR (qPCR) assays for genes in microorganisms that are asso-
ciated with feces of humans or other animals to detect and/or
quantify fecal sources.
Many MST studies are initiated with the ultimate goal of
improving water quality, which requires remedial action to
decrease pollution inputs from identified sources. Pollution
that is linked to human sources, for example, leaking sewers,
via human-associated assays must be resolved by infrastruc-
ture improvements. Pollution by nonhuman sources may be
indicated by lack of detection of genes (markers) from
human-associated microorganisms, and/or by detection of
markers associated with specific animal sources, for example,
cattle, poultry, or dogs. Mitigation of animal pollution sources
can be relatively simple, such as fencing cattle away from a
river (16), moving feedlot operations, or educating pet own-
ers about waste disposal. In other cases remedial steps may
require greater investment, for example, changing disposal
practices for land-applied manure (18). Fecal inputs from
wildlife may be particularly difficult to control, although
measures such as using dogs to chase gulls from beaches
have led to water quality improvements (34). Decisions about
reducing fecal contamination in water bodies contaminated
with nonhuman sources require balancing calculations of
human health risk from the sources identified, and the cost
of various possible solutions. In some circumstances, say,
when prioritizing remedial action for several watersheds or
water bodies, one may also take into account the use of the
water body (does primary contact recreation occur with any
frequency?) and/or potential impacts on ecosystem services
and health.
The MST studies carried out prior to 2005 frequently
relied on typing of E. coli or enterococci isolates from possible
fecal sources and creation of large databases of typed isolates
to which isolates from water were compared (e.g., 16, 17,
35, 36). The shared drawbacks of these “library-dependent”
methods eventually became apparent (37, 38) and included
the presence of cosmopolitan bacteria that were not host-
specific, greater genetic diversity among indicator bacteria
than was anticipated, and relatively small library and sample
sizes that did not adequately capture FIB diversity (35, 38–
40). In contrast, most current MST studies utilize methods
that are both library-independent and culture-independent.
Many of the methods rely on PCR or qPCR to amplify genes
belonging to microorganisms associated with a particular host
group (reviewed in [10]).
A frequently-cited criticism of PCR-based methods is that
they can detect nucleic acid in nonviable organisms. The
untested inference is that detection of nonviable microorgan-
isms compromises the effectiveness of the source identifier
(PCR target) for MST. This concern is not as pressing in sam-
ples that are concentrated by filtration compared to those that
are directly processed, as free nucleic acids will likely pass
through the filter. Furthermore, a lively topic of discussion
among microbiologists is the question of what constitutes a
“viable” bacterium, as injured or stressed bacteria may not
 3.4.1. The Evolving Science of Microbial Source Tracking ▪ 3.4.1-3
be readily culturable on standard media but may still be capa-
ble of infection, or growth under more permissive conditions
(41, 42). Thus, the ability to detect nonculturable bacteria
can be a positive attribute under many environmental condi-
tions. Another major consideration is the purpose of the
study. If it is simply to determine fecal contamination sources,
the viability of the source identifier may well be a moot point.
If the investigators plan to link MST markers to human
health outcomes or total maximum daily loads (TMDLs),
for example, in an epidemiology study or in case of remedial
action, viability may be an important consideration. To
date, little work has been done to test the hypothesis that
detection of nonviable bacteria in environmental waters by
qPCR confounds interpretation of the data. PCR methods
lie at one end of the viability spectrum, as they detect nucleic
acid without regard for cellular viability. Conversely, cultur-
ing on selective and/or differential media represents a stressful
environment in which cells that are nutrient-deprived or
under other environmental stress may not grow, even though
there is evidence that some species remain infectious (43).
Until very recently, culture methods have been required
worldwide for FIB enumeration; however, the EPA now
approves a qPCR method for enumeration of enterococci
(3). While PCR and culture approaches have biases that
must be accounted for when interpreting study results, both
techniques provide useful tools for environmental microbiol-
ogists and MST studies.
While some MST methods have shown great promise for
determining sources of fecal contamination in the environ-
ment, none meet all the characteristics of an ideal “source
identifier” (10, 11, 37). Each suffers to some extent in devia-
tions from ideal performance in terms of specificity for the tar-
get host type, or sensitivity to diluted fecal pollution.
Analytical concerns that apply to all PCR-based methods
include the potential for compounds in samples to inhibit
the PCR or recovery of DNA, lab-to-lab variation based on
the instrument used (44), and application of consistent crite-
ria for acceptable data (see Chapter 3.4.2). Researchers who
use these methods should adhere to the MIQE standards for
qPCR (45) for all MST assays, and publications should
include the correlation coefficient and efficiency of amplifi-
cation (E) for all standard curve. Tests for contamination
should include extraction blanks as well as no template con-
trols. Inhibition and the efficiency of DNA recovery should
be checked. Many researchers use salmon sperm DNA carried
through the nucleic acid extraction process to check nucleic
acid recovery (46), and this DNA may also serve as a “carrier”
to help bring low levels of environmental nucleic acid
through the process. Inhibition in the recovered nucleic
acid can be checked by spiking a known amount of engi-
neered internal amplification control or irrelevant target
(e.g., Vibrio DNA in a sample derived from freshwater) into
the PCR reaction tube.
The decision about how to report data influences the pre-
sentation of results and can impede comparisons among stud-
ies. Investigators in a recent multilab source identification
protocol project debated extensively about whether to express
qPCR results in terms of mass of template DNA or mass of
feces (44, 47). While the use of library-independent qPCR
coupled with source-specific primers has significantly
advanced the field of MST and allowed beach and river man-
agers to better devise remediation strategies and more cor-
rectly assess beach closures and health risks in near real
time, the technology does have limitations (10, 47).
The emergence of culture-independent tools such as
qPCR to determine microbial community structure and
sources, sinks, and relative proportions of fecal bacteria in
water bodies holds promise (48). One disadvantage of the
“marker” approach, where the target is usually a gene belong-
ing to a microorganism that is strongly host-associated, is that
markers do not yet exist for all animals whose waste may
impact a water body. Another strategy, the use of high-
throughput DNA sequence analysis, has proven useful for
identifying new markers (49, 50) and for MST analyses (51,
52). PCR assays can be developed for candidate markers, fol-
lowed by performance testing. Other approaches to marker
discovery such as terminal restriction fragment length poly-
morphisms (T-RFLPs) have also been employed (53).
The microbial community-based technologies being eval-
uated include T-RFLP analyses (54), microarrays (54, 55),
and high-throughput DNA sequencing analyses (52, 56).
These three methods, which were recently compared as part
of a large, multilaboratory study known as the Source Identi-
fication Protocol Project (SIPP) study (47, 57). The technol-
ogies theoretically have the ability to determine the presence
of fecal indicator bacteria as well as those occupying the intes-
tinal tract of key animals occupying a given habitat; however,
they differ on the depth of community structure that can be
resolved. Thus, in an approach that resembles that of the
library-dependent methods, these tools attempt to make cor-
respondence between the totality of microorganisms in the
GI tracts of animals inhabiting watersheds and those found
in water bodies. To avoid confusion with past terminology,
we refer to these as fecal DNA libraries.
One of the first microbial community techniques to be
assessed for MST was T-RFLP analysis (54, 58–60). Here
DNA extracted from an environmental sample or fecal source
is used as template for PCR, typically using 16S rRNA gene
primers. The PCR products are digested with restriction
enzymes, and since at least one of the primers is labeled
with a fluorescent dye, the result terminal-labeled restriction
fragments (TRFs) are separated by size and abundance using
DNA sequencers or methods like high-performance liquid
chromatography. Like other culture-independent, DNA
sequence-based approaches, this technology is not affected
by our inability to grow the vast majority of microorganisms
in environmental and animal fecal samples.
Microarray-based gene detection platforms may be geared
toward the phylogenetic composition of DNA in an environ-
mental sample, for example, the Phylochip (54, 57). Here,
the fundamental analytical strategy is shared with T-RFLP,
in that the goal is to assess the microbial community structure
in a water sample and match it to that found in fecal samples
using statistical approaches. Microarrays may also directly tar-
get FIB, MST markers, and pathogens (55). An advantage to
the FIB/pathogen microarray and microfluidic platforms (61)
is that specific oligonucleotides for genes to detect FIB, MST
source identifiers, and pathogens provide highly targeted data
for source and risk analysis.
Metagenomic approaches have been employed in MST
studies (50–52, 56, 62). Reduction of costs of DNA sequence
analysis and the simultaneous increase in the number of
sequence reads that can be obtained per run have driven
the sharp increase in usage of such approaches. Current tech-
nology limits DNA sequence length, and therefore complete
16S RNA sequences are not typically obtained using this
analysis. Thus, select regions of 16S rRNA, the V4, V5, or
V6 (or a combination thereof ) variable regions (63) are typ-
ically amplified by PCR prior to sequence analyses. Analysis
of the tens of thousands of sequences typically produced
with such technology is aided with tools such as the
web pyrosequencing-based interface PyroMiST, that allows
3.4.1-4 ▪ MICROBIAL SOURCE TRACKING
assignment on fecal inputs into waterways based on shared
operational taxonomic units (OTUs) between fecal and envi-
ronmental bacterial (51).
More recently, new sequencing technologies based on the
Illumina platforms, the Hiseq 2000 and the Miseq systems,
allows for very high-throughput DNA sequence analysis at
relatively low costs. For example, the HiSeq system can pro-
duce about 200 million paired-end reads (2 × 250 bp) per
lane in about 7–10 days, whereas the MiSeq system produces
5–12 million, 2 × 250 bp, paired-end reads overnight. The
MiSeq chemistry was recently enhanced to produce 2 ×
250 bp sequences, allowing for greater taxonomic resolution
of OTUs. The great sequence depth provided by either
technology lends itself to barcoded multiplexing, allowing
for the simultaneous analysis of a large number of samples
and thus reduced costs per sample. For aquatic systems about
250,000 sequences per sample are usually sufficient to
adequately describe microbial diversity. This technology
can be easily used to (a) define new primers that identify spe-
cific sources of fecal inputs into waterways (64), (b) examine
genetic diversity of microbial constituents in samples at a
depth that was previously unattainable, and (c) allowing
determination of sources of fecal contamination in waterways
by referencing to fecal source libraries. Recently, this technol-
ogy was used as part of the Minnesota Mississippi Metage-
nome Project (http://www.cbs.umn.edu/m3p) allowing
determination of input sources and the relationship between
species diversity and the river’s microbiota and land use
practices (65). Accurate analysis requires the generation of
large fecal DNA libraries. Like other library-dependent
methods, care must be taken to evaluate the relevance and
performance of libraries across geographic ranges. How-
ever, since the cost and time to generate new fecal DNA
libraries is relatively small, each region or watershed can
have its own source library constructed for MST purposes.
One of the largest drawbacks of this technology is the bioin-
formatics and computational and storage intensive analyses
of hundreds of millions of DNA sequences. Fortunately,
the Mothur (http://www.mothur.org/), Qime (http://qiime.
org/), and UniFrac (http://bmf.colorado.edu/unifrac/) soft-
ware packages have facilitated statistical analyses of these
very large data sets.
Given the rapid advances in gene sequencing technolo-
gies and rapidly decreasing costs, sequencing should become
increasingly affordable. More important, the sequences that
are produced will have fewer errors and will likely be much
longer, allowing for near full-length sequence analysis of
16S rDNAs. Furthermore, even greater sequence numbers
will be produced in less time. This will likely allow for (a)
DNA sequence analyses permitting quantification of all
microbial DNAs to the species level, rather than to the family
or genus level analyses permitted by current shorter DNA
sequences, (b) direct determination of pathogens in aquatic
systems by growth- and qPCR-independent means, and (c)
allow for better management of resources to prevent fecal
inputs into waterways. This future, however, is predicated
on the development of faster and less expensive computers
and the availability of bioinformatics tools needed to easily
analyze these massive data sets.
CURRENT STATUS AND FUTURE
DIRECTIONS IN MST
MST methods are continuously being added to the repertoire
or improved to provide more accurate and reliable tools to
identify sources of fecal contamination in water bodies.
Many methods were recently evaluated in the SIPP study,
where 27 laboratories compared methods for their ability to
identify sources of fecal pollution from blinded water samples
containing either one or two different fecal pollution sources
(47, 66–70). Results from the SIPP study demonstrated that
methods are currently available that can identify whether par-
ticular host sources, including humans, cows, and gulls, have
contributed to contamination in a body of water (68, 70).
Fecal source allocation to predict the percent contribution
of contamination from various pollution sources, which is
needed for TMDL applications, has proven to be more prob-
lematic and efforts to link source-associated markers with FIB
numbers have not yet been successful (48, 57). Although
additional research is needed to field validate and determine
the robustness of the SIPP results, it is clear that MST can be a
useful tool for resource managers tasked with bringing water
bodies into compliance with regulatory standards (44). Fur-
thermore, it is apparent that the discipline of MST will con-
tinue to evolve and improve over time, especially in areas
such as community analysis (57, 71). While the SIPP evalua-
tion of MST assays provided several accurate and sensitive
methods to aid water quality management, it also identified
important knowledge gaps that need to be addressed if these
methods are to be used in routine field implementation (70).
The knowledge gaps and recommendations to advance all of
watershed science include:
• standardize existing MST methodologies for consistent
and reliable application across multiple laboratories;
• determine appropriate performance criteria (e.g., specific-
ity and sensitivity) to be applied to new methods;
• improve concentration methods to improve detection of
dilute targets;
• determine the applicability of multiple, rapid MST assays
to broad geographic areas, assessing the impact of environ-
mental sample matrices on performance;
• better understand the relationships among MST markers,
FIB, and pathogens, including the influence of decay, per-
sistence, and transport in different aquatic systems;
• develop sensitive procedures to detect host-associated
pathogens such as viruses in water;
• capitalize on technological advances, including digital
droplet PCR, microfluidics PCR, microarray, high-
throughput DNA sequence analysis, and field deployable
detection systems.
Addressing the methodological and knowledge gaps
listed here will help inform site-specific water quality criteria
establishment with approaches such as quantitative microbial
risk assessment (QMRA) (72) and natural source exclusion
(http://ftp.sccwrp.org/pub/download/SOURCE_ID_WORK
SHOP/Session4.4_Gold_SMPier.pdf). Before new approaches
like natural source exclusion and QMRA can be imple-
mented with confidence, important research questions must
be addressed, including (a) how accurate are present techni-
ques for determining whether the fecal signature at a given
site is human or nonhuman in origin; (2) what are the relative
health risks associated with human and nonhuman fecal sour-
ces; and (3) what is the level of scientific uncertainty in using
this information in a management context (73, 74)? Such
approaches will likely face extensive questions, and may be
controversial when they are considered for implementation
in regulatory or legal contexts. Such consideration will only
be possible when the science is able to fully support their
feasibility in recreational waters (31).
 3.4.1. The Evolving Science of Microbial Source Tracking ▪ 3.4.1-5
Each month brings a suite of newly published MST
studies, and every year new methods are published, some
for hosts for which methods exist and others for new host
species. Thus, it is important to be current with the scien-
tific literature on new assays and their performance as
they emerge. It is also crucial to recognize caveats in this
area of research and practice, including (a) no single molec-
ular marker singularly accounts for FIB, human waste,
and pathogens; (b) most assays are insufficiently tested to
know their absolute specificity to target fecal material and
performance in different geographical regions; and (c) the
ecology, including persistence and degradation rates, of
FIB, pathogens, and MST markers is not well understood.
These knowledge gaps currently preclude definitive linkage
of MST marker quantities to far-upstream sources or predic-
tion of health consequences associated with their presence
(75). These factors support the contention that although
nucleic acid–based assays for assessing the presence of
human and other fecal sources are increasingly powerful
tools for tracking fecal sources that contribute to high
FIBs and human health risks, guidance is needed for
when and how to employ them and how best to interpret
the results. Such guidance is an important goal of the
following chapters in this section of the manual devoted
to MST.
REFERENCES
1. Dorfman M, Rosselot KS. 2010. Testing the Waters: A Guide
to Water Quality at Vacation Beaches. Natural Resources
Defense Council, Washington, DC.
2. Dorfman M, Haren A. 2013. Testing the Waters 2013: Exec-
utive Overview. Natural Resources Defense Council, Wash-
ington, DC.
3. USEPA. 2012. Recreational Water Quality Criteria. EPA:
820-F-12-058, U.S. Environmental Protection Agency—
Office of Water, Washington, DC.
4. NOAA. 2005. Welfare Estimates for Five Scenarios of Water
Quality Change in Southern California: A Report from the
Southern California Beach Valuation Project. National Oce-
anic and Atmospheric Administration, Washington, DC.
5. Allender-Hagedorn A. 2011. Public perception of and pub-
lic participation in microbial source tracking, p 283–300. In
Hagedorn C, Blanch A, Harwood VJ (eds.), Microbial Source
Tracking: Methods, Applications, and Case Studies. Springer-
Verlag, New York, NY.
6. Given S, Pendleton LH, Boehm AB. 2006. Regional public
health cost estimates of contaminated coastal waters: a case
study of gastroenteritis at southern California beaches. Envi-
ron Sci Technol 40:4851–4858.
7. U.S. EPA. 2007. Report of the experts scientific workshop
on critical research needs for the development of new or
revised recreational water quality criteria. EPA:823-R-
07-006. U.S. Environmental Protection Agency, Office of
Water, Washington, DC.
8. Staley C, Gordon KV, Schoen ME, Harwood VJ. 2012.
Performance of two human-associated microbial source
tracking qPCR methods in various Florida water types and
implications for microbial risk assessments. Appl Environ
Microbiol, 78:7317–7326.
9. Schoen ME, Soller JA, Ashbolt NJ. 2011. Evaluating the
importance of fecal sources in human-impacted waters.
Water Res 45:2670–2680.
10. Harwood VJ, Staley C, Badgley BD, Borges K, Korajkic A.
2014. Microbial source tracking markers for detection of
fecal contamination in environmental waters: relationships
to pathogens and human health outcomes. FEMS Microbiol
Rev 38:1–40.
11. U.S. EPA. 2005. Microbial source tracking guide docu-
ment. EPA:600-R-05-064. U.S. Environmental Protection
Agency, Office of Research and Development, Cincinnati,
OH.
12. Boehm AB, Fuhrman JA, MRSE RD, Grant SB. 2003.
Tiered approach for identification of a human fecal pollu-
tion source at a recreational beach: case study at Avalon
Cay, Catalina Island, California. Environ Sci Technol 37:
673–680.
13. Whitlock JE, Jones DT, Harwood VJ. 2002. Identification
of the sources of fecal coliforms in an urban watershed using
antibiotic resistance analysis. Water Res 36:4273–4282.
14. Grant SB, Sanders BF, Boehm AB, Redman JA, Kim JH,
Mrše R.D, Chu AK, Gouldin M, McGee CD, Gardiner
NA, Jones BH, Svejkovsky J, Leipzig GV, Brown A.
2001. Generation of Enterococci bacteria in a coastal salt-
water marsh and its impact on surf zone water quality. Envi-
ron Sci Technol 35:2407–2416.
15. Ahmed W, Sritharan T, Palmer A, Sidhu JP, Toze S. 2013.
Evaluation of bovine feces-associated microbial source track-
ing markers and their correlations with fecal indicators and
zoonotic pathogens in a Brisbane, Australia reservoir. Appl
Environ Microbiol 79:2682–2691.
16. Hagedorn C, Robinson SL, Filtz JR, Grubbs SM, Angier
TA, Beneau RB. 1999. Determining sources of fecal pollu-
tion in a rural Virginia watershed with antibiotic resistance
patterns in fecal streptococci. Appl Environ Microbiol 65:
5522–5531.
17. Harwood VJ, Whitlock J, Withington VH. 2000. Classifi-
cation of the antibiotic resistance patterns of indicator bac-
teria by discriminant analysis: use in predicting the source of
fecal contamination in subtropical Florida waters. Appl Envi-
ron Microbiol 66:3698–3704.
18. Weidhass JL, Macbeth TW, Olsen RL, Harwood VJ. 2011.
Correlation of quantitative PCR for a poultry-specific Brevi-
bacterium marker gene with bacterial and chemical indica-
tors of water pollution in a watershed impacted by land
application of poultry litter. Appl Environ Microbiol 77:
2094–2102.
19. Burnes BS. 2003. Antibiotic resistance analysis of fecal coli-
forms to determine fecal pollution sources in a mixed-use
watershed. Environ Monit Assess 85:87–98.
20. Badgley BD, Nayak BN, Harwood VJ. 2011. The impor-
tance of sediment and submerged aquatic vegetation as
potential habitats for persistent strains of enterococci in a sub-
tropical watershed. Water Res 44:5857–5866.
21. Byappanahali M, Nevers MN, Korajkic A, Staley ZS, Har-
wood VJ. 2012. Enterococci in the environment. Microbiol
Mol Rev 76:685–706.
22. Whitman RL, Harwood VJ, Edge TA, Nevers MB, Byap-
panahali M, Vijayavel K, Brandão J, Sadowsky MJ, Alm
EW, Crowe A, Ferguson D, Ge Z, Halliday E, Kinzelman
J, Kleinheinz G, Przybyla-Kelly K, Staley C, Staley Z,
Solo-Gabriele HM. 2014. Microbes in beach sands: integrat-
ing environment, ecology, and public health. Rev Environ Sci
Bio/Tech 13:329–368.
23. Weiskel PK, Howes BL, Heufelder GR. 1996. Coliform
contamination of a coastal embayment: sources and trans-
port pathways. Environ Sci Technol 30:1872–1881.
24. Whitman RL, Shively DA, Pawlik H, Nevers MB, Byap-
panahali MN. 2003. Occurrence of Escherichia coli and
Enterococci in Cladophora (Chlorphyta) in nearshore water
and beach sand of Lake Michigan. Appl Environ Microbiol
69:4714–4719.
25. Sauer EP, VandeWalle JL, Bootsma MJ, McLellan SL.
2011. Detection of the human specific Bacteroides
genetic marker provides evidence of widespread sewage
3.4.1-6 ▪ MICROBIAL SOURCE TRACKING
contamination of stormwater in the urban environment.
Water Res 45:4081–4091.
26. Mallin MA, Johnson VL, Ensign SH. 2009. Comparative
impacts of stormwater runoff on water quality of an urban,
a suburban, and a rural stream. Environ Monit Assess 159:
475–491.
27. Jiang S, Noble R, Chui WP. 2001. Human adenovirus and
coliphages in urban runoff-impacted coastal waters of South-
ern California. Appl Environ Microbiol 67:179–184.
28. Petersen TM, Rifai H, Suarez M, Stein A. 2005. Bacteria
loads from point and nonpoint sources in an urban water-
shed. J Environ Engin 131:1414–1425.
29. Dorsey JH. 2010. Improving water quality through Califor-
nia’s Clean Beach Initiative: an assessment of 17 projects.
Environ Monit Assess 166:95–111.
30. Stewart JR, Gast RJ, Fujioka RS, Solo-Gabriele HM,
Meschke JS, Amaral-Zettler LA, del Castillo E, Polz MF,
Collier TK, Strom MS, Sinigalliano CD, Moeller PDR,
Holland AF. 2008. The coastal environment and human
health: microbial indicators, pathogens, sentinels, and reser-
voirs. Environ Health 7(Suppl2):S3.
31. Soller JA, Schoen ME, Bartrand T, Ravenscroft JE,
Ashbolt NJ. 2010. Estimated human health risks from
exposure to recreational waters impacted by human and non-
human sources of faecal contamination. Water Res 44:
4674–4691.
32. U.S. EPA. 2009. Review of published studies to charac-
terize relative risks from different sources of fecal con-
tamination in recreational water. EPA:822-R-09-001, U.
S. Environmental Protection Agency, Office of Water,
Washington, DC.
33. Propst CW, Harwood VJ, Morrison G. 2011. Case studies
of urban and suburban watersheds. In Hagedorn C, Blanch
AR, Harwood VJ (eds.), Microbial Source Tracking: Methods,
Applications, and Case Studies. Springer Scientific,
New York, NY.
34. Converse RR, Kinzelman JL, Sams EA, Hudgens E,
Dufour AP, Ryu H, Santo-Domingo JW, Kelty CA,
Shanks OC, Siefring SD, Haugland RA, Wade TJ. 2012.
Dramatic improvements in beach water quality following
gull removal. Environ Sci Technol 46:206–213.
35. Wiggins BA, Cash PW, Creamer WS, Dart SE, Garcia
PP, Gerecke TM, Han J, Henry BL, Hoover KB, John-
son EL, Jones KC, McCarthy JG, McDonough JA,
Mercer SA, Noto MJ, Park H, Phillips MS, Purner
SM, Smith BM, Stevens EN, Varner AK. 2003. Use
of antibiotic resistance analysis for representativeness test-
ing of multiwatershed libraries. Appl Environ Microbiol 69:
3399–3405.
36. Dombek PE, Johnson LK, Zimmerley ST, Sadowsky MJ.
2000. Use of repetitive DNA sequences and the PCR to dif-
ferentiate Escherichia coli isolates from human and animal
sources. Appl Environ Microbiol 66:2572–2577.
37. Stoeckel DM, Harwood VJ. 2007. Performance, design, and
analysis in microbial source tracking studies. Appl Environ
Microbiol 73:2405–2415.
38. Stoeckel DM, Mathes MV, Hyer KE, Hagedorn C, Kator
H, Lukasik J, O’Brien TL, Fenger TW, Samadpour M,
Strickler KM, Wiggins BA. 2004. Comparison of seven pro-
tocols to identify fecal contamination sources using Escheri-
chia coli. Environ Sci Technol 38:6109–6117.
39. Griffith JF, Weisberg SB, McGee CD. 2003. Evalua-
tion of microbial source tracking methods using mixed
fecal sources in aqueous test samples. J Wat Health 1:
141–151.
40. Johnson LK, Brown MB, Carruthers EA, Ferguson JA,
Dombek PE, Sadowsky MJ. 2004. Sample size, library com-
position, and genotypic diversity among natural populations
of Escherichia coli from different animals influence accuracy of
determining sources of fecal pollution. Appl Environ Micro-
biol 70:4478–4485.
41. Heim S, Lleo MM, Bonato B, Guzman CA, Canepari P.
2002. The viable but nonculturable state and starva-
tion are different stress responses of Enterococcus faecalis, as
determined by proteome analysis. J Clin Microbiol 184:
6739–6745.
42. Kell DB, Kaprelyants AS, Weichart DH, Harwood CR,
Barer MR. 1998. Viability and activity in readily culturable
bacteria: a review and discussion of the practical issues. Anto-
nie Van Leeuwenhoek 73:169–187.
43. Oliver JD. 2010. Recent findings on the viable but noncul-
turable state in pathogenic bacteria. FEMS Microbiol Rev 34:
415–425.
44. Ebentier DL, Hanley KT, Cao Y, Badgley BD, Boehm
AB, Ervin J, Goodwin KD, Gourmelon M, Griffith
JF, Holden PA, Kelty CA, Lozach S, McGee C, Peed
L, Raith M, Ryu H, Sadowsky MJ, Scott E, Santo-
domingo JW, Schriewer A, Sinagalliano CD, Shanks
OC, Van De Werfhorst L, Wang D, Wuertz S, Jay J.
2013. Evaluation of the repeatability and reproducibility
of a suite of qPCR-based microbial source tracking meth-
ods. Water Res 47:6839–6848.
45. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J,
Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley GL,
Vandesompele J, Wittwer CT. 2009. The MIQE guidelines:
minimum information for publication of quantitative real-
time PCR experiments. Clin Chem 55:611–622.
46. Shanks OC, Peed L, Sivaganesan M, Haugland RA, Chern
EC. 2014. Human fecal source identification with real-time
quantitative PCR. Meth Mol Biol 1096:85–99.
47. Boehm AB, Van De Werfhorst L, Griffith JF, Holden P,
Jay J, Shanks OC, Wang D, Weisberg SB. 2013. Per-
formance of forty-one microbial source tracking methods:
a twenty-seven lab evaluation study. Water Res 47:
6812–6828.
48. Wang D, Farnleitner A, Field KG, Green HC, Shanks OC,
Boehm AB. 2013. Enterococcus and Escherichia coli fecal
source apportionment with microbial source tracking
genetic markers: is it feasible? Water Res 47:6849–6861.
49. McLellan SL, Eren AM. 2014. Discovering new indicators
of fecal pollution. Trends Microbiol 22:697–706.
50. McLellan SL, Huse SM, Mueller-Spitz S, Andreishcheva
E, Sogin ML. 2010. Diversity and population structure of
sewage-derived microorganisms in wastewater treatment
plant influent. Environ Microbiol 12:378–392.
51. Unno T, Di DY, Jang J, Suh YS, Sadowsky MJ, Hur HG.
2012. Integrated online system for a pyrosequencing-based
microbial source tracking method that targets Bacteroidetes
16S rDNA. Environ Sci Technol 46:93–98.
52. Unno T, Jang J, Han D, Kim JH, Sadowsky MJ, Kim O,
Chun J, Hur H. 2010. Use of barcoded pyrosequencing
and shared OTUs to determine sources of fecal bacteria in
watersheds. Environ Sci Technol 44:7777–7782.
53. Weidhass JL, Macbeth TW, Olsen RL, Sadowsky MJ,
Norat D, Harwood VJ. 2010. Identification of a Brevibacte-
rium marker gene specific to poultry litter and development
of a quantitative PCR assay. Appl Environ Microbiol 109:
334–347.
54. Cao Y, Wu CH, Andersen GL, Holden PA. 2011. Com-
munity analysis-based methods. p 251–281. In Hagedorn
C, Blanch AR, Harwood VJ (eds.), Microbial Source Track-
ing: Methods, Applications, and Case Studies. Springer Scien-
tific, New York, NY.
55. Li X, Harwood VJ, Nayak BN, Staley C, Sadowsky MJ,
Weidhass J. 2015. A novel microbial source tracking micro-
array for pathogen detection and fecal source identifica-
tion in environmental systems. Environ Sci Technol 49:
7319–7329.
 3.4.1. The Evolving Science of Microbial Source Tracking ▪ 3.4.1-7
56. Staley C, Gould TJ, Wang P, Phillips J, Cotner JB,
Sadowsky MJ. 2014. Bacterial community structure is
indicative of chemical inputs in the Upper Mississippi
River. Front Microbiol 5:524.
57. Cao Y, Van De Werfhorst LC, Dubinsky EA, Badgley BD,
Sadowsky MJ, Andersen GL, Griffith JF, Holden PA.
2013. Evaluation of molecular community analysis methods
for discerning fecal sources and human waste. Water Res 47:
6862–6872.
58. Kent AD, Smith DJ, Benson BJ, Triplett EW. 2003. Web-
based phylogenetic assignment tool for analysis of terminal
restriction fragment length polymorphism profiles of micro-
bial communities. Appl Environ Microbiol 69:6768–6776.
59. Shyu C, Soule T, Bent SJ, Foster JA, Forney LJ. 2007.
MiCA: a web-based tool for the analysis of microbial com-
munities based on terminal-restriction fragment length poly-
morphisms of 16S and 18S rRNA genes. Microb Ecol 53:
562–570.
60. Field KG, Chern EC, Dick LK, Fuhrman JA, Griffith JF,
Holden PA, LaMontagne MG, Le J, Olson B, Simonich
MT. 2003. A comparative study of culture-independent
library-independent genotypic methods of fecal source
tracking. Journal of Water Health 1:181–194.
61. Ishii S, Segawa T, Okabe S. 2013. Simultaneous quantifica-
tion of multiple food and waterborne pathogens by use of
microfluidic quantitative PCR. Appl Environ Microbiol 79:
2891–2898.
62. Lee JH, Yi H, Chun J. 2011. rRNASelector: a computer
program for selecting Ribosomal RNA encoding sequences
from metagenomic and metatranscriptomic shotgun libra-
ries. J Microbiol 49:689–691.
63. Baker GC, Smith JJ, Cowan DA. 2003. Review and
re-analysis of domain-specific 16S primers. J Microbiol Meth
55:541–555.
64. Shanks OC, Kelty CA, Archibeque S, Jenkins M, New-
ton RJ, McLellan SL, Huse SM, Sogin ML. 2011. Com-
munity structures of fecal bacteria in cattle from different
animal feeding operations. Appl Environ Microbiol 77:
2992–3001.
65. Staley ZR, Chase E, Mitraki C, Crisman TL, Harwood VJ.
2013. Microbial water quality in freshwater lakes with differ-
ent land uses. J Appl Microbiol 115:1240–1250.
66. Harwood VJ, Boehm AB, Sassoubre LM, Vijayavel K,
Stewart JR, Fong TT, Caprais MP, Converse RR, Diston
D, Ebdon J, Fuhrman JA, Gourmelon M, Gentry-Shields
J, Griffith JF, Kashian DR, Noble RT, Taylor H, Wicki
M. 2013. Performance of viruses and bacteriophages for fecal
source determination in a multi-laboratory comparative
study. Water Res 47:6929–6943.
67. Layton BA, Cao Y, Ebentier DL, Hanley K, Ballesté E,
Brandão J, Byappanahali M, Converse RR, Farnleitner
AH, Gentry-Shields J, Gidley ML, Gourmelon M, Lee
CS, Lee J, Lozach S, Madi T, Meijer WG, Noble R, Peed
L, Reischer GH, Rodrigues R, Rose JB, Schriewer A, Sini-
galliano C, Srinivivasan S, Stewart J, Van De Werfhorst
LC, Wang D, Whitman R, Wuertz S, Jay J, Holden PA,
Boehm AB, Shanks O, Griffith JF. 2013. Performance of
human fecal-associated PCR-based assays: an international
source identification method evaluation. Water Res 47:
6897–6908.
68. Schriewer A, Goodwin KD, Sinagalliano CD, Cox AM,
Wanless D, Bartkowiak J, Ebentier DL, Hanley KT, Ervin
J, Deering LA, Shanks OC, Peed LA, Meijer WG, Griffith
JF, Santo-Domingo JW, Jay JA, Holden PA, Wuertz S.
2013. Performance evaluation of canine-associated Bacteroi-
dales assays in a multi-laboratory comparison study. Water Res
47:6909–6920.
69. Sinigalliano CD, Ervin JS, Van De Werfhorst LC, Badg-
ley BD, Ballesté E, Bartkowiak J, Boehm AB, Byappana-
hali M, Goodwin KD, Gourmelon M, Griffith JF,
Holden PA, Jay J, Layton BA, Lee CS, Lee J, Meijer
WG, Noble R, Raith M, Ryu H, Sadowsky MJ,
Schriewer A, Wang D, Wanless D, Whitman R, Wuertz
S, Santo-Domingo JW. 2013. Multi-laboratory evaluations
of the performance of Catellicoccus marimammalium PCR
assays developed to target gull fecal sources. Water Res
47:6883–6896.
70. Stewart JR, Boehm AB, Dubinsky EA, Fong T, Goodwin
KD, Griffith JF, Noble RT, Shanks OC, Vijayavel K,
Weisberg SB. 2013. Recommendations following a multi-
laboratory comparison of microbial source tracking methods.
Water Res 47:6829–6838.
71. Dubinsky EA, Esmaili L, Hulls JR, Cao Y, Griffith JF,
Andersen GL. 2012. Application of phylogenetic microar-
rary analysis to discriminate sources of fecal pollution. Envi-
ron Sci Technol 46:4340–4347.
72. Schoen ME, Ashbolt NJ. 2010. Assessing pathogen risk to
swimmers at non-sewage impacted recreational beaches.
Environ Sci Technol 44:2286–2291.
73. USEPA. 2010. Quantitative microbial risk assessment to esti-
mate illness in freshwater impacted by agricultural animal sources
of fecal contamination. U.S. Environmental Protection
Agency, Office of Water, Washington, DC.
74. Ashbolt NJ, Schoen ME, Soller JA, Roser DJ. 2010. Pre-
dicting pathogen risks to aid beach management: the Real
value of quantitative microbial risk assessment (QMRA).
Water Res 44:4692–4703.
75. Girones R, Ferrús MA, Alonso JL, Rodriguez-Manzano J,
Calgua B, de Abreu Corrêa A, Hundesa A, Carratala A,
Bofill-Mas S. 2010. Molecular detection of pathogens in
water—the Pros and cons of molecular techniques. Water
Res 44:4325–4339.
 Validation of Microbial Source Tracking Markers
and Detection Protocols: Considerations for
Effective Interpretation
ASJA KORAJKIC, DON STOECKEL, AND JOHN F. GRIFFITH
3.4.2
This chapter examines the interpretation of quantitative and
qualitative data obtained for host-associated markers used to
evaluate sources of fecal contamination to the environment.
The term microbial source tracking (MST) has long been used
to describe these methods (1, 2), though the terms bacterial
source tracking (3) and fecal source tracking (4) are also used.
The purpose of this chapter is to describe characteristics of
various host-associated markers used for MST, the perform-
ance of detection protocols (starting with the sample process-
ing, followed by DNA extraction, and finally marker
detection) used to generate data, and effective interpretation
of results within the context of the described markers and
detection protocols used. Although targets other than
genetic markers have been applied to evaluate fecal contam-
ination sources, the focus of this chapter is genetic markers.
The concepts described can be readily applied to other tar-
gets, including cultivated host-associated bacteria and chem-
icals found in wastewater.
BACKGROUND
Overview of Method Validation
Two common considerations in method validation are the
desired level of resolution among fecal contamination sources
and the need for quantitative results. The end user’s desired
level of resolution of fecal contamination sources is a critical
parameter to define when evaluating the sensitivity and spe-
cificity of MST markers (see following section for details).
The level of resolution defines the types (host species
sampled) and extent (number and distribution) of reference
feces samples to analyze to calculate sensitivity and specific-
ity. Level of resolution also determines the number of detec-
tion protocols needed. For instance, if discrimination
between two host types is required, then the minimum num-
ber of detection protocols will be two—one for each host. On
the other hand, if detection of a single host is needed, then
one detection protocol, specific for that particular host, may
be sufficient. Similarly, when defining necessary validation
tests for detection protocol performance (e.g., limit of detec-
tion), the end user must determine if presence-absence data
are satisfactory, or whether quantitative data are required.
As described in the following sections, validation of both
marker characteristics and detection protocol performance
is essential for effective interpretation of MST results.
doi:10.1128/9781555818821.ch3.4.2
3.4.2-1
The specific approach taken to validate MST markers
and detection protocols should be dictated by study objectives
(5–7). For example, if the objective of the study is to evaluate
regulatory compliance then the selected validation tests
(i.e., limit of detection) must relate directly to regulatory fecal
indicator bacteria or related criteria. On the other hand, an
epidemiologic evaluation of the relation between marker
concentration and public health outcomes might be required
to validate interpretations in the context of public health out-
comes (risk of illness attributable to fecal contamination from
a source) (8). The context of the interpretation (e.g., regula-
tion of water quality versus public health implications) may
be linked or independent, depending on the user’s goals. In
each case, the validation tests must be tailored to meet the
specific requirements of the study.
Validation of MST Markers
Proposal of Novel Markers
Each proposed marker carries with it a reasonable expectation
for host specificity at some level of resolution (e.g., 9). This
expectation may come from prior work with cultivated organ-
isms (10), in silico evaluations that indicate that the marker
has host specificity (11), or the specific methodologies used
for marker discovery (e.g., subtractive methods in which
shared sequences are eliminated from consideration, leaving
only putative host-specific sequences) (12, 13). Frequently,
the proposed marker isn’t thoroughly characterized with
regard to host specificity when it is first introduced (14). In
some cases, only after substantial proof-of-concept utilization
has been completed is more comprehensive evaluation of host
specificity undertaken (15, 16). Extensive evaluation of host
specificity should be done as soon as practical to avoid selec-
tion of markers that are inappropriate to the experimental
setting, such as when a putatively human-specific marker is
found to cross-react with a locally abundant nontarget host
(17). Table 1 contains a list of selected MST markers, their
putative host range, and known cross-reactivities with feces
from nontarget hosts.
Characteristics of Markers
In cases where the full host range of a marker has not been
evaluated, the rationale provided for the expected host range
can offer some insight into the potential of the marker to
resolve hosts. For example, some microbial taxa of the
3.4.2-2 ▪ MICROBIAL SOURCE TRACKING

TABLE 1 Host range and known cross-reactivity of MST markers

Intended Method
Marker ID Target organism Target sequence Nontarget host range
target reference
Human HF 183 Bacteroides spp. 16S rDNA Opossum (17) chicken (11, (21) (end-point
16), dog and cat (11, 18, PCR), (11)
19), rainbow trout (20), (qPCR)
turkey (14), deer
HuBac Swine, dog, cattle (22), Nile (22) (qPCR)
tilapia, channel catfish,
rainbow trout, Atlantic
salmon (20)
BacHum-UCD Cow, dog, horse (23), (24) (qPCR)
rainbow trout (20)
BacH Cat (25), Nile tilapia, (25) (qPCR)
channel catfish, rainbow
trout (20)
HumanBac1 Pigs, cattle, dogs (26) (27) (qPCR)
BacHuman Pig, dog, cat (26) (26) (qPCR)
BuniF2 B. uniformis Elk, pig, sheep, chicken (28) (qPCR)
(16, 28)
BfragF1 B. fragilis Raccoon, sheep, chicken,
cat, dolphin, sea lion
(16, 28)
BvulgF1 B. vulgatus Sheep, chicken, dog, cat,
sea lion, elephant seal
(16, 28)
PcopriF1 Prevotella copri Beef cattle, pig, sheep, dog,
cat (16, 28)
BsteriF1 B. stercoris Dog, cat (16, 28)
BthetaF2 B. thetaiotamicron Pig, chicken, dog, cat (28)
HumM2 Bacteroides spp. Hypothetical protein Sheep, elk (29) (29) (qPCR),
(30) (end-
point PCR)
HumM3 Putative sigma factor Chicken (29) (29) (qPCR),
(30) (end-
point PCR)
B. thetaiotaomicron α-1,6-mannanase Dog (14) (14) (end-point
PCR), (31)
(qPCR)
B. fragilis gyrB Pig (26) (26) (qPCR)
esp Enterococcus faecium Enterococcus surface Dog, seagulls, horses, (34) (end-point),
protein (esp) sea lions, seals (32, 33) (35) (qPCR)
nifH Methanobrevibacter nifH Birds (36), cattle, dog (38) (end-point),
smithii seagull (37) (36) (qPCR)
HPyV Polyomaviruses JC T antigen None identified (39) (end-point),
and BK (18) (qPCR)
Mitochondria NADH dehydrogenase None identified (40) (end-point),
subunit 5 (41) (qPCR)
Mitochondria Cytochrome b None identified (42) (qPCR)
Pig Pig-1-Bac Bacteroidales 16S rDNA None identified (43) (qPCR)
Pig-2-Bac Bacteroidales 16S rDNA None identified
PAdV Adenovirus Viral genome None identified (44) (qPCR)
Cattle BacCow-UCD Bacteroidales 16S rDNA Horse (24) (24) (qPCR)
BoBac Bacteroidales 16S rDNA Dog (22), human (24) (22) (qPCR)
CowM2 Bacteroidales-like Protein involved in energy None identified (45) (end point),
metabolism and electron (46) (qPCR)
transport
CowM3 Bacteroidales-like Protein involved in None identified
degradation
of surface polysaccharides
and lipopolysaccharides
BacBovine Bacteroidales 16S rDNA Deer, dog (47) (47) (qPCR)
(Continued on next page)
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-3

TABLE 1 (Continued )
Intended Method
Marker ID Target organism Target sequence Nontarget host range
target reference
Dog BacCan-UCD Bacteroidales 16S rDNA Human, cat (24) (24) (qPCR)
Mitochondria NADH dehydrogenase None identified (41) (qPCR)
subunit 5
Gull Gull-2 Catellicoccus 16S rDNA Pelicans, other shorebirds (48) (qPCR)
marimmalium
Gull GFC Catellicocccus 16S rDNA Sheep (49) (49) (qPCR)
marimmalium
Bird GFD Helicobacter spp. 16S rDNA Gulls, geese, chickens,
ducks (49)
Canada CGOF1-Bac Bacteroidales 16S rDNA Pigeon (50) (50) (qPCR)
goose CGOF2-Bac Bacteroidales 16S rDNA None identified
Mitochondria NADH dehydrogenase None identified (41) (qPCR)
subunit 2
Poultry LA35 Brevibacterium avium 16S rDNA Goose, duck, sewage (51) (51) (qPCR)
Sandhill Crane 1 Lactobacillales 16S rDNA Geese, chicken and shore (52) (qPCR)
crane birds (52)
Muskrat MuBa01 Bacteroidales 16S rDNA None identified (53) (qPCR)
Cat Mitochondria NADH dehydrogenase None identified (41) (qPCR)
subunit 5
Deer Mitochondria Cytochrome b None identified (41) (qPCR),
(42) (qPCR)
Sheep Mitochondria Cytochrome b None identified (42) (qPCR)
Ruminant Rum2Bac Bacteroidales 16S rDNA Multiple (43) (qPCR)
General AllBac Bacteroidales 16S rDNA Multiple (22) (qPCR)
fecal
General GenBac3 Bacteroidales 16S rDNA Multiple (54) (qPCR)
fecal
The listing of nontarget host range is not intended to be a comprehensive evaluation of specificity, but rather to indicate where there is a known risk of cross
amplification.

gastrointestinal tract are thought to co-evolve with hosts
(55). Other microorganisms appear to be adapted to thrive
in hosts that share a particular set of chemical and physical
conditions in their digestive system, such as fowl, or to have
species-level host specificity (13).
Various regions of the microbial genome have been inves-
tigated for host specific genetic markers. A common target of
these markers is the variable region of DNA coding the 16S
ribosomal RNA (rDNA) (Table 1), which is usually present
in multiple copies per genome and has historically been used
for phylogenetic analysis (56). The relevance of this sequence
for host specificity stems from the hypothesis that some
fecal bacterial species are uniquely adapted to their hosts.
The usefulness of this target is enhanced by the large number
of readily available 16S rDNA sequences in collections
such as the Ribosomal Database Project (57–59). Sequences
that encode functional genes that participate in the host–
microbe interaction also are common targets (30, 31) and
an alternative to the phylogenetic (housekeeping) gene
sequences.
Although the emphasis of this chapter is on genetic
markers, not all markers are based on DNA sequences. Phe-
notypic (biochemical) characteristics associated with puta-
tive host-specific species have been used to selectively
cultivate these organisms (60). In addition to selective media,
carbon utilization profiles (61) and antibiotic resistance pro-
files (62) are examples of biochemical tests for microbial
characteristics that were previously thought to be informative
about microbial association with a particular host. Examina-
tion of phenotypic characteristics have historical significance
and were sometimes the basis for genetic marker discovery
efforts (such as the cultivation of host-associated Bacteroidales
[10]) prior to proposal of the Bacteroidales-based markers, but
have proven ineffective as microbial source identification
tools (9, 21, 63).
In addition to the aforementioned microbiological targets,
other nonmicrobial markers are sometimes used to evaluate
fecal contamination sources. These include host mitochon-
drial DNA shed with feces (42), diagnostic ratios of fecal
sterols (64), human wastewater-associated chemicals such as
caffeine (65), and optical brighteners (66, 67). Although
these targets have been used in field studies and may provide
valuable information for the toolbox approach to MST, the
scope of this chapter is limited to microbial targets that are
detected using molecular methods, in particular the genetic
markers.
Validation of Detection Protocols
In current usage, detection protocols for MST markers are
based primarily on PCR. PCR can be applied in either
presence-absence (end-point) or quantitative format. Other
molecular MST methods have also been used to discriminate
among fecal sources, including whole-community finger-
printing (68), phylogenetic microarrays (69), and sequencing
3.4.2-4 ▪ MICROBIAL SOURCE TRACKING
for community phylogenetic characterization (70). While
these whole-community methods may come into more
common use in the future as these technologies become
more accessible and costs decrease, this section focuses on
PCR-based detection because it is a mature technology in
wide use.
Basic quality measures for MST detection protocols are
based on well-defined quality control measurements, includ-
ing positive and negative controls and, in the case of quanti-
tative PCR (qPCR), standard curves generated from reference
material. The quality control parameters that can be derived
from these measurements include the limit of detection,
upper and lower limits of quantification, and extent of matrix
inhibition (if present). Limit of detection (LOD) of a detec-
tion protocol is a quantitative expression of the lowest con-
centration of target that can be detected. LOD is different
from and should not be confused with the limit of quantifica-
tion, which is defined as the range of concentration over
which the relation between log concentration and threshold
cycle of the quantitative PCR is linear. Each of these quality
measures may be used to compare detection protocols and
evaluate alterations intended to enhance performance.
Finally, disparate detection protocols have different sensi-
tivities to matrix inhibition of the PCR. Inhibition occurs
when substances present in a water sample interfere with
PCR amplification, leading to target underestimation, and
can be caused by a range of physical, biological, and chemical
mechanisms. Inhibitory substances commonly found in ambi-
ent water include humic and fulvic acids, carbohydrates, and
metal cations (71, 72). To alleviate or limit inhibition, many
qPCR protocols for MST begin with a DNA purification pro-
tocol to remove inhibitors, commonly using a commercial
DNA extraction kit. Although many of these protocols are
effective, it should be noted that the extra laboratory steps
associated with commercially available kits may add to meas-
urement error and require the addition of an extraction
control consisting of a known amount of nontarget DNA
to adjust for target DNA lost during extraction (73–76).
Other options for minimizing the effect of inhibition include
dilution of the inhibited sample (77) and use of internal con-
trols (46, 73, 78).
Ratios
The ratio of culturable fecal coliforms to fecal streptococci
(FC/FS) was first proposed as a way to distinguish human fecal
sources from other fecal contamination in the 1960s (79).
Higher ratios (4.0 and above) were thought to indicate
human fecal pollution, while lower ratios (0.7 and below)
were thought to indicate fecal pollution of nonhuman origin.
Intermediate values were believed to indicate a mixture of
the two sources (80). Subsequent research showed that (a)
the diagnostic FC/FS ratios are not consistent in human
and nonhuman fecal material (81) and (b) differential sur-
vival of the two groups of organisms in the environment com-
plicates the interpretation, since the FC/FS ratio tends to
change in the environment over time (82, 83). For these
reasons, many researchers abandoned this approach and are
hesitant to adopt other ratio-based approaches without thor-
ough evaluation for stable ratios and differential decay rates
between the target organisms.
With the advent of new MST technologies (i.e., qPCR)
the ratio approach to source apportionment was revisited.
Specifically, ratios of host-specific Bacteroidales marker con-
centrations to either E. coli densities (84) or general Bacteroi-
dales marker concentrations (85) have been evaluated. The
authors of each study concluded that the ratio approach
may be a viable tool for fecal source apportionment;
however, the clarity of interpretations may be affected by
the imperfect host specificity of most MST markers. Further-
more, like previous approaches using ratio-based methods,
this strategy is complicated by the potential for differential
decay of markers and fecal indicators in the environ-
ment. Hence, at the present time, approaches that utilize
MST markers for source allocation are still exploratory and
quantitative allocation of sources is not supported by most
current methodologies.
Method Comparison Studies
MST methods can be broadly divided into those that are
library-dependent (typically based on cultured isolates) and
library-independent (typically based on diagnostic DNA
sequences in target bacteria) (6). The former generally rely
on biochemical or antibiotic resistance based typing of iso-
lates, which are usually enterococci or E. coli, followed by
comparisons of patterns (fingerprints) of isolates from known
fecal sources to those from ambient water. Depending on the
fingerprint type, library-dependent methods can be classified
as genotypic (e.g., pulsed field gel electrophoresis, BOX-PCR,
ribotyping) or phenotypic (e.g., carbon source utilization
profiles, antibiotic resistance analyses). Earlier validation
attempts identified several important limitations of library-
dependent methods, including low correct classification rates
into appropriate source categories and high percentage of false
positives (i.e., sources not present were identified as being
present) (63, 86–88). Since then, mainstream methodology
has shifted to library-independent MST techniques, specifi-
cally those targeting host-associated microorganisms via
PCR. More recently, MST method development has focused
on the discovery of novel qPCR markers (18), the conversion
of existing end-point PCR assays into quantitative PCR for-
mat (29), and community analysis methods, such as terminal
restriction fragment length polymorphism, DNA microarrays,
and high-throughput DNA sequencing. While not yet
mature, community analysis methods have the potential to
allow identification of all sources of fecal contamination in
a water sample, including those for which no source-associ-
ated PCR markers have been developed (89).
The largest method comparison study to date included 27
participating laboratories evaluating the sensitivity and spe-
cificity of 41 different MST methods. The methods included
host-associated library independent markers using both end-
point and qPCR assays, as well as community based methods
(e.g., PhyloChip, terminal restriction fragment length poly-
morphism, and DNA sequencing) (90). Each participating
laboratory analyzed a common set of standardized, blinded
water samples containing a single fecal source, two sources,
or a diluted single source using their in-house protocols.
Results were analyzed in both binary (all methods treated as
presence/absence) and quantitative mode, and qPCR data
was normalized to enterococci culture-based results, as well
as mass of DNA and general Bacteroidales concentrations.
Sensitivity and specificity were influenced by the mode in
which results were analyzed (i.e., binary or quantitative)
while different normalization options (i.e., culturable entero-
cocci, DNA mass, or general Bacteroidales) did not appear to
influence the results (90).
TECHNIQUES FOR MST METHOD
VALIDATION
This section focuses on specific techniques used for validation
of marker specificity and sensitivity. These include detection
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-5
protocol LOD, detection protocol accuracy of quantification,
and other approaches used in methods comparison studies.
Validation of Marker Sensitivity and Specificity
Sensitivity and specificity are crucial performance criteria for
MST markers (91, 92). Specificity of a marker is a measure-
ment of false positive rate (specifically, 1 minus the false-
positive rate as defined in [91]) and is usually expressed as a
percentage. Markers that are found in the feces of nontarget
hosts have less than 100% specificity. Sensitivity, which is
also frequently expressed as a percentage, is 1 minus the fre-
quency of false negatives—and reflects the distribution of
the marker in target fecal samples. Markers that are not ubiq-
uitously distributed within a host population (or within a
given source of fecal material, such as domestic wastewater)
have less than 100% sensitivity. Both specificity and sensitiv-
ity testing can be carried out on individual fecal samples or
composite samples from a given source; the choice is often
governed by the goals and logistical constraints of individual
studies (reviewed in [6]). The amount and type of source
material used can influence the assessment of sensitivity
and specificity, as described below (see “Reporting” section).
Sensitivity and specificity of marker DNA as measured
by a proposed primer pair can be evaluated in silico prior
to undertaking laboratory measurements. This is accom-
plished by computer generated alignments (e.g., BLAST) of
the target sequence (marker specificity) or primer(s)/probe
(s) (detection protocol specificity) against established data-
bases (e.g., NCBI). While these theoretical tests are not a
substitute for controlled laboratory tests, they are a useful
screening tool and may indicate potentially problematic
cross-reactivity.
Sensitivity of a marker is empirically measured by testing a
collection of fecal samples from the target host, and specificity
is measured by testing a collection of fecal samples from var-
ious nontarget species. Formal guidance on the range of differ-
ent species and number of samples that should be collected is
lacking; however, it is desirable to test as many samples as pos-
sible considering the value of this information to interpreta-
tion of results. The level of effort required to conduct
specificity testing may be reduced by using pooled fecal mate-
rial (instead of individual samples); one caveat is that testing
pooled samples may mask cross-reactivity with a small num-
ber of nontarget species if the marker is present, but not in suf-
ficient concentrations to be detected (6).
The number of samples needed to accurately assess specif-
icity and sensitivity of a marker depends on the data quality
needs of the study. In theory, hypergeometric tables could
be used to calculate the minimum number of animals that
must be tested to reliably detect a marker in a nontarget pop-
ulation for specificity testing (93). For example, the hyper-
geometric tables specify that for a detection protocol with
100% analytical sensitivity (e.g., always detects when marker
is present), approximately 300 animals should be sampled to
be 95% certain of detecting the marker if it is present in 1% of
the nontarget population (e.g., specificity of the marker is
99%). In some recent studies, the number of composite sam-
ples collected to evaluate specificity and sensitivity generally
varied between 12 and 20 (24, 25, 28, 94), but some investi-
gators collected considerably more (i.e., 54 samples [16]).
Given that the distribution of a marker in host populations
can vary over space and time, it is important to establish
marker specificity within broad geographic regions and verify
expectations in a new region prior to the commencement of
the study.
Evaluation of Detection Protocol Performance
LOD
The analytical limit of detection (ALOD) for a given protocol
is the lowest amount of target that results in an observable
positive result. For end-point PCR protocols, the LOD often
is measured by dilution to extinction of a positive control
sample, for example, sewage or fecal material. In these sam-
ples, ALOD is reported as the dilution of sewage at which
the marker is no longer detectable, or the minimum mass of
fecal material that can be detected in a sample; the choice
is governed by the specific needs of the study. For qPCR,
the LOD can be estimated in a similar way but may also be
adjusted by the threshold cycle above which product is
detected with some frequency (e.g., 5%) in negative control
samples if false positives occur above the level where the pos-
itive control is consistently (e.g., 95%) detected (95).
Since sensitivity of a detection protocol is influenced by
physicochemical properties of a matrix (e.g., presence and
concentration of interfering substances such as humic and
tannic acids in ambient waters), the ALOD can be calculated
with and/or without the effects of environmental interferents.
The pure laboratory LOD of the detection protocol is meas-
ured by use of distilled water or buffered solution (e.g.,
phosphate-buffered saline or buffered water) as the diluent.
This type of LOD can also be referred to as method LOD
(MLOD) (96). The process LOD (PLOD) of the detection
protocol is measured by use of the matrix (ambient water)
as the diluent. The ALOD often is reported in terms related
to the analytical reaction (copy number per reaction). The
MLOD and PLOD, on the other hand, are reported in terms
related to the volume of water processed for analysis (e.g.,
copy number per liter).
LOD is typically not measured for every sample. Instead,
positive control samples are included in each analysis to con-
firm that the detection protocol continues to perform effec-
tively. The quantity and type of DNA utilized for positive
control samples is based on the data quality needs of the study
and the availability of suitable reference material (97). In
general, either biologically generated DNA (such as sewage
dilutions or genomic DNA extracted from a positive control
culture) or synthesized sequences (such as plasmids) are
spiked into the positive control reaction at a concentration
approximately 10-fold higher than the LOD (84). In cases
where matrix inhibition is suspected, a matrix spike sample
that includes positive control DNA can be analyzed in paral-
lel with the unspiked sample to ensure that a negative result
was not the result of matrix inhibition.
Limits of Quantification
Absolute quantification of marker copy number in environ-
mental samples can be challenging; thus, validation of quan-
tification should be undertaken. Previous research indicated
that one of the main factors impacting variability in recovery
efficiencies is inconsistent DNA recovery (98). To validate
the implicit assumption of full recovery, spike-and-recovery
controls (sometimes also referred to as sample processing con-
trols or exogenous positive controls) are included to measure
recovery efficiency. The positive control material spiked into
the sample goes through the whole protocol (e.g., process
starting with the sample concentration, followed by DNA
extraction and PCR amplification).
There are two general approaches to spike-and-recovery
controls. The first one involves addition of material (e.g.,
whole cells, plasmid, wild-type DNA extract containing the
target sequence) to the sample, while the second approach
3.4.2-6 ▪ MICROBIAL SOURCE TRACKING
calls for the addition of similar material that carries a surrogate
sequence instead of the target sequence. The latter is the more
efficient approach, since addition of target material requires
analysis of duplicate samples (one unseeded and one seeded
with the known concentration of the target) and carries
an added risk of cross-contamination. Addition of nontarget
surrogate eliminates the need for duplicate samples as it
allows for the simultaneous assessment of the spike-and-
recovery control and the target sequence ( providing that
the detection methods are compatible) and generally does
not pose a contamination risk.
Addition of nontarget surrogate as the spike-and-recovery
control carries some risks of increased error, since inherent
differences between the surrogate and the intended target
may complicate interpretation of results. To date, several
different surrogates have been published for use in MST. A
commonly used surrogate is salmon testes DNA (78, 94),
though other researchers use surrogates that more closely
resemble bacterial targets, such as whole cells of Pantoea
stewartii (99), Lactococcus lactis (100), or engineered plasmid
carried in E. coli cells (99).
While DNA extraction efficiency and PCR inhibition can
be assessed in parallel with a single spike, assessment of PCR
interferences (i.e., inhibition by the matrix components) is
generally completed in a separate step through internal ampli-
fication controls (IACs). IACs, also known as competitive
internal positive controls (28), are used to confirm that vari-
ous matrix components which may coconcentrate with the
sample and/or be a carried over from DNA extraction do
not inhibit the (q)PCR reaction. The IAC is different from
matrix spikes (which provide a binary measure of inhibition
as discussed in the section detailing LOD techniques) because
IACs are used to estimate reduction in amplification and
therefore inaccurate estimates of original marker concentra-
tion. In practical terms, IACs generally are modified versions
of the target (i.e., plasmid with target sequence insert) that
are amplified and detected with the same primers but a differ-
ent probe than the target. Because they closely resemble the
target, IACs are expected to have similar amplification
behavior and thus adequately measure inhibitory effects on
amplification.
When employing an IAC technique, inhibition is meas-
ured through a multiplex qPCR reaction that utilizes both
the probe for the target and a second probe for the IAC.
The IAC probe is labeled with a reporter molecule that fluo-
resces and is detected at a different wavelength than that of
the target probe, thus simultaneously recording the amplifica-
tion of the target and the IAC. Since the initial concentra-
tion of the IAC is controlled, inhibition can be measured
by detection of unrealistically low amounts of the IAC.
This approach has been used successfully to assess inhibition
in several different water types (96, 101). One of the caveats
of multiplex reactions is the potential for competition for
some of the reagents (e.g., deoxynucleoside triphosphates,
primers, and polymerase enzyme). A recent study offers
some insight with respect to delineation between amplifica-
tion interference (i.e., inhibition) and competition (102).
Briefly, in addition to the standard curve, this technique
requires repeated measurements of IAC of a known copy
number that can be reliably detected by the assay (e.g., 25
copies per reaction) (102). Amplification interference is
then defined as the IAC cycle threshold (Ct) values in the
actual samples that are greater than the mean of control
IACs plus 1.5 Cts (101, 102). Delineation between competi-
tion for reagents and inhibition is achieved by establishing
IAC range of quantification (range of standard curve in a
multiplex reaction where IAC concentrations are not sig-
nificantly different from each other) and finding the intersec-
tion between the standard curve and the upper limit of the
IAC range of quantification known as the competition
threshold (102). Samples that are above the competition
threshold are considered inhibited, whereas those that are
below are considered to be a product of competition (102).
Another way to assess potential inhibition that cir-
cumvents multiplex reactions and the issue of reagent compe-
tition is to dilute the sample with water, commonly performed
in 1:5 or 1:10 ratio. Since the theoretical amplification of
product in the PCR is twofold per cycle, a 1:10 dilution
should have a threshold cycle that is 3.3 lower than the undi-
luted sample. This approach can be very useful for detection
of inhibition, so long as the concentration of target in the
undiluted sample is sufficiently high that target remains
detectable in the dilution.
The purpose of the standard curve in qPCR analyses is
twofold. The first purpose is to allow comparative quantifica-
tion of the target concentrations in a sample. The second is as
a diagnostic tool to assess the continued performance of the
detection protocol, including equipment performance. Irre-
spective of the desired use, the same principles apply. Stand-
ard curves are constructed from plasmid DNA, synthesized
DNA, or genomic DNA extracts. While only three serial
10-fold dilutions typically are required to construct a standard
curve, a broader range (five to six dilutions) is preferred to
cover the range of concentrations that might be found in
the environment. It is critical that the range of dilutions
brackets the concentrations at which the target is likely to
be detected.
The output generated by the standard curve (coefficient
of determination or R 2 value, slope of the curve, and amplifi-
cation efficiency derived from slope) can inform the user of
the accuracy and precision of the detection protocol and
point to necessary adjustments or modifications. Coefficient
of determination is a measure of the linearity of the curve;
curves with R 2 values greater than 0.985 generally are consid-
ered to have acceptable quality (103). Lower values can indi-
cate either issues with the PCR chemistry—necessitating
optimization of the assay (e.g., of primer/probe concentra-
tions)—or saturation, which can be relieved by dilution of
the template. Amplification efficiency (AE) criteria are based
on the theoretical doubling of the PCR product with each
amplification cycle and can be derived from the slope of
the curve. Amplification efficiency is affected by variables
such as primer/probe design, reaction conditions, presence
of inhibitors, competition for reagents, and the like. For the
ideal PCR, in which product is doubled with each amplifica-
tion cycle, the slope of the standard curve will equal −3.33,
The AE for this slope can be calculated using the following
equation [% AE = (10−1/slope−1) × 100], resulting in 100%
calculated amplification efficiency. In practice, AE values
ranging from 80% to 110% (103) are generally considered
acceptable.
Interlab Performance Comparisons
Historic reports highlighted the need for interlaboratory per-
formance comparisons of MST detection protocols if they are
to mature into effective monitoring and regulatory tools.
Interlaboratory comparisons are needed to validate transfer-
ability of the detection protocol, including reproducibility of
results and sensitivity analysis to identify sources of variation.
A recent study investigating interlaboratory performance
of two qPCR detection protocols (in simplex and multi-
plex formats) indicated low variability (<10% coefficient of
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-7
variation) among eight participants, regardless of the format
(101). The authors suggested that the majority of the
observed variability was driven by analyst deviations from
the standard protocols, specifically incorrect implementation
of qPCR protocols. These results reinforce the need for stand-
ardization and extensive validation as highlighted in earlier
reports (6, 37).
Similar conclusions were reached by Ebentier et al., who
investigated intra- and interlaboratory variability (i.e., repeat-
ability and reproducibility) of several different human- and
cattle-associated MST markers. Despite use of standardized
protocols, some markers exhibited higher reproducibility
than others, a finding that was attributed to low concentra-
tion of marker in the source material in instances where repro-
ducibility was low (94). Another recent study evaluated
sensitivity and specificity of 41 different MST assays by 27
participants where each laboratory followed their laboratory-
specific detection protocols (90). While conclusions sug-
gest that despite differences in detection protocols some
approaches clearly outperformed others, it is noteworthy
that interpretation of the values close to or at the method
LOD varied widely among the participants. These findings
emphasized the need for common, standardized guidance
on data reporting.
Interpretation Paradigm
Validated MST detection protocols can be (and have been)
developed using the techniques described here. A crucial
next step in the assessment of MST method performance is
testing against known levels of target seeded into ambient
waters (96). At this step in the process, “target” should be
the local potential source of fecal contamination (i.e., sew-
age) rather than lab-grown pure cultures of the organism or
plasmids carrying the target gene. This step is imperative since
(a) host distribution of various markers is known to vary geo-
graphically and (b) it allows assessment of potential matrix
interference from local ambient waters. Rather than seeding
target at unrealistically high concentrations, investigators
are advised to prepare several dilutions of target. This conven-
tion sets realistic expectations and allows assessment of
the assay sensitivity in the particular locale of the study.
Provided successful analysis of these “fecal soup” samples,
the final recommended step of validation is to field test the
entire method at sites that are known (or highly suspected)
to be contaminated and contrast those results with relatively
unimpacted sites.
GENERAL GUIDELINES FOR EFFECTIVE
INTERPRETATION
A thorough understanding of limitations associated with both
method performance and characteristics of the selected
markers are needed to effectively employ MST markers and
interpret results accurately.
Limitations Based on Method Performance
LOD
The presence or absence of a marker is assumed to be indica-
tive of the presence or absence of feces from the associated
host. To meet this directive, it is important to ensure that
the false-negative and false-positive rates of the method are
understood. The LOD is measured to estimate the lowest
amount of marker that can be detected, and hence the asso-
ciated level of fecal contamination. This information, in turn,
is essential to answer the “so what” question; for example, so
what if contamination from a particular source is detected?
Does detection of the marker necessarily mean that an unac-
ceptable level of fecal contamination is present? Similarly,
does detection necessarily mean that the fecal contamina-
tion came from the indicated source, based on specificity?
In addition, so what if a source of fecal contamination is
not detected? Does that mean the fecal contamination defi-
nitely does not come from a particular source, based on the
sensitivity of the marker? Understanding the LOD in target
and nontarget sources of fecal contamination is essential
to making critical interpretations about marker presence or
absence in a sample.
Quantification
Quantification of MST markers is not trivial and poses one of
the greatest challenges to interpretation of MST results. A
lengthy sequence of calculations is required to estimate the
amount of a given MST marker in a water sample. Error
must be appropriately addressed at all levels of the calculation
to effectively interpret the indicated amount of fecal contam-
ination in the sample. Steps in this calculation include the
following.
Processing efficiency: Losses of marker may occur during
each step of processing. Specifically, volumetric losses of orig-
inal sample are commonly encountered in extraction steps,
but may be accounted and corrected for using proper controls
(6). Other losses can be more challenging to detect and cor-
rect, but are essential to consider because they change the
concentration of DNA relative to solution and can affect
the DNA mass balance calculation. These losses include
incomplete recovery of DNA from the original matrix during
concentration and incomplete recovery of DNA relative to
the carrier matrix during various steps of purification. Finally,
DNA can be removed from solution by adsorption to con-
tainer walls (104) and undergo degradation during storage
(105); these factors can be particularly important when
dealing with samples containing low concentrations of the
target. These losses can be quantified and corrected for by
including a spike-and-recovery control in the sample prior
to processing (78, 99); however, one control may not be
appropriate for all of the detection protocols employed in a
given study. The cost of controls can mount rapidly if some
concession to practicality is not made, for example, using a
“one-size-fits-all” control.
Quantification range: The standard curve for estimation of
marker concentration based on threshold cycle has upper and
lower limits of quantification. Samples above the upper limit
of quantification must be diluted to obtain credible results.
Samples below the lower limit of quantification can be qual-
itative and reported as detected, but not quantified, provided
there is an interval between the lower LOD and the lower
limit of quantification. Error in concentration estimation
can be estimated by several methodologies (106).
Detection efficiency: In some cases, marker presence can
be masked by the presence of PCR-inhibiting chemicals, or
marker concentration can be underestimated as a result of
amplification inefficiency. As discussed in previous sections,
positive-control DNA can be spiked into a matrix control
sample, or dilutions can be run to ensure that the change in
concentration is detected by the expected change in thresh-
old cycle value. Amplification efficiency can be evaluated
in one of two ways: the linear portion of each amplification
curve can be inspected to ensure that the marker shows
approximately twofold concentration increase after each
cycle, or the slope of the standard curve can be used to calcu-
late amplification efficiency (6).
3.4.2-8 ▪ MICROBIAL SOURCE TRACKING
Accurate quantification of marker concentration is possi-
ble only if the data are subjected to these quality control eval-
uations and only if mathematical corrections for losses and
inefficiencies can be performed. Interpretation of the marker
concentration, then, depends on evaluation of limitations
based on marker characteristics, in particular the relationship
between marker concentration and the required measure of
fecal contamination level (fecal mass, relevant indicator
organisms, or public health risk increase imposed by associ-
ated fecal-oral pathogens). These issues are addressed in the
next section.
Reporting
Here we pull together many of the issues that relate study
objectives to specific validation processes. The selection of
reporting units may seem trivial and is, in fact, covered in
other chapters dealing with study design and data interpreta-
tion. When applied to water samples, two sets of reporting
units are commonly used: marker copy number per volume
of water, and marker copy number per mass of DNA
extracted. The benefit of the former units (often reported
as marker/liter or /100 ml) is the use of environmentally
relevant units, in the sense that other regulatory measure-
ments (e.g., nutrients, fecal indicator bacteria) are also based
on concentration. These units are sometimes used in studies
that measure contamination upstream and downstream
from a suspected source, but interpretation of fecal content
in terms of fecal indicator bacteria is not possible without first
obtaining a validated conversion of marker copies to fecal
indicator bacteria in fecal source material (see next section).
Other validation steps that are required when reporting MST
marker concentration in these units are covered in the prior
section on quantification, which used these units as the
expected outcome.
Some researchers (i.e., 16, 29, 46, 90) prefer to report
marker concentration in terms of DNA mass extracted.
This strategy allows the researcher to bypass some of the con-
version factors required to relate MST marker content to vol-
ume of water extracted. These units essentially normalize the
level of fecal DNA containing the host-associated MST
marker to the background of environmental DNA. If one
assumes that enrichment with fecal contamination from a
particular source results in higher overall marker content,
these units allow calculation of relative contamination level
from a given fecal source. These interpretations may require
additional validation, in particular when comparing levels
of fecal contamination from one sample to another, the
implicit assumption is that the background DNA level (envi-
ronmental DNA) is essentially constant between the two.
For many applications, including total maximum daily
load and source apportionment, the most practical unit for
MST studies would relate the marker content to the fecal
indicator bacteria content. It is not practical to obtain these
measures in one test (e.g., measurement of marker copy num-
ber per unit volume does not give direct information about
the concentration of fecal indicator bacteria in colony-
forming units). Rather, the ratio of copy number to fecal indi-
cator bacteria in fecal samples must be obtained. The benefit
of utilizing these ratios is the ability to estimate the amount of
fecal contamination in units that are relevant in the regula-
tory context. Extensive validation is required to make these
interpretations (84), including measurement of error associ-
ated with each step of the calculations and incorporation
of uncertainty in the results. A further validation required
for this interpretation is to test the potential for differential
decay rates of marker and fecal indicator bacteria in the envi-
ronment (107).
A similar approach could be applied to measure marker
concentration in fecal source material in units of marker
copy number per mass DNA extracted. By use of parallel cal-
culations (mass DNA extracted per mass fecal material) it
would be possible to calculate the MST marker concentration
per mass fecal material, and thereby estimate fecal mass
loading to a water sample based on detected MST marker
concentration. These calculations are dependent on DNA
extraction efficiency; thus, validation would include exten-
sive evaluation and control of extraction efficiency. The use-
fulness of such calculations depends on the consistency of the
relationship between the organisms bearing the marker and
the bulk fecal flora, which is not understood for any marker
at this time.
The objectives of a study dictate in part the selection of
reporting units because reporting units are selected to allow
effective data interpretation. The steps required to obtain
data in the desired reporting units are clearly defined, and
validation steps to enhance interpretation were outlined in
this section.
Limitations Based on Marker Characteristics
Coverage
Ideally, an MST marker is found in all host populations in
similar concentrations and is temporally and geographically
stable (i.e., distributions do not change over time and in dif-
ferent regions). However, that is rarely the case, as many
markers show variable distribution in host populations and
may be affected by factors such as diet (15). For example, in
the case of cattle it was found that distribution of markers var-
ied considerably among herds based on whether the cattle
were fed grain, grass, or a processed feed (15). Hence prior
to starting a study, distribution of a given marker must be
ascertained by testing it against a wide range of target animals
in the study watershed.
Studies have shown that the vast majority of MST markers
have imperfect host specificity and cross-react to some degree
with nontarget hosts (Table 1). The range of nontarget hosts
and the concentrations of the target detected in these hosts
can both affect the performance of the marker. Prior to the
utilization of a marker in a new study area (or in the same
area after some time has elapsed), it is advisable that cross-
reactivity against other putative sources in the study area be
determined. For example, in the following simplified
hypothetical scenario, watershed A is suspected of being
impacted by improperly functioning septic tank systems as
well as surface runoff from a nearby dairy farm. In this
instance, establishing the potential for cross-reactivity of
human-associated MST marker(s) against cattle feces (and
vice versa) is of utmost importance. Furthermore, since distri-
bution of markers among hosts (intended and nontarget)
varies across different populations, both markers should be
tested against a range of other nontarget hosts (both wild
and domestic) to accurately characterize potential for cross-
reactivity.
Many MST markers are present in their intended hosts at
lower concentrations (as compared to fecal indicator bacteria
and other markers targeting the same host) or they can even
be absent from a segment of the population. This is especially
true for viral markers (108, 109). Hence, if one intends to use
these markers, which often are highly host-specific, the pop-
ulation affecting the source must be considered. For example,
septic tank systems contain waste from a significantly smaller
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-9
number of individuals than do municipal wastewater treat-
ment plants. Thus, some markers such as viruses, which are
only shed by infected individuals, are less likely to be found
in septic systems. Therefore, it is important to tailor the selec-
tion of an MST marker to the suspected source.
Stability
Temporal and geographic stability of a marker—in other
words, consistent presence in host populations in the study
region—are needed for successful application of any MST
marker. Hence prior to utilization of any new MST detection
protocols and absent data characterizing performance of the
marker in a particular locale, marker stability should be veri-
fied anew through extensive testing of the host populations in
the study area.
Previous studies have shown differential decay of various
MST markers, which may potentially limit their usefulness,
especially in instances where the marker may decay faster
than the fecal indicator bacteria and/or pathogens or is sub-
ject to different selective pressures. Specifically, environmen-
tal factors such as salinity, temperature, solar irradiation, and
predation are known to play important roles in marker decay.
Persistence of various Bacteroidales human-associated markers
was shown to be inversely proportional to temperature
and predation (10, 11, 27, 107, 110, 111), while elevated sal-
inities (i.e., estuarine and marine waters) apparently
extended persistence (27, 112). Studies on solar irradiation
were for the most part carried out under laboratory settings
or in closed systems (e.g., microcosms), neither one of which
can accurately depict complexity of the true environmental
conditions. Furthermore, these studies yielded conflicting
results (some investigators reported a negative effect, while
others did not find it to be an important factor) (27, 112–
118). Additional decay studies designed to closely mimic
ambient environmental conditions are needed for a more
accurate (and definitive) description of the effects of solar
irradiation.
Considering that both temperature and native microbiota
affect decay of MST markers, samples for MST analyses
should be treated in the same manner as samples for culture-
based analyses. To avoid degradation during storage, samples
should be kept on ice following collection and further proc-
essing (e.g., concentration on membrane filters, DNA extrac-
tion and [q]PCR) should be completed as soon as possible.
Following processing, concentrated samples should be stored
at −80°C to avoid further degradation of genetic material.
Note that storage of bacterial cells on filters at −80°C for
up to several months is a common practice that has not so
far been shown to result in DNA degradation.
SUMMARY
The goal of this chapter is to provide an overview of MST
marker characteristics, describe performance criteria of detec-
tion protocols used, and offer guidelines for the effective
interpretation of the results. Since the trend in the research
community has shifted toward (q)PCR detection of MST
markers targeting either a variable region of the 16S rDNA
or functional genes involved in host–microbe interactions,
the focus is on validation of these specific targets and proto-
cols used to detect them. The most basic performance criteria
applied to MST markers are based on the sensitivity and spe-
cificity characteristics, which are assessed by testing the
marker against fecal material from a broad range of target
and nontarget hosts. While some information may be gath-
ered through in silico, theoretical testing (e.g., hypergeometric
tables, NCBI/BLAST searches), empirical data are needed to
accurately assess performance of a given marker.
Limitations associated with any MST marker can be
broadly divided into limitations based on method perform-
ance and limitations based on marker characteristics. The for-
mer include LODs for a given protocol and quantification
issues, such as processing efficiency, range of detection, and
detection efficiency, as well as selection of reporting units.
It is important that LOD and range of quantification include
the dynamic range in which MST marker is expected to be
found. Furthermore, utilization of appropriate quality assur-
ance/quality control measurements will allow for calculation
of volumetric loss during processing and the assessment
of DNA extraction efficiency and PCR inhibition issues,
which are instrumental for accurate interpretation of results.
Limitations driven by marker characteristics include uneven
distribution among intended host populations, shared dis-
tribution among target and nontarget host species, and tem-
poral/geographic stability. In summary, exact requirements
for the performance validation and effective application of
MST markers will be largely defined by the needs of individ-
ual studies. Hence it is important to clearly define study goals
and expectations and select MST markers and validation
pathways that will meet the specific requirements of the study.
DISCLAIMER
The U.S. Environmental Protection Agency through its
Office of Research and Development funded and managed
or partially funded and collaborated in the research described
here. It has been subjected to Agency review and approved for
publication.
REFERENCES
1. Simpson JM, Santo Domingo JW, Reasoner DJ. 2002.
Microbial source tracking: state of the science. Environ Sci
Technol 36:5279–5288.
2. Samadpour M, Chechowitz N. 1995. Little Soos Creek
microbial source tracking—a survey. University of Wash-
ington, Department of Environmental Health.
3. Wiggins BA. 1996. Discriminant analysis of antibiotic
resistance patterns in fecal streptococci, a method to differ-
entiate human and animal sources of fecal pollution in nat-
ural waters. Appl Environ Microbiol 62:3997–4002.
4. Field KG, Samadpour M. 2007. Fecal source tracking, the
indicator paradigm, and managing water quality. Water Res
41:3517–3538.
5. Santo Domingo J, Edge T, Griffith J, Hansel J, Harwood
VJ, Jenkins M, Layton A, Molina M, Nakatsu C, Oshiro
R, Sadowsky M, Shanks O, Stelma G, Stewart J, Stoeckel
D, Wiggins B, Wilbur J. 2005. Microbial source tracking
guide document. EPA/600-R-05-064. U.S. EPA.
6. Harwood VJ, Stoeckel DM. 2011. Performance criteria, p
7–30. In Hagedorn C, Blanch AR, Harwood VJ (eds),
Microbial Source Tracking: Methods, Applications, and Case
Studies. Springer, New York.
7. Santo Domingo JW, Sadowsky MJ. 2007. Microbial Source
Tracking. ASM Press, Washington, DC.
8. Wade TJ, Sams E, Brenner KP, Haugland R, Chern E,
Beach M, Wymer L, Rankin CC, Love D, Li Q, Noble
R, Dufour AP. 2010. Rapidly measured indicators of rec-
reational water quality and swimming-associated illness at
marine beaches: a prospective cohort study. Environ Health
9:66.
9. Bernhard AE, Field KG. 2000. Identification of nonpoint
sources of fecal pollution in coastal waters by using
3.4.2-10 ▪ MICROBIAL SOURCE TRACKING
host-specific 16S ribosomal DNA genetic markers from
fecal anaerobes. Appl Environ Microbiol 66:1587–1594.
10. Kreader CA. 1998. Persistence of PCR-detectable Bacter-
oides distasonis from human feces in river water. Appl Envi-
ron Microbiol 64:4103–4105.
11. Seurinck S, Defoirdt T, Verstraete W, Siciliano SD.
2005. Detection and quantification of the human-specific
HF183 Bacteroides 16S rRNA genetic marker with real-
time PCR for assessment of human faecal pollution in fresh-
water. Environ Microbiol 7:249–259.
12. Shanks OC, Nietch C, Simonich M, Younger M, Rey-
nolds D, Field KG. 2006. Basin-wide analysis of the
dynamics of fecal contamination and fecal source identifi-
cation in Tillamook Bay, Oregon. Appl Environ Microbiol
72:5537–5546.
13. Lu J, Santo Domingo J, Shanks OC. 2007. Identification
of chicken-specific fecal microbial sequences using a meta-
genomic approach. Water Res 41:3561–3574.
14. Carson CA, Christiansen JM, Yampara-Iquise H, Ben-
son VW, Baffaut C, Davis JV, Broz RR, Kurtz WB, Rog-
ers WM, Fales WH. 2005. Specificity of a Bacteroides
thetaiotaomicron marker for human feces. Appl Environ
Microbiol 71:4945–4949.
15. Shanks OC, White K, Kelty CA, Hayes S, Sivaganesan
M, Jenkins M, Varma M, Haugland RA. 2010. Perform-
ance assessment PCR-based assays targeting Bacteroidales
genetic markers of bovine fecal pollution. Appl Environ
Microbiol 76:1359–1366.
16. Shanks OC, White K, Kelty CA, Sivaganesan M, Blan-
non J, Meckes M, Varma M, Haugland RA. 2010. Per-
formance of PCR-based assays targeting Bacteroidales
genetic markers of human fecal pollution in sewage and
fecal samples. Environ Sci Technol 44:6281–6288.
17. Kirs M, Harwood VJ, Fidler AE, Gillespie PA, Fyfe WR,
Blackwood AD, Cornelisen CD. 2011. Source tracking
faecal contamination in an urbanised and a rural waterway
in the Nelson-Tasman region, New Zealand. New Zealand
J Mar Freshw Res 45:43–58.
18. McQuaig SM, Scott TM, Lukasik JO, Paul JH, Harwood
VJ. 2009. Quantification of human polyomaviruses JC
Virus and BK Virus by TaqMan quantitative PCR and
comparison to other water quality indicators in water and
fecal samples. Appl Environ Microbiol 75:3379–3388.
19. Ahmed W, Yusuf R, Hasan I, Goonetilleke A, Gardner
T. 2010. Quantitative PCR assay of sewage-associated Bac-
teroides markers to assess sewage pollution in an urban lake
in Dhaka, Bangladesh. Can J Microbiol 56:838–845.
20. McLain JE, Ryu H, Kabiri-Badr L, Rock CM, Abbaszade-
gan M. 2009. Lack of specificity for PCR assays targeting
human Bacteroides 16S rRNA gene: cross-amplification
with fish feces. FEMS Microbiol Lett 299:38–43.
21. Bernhard AE, Field KG. 2000. A PCR assay to discrimi-
nate human and ruminant feces on the basis of host differ-
ences in Bacteroides-Prevotella genes encoding 16S rRNA.
Appl Environ Microbiol 66:4571–4574.
22. Layton A, McKay L, Williams D, Garrett V, Gentry R,
Sayler G. 2006. Development of Bacteroides 16S rRNA
gene TaqMan-based real-time PCR assays for estimation
of total, human, and bovine fecal pollution in water. Appl
Environ Microbiol 72:4214–4224.
23. Silkie SS, Nelson KL. 2009. Concentrations of host-
specific and generic fecal markers measured by quantitative
PCR in raw sewage and fresh animal feces. Water Res 43:
4860–4871.
24. Kildare BJ, Leutenegger CM, McSwain BS, Bambic DG,
Rajal VB, Wuertz S. 2007. 16S rRNA-based assays for
quantitative detection of universal, human-, cow-, and
dog-specific fecal Bacteroidales: a Bayesian approach. Water
Res 41:3701–3715.
25. Reischer GH, Kasper DC, Steinborn R, Farnleitner AH,
Mach RL. 2007. A quantitative real-time PCR assay for the
highly sensitive and specific detection of human faecal
influence in spring water from a large alpine catchment
area. Lett Appl Microbiol 44:351–356.
26. Lee CS, Lee J. 2010. Evaluation of new gyrB-based real-
time PCR system for the detection of B. fragilis as an indi-
cator of human-specific fecal contamination. J Microbiol
Meth 82:311–318.
27. Okabe S, Okayama N, Savichtcheva O, Ito T. 2007.
Quantification of host-specific Bacteroides-Prevotella 16S
rRNA genetic markers for assessment of fecal pollution in
freshwater. Appl Microbiol Biotechnol 74:890–901.
28. Haugland RA, Varma M, Sivaganesan M, Kelty C, Peed
L, Shanks OC. 2010. Evaluation of genetic markers from
the 16S rRNA gene V2 region for use in quantitative detec-
tion of selected Bacteroidales species and human fecal waste
by qPCR. Syst Appl Microbiol 33:348–357.
29. Shanks OC, Kelty CA, Sivaganesan M, Varma M, Haug-
land RA. 2009. Quantitative PCR for genetic markers of
human fecal pollution. Appl Environ Microbiol 75:
5507–5513.
30. Shanks OC, Domingo JW, Lu J, Kelty CA, Graham JE.
2007. Identification of bacterial DNA markers for the
detection of human fecal pollution in water. Appl Environ
Microbiol 73:2416–2422.
31. Yampara-Iquise H, Zheng G, Jones JE, Carson CA.
2008. Use of a Bacteroides thetaiotaomicron-specific
alpha-1-6, mannanase quantitative PCR to detect human
faecal pollution in water. J Appl Microbiol 105:1686–1693.
32. Layton BA, Walters SP, Boehm AB. 2009. Distribution
and diversity of the enterococcal surface protein (esp)
gene in animal hosts and the Pacific coast environment.
J Appl Microbiol 106:1521–1531.
33. Whitman RL, Przybyla-Kelly K, Shively DA, Byappana-
halli MN. 2007. Incidence of the enterococcal surface pro-
tein (esp) gene in human and animal fecal sources. Environ
Sci Technol 41:6090–6095.
34. Scott TM, Jenkins TM, Lukasik J, Rose JB. 2005. Poten-
tial use of a host associated molecular marker in Enterococ-
cus faecium as an index of human fecal pollution. Environ
Sci Technol 39:283–287.
35. Ahmed W, Stewart J, Gardner T, Powell D. 2008. A real-
time polymerase chain reaction assay for quantitative
detection of the human-specific enterococci surface pro-
tein marker in sewage and environmental waters. Environ
Microbiol 10:3255–3264.
36. Johnston C, Ufnar JA, Griffith JF, Gooch JA, Stewart
JR. 2010. A real-time qPCR assay for the detection of
the nifH gene of Methanobrevibacter smithii, a potential
indicator of sewage pollution. J Appl Microbiol 109:
1946–1956.
37. Harwood VJ, Brownell M, Wang S, Lepo J, Ellender RD,
Ajidahun A, Hellein KN, Kennedy E, Ye XY, Flood C.
2009. Validation and field testing of library-independent
microbial source tracking methods in the Gulf of Mexico.
Water Res 43:4812–4819.
38. Ufnar JA, Wang SY, Christiansen JM, Yampara-Iquise
H, Carson CA, Ellender RD. 2006. Detection of the
nifH gene of Methanobrevibacter smithii: a potential tool to
identify sewage pollution in recreational waters. J Appl
Microbiol 101:44–52.
39. McQuaig SM, Scott TM, Harwood VJ, Farrah SR,
Lukasik JO. 2006. Detection of human-derived fecal pol-
lution in environmental waters by use of a PCR-based
human polyomavirus assay. Appl Environ Microbiol 72:
7567–7574.
40. Martellini A, Payment P, Villemur R. 2005. Use of eukary-
otic mitochondrial DNA to differentiate human, bovine,
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-11
porcine and ovine sources in fecally contaminated surface
water. Water Res 39:541–548.
41. Caldwell JM, Levine JF. 2009. Domestic wastewater influ-
ent profiling using mitochondrial real-time PCR for source
tracking animal contamination. J Microbiol Meth 77:
17–22.
42. Schill WB, Mathes MV. 2008. Real-time PCR detection
and quantification of nine potential sources of fecal
contamination by analysis of mitochondrial cytochrome b
targets. Environ Sci Technol 42:5229–5234.
43. Mieszkin S, Furet JP, Corthier G, Gourmelon M. 2009.
Estimation of pig fecal contamination in a river catchment
by real-time PCR using two pig-specific Bacteroidales 16S
rRNA genetic markers. Appl Environ Microbiol 75:
3045–3054.
44. Hundesa A, de Motes CM, Albinana-Gimenez N,
Rodriguez-Manzano J, Bofill-Mas S, Sunen E, Girones
RR. 2009. Development of a qPCR assay for the quanti-
fication of porcine adenoviruses as an MST tool for swine
fecal contamination in the environment. J Virol Meth
158:130–135.
45. Shanks OC, Santo Domingo JW, Lamendella R, Kelty
CA, Graham JE. 2006. Competitive metagenomic DNA
hybridization identifies host-specific microbial genetic
markers in cow fecal samples. Appl Environ Microbiol 72:
4054–4060.
46. Shanks OC, Atikovic E, Blackwood AD, Lu J, Noble RT,
Domingo JS, Seifring S, Sivaganesan M, Haugland RA.
2008. Quantitative PCR for detection and enumeration
of genetic markers of bovine fecal pollution. Appl Environ
Microbiol 74:745–752.
47. Lee DY, Weir SC, Lee H, Trevors JT. 2010. Quantitative
identification of fecal water pollution sources by TaqMan
real-time PCR assays using Bacteroidales 16S rRNA genetic
markers. Appl Microbiol Biotechnol 88:1373–1383.
48. Lu JR, Santo Domingo JW, Lamendella R, Edge T, Hill S.
2008. Phylogenetic diversity and molecular detection
of bacteria in gull feces. Appl Environ Microbiol 74:
3969–3976.
49. Green HC, Dick LK, Gilpin B, Samadpour M, Field KG.
2011. Genetic markers for rapid PCR-based identifica-
tion of gull, Canada goose, duck, and chicken fecal
contamination in water. Appl Environ Microbiol 78:
503–510.
50. Fremaux B, Boa T, Yost CK. 2010. Quantitative real-time
PCR assays for sensitive detection of Canada goose-specific
fecal pollution in water sources. Appl Environ Microbiol 76:
4886–4889.
51. Weidhaas JL, Macbeth TW, Olsen RL, Sadowsky MJ,
Norat D, Harwood VJ. 2010. Identification of a Brevibac-
terium marker gene specific to poultry litter and develop-
ment of a quantitative PCR assay. J Appl Microbiol 109:
334–347.
52. Ryu H, Lu JR, Vogel J, Elk M, Chavez-Ramirez F, Ash-
bolt N, Domingo JS. 2012. Development and evaluation of
a quantitative PCR assay targeting sandhill crane (Grus
canadensis) fecal pollution. Appl Environ Microbiol 78:
4338–4345.
53. Marti R, Zhang Y, Lapen DR, Topp E. 2011. Develop-
ment and validation of a microbial source tracking marker
for the detection of fecal pollution by muskrats. J Microbiol
Methods 87:82–88.
54. Dick LK, Field KG. 2004. Rapid estimation of numbers of
fecal Bacteroidetes by use of a quantitative PCR assay for 16S
rRNA genes. Appl Environ Microbiol 70:5695–5697.
55. Buhnik-Rosenblau K, Matsko-Efimov V, Jung M, Shin
H, Danin-Poleg Y, Kashi Y. 2012. Indication for Co-
evolution of Lactobacillus johnsonii with its hosts. BMC
Microbiol 12:149.
56. Pace NR, Sapp J, Goldenfeld N. 2012. Phylogeny and
beyond: Scientific, historical, and conceptual significance
of the first tree of life. Proc Natl Acad SciUSA 109:
1011–1018.
57. Mara DD, Oragui JI. 1983. Sorbitol-fermenting bifidobac-
teria as specific indicators of human faecal pollution. J Appl
Bacteriol 55:349–357.
58. Rhodes MW, Kator H. 1999. Sorbitol-fermenting bifido-
bacteria as indicators of diffuse human faecal pollution in
estuarine watersheds. J Appl Microbiol 87:528–535.
59. Savill MG, Murray SR, Scholes P, Maas EW, McCormick
RE, Moore EB, Gilpin BJ. 2001. Application of polymer-
ase chain reaction (PCR) and TaqMan PCR techniques to
the detection and identification of Rhodococcus coprophilus
in faecal samples. J Microbiol Meth 47:355–368.
60. Bonjoch X, Balleste E, Blanch AR. 2005. Enumeration of
bifidobacterial populations with selective media to deter-
mine the source of waterborne fecal pollution. Water Res
39:1621–1627.
61. Moussa SH, Massengale RD. 2008. Identification of the
sources of Escherichia coli in a watershed using carbon-
utilization patterns and composite data sets. J Water Health
6:197–207.
62. Hagedorn C, Robinson SL, Filtz JR, Grubbs SM, Angier
TA, Reneau RB Jr. 1999. Determining sources of fecal
pollution in a rural Virginia watershed with antibiotic
resistance patterns in fecal streptococci. Appl Environ
Microbiol 65:5522–5531.
63. Griffith JF, Weisberg SB, McGee CD. 2003. Evaluation of
microbial source tracking methods using mixed fecal sour-
ces in aqueous test samples. J Water Health 1:141–151.
64. Leeming R, Ball A, Ashbolt N, Nichols PD. 1996. Using
faecal sterols from humans and animals to distinguish faecal
pollution in receiving waters. Water Res 30:2893–2900.
65. Gourmelon M, Caprais MP, Mieszkin S, Marti R, Wery
N, Jarde E, Derrien M, Jadas-Hecart A, Communal PY,
Jaffrezic A, Pourcher AM. 2010. Development of micro-
bial and chemical MST tools to identify the origin of the
faecal pollution in bathing and shellfish harvesting waters
in France. Water Res 44:4812–4824.
66. Hartel PG, Rodgers K, Moody GL, Hemmings SN,
Fisher JA, McDonald JL. 2008. Combining targeted sam-
pling and fluorometry to identify human fecal contamina-
tion in a freshwater creek. J Water Health 6:105–116.
67. Dickerson JW Jr, Hagedorn C, Hassall A. 2007. Detec-
tion and remediation of human-origin pollution at two
public beaches in Virginia using multiple source tracking
methods. Water Res 41:3758–3770.
68. Cao Y, Van De Werfhorst LC, Sercu B, Murray JL, Hol-
den PA. 2012. Application of an integrated community
analysis approach for microbial source tracking in a coastal
creek. Environ Sci Technol 45:7195–7201.
69. Wu CH, Sercu B, Van de Werfhorst LC, Wong J, DeSan-
tis TZ, Brodie EL, Hazen TC, Holden PA, Andersen GL.
2010. Characterization of coastal urban watershed bacterial
communities leads to alternative community-based indica-
tors. PLoS One 5:e11285.
70. Unno T, Jang J, Han D, Kim JH, Sadowsky MJ, Kim OS,
Chun J, Hur HG. 2010. Use of barcoded pyrosequencing
and shared OTUs to determine sources of fecal bacteria
in watersheds. Environ Sci Technol 44:7777–7782.
71. Gentry-Shields J, Wang A, Cory RM, Stewart JR. 2013.
Determination of specific types and relative levels of QPCR
inhibitors in environmental water samples using excita-
tion-emission matrix spectroscopy and PARAFAC. Water
Res 47:3467–3476.
72. Green HC, Field KG. 2012. Sensitive detection of sam-
ple interference in environmental qPCR. Water Res 46:
3251–3260.
3.4.2-12 ▪ MICROBIAL SOURCE TRACKING
73. Noble RT, Blackwood AD, Griffith JF, McGee CD,
Weisberg SB. 2010. Comparison of rapid quantitative
PCR-based and conventional culture-based methods for
enumeration of Enterococcus spp. and Escherichia coli in rec-
reational waters. Appl Environ Microbiol 76:7437–7443.
74. Haugland RA, Siefring S, Lavender J, Varma M. 2012.
Influences of sample interference and interference controls
on quantification of enterococci fecal indicator bacteria in
surface water samples by the qPCR method. Water Res 46:
5989–6001.
75. U.S. Environmental Protection Agency. 2010. Method B:
Bacteroidales in Water by TaqMan Quantitative Polymer-
ase Chain Reaction (qPCR) Assay. EPA-822-R-10-003.
76. U.S. Environmental Protection Agency. 2010. Method
A: Enterococci in Water by TaqMan Quantitative Poly-
merase Chain Reaction (qPCR) Assay. EPA-821-R-
10-004.
77. Halliday E, Griffith JF, Gast RJ. 2010. Use of an exoge-
nous plasmid standard and quantitative PCR to monitor
spatial and temporal distribution of Enterococcus spp. in
beach sands. Limnol Oceanog Meth/ASLO 8:146–154.
78. Haugland RA, Siefring SC, Wymer LJ, Brenner KP,
Dufour AP. 2005. Comparison of Enterococcus measure-
ments in freshwater at two recreational beaches by quanti-
tative polymerase chain reaction and membrane filter
culture analysis. Water Res 39:559–568.
79. Geldreich EE. 1969. Water pollution. Microbiol JWater Poll
Cont Fed 41:1053–1069.
80. Geldreich EE. 1976. Microbiology of water. J Water Poll
Cont Fed 48:1338–1356.
81. Pourcher AM, Devriese LA, Hernandez JF, Delattre JM.
1991. Enumeration by a miniaturized method of Escherichia
coli, Streptococcus bovis and enterococci as indicators of the
origin of faecal pollution of waters. J Appl Bacteriol 70:
525–530.
82. Mara DD, Oragui JI. 1981. Occurrence of Rhodococcus
coprophilus and associated actinomycetes in feces, sewage,
and freshwater. Appl Environ Microbiol 42:1037–1042.
83. McFeters GA, Bissonnette GK, Jezeski JJ, Thomson CA,
Stuart DG. 1974. Comparative survival of indicator bacte-
ria and enteric pathogens in well water. Appl Microbiol 27:
823–829.
84. Stoeckel DM, Stelzer EA, Stogner RW, Mau DP. 2011.
Semi-quantitative evaluation of fecal contamination
potential by human and ruminant sources using multiple
lines of evidence. Water Res 45:3225–3244.
85. Wang D, Silkie SS, Nelson KL, Wuertz S. 2010. Estimat-
ing true human and animal host source contribution in
quantitative microbial source tracking using the Monte
Carlo method. Water Res 44:4760–4675.
86. Myoda SP, Carson CA, Fuhrmann JJ, Hahm BK,
Hartel PG, Yampara-Lquise H, Johnson L, Kuntz RL,
Nakatsu CH, Sadowsky MJ, Samadpour M. 2003. Com-
parison of genotypic-based microbial source tracking meth-
ods requiring a host origin database. J Water Health 1:
167–180.
87. Stoeckel DM, Mathes MV, Hyer KE, Hagedorn C, Kator
H, Lukasik J, O’Brien TL, Fenger TW, Samadpour M,
Strickler KM, Wiggins BA. 2004. Comparison of seven
protocols to identify fecal contamination sources using
Escherichia coli. Environ Sci Technol 38:6109–6117.
88. Harwood VJ, Wiggins B, Hagedorn C, Ellender RD,
Gooch J, Kern J, Samadpour M, Chapman AC, Robinson
BJ, Thompson BC. 2003. Phenotypic library-based micro-
bial source tracking methods: efficacy in the California col-
laborative study. J Water Health 1:153–166.
89. Rigsbee L, Agans R, Foy BD, Paliy O. 2011. Optimizing
the analysis of human intestinal microbiota with phyloge-
netic microarray. FEMS Microbiol Ecol 75:332–342.
90. Boehm AB, Werfhorst LCVD, Griffith JF, Holden PA,
Jay JA, Shanks OC, Wang D, Weisberg SB. 2013.
Performance of forty-one microbial source tracking meth-
ods: a twenty-seven lab evaluation study. Water Res 47:
6812–6828.
91. Stoeckel DM, Harwood VJ. 2007. Performance, design,
and analysis in microbial source tracking studies. Appl Envi-
ron Microbiol 73:2405–2415.
92. Harwood VJ, Staley C, Badgley BD, Borges K, Korajkic
A. 2014. Microbial source tracking markers for detection of
fecal contamination in environmental waters: relationship
to pathogens and human health outcomes. FEMS Microbiol
Rev 38:1–40.
93. Johnson WO, Su CL, Gardner IA, Christensen R. 2004.
Sample size calculations for surveys to substantiate freedom
of populations from infectious agents. Biometrics 60:
165–171.
94. Ebentier DL, Hanley KT, Cao Y, Badgley BD, Boehm
AB, Ervin JS, Goodwin KD, Gourmelon M, Griffith
JF, Holden PA, Kelty CA, Lozach S, McGee C, Peed
LA, Raith M, Ryu H, Sadowsky MJ, Scott EA, Santo
Domingo J, Sinigalliano CD, Shanks OC, Werfhorst
LCVD, Wang D, Wuertz S, Jay JA. 2012. Evaluation of
the repeatability and reproducibility of a suite of qOCR-
based microbial source tracking methods. Water Res 47:
6839–6848.
95. Harwood VJ, Gordon KV, Staley C. 2011. Validation
of rapid methods for enumeration of markers for human
sewage contamination in recreational water. Water Envi-
ronment Research Foundation, Alexandria, VA,
PATH3C09.
96. Staley C, Gordon KV, Schoen ME, Harwood VJ. 2012.
Performance of two quantitative PCR methods for micro-
bial source tracking of human sewage and implications for
microbial risk assessment in recreational waters. Appl Envi-
ron Microbiol 78:7317–7326.
97. Cao YP, Sivaganesan M, Kinzelman J, Blackwood AD,
Noble RT, Haugland RA, Griffith JF, Weisberg SB.
2013. Effect of platform, reference material, and quantifica-
tion model on enumeration of Enterococcus by quantita-
tive PCR methods. Water Res 47:233–241.
98. Stewart JR, Gast RJ, Fujioka RS, Solo-Gabriele HM,
Meschke JS, Amaral-Zettler LA, Del Castillo E, Polz
MF, Collier TK, Strom MS, Sinigalliano CD, Moeller
PD, Holland AF. 2008. The coastal environment and
human health: microbial indicators, pathogens, sentinels
and reservoirs. Environ Health 7(Suppl 2):S3.
99. Stoeckel DM, Stelzer EA, Dick LK. 2009. Evaluation of
two spike-and-recovery controls for assessment of extrac-
tion efficiency in microbial source tracking studies. Water
Res 43:4820–4827.
100. Siefring S, Varma M, Atikovic E, Wymer L, Haugland
RA. 2008. Improved real-time PCR assays for the detec-
tion of fecal indicator bacteria in surface waters with differ-
ent instrument and reagent systems. J Water Health 6:
225–237.
101. Shanks OC, Sivaganesan M, Peed L, Kelty CA, Black-
wood AD, Greene MR, Noble RT, Bushon RN, Stelzer
EA, Kinzelman J, Anan’eva T, Sinigalliano C, Wanless
D, Griffith J, Cao Y, Weisberg S, Harwood VJ, Staley
C, Oshima KH, Varma M, Haugland RA. 2012. Interla-
boratory comparison of real-time PCR protocols for quan-
tification of general fecal indicator bacteria. Environ Sci
Technol 46:945–953.
102. Kelty CA, Varma M, Sivaganesan M, Haugland
RA, Shanks OC. 2012. Distribution of genetic mar-
ker concentrations for fecal indicator bacteria in sewage
and animal feces. Appl Environ Microbiol 78:4225–
4232.
 3.4.2. Validation of Microbial Source Tracking Markers and Detection Protocols ▪ 3.4.2-13
103. Sigma Aldrich. qPCR Technical Guide. http://www.sig
maaldrich.com/content/dam/sigma-aldrich/docs/Sigma/
General_Information/qpcr_technical_guide.pdf.
104. Teo IA, Choi JW, Morlese J, Taylor G, Shaunak S. 2002.
LightCycler qPCR optimisation for low copy number target
DNA. J Immunol Meth 270:119–133.
105. Lee SB, Crouse CA, Kline MC. 2010. Optimizing storage
and handling of DNA extracts. Forensic Sci Rev 22:
131–144.
106. Sivaganesan M, Seifring S, Varma M, Haugland RA,
Shanks OC. 2008. A Bayesian method for calculating real-
time quantitative PCR calibration curves using absolute
plasmid DNA standards. BMC Bioinformatics 9:120.
107. Dick LK, Stelzer EA, Bertke EE, Fong DL, Stoeckel
DM. 2010. Relative decay of Bacteroidales microbial
source tracking markers and cultivated Escherichia coli
in freshwater microcosms. Appl Environ Microbiol 76:
3255–3262.
108. Noble RT, Allen SM, Blackwood AD, Chu W, Jiang SC,
Lovelace GL, Sobsey MD, Stewart JR, Wait DA. 2003.
Use of viral pathogens and indicators to differentiate
between human and non-human fecal contamination in
a microbial source tracking comparison study. J Water
Health 1:195–207.
109. Harwood VJ, Boehm AB, Sassoubre LM, Vijayavel K,
Stewart JR, Fong TT, Caprais MP, Converse RR,
Diston D, Ebdon JE, Fuhrman JA, Gourmelon M,
Gentry-Shields J, Griffith JF, Kashian D, Noble RT,
Taylor H, Wicki M. 2013. Performance of viruses and
bacteriophages for fecal source determination in a multi-
laboratory, comparative study. Water Res 47:6929–6943.
110. Balleste E, Bonjoch X, Belanche LA, Blanch AR. 2010.
Molecular indicators used in the development of predictive
models for microbial source tracking. Appl Environ Micro-
biol 76:1789–1795.
111. Korajkic A, McMinn BR, Shanks OC, Sivaganesan M,
Fout GS, Ashbolt NJ. 2014. Biotic interactions and
sunlight affect persistence of fecal indicator bacteria and
microbial source tracking genetic markers in the upper Mis-
sissippi river. Appl Environ Microbiol 80:3952–3961.
112. Green HC, Shanks OC, Sivaganesan M, Haugland RA,
Field KG. 2011. Differential decay of human fecal Bacter-
oides in marine and freshwater. Environ Microbiol 13:
3235–3249.
113. Bae S, Wuertz S. 2009. Rapid decay of host-specific fecal
Bacteroidales cells in seawater as measured by quantitative
PCR with propidium monoazide. Water Res 43:4850–4859.
114. Bae S, Wuertz S. 2012. Survival of host-associated bacter-
oidales cells and their relationship with Enterococcus spp.,
Campylobacter jejuni, Salmonella enterica serovar Typhimu-
rium, and adenovirus in freshwater microcosms as measured
by propidium monoazide-quantitative PCR. Appl Environ
Microbiol 78:922–932.
115. Sokolova E, Astrom J, Pettersson TJ, Bergstedt O, Her-
mansson M. 2012. Decay of Bacteroidales genetic markers
in relation to traditional fecal indicators for water quality
modeling of drinking water sources. Environ Sci Technol
46:892–900.
116. Walters SP, Yamahara KM, Boehm AB. 2009. Persistence
of nucleic acid markers of health-relevant organisms in sea-
water microcosms: implications for their use in assessing
risk in recreational waters. Water Res 43:4929–4939.
117. Korajkic A, Wanjugi P, Harwood VJ. 2013. Indigenous
microbiota and habitat influence Escherichia coli survival
more than sunlight in simulated aquatic environments.
Appl Environ Microbiol 79:5329–5337.
118. Korajkic A, McMinn BR, Harwood VJ, Shanks OC, Fout
GS, Ashbolt NJ. 2013. Differential decay of enterococci
and Escherichia coli originating from two fecal pollution
sources. Appl Environ Microbiol 79:2488–2492.
 Overview of Microbial Source Tracking Methods Targeting
Human Fecal Pollution Sources
ORIN C. SHANKS, HYATT GREEN, ASJA KORAJKIC, AND
KATHARINE G. FIELD
3.4.3
Exposure to human fecal waste can be a public health risk due
to the presence of human pathogens. Human fecal pollu-
tion can be introduced into water resources from damaged
sewer lines, faulty septic systems, combined sewer overflows,
illicit dumping activities, and even recreational bathers.
Ensuring public safety and adequate water quality therefore
requires methods that can confirm the presence or absence
as well as the concentration of human fecal pollution sources.
This chapter provides an overview of human-associated fecal
source identification methods commonly used to assess water
quality. In addition to describing various methods, this chap-
ter also discusses factors to consider for method selection.
BRIEF HISTORY
Like many other research fields, the advent of PCR allowed
microbial source tracking (MST) methodologies to shift focus
from bacterial cultivation approaches to direct isolation of
DNA from environmental samples. Early method develop-
ment studies utilized cultivation of bacteria to characterize
physiological attributes such as carbon utilization (1, 2) or
antimicrobial resistance (3–6), while others combined culti-
vation with PCR or other molecular technologies such as
repetitive sequence-based PCR (7), ribotyping (5), and pulse
field gel electrophoresis (8) to identify human fecal pollution
sources. In addition to relying on cultivation of bacteria iso-
lated from environmental samples, these methods required
an a priori data set or library compiled from different animal
fecal sources to interpret results. However, the utility of these
methods was questioned when these physiological or genetic
traits were shown to not be exclusive to human-associated
bacteria (9) and methods performed poorly in comparative
studies (10, 11).
As a result, human-associated MST method develop-
ment efforts shifted from library-based approaches to strat-
egies that are not restricted to easily cultivated bacteria and
large reference data sets. Direct PCR amplification of DNA
from bacterial phyla such as Bacteroidales, Bifodobacterium,
and Lachnospriaceae isolated from environmental water sam-
ples are now among the most popular approaches (12). Alter-
native gene targets for PCR, such as viral and human
mitochondrial nucleic acid, have also shown great promise
due to their reliance on fecal bacterial or host cells for repli-
cation. Additionally, because PCR-based method data inter-
pretation can be as simple as identifying the presence or
doi:10.1128/9781555818821.ch3.4.3
3.4.3-1
absence of PCR products on an agarose gel compared to
statistical analysis of complex patterns among library data
sets, PCR methods are considered more user-friendly. Last,
the use of quantitative real-time PCR (qPCR) has rapidly
increased in recent years and offers the ability to estimate con-
centrations of a human-associated MST gene target.
New and emerging technologies, such as microarrays and
next-generation sequencing, may play important roles in
future human-associated MST applications (13). These
community-based methods can test hundreds to millions of
bacteria in each environmental sample and may provide
greater confidence in human fecal pollution source identifica-
tion. The development and validation of these methods will
receive much attention in coming years.
Validation Efforts
Method validation refers to a series of studies designed to
ensure that a method meets the requirements for MST
application (see Chapter 4.1.2). Common considerations in
human-associated MST method validation include sensitiv-
ity, specificity, and reproducibility. Sensitivity refers to a
true positive rate (identification of human fecal sources),
and specificity describes the true negative rate (identification
of waste from nonhuman animal groups). Reproducibility refers
to the degree of agreement between findings conducted on
replicate samples in different laboratories. In general, human-
associated MST methods that consistently identify human
fecal material from a broad geographic range of populations
(high sensitivity) and that seldom elicit false identifications
from nonhuman animal groups (high specificity) with proto-
cols that are easy to replicate across laboratories with a low
level of interlaboratory variability (good reproducibility) are
desirable.
Earlier human-associated MST method validation studies
identified several limitations to many culture-based MST
methods, most notably poor specificity and high levels of
interlaboratory variability (10). Collectively, these studies
indicated that these methods were insufficient to correctly
identify human fecal sources and signaled a movement of the
mainstream methodology toward gene-specific PCR-based
methods. Hence recent human-associated MST method val-
idation study efforts have focused on PCR-based methods,
such as those described in this chapter. Based on a single lab-
oratory validation study designed to explore the specificity
and sensitivity of human-associated PCR-based methods, it
3.4.3-2 ▪ MICROBIAL SOURCE TRACKING

was found that many technologies are highly specific with
human-associated gene targets broadly distributed across the
United States (14). A multiple-laboratory study found that
when protocols are standardized, variability among partici-
pants was low, suggesting that human-associated PCR-based
methods can be highly reproducible (15). The largest multi-
ple laboratory validation study to date was recently completed
in southern California, and results confirmed previous find-
ings, as well as highlighting the need for standardized guid-
ance on data analysis and interpretation (16).
METHOD OVERVIEWS
Many methods are available that are reported to be closely
associated with human fecal pollution using a variety of differ-
ent technologies. The following section describes some of the
most popular methods that employ PCR, bacterial commun-
ity characterization, and chemical methods.
PCR-Based Methods
PCR is a biochemical technology developed to amplify a few
copies of a DNA fragment, producing many identical nucleic
acid molecules from a small amount of starting material (tem-
plate). The method relies on an iterative thermal cycling
process, consisting of repeated heated and cooling steps to
achieve DNA synthesis. Short fragments of DNA referred
to as oligonucleotides containing complementary sequences
to a region of DNA along with a DNA polymerase are key
components that enable highly selective and reproducible
amplification of a given template. Products are usually visual-
ized on a gel, and the results are scored as presence or absence
of the target sequence. qPCR methods follow the same gen-
eral principle as PCR, but the accumulation of target DNA
sequences is measured after each thermal cycle, allowing for
simultaneous amplification and quantification. PCR and
qPCR methods for detection of human waste use custom-
designed oligonucleotides to selectively target regions of
DNA that are associated with human fecal pollution sources.
These target DNA sequences typically originate from fecal
bacteria, human viruses, and human mitochondria.
Bacterial Targets
Human intestinal bacteria represent a diverse and abun-
dant source of candidate human-associated DNA targets for
PCR-based applications. A growing body of evidence suggests
that human intestinal bacteria are significantly different from
extraintestinal organisms (17). Some phyla exhibit strong
evolutionary associations with their host (18), which suggests
the existence of human-associated bacterial lineages and
coevolution of the host and the intestinal microbiome. Based
on recent validation studies (14, 16, 19), certain members of
the order Bacteroidales have been shown to be closely associ-
ated with human fecal pollution sources (Table 1). Methods
targeting the 16S rRNA gene of these strictly anaerobic and
numerically abundant members make up the majority of these
approaches. However, nonribosomal genes directly involved
in bacterium–host interactions from Bacteroidales have also
been shown to be strongly associated with human intestinal
communities (20–22).
Viral Targets
Utilization of viral targets in the MST field is an attractive
alternative because of the high specificity of virus–host inter-
actions. Methods used to characterize viral targets rely on an
array of technologies ranging from culture-based methods to
different PCR approaches depending on the viral target
(Table 2). Studies assessing performance of viral MST targets
generally report very high specificities (23–26); however, that
is usually offset by low sensitivity as most of the viral methods
reported are present in human sources at considerably lower
concentrations compared to bacterial MST genetic targets.
Mitochondrial Targets
Human-associated mitochondrial DNA (mtDNA) MST
methods attempt to improve sensitivity and specificity com-
pared with bacterial and viral methods by targeting DNA
from a eukaryotic organelle. mtDNA originates from somatic
cells that often contain thousands of copies of the mitochon-
drial genome. In contrast, genes targeted for bacterial MST
applications often contain just a few copies of a particular
gene. Because exfoliated colonic epithelial cells can contain
thousands of mitochondria (27) and can be shed at rates as
high as 4 × 107 cells per g feces (wet weight) (28), MST meth-
ods targeting mtDNA may become more sensitive than other
approaches. In addition, mtDNA normally evolves rapidly at
the DNA sequence level and often provides enough variation
to identify species (29). Also, since mitochondria are highly
host-specific and cannot colonize or grow in environments
outside their native host, MST applications may be highly
specific.
In light of these potential advantages, a range of methods
that target human-associated mtDNA are available, includ-
ing PCR-based approaches (30, 31) and a dot-blot tech-
nology (32). To date, there have been very few published

TABLE 1 Selected PCR-based human-associated MST methods targeting bacterial genes

Target References
Assay Platform Chemistry
Bacteria Gene Method Application
HF183 Bacteroides spp. 16S rRNA PCR End-Point 37 72
qPCR SYBR 58 73
qPCR TaqMan 74 50
HF134 PCR End-Point 37 72
BacH qPCR TaqMan 75 76
BacHum qPCR TaqMan 77 65
BsteriF1 B. stericoris qPCR TaqMan 74 50
gyrB B. fragilis gyrB qPCR TaqMan 20 78
BtH B. thetaiotomicron α-1-6 mannanse gene qPCR TaqMan 22 79
HumM2 Bacteroides-like hypothetical protein qPCR TaqMan 80 50
 3.4.3. Microbial Source Tracking Methods Targeting Human Fecal Pollution Sources ▪ 3.4.3-3
TABLE 2 Selected human-associated MST methods targeting viruses
Assay Target
Human polyomaviruses Small t-antigen of human
polyomaviruses (HPyVs)
BK and JC
Pepper mild mottle virus Viral genome
GB-124 Bacteroides fragilis
Enterophage Enterococcus faecalis
Coliphage typing Viral genome
studies that directly test the performance of these mtDNA
methods relative to each other or other MST approaches. Bal-
leste and colleagues compared a human-associated mtDNA
assay (Humito) to a bacterial method targeting Bifodobacte-
rium adolescentis and reported a considerably higher sensitiv-
ity and specificity for the bacterial MST approach (33). Schill
and Mathes showed that untreated sewage contains concen-
trations of mtDNA cytochrome B genes at about two to
four orders of magnitude lower than the human-associated
HF183 gene target (34). The comparatively low abundance
in sewage suggests that mtDNA gene targets may also be
more difficult to detect in environmental water samples com-
pared to many bacterial targets, such as Bacteroidales. Another
potential challenge is that somatic cells not associated with
the intestinal tract, such as dead skin or hair cells, are shed
at high numbers by bathers and other humans in contact
with water, thus potentially lowering diagnostic specificity
as a fecal indicator and posing an increased risk of human
mtDNA contamination from laboratory technicians (35).
Finally, bovine mtDNA has been observed at one to two
orders of magnitude lower than human mtDNA in human
feces possibly due to beef consumption (30) and less than
an order of magnitude lower in wastewater (36). As a result,
it is recommended that additional validation studies are
completed that directly compare mtDNA and bacterial
MST methods to establish their efficacy for future MST
applications.
Bacterial Community-Based Methods
PCR-based methods rely heavily on the detection or quanti-
fication of monophyletic groups of bacteria or viruses, but
community-based methods may have advantages over PCR
methods because the diagnostic burden is not placed on a sin-
gle bacterial group. Instead, diagnosis relies on the overall
community similarity between reference fecal sources and
uncharacterized environmental samples. Community-based
methods are culture-independent, vary based on the techni-
ques used to generate sample profiles (e.g., restriction analysis,
microarrays, direct sequencing), and sometimes rely on ordi-
nation, cluster analysis, or other advanced statistical methods
to compare community profiles. The methods also allow the
identification of multiple fecal pollution sources simultane-
ously, not just those of human origin.
Terminal Restriction Fragment Length
Polymorphism (T-RFLP)
T-RFLP uses post-PCR restriction enzyme digestion to obtain
microbial community profiles and has been used for many
years to compare microbial communities and identify impor-
tant microbial taxa relevant to MST (37). More recently, Cao
and colleagues used both universal and Bacteroidales-targeted
T-RFLP community profiles along with visual and numerical
Methodology Chemistry Method reference
PCR End-Point 23
qPCR TaqMan 13
qPCR TaqMan 24
Culture-based Double agar overlay 81
Culture-based Double agar overlay 82
qPCR TaqMan 83
interpretation to correctly classify sources in a recent MST
validation study (38). However, the methods were less sensi-
tive than some PCR-based approaches and demonstrated
high levels of interlaboratory variability, presumably because
of the multiple-step protocol.
Phylochip
Preformatted microarrays, such as Phylochip, offer significant
advantages over T-RFLP and other methods because of the
large number of 16S rRNA gene sequences represented on
the array (n = 59,316) although in two studies only a fraction
of the Phylochip sequences were used for fecal source tracking
(n = 1,058 [39]; n = 503 [40]). Both approaches used fecal
samples to identify sequences that might be unique or
enriched in fecal sources to serve as identifier taxa. Dubinsky
and colleagues were able to use 541 human identifier taxa to
reliably detect human contaminants from an area affected by
a known sewage spill (39). Alternatively, Wu and colleagues
took significantly higher ratios of Bacteroidetes, Bacilli, and
Clostridia to α-proteobacteria to indicate the presence of
human contaminants (40). In a comparative study, depend-
ing on how the data were analyzed, Phylochip methods out-
performed other community based methods (T-RFLP and
next-generation sequencing) by showing 100% specificity
as opposed to 98% (38). Because previously defined sequen-
ces are present on the Phylochip, environmental 16S rRNA
gene sequences that hybridize to the array are almost immedi-
ately classified with relatively little computational analysis.
Currently, Phylochip MST methods are limited because
greater than 98% of the array is composed of taxa irrelevant
to source tracking and all taxa relevant to source tracking
are not represented (39).
Next-Generation Sequencing (NGS)
NGS technologies have also been used for MST (41, 42) and
are subject to similar limitations as other community profiling
methods. Furthermore, greater computational processing and
expertise in data analysis are needed to analyze the large num-
ber of DNA sequences generated, although some tools have
been designed to facilitate interpretation (41, 42). There
are relative few field studies available to date; however, New-
ton and collaborators recently used SourceTracker (41) to
identify combined sewer overflow events in an urban coastal
community (43). NGS community methods performed rela-
tively well in a recent validation study in terms of diagnostic
specificity (38), but it is unlikely that without significant
advances in data storage and computational analyses these
methods will surpass PCR-based approaches in terms of sim-
plicity and cost.
Because this class of MST methods is relatively new,
little information about their performance compared with
other methods is available. A potential challenge for most
3.4.3-4 ▪ MICROBIAL SOURCE TRACKING

community MST approaches is that signals from dominant
bacteria can overwhelm those of fecal indicator bacteria
when fecal pollution is highly diluted (38). More information
concerning the analytical sensitivity and limits of detection
of community-based MST methods is needed before these
approaches should be relied solely on for MST.
Chemical Methods
Chemical-based MST methods are designed to detect meta-
bolic by-products (e.g., coprostanol, caffeine, ratios of fecal
stanols, and other synthetic compounds) or other biochemi-
cals associated with human waste (44–46). As a group, these
methods are generally less sensitive than methods that target
microorganisms, as they require elevated concentrations of
chemicals that can be found only in close proximity to point
sources of human fecal pollution and their usefulness is usu-
ally limited to highly urbanized areas. Furthermore, interfer-
ence by other, natural substances present in the watershed
(e.g., humic acids) and relatively rapid degradation by ultra-
violet irradiation further limit their utility (47, 48). However,
in the appropriate areas, chemical-based methods can repre-
sent a rapid, cost-effective alternative to more expensive
and time-consuming molecular technologies.
FACTORS TO CONSIDER FOR METHOD
SELECTION
As already demonstrated, many human-associated MST
methods are available. Comparison studies have shown that
no single method is superior under all possible conditions.
Identification of the most appropriate method(s) for a partic-
ular application requires careful attention to a number of fac-
tors. In general, an optimal human-associated MST method
should be highly specific for human fecal pollution sources
and yield positive results to known human reference materials
collected from the watershed of interest on a consistent basis.
Very few methods appear to be 100% specific for human fecal
pollution sources. As a result, it is imperative to pay close
attention to which nonhuman animal sources are known to
elicit a false positive detection for a particular method. If
these sources are absent in the watershed of interest, it may
be appropriate to use the respective method even though it
may have poor specificity in other locations. As an extreme
example, a human-associated MST method may incorrectly
identify kangaroo feces as human, leading to a low specificity
estimate, but this lack of specificity may be irrelevant for
intended uses outside of Australia or areas impacted by zoos
housing these animals. Although an MST method may per-
form well in terms of specificity and sensitivity in the identi-
fication of reference fecal samples, there are a number of other
factors to consider when applying methods to real-world sit-
uations. This section introduces several issues that must be
accounted for, including defining the intended use applica-
tion, the importance of detection versus quantification of
the human indicator, the potential for indicator measure-
ment interference due to the local water sample matrix, and
evidence for correlations between human indicators and
regulated ambient water quality criteria.
Intended Application
The most important factor to consider before committing
to the use of a particular human-associated MST method is
the intended application. The intended application involves
not only clearly defining the problem but also careful consid-
eration of the desired outcome. For example, a watershed may
exhibit a chronic fecal pollution problem based on levels of
cultivated enterococci, a general fecal indicator that suggests
the presence of fecal pollution, but provides no information
about the fecal animal source(s) present. The watershed man-
ager performs a sanitary survey and determines that candidate
fecal pollutions sources include potentially leaky sewer lines
and/or native Canada goose populations. Funding is avail-
able to repair sewer lines, if it can be determined that human
fecal pollution is a contributing source. In this scenario, the
desired outcome is to determine the presence or absence of
human sewage pollution in the watershed. Method selection,
sampling strategies, and data analyses would be considerably
different from a scenario where the desired outcome is to
estimate the proportion of the total fecal pollution load
attributed to human and goose. Thus, failure to adequately
define the problem and desired outcomes can lead to improper
method selection, improper study design, and inconclu-
sive results. Table 3 lists some of the most common MST
applications.
Qualitative or Quantitative
Some MST methods generate only qualitative results; how-
ever, many approaches can yield both qualitative and quanti-
tative data. Qualitative MST data refer to results based on the
presence or absence of a human indicator, whereas quantita-
tive results indicate the estimated concentration of a human
fecal indicator. In general, quantitative methods are more
expensive and time consuming, and results are more difficult
to confirm. However, qualitative approaches typically require
more sampling and replicate measurements compared to
quantitative methods to achieve a similar level of confi-
dence in results. Selection of either detection or quantifica-
tion hinges on the particular problem and desired outcome
of a MST study, as well as access to resources and expertise
(many laboratories have a conventional thermocycler;

TABLE 3 MST applications

Problem Possible Strategy
Public health risk prediction Determine concentration of human indicator associated with public health risk.
Source exclusion Find evidence that human waste is or is not present over sampling duration.
Site prioritization Rank multiple sites based on degree of human fecal pollution present.
Fecal source allocation Determine proportion of human fecal pollution relative to other sources present.a
Non–point source identification Identify presence or absence of human fecal pollution source.
Best management practice evaluation Comparison of human fecal pollution indicator before and after implementation of BMP.
Hazardous event assessment Rapid determination of presence of human fecal pollution.
a
Fecal source allocation is at present still highly experimental.
 3.4.3. Microbial Source Tracking Methods Targeting Human Fecal Pollution Sources ▪ 3.4.3-5
fewer have the more expensive qPCR instruments). For exam-
ple, a hazardous event assessment application may only
require rapid determination of the presence or absence of
human fecal pollution in a flooded drinking water reservoir
(qualitative), where as a fecal source allocation application
necessitates an accurate estimate of a human indicator
(quantitative).
Matrix Interference
Matrix interference, or the ability of substances present in
environmental water samples to influence the measurement
of MST indicators, represents one of the key obstacles for
any MST application. For example, all PCR-based MST stud-
ies have to consider the possibility that components native to
environmental samples, such as particulate matter or humic
acids, can physically or chemically interfere with the detec-
tion and quantification of the respective gene target and
must adjust methods and study design accordingly. Research-
ers must be most cautious when comparing sample-to-sample
quantities because the recovery of purified DNA and the
coextraction of inhibitory substances can vary significantly
between samples and distort results if unaccounted for.
Many methods for estimating DNA recovery (49–51) and
detecting PCR inhibition (49, 50, 52, 53) have been devel-
oped and tested to varying degrees. The current thinking is
that the complete removal or sequestration of inhibitors with
specialized reagents along with controls for remaining inhib-
itors combined with an estimate of DNA extraction recovery
provide the most accurate results (49). In addition, chemical
methods, mainly the detection of optical brighteners, can be
susceptible to matrix interference. Significant background
levels of fluorescence, possibly due to the presence of aromatic
compounds, have been observed in the absence of human
contaminants; however, an increased diagnostic threshold
seemed to mitigate the effects of background fluorescence
on sample diagnosis (54). Even though it remains unclear
which matrix interference strategy is superior, it is critical
that at least one approach is used to prevent false negative
results or underestimation of contaminants.
Fate and Transport
Information collected through laboratory validation studies
has considerably increased our confidence in the sensitivity,
specificity, and reproducibility of human-associated MST
methodologies. However, once human fecal pollution is dis-
charged into the environment, it immediately begins a proc-
ess of decomposition, dilution, and transport that can impact
the interpretation of results. The importance of fecal pollu-
tion decomposition is closely linked to the intended use of
a particular method. For example, if the intended use is to
identify the presence or absence of human fecal pollution
in a water body at the point of use (i.e., a recreational beach),
it is not necessary to describe the fate and transport of the
microorganisms and chemical present in sewage. It is only
critical to identify the presence or absence of contaminants
in water samples collected directly from the water body of
interest. However, if the intended use involves estimating
the proportion of the fecal load associated with human sour-
ces, then a clear understanding of the decomposition of
human-associated indicators relative to measurements used
to estimate the total fecal load become paramount. Thus, it
becomes critical to determine whether a particular intended
use requires fate and transport information.
Unfortunately for MST applications requiring fate
and transport information, relatively little is known in this
area. Data obtained on environmental stability of MST
genetic targets suggest that various abiotic and biotic factors
influence persistence. For example, several studies have
shown that temperature is inversely proportional to persis-
tence (i.e., extended persistence at lower temperatures) of
several different human-associated MST indicators (33, 55–
58). Water type (marine, freshwater, etc.) also appears to
play an important role, as increased persistence was noted
for marine waters compared with freshwaters, presumably
due to the reduced predation in the former (52, 59). There
is no clear agreement among researchers with respect to the
effects of sunlight and associated UV irradiation; some studies
did not find it to be an important factor, whereas others
showed increased decay in the presence of sunlight (52, 60–
62). Studies investigating the effect of biotic interactions
( predation and competition) on persistence are rare; to date
it has been shown that reduced predation facilitates extended
persistence (56).
Correlation to Regulated Ambient Water
Quality Criteria
Some intended applications aim to link the data generated
from MST methods with the parameters used to regulate
ambient water quality criteria, including culture-based meas-
urements of enterococci and E. coli concentrations. Although
benefits of such an approach would be tremendous, at the
present time accurate and sensitive methods for fecal source
allocation are still lacking. One experimental approach inves-
tigated the potential use of a ratio between the estimated
human/ruminant-associated MST indicators and E. coli con-
centrations (63). However, the feasibility of a ratio approach
remains unclear at this time, as the potential for differential
decay of human-associated MST indicators compared to gen-
eral fecal indicator bacteria in environmental samples exists,
especially when samples contain contaminants from multiple
sources. This notion was confirmed for culture-based meas-
urements of enterococci and E. coli general fecal indicators
in a recent study reporting that large differences in the decay
of these different indicators introduces too much variability
for adequate fecal source apportionment (64). However,
the same study also suggested that differences in decay rates
were reduced when using qPCR measurements of general
fecal and human-associated indicators targeting Bacteroidales
and that fecal source apportionment may be possible if all
sources of human fecal contamination are a result of recent
events (65).
FUTURE RESEARCH NEEDS
Although a great deal is known about human-associated
MST methods, many research gaps remain before implemen-
tation is possible on a wide scale. One of the largest obstacles
for human MST implementation is standardization of MST
laboratory protocols and data analysis procedures. A recent
validation study of 23 human-associated MST methods indi-
cated a broad range of observed method performance across
laboratories implementing the same method (15, 16), and
that standardization of some laboratory protocols (15) as
well as data procedures (15, 66) increased interlaboratory
reproducibility. However, considerable variation was still
apparent in some instances even after partial laboratory stand-
ardization and complete standardization of data analysis
procedures, suggesting that more stringent control of fac-
tors that contribute to interlaboratory variability are needed.
For example, qPCR-based MST methods may require stand-
ardization of thermal cycler instrumentation, equipment, and
procedures for nucleic acid extraction and more rigorous
3.4.3-6 ▪ MICROBIAL SOURCE TRACKING
requirements for technician proficiency (15). Another
research gap is the characterization of the link between
human indicators, pathogens, and public health risk.
Although this link is not necessary for many of the intended
uses of a human-associated method, it is absolutely vital for
applications that assess public health risk. Limited data is
available for the co-occurrence of human pathogens and
human indicators (67), as well as the application of human-
associated indicators in epidemiology studies. Quantitative
microbial risk assessment modeling is currently being
explored as a way to predict public health risk (68, 69), but
more research is needed to substantiate this approach.
Another promising line of research seeks to integrate human
indicator measurements with GIS mapping and hydrologic
modeling. Advances in mapping technologies and character-
ization of surface runoff due to rain events provides important
information about adjacent land use and indicator/pathogen
transport mechanisms necessary to build a framework for
interpretation of human indicator data for many MST appli-
cations. While some field studies have been completed (50,
70), additional research is warranted to develop integrative
tools for water quality managers. Finally, it is important to
continue research using emerging technologies such as mas-
sive parallel sequencing (41, 42, 71) and microarray (39)
methods. These techniques are revolutionizing our under-
standing of the diversity, composition, and interactions of
fecal bacterial communities within hosts and when dis-
charged into the environment.
In summary, there are currently many human-associated
approaches designed to detect and/or quantify human fecal
pollution in environmental waters. The majority of these
methods employ PCR technologies; however, several chem-
ical and emerging bacterial community-based approaches
are available. The selection of the most appropriate method
represents a key challenge for end users and requires a clear
understanding of the strengths and weaknesses of the avail-
able methods, intended application, and desired outcomes
of a particular application.
REFERENCES
1. Hagedorn C, Crozier JB, Metnz MA, Booth AM, Graves
AK, Nelson NJ, Reneau RBJ. 2003. Carbon source utliza-
tion profiles as a method to identify faecal pollution sources
in water. J Appl Microbiol 94:792–799.
2. Konopka A, Oliver L, Turco RFJ. 1998. The use of carbon
substrate utilization patterns in environmental and ecologi-
cal microbiology. Microb Ecol 35:103–115.
3. Hagedorn C, Robinson SL, Filtz JR, Grubbs SM, Angier
TA, Reneau RBJ. 1999. Determining sources of fecal pollu-
tion in a rural Virginia watershed with antibiotic resistance
patterns in fecal streptococci. Appl Environ Microbiol 65:
5522–5531.
4. Harwood VJ, Whitlock J, Withington V. 2000. Classifica-
tion of antibiotic resistance patterns of indicator bacteria by
discriminant analysis: use in predicting the source of fecal
contamination in subtropical waters. Appl Environ Microbiol
66:3698–3704.
5. Parveen S, Portier KM, Robinson K, Edmiston L, Tamplin
ML. 1999. Discriminant analysis of ribotype profiles of
Escherichia coli for differentiating human and nonhuman
sources of fecal pollution. Appl Environ Microbiol 65:
3142–3147.
6. Wiggins BA, Andrews RW, Conway RA, Corr CL,
Dobratz EJ, Dougherty DP, Eppard JR, Knupp SR,
Limjoco MC, Mettenburg JM, Rinehardt JM, Sonsino J,
Torrijos RL, Zimmerman ME. 1999. Use of antibiotic
resistance analysis to identify nonpoint sources of fecal poll-
tuion. Appl Environ Microbiol 65:3483–3486.
7. Dombek PE, Johnson LK, Zimmerley ST, Sadowksy MJ.
2000. Use of repetitive DNA sequences and PCR to differ-
entiate Escherichia coli isolates from human and animal sour-
ces. Appl Environ Microbiol 66:2572–2577.
8. Myoda SP, Carson CA, Fuhrmann JJ, Hahm BK, Hartel
PG, Yampara-Iquise H, Johnson L, Kuntz RL, Nakatsu
CH, Sadowsky MJ, Samadpour M. 2003. Comparison of
genotypic-based microbial source tracking methods requir-
ing a host origin database. J Water Health 1:167–180.
9. Souza V, Rocha M, Valera A. 1999. Genetic structure of
natural populations of Escherichia coli in wild hosts on differ-
ent continents. Appl Environ Microbiol 65:3373–3385.
10. Griffith JF, Weisberg SB, McGee C. 2003. Evaluation of
microbial source tracking methods using mixed fecal sources
in aqueous test samples. J Water Health 1:141–151.
11. Samadpour M, Roberts MC, Kitts CL, Mulugeta W, Alfi
D. 2005. The use of ribotyping and antibiotic resistance pat-
terns for identification of host sources of Escherichia coli
strains. Lett Appl Microbiol 40:63–68.
12. Layton BA, Cao Y, Ebentier DL, Hanley K, Ballesté E,
Brandão J, Byappanahalli M, Converse R, Farnleitner
AH, Gentry-Shields J, Gidley ML, Gourmelon M, Lee
CS, Lee J, Lozach S, Madi T, Meijer WG, Noble R, Peed
L, Reischer GH, Rodrigues R, Rose JB, Schriewer A, Sini-
galliano C, Srinivasan S, Stewart J, Van De Werfhorst LC,
Wang D, Whitman R, Wuertz S, Jay J, Holden PA,
Boehm AB, Shanks O, Griffith JF. 2013. Performance of
human fecal anaerobe-associated PCR-based assays in a
multi-laboratory method evaluation study. Water Res 47:
6897–6908.
13. Rosario K, Symonds EM, Sinagalliano CD, Stewart JR,
Breitbart RA. 2009. Pepper mild mottle virus as an indicator
of fecal pollution. Appl Environ Microbiol 75:7261–7267.
14. Shanks OC, White K, Kelty CA, Sivaganesan M, Blannon
J, Meckes M, Varma M, Haugland RA. 2010. Performance
of PCR-based assays targeting Bacteroidales genetic markers
of human fecal pollution in sewage and fecal samples. Envi-
ron Sci Technol 44:6281–6288.
15. Ebentier DL, Hanley KT, Cao Y, Badgley BD, Boehm AB,
Ervin JS, Goodwin KD, Gourmelon M, Griffith JF, Holden
PA, Kelty CA, Lozach S, McGee C, Peed LA, Raith M, Ryu
H, Sadowsky MJ, Scott EA, Domingo JS, Schriewer A,
Sinigalliano CD, Shanks OC, Van De Werfhorst LC,
Wang D, Wuertz S, Jay JA. 2013. Evaluation of the repeat-
ability and reproducibility of a suite of qPCR-based microbial
source tracking methods. Water Res 47:6839–6848.
16. Boehm AB, Van De Werfhorst LC, Griffith JF, Holden
PA, Jay JA, Shanks OC, Wang D, Weisberg SB. 2013.
Performance of forty-one microbial source tracking methods:
a twenty-seven lab evaluation study. Water Res 47:
6812–6828.
17. Ley RE, Lozupone CA, Hamady M, Knight R, Gordon JI.
2008. Worlds within worlds: evolution of hte vertebrate gut
microbiota. Nat Rev Microbiol 6:776–788.
18. Ley RE, Hamady M, Lozupone CA, Turnbaugh PJ, Ramey
RR, Bircher SB, Schlegel ML, Tucker TA, Schrenzel MD,
Knight R, Gordon JI. 2008. Evolution of mammals and
their gut microbes. Science 320:1647–1651.
19. Ahmed W, Goonetilleke A, Powell D, Gardner T. 2009.
Evaluation of multiple sewage-associated Bacteroides PCR
markers for sewage pollution tracking. Water Res 43:
4872–4877.
20. Lee CS, Lee J. 2010. Evaluation of new gyrB-based real-time
PCR system for the detection of B. fragilis as an indicator of
human-specific fecal contamination. J Microbiol Meth 82:
311–318.
 3.4.3. Microbial Source Tracking Methods Targeting Human Fecal Pollution Sources ▪ 3.4.3-7
21. Shanks OC, Santo Domingo JW, Lu J, Kelty CA, Graham
JE. 2007. Identification of bacterial DNA markers for the
detection of human fecal pollution in water. Appl Environ
Microbiol 73:2416–2422.
22. Yampara-Iquise H, Zheng G, Jones JE, Carson CA. 2008.
Use of a Bacteroides thetaiotaomicron-specific alpha-1-6, man-
nanase quantitative PCR to detect human faecal polltion in
water. J Appl Microbiol 105:1686–1693.
23. McQuaig SM, Scott TM, Harwood VJ, Farrah SR,
Lukasik J. 2006. Detection of human-derived fecal pollu-
tion in environmental waters by use of a PCR-based
human polyomavirus assay. Appl Environ Microbiol 72:
7567–7574.
24. McQuaig S, Scott TM, Harwood VJ, Farrah SR, Lukaskik
JO. 2009. Quantification of human polyomaviruses JC virus
and BK virus by TaqMan quantitative PCR and comparison
to other water quality indicators in water and fecal samples.
Appl Environ Microbiol 75:3379–3388.
25. Noble RT, Weisberg SB, Leecaster MK, McGee CD, Rit-
ter KJ, Walker KO, Vainik PM. 2003. Comparison of beach
bacterial water quality indicator measurement methods.
Environ Monit Assess 81:301–312.
26. Harwood VJ, Boehm AB, Sassoubre LM, Vijayavel K,
Stewart JR, Fong TT, Caprais MP, Converse RR, Diston
D, Ebdon JE, Fuhrman JA, Gourmelon M, Gentry-Shields
J, Griffith JF, Kashian DR, Noble RT, Taylor HD, Wicki
M. 2013. Performance of viruses and bacteriophages for fecal
source determination in a multiple laboratory, comparative
study. Water Res 47:6929–6943.
27. Wang Y, Liu VW, Xue WC, Tsang PC, Cheung AN, Ngan
HY. 2005. The increase of mitochondrial DNA content
in endometrial adenocarcinoma cells: a quantitative study
using laser-captured microdissected tissues. Gynecol Oncol
98:104–110.
28. Kamra A, Kessie G, Chen JH, Kalavapudi S, Shores R,
McElroy I, Gireesh T, Sudhakaran PR, Dutta SK, Nair
PP. 2005. Exfoliated colonic epithelial cells: surrogate tar-
gets for evaluation of bioactive food components in cancer
prevention. J Nutr 135:2719–2722.
29. Wilson AC, Cannon RL, Carr SM, George MJ, Gyllensten
VB, Helm-Bychowski KM, Higuchi R, Palumbi SR,
Prager EM, Sage RD, Stoneking M. 1985. Mitochondrial
DNA and two perspectives on evolutionary genetics.
Biol J Linnean Soc 26:375–400.
30. Caldwell JM, Raley ME, Levine JF. 2007. Mitochondrial
multiplex real-time PCR as a source tracking method in
fecal-contaminated effluents. Environ Sci Technol 41:
3277–3283.
31. Martellini A, Payment P, Villemur R. 2005. Use of eukary-
otic mitochondrial DNA to differentiate human, bovine,
porcine and ovine sources in fecally contaminated surface
water. Water Res 39:541–548.
32. Kortbaoui R, Locas A, Imbeau M, Payment P, Villemur R.
2009. Universal mitochondrial PCR combined with species-
specific dot-blot assay as a source-tracking methond of human,
bovine, chicken, ovine, and porcine in fecal-contaminated
surface water. Water Res 43:2002–2010.
33. Balleste E, Bonjoch X, Belanche LA, Blanch AR. 2010.
Molecular indicators used in the development of predictive
models for microbial source tracking. Appl Environ Microbiol
76:1789–1795.
34. Schill WB, Mathes MV. 2008. Real-time PCR detection
and quantification of nine potential sources of fecal contam-
ination by analysis of mitochondrial cytochrome b targets.
Environ Sci Technol 42:5229–5234.
35. Baker-Austin C, Rangdale R, Lowther J, Lees DN. 2010.
Application of mitochondrial DNA analyiss for microbial
source tracking purposes in shellfish harvesting waters. Water
Sci Technol 61:1–7.
36. Caldwell JM, Levine JF. 2009. Domestic wastewater influent
profiling using mitochondrial real-time PCR for source
tracking animal contamination. J Microbiol Meth 77:17–22.
37. Bernhard AE, Field KG. 2000. A PCR assay to discriminate
human and ruminant feces on the basis of host differences in
Bacteroides-Prevotella genes encoding for 16S rRNA. Appl
Environ Microbiol 66:4571–4574.
38. Cao Y, Van De Werfhorst LC, Dubinsky EA, Badgley BD,
Sadowsky MJ, Andersen GL, Griffith JF, Holden PA.
2013. Evaluation of molecular community analysis methods
for discerning fecal sources and human waste. Water Res 47:
6862–6872.
39. Dubinsky EA, Esmaili L, Hulls JR, Cao Y, Griffith JF,
Andersen GL. 2012. Application of phyogentic microarray
analysis to discriminate sources of fecal pollution. Environ Sci
Technol 46:4340–4347.
40. Wu CH, Sercu B, Van De Werfhorst L, Wong J, deSantis
TZ, Brodie EL, Hazen TC, Holden PA, Andersen GL.
2010. Characterization of coastal urban watershed bacterial
communities leads to alternative community-based indica-
tors. PLoS One 5:e11285.
41. Knights D, Kuczynski J, Charlson ES, Zaneveld J, Mozer
MC, Collman RG, Bushman FD, Knight R, Kelly ST.
2011. Bayesian comunity-wide culture-independent micro-
bial source tracking. Nat Meth 8:761–763.
42. Unno T, Jang J, Han D, Ha Kim J, Sadowsky MJ, Kim O,
Chun J, Hur H. 2010. Use of barcoded pyrosequencing and
shared OTUs to determine source of fecal bacteria in water-
sheds. Environ Sci Technol 44:7777–7782.
43. Newton RJ, Bootsma MJ, Morrison HG, Sogin ML,
McLellan SL. 2013. A microbial signature approach to iden-
tify fecal pollution in the waters off an urbanized coast of
Lake Michigan. Microb Ecol 65:1011–1023.
44. Derrien M, Jarde E, Grauau G, Pourcher A, Gourmelon M,
Jadas-Hecart A, Wickmann AC. 2012. Development of
microbial and chemical microbial source tracking tools to
identify the origin of the faecal pollution in bathing and
shellfish harvesting waters in France. Water Res 44:
4812–4824.
45. Mieszkin S, Yala JF, Joubrel R, Gourmelon M. 2010. Phy-
logenetic analysis of Bacteroidales 16S rRNA gene sequences
from human and animal effluents and assessment of rumaint
faecal pollution by real-time PCR. J Appl Microbiol 108:
974–984.
46. McDonald JL, Hartel PG, Gentit LC, Belcher CN, Gates
KW, Rodgers K, Fisher JA, Smith KA, Payne KA. 2006.
Identifying sources of fecal contamination inexpensively
with targeted sampling and bacterial source tracking.
J Environ Qual 35:889–897.
47. Boving TB, Merrit DL, Boothryod JC. 2004. Fingerprint-
ing sources of bacteria input into small residential water-
sheds: fate of flourescent whitening agents. Environ Geol
46:228–232.
48. Cao Y, Griffith JF, Weisberg SB. 2009. Evaluation of opti-
cal brightener photodecay characteristics for detection of
human fecal contamination. Water Res 43:2273–2279.
49. Green HC, Field KG. 2012. Sensitive detection of sample
interference in environmental qPCR. Water Res 46:
3251–3260.
50. Peed LA, Nietch CT, Kelty CA, Meckes M, Mooney T,
Sivaganesan M, Shanks OC. 2011. Combining land use infor-
mation and small stream sampling with PCR-based methods
for better characterization of diffuse sources of human
fecal pollution. Environ Sci Technol 45:5662–5659.
51. Rajal VB, McSwain BS, Thompson DE, Leutenegger CM,
Kildare BJ, Wuertz S. 2007. Validation of hollow fiber
ultrafiltration and real-time PCR using bacteriophage PP7
as surrogate for the quantification of viruses from water sam-
ples. Water Res 41:1411–1422.
3.4.3-8 ▪ MICROBIAL SOURCE TRACKING
52. Green H, Shanks OC, Sivaganesan M, Haugland R, Field
K. 2011. Differential decay of human faecal Bacteroides in
marine and freshwater. Environ Microbiol 13:3235–3249.
53. Yamahara KM, Layton AL, Santoro AE, Boehm AB. 2007.
Beach sands along the California coast are diffuse sources of
fecal bacteria to coastal waters. Environ Sci Technol 41:
4515–4521.
54. Dickerson JW, Hagedorn C, Hassall A. 2007. Detection
and remediation of human-origin pollution at two public
beaches in Virginia using multiple source tracking methods.
Water Res 41:3758–3770.
55. Bonjoch X, Balleste E, Blanch AR. 2004. Multiplex PCR
with 16S rRNA gene-targeted primers of Biofidobacterium
spp. to identify sources of fecal pollution. Appl Environ
Microbiol 70:3171–3175.
56. Dick LK, Stelzer EA, Bertke EE, Fong DL, Stoeckel DM.
2010. Relative decay of Bacteroidales microbial source track-
ing markers and cultivated Escherichia coli in freshwater
microcosms. Appl Environ Microbiol 76:3255–3262.
57. Okabe S, Okayama N, Savichtcheva O, Ito T. 2007. Quan-
tification of host-specific Bacteroides-Prevotella 16S rRNA
genetic markers for assessment of fecal pollution in fresh-
water. Appl Microbiol Biotechnol 74:890–901.
58. Seurinck S, Defoirdt T, Verstraete W, Siciliano S. 2005.
Detection and quantification of human-specific HF183 Bac-
teroides 16S rRNA genetic marker with real-time PCR for
assessment of human faecal pollution in freshwater. Environ
Microbiol 7:249–259.
59. Okabe S, Shimazu Y. 2007. Persistences of host-specific
Bacteroides-Prevotella 16S rRNA genetic markers in environ-
mental waters: effects of temperature and salinity. Appl
Microbiol Biotechnol 76:935–944.
60. Bae S, Wuertz S. 2009. Rapid decay of host-specific fecal
Bacteroidales cells in seawater as measured by quantitative
PCR with propidium monoazide. Water Res 43:4850–4859.
61. Sokolova E, Astrom J, Pettersson TJ, Bargstedt O, Her-
mansson M. 2012. Decay of Bacteroidales genetic markers
in relation to traditional fecal indicators for water quality
modeling of drinking water sources. Environ Sci Technol
46:892–900.
62. Walters SP, Field KG. 2009. Survival and persistence of
human and ruminant-specific faecal Bacteroidales in fresh-
water microcosms. Environ Microbiol 11:1410–1421.
63. Stoeckel DM, Stelzer EA, Stogner RW, Mau DP. 2011.
Semi-quantitative evaluation of fecal contamination poten-
tial by human and ruminant sources using multiple lines of
evidence. Water Res 45:3225–3244.
64. Wang D, Farnleitner AH, Field KG, Green HC, Shanks
OC, Boehm AB. 2013. Enterococcus and Escherichia coli fecal
source apportionment with microbial source tracking
genetic markers—is it feasible? Water Res 47:6849–6861.
65. Wang D, Silkie SS, Nelson KL, Wuertz S. 2010. Estimating
true human and animal host source contribution in quanti-
tative microbial source tracking using the Monte Carlo
method. Water Res 44:4760–4775.
66. Raith MR, Kelty CA, Griffith JF, Schriewer A, Wuertz S,
Mieszkin S, Gourmelon M, Reischer GH, Farnleitner AH,
Ervin JS, Holden PA, Ebentier DL, Jay JA, Wang D,
Boehm AB, Aw TG, Rose JB, Balleste E, Meijer WG,
Sivaganesan M, Shanks OC. 2013. Comparison of PCR
and quantitative real-time PCR methods for the character-
ization of ruminant and cattle fecal pollution sources. Water
Res 47:6921–6928.
67. Walters SP, Gannon VPG, Field KG. 2007. Detection of
Bacteroidales fecal indicators and the zoonotic pathogens
E. coli O157:H7, Salmonella, and Campylobacter in river
water. Environ Sci Technol 41:1856–1862.
68. Soller JA, Schoen ME, Bartrand T, Ravenscroft JE,
Ashbolt NJ. 2010. Estimated human health risks from
exposure to recreational waters impacted by human and non-
human sources of faecal contamination. Water Res 44:
4674–4691.
69. Viau EJ, Lee D, Boehm AB. 2011. Swimmer risk of gastro-
intestinal illness from exposure to tropical coastal waters
impacted by terrestrial dry-weathery runoff. Environ Sci Tech-
nol 45:7158–7165.
70. Gentry RW, Layton AC, McKay L, McCarthy JF,
Williams DE, Koirala SR, Sayler GS. 2007. Efficacy of
Bacteroides measurements for reducting the statistical uncer-
tainty associated with hydrologic flow and fecal loads in a
mixed use watershed. J Environ Qual 36:1324–1330.
71. Lee JE, Lee S, Sung J, Ko G. 2010. Analysis of human and
animal fecal microbiota for microbial source tracking. ISME
J 5:362–365.
72. Shanks OC, Nietch C, Simonich MT, Younger M, Rey-
nolds D, Field KG. 2006. Basin-wide analysis of the dynam-
ics of fecal contamination and fecal source identification in
Tillamook Bay, Oregon. Appl Environ Microbiol 72:
5537–5546.
73. Sercu B, Van De Werfhorst L, Murray J, Holden PA.
2009. Storm drains are sources of human fecal pollution dur-
ing dry weather in three urban soutner California water-
sheds. Environ Sci Technol 43:293–298.
74. Haugland RA, Varma M, Kelty CA, Peed L, Sivaganesan
M, Shanks OC. 2010. Evaluation of genetic markers from
the 16S rRNA gene V2 region for use in quantitative detec-
tion of selected Bacteroidales species and human fecal waste
by real-time PCR. Syst Appl Microbiol 33:348–357.
75. Reischer GH, Kasper DC, Steinborn R, Farnleitner AH,
Mach RL. 2007. A quantitative real-time PCR assay for
the highly sensitive and specific detection of human faecal
influence in spring water from a large alpine catchment
area. Lett Appl Microbiol 44:351–356.
76. Reischer GH, Haider JM, Sommer R, Stadler H, Keiblin-
ger KM, Hornek R, Zerobin W, Mach RL, Farnleitner AH.
2008. Quantitative microbial faecal source tracking with
sampling guided by hydrological catchment dynamics. Envi-
ron Microbiol 10:2598–2608.
77. Kildare BJ, Leutenegger CM, McSwain BS, Bambic DG,
Rajal VB, Wuertz S. 2007. 16S rRNA-based assays for
quantitative detection of universal, human-, cow-, and dog-
specific fecal Bacteroidales: a Bayesian approach. Water Res
41:3701–3715.
78. Lee CS, Lee J. 2012. Application of host-specific source-
tracking tools for rapid identification of fecal contamination
in fresh produce by humans and livestock. Sci Food Agric 93:
1089–10965.
79. Srinivasan S, Aslan A, Xagoraraki I, Alocilja E, Rose JB.
2011. Escherichia coli, enterococci, and Bacteroides thetaiotao-
micron qPCR signals through wastewater and septage treat-
ment. Water Res 45:2561–2572.
80. Shanks OC, Kelty CA, Sivaganesan M, Varma M,
Haugland RA. 2009. Quantitative PCR for genetic markers
of human fecal pollution. Appl Environ Microbiol 75:
5507–5513.
81. Ebdon JE, Sellwood J, Shore J, Taylor HD. 2012. Phages of
Bacteroidales (GB-124): a novel tool for viral waterborne dis-
ease control? Environ Sci Technol 46:1163–1169.
82. Santiago-Rodriquez TM, Davila C, Gonzalez J, Bonilla N,
Marcos P, Urdaneta M, Cadete M, Monteiro S, Santos R,
Domingo JS, Toranzos GA. 2010. Characterization of
Enterococcus faecalis-infecting phages (enterophages) as
markers of human fecal polltuion in recrational waters. Water
Res 44:4716–4725.
83. Kirs M, Smith DC. 2007. Multiplex quantitative real-time
reverse transcriptase PCR for F+-specific RNA coliphages:
a method for use in microbial source tracking. Appl Environ
Microbiol 73:808–814.
 Methods of Targeting Animal Sources of Fecal
Pollution in Water
ANICET R. BLANCH, ELISENDA BALLESTÉ, JENNIFER WEIDHAAS,
JORGE SANTO DOMINGO, AND HODON RYU
3.4.4
The determination of fecal pollution sources in waters is
an essential subject in the management of catchments.
Although traditional microbial water analyses using indicator
microorganisms have been used for water health management
for more than a century, it is known that they cannot pro-
vide information about the origin of fecal pollution. The dis-
tinction between anthropogenic and nonanthropogenic
(animal) fecal pollution in water would greatly support assess-
ment of health risks from water exposure based on knowledge
of the host specificity of many pathogens. For example,
human sewage could constitute a greater health risk to
humans than waste of animal origin (1, 2). However, there
are some exceptions because some pathogens (named zoono-
sis) can infect and cause clinical disease in both humans and
animals (3). Therefore, fecal pollution source assessment
could support and determine different water management
strategies, treatment measures, and policies to prevent or
decrease fecal inputs in water and remediation of environ-
mental contaminant at the source.
In this chapter, proposed chemical and biological micro-
bial source tracking (MST) indicators for the determination
of animal fecal sources are discussed. The biological indica-
tors are grouped based on the phylogenetic description of
the proposed target (eukarya, bacteria, and virus). A compre-
hensive description for each proposed target is provided and
the developed methodologies employed are presented.
Emphasis is placed on the validation and applicability of
each proposed method and animal-MST indicator. More-
over, each proposed target is critically reviewed concerning
environmental factors such prevalence, resistance to different
water treatments, and environmental persistence. New
molecular approaches for animal-MST targets based on meta-
genomics are also presented in a specific section by analyzing
limitations and strengths at the present stage of the method-
ology. Finally, MST assay implementation on practical cases,
their contribution to the assessment of maximum fecal load of
water bodies, and their relationship to traditional microbial
indicators and waterborne pathogens is examined.
MST TARGETS FOR ANIMAL SOURCES
Chemical Targets
Several authors have reported on the use of chemical markers
for identifying fecal contamination in the environment and
doi:10.1128/9781555818821.ch3.4.4
3.4.4-1
their potential advantages and disadvantages over biologi-
cally based methods (4 and references therein). Chemical
markers of animal fecal contamination in the environment
include fecal sterols and their metabolic products, stanols, flu-
orescent compounds, and bile acids.
The metabolism of sterols in animal guts form fecal stanols
with end-product compositions and ratios varying based on
animal diet and gastrointestinal flora (5). Fecal sterols utilized
as chemical markers are presented in Table 1. Animal feces
contains considerable amounts of sterols and their metabolic
products (5), and it has been shown that the steroid finger-
print in animal manure can be transferred into runoff water
from amended soils (6) similar to what has been reported by
others (7, 8). Chemical source tracking methods are based
on detection of one particular fecal stanol or sterol or evalua-
tion of ratios of these compounds. For example, it has been
shown that herbivore feces (e.g., bovine, equine, and ovine),
compared with human feces, have greater relative proportions
of 5β-campestanol and 5β-stigmastanol from the metabolism
of campesterol and sitosterol (5). In contrast, dog feces
contains predominately cholesterol and bird feces contains
cholesterol and sitosterol (5). Similarly, 24-ethylcoprostanol
was shown to be in greater concentration than coprostanol in
animals and when used in combination with other parameters
correctly identified the source of contamination with a 6.5%
error rate (9). Others have used ratios of various fecal sterols
and stanols to determine source of fecal pollution (10). For
example, it has been shown that ratios of coprostanol plus epi-
coprostanol divided by cholesterol above 3.7 are indicative of
porcine feces, but ratios less than 0.7 are more indicative of
the presence of chicken and/or bovine feces (11). Similarly,
campesterol plus sitosterol divided by cholesterol ratios
greater than 1.5 indicate porcine, chicken, or bovine fecal
material, but ratios less than 1.0 are indicative of human feces
(11). Also, it has been shown that the ratio of epicoprostanol
divided by the combined value of coprostanol plus cholesterol
when greater than 0.1 is indicative of bovine, equine, or deer,
whereas values of this ratio that are less than 0.01 are indica-
tive of human feces (12). In other studies, ratios of coprosta-
nol versus the sum of all fecal sterols in a sample has been
shown to indicate nonhuman fecal impacts if the ratio is
less than 1.0 (13). Although various studies have used fecal
sterols and stanols directly or their ratios for identifying fecal
impacts to water bodies (5, 9–16), it has been suggested that
other water quality and monitoring parameters should be
3.4.4-2 ▪ MICROBIAL SOURCE TRACKING
TABLE 1 Chemical compounds proposed as MST host-specific
markers
Sterols and Fluorescent
Bile acids
stanols compounds
5α-cholestane Nor-cholic acid Tryptophane
Coprostanol Lithocholic acid Fluvic acids
Coprostanone Deoxycholic acid Humic acids
Epicoprostanol Cholic acid
Cholesterol Chenodeoxycholic
acid
Cholestanol Hyodeoxycholic acid
Stigmasterol Ursodeoxycholic acid
Stigmastanol Ursodeoxycholic acid
b-Sitosterol
Fucosterol
Campesterol
Demosterol
included in the evaluation to correctly classify the source of
fecal pollution (9).
Bile acids are excreted in the feces, and the composition of
bile acids released varies from species to species. Several stud-
ies have evaluated the use of bile acids for source tracking and
have been shown to be fairly specific in distinguishing
between animal and human feces in water. For example,
bile acids used in chemical source tracking have included
deoxycholic acid for ruminant bovines, lithocholic acid for
canines and humans, and the presence of hyocholic acid
but not deoxycholic acid for pigs (7, 8, 17–21).
Fluorescent compounds such as tryptophane-like fluores-
cence as opposed to fulvic/humic fluorescence from farm
waste has also been used to identify wastes from dairy bovine
slurry (22, 23). Compared to fecal sterols and stanols and stud-
ies involving bile acids, relatively few researchers have
reported on the use of fluorescent compounds for source
tracking studies.
Gas chromatography–mass spectrometry is the common
detection methods for fecal sterols and stanols and bile acids.
However, there are not standard methods for detection of
these compounds and some stanols and sterols produce very
similar analytical peaks, making differentiation of compounds
difficult in complex samples (4). Analytical detection limits
for these compounds are in the ng/liter range (4, 10, 13). Flu-
orescence measurements are typically taken with lumines-
cence spectrometers (22, 23). Known drawbacks to this
method of detection of fluorescent compounds is the inability
to distinguish individual compounds and the relatively high
detection limits.
One significant benefit to using chemical markers com-
pared with biological markers is that these compounds are
not subject to regrowth in the environment or persistence
of viable but nonculturable organisms. Additional benefits
to the use of fecal sterols include the relatively high concen-
trations in animal feces (5, 15) and the sensitive analytical
methods available (10, 13). Some drawbacks associated
with the use of chemical indicators for fecal source tracking
are that some are not specific to fecal sources; chemicals are
subject to microbial degradation once in the environment
(14); the potential for natural background concentrations;
the high cost of analytical techniques (20); poor correlation
with fecal indicator bacteria and pathogens in water (24);
and unknown persistence of some of the compounds in
environment.
Biological Targets
Eukaryotic Targets
DNA Mitochondrial Targets
Mitochondrial DNA (mtDNA) from host epithelial cells
shed from the gut lumen and present in feces have been used
as targets in MST studies. mtDNA targets have been sug-
gested to be good MST targets for several reasons. First, mito-
chondria are unique to the host species, and therefore
methods targeting these genes detect the presence of the
host species itself, rather than microorganisms associated
with a particular host. Second, multiple mitochondria are
found in eukaryotic cells, each containing multiple copies
of its genome. Known drawbacks to this method include
the potential for false positives: first by water contaminated
by nonfecal sources of mtDNA such as from skin, fur, or
hair; and second the potential for meat carryover in human
feces, from kitchen garbage disposals and industrial process
wastewater (25). Contamination of water from animal skin,
fur, or hair would indicate animal activity in and around
a water body, but not necessarily fecal contamination.
The potential for carryover of meat in municipal and indus-
trial wastewater is largely unknown, warranting further
investigation.
To date, published mtDNA targets include quantitative
PCR (qPCR) primers and probes targeting sequences for
the cytochrome b, cytochrome c, mitochondrial 12S and
16S ribosomal RNA, and NADH dehydrogenase subunit 2
or 5 genes on the mtDNA genome specific to humans,
bovine, porcine, ovine, chicken, horse, dog, domestic cat,
Canada goose, and deer (see Table 2) (26, 28–31, 33, 84).
In addition, so-called universal mtDNA targets have been
published (31, 32, 84). To date, a few qPCR mtDNA assays
have been used outside of the laboratories in which the meth-
ods were originally developed (27). Therefore, there is limited
data on the application of these assays for tracking animal
sources in real-world applications. The detection limits
reported for these methods have been estimated to range
from 1.8 × 104 to 1 × 106 gene copies (GC) /100 ml water
(29, 30, 32) or 1.6–90 mg feces/100 ml, assuming 1.1 × 107
GC/g feces (29). Therefore future work with mtDNA in
MST should include validation of the mtDNA qPCR assays
in interlaboratory studies using also environmental samples.
Protozoa: Cryptosporidium and Giardia
Cryptosporidium spp. are pathogenic protozoan parasites
which have a broad host range, including mammals, birds,
reptiles, and fish (85, 86). The first report of Cryptosporidium
associated disease was documented in a turkey flock more
than 50 years ago (87). At least two other species
(C. meleagridis and C. baileyi) have been identified in chicken
and other birds (88, 89). Besides avian-specific assays, several
species of Cryptosporidium have been found in a variety of ani-
mals, including C. andersoni (cattle), C. canis (dogs), C. felis
(cats), C. galli (birds), C. hominis (humans), C. molnari
(fish), C. muris (rodents and some other mammals),
C. parvum (ruminants and humans), C. wrairi (guinea
pigs), C. saurophilum (lizards and snakes), and C. serpentis
(snakes and lizards) (Xiao et al., 2004). To date, there are
16 valid named species and >40 genotypes of Cryptosporidium
(90). While some genotypes are referred to as wildlife geno-
types, some of which might potentially be different species,
TABLE 2 Mithocondrial and microbial MST indicators proposed for the detection of different animal fecal sources
Target organism/ Sensitivity of Environmental Geographic Environmental
Marker name Source Technique Specificity References
gene technique detection limit areas tested persistence
Pomito3 and mtDNA cytochrome c Swine PCR 103–104 cells/l 0.81 ND Quebec, <5 days in 26
Pomito11 oxidase subunit II, with PCR Canada, wastewater 27
tRNA-Lys, ATP 101–102 cells/l France,
synthase Fo subunits with nested Sweden, UK,
6–8 PCR Cyprus
Bomito1 and mtDNA: Nadlt5 Cow PCR 103–104 cells/l 0.69 ND Quebec, <5 days in 26
Bomito11 with PCR Canada, wastewater 27
101–102 cells/l France,
with nested Sweden, UK,
PCR Cyprus
Ovmito1 and mtDNA: Nadlt5 Sheep PCR 103–104 cells/l ND ND Quebec, Canada ND 26
Ovmito11 with PCR

3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-3

101–102 cells/l
with nested
PCR
Bovine mtDNA: NADH Cow TaqMan qPCR ND 0.89 2 × 107 cells/liter or USA, North ND 28, 29
dehydrogenase 20 mg feces/100 ml Carolina
subunit 5
Swine mtDNA: NADH Pig TaqMan qPCR ND 0.79 2 × 107 cells/liter or USA, North ND 28, 29
dehydrogenase 20 mg feces/100 ml Carolina
subunit 5
Dog NADH dehydrogenase Dog TaqMan qPCR ND 1.0 ND USA ND 28
subunit 5
Cat NADH dehydrogenase Cat TaqMan qPCR ND 1.0 ND USA ND 28
subunit 5
Canada goose mtDNA: NADH Canada goose TaqMan qPCR ND 0.66 ND USA ND 28
dehydrogenase
subunit 2
White-tailed deer mtDNA: cytochrome b White-tailed TaqMan qPCR ND 0.68 ND USA ND 28
deer
Cow mtDNA: cytochrome b Cow TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
Virginia and
Ohio
Pig mtDNA: cytochrome b Pig TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
Virginia and
Ohio
Chicken mtDNA: cytochrome b Chicken TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
Virginia and
Ohio
Sheep mtDNA: cytochrome b Sheep TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
Virginia and
Ohio
(Continued on next page)
TABLE 2 Mithocondrial and microbial MST indicators proposed for the detection of different animal fecal sources (Continued)

3.4.4-4 ▪ MICROBIAL SOURCE TRACKING

Target organism/ Sensitivity of Environmental Geographic Environmental
Marker name Source Technique Specificity References
gene technique detection limit areas tested persistence
Horse mtDNA: cytochrome b Horse TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
Virginia and
Ohio
White-tailed deer mtDNA: cytochrome b White-tailed TaqMan qPCR ND ND 2 mg feces/100 ml USA, West ND 30
deer Virginia and
Ohio
Canada goose mtDNA: cytochrome b Canada goose TaqMan qPCR ND ND 2 mg feces/ 100 ml USA, West ND 30
Virginia and
Ohio
Dog mtDNA: cytochrome b Dog TaqMan qPCR ND ND 2 mg feces/ 100 ml USA, West ND 30
Virginia and
Ohio
Ovmito-2 mtDNA: cytochrome c Sheep PCR, dot-blot ND 1b 0.1 mg feces/100 mla Quebec, Canada ND 31
oxidase subunit 3- hybridization
ND3
Bomito-2 mtDNA: cytochrome b Cow PCR, dot-blot ND 1b 0.1 mg feces/100 mla Quebec, Canada ND 31
hybridization
Pomito-4 mtDNA: NADH Swine PCR, dot-blot ND 1b 0.1 mg feces/100 mla Quebec, Canada ND 31
dehydrogenase hybridization
subunit 2
Ckmito-1 mtDNA: 16S rRNA Chicken PCR, dot-blot ND 1b 0.1 mg feces/100 mla Quebec, Canada ND 31
hybridization
cytB mtDNA: cytochrome b Swine, sheep, PCR ND 0.8 ND UK ND 32
cow, goat
cytB and cow mtDNA: cytochrome b Cow TaqMan qPCR ND ND ND UK >11 days 32, 33
cytB and sheep mtDNA: cytochrome b Sheep TaqMan qPCR ND ND ND UK 7 days 32, 33
cytB and pig mtDNA: cytochrome b Pig TaqMan qPCR ND ND ND UK 7 days 32, 33
CF128 Bacteroidales 16S Ruminant PCR 1·10−12 g total 41–100% ND USA, Canada, T90: 1.25–1.66// 27, 34–41
rRNA gene DNA UK, Ireland, detec. 14 days
Portugal,
France, Spain,
Australia
CF193 Bacteroidales 16S Ruminant PCR 1·10−12 g of total 99–100% ND USA, Canada, T90: 1 day//det. 27, 34, 36,
rRNA gene DNA France, Spain, 3–6 days 38, 41
BacCow Bacteroidales 16S Ruminant TaqMan 32 GC/reaction 40–95% ND USA T90: 3 days//detec 42–45
rRNA gene 6 days
BacR Bacteroidales 16S Ruminant TaqMan Detected until 89–100% ND Austria, France, T90: 1.4–10.6// 46–49
rRNA gene filtered volume Canada 7–>14 days
with 2·10−8 g
wet feces
Rum-2-Bac Bacteroidales 16S Ruminant TaqMan ND 100% ND France ND 50
rRNA gene
(Continued on next page)
TABLE 2 (Continued)
Target organism/ Sensitivity of Environmental Geographic Environmental
Marker name Source Technique Specificity References
gene technique detection limit areas tested persistence
CI125f, Bac708r, Bacteroidales 16S Bovine Scorpion probe ND ND ND Austria ND 51
1408 r rRNA gene
Cow-Bac 1,2,3 Bacteroidales 16S Bovine SYBR green 6.2 GC/reaction 100% ND Japan T90: 2–16 day//det: 52
rRNA gene 7–21 days
YCF Bacteroidales 16S Bovine SYBR green 60 GC/reaction 100% 8.5 × 10 to 4.2 × 102 Korea ND 53
rRNA gene cells per 100 ml
river water samples
BoBac Bacteroidales 16S Bovine TaqMan ND 47.4–97% 0.1 mg feces/liter USA ND 36, 54
rRNA gene
Bac2 Bacteroidales HDIG Bovine PCR 1·10−12 g of total 100% ND USA ND 36, 55
domain protein DNA
Bac3 Bacteroidales sialic Bovine PCR 1·10−12 g of total 98.9% ND USA ND 36, 55

3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-5

acid-specific DNA
9-O-acetylesterase
secretory prot
CowM2 Bacteroidales HDIG Bovine TaqMan 25 GC/reaction 100% ND USA T99: <2–4 days 36, 56
domain protein
CowM3 sialic acid-specific Bovine TaqMan 25 GC/reaction 98–100% ND USA ND 36, 48, 56
9-O-acetylesterase
secretory prot
EF447F/990R Bacteroidales 16S Elk PCR ND ND ND USA ND 57
rRNA gene
HoF597 Bacteroidales 16S Horse PCR ND 100% ND USA ND 57
rRNA gene
HorseBact Bacteroidales 16S Horse TaqMan ND 100% ND USA ND 43
rRNA gene
PF163 Bacteroidales 16S Pig PCR 1·10−15–1·10−12 77–100% ND USA, Canada, ND 39, 41,
rRNA gene g/DNA extract France 57–64
PigBac1,2 Bacteroidales 16S Pig qPCR 1·10−12 g/DNA 54–100% ND USA, Japan, T90: 2–16 day//det: 58, 65
rRNA gene extract France 11–28 days
Pig-1-Bac Bacteroidales 16S Pig TaqMan 1.6 GC/reaction 100% ND France ND 47
rRNA gene
Pig-2-Bac Bacteroidales 16S Pig TaqMan 1.6 GC/reaction 95–100% 3.2·104 GC/gr feces France, T90: 2 days//detect 47, 66
rRNA gene and 3.2·103 GC/ Canada 27 days
100 ml
DF475F/Bac708R Bacteroidales 16S Dog PCR ND ND ND USA ND 57
rRNA gene
BacCan Bacteroidales 16S Dog TaqMan 30 GC/reaction 35–86% ND USA T99: 4.25 days 42, 44, 45
rRNA gene Detec 11 days
(Continued on next page)
TABLE 2 Mithocondrial and microbial MST indicators proposed for the detection of different animal fecal sources (Continued)

3.4.4-6 ▪ MICROBIAL SOURCE TRACKING

Target organism/ Sensitivity of Environmental Geographic Environmental
Marker name Source Technique Specificity References
gene technique detection limit areas tested persistence
MuBa01 Bacteroidales 16S Muskrat TaqMan ND 100% 1.6·103 GC/ gr feces Canada ND 66
rRNA gene and 7.9·102 GC/
100 ml
CGOF1-Bac Bacteroidales 16S Geese TaqMan ND 100% 10 GC/rxn Canada ND 67
rRNA gene
CGOF2-Bac Bacteroidales 16S Geese TaqMan 10 GC/reaction 100% ND Canada ND 67
rRNA gene
Bif-PL-Probe Bifdobacterium Poultry TaqMan 25 GC/reaction 100% ND Spain ND 68
saeculare 16 s
rRNA gene
Bif-CW-Probe Bifdobacterium Bovine TaqMan 46 GC/reaction 100% ND Spain ND 68
pseudolongum 16 s
rRNA gene
Bif-PG-Probe Neoscardovia arbecensis Pig TaqMan 21 GC/reaction 100% ND Spain ND 68
16 s rRNA gene
Bif. thermacidophilum Bif. thermacidophilum Pig Nested-PCR ND 100% ND France ND 69
ssp. porcinum ssp. porcinum 16S
rRNA gene
Brevibacterium sp. 16S rRNA Poultry SYBR qPCR ND 0.93 100 mg poultry USA, Arkansas, ND 70, 71
LA35 litter/100 mL Oklahoma
F-RNA phages sum of Host-specific phages Animal lysis-plaque 1 PFU/plate or Animal source Samples could be USA, Europe, Summer T90 62 h, 59–64
genogrous formation <10 copies/ when sum concentrated and South Africa, winter T90
I and IV technique, reaction of the genogroups plated for PFU South 323 h
plaque target or serogroups count with America
hybridization, sequences I + IV higher sensitivity of the
qPCR, carried in than sum of culture method
multiplex plasmids for genogroups (1 PFU/plate)
multiplex multiplex or serogroups
real-time II + III
RT-PCR
Bacteroides phages Host-specific phages Human lysis-plaque 1 PFU/plate 70–100% of Samples could be UK, Spain, Summer T90 140 62, 72–74
bovine, formation samples of concentrated and Hawaii h, winter T90
porcine, technique fecal sources plated for PFU 417 h
poultry coinciding count with
with their sensitivity of the
origin culture method
(1 PFU/plate)
Bovine nteroviruses 50 nontranslated region Bovine TaqMan real 0.5 and 2 viral 76% cattle ND USA ND 75
(BEV-2) sequence of BEV time PFU in 14 µl of 50% deer
PS87 RT-PCR fecal extract 40% oysters
25% goose
(Continued on next page)
TABLE 2 (Continued)
Target organism/ Sensitivity of Environmental Geographic Environmental
Marker name Source Technique Specificity References
gene technique detection limit areas tested persistence
Bovine enteroviruses 50 nontranslated region Bovine TaqMan ND 78% cattle, 65% 12 BEV RNA Spain ND 76
(BEV-2) sequence of BEV qRT-PCR horses, 60% molecules/ml of
PS87 goats, water sample
36% sheep
Techoviruses (PTV) 50 noncoding region Porcine TaqMan ND 100% porcine 6.8 × 103 molecules of Spain ND 77
sequence of PTV-1 qRT-PCR PTV RNA/ml of
initial sample
Bovine adenovirus Hexon gene Bovine Nested PCR ND 75% bovine ND Spain ND 78
(BAdV) pooled feces
Porcine adenovirus Hexon gene Porcine Nested PCR ND 70% porcine ND Spain ND 78
(BAdV) pooled feces
Porcine adenovirus 68 bp region of hexon Porcine TaqMan qPCR ND 76–100% swine 1–10 GC in 4.2 ml Spain ND 79

3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-7

(PAV) gene pooled fecal sewage, or 1 liter of
samples, 100% river water or 0.1 g
porcine of stools
slaughter-house
wastewater
Bovine adenovirus Hexon gene of Bovine TaqMan qPCR 25 GC per qPCR 100% bovine USA ND 80
(BAdV) subgroup 2 tube reaction manure and
including serotypes 20% of cattle
4–8 feces
Bovine polyomavirus VP1 and agnoprotein Bovine Nested PCR 95% bovine ND Spain ND 81
(BPyV) regions slaughterhouse
wastewaters
Bovine polyomavirus 77 bp fragment of the Bovine TaqMan qPCR 10 GC per qPCR 30% bovine ND USA, Spain ND 82, 83
(BPyV) BPyV VP1gene tube reaction urine, 90%
bovine
slaughterhouse
wastewaters,
100% bovine
manure
a
Assuming 1.4 × 107 GC mtDNA/g feces.
b
<6 nontarget fecal samples tested.
ND: no data. GC: gen copies. CFU: colony forming units. PFU: plaque forming units.
3.4.4-8 ▪ MICROBIAL SOURCE TRACKING
some species and genotypes are considered host-specific, sug-
gesting that Cryptosporidium genotyping is a promising tool
for tracking contamination sources.
Cryptosporidium genotyping has been used to determine
the relative importance of different sources of this parasite
in environmental waters (91–95). Specifically, from a public
health standpoint, special attention has been focused on
human pathogenic species and human-specific genotypes.
At least eight species infect humans: C. hominis, C. parvum,
C. meleagridis, C. felis, C. canis, C. muris, C. suis, and the
Cryptosporidium cervine genotype (96). The first two species,
however, C. hominis and C. parvum, are the most dominant
species associated with Cryptosporidium outbreaks worldwide
(90, 97, 98) and therefore are a useful target for human fecal
sources.
Another protozoan group that has received attention in
the microbial water quality community is Giardia. In contrast
with Cryptosporidium, a limited number of Giardia species are
recognized and many of them have been detected in different
animals, including mammals, fish, and reptiles (99–102).
From a source tracking standpoint, G. duodenalis has received
the most attention due to its prevalence in mammals. Also,
some G. duodenalis genotypes or assemblages seem to be
more frequently found in some hosts than in others, suggest-
ing that Giardia genotypes could be used as evidence for some
hosts. PCR and sequencing assays targeting the SSU rRNA
and gdh, tpi, and bg genes have been commonly used as
molecular tools for the genotyping and subtyping of
G. duodenalis in fecal and water samples (103). Recently, real-
time PCR assays targeting gdh, tpi, bg, and open reading
frame C4 sequences of G. duodenalis have been developed,
showing relatively high specificity and sensitivity (104,
105). However, due to their broad host range, the use of
this protozoan group in source tracking studies has been
very limited. Further use of multilocus genotyping analysis
for the characterization of G. duodenalis from a variety of
host animals may enhance their use in source identification
(106). Future genome sequencing studies should shed more
light on the taxonomical affiliation of Giardia and their
potential use in MST.
Like most pathogens, the occurrence and abundance of
Cryptosporidium and Giardia in environmental waters is low
(91, 95, 107–109). For example, Yang et al. (95) reported
that Cryptosporidium oocysts and Giardia cysts were detected
in 8% (8/90) and 13% (12/90) of 10-liter river water samples
in the United States. One oocyst was observed in seven of
eight positive samples, whereas cysts ranged from 1 to 50
cysts/10 liter. In another U.S. study (108), 10% and 27% of
surface water samples (n = 162) were positive for Cryptospori-
dium oocysts and Giardia cysts, respectively. More recently,
Feng et al. (91) reported that the concentrations of Crypto-
sporidium and Giardia in river water of China averaged 5.2
oocysts/10 liters and 4 cysts/10 liters in positive samples,
respectively. Therefore, from an environmental monitoring
perspective, MST methods relying on protozoan targets are
facing the significant challenge of having to deal with low
detection limits.
Bacterial Targets
Bacteroidales and Related Genera
Bacteroidales together with Firmicutes are the most domi-
nant bacterial taxa in warm-blooded animal guts, outnumber-
ing coliforms by two to three orders of magnitude (110, 111).
Although gut microbiota may vary depending on animal age,
health status, diet, or geographical location, these groups have
been widely found in bovine, ovine, equine, porcine, and
chickens (112–116); however, the group remains less abun-
dant in avian species like gull, goose, duck, and turkey
(117–119). Bacteroidales order contains Gram-negative,
nonspore forming, and strictly anaerobic bacteria from genera
like Bacteroides, Prevotella, and Parabacteroides, which are cur-
rently the targets of most of the broadly used fecal source
tracking markers.
Bacteroidales host specificity was early on documented
using microbial community profiling methods such as length
heterogeneity PCR, terminal restriction fragment length
polymorphism, and 16S rRNA gene sequence analysis
(120, 121). 16S rRNA gene sequence analyses lead to the
identification of different animal host-specific Bacteroidales
markers that were later on used in PCR or qPCR assay devel-
opment. Most of these assays target the 16S rRNA gene since
it contains variable regions allowing the discrimination of
bacteria to subspecies level and has multiple copies in each
cell, which enhances target detection in environmental
samples. Comparison of Bacteroidales 16S rRNA gene
sequences in clone libraries generated from different animal
fecal samples have driven the development of ruminant-
(34), porcine- and equine-specific (57) PCR assays, and
bovine- (51, 53, 54, 65), ruminant- (42, 46, 50), porcine-
(47, 65), equine- (43), dog- (42) and geese-specific (67)
qPCR assays (Table 2). In other cases genome and fragment
enrichment subtractive hybridization has been used to obtain
the source specific bacteria to detect elk or dog sources
detected by PCR markers (122). The concentration of each
of these markers in fecal samples may differ depending on
the age, type, diet or geographic location. However, these
markers usually showed high levels in fecal samples from
the target host ranging from 108 to 109 GC/gram of feces
(Table 2).
Bacteroidales markers targeting genes other than 16S
rRNA have also been detected by PCR (e.g., Bac2 and
Bac3) and by qPCR (e.g., CowM2 and CowM3) (Table 2).
These markers codify for membrane associated and secreted
proteins that might be implicated in host-microbial interac-
tions (55, 56). These function-specific assays have been
shown to be highly specific, but may exhibit lower sensitivity
and occurrence in the host feces than 16S rRNA gene based
markers, which in turn is becoming a problem as the marker is
diluted in the environment.
Several animal associated Bacteroidales markers have
been widely applied for water monitoring around the world
(35, 123–127). One important caveat is that there are no
standardized protocols to detect Bacteroidales host-specific
markers, which may contribute to the different results
obtained by different laboratories. One specific problem asso-
ciated with some of these markers is cross-reaction with other
fecal sources (specificity). Host-specificity of various Bacter-
oidales markers may differ with respect to geographical loca-
tion as documented in several studies. For example, Shanks
et al. (36) evaluated seven ruminant and bovine markers,
while Lamendella et al. (58) studied three porcine associated
markers. None of the tested markers achieved 100% sensitiv-
ity and specificity, which contrasted with the results obtained
in the original studies. These differences can be the result of
factors such as diet and host age. While non-targeted samples
may generate significant false positive signals, the abundance
of such an MST marker is often much lower than in the tar-
geted sample, and therefore, non-specific amplification might
be negligible once the samples reach environmental
waters (43). In those cases in which the performance of the
assays differ significantly from the area where the methods
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-9
were developed, a combination of MST makers through pre-
dictive models could improve the value of Bacteroidales
methods (27).
Bacteroidetes species are strictly anaerobic bacteria, thus
their potential for regrowth in the environment is limited,
which is a prerequisite for a microbial source tracking marker.
However, this anaerobic characteristic can accelerate their
die-off outside the host, hampering the persistence of these
markers in the environment and limiting their use to identify
recent pollution events (48). Several studies have focused on
describing the persistence of these markers in the environ-
ment and evaluating the main factors involved in their die-off
(49, 52, 66, 128, 129). Environmental factors reported to
affect Bacteroidetes survival include temperature, sunlight,
dissolved oxygen concentration, salinity and predation. The
difficulty in controlling the main environmental factors and
their interaction can lead to slightly different decisions
regarding the experiment design. For instance, different Bac-
teroides decay rates have been observed in situ versus in con-
trolled microcosm conditions (37). For example, differences
noted on the effect of sunlight on MST marker persistence
are likely due to the way the persistence study was conducted
(38, 44, 129). Several studies have concluded that Bacteroi-
detes decay rates increase with higher temperature (49, 52,
128, 130) while increase in salinity may promote longer sur-
vival in seawater (52, 131, 132). The T90 (time needed to
achieve one logarithm reduction from the initial population)
values vary depending on the marker and environmental con-
ditions tested but have been reported to range from 1 to 10
days. Moreover, Bacteroidales signals have been detected by
PCR or qPCR as long 14 days (38, 44, 48, 49, 52, 66). Addi-
tionally, grazers’ predation and viruses’ lysis have been
reported to be one of the strongest influences on Bacteroidales
numbers, especially at warmer temperatures (37, 48, 128, 130,
133). The specific composition and abundance of grazers and
viruses is likely related to the water matrix and may influence
the repeatability of experiments.
The potential of MST markers to predict the presence of
pathogens has been analyzed either by spiking a water sample
(with feces or pathogens) and following their decay, or by
comparing the levels of MST and pathogens in environmen-
tal water samples. The persistence of Bacteroidales source
tracking markers have equaled or exceeded the reported per-
sistence of Campylobacter and E. coli O157:H7 in some stud-
ies, while results have been more variable for Salmonella (39,
45, 134). Some studies suggest that MST methods can be used
to predict pathogens (135), but in other cases the correlation
between Bacteroidales and pathogens was non-existent (136,
137). In general, there is insufficient data on survival and cor-
relation of MST markers and pathogens. Open access to data-
bases in which data from survival studies is shared will allow
researchers the opportunity to compare and select decay rates
for the microorganism of interest and under environmental
conditions similar to their study site (132).
Bifidobacterium and related genera
Bifidobacterium spp. are rod-shaped, Gram-positive and
strict anaerobic bacteria that belonging to the Actinobacteria
phyla and that are commonly associated with human and ani-
mal intestines. Although less abundant than Bacteroidetes
group, Actinobacteria are always detected in human and
animal guts, being particularly abundant during the first
two years of breast-feeding infants. Bifidobacteria were ini-
tially investigated for microbial source tracking purposes in
the early 1980s, and specifically suggested to be good indica-
tors of sewage pollution (138). Techniques to detect
bifidobacteria as an indicator of human fecal pollution typi-
cally include both culture dependent methods (to detect sor-
bitol-fermenting bifidobacteria) (139, 140) and culture
independent methods targeting human specific species as
B. adolescentis and dentium (141, 142). The analyses of micro-
bial communities from different animal sources by length
heterogeneity PCR, terminal restriction fragment length
polymorphism, denaturing gradient gel electrophoresis, and
capillary electrophoresis single-strand conformation poly-
morphism have also shown strong animal host specificity
for this group (69, 121, 143), and thus specific species have
been reported to be associated with animals (144). Animal-
specific Bifidobacteria includes B. saeculare strongly associated
with poultry and B. pseudolongum associated with bovine
fecal pollution (68, 143). TaqMan qPCR-based methods
detected these organisms with 100% sensitivity and specific-
ity and with an abundance 105–106 GC/ml for the poultry
marker in slaughterhouse wastewater and 103–107 GC/ml
for the bovine marker (68). Additionally, the screening of
porcine slurries by an specific hybridization oligonucleotide
probe aided in the isolation of the new species Neoscardovia
arbecensis, which was found to be strongly associated to por-
cine fecal pollution (145). A TaqMan qPCR-based method
to detect this bacterium (68) reported a concentration rang-
ing from 105 to 108 GC/ml in slaughterhouse wastewater. B.
thermocidophilum ssp. porcinum has also been associated with
porcine feces and the PCR used for its detection as an MST
marker was reported to exhibit good performance (69).
Bifidobacteria as anaerobic bacteria are likely to have a low
capability to survive in the environment after being shed in
feces. Although some studies have used a two-step nested
PCR approach to enhance the chances of environmental
detection (142, 146), recently some host-specific Bifidobacte-
rium qPCRs have been developed showing a high sensitivity
on environmental samples (68). There are no studies evaluat-
ing the persistence of animal-associated bifidobacteria in
environmental waters. However Marti et al. (69) observed
Bifidobacterium spp. to be the less impacted bacterial popu-
lations on porcine manure after an aerobic treatment.
B. thermocidophilum ssp. porcinum was detected in lagoons
treating pig slurries with retention time of 5 days. Persistence
studies performed with human-associated species indicates a
better survival of bifidobacteria at low (T90: 4.2–4.6 days
at 4–10°C) than at high temperatures (T90: 1.7–1.8 days
at 21–25°C) (147). Similar decay rates were obtained by
another study (T90: 3.4–3.8 days at 18.5°C) in sea- and fresh-
water samples, in which Bifidobacteria were detected up to 13
and 20 days of incubation, respectively (131).
Brevibacterium
Brevibacterium spp. are Gram-positive bacteria commonly
found in soils, and which members belong to the Brevibacter-
iaceae family (Actinomycetales). A SYBR green qPCR assay
specific to poultry litter (i.e., chicken and turkey litter) has
been developed targeting the 16S rRNA gene of Brevibacte-
rium sp. LA35 (70). The Brevibacterium sp. LA35 marker
was validated in target samples (17 poultry litter samples,
23 individual chicken and turkey fecal samples) and nontar-
get samples (116 samples from bovine, porcine, ducks, geese,
humans). The SYBR green qPCR assay was shown to have a
detection limit of 30 GC per reaction and 95% specificity,
with few false positives in goose fecal samples and human
wastewater samples. Independent confirmation of its high
level of sensitivity and host specificity was performed with
DNA extracts from diverse animals (e.g., poultry litter; non-
target feces from cows, waterfowl, swine; and stream samples).
3.4.4-10 ▪ MICROBIAL SOURCE TRACKING
Thus far, the poultry litter marker was only been detected in
environmental samples expected to be impacted by poultry
litter. To date the published literature reports testing for
this poultry marker in litter and poultry fecal samples from
the United States (Oklahoma, Georgia, Florida, Utah, and
Minnesota) but limited geographical water sample testing
(e.g., Oklahoma and Arkansas). Further testing for the
marker in additional watersheds is recommended as well as
additional nontarget fecal sample testing (e.g., domesticated
pets, additional wild fowl).
Rhodococcus coprophilus
Rhodococcus coprophilus is an aerobic, nonsporulating,
nonmotile Gram-positive bacterium (148, 149), which was
one of the first Actinomycetes used in MST studies (150–
153). A natural inhabitant of herbivore feces (e.g., bovine,
ovine, porcine, equine, donkeys, waterfowl, and chickens;
151–153), R. coprophilus has been suggested to be a good indi-
cator of contamination of aquatic environments by animals as
compared with humans. Specifically, Rowbotham and Cross
(148, 149) identified R. coprophilus in feces of domesticated
herbivores and in runoff from agricultural pastures but its
absence from human feces. Others have found that
R. coprophilus counts were correlated with fecal streptococci
in aquatic habitats (152) and that this organism was more
closely associated with animal fecal pollution in water rather
than human fecal pollution (150, 154). Its persistence in
water and sediment, up to 120 days, has been shown to be lon-
ger than that of common fecal indicator bacteria such as fecal
streptococci (152); however, this persistence could be a prob-
lem as its presence in the environment is not necessarily indi-
cative of recent fecal contamination of waterways nor risk due
to exposure to pathogens.
Methods for enumeration of R. coprophilus include a cul-
turing method (150–152, 155) requiring up to 21 days for a
presumptive result and more recently PCR (153) and qPCR
(153, 156) assays targeting the organism’s 16S rRNA gene.
Detection methods for R. coprophilus by the various methods
are 60 CFU per PCR reaction (153) and 10 CFU/ml sample
for qPCR (156) or 1 CFU per qPCR reaction (153). There is
some evidence that R. coprophilus detected by culture-based
methods is correlated with fecal coliforms and enterococci
in fecally impacted watersheds (157). However in this study,
R. coprophilus was only correlated with these indicator organ-
isms in one of three rivers in Massachusetts, therefore the
authors suggested that this organism is not suitable for routine
monitoring alone. Rather, they suggested a weight of evi-
dence approach should be used.
Enterococcus, Streptococcus, Catellicoccus, and
Helicobacter
Most bacterial MST markers have targeted Gram-negative
bacteria, namely, Bacteriodales species and E. coli. However, as
mentioned earlier, Gram-positive bacteria have also been
suggested as potential fecal source identifiers. For example,
some studies have targeted members of the order Lactobacil-
lales. Specifically, species of the genus Enterococcus have been
used as indicators of fecal pollution for several decades and,
more recently, as the target of source tracking studies, partic-
ularly in library-dependent MST studies (158). Some enter-
ococci species, such as Enterococcus faecalis, have been
proposed as a potential indicator of human and poultry fecal
pollution (159, 160). However, a qPCR-based study showed
that E. faecalis, E. casseliflavus, E. gallinarum, and E. faecium–
like species are present in the feces of many different animals
(161). It should also be noted that E. gallinarium was
originally isolated from chicken, but its occurrence in birds
is not very well documented. Moreover, E. gallinarium and
E. casseliflavus 16S rRNA gene sequences are nearly identical,
and therefore 16S rRNA gene assays are not useful at discrim-
inating these two species. In fact, while enterococci are a
phylogenetically diverse group consisting of more than 22
species, 16S rRNA gene sequences have not been used as a
target of host-specific assays. Other gene targets such as super-
oxide dismutase (sodA) (162) might be useful at determining
the occurrence and relative abundance of E. gallinarium and
other enterococci species and therefore providing a better
framework for enterococci-based MST studies.
A gene associated with an enterococci surface protein
(i.e., esp gene) was proposed as a human-specific marker
(163). Different from most host-specific PCR assays, detec-
tion of the esp gene relies on a culture-enrichment step fol-
lowed by the PCR amplification step. As a result, the
method can only be used as a presence/absence assay, not to
quantify source-specific loads. A few studies have demon-
strated that esp detection could be useful in scenarios in which
wastewater treatment plants are the primary fecal sources
(164, 165). Thus far, no applied studies have shown that
the esp assay can be used when septic tanks are implicated
as fecal sources. On the other hand, several studies have dem-
onstrated that the esp assay cross-amplifies with nontargeted
sources such as dog, mice, and waterfowl (166–168), thus lim-
iting the value of this assay in environmental applications as a
human marker. Additionally, E. faecium isolates lacking the
esp gene have been shown to exhibit higher transport rates
in sand as compared to wild types suggesting source tracking
approaches targeting this gene might not be effective in
groundwater systems (169).
Another promising MST target is Catellicoccus marimam-
malium. Although it was originally isolated from dead marine
mammals (170), several studies have shown that C. marimam-
malium and close relatives are frequently detected in different
waterfowl hosts (117, 171). Thus far, different C. marimam-
malium–based host-specific assays have been developed and
used in environmental monitoring applications (117, 171,
172). Interestingly, attempts to grow C. marimammalium
from environmental waters have not been fruitful (unpub-
lished data). Genome sequence analysis has indicated that
this species has a reduced metabolic network (173), suggest-
ing that indeed this species might be transitioning into an
obligate symbiosis with specific hosts, particularly birds, and
explaining in part why it is difficult to grow on basic culture
media.
Other lactobacilli species have been isolated from several
animals, suggesting that they might be potential targets for
MST methods. For example, Streptococcus equi has been pri-
marily isolated from horses, although zoonotic cases have
been reported (174). Additionally, the following species
have also been isolated from animals: S. gallolyticus ssp. gallo-
lyticus and S. gallinaceus from poultry, S. pluranimalium from
cattle (175), S. porcorum and S. plurextorum from pigs (176,
177), and S. castoreus from beaver (178). In most cases, these
species have been associated with animal disease, yet little is
known on their occurrence in other hosts or their abundance
in fecal sources.
Phenotypic Library-Dependent Methods Based on
Microbial Antibiotic Resistance or Carbon Source
Utilization
Two phenotypic profiling techniques have been suggested
to discriminate between human and animals sources. These
techniques are based on differences in carbon utilization
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-11
profiles (179) and antibiotic resistance profiles (158) of fecal
indicator bacteria (FIB), specifically E. coli and enterococci.
In these methods, the profiles of FIB strains isolated from fecal
samples are compared with those isolated from water sam-
ples. The rationale of using antibiotic resistance profiles
stands from the assumption that human and domesticated
animals are exposed to different antibiotics while wildlife
exposure to antibiotics is relatively minimal (180, 181).
The rationale used in carbon utilization profile–based
methods is related to the potential differences in dietary
regime and gut metabolic potential of the different hosts.
In many regards, the use of phenotypic profiles reenergized
the field of source tracking, which had initially depended
on the fecal streptococci and fecal coliform ratios (182).
Although many of the initial papers showed promising
results, the need for large databases for each study precluded
their widespread use in environmental applications. Yet the
identification of resistance profiles in fecal bacteria is
important to public health in light of the increase in
antibiotic-resistant bacteria, a trend that is likely to con-
tinue in years to come (183). On the other hand, the direct
detection of antibiotic resistance genes in environmental
samples using PCR techniques (184) and metagenomic
approaches (58) might prove to shed light on the impor-
tance of fecal pollution sources in the environmental distri-
bution and mobilization of these genes.
Genotypic Library-Dependent Methods
Besides phenotypic methods, there are library-dependent
methods based on genotypic profiles. In genotypic-based
approaches, bacterial strains are isolated and DNA from
each strain is used to generate fingerprints based on molecular
techniques such as repetitive sequence-based PCR (185,
186), enterobacterial repetitive intergenic consenses PCR
(185, 187), pulsed-field gel electrophoresis (187, 188),
16S-23S rRNA intergenic spacer region PCR (186), and
ribotyping (189). The use of these techniques gained popu-
larity more than 10 years ago as some of them were used in epi-
demiological studies to identify the strains associated to
waterborne and foodborne outbreaks. Several studies have
used these techniques to determine the presence of animal
sources in waters, using E. coli and enterococci isolates as
the primary source identifiers (190, 191). Culture-dependent
genotypic methods are not used as much in environmental
applications due to the number of isolates that are needed
to statistically determine their performance (192), the poor
classification of environmental isolates (193), and the diffi-
culties associated with the use of gel imaging to accurately
identify unique patterns (194).
Viral Targets
Bacteriophages
Bacteriophages are estimated as the most abundant
microorganisms on the planet (around 1031), since according
to electron microscope observations, phages are one order of
magnitude more abundant than bacteria in the oceans
(i.e., 1030 total viral particles) (195–198). The abundance
of bacteriophages suggests they may be of interest as targets
of MST. Bacteriophages infecting selected host strains related
to human or animal fecal sources would be especially
attractive as MST targets. Usually, these host strains are com-
mensal microbiota associated with the intestinal tract of
humans and/or animal species. Some bacteriophages are
viable in water for long periods and could be detected when
culture or molecular methods are available. Moreover,
bacteriophages were first proposed as alternative indicators
for fecal pollution (199), and since then they have been
extensively studied as surrogates of fecal and viral pollution
(200–202). Bacteriophages could solve the shortcoming of
fecal bacteria as indicators for waterborne viral pathogens.
For example, bacteriophages have many structural similarities
with enteric viruses, suggesting they may more closely model
environmental persistence and, more important, resistance of
pathogenic viruses to treatment process (201, 202). More-
over, the detection and enumeration of bacteriophages could
be done by simple and inexpensive techniques, which do not
constitute any health risk to analysts and could be performed
in any basic microbial laboratory.
Qualitative ( presence-absence enrichment test) and
quantitative (direct quantitative plaque assays) methods are
well established (203), and standardized protocols exists for
some particular bacteriophages or groups of bacteriophages
(59–61, 72, 204). Although molecular techniques have
been developed for the detection of enteric bacteriophages
(205–207), their use is limited since the numbers of viruses
in some water environments may be low and not readily
detectable (208) and due to the large diversity of bacterio-
phages included in some of the groups studied as indicators.
However, for both the standardized and the molecular meth-
ods, large sample volumes ranging from 100 to 1000 ml could
be taken by using practical recovery methods (209, 210),
which could improve the detection limits for certain meth-
odological approaches.
Presently, two bacteriophage groups have been proposed
and extensively studied as animal fecal source MST indica-
tors: ratios of the main four serotypes/genotypes of F-specific
RNA coliphages (211, 212) and bacteriophages infecting
various species of Bacteroides (73, 74, 213). Recently, Purnell
et al. (214) and Santiago-Rodriguez et al. (126) also investi-
gated bacteriophages infecting certain Enterococcus host
strains. The simultaneous detection of total somatic coli-
phages with either bacteriophage groups has been suggested
to provide essential data in studies of MST on the total
load of fecal contamination in water (9). Moreover, the ratio
of total somatic coliphages and host-specific bacteriophages
have been identified as important variables in many predic-
tive models based on inductive machine learning (215),
such as the ratio of somatic coliphages to GA17 human-
related Bacteroides phages (216) or somatic coliphages to
PL122 poultry-related Bacteroides phages (74).
F-Specific RNA Coliphages. Phages infecting E. coli hosts
are usually named coliphages. Somatic coliphages infect their
host through cell wall receptors, while F-specific bacterio-
phages (also named F-specific or “male-specific” coliphages)
infect their hosts through receptors on F pili. F-specific bac-
teriophages belong to the families Leviviridae (RNA
genome) and Inoviridae (DNA genome). F-specific RNA
coliphages were subdivided into four groups by serological
typing (217). Later, these serological groups were further
identified as analogous to genogroups (211, 218). Although
association of F-RNA serogroups with a particular fecal
source is not conclusive, groups II and III have been usually
associated with human fecal sources and groups I and IV
with animal fecal origins (212, 219, 220). However, gen-
ogroup I has been frequently isolated from municipal sewage,
and it is not clear if this is due to animal wastewater in the
municipal sewage or if it is also associated with humans
(221–223). F-RNA coliphages are in low abundance in feces,
and their environmental multiplication seems to be unlikely
because of the production of F pili is temperature-dependent
3.4.4-12 ▪ MICROBIAL SOURCE TRACKING
and is not expressed below 25°C (224). Such low occurrence
could reduce their application for MST studies in some envi-
ronments. Total concentration of F-RNA bacteriophages
in raw domestic wastewater could be up 5 × 105 plaque form-
ing units (PFU)/100 ml.
Standard methods based on incubating these coliphages
in a medium with the host bacteria and later observation of
lysis-plaque formation are available (59–61). These tradi-
tional plaque formation methods have been improved for a
rapid analysis using latex agglutination after 2 to 5 h enrich-
ment steps (225) or linking host lysis to changes in color
linked to β-galactosidase induction for the presence/absence
determinations (226). Recently, several PCR-based
approaches have been also developed to distinguish the four
genogroups of F-RNA coliphages (205, 206, 227, 228).
Some ratios or summation of certain genogroups have
been proposed to be used as indicators of animal fecal sources
in statistical or predictive MST analyses, for instance, the
ratio of low F-RNA II to F-RNA I (206) and the sum of
F-RNA I and F-RNA IV (9). However, low occurrence of
F-RNA in feces and differential persistence of genogroups
in water environments (Table 2) (usually genogroup I persist
longer, while genogroup IV is the least persistent) are serious
limitations to the use of such ratios or combinations in routine
monitoring (62–64, 222). Though the enrichment of coli-
phages could increase their detection, this practice is discour-
aged because it seems to decrease the representativeness of
coliphage diversity in a water sample, and consequently it
could disguise the proportions of genogroups (223). Never-
theless, large and multilaboratory comparison studies of pro-
posed MST indicators showed that the distribution of
F-RNA genotypes is still among the best predictive MST
indicators (9, 229).
Bacteriophages Infecting Bacteroides. Bacteroides phages
are double-stranded DNA somatic phages, most of which
have been associated to Siphoviridae due to their morpholog-
ical and genomic features (202). Bacteroides phages have been
detected in wastewater and feces and have been described
infecting strains of B. fragilis, B. thetaiotaomicron, B. rumini-
cola, and B. ovatus (73, 74, 202). Due to the limited growth
of their bacterial hosts, it is believed that Bacteroides phages
do not replicate outside the intestinal track (230), which
will further support the possible coevolution of Bacteroides
phages and the bacterial host of origin (231–233).
Plaque assay and enrichment methods for the detection
and enumeration Bacteroides phages are documented (72).
Concentration techniques are available allowing for sampling
of large volumes of water (209). The use of different host
strains related to a specific fecal source has been related to dis-
tinct climatic characteristics and/or diets on geographical
areas (234, 235). An easy three-step method for the isolation
of the most adequate host strain in a particular area has been
also described (73) and can be applied with appropriate host
strains (72). Recently, Bacteroides host strains for specific
detection of different animal species sources have been iso-
lated following this approach (74). Host strains for the detec-
tion and enumeration of phages infecting Bacteroides specific
for bovine (CW50 strain), porcine (PG76 strain), and poultry
(PL122) fecal pollutions were isolated. These host strains
allowed the detection of phages in 70–100% of samples of
fecal sources coinciding with their origin, with 0–20% mis-
classification of the nontarget fecal samples. However, in
this study host strains isolates from porcine wastewater were
also detecting Bacteroides phages in 50% of human sewage
samples, but in the worst case with ratios of somatic
coliphages:phages infecting porcine strain PG76 about 2 log
units lower than the ratio of somatic coliphages:phages
detected by human-specific strains. On the other hand, those
host strains detecting phages from other origins other than
their own showed always very low concentrations. Results
reported by Purnell et al. (214) with host strain for different
animals give results comparable for the hosts reported by
Gomez-Donate et al. (74) regarding both numbers and per-
centage of positive samples. New animal MST targets based
on bacteriophages infecting Bacteroides could be methodolog-
ically applied using the ratio of somatic coliphages to the Bac-
teroides host-specific phage as described for phages detected by
human-specific hosts (216). Such a combination has been
shown to be a powerful tool for resolving complex mixtures
of fecal-polluted water samples. They could facilitate a quan-
titative analysis of the mixture components and even help
discern the effects of dilution or die-off when applying
MST predictive models (236).
Studies on resistance to natural and anthropogenic stres-
sors (heat, UV radiation, and chemical disinfectants) have
shown that B. fragilis phages present higher resistance than
do traditional bacterial indicators and are similar to most
resistant viruses and other bacteriophages groups (62). This
observation is supported by the slow die-off of Bacteroides
phages (Table 2) in different types of water and water-related
matrices (62, 202).
Animal Viruses
Similar to their human counterparts, animal viruses have
been suggested as MST indicators. Most of them belong to
enteric viruses that are usually transmitted via the fecal–
oral route and that infect and replicate in the host gastrointes-
tinal tract. Enteric viral infections in animals are normally
asymptomatic but can result in mild to severe disease (237).
Additionally, nontraditional enteric viruses transmitted by
routes other than fecal–oral (i.e., polyomaviruses by urine–
oral route) have also been proposed as MST indicators. How-
ever, the adoption of traditional and nontraditional viral
groups as indicators has been questioned because they might
not fulfill some of the necessary features of an ideal fecal indi-
cator: high concentration in fecal materials, simple analytical
procedure, high survival rates in environment and treat-
ments, low risk to analyst, rapid results, and low cost (238).
Additionally, certain viruses present seasonal distributions
such as adenoviruses, which have been consistently detected
in river water during summer and fall months, although not
detected in the winter (239). This seasonality distribution
is an important limitation for their application as MST
indicators.
Cell culture has been the most widely used technique to
determine the presence of enteric viruses in environmental
samples. These methods are used not only to isolate the viral
agent but also to determine their infectivity. However, the
application of molecular techniques in the 1990s in MST
studies allowed for the development of presence/absence,
semiquantitative, and quantitative PCR-based methods.
Thus providing practical approaches for determining the
occurrence of different human and animal-specific viruses
in environmental samples (240, 241).
Data on the concentration of these proposed animal
viruses in the environment are limited. Moreover, the com-
parison of results from different studies is difficult due to dif-
ferent sample types, sample processing, and enumeration
techniques (82). Though most of the developed PCR-based
techniques for enumeration of animal viruses in aquatic envi-
ronments include control materials for standard curves, many
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-13
assays have been conducted without environmental matrix
controls for estimation of the performance and effectiveness
of the developed methodology. Therefore, the extrapolation
of results to estimate the concentration of viruses in water
samples may be unreliable. Other studies on bovine adenovi-
ruses (BAdVs) have been performed spiking environmental
samples with the plasmid vector of the cloned target which
was applied to prepare standards for qPCR, to estimate the
detection limits (80, 82). It was concluded that sensitivity
of the defined molecular assay was not affected or slightly
affected by the fecal matrix of samples. Differences in the
physical and chemical features of plasmids and viral particles
in environmental complex water matrices were not reported.
In contrast, assays on bovine enteroviruses (BEVs) have been
able to successfully establish the detection limit of the devel-
oped molecular methodologies by spiking environmental
samples with known amounts of BEV RNA or BEV PFU
and applying established protocols for concentration, extrac-
tion, and PCR-based enumeration. In this way one can deter-
mine the efficiency of the approach and later calculate
properly the concentration of BEV in fecal-polluted environ-
mental water or fecal samples (75, 76). These studies deter-
mined limits of detection around 103 molecular targets or
PFU/100 ml which are in agreement with other detection
limits of current molecular techniques for different microbial
indicators and pathogens (208).
Three main groups of viruses have been proposed as poten-
tial indicators MST (enterovirus, adenovirus, and polyomavi-
rus). Their main characteristics and developed methods for
application on MST studies are described below.
Enteroviruses. Enteroviruses are single-stranded RNA
viruses with an icosahedral capsid ranging from 20 to 30
nm in diameter. Mostly nonpolioviruses have been associated
with human infections, but around 30% are associated with
animal infections (77). Animal-specific enteroviruses usually
cause asymptomatic infections and are excreted in large num-
ber in feces of infected animals. Most enteroviruses are good
MST markers as they show a narrow host range, usually hav-
ing a strong preference for infection of one or a few related ani-
mal species. BEVs and porcine teschoviruses group (initially
classified as porcine enteroviruses), both belonging to the
family Picornaviridae, have been suggested for their applica-
tion as MST indicators (75–77). At present, two BEV sero-
types are recognized (242), serotype 1 and 2 (BEV-1 and
BEV-2, respectively). A reverse transcription (RT) PCR
amplification of BEV RNA from feces and water samples
has been developed (75). The sequences of primers were
taken from the 50 nontranslated region sequence of BEV
PS87 (GenBank accession no. X79368). The detection limit
of the RT-PCR was between 5 and 10 copies of the RNA
genome and was estimated by adding known concentrations
of BEV RNA to the RT-PCR mixture. Additionally, the
methods sensitivity in environmental samples was assessed
before samples were analyzed. Mixtures of known concentra-
tions of BEV-2 PFU were spiked into fecal extract coming
from free BEV environmental samples. Then the sensitivity
of the extraction plus RT-PCR amplification was between
0.5 and 2 viral PFUs in 14 µl fecal extract. Bovine, deer,
and goose fecal samples; runoff streams, standing water in pas-
tures, and watering tanks from bovine farms; and oyster sam-
ples downstream from farms were analyzed. A high percentage
(76%) of bovine samples showed positive results by RT-PCR,
indicating that BEV-2 was endemic in the studied region.
Because BEV is shed in large numbers by infected animals,
environmental water samples tested showed concentration
of BEV at least 1.4 × 104 PFU/liter by cell line analyses.
The detection of BEV-2 by RT-PCR in water and oyster
downstream samples, 50% of deer fecal samples, and one of
four goose fecal samples indicated a limited host specificity,
though the correlation between cytophatic effects on cell cul-
ture assays (Madin-Darby bovine kidney cells) and RT-PCR
results was poor for some samples. Further nucleotide
sequencing of the amplified products from positive samples
and phylogenetic analyses showed at least three groups of
sequences within the studied BEV population. It was deter-
mined that BEV was endemic in the bovine population and
the bovine feces contaminated the water surrounding the
farm and oysters collected downstream. Deer and goose sam-
ples were also clustered into phylogenetic groups of bovine
samples, suggesting limitations on host specificity. Further
studies are needed on BEV sequencing variability and host
specificity to select reliable bovine-specific molecular targets
for MST studies. Moreover, though the BEV-specific
RT-PCR could be useful to detect animal fecal pollution,
low concentrations of BEV in environmental samples
requires efficient concentration of water samples for a success-
ful detection.
A TaqMan real-time RT-PCR method was developed for
the detection of BEV using the BEV-2 strain PS87 (76). The
sensitivity of the TaqMan assay was able to detect 11.6 RNA
equivalent molecules per tube. Authors evaluated the effec-
tiveness of the whole procedure (concentration by
filtration-elution with electropositive filters, extraction, and
real-time RT-PCR) by spiking distilled water with viral
RNA stock from propagated BEV-2 PS87 in MDBK cell
line cultures. The detection limit for the whole protocol
was equal or greater than 12 BEV RNA molecules/ml of water
sample. Fecal samples (mostly from cows but also from goats,
sheep, donkeys, and horses that share pastures) and water
samples surrounding bovine farms from different areas were
taken. A high percentage of bovine fecal samples (78%)
showed positive results for the RT-PCR, with higher BEV
positive results for bovine samples coming from intensive
livestock farms. Phylogenetic studies of the amplified positive
BEV samples showed also high heterogeneity in bovine sam-
ples, but water and bovine samples were closely related. In this
study BEV-like sequences were also detected on equine
(60%), goats (60%), and ovine (36%) samples. Though the
developed molecular method proved to be useful for the
detection of animal fecal contamination, the used sequence
target allowed a moderate host specificity. More phylogenetic
studies on BEV (both groups 1 and 2) are needed to discern
narrow molecular host-associated targets which allow BEV
to be a suitable bovine MST indicator.
Porcine teschoviruses (PTVs), which were formerly classi-
fied as porcine enteroviruses, have been postulated as porcine
MST indicators (77). A TaqMan real-time RT-PCR was
developed for a specific and quantitative detection of PTVs
by defining primers and probe on the 50 noncoding region
of PTV-1 (GenBank accession no. AF231769). The detec-
tion limit of the developed methodology was determined by
using a standard PTV-1 strain, which was propagated in
PK-15 cell monolayers, and later its stock was used to spiked
water and fecal samples. The assay was able to detect approx-
imately 6.8 × 103 molecules of PTV RNA/ml of initial sam-
ple. Environmental samples (sewage, pig slurry, farm ponds)
were if necessary concentrated by filtration-elution through
electropositive filters. Feces from bovine, ovine, and goats
were also analyzed. Only fecal samples from porcine were pos-
itive for the PTV real-time RT-PCR. PTVs were detected at
104–105 viral particles/ml of the wastewater effluent from a
3.4.4-14 ▪ MICROBIAL SOURCE TRACKING
farm. The concentration of PTV RNA decreases up to 103
times in receiving stream water samples due to the dilution
effect, becoming undetectable 3,100 m downstream from
the spillage. The developed assay was shown to be useful for
the detection of porcine fecal pollution at least at the point
of release and immediately downstream in the receiving
water. The method needs to be tested for its usefulness in dif-
ferent geographical areas, where the prevalence of PTV in
porcine populations may diverge, attending to the endemicity
of the disease.
Adenoviruses. Adenoviruses are nonenveloped, double-
stranded DNA viruses ranging from 90 to 100 nm in diame-
ter. All adenoviruses associated with a human or animal
host belong to the genus Mastadenovirus. They have been
shown to be up to 60 times more resistant to UV irradiation
than are RNA viruses (243). Animal adenoviruses infect a
wide range of species, including mammals, birds, reptiles,
amphibians, and fish. Similarly to the proposal of human
adenoviruses as MST indicators for human fecal contamina-
tion, BAdVs and porcine adenoviruses (PAVs) have been
proposed respectively as bovine and porcine fecal contamina-
tion indicators (78).
Two nested PCRs were developed for the detection of viral
DNA of BAdV and PAV based on sequences of the hexon
gene (78). Quality control of amplification methods was
done by negative controls and sequencing the positive results.
No data were reported related to sensitivity of the methods.
Small sample sizes were taken (1 g of pooled feces from por-
cine or bovine, or 40 ml of urban sewage), and concentration
by centrifugation, elution, and DNA extraction were per-
formed. The number of detectable viral particles in porcine
and bovine feces was 101–103 and 101–104 genome equiva-
lents/g of pooled fecal sample, respectively. Though BAdV
was detected in 75% bovine fecal samples and PAV in 70%
porcine fecal samples, the selected targets were not detected
in any of the urban sewage analyzed. Phylogenetic studies
showed high variability of BAdV isolates indicating that sev-
eral serotypes may coexist in certain individuals thus limiting
the utility of BAdV as an MST indicator. In contrast, phylo-
genetic analysis showed a similarity between the PAV isolates,
suggesting the existence of a stable variant of a specific sero-
type in the geographical area. However, their sequences
showed minor similarity with the reference strains used in
this study, which is also hampering the use of the selected
molecular target as a universal porcine MST indicator.
Nested-PCR with BAdV and PAV primers was applied to
analyze two slaughterhouse wastewaters, one processing
both bovines and ovines and another mainly processing
porcine and less bovine (81). They reported 100% detection
of PAV in 10 bovine-porcine slaughterhouse wastewater
samples but less than 5% of positive results for BAdV when
analyzing both slaughterhouse wastewaters. No data on the
dilution effects and environmental persistence of BAdVs
and PAVs were provided, though seven out nine tested
river water samples (Ter and Llobregat Rivers, Catalonia,
Spain) were reported to be positive for PAV at concentrations
up to 101 genome equivalents/liter. Later, authors conti-
nued working on PAVs, developing a new TaqMan qPCR
targeting a 68-bp region of PAdV hexon gene (79). Standard
curves for the qPCR were done using transformed E. coli
JM109 cells containing a sequence of the PAV-3 hexon.
Pooled porcine fecal samples (1 g), porcine slaughterhouse
wastewater (40 ml), urban sewage (40 ml), and river water
(100 liters) samples were taken. Concentration and elution
of samples was performed as needed. Enzymatic inhibition
of the qPCR by environmental samples was checked by
running multiple dilutions of samples and adding known
amounts of target DNA to environmental samples. Though
the efficiency of the complete protocol (concentration, elu-
tion, extraction, and enumeration) was not performed, the
sensitivity of the qPCR assay was 1–10 genome copies in
4.2 ml of sewage, or 1 liter of river water, or 0.1 g of stools.
PAV was detected in 76–100% of porcine feces from different
areas and 100% of porcine slaughterhouse wastewater. The
concentration of PAV was 5.58–7.25 × 105 GC/g of feces
and 1.56 × 103 GC/ml.
A TaqMan qPCR for the detection of BAdV has been also
developed based on targeting the hexon gene of subgroup 2,
which includes BAdV serotypes 4–8 (80). To determine how
the environmental matrix would affect the sensitivity of the
developed qPCR, plasmid DNA carrying the cloned BAdV
hexon gene was also used to spike negative bovine feces at
four different levels of dilution. It was concluded the fecal
matrices had a slight effect on the sensitivity of the BAdV2
qPCR. BAdV was detected in all manure analyzed at concen-
trations around 103–104 GC/ml, and about 1–3 log units
lower concentrations in farm tile drainage samples. BAdV
was detected in 20% of the bovine feces analyzed.
All these studies support the need to study the geograph-
ical distribution of animal adenoviruses to determine their
relative abundance in target species related to the epidemio-
logical prevalence of the viruses and to assess the seasonality
usually associated with adenoviruses. The use of source feces
may not be feasible for defining universal host animal species
fecal indicators because of the need to study many samples
and the high variability observed between individual feces.
Therefore, the occurrence and concentrations of proposed
host-specific universal targets may be better studied in
heavily contaminated waters such as wastewater from slaugh-
terhouses and slurries emanating from at least 10 individual
farm animals (9).
Polyomaviruses. Polyomaviridae are a family of double-
stranded DNA viruses that infect humans and animals (birds,
rodents, bovine, nonhuman primates). They are highly host
specific as it has been reported for human-related wastewater
samples (244–248). Rangan et al. (249) reported a bovine
polyomaviruses-like (BPyV) as contaminant of macaque kid-
ney cell cultures. Later, BPyVs were proposed as a bovine fecal
source indicator in water (81). They developed a nested PCR
based on VP1 and agnoprotein regions which allowed detect-
ing BPyV in 21 out of the 22 bovine-ovine-porcine slaughter-
house wastewaters at 101 and 102 genome equivalents/ml of
sample. BPyVs were not detected in any of the urban sewage
samples analyzed. BPyVs were also detected in most of the
river water samples at concentrations ranging from 101 to
102 genome equivalents per 5 liters. BPyVs were found in dif-
ferent seasons of the year, which may support they do not
present seasonality, and certain heterogeneity was found in
the phylogenetic analyses of the isolates. Later, these authors
developed a TaqMan qPCR based on a 77-bp region of the
bovine polyomavirus VP1 gene (83). The sensitivity was
only estimated for the qPCR assay (1–10 plasmid copies of
standard material per test tube), and according to the volume
of water and fecal matter analyzed it correspond to 1 ml of
urine, 4 ml of sewage, and 4 ml of river water. BPyVs were
found in 30% of bovine urine samples at mean values of
104 GC/liter, 90% of bovine slaughterhouse wastewaters at
103 GC/liter, and 50 of river water samples at 102 GC/liter.
BPyVs were not detected on 10 studied samples of bovine
feces, and not on urban sewage where HPyVs were detected
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-15
in 100% of samples at values around 105 GC/liter. Wong and
Xagoraraki (82) compared the occurrence and concentra-
tions of BAdV and BPyV in bovine feces and manure
by using a similar TaqMan qPCR. BPyVs were around 106
GC/ml in 100% dairy manure samples and BAdV was at 1
log unit lower. Moreover, all studied manure samples showed
positive results for BPyV but not for BAdV. Consequently,
BPyV may be a better host-specific fecal source indicator
than BAdV. Moreover, both bovine viruses were compared
with bacterial indicators (Bacteroidetes, E. coli, and entero-
cocci). The concentrations of BAdV were significantly lower
that bacterial indicators in all manure samples, but BPyV
showed a significant positive correlation with the concentra-
tions of E. coli and enterococci with lower nonsignificant
BPyV concentrations.
All these studies suggest that BPyV could be a better
bovine source-tracking indicator than BAdV. However,
the environmental persistence of BPyVs is still unknown,
although they may be resistant to environmental degradation
as the human polyomavirus JCPyV had a T90 of 63.9 days
(250). Thus, studies on the resistance of BPyV to natural
inactivation processes should be done to assess its survival
in different environmental matrices. Additionally, more
information on the prevalence of infections and shedding
rates of animal polyomaviruses in widely distributed geo-
graphical areas are necessary. Environmental persistence,
prevalence of infections, and shedding rates are essential
data to assess the applicability of bovine polyomaviruses as
source-tracking indicators.
NEW MOLECULAR APPROACHES AND
FUTURE PERSPECTIVES: METAGENOMICS
Methods used to track fecal sources have shifted from culture-
dependent techniques to molecular assays in which sources
are identified directly from environmental water DNA
extracts. Most methods rely on conventional PCR or qPCR
techniques for the detection of host-specific markers. New
PCR-based methods are emerging, such as digital PCR, that
have the potential for low copy number discrimination, no
need for internal standards, simultaneous processing of multi-
ple host-specific assays, and higher throughput than conven-
tional PCR methods (251). While digital PCR methods have
yet to be used in MST applications, this will certainly change
in the near future.
Using 16S rRNA gene microarrays, Andersen and col-
leagues generated community profiles and identified signa-
ture sequences associated with different animals (252). This
approach is different from most current MST approaches
because it uses multiple bacterial populations and multiple
regions of the 16S rRNA gene to determine the presence of
fecal sources. It is possible to determine the presence of multi-
ple fecal sources in a sample using microarrays, although it is
difficult to accurately quantify fecal loads as they rely on an
initial amplification step with universal PCR primers prior
to the hybridization step. Additionally, as the current proto-
type has more than 500,000 targets per microarray, there are
computational and bioinformatics needs associated with this
approach that are not commonly available in most environ-
mental microbiology laboratories. However, this approach is
useful at identifying a wide diversity of bacterial and archaeal
16S rRNA targets per source, information that can then be
used to develop novel PCR-based assays or microarray-based
MST assays with a limited number of microbial targets.
Recent innovations in sequencing technologies have
resulted in better microbial characterization of different
natural matrices, including different host feces (253), waste-
water treatment plants (254), and aquatic environments
(255). In fact, it is now possible to generate millions of
sequencing reads per sample using sequencing approaches
developed by new technologies (256). As a result, the detec-
tion of sequences associated with rare members, such as host-
specific populations and pathogens, can theoretically be
achieved. The primary focus in MST related studies have
been on deeper sequencing of 16S rRNA gene libraries
coupled with barcoded primers and on MST studies focusing
on human fecal sources. It is likely that water samples
impacted with other fecal sources will be soon examined
using this approach it is possible to generate sequencing infor-
mation for virtually hundreds of samples per experiment for
considerably less money than conventional Sanger chemistry
approaches. Altogether, this provides an opportunity to
detect targets for novel host-specific assays when using uni-
versal eubacterial primers or to generate community profiles
similar to the microarray approach. When coupled with
group-specific primers (e.g., Bacteroidales) it may provide
the means for assessing the relative abundance of host-specific
bacterial populations within watersheds impacted by multiple
sources. Like microarrays, the computational and informatic
needs associated with next-generation sequencing (NGS)
technologies have limited this approach to a small number
of MST laboratories. However, recently available software
algorithms can be used with large pyrosequencing databases
developed for source tracking purposes (257, 258). As the
number of laboratories using NGS increases and the infor-
matic tools are more accessible, it is reasonable to assume
that sequencing will be an even more important tool in future
MST studies.
For the most part, MST methods have relied on 16S rRNA
gene sequences. One approach recently used to discover host-
specific markers focused on using a metagenomic approach
coupled with the positive selection of DNA sequences to
determine the presence of functional genes that might be
involved in host–microbial interactions (55). This approach
has been used to develop human-, cattle-, and chicken-spe-
cific assays (113, 259), which have been applied in a few
watershed studies (260). While function-based assays might
not have the same detection limit of assays targeting multi-
copy genes, these can be used to further complement the
16S rRNA gene-based assays due to their high level of host
specificity. Metagenomic approaches that couple DNA
hybridization methods promise to add information on the
mechanisms of host specificity and on the prevalence of
markers in environmental samples.
Metagenomic sequencing has also been used to study
fecal, clinical, and environmental viromes (261–263). This
type of molecular survey has provided information that is crit-
ical for better understanding the diversity of unculturable or
difficult to culture enteric human viruses ( pathogens). In
fact, Svraka et al. (264) were able to identify the DNA and
RNA viruses in samples for which viral agents were previously
unidentified. The simultaneous identification of diverse
phages present in fecal samples can also be performed using
metagenomic sequencing (265). As some phages can be
used as surrogate targets for human virus detection, phage
metagenomes can effectively be used in fate and transport
studies in which direct detection of enteric viruses is chal-
lenging. Viral metagenome studies can also reveal the viral
diversity dynamics associated with water treatment processes.
Recently, Tamaki et al. (266) studied the presence of DNA
viruses in a wastewater treatment plant in Singapore and
found that most viruses in these environments are novel.
3.4.4-16 ▪ MICROBIAL SOURCE TRACKING
Similar studies targeting RNA viruses are needed as they rep-
resent an important group of human pathogens.
Human viruses that survive water treatment can become
relevant environmental targets in source tracking studies.
However, not all sewage-specific assays need to target human
viruses. For example, viral metagenome sequencing was used
to identify a plant virus (i.e., pepper mild mottle virus) sug-
gested to be a potential indicator of human sewage due to
its widespread occurrence and abundance in wastewater
(267). This approach has yet to be used in the detection of
animal markers but promises to be relevant if they might cor-
relate with the host dietary regime. In summary, in the near
future, viral metagenomics will have relevant implications
not only in source tracking but also in many other areas asso-
ciated with public health and microbial risk assessment.
Although metagenomic studies are becoming increasingly
popular, genome sequencing will also be very important for
the advancement of microbial water quality and source track-
ing. Currently, the genomes of fecal bacteria such as E. coli,
Bacteriodes spp., and Enterococcus spp. have been sequenced,
and many of their genes have been annotated (268–274).
Genome information is critical to determine the presence
of fecal bacterial genes in metagenomes, to develop detection
methods based on species-specific single-copy genes, and to
better understand genetic mechanisms relevant to environ-
mental survival and host specificity. Thus far, only a few
genomes of host-specific populations targeted in source
tracking studies have been completed, specifically C. mari-
mammalium (173), Methanobrevibacter smithii (275), and
Cryptosporidium parvum (276). Some of the host-specific pop-
ulations have yet to be cultured, precluding genome sequenc-
ing. Thus, future efforts should be focused on isolating such
populations as pure cultures, or alternatively, using single-cell
genome sequencing (277, 278).
IMPLEMENTATION IN ROUTINE MST
ANALYSES: RELATIONSHIP TO TRADITIONAL
FIB AND PATHOGENS
Role of MST in TMDL Development
The value of MST assays is unquestionably linked to their
accuracy at identifying the primary fecal sources impacting
environmental waters (279). From the U.S. perspective, ideal
MST assays must be useful at determining fecal loadings as per
the U.S. EPA’s total maximum daily load (TMDL) regula-
tions and correlate with the detection of human pathogens
or less ideally FIB. To date, 13 states have used MST methods
at some stage of TMDL development (e.g., source assessment,
load partitioning, or model calibration) (280–293). Addi-
tionally several states have developed protocols on the use
of MST in TMDL development (294, 295). Future use of
MST in the TMDL process may include the TMDL imple-
mentation planning and monitoring of load reductions
once further studies are conducted to understand the fate
and transport of these MST markers and their correlation
with pathogens. Additionally studies are needed to improve
the sampling theory for estimating fecal source inputs at
desired levels of confidence.
Currently there are no standard methods (e.g., EPA,
American Society for Testing and Materials, International
Organization for Standardization) for the detection of MST
markers. Therefore interpretation of the MST results with
respect to existing FIB-based water quality criteria is diffi-
cult—not only because current FIB-based criteria are culture
based while MST methods target specific genes by qPCR, but
also because MST methods target specific organisms whereas
FIB represents much larger groups of microorganisms which
may or may not include FIB. Additionally, most MST meth-
ods have been tested in only a handful of watersheds over lim-
ited temporal scales and environmental conditions, reducing
confidence in MST marker utility in other watersheds.
Utilizing Multiple Methods for Increased
Confidence in MST
The use of multiple molecular or chemical assays per host
should be favored over single assays (9, 296, 297). Targeting
multiple bacterial groups and/or species should be considered
instead of targeting the same bacterial species unless different
genes are used as the target. Overall, the approach of using
multiple assays per host theoretically increases the confidence
that a particular source(s) is present in a water body and
should aid in assessing its overall relative importance.
Undoubtedly, this will complicate the analysis and the cost
of MST studies, but should increase the confidence of identi-
fying a source and its overall importance in fecal loadings.
Additionally, MST methods should be just one tool used in
a weight-of-evidence approach to identify sources of contam-
ination in a watershed. For example, MST data could be com-
bined with GIS, sanitary surveys, and traditional water quality
monitoring for pathogens, FIB, or other parameters (298).
Similar to human fecal pollution, tracking fecal pollution
with the currently available animal-specific markers is an
intense area of research. Some assays have been tested by mul-
tiple laboratories and evaluated using samples from multiple
geographic locations. However, most animal assays have
been used by few laboratories and in limited environmental
applications. The sources could be associated with animal
fecal pollution could vary: from feces (i.e., when animals
have direct access to watersheds) to spread manure, over-
flowed waste lagoons, and waste from animal operations not
under compliance. Human sources can be linked to
swimmers, but are mostly limited to wastewater treatment
plants and leaky septic tanks.
Direct access to the watershed might explain more fre-
quent detection of fecal pollution from domesticated animals,
but this might be one of several factors for the poor perform-
ance of a marker when used against environmental waters. For
example, little is known on the survival of the targeted bacte-
rial species during the manuring process, in manured soils,
and once the animal source identifiers reach environmental
waters. In the future, evaluation of these markers should
include climatological and hydrological factors, which may
better explain their occurrence in environmental waters.
Quality Assurance and Quality Control for
Routine MST
Quality assurance (QA) and quality control (QC) issues
associated with MST applications can be grouped into two
categories: (a) methodological QA/QC during assay develop-
ment, and (b) routine sample collection and handling QA/
QC during assay implementation. Currently, there are no uni-
versally accepted performance criteria for MST indicator
assays with respect to specificity of methods, method sensitiv-
ity (i.e., percentage cross reactivity with nontarget feces), and
method detection limits. MST signals can be reported in
different ways such as cell equivalent and gene copies per
reaction. Moreover, MST signals have been be normalized
using different approaches, namely, per volume of DNA
extracts used or by normalizing for the amount of DNA
template. These different approaches preclude a direct
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-17
comparison of MST methods between different laboratories
and across different geographic locations.
During assay development and prior to the use of MST
assays in routine monitoring, it is necessary to determine
the assay’s detection limits, host distribution, and host specif-
icity, preferentially with fecal samples collected from areas
nearby the body of water under study. As most assays are
not 100% host specific, measuring the relative abundance
of the MST indicators against samples from targeted and non-
targeted hosts is also important. This information is useful to
rule out the contribution of other sources that might cross-
react with a given assay. Increasingly, Bayes’s theorem has
been used to determine the conditional probability that an
MST assay can correctly detect an MST indicator in environ-
mental samples (42, 58, 71, 125, 127).
General good laboratory practice should be followed dur-
ing the assay development and during utilization in routine
monitoring studies. For example, control samples should be
utilized along the entire MST process and should include
sample collection and handling controls (field blanks and
contamination controls), DNA extraction efficiency controls
(spike and recovery controls; 299), negative DNA extraction
controls, and qPCR analysis positive and negative controls
(PCR inhibition controls, amplification positive control).
Standards should be included within each PCR plate, and
dilutions of each DNA extract should be used to more accu-
rately determine the relative abundance of the markers in
each water and fecal sample tested. This approach will also
address the level of PCR inhibition within a sample. DNA
extracted from fecal sources and genomic DNA from targeted
bacterial populations can be used as positive controls; how-
ever, genomic DNA might not be suitable to calculate gene
copies. Instead, plasmids ( preferentially linearized) contain-
ing the targeted gene sequence should be used to calculate
gene copies. Although theoretically possible to use gene cop-
ies to calculate genome or cell equivalents, this is not recom-
mended for a number of reasons. Utilization of TaqMan or
SYBR-based qPCR assays requires different analysis and
data reduction techniques. For example, in SYBR-based
assays the presence of double peaks needs to be verified for
each sample by looking at the melting curve plots. Signals
with double peaks should be reported as presence/absence
data and not used to calculate signal intensity.
After method development and validation in the labora-
tory or in a small watershed study, further QA/QC issues arise
during the implementation of the methods in the field. For
example, there is a lack of standardization of sample collec-
tion and handling methods. It is expected that the number
and location of sampling sites within a watershed will affect
the ability of MST studies to accurately determine the source
and location of fecal inputs. For accurate watershed assess-
ment with MST approaches, statistical methods should be
used to determine the minimum number of samples required
to determine the source distribution within a watershed at a
desired confidence level. Other factors that could influence
the variability between samples during MST sampling should
also be taken into account. For example, seasonal changes in
temperature, base flow, storm characteristics, landscape, and
animal management practices in different seasons could influ-
ence the variability between samples.
Challenges to Incorporating MST in Routine
Monitoring
Two major limitations to incorporating MST in routine mon-
itoring, in addition to the challenges listed already, include
the equipment and technical expertise required for collec-
tion, sample analysis, and data evaluation involved in MST
studies. Currently, there are a few commercial laborato-
ries available for routine MST sample analysis. While MST
methods could potentially be used for rapid turnaround mon-
itoring of environmental waters to determine if beach closure
advisories are necessary, there are few laboratories available to
perform these rapid analyses. Turnaround times for qPCR
analyses may be as little as 6 h; however, if analytical labora-
tories are not located geographically close to the site of sample
collection, an additional 12–18 h may be required for ship-
ment of samples. Finally, standardization of reports from ana-
lytical laboratories has not been established, as has been seen
in the chemical analysis fields (e.g., gas chromatography). At
a minimum, laboratory reports from commercial analytical or
academic laboratories performing MST assays should include
the information suggested in the “Minimum Information for
Publication of Quantitative Real-Time PCR Experiments”
guidance (300) and laboratories should follow the EPA guid-
ance “Quality Assurance/Quality Control Guidance for
Laboratories Performing PCR Analyses on Environmental
Samples” (EPA 815-B-04-001, 2004).
DISCUSSION
Although initially MST studies focused on developing indi-
cators to differentiate human from nonhuman fecal pollution,
research began to propose indicators to detect fecal sources of
specific animal species. Identifying all potential fecal sources
of pollution in a watershed is necessary for development of the
most effective catchment management and remediation strat-
egies. Currently, there is a wide variety of chemical-based
methods as well as culture-dependent and culture-independ-
ent methods that have been proposed to track animal fecal
sources on. No single method can be used universally due
to the many issues already discussed, and as a result a combi-
nation of methods is recommended to increase the confi-
dence of identifying the primary fecal sources.
Each method has its own advantages and disadvantages.
The major advantages that chemical indicators exhibit com-
pared to biological MST indicators are that they are not
subject to possible regrowth in the environment and the
available analytical methods are very sensitive. However,
the previously proposed sterols are not enough specific to a
single fecal source. Ratios of different sterols may more accu-
rately identify various animal sources, although additional
confirmation by other MST indicators is recommended for
further identification of sources of fecal contamination. An
important concern is that there are no chromatographic
and mass spectrometry standard methods for common detec-
tion of fecal sterols and bile acids. Additional issues limiting
the utility of chemical indicators include the potential for
biodegradation once in environment, relative low knowledge
level of their environmental persistence, the potential for
natural background concentrations in waters, their lack of
correlation with pathogens, and the high cost and the need
of qualified personnel to perform their analysis.
MST molecular markers based in mtDNA show a high
host specificity and a large number of targets in the cells of
the host. On the other hand, false positives have been
observed which could be related to the accumulation and
mixing of mitochondrial targets through the formulation of
feeds or composition of foods. The recent development of a
mitochondrial microarray (301) has been suggested to cir-
cumvent some of the problems associated with current mito-
chondrial assays.
3.4.4-18 ▪ MICROBIAL SOURCE TRACKING
Several bacterial markers that are animal-host associated
have been proposed. Among these are members of the Bacter-
oidales family and the Bifidobacterium genus. However, the
lack of standardized protocols for their detection can lead to
variable results between laboratories due to the type of
reagents, equipment, and amplification conditions utilized.
Clearly, the development of standard protocols and interla-
boratory validation studies of proposed assays is necessary
for assuring reproducibility and therefore, future implementa-
tion of bacterial host-specific assays (302). In addition, DNA
processing methodologies may include bias in the sample
concentration, DNA extraction, DNA purification, and
amplification, thus controls such as spike and recovery should
be considered. Several studies related to the application of
bacterial markers have applied efficiency controls for sample
concentration and DNA extraction (43, 299, 303) and to
detect the presence of PCR inhibitors (internal amplification
control) (56, 304). While the use of internal amplification
controls can identify amplification competition problems
with the target DNA, alternatively, dilution of the samples
represents a reliable and simple alternative to reduce inhibi-
tors (305). However, the sample dilution to limit PCR inhib-
ition may lead to false negative results.
Most qPCR-based methods require concentrations around
103 GC/100 ml of water for positive identification of target
genes in an environmental sample. Normally, processed water
volumes may vary between 100 and 1000 ml filtered through
polycarbonate or nitrocellulose membranes with a pore size of
0.22–0.45 µm. DNA is extracted from the membranes using
different commercial kits depending on laboratory prefer-
ence. The final step is the DNA amplification and visualiza-
tion of PCR products either by end-point or real-time qPCR.
DNA amplification protocols have been defined by every
method development laboratory (Table 1) and include varia-
tions regarding the PCR chemistry and equipment used. In
case of qPCR, some methods rely on PCR product detection
via SYBR green assays or TaqMan assays. It should be noted
that one important disadvantage of using SYBR green–based
assays is that the presence of double peaks limits their use as
presence/absence assays rather than quantitative assays.
MST animal indicators have been developed based on
molecular targets associated to certain animal viruses with
high host specificity. The detection and the enumeration of
animal viruses as MST indicators depend on their prevalence
in animal or human populations (fraction of the population
that at some particular time has a given disease). This preva-
lence is related to the disease occurrence (epidemiology)
attending to the geographical and demographical distribu-
tion. Consequently, unlike commensal gut microbial popula-
tions proposed as MST indicators (e.g., Bacteroidetes and
Bifidobacterium), the presence and concentration of viral
pathogens in water depend on the epidemiology of the disease
in the region under study. They are usually found in low con-
centrations in the environment, which could result in levels
below the technique’s detection limits (306). Only those
viruses at higher levels in fecal sources could have a chance
of being detected in environmental waters without significant
concentration of samples. Accordingly, large volumes of
water (10–100 liters) need to be concentrated (i.e., via
filtration-elution, filtration-organic flocculation, ultracentri-
fugation, and elution) for MST studies before enumeration
by cultured host cells or molecular techniques are applied
(76, 77, 81, 82). However, the concentration of large
volumes of environmental water often increases the concen-
tration of PCR inhibitors, and therefore water DNA extract
would require serial dilutions to relive such inhibitors.
Consequently, the type of water sample or environmental
matrix under evaluation is an important consideration in
deciding the volume of sample to be analyzed. While pristine
waters could be sampled and concentrated using hundreds of
liters, smaller volumes (some milliliters up to several liters) of
storm water, runoff, wastewater, or similar highly eutrophic
water should be sampled to avoid concentration of PCR
inhibitors (237). The major disadvantage of using viruses as
MST targets is their low concentration in the environment
which could result in concentrations below the detection lim-
its of the available molecular tools. Despite these limitations
and challenges, animal viruses offer high levels of host
specificity.
In many regards, the concept of fecal source tracking con-
tinues to be a moving target. MST evolved from a general
observation that some conventional fecal bacterial indicators
were more prevalent in human than in animal feces, to the
current PCR-based methods that detect host-specific popula-
tions directly from the water samples without requiring culti-
vation. Undoubtedly, biotechnological advances will shape
MST and the methods that are currently favored will be
replaced with others that are faster and that can simultane-
ously detect multiple targets. As funding is needed to pursue
MST studies and improve current approaches/microbial tar-
gets, some important issues remain and therefore must be con-
sidered. Some of these issues and research needs are
summarized below.
ADOPTION OF STANDARD ANALYTICAL
METHODS
• Reduction or quantification of MST uncertainty using
Bayes’s theorem.
• Adoption of standard sampling approaches and statistical
support of sampling approach.
• Little is known on the bacterial diversity and ecology of
most MST microbial targets.
• Assays have not been tested across different geographic
locations. In fact, most methods have been developed in
temperate regions and their performance is vastly
unknown in the tropics.
• Most MST has been developed and tested in the United
States, although the number of studies in Europe, Aus-
tralia, Japan, and Korea are increasing.
• Most studies do not conduct sanitary surveys.
• Most studies test a limited number of fecal samples.
• Most studies do not have an adequate temporal
component.
• Most studies do not incorporate landscape as part of the
performance evaluation.
• Compared to domesticated animals, there are fewer assays
for wildlife and waterfowl.
• There is very little information on the predictive value of
MST assays as far as identifying the sources of pathogens,
particularly on zoonotic assays. Thus epidemiology studies
are needed to quantify the relationship between the den-
sity of indicators of animal fecal contamination in waters
and risk to humans from exposure to zoonotic diseases.
In summary, only proposed animal-host MST indicators with
high specificity and sensitivity that are present at high con-
centrations at fecal pollution should be considered for MST
studies. Standardization of methods is an essential step to
resolve variability in results between analysts and laboratories.
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-19
This method standardization will facilitate the validation
of these animal MST indicators through interlaboratory stud-
ies. Methods of sample pretreatment (biomass concentration,
DNA extraction, DNA elution) should also be improved and
considered when estimating the environmental detection
limits of a methodology to avoid confusion with sensitivity
of the applied technique. Environmental persistence of
MST markers is also a key issue to resolve complex situations
when there are old fecal pollution contributions in waters.
The combination of traditional water microbial indicators
and new animal-host specific indicators are necessary to
resolve mixtures of fecal sources and properly correlate indica-
tors with pathogens presence in fecally impacted water.
REFERENCES
1. Schoen M, Ashbolt N. 2010. Assessing pathogen risk to
swimmers at non-sewage impacted recreational beaches.
Environ Sci Technol 44:2286–2291.
2. Schoen M, Soller J, Ashbolt N. 2011. Evaluating the
importance of faecal sources in human-impacted waters.
Water Res 45:2670–2680.
3. Morse S, Mazet J, Woolhouse M, Parrish C, Carroll D,
Karesh W, Zambrana-Torrelio C, Lipkin W, Daszak P.
2012. Prediction and prevention of the next pandemic zoo-
nosis. Lancet 380:1956–1965.
4. Hagedorn C, Weisberg S. 2009. Chemical-based fecal
source tracking methods: current status and guidelines for
evaluation. Rev Environ Sci Biotechnol 8:275–287.
5. Leeming R, Ball A, Ashbolt N, Nichols P. 1996. Using
faecal sterols from humans and animals to distinguish faecal
pollution in receiving waters. Water Res 30:2893–2900.
6. Jaffrezic A, Jardé E, Pourcher A-M, Gourmelon M, Cap-
rais M-P, Heddadj D, Cottinet P, Bilal M, Derrien M,
Marti R, Mieszkin S. 2011. Microbial and chemical
markers: runoff transfer in animal manure–amended soils.
J Environ Qual 40:959–968.
7. Tyagi P, Edwards D, Coyne M. 2008. Use of serol and bile
acid biomarkers to identify domesticated animal sources of
fecal pollution. Water Air Soil Poll 187:263–274.
8. Tyagi P, Edwards D, Coyne M. 2009. Fecal sterol and bile
acid biomarkers: runoff concentrations in animal waste-
amended pastures. Water Air Soil Poll 198:45–54.
9. Blanch AR, Belanche-Munoz L, Bonjoch X, Ebdon J,
Gantzer C, Lucena F, Ottoson J, Kourtis C, Iversen A,
Kuhn I, Moce L, Muniesa M, Schwartzbrod J, Skraber
S, Papageorgiou GT, Taylor H, Wallis J, Jofre J. 2006.
Integrated analysis of established and novel microbial and
chemical methods for microbial source tracking. Appl Envi-
ron Microbiol 72:5915–5926.
10. Furtula V, Osachoff H, Derksen G, Juahir H, Colodey A,
Chambers P. 2012. Inorganic nitrogen, sterols and bacte-
rial source tracking as tools to characterize water quality
and possible contamination sources in surface water. Water
Res 46:1079–1092.
11. Jardé E, Gruau G, Mansuy-Huault L. 2007. Detection of
manure-derived organic compounds in rivers draining agri-
cultural areas of intensive manure spreading. Appl Geochem
22:1814–1824.
12. Standley LJ, Kaplan LA, Smith D. 2000. Molecular tracers
of organic matter sources to surface water resources. Environ
Sci Technol 34:3124–3130.
13. Noblet JA, Young DL, Zeng EY, Ensari S. 2004. Use of
fecal steroids to infer the sources of fecal Indicator bacteria
in the lower Santa Ana river watershed, California: sewage
is unlikely a significant source. Environ Sci Technol 38:
6002–6008.
14. Isobe KO, Tarao M, Zakaria MP, Chiem NH, Minh LY,
Takada H. 2002. Quantitative application of fecal sterols
using gas chromatography–mass spectrometry to investi-
gate fecal pollution in tropical waters: western Malaysia
and Mekong Delta, Vietnam. Environ Sci Technol 36:
4497–4507.
15. Leeming R, Nichols PD. 1996. Concentrations of copro-
stanol that correspond to existing bacterial indicator guide-
line limits. Water Res 30:2997–3006.
16. Gilpin BJ, James T, Nourozi F, Saunders D, Scholes P,
Savill MG. 2003. The use of chemical and molecular
microbial indicators for faecal source identification. Water
Sci Technol 47:39–43.
17. Evershed Richard P, Bethell Philip H. 1996. Application
of multimolecular biomarker techniques to the identi-
fication of fecal material in archaeological soils and
sediments, p. 157–172. In Archaeological Chemistry. ACS
Symposium Series, vol. 625. American Chemical Society,
Washington, DC.
18. Simpson IA, van Bergen PF, Perret V, Elhmmali MM,
Roberts DJ, Evershed RP. 1999. Lipid biomarkers of man-
uring practice in relict anthropogenic soils. Holocene 9:
223–229.
19. Seurinck S, Verstraete W, Siciliano S. 2005. Microbial
source tracking for identification of fecal pollution. Rev
Environ Sci Biotechnol 4:19–37.
20. Elhmmali MM, Roberts DJ, Evershed RP. 1999. Com-
bined analysis of bile acids and sterols/atanols from riverine
particulates to assess sewage discharges and other fecal sour-
ces. Environ Sci Technol 34:39–46.
21. Elhmmali MM, Roberts DJ, Evershed RP. 1997. Bile acids
as a new class of sewage pollution indicator. Environ Sci
Technol 31:3663–3668.
22. Bilal M, Jaffrezic A, Dudal Y, Le Guillou C, Menasseri S,
Walter C. 2010. Discrimination of farm waste contamina-
tion by fluorescence spectroscopy coupled with multivari-
ate analysis during a biodegradation study. J Agric Food
Chem 58:3093–3100.
23. Baker A. 2002. Fluorescence properties of some farm
wastes: implications for water quality monitoring. Water
Res 36:189–195.
24. Düreth S, Herrmann R, Pecher K. 1986. Tracing
faecal pollution by coprostanol and intestinal bacteria
in an ice-covered finnish lake loaded with both
industrial and domestic sewage. Water Air Soil Poll 28:
131–149.
25. Caldwell JM, Payment P, Villemur R. 2011. Mitochon-
drial DNA as source tracking markers of fecal contamina-
tion, p. 642. In Hagedorn C, Blanch AR, Harwood VJ
(eds), Microbial Source Tracking: Methods, Applications and
Case Studies. Springer, New York.
26. Martellini A, Payment P, Villemur R. 2005. Use of
eukaryotic mitochondrial DNA to differentiate human,
bovine, porcine and ovine sources in fecally contaminated
surface water. Water Res 39:541–548.
27. Balleste E, Bonjoch X, Belanche LA, Blanch AR. 2010.
Molecular indicators used in the development of predictive
models for microbial source tracking. Appl Environ Micro-
biol 76:1789–1795.
28. Caldwell JM, Levine JF. 2009. Domestic wastewater influ-
ent profiling using mitochondrial real-time PCR for source
tracking animal contamination. J Microbiol Meth 77:
17–22.
29. Caldwell JM, Raley ME, Levine JF. 2007. Mitochondrial
multiplex real-time PCR as a source tracking method in
fecal-contaminated effluents. Environ Sci Technol 41:
3277–3283.
30. Schill WB, Mathes MV. 2008. Real-time PCR detection
and quantification of nine potential sources of fecal
3.4.4-20 ▪ MICROBIAL SOURCE TRACKING
contamination by analysis of mitochondrial cytochrome b
targets. Environ Sci Technol 42:5229–5234.
31. Kortbaoui R, Locas A, Imbeau M, Payment P, Villemur
R. 2009. Universal mitochondrial PCR combined with
species-specific dot-blot assay as a source-tracking method
of human, bovine, chicken, ovine, and porcine in fecal-
contaminated surface water. Water Res 43:2002–2010.
32. Baker-Austin C, Morris J, Lowther JA, Rangdale R, Lees
DN. 2009. Rapid identification and differentiation of agri-
cultural faecal contamination sources using multiplex
PCR. Lett Appl Microbiol 49:529–532.
33. Baker-Austin C, Rangdale R, Lowther J, Lees D. 2010.
Application of mitochondrial DNA analysis for microbial
source tracking purposes in shellfish harvesting waters.
Water Sci Technol 61:1–7.
34. Bernhard AE, Field KG. 2000. A PCR assay To discrimi-
nate human and ruminant feces on the basis of host differ-
ences in Bacteroides-Prevotella genes encoding 16S rRNA.
Appl Environ Microbiol 66:4571–4574.
35. Ahmed W, Powell D, Goonetilleke A, Gardner T.
2008. Detection and source identification of faecal
pollution in non-sewered catchment by means of host-
specific molecular markers. Water Sci Technol 58:
579–586.
36. Shanks OC, White K, Kelty CA, Hayes S, Sivaganesan
M, Jenkins M, Varma M, Haugland RA. 2010. Perform-
ance assessment PCR-based assays targeting bacteroidales
genetic markers of bovine fecal pollution. Appl Environ
Microbiol 76:1359–1366.
37. Balleste E, Blanch AR. 2010. Persistence of Bacteroides
species populations in a river as measured by molecular
and culture techniques. Appl Environ Microbiol 76:
7608–7616.
38. Walters SP, Field KG. 2009. Survival and persistence
of human and ruminant-specific faecal Bacteroidales in
freshwater microcosms. Environ Microbiol 11:1410–1421.
39. Fremaux B, Gritzfeld J, Boa T, Yost CK. 2009. Evaluation
of host-specific Bacteroidales 16S rRNA gene markers as a
complementary tool for detecting fecal pollution in a prai-
rie watershed. Water Res 43:4838–4849.
40. Gawler AH, Beecher JE, Brandao J, Carroll NM, Falcao
L, Gourmelon M, Masterson B, Nunes B, Porter J, Rince
A, Rodrigues R, Thorp M, Walters JM, Meijer WG.
2007. Validation of host-specific Bacteriodales 16S rRNA
genes as markers to determine the origin of faecal pollution
in Atlantic Rim countries of the European Union. Water
Res 41:3780–3784.
41. Gourmelon M, Caprais MP, Segura R, Le Mennec C,
Lozach S, Piriou JY, Rince A. 2007. Evaluation of two
library-independent microbial source tracking methods to
identify sources of fecal contamination in French estuaries.
Appl Environ Microbiol 73:4857–4866.
42. Kildare BJ, Leutenegger CM, McSwain BS, Bambic DG,
Rajal VB, Wuertz S. 2007. 16S rRNA-based assays for
quantitative detection of universal, human-, cow-, and
dog-specific fecal Bacteroidales: a Bayesian approach. Water
Res 41:3701–3715.
43. Silkie SS, Nelson KL. 2009. Concentrations of host-
specific and generic fecal markers measured by quantitative
PCR in raw sewage and fresh animal feces. Water Res 43:
4860–4871.
44. Bae S, Wuertz S. 2009. Rapid decay of host-specific
fecal Bacteroidales cells in seawater as measured by quantita-
tive PCR with propidium monoazide. Water Res 43:
4850–4859.
45. Bae S, Wuertz S. 2012. Survival of host-associated bacter-
oidales cells and their relationship with Enterococcus spp.,
Campylobacter jejuni, Salmonella enterica serovar Typhimu-
rium, and adenovirus in freshwater microcosms as measured
by propidium monoazide-quantitative PCR. Appl Environ
Microbiol 78:922–932.
46. Reischer GH, Kasper DC, Steinborn R, Mach RL, Farn-
leitner AH. 2006. Quantitative PCR method for sensitive
detection of ruminant fecal pollution in freshwater and
evaluation of this method in alpine karstic regions. Appl
Environ Microbiol 72:5610–5614.
47. Mieszkin S, Furet JP, Corthier G, Gourmelon M. 2009.
Estimation of pig fecal contamination in a river catchment
by real-time PCR using two pig-specific Bacteroidales 16S
rRNA genetic markers. Appl Environ Microbiol 75:
3045–3054.
48. Tambalo DD, Fremaux B, Boa T, Yost CK. 2012. Persis-
tence of host-associated Bacteroidales gene markers and
their quantitative detection in an urban and agricultural
mixed prairie watershed. Water Res 46:2891–2904.
49. Sokolova E, Astrom J, Pettersson TJ, Bergstedt O, Her-
mansson M. 2012. Decay of Bacteroidales genetic markers
in relation to traditional fecal indicators for water quality
modeling of drinking water sources. Environ Sci Technol
46:892–900.
50. Mieszkin S, Yala JF, Joubrel R, Gourmelon M. 2010. Phy-
logenetic analysis of Bacteroidales 16S rRNA gene sequen-
ces from human and animal effluents and assessment of
ruminant faecal pollution by real-time PCR. J Appl Micro-
biol 108:974–984.
51. Stricker AR, Wilhartitz I, Farnleitner AH, Mach RL.
2008. Development of a Scorpion probe-based real-
time PCR for the sensitive quantification of Bacteroides
sp. ribosomal DNA from human and cattle origin and
evaluation in spring water matrices. Microbiol Res 163:
140–147.
52. Okabe S, Shimazu Y. 2007. Persistence of host-specific
Bacteroides-Prevotella 16S rRNA genetic markers in envi-
ronmental waters: effects of temperature and salinity.
Appl Microbiol Biotechnol 76:935–944.
53. Jeong JY, Park HD, Lee KH, Hwang JH, Ka JO. 2010.
Quantitative analysis of human and cow-specific 16S
rRNA gene markers for assessment of fecal pollution in
river waters by real-time PCR. J Microbiol Biotechnol 20:
245–253.
54. Layton A, McKay L, Williams D, Garrett V, Gentry R,
Sayler G. 2006. Development of Bacteroides 16S rRNA
gene TaqMan-based real-time PCR assays for estimation
of total, human, and bovine fecal pollution in water. Appl
Environ Microbiol 72:4214–4224.
55. Shanks OC, Santo Domingo JW, Lamendella R, Kelty
CA, Graham JE. 2006. Competitive metagenomic DNA
hybridization identifies host-specific microbial genetic
markers in cow fecal samples. Appl Environ Microbiol 72:
4054–4060.
56. Shanks OC, Atikovic E, Blackwood AD, Lu J, Noble RT,
Domingo JS, Seifring S, Sivaganesan M, Haugland RA.
2008. Quantitative PCR for detection and enumeration
of genetic markers of bovine fecal pollution. Appl Environ
Microbiol 74:745–752.
57. Dick LK, Bernhard AE, Brodeur TJ, Santo Domingo JW,
Simpson JM, Walters SP, Field KG. 2005. Host distribu-
tions of uncultivated fecal Bacteroidales bacteria reveal
genetic markers for fecal source identification. Appl Environ
Microbiol 71:3184–3191.
58. Lamendella R, Santo Domingo JW, Yannarell AC, Ghosh
S, Di Giovanni G, Mackie RI, Oerther DB. 2009.
Evaluation of swine-specific PCR assays used for fecal
source tracking and analysis of molecular diversity of
swine-specific “bacteroidales” populations. Appl Environ
Microbiol 75:5787–5796.
59. ISO. 1995. ISO 10705-1: Water quality Detection and enu-
meration of bacteriophages part 1: Enumeration of F-specific
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-21
RNA bacteriophages. International Organization for Stand-
ardization, Geneva.
60. U.S. EPA. 2001. Method 1601: Male-specific (F+) and
somatic coliphage in water by two-step enrichment proce-
dure, vol. 821-R-01-030. U.S. EPA, Office of Water,
Washington, DC.
61. U.S. EPA. 2001. Method 1602. Male-specific (F+) and
somatic coliphage in water by single agar layer (SAL) pro-
cedure, vol. 821-R-01-029. U.S. EPA, Office of Water,
Washington, DC.
62. Duran AE, Muniesa M, Mendez X, Valero F, Lucena F,
Jofre J. 2002. Removal and inactivation of indicator bac-
teriophages in fresh waters. J Appl Microbiol 92:338–347.
63. Muniesa M, Payan A, Moce-Llivina L, Blanch AR, Jofre
J. 2009. Differential persistence of F-specific RNA phage
subgroups hinders their use as single tracers for faecal source
tracking in surface water. Water Res 43:1559–1564.
64. Long SC, Sobsey MD. 2004. A comparison of the survival
of F+RNA and F+DNA coliphages in lake water micro-
cosms. J Water Health 2:15–22.
65. Okabe S, Okayama N, Savichtcheva O, Ito T. 2007.
Quantification of host-specific Bacteroides-Prevotella 16S
rRNA genetic markers for assessment of fecal pollution in
freshwater. Appl Microbiol Biotechnol 74:890–901.
66. Marti R, Mieszkin S, Solecki O, Pourcher AM, Hervio-
Heath D, Gourmelon M. 2011. Effect of oxygen and tem-
perature on the dynamic of the dominant bacterial popula-
tions of pig manure and on the persistence of pig-associated
genetic markers, assessed in river water microcosms. J Appl
Microbiol 111:1159–1175.
67. Fremaux B, Boa T, Yost CK. 2010. Quantitative real-time
PCR assays for sensitive detection of Canada goose-specific
fecal pollution in water sources. Appl Environ Microbiol 76:
4886–4889.
68. Gomez-Donate M, Balleste E, Muniesa M, Blanch AR.
2012. New molecular quantitative PCR assay for detection
of host-specific bifidobacteriaceae suitable for microbial
source tracking. Appl Environ Microbiol 78:5788–5795.
69. Marti R, Dabert P, Pourcher AM. 2009. Pig manure
contamination marker selection based on the influence
of biological treatment on the dominant fecal microbial
groups. Appl Environ Microbiol 75:4967–4974.
70. Weidhaas J, Macbeth T, Olsen R, Sadowsky M, Norat D,
Harwood V. 2010. Identification of a poultry litter-specific
DNA marker gene and development of a 16S rRNA-based
quantitative PCR assay. J Appl Microbiol 109:334–347.
71. Weidhaas J, Macbeth T, Olsen R, Harwood V. 2011.
Correlation of Quantitative PCR for a Poultry-Specific Bre-
vibacterium Marker Gene with Bacterial and Chemical
Indicators of Water Pollution in a Watershed Impacted
by Land Application of Poultry Litter. App Environ Micro-
biol 77:2094–2102.
72. ISO. 2001. ISO 10705-4: Water quality. Detection and
enumeration of bacteriophages part 4: Enumeration of bac-
teriophages infecting Bacteroides fragilis. International
Organization for Standardization, Geneva.
73. Payan A, Ebdon J, Taylor H, Gantzer C, Ottoson J, Papa-
georgiou GT, Blanch AR, Lucena F, Jofre J, Muniesa M.
2005. Method for isolation of Bacteroides bacteriophage
host strains suitable for tracking sources of fecal pollution
in water. Appl Environ Microbiol 71:5659–5662.
74. Gomez-Donate M, Payan A, Cortes I, Blanch AR,
Lucena F, Jofre J, Muniesa M. 2011. Isolation of bacterio-
phage host strains of Bacteroides species suitable for tracking
sources of animal faecal pollution in water. Environ Micro-
biol 13:1622–1631.
75. Ley V, Higgins J, Fayer R. 2002. Bovine enteroviruses as
indicators of fecal contamination. Appl Environ Microbiol
68:3455–3461.
76. Jimenez-Clavero MA, Escribano-Romero E, Mansilla C,
Gomez N, Cordoba L, Roblas N, Ponz F, Ley V, Saiz
JC. 2005. Survey of bovine enterovirus in biological and
environmental samples by a highly sensitive real-time
reverse transcription-PCR. Appl Environ Microbiol 71:
3536–3543.
77. Jimenez-Clavero MA, Fernandez C, Ortiz JA, Pro J,
Carbonell G, Tarazona JV, Roblas N, Ley V. 2003.
Teschoviruses as indicators of porcine fecal contamination
of surface water. Appl Environ Microbiol 69:6311–6315.
78. Maluquer de Motes C, Clemente-Casares P, Hundesa A,
Martin M, Girones R. 2004. Detection of bovine and por-
cine adenoviruses for tracing the source of fecal contamina-
tion. Appl Environ Microbiol 70:1448–1454.
79. Hundesa A, Maluquer de Motes C, Albinana-Gimenez
N, Rodriguez-Manzano J, Bofill-Mas S, Sunen E, Girones
R. 2009. Development of a qPCR assay for the quantifica-
tion of porcine adenoviruses as an MST tool for swine fecal
contamination in the environment. J Virol Meth 158:
130–135.
80. Wong K, Xagoraraki I. 2010. Quantitative PCR assays to
survey the bovine adenovirus levels in environmental sam-
ples. J Appl Microbiol 109:605–612.
81. Hundesa A, Maluquer de Motes C, Bofill-Mas S,
Albinana-Gimenez N, Girones R. 2006. Identification
of human and animal adenoviruses and polyomaviruses
for determination of sources of fecal contamination in
the environment. Appl Environ Microbiol 72:7886–7893.
82. Wong K, Xagoraraki I. 2011. Evaluating the prevalence
and genetic diversity of adenovirus and polyomavirus in
bovine waste for microbial source tracking. Appl Microbiol
Biotechnol 90:1521–1526.
83. Hundesa A, Bofill-Mas S, Maluquer de Motes C,
Rodriguez-Manzano J, Bach A, Casas M, Girones R.
2010. Development of a quantitative PCR assay for the
quantitation of bovine polyomavirus as a microbial source-
tracking tool. J Virol Meth 163:385–389.
84. Tobe SS, Linacre AMT. 2008. A technique for the quan-
tification of human and non-human mammalian mito-
chondrial DNA copy number in forensic and other
mixtures. Forensic Sci Int Genet 2:249–256.
85. Xiao L, Morgan UM, Fayer R, Thompson RCA, Lal AA.
2000. Cryptosporidium Systematics and Implications for
Public Health. Trends Parasitol 16:287–292.
86. Xiao L, Fayer R, Ryan U, Upton SJ. 2004. Cryptospori-
dium Taxonomy: Recent Advances and Implications for
Public Health. Clin Microbiol Rev 17:72–97.
87. Slavin D. 1955. Cryptosporidium meleagridis (sp. nov.).
J Comp Pathol 65:262–266.
88. Current WL, Garcia LS. 1991. Cryptosporidiosis. Clin
Microbiol Rev 4:325–358.
89. Dubey JP, Speers CA, Fayer R. 1990. Cryptosporidiosis of
Man and Animals. CRC Press, New York.
90. Cama VA, Bern C, Sulaiman IM, Gilman RH, Ticona E,
Vivar A, Kawai V, Vargas D, Zhou L, Xiao L. 2003.
Cryptosporidium species and genotypes in HIV-positive
patients in Lima, Peru. J Eukaryot Microbiol 50(Suppl):
531–533.
91. Feng Y, Zhao X, Chen J, Jin W, Zhou X, Li N, Wang L,
Xiao L. 2011. Occurrence, source, and human infection
potential of cryptosporidium and Giardia spp. in source
and tap water in shanghai, China. Appl Environ Microbiol
77:3609–3616.
92. Jellison KL, Lynch AE, Ziemann JM. 2009. Source track-
ing identifies deer and geese as vectors of human-infectious
Cryptosporidium genotypes in an urban/suburban water-
shed. Environ Sci Technol 43:4267–4272.
93. Nichols RA, Connelly L, Sullivan CB, Smith HV.
2010. Identification of Cryptosporidium species and
3.4.4-22 ▪ MICROBIAL SOURCE TRACKING
genotypes in Scottish raw and drinking waters during a one-
year monitoring period. Appl Environ Microbiol 76:
5977–5986.
94. Oates SC, Miller MA, Hardin D, Conrad PA, Melli A,
Jessup DA, Dominik C, Roug A, Tinker MT, Miller
WA. 2012. Prevalence, environmental loading, and molec-
ular characterization of Cryptosporidium and Giardia isolates
from domestic and wild animals along the central Califor-
nia coast. Appl Environ Microbiol 78:8762–8772.
95. Yang W, Chen P, Villegas EN, Landy RB, Kanetsky C,
Cama V, Dearen T, Schultz CL, Orndorff KG, Prelewicz
GJ, Brown MH, Young KR, Xiao L. 2008. Cryptospori-
dium source tracking in the Potomac River watershed.
Appl Environ Microbiol 74:6495–6504.
96. Cama VA, Bern C, Roberts J, Cabrera L, Sterling CR,
Ortega Y, Gilman RH, Xiao L. 2008. Cryptosporidium spe-
cies and subtypes and clinical manifestations in children.
Peru Emerg Infect Dis 14:1567–1574.
97. Sulaiman IM, Hira PR, Zhou L, Al-Ali FM, Al-Shelahi
FA, Shweiki HM, Iqbal J, Khalid N, Xiao L. 2005.
Unique endemicity of cryptosporidiosis in children in
Kuwait. J Clin Microbiol 43:2805–2809.
98. Xiao L, Bern C, Limor J, Sulaiman I, Roberts J, Checkley
W, Cabrera L, Gilman RH, Lal AA. 2001. Identification
of 5 types of Cryptosporidium parasites in children in Lima,
Peru. J Infect Dis 183:492–497.
99. Caccio SM, Thompson RC, McLauchlin J, Smith HV.
2005. Unravelling Cryptosporidium and Giardia epidemiol-
ogy. Trends Parasitol 21:430–437.
100. Feng Y, Xiao L. 2011. Zoonotic potential and molecular
epidemiology of Giardia species and giardiasis. Clin Micro-
biol Rev 24:110–140.
101. Monis PT, Caccio SM, Thompson RC. 2009. Variation in
Giardia: towards a taxonomic revision of the genus. Trends
Parasitol 25:93–100.
102. Thompson RC, Monis PT. 2004. Variation in Giardia:
implications for taxonomy and epidemiology. Adv Parasitol
58:69–137.
103. Wielinga CM, Thompson RC. 2007. Comparative evalu-
ation of Giardia duodenalis sequence data. Parasitol 134:
1795–1821.
104. Almeida A, Pozio E, Caccio SM. 2010. Genotyping of
Giardia duodenalis cysts by new real-time PCR assays for
detection of mixed infections in human samples. Appl Envi-
ron Microbiol 76:1895–1901.
105. Guy RA, Payment P, Krull UJ, Horgen PA. 2003. Real-
time PCR for quantification of Giardia and Cryptosporidium
in environmental water samples and sewage. Appl Environ
Microbiol 69:5178–5185.
106. Caccio SM, Beck R, Lalle M, Marinculic A, Pozio E.
2008. Multilocus genotyping of Giardia duodenalis reveals
striking differences between assemblages A and B.
Int J Parasitol 38:1523–1531.
107. LeChevallier MW, Norton WD, Lee RG. 1991. Occur-
rence of Giardia and Cryptosporidium spp. in surface water
supplies. App Environ Microbiol 57:2610–2616.
108. Ryu H, Abbaszadegan M. 2008. Long-term study of Cryp-
tosporidium and Giardia occurrence and quantitative micro-
bial risk assessment in surface waters of Arizona in the
USA. J Water Health 6:263–273.
109. Ryu H, Alum A, Abbaszadegan M. 2005. Microbial char-
acterization and population changes in nonpotable
reclaimed water distribution systems. Environ Sci Technol
39:8600–8605.
110. Ley RE, Lozupone CA, Hamady M, Knight R, Gordon JI.
2008. Worlds within worlds: evolution of the vertebrate gut
microbiota. Nat Rev Microbiol 6:776–788.
111. Eckburg PB, Bik EM, Bernstein CN, Purdom E, Dethlef-
sen L, Sargent M, Gill SR, Nelson KE, Relman DA. 2005.
Diversity of the human intestinal microbial flora. Science
308:1635–1638.
112. Perkins GA, den Bakker HC, Burton AJ, Erb HN,
McDonough SP, McDonough PL, Parker J, Rosenthal
RL, Wiedmann M, Dowd SE, Simpson KW. 2012. Equine
stomachs harbor an abundant and diverse mucosal micro-
biota. Appl Environ Microbiol 78:2522–2532.
113. Lu J, Santo Domingo J, Shanks OC. 2007. Identification
of chicken-specific fecal microbial sequences using a meta-
genomic approach. Water Res 41:3561–3574.
114. Lee JE, Lee S, Sung J, Ko G. 2011. Analysis of human and
animal fecal microbiota for microbial source tracking.
ISME J 5:362–365.
115. Shanks OC, Kelty CA, Archibeque S, Jenkins M, New-
ton RJ, McLellan SL, Huse SM, Sogin ML. 2011. Com-
munity structures of fecal bacteria in cattle from different
animal feeding operations. Appl Environ Microbiol 77:
2992–3001.
116. Kim HB, Borewicz K, White BA, Singer RS, Sreevatsan
S, Tu ZJ, Isaacson RE. 2011. Longitudinal investigation of
the age-related bacterial diversity in the feces of commer-
cial pigs. Vet Microbiol 153:124–133.
117. Lu J, Santo Domingo JW, Lamendella R, Edge T, Hill S.
2008. Phylogenetic diversity and molecular detection
of bacteria in gull feces. Appl Environ Microbiol 74:
3969–3976.
118. Jeter SN, McDermott CM, Bower PA, Kinzelman JL,
Bootsma MJ, Goetz GW, McLellan SL. 2009. Bacteroidales
diversity in ring-billed gulls (Laurus delawarensis) residing at
Lake Michigan beaches. Appl Environ Microbiol 75:
1525–1533.
119. Kelty CA, Varma M, Sivaganesan M, Haugland RA,
Shanks OC. 2012. Distribution of genetic marker concen-
trations for fecal indicator bacteria in sewage and animal
feces. Appl Environ Microbiol 78:4225–4232.
120. Fogarty LR, Voytek MA. 2005. Comparison of
Bacteroides-Prevotella 16S rRNA genetic markers for fecal
samples from different animal species. Appl Environ Micro-
biol 71:5999–6007.
121. Bernhard AE, Field KG. 2000. Identification of non-
point sources of fecal pollution in coastal waters by using
host-specific 16S ribosomal DNA genetic markers
from fecal anaerobes. Appl Environ Microbiol 66:
1587–1594.
122. Dick LK, Simonich MT, Field KG. 2005. Microplate sub-
tractive hybridization to enrich for bacteroidales genetic
markers for fecal source identification. Appl Environ Micro-
biol 71:3179–3183.
123. Mauffret A, Caprais MP, Gourmelon M. 2012. Relevance
of bacteroidales and f-specific RNA bacteriophages for
efficient fecal contamination tracking at the level of
a catchment in france. Appl Environ Microbiol 78:
5143–5152.
124. Chase E, Hunting J, Staley C, Harwood VJ. 2012. Micro-
bial source tracking to identify human and ruminant sour-
ces of fecal pollution in an ephemeral Florida river. J Appl
Microbiol 113:1396–1406.
125. Reischer G, Kollanur D, Vierheilig J, Wehrspaun C,
Mach R, Sommer R, Stadler H, Farnleitner A. 2011.
Hypothesis-driven approach for the identification of fecal
pollution sources in water resources. Environ Sci Technol
45:4038–4045.
126. Santiago-Rodriguez TM, Tremblay RL, Toledo-
Hernandez C, Gonzalez-Nieves JE, Ryu H, Santo Domi-
ngo JW, Toranzos GA. 2012. Microbial quality of tropical
inland waters and effects of rainfall events. Appl Environ
Microbiol 78:5160–5169.
127. Jenkins MW, Tiwari S, Lorente M, Gichaba CM, Wuertz
S. 2009. Identifying human and livestock sources of fecal
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-23
contamination in Kenya with host-specific Bacteroidales
assays. Water Res 43:4956–4966.
128. Kreader CA. 1998. Persistence of PCR-detectable Bacter-
oides distasonis from human feces in river water. Appl Envi-
ron Microbiol 64:4103–4105.
129. Walters SP, Yamahara KM, Boehm AB. 2009. Persistence
of nucleic acid markers of health-relevant organisms in sea-
water microcosms: implications for their use in assessing
risk in recreational waters. Water Res 43:4929–4939.
130. Dick LK, Stelzer EA, Bertke EE, Fong DL, Stoeckel
DM. 2010. Relative decay of Bacteroidales microbial
source tracking markers and cultivated Escherichia coli in
freshwater microcosms. App Environ Microbiol 76:
3255–3262.
131. Jeanneau L, Solecki O, Wery N, Jarde E, Gourmelon M,
Communal PY, Jadas-Hecart A, Caprais MP, Gruau G,
Pourcher AM. 2012. Relative decay of fecal indicator bac-
teria and human-associated markers: a microcosm study
simulating wastewater input into seawater and freshwater.
Environ Sci Technol 46:2375–2382.
132. Schulz CJ, Childers GW. 2011. Fecal bacteroidales diver-
sity and decay in response to variations in temperature and
salinity. Appl Environ Microbiol 77:2563–2572.
133. Bell A, Layton AC, McKay L, Williams D, Gentry R,
Sayler GS. 2009. Factors influencing the persistence of
fecal Bacteroides in stream water. J Environ Qual 38:
1224–1232.
134. Walters SP, Gannon VP, Field KG. 2007. Detection of
Bacteroidales fecal indicators and the zoonotic pathogens
E. coli 0157:H7, salmonella, and campylobacter in river
water. Environ Sci Technol 41:1856–1862.
135. Schriewer A, Miller WA, Byrne BA, Miller MA, Oates S,
Conrad PA, Hardin D, Yang HH, Chouicha N, Melli A,
Jessup D, Dominik C, Wuertz S. 2010. Presence of Bacter-
oidales as a predictor of pathogens in surface waters of the
central California coast. Appl Environ Microbiol 76:
5802–5814.
136. Walters SP, Thebo AL, Boehm AB. 2011. Impact of
urbanization and agriculture on the occurrence of bacterial
pathogens and stx genes in coastal waterbodies of central
California. Water Res 45:1752–1762.
137. Rogers SW, Donnelly M, Peed L, Kelty CA, Mondal S,
Zhong Z, Shanks OC. 2011. Decay of bacterial pathogens,
fecal indicators, and real-time quantitative PCR genetic
markers in manure-amended soils. Appl Environ Microbiol
77:4839–4848.
138. Resnick IG, Levin MA. 1981. Assessment of bifidobacte-
ria as indicators of human fecal pollution. Appl Environ
Microbiol 42:433–438.
139. Bonjoch X, Balleste E, Blanch AR. 2005. Enumeration of
bifidobacterial populations with selective media to deter-
mine the source of waterborne fecal pollution. Water Res
39:1621–1627.
140. Mara DD, Oragui JI. 1983. Sorbitol-fermenting bifidobac-
teria as specific indicators of human faecal pollution. J Appl
Bacteriol 55:349–357.
141. Bonjoch X, Balleste E, Blanch AR. 2004. Multiplex PCR
with 16S rRNA gene-targeted primers of Bifidobacterium
spp. to identify sources of fecal pollution. Appl Environ
Microbiol 70:3171–3175.
142. King EL, Bachoon DS, Gates KW. 2007. Rapid detec-
tion of human fecal contamination in estuarine environ-
ments by PCR targeting of Bifidobacterium adolescentis.
J Microbiol Meth 68:76–81.
143. Balleste E, Blanch AR. 2011. Bifidobacterial diversity and
the development of new microbial source tracking indica-
tors. Appl Environ Microbiol 77:3518–3525.
144. Biavati B, Mattarelli P. 2006. The family Bifidobacteria-
ceae, p. 322–382. In Dworkin M, Falkow S, Rosenberg E,
Schleifer K-H, Stackebrandt E (eds), The Prokaryotes, vol.
3: Archaea. Bacteria: Firmicutes, Actinomycetes. Springer,
New York.
145. Garcia-Aljaro C, Balleste E, Rossello-Mora R, Cifuentes
A, Richter M, Blanch AR. 2012. Neoscardovia arbecensis
gen. nov., sp. nov., isolated from porcine slurries. Syst
Appl Microbiol 35:374–379.
146. Lamendella R, Santo Domingo JW, Kelty C, Oerther DB.
2008. Bifidobacteria in feces and environmental waters.
Appl Environ Microbiol 74:575–584.
147. Bonjoch X, Lucena F, Blanch AR. 2009. The persistence
of bifidobacteria populations in a river measured by molec-
ular and culture techniques. J Appl Microbiol 107:
1178–1185.
148. Rowbotham TJ, Cross T. 1977. Rhodococcus coprophilus
sp. nov.: an aerobic nocardioform actinomycete belonging
to the “rhodochrous” complex. J Gen Microbiol 100:
123–138.
149. Rowbotham TJ, Cross T. 1977. Ecology of Rhodococcus
coprophilus and associated actinomycetes in fresh water
and agricultural habitats. J Gen Microbiol 100:231–240.
150. Jagals P, Grabow W, de Villiers JC. 1995. Evaluation of
indicators for assessment of human and animal faecal pollu-
tion of surface run-off. Water Sci Technol 31:235–241.
151. Mara DD, Oragui JI. 1981. Occurrence of Rhodococcus cop-
rophilus and associated actinomycetes in feces, sewage, and
freshwater. Appl Environ Microbiol 42:1037–1042.
152. Oragui JI, Mara DD. 1983. Investigation of the survival
characteristics of Rhodococcus coprophilus and certain fecal
indicator bacteria. Appl Environ Microbiol 46:356–360.
153. Savill MG, Murray SR, Scholes P, Maas EW, McCormick
RE, Moore EB, Gilpin BJ. 2001. Application of polymer-
ase chain reaction (PCR) and TaqMan™ PCR techniques
to the detection and identification of Rhodococcus coprophi-
lus in faecal samples. J Microbiol Meth 47:355–368.
154. Mara DD, Oragui JI. 1985. Bacteriological methods for
distinguishing between human and animal faecal pollution
of water: results of fieldwork in Nigeria and Zimbabwe. Bull
World Health Org 63:773–783.
155. Lechevalier HA.1986. Nocardioforms, p. 1458–1506.
In Sneath P, Mair N, Sharp M, Holt J (eds), Bergey’s
Manual of Systematic Bacteriology, vol. 2. Williams and
Wilkins, Baltimore, MD.
156. Wicki M, Auckenthaler A, Felleisen R, Liniger M,
Loutre C, Niederhauser I, Tanner M, Baumgartner A.
2012. Improved detection of Rhodococcus coprophilus with
a new quantitative PCR assay. Appl Microbiol Biotechnol
93:2161–2169.
157. Plummer J, Long S. 2007. Monitoring source water for
microbial contamination: evaluation of water quality meas-
ures. Water Res 41:3716–3728.
158. Wiggins BA. 1996. Discriminant analysis of antibiotic
resistance patterns in fecal streptococci, a method to
differentiate human and animal sources of fecal pollu-
tion in natural waters. Appl Environ Microbiol 62:
3997–4002.
159. Wheeler AL, Hartel PG, Godfrey DG, Hill JL, Segars
WI. 2002. Potential of Enterococcus faecalis as a human
fecal indicator for microbial source tracking. J Environ
Qual 31:1286–1293.
160. Kuntz RL, Hartel PG, Rodgers K, Segars WI. 2004. Pres-
ence of Enterococcus faecalis in broiler litter and wild bird
feces for bacterial source tracking. Water Res 38:
3551–3557.
161. Ryu H, Henson M, Elk M, Toledo-Hernandez C, Griffith
J, Blackwood D, Noble R, Gourmelon M, Glassmeyer S,
Santo Domingo J. 2012. Development of quantitative
PCR assays targeting 16S rRNA gene of Enterococcus spp.
and their application to the identification of Enterococcus
3.4.4-24 ▪ MICROBIAL SOURCE TRACKING
species in environmental samples. Appl Environ Microbiol
79:196–204.
162. Jackson CR, Fedorka-Cray PJ, Barrett JB. 2004. Use of a
genus- and species-specific multiplex PCR for identifica-
tion of enterococci. J Clin Microbiol 42:3558–3565.
163. Scott TM, Jenkins TM, Lukasik J, Rose JB. 2005. Poten-
tial use of a host associated molecular marker in Enterococ-
cus faecium as an index of human fecal pollution. Environ
Sci Technol 39:283–287.
164. Ahmed W, Stewart J, Powell D, Gardner T. 2008. Evalu-
ation of the host-specificity and prevalence of enterococci
surface protein (esp) marker in sewage and its application
for sourcing human fecal pollution. J Environ Qual 37:
1583–1588.
165. McQuaig SM, Scott TM, Harwood VJ, Farrah SR,
Lukasik JO. 2006. Detection of human-derived fecal
pollution in environmental waters by use of a PCR-based
human polyomavirus assay. Appl Environ Microbiol 72:
7567–7574.
166. Lanthier M, Scott A, Lapen DR, Zhang Y, Topp E. 2010.
Frequency of virulence genes and antibiotic resistances in
Enterococcus spp. isolates from wastewater and feces of
domesticated mammals and birds, and wildlife. Can J
Microbiol 56:715–729.
167. Layton BA, Walters SP, Boehm AB. 2009. Distribution
and diversity of the enterococcal surface protein (esp)
gene in animal hosts and the Pacific coast environment.
J Appl Microbiol 106:1521–1531.
168. Whitman RL, Przybyla-Kelly K, Shively DA, Byappana-
halli MN. 2007. Incidence of the enterococcal surface pro-
tein (esp) gene in human and animal fecal sources. Environ
Sci Technol 41:6090–6095.
169. Johanson JJ, Feriancikova L, Xu S. 2012. Influence of
enterococcal surface protein (esp) on the transport of Enter-
ococcus faecium within saturated quartz sands. Environ Sci
Technol 46:1511–1518.
170. Lawson PA, Collins MD, Falsen E, Foster G. 2006. Cat-
ellicoccus marimammalium gen. nov., sp. nov., a novel
Gram-positive, catalase-negative, coccus-shaped bacte-
rium from porpoise and grey seal. Int J Syst Evol Microbiol
56:429–432.
171. Ryu H, Lu J, Vogel J, Elk M, Chavez-Ramirez F, Ashbolt
N, Santo Domingo J. 2012. Development and evaluation
of a quantitative PCR assay targeting sandhill crane
(Grus canadensis) fecal pollution. Appl Environ Microbiol
78:4338–4345.
172. Ryu H, Griffith JF, Khan IU, Hill S, Edge TA, Toledo-
Hernandez C, Gonzalez-Nieves J, Santo Domingo J.
2012. Comparison of gull feces-specific assays targeting
the 16S rRNA genes of Catellicoccus marimammalium and
Streptococcus spp. Appl Environ Microbiol 78:1909–1916.
173. Weigand MR, Ryu H, Bozcek L, Konstantinidis KT,
Santo Domingo J. 2013. Draft genome sequence of Catelli-
coccus marimammalium, a novel species commonly found in
gull feces. Genome Announce 1:e00019-12
174. Bisgaard M, Bojesen AM, Petersen MR, Christensen H.
2012. A major outbreak of Streptococcus equi subsp. zooepi-
demicus infections in free-range chickens is linked to
horses. Avian Dis 56:561–566.
175. Devriese LA, Vandamme P, Collins MD, Alvarez N, Pot
B, Hommez J, Butaye P, Haesebrouck F. 1999. Streptococ-
cus pluranimalium sp. nov., from cattle and other animals.
Int J Syst Bacteriol 49(Pt 3):1221–1226.
176. Vela AI, Casamayor A, Sanchez Del Rey V, Dominguez
L, Fernandez-Garayzabal JF. 2009. Streptococcus plurexto-
rum sp. nov., isolated from pigs. Int J Syst Evol Microbiol
59:504–508.
177. Vela AI, Sanchez V, Mentaberre G, Lavin S, Dominguez
L, Fernandez-Garayzabal JF. 2011. Streptococcus porcorum
sp. nov., isolated from domestic and wild pigs. Int J Syst Evol
Microbiol 61:1585–1589.
178. Lawson PA, Foster G, Falsen E, Markopoulos SJ, Collins
MD. 2005. Streptococcus castoreus sp. nov., isolated from a
beaver (Castor fiber). Int J Syst Evol Microbiol 55:
843–846.
179. Hagedorn C, Crozier JB, Mentz KA, Booth AM, Graves
AK, Nelson NJ, Reneau RB Jr. 2003. Carbon source uti-
lization profiles as a method to identify sources of faecal pol-
lution in water. J Appl Microbiol 94:792–799.
180. Simonsen GS, Smabrekke L, Monnet DL, Sorensen TL,
Moller JK, Kristinsson KG, Lagerqvist-Widh A, Torell E,
Digranes A, Harthug S, Sundsfjord A. 2003. Prevalence
of resistance to ampicillin, gentamicin and vancomycin
in Enterococcus faecalis and Enterococcus faecium isolates
from clinical specimens and use of antimicrobials in five
Nordic hospitals. J Antimicrob Chemother 51:323–331.
181. Witte W. 2000. Selective pressure by antibiotic use in live-
stock. Int J Antimicrob. Agents. 16(Suppl 1):S19–S24.
182. Geldreich EE, Kenner BA. 1969. Concepts of fecal strep-
tococci in stream pollution. J Water Pollut Control Fed 41
(Suppl):R336+.
183. Kummerer K. 2009. Antibiotics in the aquatic environ-
ment–a review–part I. Chemosphere 75:417–434.
184. Koike S, Aminov RI, Yannarell AC, Gans HD, Krapac
IG, Chee-Sanford JC, Mackie RI. 2010. Molecular ecol-
ogy of macrolide-lincosamide-streptogramin B methylases
in waste lagoons and subsurface waters associated with
swine production. Microb Ecol 59:487–498.
185. Duan H, Chai T, Liu J, Zhang X, Qi C, Gao J, Wang Y,
Cai Y, Miao Z, Yao M, Schlenker G. 2009. Source identi-
fication of airborne Escherichia coli of swine house surround-
ings using ERIC-PCR and REP-PCR. Environ Res 109:
511–517.
186. Seurinck S, Verstraete W, Siciliano SD. 2003. Use of
16S-23S rRNA intergenic spacer region PCR and repeti-
tive extragenic palindromic PCR analyses of Escherichia
coli isolates to identify nonpoint fecal sources. App Environ
Microbiol 69:4942–4950.
187. Casarez EA, Pillai SD, Di Giovanni GD. 2007. Genotype
diversity of Escherichia coli isolates in natural waters deter-
mined by PFGE and ERIC-PCR. Water Res 41:3643–3648.
188. McLellan SL, Daniels AD, Salmore AK. 2003. Genetic
characterization of Escherichia coli populations from host
sources of fecal pollution by using DNA fingerprinting.
Appl Environ Microbiol 69:2587–2594.
189. Moore DF, Harwood VJ, Ferguson DM, Lukasik J, Han-
nah P, Getrich M, Brownell M. 2005. Evaluation of anti-
biotic resistance analysis and ribotyping for identification
of faecal pollution sources in an urban watershed. J Appl
Microbiol 99:618–628.
190. Lyautey E, Lu Z, Lapen DR, Wilkes G, Scott A, Berkers
T, Edge TA, Topp E. 2010. Distribution and diversity of
Escherichia coli populations in the South Nation River
drainage basin, eastern Ontario, Canada. App Environ
Microbiol 76:1486–1496.
191. Ma HJ, Fu LL, Li JR. 2011. Differentiation of fecal Escher-
ichia coli from human, livestock, and poultry sources by
rep-PCR DNA fingerprinting on the shellfish culture
area of East China Sea. Curr Microbiol 62:1423–1430.
192. Ritter KJ, Carruthers E, Carson CA, Ellender RD, Har-
wood VJ, Kingsley K, Nakatsu C, Sadowsky M, Shear B,
West B, Whitlock JE, Wiggins BA, Wilbur JD. 2003.
Assessment of statistical methods used in library-based
approaches to microbial source tracking. J Water Health 1:
209–223.
193. Stoeckel DM, Mathes MV, Hyer KE, Hagedorn C,
Kator H, Lukasik J, O’Brien TL, Fenger TW,
Samadpour M, Strickler KM, Wiggins BA. 2004.
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-25
Comparison of seven protocols to identify fecal con-
tamination sources using Escherichia coli. Environ Sci
Technol 38:6109–6117.
194. Johnson LK, Brown MB, Carruthers EA, Ferguson JA,
Dombek PE, Sadowsky MJ. 2004. Sample size, library
composition, and genotypic diversity among natural popu-
lations of Escherichia coli from different animals influence
accuracy of determining sources of fecal pollution. Appl
Environ Microbiol 70:4478–4485.
195. Bergh O, Borsheim KY, Bratbak G, Heldal M. 1989. High
abundance of viruses found in aquatic environments.
Nature 340:467–468.
196. Fuhrman JA. 1999. Marine viruses and their biogeochem-
ical and ecological effects. Nature 399:541–548.
197. Whitman WB, Coleman DC, Wiebe WJ. 1998. Prokar-
yotes: the unseen majority. Proc Natl Acad Sci USA 95:
6578–6583.
198. Waldor MK, Friedman DI, Adhya SL. 2005. Phages: Their
Role in Bacterial Pathogenesis and Biotechnology. ASM Press,
Washington, DC.
199. Guelin A. 1948. Etude quantitative de bacteriophages
typhiques. VI. Dans les eux. Ann Inst Pasteur 79:447.
200. IAWPRC. 1991. Bacteriophages as model viruses in water
quality control. Water Res 25:17.
201. Grabow WOK. 2001. Bacteriophages: update on applica-
tion as models for viruses in water. Water SA 27:18.
202. Jofre J, Stewart JR, Grabow W. 2011. Phage methods, pp
137–156. In Hagedorn C, Blanch AR, Harwood VN (eds),
Microbial Source Tracking: Methods, Applications, and Case
Studies:. Springer, New York.
203. Adams MH. 1959. Bacteriophages. Interscience, New York.
204. ISO. 2000. ISO 10705-2: Water quality. Detection and enu-
meration of bacteriophages part 2: enumeration of somatic col-
iphages. International Organization for Standardization,
Geneva.
205. Wolf S, Hewitt J, Rivera-Aban M, Greening GE. 2008.
Detection and characterization of F+ RNA bacteriophages
in water and shellfish: application of a multiplex real-time
reverse transcription PCR. JVirol Meth 149:123–128.
206. Ogorzaly L, Tissier A, Bertrand I, Maul A, Gantzer C.
2009. Relationship between F-specific RNA phage gen-
ogroups, faecal pollution indicators and human adenovi-
ruses in river water. Water Res 43:1257–1264.
207. Puig M, Jofre J, Girones R. 2000. Detection of phages
infecting Bacteroides fragilis HSP40 using a specific DNA
probe. J Virol Meth 88:163–173.
208. Jofre J, Blanch AR. 2010. Feasibility of methods based on
nucleic acid amplification techniques to fulfil the require-
ments for microbiological analysis of water quality. J Appl
Microbiol 109:1853–1867.
209. Mendez J, Audicana A, Isern A, Llaneza J, Moreno B,
Tarancon ML, Jofre J, Lucena F. 2004. Standardised eval-
uation of the performance of a simple membrane
filtration-elution method to concentrate bacteriophages
from drinking water. J Virol Meth 117:19–25.
210. Sobsey MD, Schwab K, Kellogg DE. 1990. Simple mem-
brane filter method to concentrate and enumerate
male-specific RNA coliphages. J Am Water Works Assn
82:52–59.
211. Hsu FC, Shieh YS, van Duin J, Beekwilder MJ, Sobsey
MD. 1995. Genotyping male-specific RNA coliphages by
hybridization with oligonucleotide probes. Appl Environ
Microbiol 61:3960–3966.
212. Furuse K, Ando A, Watanabe I. 1975. Isolation and
grouping of RNA phages. A survey in the islands in the
adjacent seas of Japan. J Keio Med Soc 52:15.
213. Tartera C, Lucena F, Jofre J. 1989. Human origin of Bac-
teroides fragilis bacteriophages present in the environment.
Appl Environ Microbiol 55:2696–2701.
214. Purnell SE, Ebdon JE, Taylor HD. 2011. Bacteriophage
lysis of Enterococcus host strains: a tool for microbial source
tracking? Environ. Sci Technol 45:10699–10705.
215. Belanche-Munoz L, Blanch AR. 2007. Machine learning
methods for microbial source tracking. Envrion Model Soft-
ware 23:741–750.
216. Muniesa M, Lucena F, Blanch AR, Payan A, Jofre J.
2012. Use of abundance ratios of somatic coliphages and
bacteriophages of Bacteroides thetaiotaomicron GA17 for
microbial source identification. Water Res 46:6410–6418.
217. Miyake T, Haruna I, Shiba T, Ito YH, Yamane K. 1971.
Grouping of RNA phages based on the template specificity
of their RNA replicases. Proc Natl Acad Sci USA 68:
2022–2024.
218. Beekwilder J, Nieuwenhuizen R, Havelaar AH, van
Duin J. 1996. An oligonucleotide hybridization assay
for the identification and enumeration of F-specific
RNA phages in surface water. J Appl Bacteriol 80:179–186.
219. Havelaar AH, Furuse K, Hogeboom WM. 1986. Bacterio-
phages and indicator bacteria in human and animal faeces.
J Appl Bacteriol 60:255–262.
220. Dhillon EK, Dhillon TS. 1974. Synthesis of indicator
strains and density of ribonucleic acid-containing coli-
phages in sewage. Appl Microbiol 27:640–647.
221. Griffin DW, Stokes R, Rose JB, Paul JHI. 2000. Bacterial
indicator occurrence and the use of an F+ specific RNA
coliphage assay to identify fecal sources in Homosassa
Springs, Florida. Microb Ecol 39:56–64.
222. Schaper M, Duran AE, Jofre J. 2002. Comparative resist-
ance of phage isolates of four genotypes of F-specific RNA
bacteriophages to various inactivation processes. Appl Envi-
ron Microbiol 68:3702–3707.
223. Stewart-Pullaro J, Daugomah JW, Chestnut DE, Graves
DA, Sobsey MD, Scott GI. 2006. F+RNA coliphage typ-
ing for microbial source tracking in surface waters. J Appl
Microbiol 101:1015–1026.
224. Novotny CP, Lavin K. 1971. Some effects of temperature
on the growth of F-pili. J Bacteriol 107:12.
225. Love DC, Sobsey MD. 2007. Simple and rapid F+ coli-
phage culture, latex agglutination, and typing assay to
detect and source track fecal contamination. Appl Environ
Microbiol 73:4110–4118.
226. Ijzerman MM, Hagedorn C. 1992. Improved method for
coliphage detection based on beta-galactosidase induction.
J Virol Meth 40:31–36.
227. Kirs M, Smith DC. 2007. Multiplex quantitative real-time
reverse transcriptase PCR for F+-specific RNA coliphages:
a method for use in microbial source tracking. Appl Environ
Microbiol 73:808–814.
228. Friedman SD, Cooper EM, Casanova L, Sobsey MD,
Genthner FJ. 2009. A reverse transcription-PCR assay to
distinguish the four genogroups of male-specific (F+)
RNA coliphages. J Virol Meth 159:47–52.
229. Griffith JF, Weisberg SB, McGee CD. 2003. Evalua-
tion of microbial source tracking methods using mixed
fecal sources in aqueous test samples. J Water Health 1:
141–151.
230. Tartera C, Jofre J. 1987. Bacteriophages active against Bac-
teroides fragilis in sewage-polluted waters. Appl Environ
Microbiol 53:1632–1637.
231. Xu J, Bjursell MK, Himrod J, Deng S, Carmichael LK,
Chiang HC, Hooper LV, Gordon JI. 2003. A genomic
view of the human-Bacteroides thetaiotaomicron symbiosis.
Science 299:2074–2076.
232. Mazmanian SK, Round JL, Kasper DL. 2008. A microbial
symbiosis factor prevents intestinal inflammatory disease.
Nature 453:620–625.
233. Ebdon J, Muniesa M, Taylor H. 2007. The application of a
recently isolated strain of Bacteroides (GB-124) to identify
3.4.4-26 ▪ MICROBIAL SOURCE TRACKING
human sources of faecal pollution in a temperate river
catchment. Water Res 41:3683–3690.
234. Zoetendal EG, Collier CT, Koike S, Mackie RI, Gaskins
HR. 2004. Molecular ecological analysis of the gastrointes-
tinal microbiota: a review. J Nutr 134:465–472.
235. Dethlefsen L, Eckburg PB, Bik EM, Relman DA. 2006.
Assembly of the human intestinal microbiota. Trends Ecol
Evol 21:517–523.
236. Sanchez D, Belanche LA, Blanch AR. 2011. A software
system for the microbial source tracking problem. J Mach
Learn Res 17:7.
237. Fong TT, Lipp EK. 2005. Enteric viruses of humans and
animals in aquatic environments: health risks, detection,
and potential water quality assessment tools. Microbiol
Mol Biol Rev 69:357–371.
238. Yates MV. 2007. Classical indicators in the 21st century—
far and beyond the coliform. Water Environ Res 79:
279–286.
239. Fong TT, Griffin DW, Lipp EK. 2005. Molecular assays for
targeting human and bovine enteric viruses in coastal
waters and their application for library-independent source
tracking. Appl Environ Microbiol 71:2070–2078.
240. Griffin DW, Donaldson KA, Paul JH, Rose JB. 2003.
Pathogenic human viruses in coastal waters. Clin Microbiol
Rev 16:129–143.
241. Noble RT, Allen SM, Blackwood AD, Chu W, Jiang SC,
Lovelace GL, Sobsey MD, Stewart JR, Wait DA. 2003.
Use of viral pathogens and indicators to differentiate
between human and non-human fecal contamination in
a microbial source tracking comparison study. J Water
Health 1:195–207.
242. McCarthy FM, Smith GA, Mattick JS. 1999. Molecular
characterisation of Australian bovine enteroviruses. Vet
Microbiol 68:71–81.
243. Gerba CP, Gramos DM, Nwachuku N. 2002. Compara-
tive inactivation of enteroviruses and adenovirus 2 by
UV light. Appl Environ Microbiol 68:5167–5169.
244. Shah KV, Daniel RW, Warszawski RM. 1973. High prev-
alence of antibodies to BK virus, an SV40-related papova-
virus, in residents of Maryland. J Infect Dis 128:784–787.
245. Hirsch HH, Steiger J. 2003. Polyomavirus BK. Lancet
Infect Dis 3:611–623.
246. Polo C, Perez JL, Mielnichuck A, Fedele CG, Niubo J,
Tenorio A. 2004. Prevalence and patterns of polyomavirus
urinary excretion in immunocompetent adults and chil-
dren. Clin Microbiol Infect 10:640–644.
247. Bofill-Mas S, Pina S, Girones R. 2000. Documenting the
epidemiologic patterns of polyomaviruses in human popu-
lations by studying their presence in urban sewage. Appl
Environ Microbiol 66:238–245.
248. McQuaig S, Noble R. 2011. Viruses as tracers of fecal con-
tamination, pp 113–135. In Hagedorn C, Blanch A, Har-
wood VJ (eds), Microbial Source Tracking: Methods,
Applications, and Case Studies:. Springer, New York.
249. Rangan SR, Lowrie RC, Roberts JA, Johnston PB, War-
rick RP. 1974. Virus from stumptailed monkey (Macaca
arctoides) kidney cultures. Lab Anim Sci 24:211–217.
250. Bofill-Mas S, Albinana-Gimenez N, Clemente-Casares P,
Hundesa A, Rodriguez-Manzano J, Allard A, Calvo M,
Girones R. 2006. Quantification and stability of human
adenoviruses and polyomavirus JCPyV in wastewater matri-
ces. Appl Environ Microbiol 72:7894–7896.
251. Tadmor AD, Ottesen EA, Leadbetter JR, Phillips R.
2011. Probing individual environmental bacteria for viruses
by using microfluidic digital PCR. Science 333:58–62.
252. Dubinsky EA, Esmaili L, Hulls JR, Cao Y, Griffith JF,
Andersen GL. 2012. Application of phylogenetic microar-
ray analysis to discriminate sources of fecal pollution. Envi-
ron Sci Technol 46:4340–4347.
253. Jeong JY, Park HD, Lee KH, Weon HY, Ka JO. 2011.
Microbial community analysis and identification of alter-
native host-specific fecal indicators in fecal and river water
samples using pyrosequencing. J Microbiol 49:585–594.
254. McLellan SL, Huse SM, Mueller-Spitz SR, Andreish-
cheva EN, Sogin ML. 2010. Diversity and population
structure of sewage-derived microorganisms in waste-
water treatment plant influent. Environ Microbiol 12:
378–392.
255. Unno T, Jang J, Han D, Kim JH, Sadowsky MJ, Kim OS,
Chun J, Hur HG. 2010. Use of barcoded pyrosequencing
and shared OTUs to determine sources of fecal bacteria
in watersheds. Environ Sci Technol 44:7777–7782.
256. Loman NJ, Misra RV, Dallman TJ, Constantinidou C,
Gharbia SE, Wain J, Pallen MJ. 2012. Performance com-
parison of benchtop high-throughput sequencing plat-
forms. Nat Biotechnol 30:434–439.
257. Knights D, Kuczynski J, Charlson ES, Zaneveld J, Mozer
MC, Collman RG, Bushman FD, Knight R, Kelley ST.
2011. Bayesian community-wide culture-independent
microbial source tracking. Nat Meth 8:761–763.
258. Unno T, Di DY, Jang J, Suh YS, Sadowsky MJ, Hur HG.
2012. Integrated online system for a pyrosequencing-based
microbial source tracking method that targets Bacteroidetes
16S rDNA. Environ Sci Technol 46:93–98.
259. Shanks OC, Domingo JW, Lu J, Kelty CA, Graham JE.
2007. Identification of bacterial DNA markers for the
detection of human fecal pollution in water. Appl Environ
Microbiol 73:2416–2422.
260. Lee YJ, Molina M, Santo Domingo JW, Willis JD, Cyter-
ski M, Endale DM, Shanks OC. 2008. Temporal assess-
ment of the impact of exposure to cow feces in two
watersheds by multiple host-specific PCR assays. Appl Envi-
ron Microbiol 74:6839–6847.
261. Mokili JL, Rohwer F, Dutilh BE. 2012. Metagenomics
and future perspectives in virus discovery. Curr Opin Virol
2:63–77.
262. Reyes A, Semenkovich NP, Whiteson K, Rohwer F, Gor-
don JI. 2012. Going viral: next-generation sequencing
applied to phage populations in the human gut. Nat Rev
Microbiol 10:607–617.
263. Willner D, Haynes MR, Furlan M, Hanson N, Kirby B,
Lim YW, Rainey PB, Schmieder R, Youle M, Conrad
D, Rohwer F. 2012. Case studies of the spatial heterogene-
ity of DNA viruses in the cystic fibrosis lung. Am J Respir
Cell Mol Biol 46:127–131.
264. Svraka S, Rosario K, Duizer E, van der Avoort H, Breit-
bart M, Koopmans M. 2010. Metagenomic sequencing for
virus identification in a public-health setting. J Gen Virol
91:2846–2856.
265. Breitbart M, Hewson I, Felts B, Mahaffy JM, Nulton J,
Salamon P, Rohwer F. 2003. Metagenomic analyses of
an uncultured viral community from human feces.
J Bacteriol 185:6220–6223.
266. Tamaki H, Zhang R, Angly FE, Nakamura S, Hong PY,
Yasunaga T, Kamagata Y, Liu WT. 2012. Metagenomic
analysis of DNA viruses in a wastewater treatment plant
in tropical climate. Environ Microbiol 14:441–452.
267. Rosario K, Symonds EM, Sinigalliano C, Stewart J,
Breitbart M. 2009. Pepper mild mottle virus as an indicator
of fecal pollution. Appl Environ Microbiol 75:7261–7267.
268. Cerdeno-Tarraga AM, Patrick S, Crossman LC, Blakely
G, Abratt V, Lennard N, Poxton I, Duerden B, Harris B,
Quail MA, Barron A, Clark L, Corton C, Doggett J, Hol-
den MT, Larke N, Line A, Lord A, Norbertczak H,
Ormond D, Price C, Rabbinowitsch E, Woodward J, Bar-
rell B, Parkhill J. 2005. Extensive DNA inversions in the
Bacillus fragilis genome control variable gene expression.
Science 307:1463–1465.
 3.4.4. Methods of Targeting Animal Sources of Fecal Pollution in Water ▪ 3.4.4-27
269. Cuiv PO, Klaassens ES, Durkin AS, Harkins DM, Foster
L, McCorrison J, Torralba M, Nelson KE, Morrison M.
2011. Draft genome sequence of Bacteroides vulgatus
PC510, a strain isolated from human feces. J Bacteriol
193:4025–4026.
270. Kim M, Yi H, Cho YJ, Jang J, Hur HG, Chun J. 2012.
Draft genome sequence of Escherichia coli W26, an enteric
strain isolated from cow feces. J Bacteriol 194:5149–5150.
271. Kim MS, Whon TW, Roh SW, Shin NR, Bae JW. 2011.
Draft genome sequence of Bacteroides faecis MAJ27 T, a
strain isolated from human feces. J Bacteriol 193:
6801–6802.
272. Lee K, Yi H, Cho YJ, Jang J, Hur HG, Chun J. 2012. Draft
genome sequence of Escherichia coli AI27, a porcine isolate
belonging to phylogenetic group B1. J Bacteriol 194:
6640–6641.
273. Li T, Pu F, Yang R, Fang X, Wang J, Guo Y, Chang D, Su
L, Guo N, Jiang X, Zhao J, Liu C. 2012. Draft genome
sequence of Escherichia coli LCT-EC106. J Bacteriol 194:
4443–4444.
274. Palmer KL, Godfrey P, Griggs A, Kos VN, Zucker J, Des-
jardins C, Cerqueira G, Gevers D, Walker S, Wortman J,
Feldgarden M, Haas B, Birren B, Gilmore MS. 2012.
Comparative genomics of enterococci: variation in Entero-
coccus faecalis, clade structure in E. faecium, and defining
characteristics of E. gallinarum and E. casseliflavus. mBio
3:e00318-11.
275. Hansen EE, Lozupone CA, Rey FE, Wu M, Guruge JL,
Narra A, Goodfellow J, Zaneveld JR, McDonald
DT, Goodrich JA, Heath AC, Knight R, Gordon JI.
2011. Pan-genome of the dominant human gut-associated
archaeon, Methanobrevibacter smithii, studied in twins. Proc
Natl Acad Sci USA 108(Suppl 1):4599–4606.
276. Abrahamsen MS, Templeton TJ, Enomoto S, Abrahante
JE, Zhu G, Lancto CA, Deng M, Liu C, Widmer G,
Tzipori S, Buck GA, Xu P, Bankier AT, Dear PH,
Konfortov BA, Spriggs HF, Iyer L, Anantharaman V,
Aravind L, Kapur V. 2004. Complete genome sequence
of the apicomplexan, Cryptosporidium parvum. Science
304:441–445.
277. Kolinko S, Wanner G, Katzmann E, Kiemer F, B MF,
Schuler D. 2013. Clone libraries and single cell genome
amplification reveal extended diversity of uncultiva-
ted magnetotactic bacteria from marine and freshwater
environments. Environ Microbiol 15:1290–1301. doi:
10.1111/1462-2920.12004.
278. Lasken RS. 2012. Genomic sequencing of uncultured
microorganisms from single cells. Nat Rev Microbiol 10:
631–640.
279. Simpson JM, Santo Domingo JW, Reasoner DJ. 2002.
Microbial source tracking: state of the science. Environ Sci
Technol 36:5279–5288.
280. CDM. 2008. Middle Santa Ana River Bacterial Indicator
TMDL Agricultural Source Evaluation Plan. Santa Ana
Watershed Project Authority, http://www.swrcb.ca.gov/
rwqcb8/water_issues/programs/tmdl/docs/msar/08_03_24_
agsep_final_032108.pdf.
281. MDE. 2007. Total Maximum Daily Loads of Fecal Bacteria
for the Non-Tidal Gwynns Falls Basin in Baltimore City
and Baltimore County, Maryland, Maryland Department
of Environmental Protection. U.S. EPA Approval Date:
December 4, 2007.
282. MDEQ. 2008. Detroit River & Ecorse River E. coli Monitor-
ing to Support TMDL Development Final Report. Prepared for
U.S. EPA Region V and Michigan Department of Environ-
mental Quality Water Bureau.
283. Baffaut C, VW B. 2003. A bacteria TMDL for Shoal
Creek using SWAT modeling and DNA source tracking,
p. 35–40. In Total Maximum Daily Load (TMDL)
Environmental Regulations—II: Proceedings of the
November 8–12, 2003 Conference, Albuquerque, NM.
ASAE, St. Joseph, MI.
284. NHDES. 2004. TMDL Study for Bacteria in Hampton/
Seabrook Harbor. New Hampshire Department of Envi-
ronmental Services. U.S. EPA Approval Date: May 20,
2004.
285. NHDES. 2006. TMDL Study for Bacteria in Sand Dam
Village Pond Town Beach, Troy, New Hampshire. New
Hampshire Department of Environmental Services. U.S.
EPA Approval Date: November 8, 2006.
286. NJDEP. 2007. Total Maximum Daily Loads for Pathogens
to Address 17 Lakes in the Lower Delaware Water Region.
New Jersey Department of Environmental Protection. U.S.
EPA Approval Date: September 28, 2007.
287. Battelle. 2006. Final Report for Peconic Bay Pathogens
TMDL. Prepared for U.S. Environmental Protection
Agency Oceans and Coastal Protection Division. http://
www.peconicestuary.org/Pathogen_TMDL.pdf.
288. NDDH. 2005. Cannonball River, North Dakota Bacteria
Total Maximum Daily Load, North Dakota Department
of Health. U.S. EPA Approval Date: March 2005.
289. ODEQ. 2007. Bear Creek Watershed TMDL. Oregon
Department of Environmental Quality. U.S. EPA Appro-
val Date: October 2, 2007.
290. SDDENR. 2009. Fecal Coliform Bacteria Total Maximum
Daily Load (TMDL) for Beaver Creek, Fall River County,
South Dakota. South Dakota Department of Environment
and Natural Resources.
291. TCEQ. 2007. Bacteria Total Maximum Daily Load Task
Force Report. http://twri.tamu.edu/docs/bacteria-tmdl/Final
Report.pdf.
292. VADEQ. 2009. Bacteria TMDL Development for Hays
Creek, Moffatts Creek, Walker Creek, and Otts Creek in
Augusta County and Rockbridge County, Virginia.
VT-BSE Document No. 2008-0004. Virginia Department
of Environmental Quality. U.S. EPA Approval Date:
May 19, 2008. http://www.deq.virginia.gov/tmdl/apptmdls/
jamesrvr/haysec.pdf.
293. Benham B, Krometis L-A, Yagow G, Kline K, Dillaha T.
2011. Applications of microbial source tracking in the
TMDL process, p. 313–335. In Hagedorn C, Blanch AR,
Harwood VJ (eds), Microbial Source Tracking: Methods,
Applications, and Case Studies. Springer, New York.
294. TetraTech. 2011. Using Microbial Source Tracking
to Support TMDL Development and Implementation.
U.S. Environmental Protection Agency Region 10,
April. http://www.epa.gov/region10/pdf/tmdl/mst_for_tmdls_
guide_04_22_11.pdf.
295. FDEP. 2011. Implementation Guidance for the Fecal
Coliform Total Daily Maximum Loads Adopted by the
Florida Department of Environmental Protection.
Florida Department of Environmental Protection, March.
http://www.dep.state.fl.us/water/watersheds/docs/fcg_tool
kit.pdf.
296. Stewart JR, Ellender R, Gooch J, Jiang S, Myoda S, Weis-
berg S. 2003. Recommendations for microbial source track-
ing: lessons from a methods comparison study. J Water
Health 1:225–231.
297. Santo Domingo JW, Bambic DG, Edge TA, Wuertz S.
2007. Quo vadis source tracking? Towards a strategic frame-
work for environmental monitoring of fecal pollution.
Water Res 41:3539–3552.
298. Peed LA, Nietch CT, Kelty CA, Meckes M, Mooney T,
Sivaganesan M, Shanks OC. 2011. Combining land use
information and small stream sampling with PCR-based
methods for better characterization of diffuse sources of
human fecal pollution. Environ Sci Technol 45:
5652–5659.
3.4.4-28 ▪ MICROBIAL SOURCE TRACKING
299. Stoeckel DM, Stelzer EA, Dick LK. 2009. Evaluation of
two spike-and-recovery controls for assessment of extrac-
tion efficiency in microbial source tracking studies. Water
Res 43:4820–4827.
300. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J,
Kubista M, Mueller R, Nolan T, Pfaffl MW, Shipley
GL, Vandesompele J, Wittwer CT. 2009. The MIQE
guidelines: minimum information for publication of quan-
titative real-time PCR experiments. Clin Chem 55:
611–622.
301. Vuong NM, Villemur R, Payment P, Brousseau R,
Topp E, Masson L. 2013. Fecal source tracking in water
using a mitochondrial DNA microarray. Water Res 47:
16–30.
302. Shanks OC, Sivaganesan M, Peed L, Kelty CA, Black-
wood AD, Greene MR, Noble RT, Bushon RN, Stelzer
EA, Kinzelman J, Anan’eva T, Sinigalliano C, Wanless
D, Griffith J, Cao Y, Weisberg S, Harwood VJ, Staley
C, Oshima KH, Varma M, Haugland RA. 2012. Interla-
boratory comparison of real-time PCR protocols for
quantification of general fecal indicator bacteria. Environ
Sci Technol 46:945–953.
303. Siefring S, Varma M, Atikovic E, Wymer L, Haugland
RA. 2008. Improved real-time PCR assays for the detection
of fecal indicator bacteria in surface waters with different
instrument and reagent systems. J Water Health 6:
225–237.
304. Staley C, Gordon KV, Schoen ME, Harwood VJ. 2012.
Performance of two quantitative PCR methods for micro-
bial source tracking of human sewage and implications for
microbial risk assessment in recreational waters. Appl Envi-
ron Microbiol 78:7317–7326.
305. Cao Y, Griffith JF, Dorevitch S, Weisberg SB. 2012.
Effectiveness of qPCR permutations, internal controls
and dilution as means for minimizing the impact of inhib-
ition while measuring Enterococcus in environmental
waters. J Appl Microbiol 113:66–75.
306. Scott TM, Rose JB, Jenkins TM, Farrah SR, Lukasik J.
2002. Microbial source tracking: current methodology and
future directions. Appl Environ Microbiol 68:5796–5803.
 Microbial Source Tracking: Field Study
Planning and Implementation
JULIE KINZELMAN AND WARISH AHMED
3.4.5
Surface water is frequently polluted with pathogenic bacte-
ria, protozoa, and viruses. Nonpoint sources such as domestic
and wild animal defecation (1, 2), malfunctioning on-site
wastewater treatment (septic) systems (3, 4), stormwater run-
off (5, 6), and/or point sources such as industrial effluents
and sewage treatment plants (7, 8) are known to be potential
sources of such pollution. Surface water is commonly used for
recreation, shellfish harvesting, and drinking water supply.
Primary contact (ingestion or inhalation via consumption,
swimming, or surfing) and secondary contact (fishing and
boating) with fecally polluted water could cause gastrointes-
tinal, respiratory, and various other infections in humans
(9, 10).
Disease outbreaks following primary contact with polluted
surface waters have been well documented. For example, 138
waterborne diseases outbreaks related to recreational water
were reported from 1991 to 2006 in the United States (11).
Similarly five outbreaks associated with the recreation water
and exposures to fountains have been reported from 1992 to
2003 in the United Kingdom (12). The majority of these out-
breaks involved cases of gastroenteritis. Cryptosporidium spp.,
Giardia spp., and noroviruses were frequently reported as the
causative agents. Epidemiological studies undertaken by the
U.S. EPA and others reported that elevations in fecal indica-
tor bacteria (FIB), Enterococcus spp. in marine water, and
Escherichia coli in freshwater correlated well with increased
incidence of gastrointestinal illnesses in exposed individuals
(13–15). Therefore, within the United States and globally,
FIB standards or guidelines have been developed to prevent
human illness (16, 17). However, while protective of public
health, assessments solely employing FIB do not provide
information as to the sources of pollution. Identification
and remediation of pollution sources provides greater reduc-
tion in health risk over routine monitoring alone. Therefore,
many monitoring authorities have been augmenting routine
assessments of surface water with regulatory and/or voluntary
pollution source identification programs.
REGULATORY APPROACHES AND FECAL
SOURCE TRACKING
Restoration of polluted or degraded water is the desired out-
come of any regulatory program. For surface waters, this is
often accomplished through the development of total maxi-
mum daily loads (TMDLs). A TMDL is a calculation of the
doi:10.1128/9781555818821.ch3.4.5
3.4.5-1
maximum amount of pollutant that a water body can receive
and safely meet water quality standards. The TMDL program
was initiated by the U.S. Congress in sections 303(d) and 305
(b) of the 1972 Clean Water Act. In the early 1990s, the U.S.
EPA issued TMDL-related guidance. Pollution source identi-
fication is an important component of TMDL development
in which pollutants are identified and characterized. The
implementation of TMDLs has provided one of the driving
forces for the development of microbial source tracking
(MST) tools to identify the sources of fecal pollution in envi-
ronmental waters (18).
The identification and quantification of waterborne
pathogens from surface water can be a cumbersome task due
to the difficulty of detecting these microorganisms using tra-
ditional culture-based methods (18–20). Because of that,
fecal pollution of surface waters is usually assessed by monitor-
ing FIB such as E. coli and Enterococcus spp. (21, 22). These
FIB are abundant in the intestine of warm-blooded animals
and generally nonpathogenic; their presence in environmen-
tal waters indicates fecal pollution as well as the presence of
potentially pathogenic microorganisms. Despite their utility,
there are several limitations of FIB, for example, they do not
correlate well with the presence of pathogens in surface waters
(2, 23). FIB can persist and multiply in natural habitats,
such as sediments and aquatic vegetations (24–26). Another
major limitation of FIB is that they cannot provide informa-
tion regarding whether the pollution is sourced from sewage
or animal waste (18, 20, 27). Identifying the source of FIB
in surface waters is difficult because they are shed in the waste
of a wide variety of animals, including humans.
IDENTIFICATION OF POLLUTION SOURCES
The identification of major polluting source(s) is vitally
important to implement appropriate mitigation strategies to
restore polluted water and minimize potential public health
risks. Over the past decade, researchers have developed a
range of fecal source tracking (FST) tools that can be used
to distinguish human/sewage pollution from animals. These
tools are broadly categorised into MST and chemical source
tracking (CST) tools.
Advances in FST
The majority of the MST tools used in initial studies were
library-dependent, which requires the development of a
3.4.5-2 ▪ MICROBIAL SOURCE TRACKING
“library” of FIB (E. coli or Enterococcus spp.) from the feces of
suspected animals (sewage or animals) using various geno-
typic and phenotypic fingerprinting techniques. The under-
lying assumption of the library-dependent tools is that host
specificity of microorganisms is influenced by selective pres-
sures within the host animal (27, 28). Phenotypic and geno-
typic fingerprints of FIB are then compared to the library to
identify their likely host sources (29).
Despite the initial successful application of these library-
dependent MST tools (reviewed in (20)), several questions
have arisen regarding their application for FIB source track-
ing. For instance, the size and the representativeness of
the library need to be addressed prior to developing a library
for optimal performance. It has further been reported that
temporal and geographical variability exists in FIB popu-
lations, which may restrict their use for a universal library.
The other disadvantages of library-dependent tools include
lack of host specificity of FIB and their relative persis-
tence in water as well as complexity in statistical analysis
(18, 20, 30).
Another set of MST tools, not requiring the development
of a library, are known as library-independent tools. These
tools involve detection or quantification of specific genetic
marker(s) associated with microorganisms from various host
sources (30). With the aid of these tools, the presence or
absence of fecal pollution from a particular source in a water
sample can be determined. Library-independent tools could
be categorized into four groups: (1) bacterial markers such
as sewage-associated or animal feces associated Bacteroides
(31, 32); (2) viral markers such as sewage and bovine feces–
associated adenoviruses and polyomaviruses (33–35); (3)
bacterial toxin markers such as pig wastewater associated
ST1b (36), cattle wastewater-associated LTIIa E. coli toxin
gene (37), and the sewage-associated enterococcus surface
protein (esp) gene found in Enterococcus faecium (38); mito-
chondrial DNA markers such as such as human, ovine,
bovine, and swine (39); and (4) pathogenic protozoa such
as Cryptosporidium spp. (40).
Several community profile-based library-dependent MST
tools such as terminal restriction fragment length polymor-
phism, DNA microarray, and next-generation sequencing
have been used to identify the sources of fecal pollution in
water (31, 41–43). The most important features of these tools
are they target multiple source tracking taxonomic units
rather than single taxonomic unit such as Bacteroides and Pre-
votella (44). Regardless of tool, library-dependent or library-
independent, an ideal MST candidate should possess certain
desirable characteristics: host specificity, sensitivity, distribu-
tion (equal numbers in the feces), good correlation with the
pathogens, and temporal/geographical stability, among others
(45).
Chemical source tracking (CST) markers include organic
wastewater chemicals, caffeine, fecal sterols, human pharma-
ceuticals, personal care products, and optical brighteners (20,
27, 46, 47). Caffeine, because of its pervasive presence in bev-
erages and pharmaceuticals, has been suggested as an indica-
tor of human sewage (48). Fecal sterols have been used to
distinguish human from animal feces (49). During the past
decade, a number of studies have extensively surveyed the
prevalence of pharmaceuticals and personal care products
in sewage effluents and receiving surface waters (50, 51). Per-
sistent markers such as acesulfame are useful for tracing the
pathways of treated sewage, whereas biodegradable com-
pounds such as caffeine are indicators of untreated wastewater
discharge in fresh water due to sewer overflow (52). CST
markers, such as these, are predominantly used to identify
the ingress of sewage into environmental waters due to their
association with anthropogenic sources (users of such sub-
stances) versus animal sources (nonusers).
“Low-Tech” Source Tracking Tools
Advanced MST tools, using FIB or genetic markers (library-
dependent or library-independent) have provided researchers
and those interested in pollution source tracking with the
ability to increase their level of discrimination, thereby
improving the outcome of the human exposure interventions,
for example, identification and amelioration of the fecal
source or prevention of contact. However, molecular analyses
are costly and require proximity to facilities capable of under-
taking the assays, and results generated must be taken in con-
text with multiple lines of evidence (53). Other analytical
tools are available for consideration to those undertaking pol-
lutant source identification projects that are less costly and
will provide for targeted use of higher level MST tools. Exam-
ples of these approaches include physical assessments,
expanded monitoring using traditional FIB and culture-based
assays, and measurements of environmental constituents (i.e.,
pH, total suspended solids, nutrients, and dissolved oxygen).
Using a toolbox of various MST/CST techniques creates a
multiple lines of evidence approach, which increases the
robustness and accuracy of the data set and the reliability of
source prediction (54, 55). The use of redundant MST tools
(more than one human marker or a combination of markers
plus library-dependent tools) may be required to increase con-
fidence (54) in the correct or “true” assignation of the pollu-
tion source. Other techniques, such as quantitative microbial
risk assessment (QMRA) and sanitary surveys combine the
data generated from several applications/analyses to make a
prediction of health risk and guide mitigative approaches,
based on identified pollution sources (a multiple lines of evi-
dence approach) (56, 57).
The choice of tool(s) to be employed depends on several
factors including environmental complexity (likelihood of
multiple sources of contamination), availability of historic
data, previous application or validation of tools, level of dis-
crimination required, accessibility to analytical facilities
(and level of associated technical expertise), and associated
cost. Mining of historic monitoring data may lend focus to
future studies. Sanitary surveys are guided data collection
tools ( providing a uniform approach to data collection across
multiple agencies) which consider not only quantification of
FIB but also the conditions under which they rise and fall. Pri-
mary assessments of hydrodynamic and meteorological data,
in conjunction with measurements of FIB and chemical con-
stituents in the aquatic environment, can tease out multiple
pollution sources in complex environments and aid in clearly
identifying when and where MST/CST efforts can be most
beneficially applied while managing the costs associated
with large-scale source identification projects. Many of these
environmental observations or laboratory-based tests may be
achieved using freely available web-based data (e.g., National
Oceanographic Atmospheric Association National Oceano-
graphic Data Center; http://www.nodc.noaa.gov/access/
index.html), field-based estimations or measurements (e.g.,
U.S. EPA beach sanitary survey user manuals for fresh and
marine waters; http://water.epa.gov/type/oceb/beaches/sani
tarysurvey_index.cfm#great), and easy-to-perform/cost-
effective assays requiring less technical and analytical exper-
tise (e.g., IDEXX water microbiology; http://www.idexx.
com/view/xhtml/en_us/water/water-testing-solutions.jsf?con
versationId=296661). Stakeholders should work coopera-
tively with those undertaking the study to determine goals
 3.4.5. Microbial Source Tracking: Field Study Planning and Implementation ▪ 3.4.5-3
and objectives prior to making final decisions on study design,
inclusive of field sampling, generation of results, and data
analysis.
Field Study Design
Sampling locations, study duration, and sampling frequency
(to account for seasonality and other forms of temporal
variation), are critical to a well-crafted source tracking study
(58). Samples must be collected in manner that adequately
characterizes the study site(s) and is representative of and tar-
geted toward the potential pollution sources in question
(56). For example, when conducting watershed or catchment
assessments, it is important to select sampling locations that
represent different land uses (e.g., urban, suburban, or agricul-
tural), discrete branches or confluences of rivers/streams,
effluent or stormwater discharge points, and the intersection
between tributaries and receiving waters. Temporal time
scales could include seasonal assessments, wet versus dry
weather sampling events, or time series samples (to assess
diurnal variation, [59]). Sampling reliability has been shown
to be dependent on the dates and/or times of sample collec-
tion (60).
Spatial distribution studies are another way to increase the
probability of achieving a representative sampling effort (61).
Spatial distribution studies involve the collection of multiple
samples for a single analyte (or a suite of related analytes) at a
bathing beach, along a tributary segment, or throughout
municipal infrastructure distribution systems. For example, a
screening approach based on a hierarchy of sampling loca-
tions and test methods may be used to differentiate pollution
sources discharging via piped infrastructure by categorizing
them into those releasing (a) pollutants pathogenic or toxic
to humans, (b) substances adversely impacting ecosystem
health, and (c) unpolluted water ( process water or precipita-
tion) (62–64).
One recommended approach to field study design is to sep-
arate it out into discrete phases: that is, site assessment to
determine sampling locations as well as potential sources
of pollution (via GIS mapping, visualization, or the sanitary
survey process), followed by initial field sampling/sample
analysis, and finally confirmatory testing (63, 65, 66). Field
sampling should start with those locations having the greatest
pollutant source potential to those with lesser relative contri-
butions. For example, point sources such as municipal infra-
structure or industrial outlets should be screened initially as
should the mouths of tributaries and streams, as these have
been demonstrated to be frequent delivery mechanisms of
FIB to the environment. Likewise test methods should begin
with exploratory or screening techniques (FIB assessments,
detergents, ammonia, specific conductance, turbidity, for
example) and then focus on more discriminatory or confirma-
tory procedures, such as fecal sterols (49), genetic marker test-
ing (i.e., human-specific Bacteroides), or DNA profiling
though the use of library dependent MST tools based on
those initial results (30, 66). Decision trees (see Fig. 1 for
an example) are an effective tool for field study planning as
well as determining the level of discrimination desired
when employing MST tools. Following this paradigm, MST
has been appropriately and successfully employed in the con-
firmation of illicit discharges into municipal stormwater sys-
tems as part of the TMDL process and in the identification
of potential pollution sources impacting coastal recreational
waters (35, 67–69). On the other hand, missed or erroneously
identified/implicated sources will result in continued expo-
sure risk and the potential for wasted capital investment for
mitigation activities.
Data Accumulation and Statistical Analysis
A well-crafted and executed study design will yield a wealth of
data that can be used in the discrimination of pollution sour-
ces in aquatic environments. Care must be taken to present
the data in a manner that is reflective and supportive of the
initial study design. Large datasets, such as those generated
by microarrays and next generation sequencing of bacterial
community analysis may be difficult to handle and analyze
due to their vast size and complexity (70, 71). Whenever pos-
sible it is important to leverage historic data, which frequently
exists in the form of FIB concentrations, for example, fecal
coliforms, E. coli, or enterococci. Historic data can assist in
the determination of spatial distribution, trends over time
(temporal variability), or responses to environmental condi-
tions such as extreme weather events (event-based), for exam-
ple, heavy precipitation. Using a single season’s worth of data,
or data confined to a project of finite, limited duration as a
result of budgetary or other constraints, may preclude the abil-
ity of the researcher to capture the full range of possible con-
ditions under which the pollution source has the ability to
impact water quality or misconstrue the geographic spread
or influence of the source.
Whenever historic data is utilized it must conform in type
and configuration to the most recently collected data for the
purpose of statistical analysis. There are many statistical soft-
ware programs available for the analysis of data (e.g., SPSS,
SYSTAT, and SigmaPlot) as well programs and web-based
applications which are useful for visualizing and mapping
data spatially (ArcGIS and Google Earth). A suite of poten-
tial explanatory statistical analyses for gauging the response of
FIB levels to environmental conditions, comparing data from
multiple sampling periods, or within/across sampling trans-
ects or locations, includes: paired t-test, analysis of variance,
Mann-Whitney rank sum test, Wilcoxon signed rank test,
Spearman rank order correlation, regression analyses, Pear-
son’s correlation, and others. As an example, Shibata (72)
and Belanche and Blanch (73) used similar statistical analyses
in the interpretation of data from a study comparing conven-
tional and rapid monitoring tools and source tracking markers
at a marine beach dominated by non–point source pollution.
Alternatively, Dubinsky et al. (42) applied phylogenetic
microarray analysis to discriminate sources of fecal pollution.
Once the appropriate statistical analyses have been per-
formed, the next step is interpretation of the data. It is impor-
tant to note that when looking at FIB fluctuations in response
to environmental variables, the proper interpretation is one
of correlation, not causation. In other words, one is determin-
ing the amount of variability in FIB levels that can be
explained by precipitation events rather than stating that ele-
vations in FIB are caused by rainfall. In some instances these
correlations may not be straightforward or strong. When this
occurs, a weight of evidence approach may be used to associ-
ate the probability of increased human health risk to a likely
source or pollution events to a particular set of circumstances
(74, 75). In complex systems, the results of multiple or redun-
dant analyses, including MST, may be used to focus mitiga-
tion activities on sources with (a) the greatest relative
contribution or (b) most likely to result in adverse human
health outcomes if left unabated (30, 76).
TRANSLATING STUDY RESULTS INTO
ACTIONABLE ITEMS
The end product of any source tracking study is to identify
pollution sources, generally with the goal of focusing reme-
diation activities and eliminating that source. Examining
 3.4.5-4 ▪ MICROBIAL SOURCE TRACKING
FIGURE 1 Decision tree approach to identifying pollution sources based on fecal indicator bacteria (FIB) density, relationship to environmental conditions, physical assessments and proximity to
potential pollution sources. PPT, precipitation; FIB, fecal indicator bacteria; STDs, standards; IC, impervious cover; MST, microbial source tracking.
doi: 10.1128/9781555818821.ch3.4.5.f1
 3.4.5. Microbial Source Tracking: Field Study Planning and Implementation ▪ 3.4.5-5
recently collected and historic data, assessing adjacent land
use, and analyzing microbial indicator response under a vari-
ety of environmental conditions allow for the evaluation of
public health risk more effectively than by using FIB levels
alone (77). Source tracking data, as part of the larger sanitary
survey/inspection process, can assist responsible authorities in
the decision-making process to target mitigative actions (30,
56, 69). For example, Kinzelman and McLellan (69) used
sanitary survey data in addition to MST techniques to iden-
tify gulls as likely sources of pollution impacting coastal rec-
reational water quality (69). Converse et al. (78) went on
to further demonstrate that the removal of gulls from recrea-
tional water led to a direct reduction in the amount of FIB and
human pathogens recovered from nearshore water (78). The
confirmation of gulls as a likely source of pollution has led to a
variety of best management practices designed to alter their
behavior or reduce the transfer of their feces from beach sands
to surface water (69).
Decision trees are effective tools not only in the analytical
decision-making process but also for making the outcomes of
statistical analyses and conclusions drawn more accessible to
the stakeholder (79). Take for example Fig. 1, based on the
assessment of an urban river segment. In the figure we start
with a specific sampling location in which FIB concentra-
tions are assessed as a function of wet weather events, includ-
ing both 24 and 48 h antecedent precipitation. At this site,
there is some but not a 100% frequency of elevation in FIB
as a function of rainfall, indicating that not all of the water
quality failures at this location are correlated with wet weather
events, that is, precipitation-mediated delivery of stormwater
runoff from a variety of land uses. This would lead the inves-
tigator to also consider dry weather sources of pollution. In
this case the dry weather source was discharge from a munic-
ipal stormwater outfall, which was associated with greater
than 40% of the water quality standard exceedances during
the study period. Dry weather discharge from municipal con-
veyance systems can occur due to infrastructure degradation,
misconnection or cross-connection with sanitary sewers, or
from the discharge of process water from industry (such as
cooling towers). The identification of a potential illicit dis-
charge at this location using the decision tree approach would
bring to bear available MST/CST tools as confirmation of the
findings prior to final mitigation plan approval. Decision trees
may also aid in prioritizing and recommending site-specific
exposure interventions based on the severity of risk to human
or ecosystem health. In the provided example (Fig. 1), red
dashed lines lead to high or major priorities, those indicative
of potential human sewage sources, while yellow “yes”
responses lead to medium priority mitigation measures, fre-
quently related to better management of stormwater runoff.
A Multiple Barrier Approach to Human
Exposure Intervention
A multiple barrier approach designed to prevent exposure
may be an advantageous strategy to reduce human health
risk from surface waters polluted with pathogenic bacteria,
protozoa, and viruses. Once the source and means by which
humans come in contact with it have been determined
through FST studies, the implementation of appropriate bar-
riers may commence. Exposure interventions can take many
forms, from better monitoring paradigms which take into
account factors other than FIB concentration (i.e., Annapolis
Protocol or QMRA) to hard engineered structures to natural-
ized control measures to public education (76, 80, 81). Hard
engineered solutions can take the form of infrastructure
replacement or repair, constructed wetlands, retention/deten-
tion basins, and diversions (82–84). Examples of naturalized
control measures include vegetated buffer strips, dune restora-
tion, and vegetated swales (85). Public education can stand
alone in its own right or take the form of stakeholder engage-
ment. When conducting source tracking studies, whose out-
come or purpose is to alter human behavior and/or the
environment in which people reside (built or natural), it is
of utmost importance to communicate goals/objectives,
study results, and interpretations thereof in a manner that is
understandable to the stakeholder. Stakeholder groups, once
engaged in the process, may not only be accepting of change
but willing to volunteer their time and expertise to the proc-
ess (86). For example, education and outreach resulted in
volunteer participation in, and acceptance of, the installation
of constructed wetlands to mitigate stormwater pollution at
a high-priority public bathing beach on Lake Michigan (in the
United States, coastal beach priority designations are assigned
based on potential health risk, beach miles, usage, and eco-
nomic importance) (87). This mitigation strategy was ini-
tially a point of contention to some residents for fear of an
increase in vectorborne disease due to standing water (88).
CONFOUNDING FACTORS: ANALYTICAL,
FINANCIAL, LOGISTICAL, CULTURAL
BARRIERS TO IMPLEMENTATION
Even well-conceived and well-planned field studies and sam-
pling plans are subject to limitations; for example, cost con-
straints might prohibit the collection of a representative
number of samples (temporal or spatial), in which case either
Type I or Type II statistical error may well occur. Other con-
founders related to the preanalytical process include poor
sample quality either through lack of expertise or inappro-
priate sample handling, holding time/temperature, or process-
ing. All analytical methods require strict adherence to quality
assurance, quality control, and procedural requirements,
therefore interlaboratory studies and future regulatory proce-
dures require detailed standard operating procedures (e.g.,
89, 90). When considering molecular MST tools, one must
consider both preanalytical (the source, location, timing,
transport, and storage of the sample) and analytical processes
(choice of method, accuracy, precision, and source-specific
interfering substances, if any). Constituents within the sur-
face water environment can result in competition for the tar-
get substrate, target underestimation, or complete inhibition
of the assay (91, 92, also see Chapter 3.4.2).
Improper data analysis and/or interpretation may result in
the misidentification of pollutant sources or ascribe dispro-
portionate relative contributions from various contamination
sources, resulting in the application of inadequate or unsuit-
able management or intervention measures. The results of
any source tracking data, in particular those arising from
MST methodologies, must be taken in context with histori-
cal, perceptual, observational, and other measurable parame-
ters. No analytical test method is all-telling and infallible, and
practitioners must avoid false expectations of finding a “smok-
ing gun” with respect to pollution source identification based
a single or even a suite of environmental and laboratory meas-
urements. The investigation and identification of pollution
events and sources adversely impacting surface waters is a
process which requires due diligence on the part of the
researcher and patience on the part of the stakeholder.
Reliance on inadequate data and/or unsound professional
judgment may result in the unnecessary expenditure of time
and funds in pursuit of unsuitable remediation efforts or
3.4.5-6 ▪ MICROBIAL SOURCE TRACKING

incomplete abatement of the pollution source. Proper target-
ing of research data in support of the initial study design/
objectives has a higher probability of successful implementa-
tion of best management practices (69). Even with proper
application, best management measures may sometimes fail.
Cultural traditions, religious beliefs, and prohibitive costs
may preclude the implementation of the some remediation
activities (93). Fluctuations in environmental conditions,
for example, as a result of climate change, may change the
relationship between FIB, pathogens, MST markers, and
loading mechanisms such as precipitation events. Although
uncertainty complicates planning efforts, attempts have been
made to provide planning and technical guidance for the
design and implementation of climate-smart restoration proj-
ects (94, 95).
CASE STUDIES
The following brief case studies represent examples of the suc-
cessful development, application, and execution of MST/
CST studies in the identification of pollution sources
adversely impacting surface water quality and posing human
health hazards. The list of MST and CST tools that were
chosen as a part of this toolbox approach, and the results
thereof, are illustrative of some of the advantages of deploying
MST tools as part the investigative process (Table 1).
Identification and Remediation of Illicit Discharges
to an Urban River (Racine, WI, USA)
Routine monitoring of stormwater outfalls, which is required
of municipalities maintaining a National Pollutant Discharge
TABLE 1 Microbial and chemical source tracking tools (case
studies) used to identify pollution sources at the case study sites,
Racine, WI (USA) and various urban catchments in Brisbane,
Australiaa
Permit stormwater discharge permit (http://cfpub.epa.gov/
npdes/stormwater/munic.cfm), was conducted annually along
the Root River in Racine, WI (USA). Monitoring locations
were selected primarily to correspond to sizable stormwater
basins directly discharging to the river within the city limits.
Sampling was conducted once weekly throughout the year,
weather permitting. Samples were not collected if there was
no flow or ice/snow precluded safe access to the sampling
location. Screening tests included both microbial (FIB) and
chemical ( pH, specific conductance, detergents [methylene
blue active substances method], and total residual chlorine)
source tracking tools. Environmental conditions, such as
antecedent precipitation, and physical observations (e.g.,
presence of overt fecal contamination, foaming, excessive
plant growth, or degraded physical integrity of the pipe)
were noted each time that sample collection occurred. Four
of 10 outfalls exhibited frequent dry weather flow containing
elevated levels of FIB, detergents, and, intermittently, resid-
ual chlorine. Confirmatory testing was performed at these
sites using human-associated (31) and general Bacteroidales
markers (32). Three of the four sites had ratios of human
(BacHuman) to general Bacteroides (BacSpp) marker in excess
of 8%, indicating that sanitary sewer infiltration into the
stormwater conveyance system was likely (Dr. Sandra McLel-
lan, personal communication). For the purpose of this study, a
ratio of BacHuman:BacSpp of 2–8% (or greater) indicated a
sanitary source, that is, the concentration of BacHuman
marker was divided by the concentration of BacSpp marker
and multiplied by 100. Televising and/or dye testing of the
four sites with indications of human sewage contamination
revealed compromised infrastructure integrity at two of the
sites, an illicit connection to a residential unit at another
site (site 16, Fig. 2), and evidence of rodent infestation at
the fourth site. Lining of the compromised infrastructure
and separation of the illicit connection has resulted in the
near elimination of dry weather flow, improved water quality,
and removal of human health hazards (J. Kinzelman, personal
communication).

Field study
Urban Evidence of Sewage Pollution in Stormwater
site(s)
Source tracking methods
Storm water
catchments Samples in Australia
(AUS) The occurrence of sewage pollution in urban stormwater
outfalls (US)
samples from six catchments across Australia was assessed by
Microbial using both MST and CST markers (96). Stormwater samples
E. coli spp. X (n = 23) collected from catchments located in Brisbane, Mel-
Enterococcus spp. X bourne, and Sydney were assessed for the presence of human
Bacteroides (total) X adenovirus (HAdVs), human polyomavirus (HPyVs), and
the sewage-associated markers, Metahonobrevibacter smithii
Bacteroides (HF183) X X
nifH and Bacteroides HF183. In addition, the samples were
Human adenovirus X also tested for the presence of pharmaceuticals and food addi-
Human polyomavirus X tives ( paracetamol [acetominophen], salicylic acid, acesul-
M. smithii nifH X fame, and caffeine) associated with sewage pollution. Levels
Chemical of Enterococcus spp. in all stormwater samples exceeded the
recommended limits for the lowest water quality category D
pH X
(<501 Enterococcus spp. per 100 ml) under the Australian
Specific conductance X guidelines for managing risks in recreational water (97), with
Detergents X 48% samples containing >10,000 CFU per 100 ml water.
Total residual chlorine X Among the 23 stormwater samples collected, 21 (91%)
Pharmaceuticals ( paracetamol X were positive for six to eight sewage-associated markers and
and salicylic acid) the remaining 2 samples were positive for five and four
markers, respectively. This study demonstrated that storm-
Food additives (acesulfame and X
water outfalls located in urban catchments in Australia are
caffeine)
widely polluted by human sewage, as indicated by the ubiqui-
a
Bolded text represents confirmatory MST tools that target human- tous presence of both MST (including HAdVs and HPyVs)
associated markers. and CST markers. A very good consensus (>80%) between
 Key: Urban River
IC = Impervious Cover Site #16
OF = Stormwater Outfall
PPT = Precipitation
No E.coli responds Yes
DWF = Dry Weather Flow to 24hr PPT?
R2 = Degree of Determination (Regression) (r2>0.65)

= Decision Tree Path

E. coli responds to
No 48 hr PPT? Yes
(r2>0.5)
High OFs / Eroded stream
IC?? animal banks / lack of

3.4.5. Microbial Source Tracking: Field Study Planning and Implementation ▪ 3.4.5-7
E. coli feedlots? buffer strip?
E. coli exceedance >
No exceedance Yes No 50% within 24 Yes
>40% in dry hr PPT?
weather? &
>50% in 24 hr
PPT?

E. coli Low priority:

exceedance > Pollution from Nearby look at localized Low-Med priority: Reduce
No 40% in 24 hr Yes upstream sites. infrastructure areas of storm water runoff – infiltration
PPT? Insufficient buffer with DWF? improvement systems/buffers, etc.
strips? Positive hits on
CST markers ?

Medium Priority:
Storm water runoff Low-Med priority:
Medium priority:
management, Consider stream bank
Low priority: Reduce runoff in Med–High priority: Major priority: improve sites improvements –
Look at localized local area, consider Investigate E.coli Investigate source upstream vegetate/buffer strips
areas of stream bank sources upstream to of DWF using
improvement improvements determine if other human MST
sources exist. markers.

Site #16 was the subject of a MST study to determine the source of frequent dry weather E. coli elevaons. E.
coli exceeded 10,000 MPN/100 ml in > 40% of samples. The rao of human Bacteroides to general Bacteroides,
expressed as a percent, was consistenly greater than 2 (indicave of sanitary infiltraon). Positive hits for
detergents and phosphorus approacehd 100% of samples collected. A sanitary sewage misconnecon was
found in a nearby house in the ouall drainage basin. The pipe was fixed and the ouall no longer flows during
dry weather. Monitoring of upstream OF to connue. Site #16 did have a signficant posive relaonship
(p<0.05) with 24 and 48 hr PPT, but the R2 values are not above the prescribed thresholds. 16

FIGURE 2 Utilization of a decision tree approach in determining the source of fecal contamination at an urban river site; ultimately identifying and illicit connection. IC, impervious cover;
OF, stormwater outfall; PPT, precipitation; DWF, dry weather flow; R2, degree of determination (regression). doi: 10.1128/9781555818821.ch3.4.5.f2
3.4.5-8 ▪ MICROBIAL SOURCE TRACKING
the occurrence of the HF183, HAdVs, acesulfame, paraceta-
mol, and caffeine was observed. This study shows the feasibil-
ity and benefits of employing a toolbox of FST markers to
identify sewage pollution in stormwater and characterization
of potential health risks from exposure to stormwater.
REFERENCES
1. Ibekwe AM, Grieve CM. 2003. Detection and quantifica-
tion of Escherichia coli O157:H7 in environmental samples
by real-time PCR. J Appl Microbiol 94:421–431.
2. Hörman A, Rimhanen-Finne R, Maunula L, von
Bonsdorff CH, Torvela N, Heikinheimo A, Hänninen
ML. 2004. Campylobacter spp., Giardia spp., Cryptosporidium
spp., noroviruses, and indicator organisms in surface water in
Southwestern Finland, 2000–2001. Appl Environ Microbiol
70:87–95.
3. Ahmed W, Neller R, Katouli M. 2005. Evidence of septic
systems failure determined by a bacterial biochemical finger-
printing method. J Appl Microbiol 98:910–920.
4. Hartel PG, Summer JD, Hill JL, Collins JC, Entry JA
Segers WI. 2002. Geographic variability of Escherichia
coli isolates from animals in Idaho and Georgia. J Environ
Qual 31:1273–1278.
5. Noble RT, Griffith JF, Blackwood AD, Fuhrman JA,
Gregory JB, Hernandez X, Liang X, Bera AA, Schiff K.
2006. Multitiered approach using quantitative PCR to track
sources of fecal pollution affecting Santa Monica Bay, Cali-
fornia. Appl Environ Microbiol 72:1604–1612.
6. Sidhu JPS, Hodgers L, Ahmed W, Chong MN, Toze S.
2012. Prevalence of human pathogens and indicators in
stormwater runoff in Brisbane, Australia. Water Res 46:
6652–6660 .doi:10.1016/j.watres.2012.03.012.
7. Rajal VB, McSwain BS, Thomson DE, Leutenegger CM,
Wuertz S. 2007. Molecular quantitative analysis of human
viruses in California stormwater. Water Res 41:4287–4298.
8. Sercu B, Van De Werfhorst LC, Murray J, Holden PA.
2009. Storm drains are sources of human fecal pollution dur-
ing dry weather in three urban Southern California water-
sheds. Environ Sci Technol 43:293–298.
9. Gaffield SJ, Goo RL, Richards LA, Jackson RJ. 2003. Pub-
lic health effects of inadequately managed stormwater run-
off. Am J Public Health 93:1527–1533.
10. U.S. Environmental Protection Agency. 1986. Ambient
water quality criteria for bacteria. 1986 EPA 440-5-84-002.
U.S. Environmental Protection Agency, Washington, DC.
11. Schets FM, De Roda Husman AM, Havelaar AH. 2011.
Disease outbreaks associated with untreated recreational
water use. Epidemiol Infect 139:1114–1125.
12. Smith A, Reacher M, Smerdon W, Adak GK, Nichols G,
Chalmers RM. 2006. Outbreaks of waterborne infectious
intestinal disease in England and Wales, 1992–2003. Epide-
miol Infect 134:1141–1149.
13. Prüss A. 1998. Review of epidemiological studies on health
effects from exposure to recreational water. Int J Epidemiol
27:1–9.
14. Wade TJ, Calderon RL, Sams E, Beach M, Brenner KP,
Williams AH, Dufour AP. 2006. Rapidly measured indica-
tors of recreational water quality are predictive of swimming-
associated gastrointestinal illness. Environ Health Perspect
114:24–28.
15. Kay D, Fleisher JM, Salmon RL, Jones F, Wyer MD, God-
free AF, Zelenauch-Jacquotte Z, Shore R. 1994. Predicting
likelihood of gastroenteritis from sea bathing: results from
randomized exposure. Lancet 344:905–990.
16. European Union. 2006. Directive 2006/7/EC of the Euro-
pean Parliament and of the Council of 15 February 2006.
Off J Eur Union L 64/37.
17. U.S. Environmental Protection Agency. 2011a. Recrea-
tional Water Quality. Document EPA-820-D-11–002.
U.S. Environmental Protection Agency, Office of Water,
Washington, DC.
18. Field KG, Samadpour M. 2007. Fecal source tracking, the
indicator paradigm, and managing water quality. Water Res
41:3517–3538.
19. Payment P, Franco E. 1993. Clostridium perfringens and
somatic coliphages as indicators of the efficiency of drinking
water treatment for viruses and protozoan cysts. Appl Environ
Microbiol 59:2418–2424.
20. Stoeckel DM, Harwood VJ. 2007. Performance, design, and
analysis in microbial source tracking studies. Appl Environ
Microbiol 73:2405–2415.
21. U.S. Environmental Protection Agency. 1997. Method
1600: membrane filter test method for enterococci in
water. EPA/821/R-97/004. U.S. Environmental Protection
Agency, Office of Water, Washington, DC.
22. U.S. Environmental Protection Agency. 2002. Method
1603: Escherichia coli (E. coli) in water by membrane filtra-
tion using modified membrane-thermotolerant Escherichia
coli agar (modified mTEC). EPA/821/R-02/023. U.S. Envi-
ronmental Protection Agency, Office of Water, Washing-
ton, DC.
23. Anderson KL, Whitlock JE, Harwood VJ. 2005. Persis-
tence and differential survival of fecal indicators bacteria
in subtropical waters and sediments. Appl Environ Microbiol
71:3041–3048.
24. Byappanahalli M, Fowler M, Shivley D, Whitman R.
2003. Ubiquity and persistence of Escherichia coli in a
Midwestern coastal stream. Appl Environ Microbiol 69:
4549–4555.
25. Solo-Gabriele HM, Wolfert MA, Desmarais TR, Palmer
CJ. 2000. Sources of Escherichia coli in a coastal subtropical
environment. Appl Environ Microbiol 66:230–237.
26. Whitman RL, Nevers MB. 2003. Foreshore sand as a source
of Escherichia coli in nearshore water of a Lake Michigan
Beach. Appl Environ Microbiol 69:5555–5562.
27. Scott TM, Rose JB, Jenkins TM, Farrah SR, Lukasik J.
2002. Microbial source tracking: current methodology and
future directions. Appl Environ Microbiol 68:5796–5803.
28. Wiggins BA. 1996. Discriminant analysis of antibiotic resis-
tance patterns in fecal streptococci, a method to differentiate
human and animal sources of fecal pollution in natural
waters. Appl Environ Microbiol 62:3997–4002.
29. Harwood VJ, Whitlock J, Withington V. 2000. Classifica-
tion of antibiotic resistance patterns of indicator bacteria by
discriminant analysis: use in predicting the source of fecal
contamination in subtropical waters. Appl Environ Microbiol
66:3698–3704.
30. U.S. Environmental Protection Agency. 2005. Microbial
source tracking guide. Document EPA/600/R-05/064. U.S.
Environmental Protection Agency, Washington, DC.
31. Bernhard AE, Field KG. 2000. A PCR assay to discriminate
human and ruminant feces on the basis of host differences in
Bacteriodales-Prevotella genes encoding 16S rRNA. Appl
Environ Microbiol 66:4571–4574.
32. Layton A, McKay L, Williams D, Garrett V, Gentry R,
Sayler G. 2006. Development of Bacteroides 16S rRNA
gene TaqMan-based real-time PCR assays for estimation of
total, human, and bovine fecal pollution in water. Appl Envi-
ron Microbiol 72:4214–4224.
33. Fong TT, Griffin DW, Lipp EK. 2005. Molecular assays for
targeting human and bovine enteric viruses in coastal waters
and their application for library-independent source track-
ing. Appl Environ Microbiol 71:2070–2078.
34. Hundesa A, Maluquer de Motes C, Bofill-Mas S,
Albinana-Gimenez N, Girones R. 2006. Identification of
human and animal adenoviruses and polyomaviruses for
 3.4.5. Microbial Source Tracking: Field Study Planning and Implementation ▪ 3.4.5-9
determination of sources of fecal contamination in the envi-
ronment. Appl Environ Microbiol 72:7886–7893.
35. McQuaig SM, Scott TM, Lukasik JO, Paul JH, Harwood
VJ. 2009. Quantification of human polyomaviruses JC virus
and BK virus by TaqMan quantitative PCR and comparison
to other water quality indicators in water and fecal samples.
Appl Environ Microbiol 75:3379–3388.
36. Khatib LA, Tsai YL, Olson BH. 2003. A biomarker for the
identification of swine fecal pollution in water, using the
STII toxin gene from enterotoxigenic Escherichia coli. Appl
Microbiol Biotechnol 63:231–238.
37. Chern EC, Tsai Y, Olson BH. 2004. Occurrence of genes
associated with enterotoxigenic and enterohemorrhagic
Escherichia coli in agricultural waste lagoons. Appl Environ
Microbiol 70:356–362.
38. Scott TM, Jenkins TM, Lukasik J, Rose JB. 2005. Poten-
tial use of a host associated molecular marker in Enterococcus
faecium as an index of human fecal pollution. Environ Sci
Technol 39:283–287.
39. Baker-Austin C, Rangdale R, Lowther J, Lees DN. 2010.
Application of mitochondrial DNA analysis for microbial
source tracking purposes in shellfish harvesting waters. Wat
Sci Technol 61(1):1–7.
40. Jellison KL, Hemond HF, Schauer DB. 2002. Sources and
species of Cryptosporidium oocysts in the Wachusett Reser-
voir Watershed. Appl Environ Microbiol 68:569–575.
41. Savichtcheva O, Okage S. 2009. Qualitative and quan-
titative estimation of host-specific fecal pollution using
Bacteroides-Prevotella 16S rRNA genetic markers by T-RFLP
and real-time PCR analyses. Wat Sci Tech 59(9):
1831–1840.
42. Dubinsky EA, Esmaili L, Hulls JR, Cao Y, Griffith JF,
Andersen GL. 2012. Application of phylogenetic microar-
ray analysis to discriminate sources of fecal pollution. Environ
Sci Tech 45(8):4340–4347.
43. Shanks OC, McLellan S, Huse SM, Sogin ML. 2010. Char-
acterization of microbial population structures in recrea-
tional waters and primary sources of fecal pollution with a
next-generation sequencing approach. Chapter 9. In Sen
K, Ashbolt NJ (eds), Environmental Microbiology: Current
Technology and Water Applications. Caister Academic Press,
Norwich, UK.
44. Unno T, Jang J, Han D, Kim JH, Sadowsky MJ, Kim OS,
Chun J, Hur HG. 2010. Use of barcoded pyrosequencing
and shared OTUs to determine sources of fecal bacteria in
watersheds. Environ Sci Technol 44:7777–7782.
45. Ahmed W, Masters N, Toze S. 2012. Consistency in the
host specificity and host sensitivity of the Bacteroides
HF183 marker for sewage pollution tracking. Lett Appl
Microbiol 55:283–289.
46. Dickerson J, Hagedorn C, Hassall A. 2007. Detection and
remediation of human-origin pollution at two public
beaches in Virginia using multiple source tracking methods.
Wat Res 41:3758–3770.
47. Haack S, Duris J, Fogharty L, Kolpin D, Focazio M, Fur-
long E, Meyer M. 2009. Comparing wastewater chemicals,
indicator bacteria concentrations, and bacterial pathogen
genes as fecal pollution indicators. J Environ Qual 38:248–
258.
48. Burkhardt M. 1999. Determination of submicrogram-per
liter concentrations of caffeine in surface water and ground-
water samples by solid-phase extraction and liquid chroma-
tography. J AOAC Int 82:161–166.
49. Leeming R, Ball A, Ashbolt N, Nichols P. 1996. Using
fecal sterols from humans and animals to distinguish fecal
pollution in receiving waters. Water Res 30:2893–2900.
50. Glassmeyer ST, Furlon ET, Kolpin DW, Cahill JD, Zaugg
SD, Werner SL, Meyer MT, Kryak DD. 2005. Transport of
chemical and microbial compounds from known wastewater
discharges: potential for use as indicators of human fecal
contamination. Environ Sci Technol 9:5157–5169.
51. Nakada N, Kiri K, Shinohara H, Harada A, Kuroda K,
Takizawa S, Takada H. 2008. Evaluation of pharmaceuti-
cals and personal care products as water-soluble molecular
markers of sewage. Environ Sci Technol 42:6347–6353.
52. Buerge IJ, Poiger T, Muller MD, Buser HR. 2006. Com-
bined sewer overflows to surface waters detected by the
anthropogenic marker caffeine. Environ Sci Technol 40:
4096–4102.
53. McDonald J, Hartel P, Gentit L, Belcher C, Gates K,
Rodgers K, Fisher J, Smith K, Payne K. 2006. Identifying
sources of fecal contamination inexpensively with targeted
sampling and bacterial source tracking. J Environ Qual 35:
889–897.
54. Stoeckel D, Stelzer E, Stogner R, Mau D. 2011. Semi-
quantitative evaluation of fecal contamination potential
by human and ruminant sources using multiple lines of evi-
dence. Wat Res 45:3225–3244.
55. Vogel JR, Stoeckel DM, Lamendella R, Zelt RB, Santo
Domingo JW, Walker SR, Oerther DB. 2007. Identifying
fecal sources in a selected catchment reach using multiple
source-tracking tools. J Environ Qual 36:718–29.
56. Figueras MJ, Boreego JJ, Pike EB, Robertson W,
Ashbolt B. 2000. Sanitary inspection and microbiological
water quality, pp. 113–167. In Bartram J, Rees G (eds).
Monitoring Bathing Waters, a Practical Guide to the Design
and Implementation of Assessments and Monitoring Pro-
grammes. E & FN Spon, London.
57. Howard G, Pedley S, Tibatemwa S. 2006. Quantitative
microbial risk assessment to estimate health risks attribut-
able to water supply: can the technique be applied in
developing countries with limited data? J Wat Health 4:
49–65.
58. Kinzelman J. 2007. Case study: spatial distribution studies—
setting the groundwork for microbial source tracking (a case
study from Racine, WI). EPINET: Workshop on Microbial
Source Tracking. Purdue University, Chicago, IL (USA)
http://www.reeis.usda.gov/web/crisprojectpages/0201683-the-
environmental-pathogens-information-network-epinet.html.
59. Mudd D, Anan’eva T, Kinzelman J. 2012. Examination of
diurnal variation at a non-sewage impacted beach via qPCR
and culture based methods. J Environ Prot 3:1310–1317. doi:
10.4236/jep.2012.310149.
60. Whitman R, Nevers MB. 2004. Escherichia coli sampling
reliability at a frequently closed Chicago beach: monitoring
and management implications. Environ Sci Tech 38:
4241–4246.
61. Reeves R, Grant S, Mrse R, Oancea C, Sanders B, Boehm
A. 2004. Scaling and management of fecal indicator bacteria
in runoff from a coastal urban watershed in southern Califor-
nia. Environ Sci Technol 38:2637–2648.
62. Pitt R, Lalor M, Field R, Adrian D, Barbe’ D. 1993.
A user’s guide for the assessment of non-stormwater dis-
charges into separate storm drainage systems. #EPA/600/
R-92/238 and #PB93-131472. Center of Environmental
Research Information, Cincinnati, OH, and Urban Waste
Management & Research Center, New Orleans, LA.
63. U.S. Environmental Protection Agency. 1993. Investiga-
tion of inappropriate entries into storm drainage systems: a
user’s guide. Risk Reduction Engineering Laboratory, Con-
tract Number 68-C9-0033. U.S. EPA Office of Research
and Development, Cincinnati, OH.
64. Lalor M. 1994. Assessment of non-stormwater discharges
to storm drainage systems in residential and commercial
land use areas. PhD dissertation. Vanderbilt University, Nash-
ville, TN.
65. Wang JD, Solo-Gabriele HM, Abdelzaher AM, Fleming
LE. 2010. Estimation of enterococci input from bathers
3.4.5-10 ▪ MICROBIAL SOURCE TRACKING
and animals on a recreational beach using camera images.
Mar Pollut Bull 60:1270–1278.
66. Kinzelman JL, Field KG, Green HC, Harwood VJ,
McPhail CD. 2012. Indicators, sanitary surveys, and source
attribution techniques, pp 319–360. In Dufour A, Batram
J, Bos R, Gannon V (eds), Animal Waste, Water Quality,
and Human Health. U.S. EPA, Washington, DC; WHO,
Geneva; IWA Publishing, London.
67. Michigan Department of Environmental Quality (MDEQ).
2008. Total maximum daily load for E. coli for the Ecorse
River Watershed.Wayne County, MI. Michigan Department
of Environmental Quality. http://www.sourcemolecular.
com/pdfs/links/wb-swas-tmdl-ecorseckecoli_268861_7.pdf.
68. U.S. Environmental Protection Agency. 2011b. Using
microbial source tracking to support TMDL development and
implementation. U.S. Environmental Protection Agency,
Region 10, Watersheds Unit, Seattle, WA.
69. Kinzelman J, McLellan S. 2009. Success of science-based
best management practices in reducing swimming bans—a
case study from Racine, Wisconsin, USA. Aquat Ecosys
Health Manage 12:187–196.
70. Quakenbush J. 2006. Computational approaches to the
analysis of DNA microarray data. IMIA Yearbook Med Infor-
matics 2006:91–103.
71. Rusk N. 2009. Focus on next-generation sequencing data
analysis. Nature Meth 6:S1.
72. Shibata T, Solo-Gabriele HM, Sinigalliano CD, Gidley
ML, Plano LRW, Fleisher JM, Wang JD, Elmir SM, Guoq-
ing H, Wright ME, Abdelzaher AM, Ortega C, Wanless D,
Garza AC, Kish J, Scott T, Hollenbeck J, Backer L, Flem-
ing LE. 2010. Evaluation of conventional and alternative
monitoring methods for a recreational marine beach with
non-point source of fecal contamination. Environ Sci Technol
44:8175–8181.
73. Belanche L, Blanch A. 2011. Statistical approaches for
modeling in microbial source tracking, p 2070–0227. In
Hagedorn C, Blanch A, Harwood VJ (eds), Microbial Source
Tracking: Methods, Applications, and Case Studies, Springer,
New York.
74. Balls M, Amcoff P, Bremer S, Casati S, Coecke S, Clothier
R, Combes R, Corvi R, Curren R, Eskes C, Fentem J, Gri-
baldo L, Halder M, Hartung T, Hoffmann S, Schechtman
L, Scott L, Spielmann H, Stokes W, Tice R, Wagner D,
Zuang V. 2006. The principles of weight of evidence valida-
tion of test methods and testing strategies. Altern Lab Anim
34:603–620.
75. Linkov I, Loney D, Cormier S, Satterstrom FK, Bridges T.
2009. Weight of evidence evaluation in environmental
assessment: review of qualitative and quantitative
approaches. Sci Tot Environ 407:5199–5205.
76. Ashbolt NJ, Schoen ME, Soller JA, Roser DJ. 2010. Pre-
dicting pathogen risks to aid beach management: the real
value of quantitative microbial risk assessment (QMRA).
Wat Res 44:4692–703.
77. Roberts GS. 2007. Using the Annapolis Protocol for risk-
based assessment of recreational water—case studies from
New Zealand and San Diego. WEFTEC 2007, Sessions
61–70, 4870–4879 (10).
78. Converse R, Kinzelman J, Sams E, Hudgens E, Dufour A,
Ryu H, Santo-Domingo J, Kelty C, Shanks O, Siefring S,
Haugland R, Wade T. 2012. Dramatic improvements in
beach water quality following gull removal. Environ Sci Tech-
nol 46:10206–10213.
79. Abbott K. 2008. Guiding remediation of the Root River,
Racine, WI, USA, through correlation of stream bank con-
dition, river morphology and infrastructure surveys to chem-
ical and microbial source tracking. PhD dissertation,
University of Surrey, Guildford, UK.
80. World Health Organization. 1999. Health based monitoring
of recreational waters: the feasibility of a new approach (the
Annapolis Protocol). WHO/SDE/WSH/99.1. World Health
Organization, Geneva.
81. Harwood VJ, Staley C, Badgley B, Borges K, Korajkic A.
2013. Microbial source tracking markers for detection of
fecal contamination in environmental waters: relationships
between pathogens and human health outcomes. FEMS
Microbiol Rev 28:1–40.
82. Heaney JP, Pitt R, Field R. 2000. Innovative urban wet-
weather flow management systems. U.S. EPA EPA/600/
R-99/029. CRC Press, FL.
83. Pennsylvania Department of Environmental Protection.
2006. Pennsylvania Stormwater Best Management Practices
Manual. Document 363-0300-002. Department of Environ-
mental Protection, Bureau of Watershed Management, State
of Pennsylvania.
84. Kurzbaum E, Kirzhner F, Armon R. 2012. Improvement of
water quality using constructed wetland systems. Rev Environ
Health 27:59–64.
85. Koski AJ, Kinzelman JL. 2010. A review of best manage-
ment practices benefitting Great Lakes coastal waters: cur-
rent success stories and future innovations. Beach Water
Quality Work Group. International Joint Commission,
ftp://ftp.chesapeakebay.net/Monitoring/Foreman/Portland
%20work/Lessons%20Learned%20BMPs/Archives%20(Old
%20Organization)/Ashlee%20Literature/Kinzelman_2.pdf
86. Whiting-Grant K, Dalton C, Dillon F. 2003. Microbial
source tracking in two southern Maine watersheds – Webhannet
River watershed report. Maine Sea Grant Report #MSG-
TR-03-01. Maine Sea Grant, Wells, ME.
87. U.S. Environmental Protection Agency. 2002. National
Beach Guidance and Required Performance Criteria, Chap-
ter 3: risk-based beach evaluation and classification process.
EPA-823-B-02-004. U.S. EPA Office of Water, Washing-
ton, DC.
88. Kinzelman J, Hiller J. 2007. Incorporating education and
outreach in the re-engineering of a stormwater outfall
impacting recreational water quality at two public bathing
beaches on Lake Michigan. Current J Mar Educ 23:13–18.
89. Harwood VJ, Brownell M, Wang S, Lepo J, Ellender RD,
Ajidahun K, Hellein N, Kennedy E, Ye X, Flood C. 2009.
Validation and field testing of library-independent microbial
source tracking methods in the Gulf of Mexico. Water Res
43:4812–4819.
90. Gordon KV, Brownell M, Wang S, Lepo JE, Mott JA,
Nathaniel RR, Kilgen M, Hellein KN, Kennedy E, Har-
wood VJ. 2013. Relationship of human-associated microbial
source tracking markers with enterococci in Gulf of Mexico
waters. Water Res 47:996–1004.
91. Cao Y, Griffith JF, Dorevitch S, Weisberg SB. 2012. Effec-
tiveness of qPCR permutations, internal controls and dilu-
tion as means for minimizing the impact of inhibition
while measuring Enterococcus in environmental waters.
J Appl Microbiol 113:66–75.
92. Haugland RA, Siefring S, Lavender J, Varma M. 2012.
Influences of sample interference and interference controls
on quantification of enterococci fecal indicator bacteria in
surface water by the qPCR method. Water Res 46:
5989–6001.
93. Kinzelman JL, McPhail CD. 2012. Exposure interventions,
p 283–318. In Dufour A, Batram J, Bos R, Gannon V (eds),
Animal Waste, Water Quality, and Human Health. U.S. EPA,
Washington, DC; WHO, Geneva; IWA Publishing,
London.
94. Roseen R, Ballestrero T, Houle J. 2007. University of New
Hampshire Stormwater Center—2007 Annual Report. Univer-
sity of New Hampshire Stormwater Center.
 3.4.5. Microbial Source Tracking: Field Study Planning and Implementation ▪ 3.4.5-11

95. Glick P, Hoffman J, Koslow M, Kane A, Inkley D. 2011.
Restoring the Great Lakes’ coastal future—technical guidance
for the design and implementation of climate-smart restoration
projects. National Wildlife Federation, Ann Arbor, MI.
96. Sidhu JPS, Ahmed W, Gernjak W, Aryal R, McCarthy
D, Palmer A, Kolotelo P, Toze S. 2013. Sewage
pollution in urban stormwater runoff as evident from
the widespread presence of multiple microbial and chem-
ical source tracking markers. Sci Tot Environ 463–464:
488–496.
97. NHMRC. 2008. The guidelines for managing risks in recrea-
tional water. NHMRC publications, Canberra, Australia.
 Fecal Indicator Organism Modeling and Microbial
Source Tracking in Environmental Waters
MEREDITH B. NEVERS, MURULEEDHARA N. BYAPPANAHALLI,
MANTHA S. PHANIKUMAR, AND RICHARD L. WHITMAN
3.4.6
The use of models to describe environmental processes has
been widely examined in a number of fields, and recently,
extensive work has focused on models to explain the move-
ment, survival, and persistence of fecal indicator organisms
(FIO) and enteric pathogens in surface waters (1–4). As FIO
and associated pathogens are released into surface or ground-
water, they are subject to widely varying environmental con-
ditions and processes that influence how long they survive
and their transport pathway, all of which directly impacts
risk to human health. Among the variables influencing
FIO are microbial transport, dilution, settling and resuspen-
sion, biological inactivation, and predation. Models can
be useful for explaining the relationship between each of
these variables and concentrations of both FIO and patho-
gens (e.g., viruses, enteric pathogenic bacteria, protozoan
parasites) and their risk to human health.
The primary routes of human exposure to waterborne ill-
nesses ( primarily gastroenteritis) are through drinking water,
recreational water, and shellfish consumption. To minimize
exposure, surface water is monitored for pathogens themselves
or their surrogates—typically FIO such as fecal coliforms,
E. coli, or enterococci—to detect instances of excessive con-
tamination, which may trigger a swimming advisory, closed
shellfish beds, or a drinking water advisory and more intensive
drinking water treatment. An understanding of where the
contamination originated is of particular importance because
FIO can be present even in the absence of recent human
fecal contamination. For example, runoff from agricultural
fields, contaminated beach sands from wildlife (bird) feces,
or other transient sources may elevate FIO concentrations.
Also important is the ability to predict under what condi-
tions similar events might occur so that appropriate actions
can be taken to limit human contact. Models that describe
the processes involved in contaminant elevation can be
used as part of microbial source tracking to locate and elimi-
nate intrusive and harmful sources of fecal contamination
and for predicting when concentrations of FIO/pathogens
might be expected to be high, based on historic monitoring
observations.
Mathematical models are used to attempt to explain or
represent natural phenomena under various conditions.
They can be used to predict future events, explain past situa-
tions or test hypotheses. Models for surface water can be
roughly divided into two types: mechanistic, or process-
based, models and empirical/statistical, or data-based, models
doi:10.1128/9781555818821.ch3.4.6
3.4.6-1
(Table 1). In surface water applications, mechanistic models
integrate the processes involved that give rise to the presence
of pathogens or FIO in a given parcel of water at a certain
point in time. Factors including current movement, flow
rates, and settling of particles might be incorporated into a
nearshore mechanistic model. Biological inactivation of the
target FIO may also be included. The resulting model is typ-
ically a three-dimensional visualization of water movement
and fluctuation of associated FIO. Because process-based
models rely heavily on physical conditions, biological proc-
esses are often difficult to incorporate, increasing uncertainty
in model results.
Data-based models rely on empirical observations of con-
ditions in the parcel of water, observations of hydromete-
orological conditions, for example, and how they relate to
the presence of pathogens and/or FIO concentration. These
observations are used to relate pathogen/FIO concentrations
and water conditions with high-frequency hydrometeoro-
logical data or grab samples collected over an extended period
of time to predict what FIO concentration might be expected
at any given time. Both model types can be used for tracking
FIO contamination to its source or forecasting future con-
tamination, based on understanding of biological, chemical,
and physical processes and historical data.
There are many applications for models in environmental
settings, including river plume direction, watershed drainage,
groundwater movement, and nearshore beach dynamics.
For most of these, select FIO (e.g., fecal coliforms, E. coli,
or enterococci) have been used as the dependent variable,
that is, the target being modeled. A great deal of work has
been devoted to surface water models with applications in
microbial contamination and FIO ecology, especially sur-
vival, inactivation, and growth and die-off (5, 12). These
have been used for determining potential contaminant sour-
ces by predicting beach closures (5), estimating the influence
of river plumes on drinking water intakes (32), and closing
shellfish areas (33, 34). The models range from simple empir-
ical models to complex 3D models and the integration of the
two types. These models are finding increasing application in
management and environmental restoration efforts (35, 36).
In recent years, efforts to track microbial contamination
have become increasingly complicated with the finding
that FIO populations can survive for extended periods in nat-
ural environments (37, 38). Microbial source tracking efforts
have also needed refinement to distinguish between human
3.4.6-2 ▪ MICROBIAL SOURCE TRACKING

TABLE 1 Examples of two types of models for FIO and application potential for microbial source tracking
Accuracy at tracking
Ease of
microbial sources and
development/use
Model type Description predicting future References Remark
(1–3, easy to
conditions(1–3, low
difficult)
to high)
Mechanistic/ 1-dimensional 3 (1–4)
process-based advection-dispersion-reaction
models
2-dimensional verticallyintegrated 3 (5–10) (8) considered
models sediment-bacteria
interactions withina
fully 3D model
Fully 3-dimensional models 3 (11–13)
Review of mechanistic models (14, 15)
Statistical/ Simple correlation models 1 1
data-based
Statistical: multiple regression 2 2 (16–23)
Advanced statistical models: 3 2 (24, 25)
artificial neural network(ANN)
models
Advanced statistical models:path 3 3
analysis
Regional models 3 1 (26, 27)
Linking statistical models with 3 3 (28–31)
additional microbial markers

and animal fecal sources, each of which poses a different risk
to human health. Any expectation that high concentrations
of FIO unequivocally signal recent human sewage pollution
has been questioned (39, 40). While much of the effort has
been devoted to microbiological exploration through molec-
ular techniques (PCR, quantitative PCR, metagenomics,
and community analysis) (41, 42), process-based modeling
presents a realistic option for advancing source tracking suc-
cess by exploring the physical and biological processes affect-
ing bacterial distribution, survival, and environmental
growth, as well as transportation mechanisms with the goal
of tracing the contamination back to its origin.
In this chapter, we explore the use of models ( process-
based and statistical) for characterizing presence, fate, and
transport of FIO in nearshore surface waters. We describe
the potential sources of FIO and pathways by which these
organisms contaminate surface waters, as well as the impor-
tance of distinguishing source(s) to accurately gauge risk to
human health. Next, we describe mechanistic or process-
based models and how they can be used to track FIO contam-
ination to its source, as well as predict future events of ele-
vated FIO in surface waters. Empirical predictive models are
described that use our understanding of microbial movement,
survival-persistence, and source to predict water quality.
Finally, direct applications of these models are described
and for determining contamination events and initiating
remediation activities. While the extent of modeling applica-
tions far exceeds the scope of this chapter, numerous other
works have explored the use of models for subsurface drainage,
groundwater contamination, and virus transport (43, 44).
The use of models allows insight into the ability of pathogens
and pathogen indicators to survive in natural environments,
and the numerous physical and biological processes that con-
tribute to the associated calculation of human health risk.
SOURCES OF FIO IN SURFACE WATERS
FIO such as fecal coliforms, E. coli, and enterococci/fecal
streptococci have historically been used as indicators of water
quality worldwide. The widely held belief that these bacteria
(e.g., coliforms) are primarily found in the gastrointestinal
tract of humans and warm-blooded animals has been chal-
lenged, particularly since the late 1970s (45). The scientific
community and regulators in water quality have now gener-
ally recognized that FIO are not limited to human, animal,
and wildlife fecal sources; instead, populations of these bacte-
ria are quite common in the environment, with little or no
fecal inputs from known sources (see reviews by [40, 46, 47]).
Broadly, FIO persisting in ambient environments have
been referred to as potential “environmental” sources. Envi-
ronmental FIO populations (which do not include wildlife)
can be divided by habitat: terrestrial or aquatic. Habitats in
which environmental FIO have been identified include
upland soil and sediments, beach sand, terrestrial and aquatic
vegetation, and water bodies (Table 2). It is significant to
note that in many instances, FIO vis-à-vis sources/substrates
are not entirely independent of each other since interactions
are common and dynamic. Some examples of such interac-
tions include upland soil with streambed, streambed with sur-
face water, and streams/creeks with receiving water body (37,
54). The complexity of these interactions complicate source
tracking efforts and necessitate an understanding of survival
and flux between habitats.
Even though FIO are ubiquitous in the environment, their
distribution and relative numbers are highly variable in
the substrates examined. Bacterial counts (most-probable
number [MPN]/CFU/g) ranging from <1 to in excess of
1.0 × 103 have been observed in subtropical (46, 48) and
temperate (37, 46, 53) soils. Patchy albeit persistent FIO
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-3

TABLE 2 FIO in ambient environments: major habitats, persistence, survival and growth characteristics, and mechanisms of transport to
surface waters

Mechanisms of physical
Habitat Specific niche/biomes Major findings (references)
transport (References)
Uplandsoils Tropical, subtropical,and • Recovered in a wide range of soils including relatively Erosion and resuspension
andsediments temperateclimates pristine areas (37, 39, 48–52) (68, 69, 70)
• Patchy distribution, but can persist at length insoil Surface runoff (71, 72, 73)
environments (37, 39, 53, 52) Stream flow (74, 68, 75)
• Simulated field and in situ studies show thatFIO
populations may grow in the under certain conditions (38,
53–56)
• Metabolically diverse and genotypically distinct
populations have been recovered from a range of soils; the
original source of soil FIO remains speculative (57, 58)
• Populations survive longer in moist sediments than in
water (59–62)
• Likely grow/replicate under certain conditions (60, 63, 64)
• Serve as a source and/or sink for these bacteria, potentially
elevating FIO densities in the surficial water during
hydrometeorological events (64–67)
Sand Freshwater and • FIO and a range of human pathogens (bacteria, parasites, Wave resuspension (87, 70,
marinebeaches in and pathogenic fungi) are found in beach sand (76–80) 88, 89)
bothtropical • Persist longer in moist and deep subsurface sand (76) Currents (1, 4, 90)
andtemperate climates • Patchy distribution (76, 77, 81) Groundwater (91, 92, 93)
• Repeated inputs from wildlife (birds) and other sources
makes remediation tasks difficult (82–86)
Vegetation Aquatic, wetland, • FIO, particularly E. coli and Enterococcus spp., are found in Wave resuspension (103)
terrestrial, and algae, submerged vegetation, and forage crops (94–100);
ornamental, plants pathogenic bacteria have also been recovered from algae
(51, 101)
• Some plant associated populations have been recovered
from forage and other crops (98, 100, 102)
• Vegetation prolongs FIO in the environment (97, 102)
Water Fresh, estuarine, and • Includes multiple sources (46, 47, 82, 104, 105) Stream flow (110, 111)
marine waters • FIO growth has been observed in nutrient-rich waters Direct surface runoff(112–
(106, 107), algae/wrack (94–97, 108), and other 114)
vegetation (102) Advection (1, 4, 115)
• Residual FIO populations in sediments influence surface Groundwater (116, 117)
water quality due to mechanical disturbance or
hydrometeorological conditions (62, 65, 67, 109)
Source: Adapted from (46).

populations have similarly been observed in other environ-
mental substrates, such as beach sand (76, 77, 81) and aquatic
and terrestrial vegetation: algae/beach wrack (94–97, 118),
forage crops (98, 100), and wetland and ornamental plants
(119, 120). The widespread occurrence and persistence of
FIO in a broad range of substrates across spatial-temporal
and ecological gradients supports the argument of in situ
growth under certain conditions, as has been shown in soil
(38, 53–55), beach sand (121, 122), and algae/beach wrack
(94–97, 108).
Historically, classical culture-based methods, such as
membrane filtration and multiple-tube fermentation/MPN,
have routinely been used to enumerate FIO in environmen-
tal waters (123). Recent developments using defined sub-
strate technology technique (used to enumerate specific
bacteria from a mixed population using a chromogenic sub-
stance), though still based on the MPN format, have been a
significant progress, largely eliminating the required multistep
confirmatory tests associated with traditional MPN methods.
Commercial defined substrate technology–based methods are
currently available for many FIO, such as total coliforms,
E. coli, and enterococci, and these methods have been
approved as suitable alternates for enumerating FIO in drink-
ing water and environmental waters (123, 124).
Several studies have shown a positive relationship
between FIO densities in sewage-impacted waters (e.g., bath-
ing beaches) and incidence of gastrointestinal illness in
swimmers (see review [125]). However, associated health risks
from FIO derived from environmental or of unknown sources
remains largely unknown. Definitive conclusions on this
relationship are difficult at this time because of limited studies
with results ranging from limited or no health risk (126) to
measurable health risk (127). There is conflicting success
with using FIO as a surrogate for pathogens (128–130).
Both culturing techniques and many of the cutting-edge
technologies used in microbial source tracking have time
and again shown that traditional FIO alone, such as E. coli
and enterococci, are simply inadequate for identifying or dis-
criminating contaminant sources (see 41, 131). Modeling
may therefore be needed to identify sources of contamination
3.4.6-4 ▪ MICROBIAL SOURCE TRACKING
to aid in more accurate identification of health risk in surface
waters. Empirical predictive models have been used in recent
years as an alternate approach in beach monitoring to address
some of the limitations of traditional monitoring tools men-
tioned previously.
MECHANISTIC/PROCESS MODELING OF
MICROBIAL MOVEMENT AND SURVIVAL AND
APPLICATION TO SOURCE IDENTIFICATION
Significant progress has been made in modeling the move-
ment and survival of FIO and pathogens in watersheds and
at recreational beaches impacted by both point and nonpoint
sources. Since transport pathways and processes are different
in watersheds and nearshore areas of lakes and oceans, differ-
ent model approaches are needed. A major focus of previous
watershed modeling efforts (68, 74, 132–144) was on charac-
terizing sources and key processes influencing FIO and patho-
gen transport and to estimate bacterial loads in support of
total maximum daily loads (133, 136, 137, 140, 143–145).
Detailed reviews on modeling the fate and transport of bacte-
rial indicators and manure-borne pathogens focusing on the
pedon (soil particle), hill slope, and watershed scales are
available (135, 145). The application of process-based, dis-
tributed hydrologic models that are fully integrated across
all flow domains (such as surface water, vadose zone, and
groundwater domains) has become popular in recent years.
A number of such watershed models are now available,
including CATHY (146), InHM (147), OpenGeoSys (148),
ParFlow (149), PAWS (150), and Wash123 (151). Maxwell
et al. (152) provide a description of these watershed models
and a comparison of their results, while Niu (153) described
an application of one such model to bacterial (E. coli) fate and
transport modeling. Despite the recent interest in distributed
hydrologic modeling, application of these models to FIO fate
and transport is still currently limited. The semi-empirical
watershed model SWAT (Soil and Water Assessment Tool)
which can account for land use/land cover and other key
watershed attributes, is a popular model for describing the
fate and transport of FIO in watersheds (134, 135, 145,
154, 155). However, the fundamental unit of analysis in
SWAT is the hydrologic response unit and not a grid cell,
which can be easily reduced in size in areas where more detail
is needed. This, together with the semi-empirical nature of
the model (that is, lack of explicit representation of flow paths
in the vadose zone, groundwater, and surface water domains),
is known to introduce limitations when detailed flow and
transport representations are needed.
Due to the importance of the processes in describing
watershed-scale transport, the release of FIO from surface-
applied manure to runoff as well as from bottom sediments
in streams to the water column above formed the topic of sev-
eral earlier modeling efforts (132, 156–163). The application
of process-based models in describing FIO fate and transport
is generally limited to the overland, stream channel, and soil
compartments (74, 132, 159, 164). Due to the complexity of
processes involved, the release of FIO from surfaces is gener-
ally not described using process-based approaches and follows
either semi-empirical or statistical approaches (74, 135, 145).
Once released from sources, FIO can be routed using fully
mechanistic or semi-physical models (74, 132, 134, 154,
165, 166). The new generation of process-based watershed
models described already have this capability; however,
when such models failed to describe the observed FIO levels
at some sites, the reasons were attributed to either insufficient
information on FIO sources in certain areas (153) or lack of
additional detail in the modeling (e.g., a network of pipes
drained the landscape and this detail was absent from the
modeling, which included only overland flow).
Process-based models describing nearshore hydrodynam-
ics and FIO fate and transport have the potential to offer
significant insights into key processes controlling microbial
pollution at beach sites (Fig. 1). The insights gained from
mechanistic models can be used to further refine statistical
or empirical models. While a process-based model pro-
vides a complex understanding of the system, the statistical
model is a useful application of these findings, using a limited
amount of data input. For example, a process-based model
may provide a three-dimensional visualization of a plume to
identify microbial transport, whereas a statistical model
instead uses turbidity measurement in the plume to estimate
bacteria concentration in a particular water cell. A number
of physical, chemical, and biological processes influence the
transport, inactivation (loss per unit time) and survival of
FIO in the nearshore environment, including current speed
and direction, particle attachment, and biological inacti-
vation. An description of these coupled processes with a
focus on their mathematical representation is available in
(14). A majority of the papers that reported the application
of process-based models in the past examined the impacts
of point source discharges in a nearshore environment.
The fluid mechanics of river plumes discharging into lakes
or marine environments has been the focus of several earlier
papers, and a good review of these models is available (167).
Almost all process-based descriptions of FIO fate
and transport regardless of model dimension (1D, 2D, or
3D) and scale are based on some form of the advection-
disperion-reaction equation (ADRE). The “reaction” terms
in these equations usually describe loss processes associated
with FIO such as settling of particle-associated FIO,
solar or light-based inactivation, and base mortality. Cur-
rently used mathematical formulations based on the ADRE
often have competing parameters—for example, mixing
parameters such as dispersion or turbulent diffusion coeffi-
cients and inactivation rates both affect FIO peaks. In many
coastal/nearshore areas for which FIO models are being
developed, it is not uncommon to find that there were no pre-
vous efforts to model hydrodynamics and solute transport,
and this introduces significant uncertainty in the selection
of mixing parameters in the ADRE; therefore, describing
the dispersion of a conservative tracer within the nearshore
environment is expected to constrain better the FIO fate
and transport models. However, tracer dispersion in the
nearshore region is a complex problem that continues to
receive attention due to the wide range of processes and
spatial and temporal scales involved (e.g., the near-field
dominated by the momentum and buoyancy of the river
plume, the far-field dominated by turbulent diffusion in the
ambient flow and the intermediate region domianted by
buoyant spreading). In addition to these disparate length
and time scales involved, several physical processes are impor-
tant to accurately describe transport. These include (a) flow
reversals induced by changes in wind speed, fetch, and direc-
tion; (b) flow anisotropy, tubulence levels, and the high-
frequency components of currents; (c) waves, wave–current
interactions, and bottom shear stress (which controls sedi-
ment resuspension); the ability to resolve vertical and lateral
shear is important for transport; and (d) temperature and
thermal dynamics in the nearshore region which influence
FIO levels both directly and indirectly (via altered flow
dynamics). As well as these factors, tracer studies are expen-
sive, time consuming, and require significant planning and
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-5

FIGURE 1 Output of a process-based model showing of E. coli distribution in water at a Chicago beach. Diagram indicates concentrations 8
hr after a resuspension event within the embayment. (a) Suspended culturable E. coli concentration (b) settled culturable E. coli concentration.
Reprinted with permission from (171). doi:10.1128/9781555818821.ch3.4.6.f1

coordination. It is also uncommon to have tracer data avail-
able over long periods of time (e.g., over a three-month
summer season when FIO monitoring takes place at beach
sites), therefore, the search for inexpensive and easily meas-
ured parameters that serve as “tracers” continues. Recently
Wendzel (168) successfully used electrical conductivity
data measured at the Ogden Dunes beaches in Indiana to con-
strain three-dimensional transport models over the summer
season in 2008. However, care should be exercised as this
approach may not work at all sites and limitations and other
successful applications are described in Schimmelpfennig
et al. (169).
Current process-based models used to describe FIO trans-
port range in complexity from simple mass balance models to
fully three-dimensional models with sediment resuspension
(170). Experience with several hydrodynamic and transport
models indicates that models that successfully describe lake-
wide circulation and transport may not describe nearshore
hydrodynamics and transport adequately, especially from
the point of FIO transport due to the need to resolve small-
scale processes close to the shoreline. A detailed and careful
evaluation of existing process-based models using high-
resolution FIO data sets is expected to help improve the app-
plication of process-based models in microbial source tracking
and identify future research needs.
Increasingly, process-based models are being developed
for microbial source tracking that incorporate biological
processes and FIO (32, 171, 172). In a study attempting to
identify the source of contamination at a drinking water
intake, Sokolva et al. (173) measured concentrations of
human pathogens, including Cryptosporidium, norovirus,
and E. coli O157:H7, and used dispersion scenarios to indi-
cate the amount of contamination contributed by known
sources as well as the predicted persistence of the pathogen
threat to the drinking water. This allowed them to dis-
tinguish the main source of contamination as an emergency
sewage overflow, with other contributions from stormwater
and onsite sewer systems (173). This type of input character-
ization permits managers to direct remediation efforts more
effectively.
Similarly, Russell et al. (172) used spatial sampling for
FIO alongside process-based models to distinguish shoreline
sources of contamination at a California beach. Here, they
determined that the source of FIO in nearshore beach water
was the shoreline itself rather than an offshore input. Using
mass-balance modeling, they were able to further distinguish
source as beach sand for E. coli and groundwater for entero-
cocci. Integrating these findings with an understanding of
shoreline interactions can be used to mitigate contamination
to surface water.
Process-based models are expected to perform better
(compared with empirical models) when applied to new sites
due to their mechanistic framework, but there are still limita-
tions. The first major limitation is related to the amount of
time and effort needed to develop, test, and apply process-
based models. If the only aim of the modeling is to generate
timely predictions, then the empirical models seem to have
an advantage here, especially if we consider the fact that
performance metrics (such as R 2 or root mean square error
[RMSE]) are comparable for both types of models. The sec-
ond limitation is related to data requirements and computer
running times. Process-based models require significantly
more data by way of model inputs. Examples include bathy-
metry, nearshore features including coastlines, metereological
forcing data, initial and boundary conditions (e.g., loading
from tributaries) and information on the location and behav-
ior of sources. Computational demands of process-based mod-
els can be significant. It has become increasingly clear that
3.4.6-6 ▪ MICROBIAL SOURCE TRACKING
unstructured grid approaches, which represent nearshore fea-
tures such as breakwalls and complex coastline shapes more
accurately, are better suited for FIO modeling since the focus
is primarily on the nearshore. This is due to the fact that stair-
step representations of nearshore features introduce addi-
tional errors that negatively impact FIO modeling results.
The concept of scale is fundamental to hydrology and most
process-based models are also limited to the scales for which
they are developed. Finally, we offer a few comments on
the formulations used to describe FIO loss/inactivation in
process-based models. In addition to parameter competition
referred to already, some formulations suffer from parameter
nonuniqeness. For example, if the total FIO loss rate is div-
ided into a settling-related loss rate based on a constant
(that is, time- and space-invariant) fraction of attached bac-
teria, base mortality, and a light-based inactivation rate,
then as long as the net loss rate remains the same, any combi-
nation of these parameters may still produce acceptable
results. One way to better constrain these formulations is by
including more detailed/fundamental process descriptions
such as sediment–bacteria interactions (174) and depth-
dependent light/inactivation formulations. In the past, model
results from such formulations revealed that the fraction of
attached bacteria is highly variable in space and time, that
the background concentrations of FIO are better described,
and that light-dependent inactivation rates can vary sig-
nificantly within the water column (174); however, these
insights come at the expense of introducing more parameters
and additional complexity into the modeling. To summarize,
the major limitations are related to data requirements, model
development and running times, formulation/parameteriza-
tion of models and scale.
EMPIRICAL MODELS: PREDICTING
PATHOGEN/FIO CONCENTRATIONS TO
PROTECT PUBLIC HEALTH
Empirical, statistical (data-based) models can range from
fairly simple to highly complex, depending on the site loca-
tion and the desired accuracy and sensitivity. In surface water
applications, statistical models are used to relate hydrome-
teorological conditions with FIO concentrations to deter-
mine what physical conditions are associated with elevated
concentrations of FIO. Results can be used to predict similar
future events or to track elevated concentrations to their
potential source(s). Simple models might incorporate a single
predictive variable, such as rainfall or turbidity, at a single
location to predict when FIO concentrations will be elevated
(36). More complicated empirical models often include
multiple predictive variables, possibly a larger geographic
area, and higher level statistical analyses (26, 28, 175).
With the incorporation of larger amounts of data—temporal
and spatial—or more precise data, model predictive capacity
often improves because the model can account for more envi-
ronmental variation.
Empirical models have been widely tested for applications,
but model success is often determined by data quality or
intensity. Furthermore, some nearshore water locations sim-
ply do not lend themselves to modeling due to variation in
the target FIO, physical or geographic constraints, biological
systems, ecological interactions, or complex contaminant
inputs. Although some empirical models can be improved
with higher resolution or higher intensity data, some situa-
tions simply cannot be effectively modeled using statistical
approaches.
Model Types
Simple Models
A simple model for bacteria contamination directly correlates
FIO or pathogen concentrations with an associated environ-
mental fluctuation. Rainfall is the most common single pre-
dictor used: an increase in rainfall frequently results in an
increase in FIO in receiving waters as a result of surface runoff
(71), storm water inputs (139), septic system leaks (176), or
sewer overflows (177, 178). Often, a simple model will suffice
in situations where the release of waste water from sewage
treatment plants is expected (e.g., combined sewer overflows)
or in other areas where there are known fecal sources, such as
manure applications in agricultural areas or feedlot opera-
tions. For example, in the event of a significant rain event,
raw sewage or primary-treated sewage may be released into
receiving waters through a combined sewer overflow or storm
sewer overflow. This type of model is often used for direct
management applications (36), perhaps triggering a preemp-
tive closing of downstream beaches and shellfish areas or a
boil-water advisory.
Sources of FIO contamination have been identified with
these simple models, particularly in areas with heavy agri-
cultural use. In an examination of rural areas in Wales, it
was determined that widespread agricultural operations
were contributing significantly to coastal beach FIO concen-
trations during rain events (110) but that land use patterns
were impacting the net transport of these bacteria. Similarly,
Kelsey et al. (179) integrated land use and rainfall runoff
to determine sources of runoff that were identified as originat-
ing from septic tanks. In a related study, it was determined
that agricultural source became a more significant source
than sewage input after the installation of wastewater plant
improvements (16). By pairing an understanding of FIO
transport triggered by rain and high-flow events, as measured
using hydrographs and tracer studies, Wyer et al. (180) were
able to track the contamination to identify cattle as the source
of contamination. The correlation between increased FIO
and rainfall itself indicates the potential for overland flow
and agricultural contamination.
Rainfall models for FIO have been used directly in man-
agement in numerous locations. Characterizing the effect of
rainfall on FIO at several beaches in Scotland led to the devel-
opment of a predictive model that uses rainfall to trigger
swimming advisories (36). Here, rain gauges installed on
several rivers are used to predict when FIO concentrations
will be elevated enough to impact downstream coastal
beaches (36). Rainfall is the general model used to regulate
US shellfish areas as well; a measured amount of rainfall
will result in the closing of shellfish beds (181). Simple rain-
fall models have been validated in this field for accurately
predicting presence of norovirus in shellfish (182) and the
presence of diarrhoetic shellfish poisoning toxins attributed
to the microalgae Dinophysis (33).
Regression Models
Like simple models, regression models are developed based on
existing historical data, and relationships between the target
(in this example FIO) and hydrometeorological parameters
are established. However, with regression models, multiple
parameters (multivariate statistics) can be used in the con-
structed equations, leading to a more refined understanding
of the relationship between changing environmental condi-
tions and FIO. Regression models rely on advanced statistics
that determine how much variation in FIO concentra-
tions can be explained by the model and by each individual
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-7
predictor. There are a number of types of regression models
that can be used, but the most widely used is multiple linear
regression, which attempts to establish a linear relationship
between independent and dependent variables ( predictors).
These models should be used with caution because there
are sets of assumptions, including normal data distribution
and lack of colinearity that must be met to maintain the val-
idity of the procedures. Other regression models that have
been used in environmental applications include Bayesian
linear regression (182, 183), partial least squares (17, 175),
recursive partitioning (regression trees) (184), and logistic
regression (18, 24).
The statistical output of a regression model provides
an assessment of how much variation is explained by each
of the significant predictors. This value gives insight for
the importance of any individual predictor on potentially
influencing FIO concentrations and therefore can be used
to indicate source of contamination. For example, a high
regression coefficient associated with wave height might
indicate that the FIO sources are the beach sand/sediment
or shoreline algae, both of which can harbor high concen-
trations of FIO (77, 97). Alternately, if wind direction
and turbidity of a nearby river outfall have a high regression
coefficient in the model, it may indicate that the source of
FIO contamination is riverborne, leading to further ex-
amination of the river to locate dominant contamination
source(s).
Predictive models at coastal beaches have been heavily
investigated in recent years as an alternate strategy for
monitoring FIO contamination, as the practice has been
encouraged by the U.S. Environmental Protection Agency
as part of the requirements for recreational water quality
monitoring (185). Findings are lending new insight to iden-
tifying sources of FIO that lead to beach swimming adviso-
ries. Typical predictive parameters include rainfall, wave
height, wind direction, solar radiation, and water turbidity,
all of which are measureable surrogates of the physical
processes influencing the survival and distribution of FIO
in the environment (186). In the Great Lakes, efforts have
been undertaken in a number of locations, including north-
ern Indiana (19), northern Illinois (187), Chicago (188),
and Cleveland (35). River locations have also been examined
for regression-based predictive models in Massachusetts
(18, 90), Arkansas, and Oklahoma (189). Marine coastal
beaches have also been examined as potential locations for
predictive modeling, including coastal California (17), Loui-
siana (25), Miami (29), and the United Kingdom (16). As a
result of these models and an understanding of FIO source,
additional actions to reduce contamination load and restore
beaches have been undertaken (e.g., native grass planting,
bird deterrence).
Generally, beaches that have shown good potential for
predictive models are those directly influenced by a point
source because of the major impact the point source has on
beach water quality (36). Beaches with more diffuse sources
of contamination are more difficult to predict with higher lev-
els of uncertainty, and these are the beaches where microbial
source tracking becomes more important. Also, beaches
where FIO concentrations follow a normal distribution may
be more readily modeled because variation can be more
predictable.
Regression modeling for shellfish beds has also been
explored to refine the rainfall/FIO relationship generally
used for management. Among the environmental parameters
also linked to elevated FIO in shellfish areas are salinity
and water temperature (190), stream flow (191), and tidal
cycle (192). In a simultaneous assessment of shellfish and
recreational waters, Gonzalez et al. (193) confirmed these
parameters and included dissolved oxygen. While these rela-
tionships can attest to the abundance of FIO contamination,
further microbial analyses would be needed to define the
health risk.
The use of regression predictive modeling has also been
explored for use on multiple beaches. These regional models
associate the larger scale processes to explain simultaneous
fluctuations in FIO at multiple beach locations. This permits
a better understanding of the influence of local contami-
nation sources on individual locations and may allow the
more effective tracing of contamination by defining the over-
all background level of FIO in nearshore water. A model
developed for Chicago beaches was able to explain up to
40% of the variation at 23 beaches using only three predic-
tors: wave height, barometric pressure, and day of the year
(27). An application of this approach at a stretch of coastline
locally influenced by several point source outfalls exhibited
the interaction between background FIO affecting all of
the beaches simultaneously (e.g., sandborne FIO resus-
pended by waves) and local FIO from a source with a limited
area of influence (e.g., creek outfall) (26). An understanding
of the larger scale fluctuations allows a better parsing of
the background and local influences on water quality and
therefore the identification of potential contamination sour-
ces. Persistent, significant sources may be identified and
examined more closely for human health risk and potential
remediation.
Advanced Statistical Models
More advanced statistical approaches have recently been
explored. Among these, artificial neural networks (ANNs)
are used to describe nonlinear relationships among predictors
and the independent variable. Applications to surface
water quality have been focused on comparisons between
ANNs and regression models (24, 25). ANN models can be
programmed for a selected number of input parameters,
like regression models, but they identify intricate relation-
ships between predictors in a nonlinear fashion. Input param-
eters are similar to those used in linear regression models,
and ANNs have shown predictive success for FIO concen-
trations in Scotland using tidal range and discharge (6); in
California using rainfall, temperature, and turbidity (90);
and along the Gulf Coast using 15 input variables (25). Initial
comparisons of ANN models for recreational water quality
indicate no significant improvement in predictive capacity
over linear regression models, but models have different pre-
dictive capacity depending on the selection of measurement
of predictive success. Their usefulness in predicting and
forecasting norovirus contamination in oyster beds has been
indicated (182). Like regression models, the ANN method
identifies conditions associated with elevated levers of
norovirus; this information can similarly be used to indicate
potential source.
Path analysis is a form of structural equation modeling that
describes the directed dependencies of individual variables.
It can be used to characterize the relationship between a num-
ber of independent variables and the dependent variable,
while accounting for the error associated with each variable.
The analysis focuses on causality, linking dependent variables
directly to the independent variable of interest. This analyt-
ical approach was used to distinguish the pathway of contam-
ination between beach sand and water at a Chicago beach
(54), as will be discussed later in this chapter.
3.4.6-8 ▪ MICROBIAL SOURCE TRACKING
Data Requirements
Predictive capacity of any empirical model depends on the
robustness of the input data. A data set that covers a wide
range of environmental conditions will be able to predict var-
iation in FIO/pathogens across a broader spectrum of condi-
tions. A more intensive data set (larger sample size) will
also provide a better estimate of each individual input varia-
ble. Models with limited data sets provide spurious results sub-
ject to misinterpretation. Numerous local, state, and federal
agencies maintain monitoring stations that have data freely
available; these might include nearshore wave buoys, weather
stations, rain gauges, or air monitoring stations. Finally, the
quality of the input data will influence predictive outcomes.
Typically, the dependent variable (in this case FIO) is
limiting in environmental applications of predictive models
because high-frequency, continuous hydrometeorological
data sets are often available, but FIO monitoring data are typ-
ically collected far less frequently. FIO data that are more
extensive and intensive will result in a more accurate model.
To develop a multiple linear regression model, often,
numerous independent variables or predictors are considered.
These independent variables can be differently aggregated
prior to application, for example, 24-h rainfall or 48-h rain-
fall. For local models, more site-specific input data are sought,
which can involve the installation of monitoring stations or
regular collection of monitoring data by on-site personnel,
such as turbidity or water temperature. The accuracy and pre-
dictive capacity will depend on the frequency of collection
for the independent variable, which in the case of FIO may
be relatively infrequent, even if multiple years of monitoring
data are available. For example, weekly sample collection will
give little insight in a regression model developed for a beach
because FIO exhibits high spatial and temporal variation
(194, 195). Attempting to develop a linear regression based
on once-a-week sampling often leads to highly variable and
unreliable predictive capability (196).
Development of empirical models often requires the
consideration of multiple years of data (19, 25). An empirical
predictive model is usually developed with a subset of data
(training set) and validated with additional data collec-
tion (validation set) in out years. This allows the verification
and stabilization of the model as well as allowing adjustments
to be made if the training data set reflects unusual conditions.
Variation in hydrometeorological conditions over seasons or
years can be significant (e.g., drought, extreme temperatures).
The best predictive models will include this variation by
incorporating a broad range for both the dependent and inde-
pendent variables.
Still, uncertainty is an important component of any mod-
eling exercise. Statistical models typically fail to incorporate
effectively biological process and ecological interactions, so
empirical models rarely account for all sources of variation.
This level of uncertainty also varies across spatial and tempo-
ral scales.
Comparing Models
Comparisons among models can be difficult if many statistical
analyses are being compared. Often simple models are devel-
oped using only a correlation coefficient: a high FIO event is
associated with high rainfall. This type of direct relationship
is fairly easy to explain, with a comparison of the correla-
tion coefficients between two models. In a multiple linear
regression, however, the model attempts to fit a line to the
relationship between the FIO and the predictors that have a
significant linear relationship to FIO concentration. Results
are reported as an R 2 value, which is a measure of how
much variation in the dependent variable is explained by
the model equation. An R 2 of 1.0 indicates that all of the var-
iation is explained in the model, whereas an R 2 of 0 indicates
the relationships among the variables are random. To com-
pare model outcomes, the R 2 can be evaluated, but other stat-
istical considerations should be made, including residuals,
data overfitting, and collinearity in the independent varia-
bles. Linear regression models are also compared with an
examination of the RMSE in the model equation. This is a
measure of error in the model: a lower RMSE indicates lower
error, a more favorable result. ANNs also produce an RMSE,
making comparisons with multiple linear regressions possible.
In addition to comparing success by way of fitting the
observations to the model, model success is also determined
based on predicting whether a sample will exceed an estab-
lished threshold. For example, many beaches in the United
States are monitored and managed for FIO, and a water
sample that exceeds a single-sample threshold will result in
the order for a swimming advisory or beach closure (197).
As such, many models are assessed by how well they predict
whether FIO are above or below that threshold. Binary logis-
tic regression can be used directly to predict the likelihood
that a water sample will exceed that threshold, using input
variables in the same way they are used for linear regression.
In multiple linear regression, however, model success can be
determined by counting the number of type I and type II
errors. Type I errors occur when the model predicts an FIO
concentration above the advisory threshold when actual
concentration is below (a false positive), and type II errors
occur when the model predicts an FIO concentration below
the advisory threshold when actual concentration is above
(a false negative). Simply calculating the number of errors
for each model approach provides the basis for a comparison
of multiple models.
MODELING PATHOGEN/FIO FOR SURFACE
WATER MONITORING: APPLIED USE
The use of models to track a source of FIO and associated
pathogens has been examined (173, 198), but as yet has not
found widespread use. To develop an effective strategy for
tracking candidate sources of contamination, the problem
must first be identified (Fig. 2), because this will determine
the tools and strategies employed. Among the potential
objectives might be the elimination of outbreaks of illness,
chronic swimming advisories, episodic contamination events,
or regional or location-specific contamination events. The
objectives of the contaminant source tracking effort must
be clearly defined, even if solutions are not readily discernible
at an early stage. It is important to identify the agent of
concern, which may be merely the FIO or the associated
pathogens. The agent may also be the microbial source-
tracking surrogate, such as genetic markers of animal contam-
ination or affected organisms (e.g., contaminated shellfish,
infected birds, and human illnesses). Markers and pathogens
themselves can be modeled given that data and analytical
requirements are feasible. Regardless, it is important to under-
stand the temporal and spatial distribution of these surrogates
and/or effects. The spatial-temporal distribution will become
important in determining source and fate of the agent or
effect in question. Often, dynamic models can be used to
help reconstruct the distribution or probable distribution of
organisms during the event(s) in question.
Once the problem and affected area have been clearly char-
acterized, hypotheses regarding distribution and transport
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-9

FIGURE 2 Diagram outlining the process that might be followed for using models to identify sources of fecal contamination and initiate
remediation activities. doi:10.1128/9781555818821.ch3.4.6.f2

are made, and collection of data for retrospective (finding the
source of a recorded contamination event) or prospective
(describing the outcome of a potential contamination event)
modeling can proceed. Although the best models are gener-
ally the most parsimonious, using the fewest predictors, it is
best to collect all variables that may be candidate forcing fac-
tors, correlates, or explanatory factors to maximize the model’s
potential. For deterministic models, hydrometeorological
information or forcing factors that will eventually drive the
model will be examined. Additional information on the
dependent variable such as FIO concentration or illness rate
will eventually be needed for calibration and validation of
this model. Data collection needs for empirical modeling
are similar, but additional information that extends beyond
hydrometeorological conditions is needed, particularly for
the dependent variable. Again, these may be direct values
such as illness rates, pathogen presence, indicator concentra-
tions, or microbial source tracking markers. It is important
that all physical, chemical, and biological data be collected
consistently so that past and future years can be used for model
improvement or validation.
The decision on whether to use process-based or empirical
models relates directly to the objective of identifying and
understanding contaminant sources. Process-based models
directly address causation but may not be able to cope with
the complexity and indeterminate variation of the biological-
physical system. Empirical models tend to explain more effec-
tively the degree of variation but not necessarily the nature
or cause of the variation; empirical models also may not be
able to identify interactive components of the system or pre-
viously undiscovered potential influences. Testing under
controlled conditions and the statistical approach of path
analysis, a structural equation model that calculates the con-
tribution of predictors to the overall model, may then be
useful (see Fig. 3). Whichever model type is chosen, it is
important that models be validated and refined.

FIGURE 3 Example of a structural equation model ( path analysis) for determining the source and pathway of FIO contamination in
nearshore beach water. The exhibited results are from E. coli data collected from 63rd Street Beach in Chicago in 2000, from several water
depths, shoreline sand, and submerged sand. Numbers on each box refer to the cumulative R 2 value, while numbers associated with the arrows
are regression weights. Significance is indicated: *, 0.05, **, 0.01; ***, 0.001. Errors (E1–E4) are included in path analysis. Reprinted with
permission from (54). doi:10.1128/9781555818821.ch3.4.6.f3
3.4.6-10 ▪ MICROBIAL SOURCE TRACKING
Once the working hypothesis has been supported
or rejected by the model, an alternate hypothesis may be
developed or information may be consolidated to develop rec-
ommendations for remediation (Fig. 2). Too often, manage-
ment options are developed without a clear understanding
of the source, fate, and transport of the contamination. With-
out this information, remediation success can be defeated or
compromised. All management action should be evaluated
retrospectively to determine the efficacy of the remediation
and to provide information for continued improvement.
Case Study: 63rd Street Beach, Chicago
The city of Chicago was concerned about the large number of
beach closings over the previous few years due to elevated
concentrations of the FIO E. coli at 63rd Street Beach ( prob-
lem identification). Suspected sources of contamination
included a nearby drainage pond, infiltration of contamina-
ted groundwater, and bird feces. In a 6-month study, samples
were collected 3 consecutive days along five transects in
beach sand and water; bird and visitor counts were obtained;
and DNA fingerprinting of E. coli and Salmonella isolates
were collected from sand and water. Additionally, water qual-
ity instruments and a weather station measured hydrome-
teorological conditions continuously (199). Through DNA
fingerprinting and empirical predictive modeling, it was dem-
onstrated that most of the beach closings were caused by bird
fecal contamination and extant FIO in the sand rather than
input from groundwater or stream inflows (agent distribution)
(77) (Fig. 3). Even more interesting, FIO in the sand were not
solely derived from birds; FIO in the foreshore sand remained
at rough equilibrium, suggesting that input/growth and mor-
tality were equivalent.
The next step to determine source was to explore models
(collect existing and new data). The predictive model for
63rd Street Beach suggested that solar radiation, wave height,
and turbidity were correlative factors with FIO concentration
but did not demonstrate causation (identify probable corre-
lates; develop data-based models). To better understand
the causative relationship, path analysis was explored. While
linear regression showed a close relationship between FIO in
the foreshore sand and water quality, the direction of the
interaction was unclear. Path analysis suggested that there
was a tendency for FIO in water to move from deeper to shal-
lower depths and then to shore sand (Fig. 3). Mechanistic
( process-based) modeling was then explored, using an acous-
tic Doppler current profiler and a laser in situ scattering
and transmissometry instrument. Results suggested that
Stokes drift was allowing material to move directionally in
near-laminar flow (identify probable forcing factors; develop
mechanistic models); FIO suspended in water moved shore-
ward during generally calm evening conditions, and the
deposited material along the shoreline could then be resus-
pended as wave activity increased during the day (171).
Thus, there was a cycling between morning and afternoon
bacterial distribution in the water column driven by wave
conditions and counteracted by the effects of solar inacti-
vation of E. coli (87, 200) (compare/improve/consolidate).
Process-based modeling again suggested that breakwalls on
both sides of the beach captured along-shore contaminants,
and once deposited in the beach area, they became trapped;
when resuspended, beach swimming advisories resulted
(171). Remediation activities by the Chicago Park District
are currently being undertaken; these include reconfigura-
tion the breakwaters to allow more flow (escape) of these
contaminants and employment of dogs to assist in gull harass-
ment to reduce the amount of input from bird feces
(remediate source, transport, fate). Following remediation,
models to validate improvement may be tested (evaluate).
This example shows the strength of combining multiple
approaches for contaminant microbial source determination.
The empirical models, process-based models, and source
tracking together established a clear understanding of how
this particular beach functions, with FIO input from birds
and sand being entrapped in the nearshore water. This infor-
mation was used directly to guide management decisions
to decrease the number of beach closings or swimming advi-
sories. Also important, it signifies the lack of a significant
human contamination source, which has implications for
monitoring and the undertaking of extensive mitigation
activities.
SUMMARY AND CONCLUSIONS
Process-based and data-based models hold great potential for
identifying sources of fecal contamination. Most of the sur-
face water modeling for FIO has been devoted to forecasting
movement and survival of these contaminants, for either
identification of source type or predictions. The body of
knowledge that has been gained as a result of these studies,
however, can easily be adapted to tracking FIO to their orig-
inal sources. The vast majority of microbial source tracking
efforts have been molecular-based techniques that rely on
DNA markers, but methods have not been consistently reli-
able in field testing, as a result of variations in sensitivity
and specificity and analytical interference (201–203), as
well as the nature of the contaminants unrelated to known
sources. As techniques continue to be improved, the use of
models for detecting and tracking FIO contamination to
determine source is a viable addition to the suite of techni-
ques and tools being developed. Further, the integration
of these developing molecular techniques with physical mod-
els could greatly enhance source tracking success. Initial
attempts to do so are showing some promise (173, 198),
and the paired use of these techniques would better highlight
the physical factors influencing the transport and fate of tar-
get contaminants. A multifaceted approach to source track-
ing appears to provide the most complete explanation of
surface water contamination, which will help best protect
public health.
We acknowledge the hard work of our colleagues in the beach com-
munity (scientists, managers, regulators) for their efforts in managing
and preserving water quality in the Great Lakes. This article is Contribu-
tion 1907 of the USGS Great Lakes Science Center.
REFERENCES
1. Boehm AB. 2003. Model of microbial transport and inac-
tivation in the surf zone and application to field measure-
ments of total coliform in northern Orange County,
California. Environ Sci Technol 37:5511–5517.
2. Boehm AB, Keymer DP, Shellenbarger GG. 2005. An
analytical model of enterococci inactivation, grazing, and
transport in the surf zone of a marine beach. Water Res
39:3565–3578.
3. Canale RP, Auer MT, Owens EM, Heidtke TM, Effler
SW. 1993. Modeling fecal coliform bacteria—II. Model
development and application. Water Res 27:703–714.
4. Grant SB, Kim JH, Jones BH, Jenkins SA, Wasyl J,
Cudaback C. 2005. Surf zone entrainment, along-shore
transport, and human health implications of pollution
from tidal outlets. J Geophys Res C 110:1–20.
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-11
5. Liu L, Phanikumar MS, Molloy SL, Whitman RL,
Shively DA, Nevers MB, Schwab DJ, Rose JB. 2006.
Modeling the transport and inactivation of E. coli and
enterococci in the near-shore region of Lake Michigan.
Environ Sci Technol 40:5022–5028.
6. Kashefipour SM, Lin B, Falconer RA. 2005. Neural net-
works for predicting seawater bacterial levels. P I Civil Eng-
Wat M 158:111–118.
7. de Brauwere A, de Brye B, Servais P, Passerat J, Deleer-
snijder E. 2011. Modelling Escherichia coli concentrations
in the tidal Scheldt River and estuary. Water Res 45:
2724–2738.
8. Gao G, Falconer RA, Lin B. 2011. Numerical modelling of
sediment-bacteria interaction processes in surface waters.
Water Res 45:1951–1960.
9. Sanders BF, Arega F, Sutula M. 2005. Modeling the
dry-weather tidal cycling of fecal indicator bacteria in
surface waters of an intertidal wetland. Water Res 39:
3394–3408.
10. Zhu X, Wang JD, Solo-Gabriele HM, Fleming LE. 2011.
A water quality modeling study of non-point sources at rec-
reational marine beaches. Water Res 45:2985–2995.
11. Hellweger FL, Masopust P. 2008. Investigating the fate
and transport of Escherichia coli in the Charles River, Bos-
ton, using high-resolution observation and modeling. J
Amer Water Resour Ass 44:509–522.
12. Thupaki P, Phanikumar MS, Beletsky D, Schwab DJ,
Nevers MB, Whitman RL. 2010. Budget analysis of
Escherichia coli at a southern Lake Michigan beach. Environ
Sci Technol 44:1010–1016.
13. Connolly JP, Blumberg AF, Quadrin JD. 1999. Modeling
fate of pathogenic organisms in coastal waters of Oahu,
Hawaii. J Environ Engineer-ASCE 125:398–406.
14. Hipsey MR, Antenucci JP, Brookes JD. 2008. A generic,
process-based model of microbial pollution in aquatic sys-
tems. Water Resour Res 44:W07408.
15. Grant SB, Sanders BF. 2010. Beach boundary layer: a
framework for addressing recreational water quality impair-
ment at enclosed beaches. Environ Sci Technol 44:8804–
8813.
16. Crowther J, Kay D, Wyer MD. 2001. Relationships
between microbial water quality and environmental condi-
tions in coastal recreational waters: the Fylde coast, UK.
Water Res 35:4029–4038.
17. Hou D, Rabinovici SJM, Boehm AB. 2006. Enterococci
predictions from partial least squares regression models in
conjunction with a single-sample standard improve the
efficacy of beach management advisories. Environ Sci Tech-
nol 40:1737–1743.
18. Eleria A, Vogel RM. 2005. Predicting fecal coliform bacte-
ria levels in the Charles River, Massachusetts, USA. J Am
Water Resour Assn 41:1195–1209.
19. Nevers MB, Whitman RL. 2005. Nowcast modeling
of Escherichia coli concentrations at multiple urban
beaches of southern Lake Michigan. Water Res 39:5250–
5260.
20. Frick WA, Ge Z, Zepp RG. 2008. Nowcasting and fore-
casting concentrations of biological contaminants at
beaches: a feasibility and case study. Environ Sci Technol
42:4218–4824.
21. Gronewold AD, Qian SS, Wolpert RL, Reckhow KH.
2009. Calibrating and validating bacterial water quality
models: a Bayesian approach. Water Res 43:2688–2698.
22. Nevers MB, Whitman RL, Frick WA, Ge Z. 2007. Inter-
action and influence of two creeks on E. coli concentrations
of nearby beaches: exploration of predictability and mech-
anisms. J Environ Qual 36:1338–1345.
23. Olyphant GA, Whitman RL. 2004. Elements of a predic-
tive model for determining beach closures on a real time
basis: the case of 63rd Street Beach Chicago. Environ Monit
Assess 98:175–190.
24. Mas DML, Ahlfeld DP. 2007. Comparing artificial neural
networks and regression models for predicting faecal coli-
form concentrations. Hydrolog Sci J 52:713–731.
25. Zhang Z, Deng Z, Rusch KA. 2012. Development of pre-
dictive models for determining enterococci levels at Gulf
Coast beaches. Water Res 46:465–474.
26. Nevers MB, Whitman RL. 2008. Coastal strategies to pre-
dict Escherichia coli concentrations for beaches along a 35
km stretch of southern Lake Michigan. Environ Sci Technol
42:4454–4460.
27. Whitman RL, Nevers MB. 2008. Summer E. coli patterns
and responses along 23 Chicago beaches. Environ Sci Tech-
nol 42:9217–9224.
28. Wong M, Kumar L, Jenkins TM, Xagoraraki I, Phaniku-
mar MS, Rose JB. 2009. Evaluation of public health risks at
recreational beaches in Lake Michigan via detection of
enteric viruses and a human-specific bacteriological
marker. Water Res 43:1137–1149.
29. Shibata T, Solo-Gabriele HM, Sinigalliano CD, Gidley
ML, Plano LRW, Fleisher JM, Wang JD, Elmir SM, He
G, Wright ME, Abdelzaher AM, Ortega C, Wanless D,
Garza AC, Kish J, Scott T, Hollenbeck J, Backer LC,
Fleming LE. 2010. Evaluation of conventional and alter-
native monitoring methods for a recreational marine beach
with nonpoint source of fecal contamination. Environ Sci
Technol 44:8175–8181.
30. Nnane DE, Ebdon JE, Taylor HD. 2011. Integrated anal-
ysis of water quality parameters for cost-effective faecal pol-
lution management in river catchments. Water Res 45:
2235–2246.
31. Shibata T, Solo-Gabriele HM, Fleming LE, Elmir S. 2004.
Monitoring marine recreational water quality using multi-
ple microbial indicators in an urban tropical environment.
Water Res 38:3119–3131.
32. Sokolova E, Pettersson TJR, Bergstedt O, Hermansson
M. 2013. Hydrodynamic modelling of the microbial water
quality in a drinking water source as input for risk reduction
management. J Hydrol 497:15–23.
33. Vale P. 2012. Two simple models for accounting mussel
contamination with diarrhoetic shellfish poisoning toxins
at Aveiro lagoon: control by rainfall and atmospheric forc-
ing. Est Coast Shelf Sci 98:94–100.
34. Wilkinson RJ, McKergow LA, Davies-Colley RJ, Ballan-
tine DJ, Young RG. 2011. Modelling storm-event E. coli
pulses from the Motueka and Sherry Rivers in the South
Island, New Zealand. NZ J Mar Freshw Res 45:369–393.
35. Francy DS. 2009. Use of predictive models and rapid meth-
ods to nowcast bacteria levels at coastal beaches. Aquat Eco-
syst Health Manage 12:177–182.
36. Stidson RT, Gray CA, McPhail CD. 2012. Development
and use of modelling techniques for real-time bathing water
quality predictions. Water Environ J 26:7–18.
37. Byappanahalli M, Fowler M, Shively D, Whitman R.
2003. Ubiquity and persistence of Escherichia coli in a mid-
western stream. Appl Environ Microbiol 69:4549–4555.
38. Solo-Gabriele HM, Wolfert MA, Desmarais TR, Palmer
CJ. 2000. Sources of Escherichia coli in a coastal subtropical
environment. Appl Environ Microbiol 66:230–237.
39. Byappanahalli MN, Roll BM, Fujioka RS. 2012. Evidence
for occurrence, persistence, and growth of Escherichia coli
and enterococci in Hawaii’s soil environments. Microbes
Environ 27:164–170.
40. Ferguson D, Signoretto C. 2011. Environmental persis-
tence and naturalization of fecal indicator organisms,
p. 379–397. In Hagedorn C, Blanch AR, Harwood VJ
(eds), Microbial Source Tracking: Methods, Applications,
and Case Studies. Springer, New York.
3.4.6-12 ▪ MICROBIAL SOURCE TRACKING
41. Hagedorn C, Blanch AR, Harwood VJ (eds). 2011.
Microbial Source Tracking: Methods, Applications, and Case
Studies. Springer, New York.
42. Unno T, Jang J, Han D, Kim JH, Sadowsky MJ, Kim OS,
Chun J, Hur H-G. 2010. Use of barcoded pyrosequencing
and shared OTUs to determine sources of fecal bacteria in
watersheds. Environ Sci Technol 44:7777–7782.
43. Bradford SA, Wang Y, Kim H, Torkzaban S, Šimů nek J.
2014. Modeling microorganism transport and survival in
the subsurface. J Environ Qual 43:421–440.
44. Yates MV, Yates SR, Wagner J, Gerba CP. 1987. Model-
ing virus survival and transport in the subsurface. J Contam
Hydrol 1:329–345.
45. Dukta BJ. 1973. Coliforms are an inadequate index of
water quality. J Environ Health 36:39–46.
46. Byappanahalli MN, Nevers MB, Korajkic A, Staley ZR,
Harwood VJ. 2012. Enterococci in the environment: a
review. Microbiol Mol Biol Rev 76:685–706.
47. Winfield MD, Groisman EA. 2003. Role of nonhost envi-
ronments in the lifestyles of Salmonella and Escherichia coli.
Appl Environ Microbiol 69:3687–3694.
48. Fujioka R, Sian-Denton C, Borja M, Castro J, Morphew
K. 1999. Soil: the environmental source of Escherichia coli
and enterococci in Guam’s streams. J Appl Microbiol Symp
Suppl 85:83S–89S.
49. Brennan FP, O’Flaherty V, Kramers G, Grant J, Richards
KG. 2010. Long-term persistence and leaching of Escheri-
chia coli in temperate maritime soils. Appl Environ Microbiol
76:1449–1455.
50. Hardina CM, Fujioka RS. 1991. Soil: the environmental
source of Escherichia coli and enterococci in Hawaii’s
streams. Environ Toxicol Water Qual 6:185–195.
51. Ishii S, Yan T, Shively DA, Byappanahalli MN, Whit-
man RL, Sadowsky MJ. 2006. Cladophora (Chlorophyta)
spp. harbor human bacterial pathogens in nearshore water
of Lake Michigan. Appl Environ Microbiol 72:4545– 4553.
52. Lasalde C, Rodriguez R, Toranzos GA, Smith HH. 2005.
Heterogeneity of uidA gene in environmental Escherichia
coli populations. J Water Health 3:297–304.
53. Ishii S, Ksoll WB, Hicks RE, Sadowsky MJ. 2006. Pres-
ence and growth of naturalized Escherichia coli in temperate
soils from Lake Superior watersheds. Appl Environ Microbiol
72:612–621.
54. Whitman RL, Nevers MB, Byappanahalli MN. 2006.
Watershed-wide distribution of Escherichia coli along south-
ern Lake Michigan: an integrated approach. Appl Environ
Microbiol 72:7301–7310.
55. Byappanahalli M, Fujioka R. 2004. Indigenous soil bacte-
ria and low moisture may limit but allow faecal bacteria to
multiply and become a minor population in tropical soils.
Water Sci Technol 50:27–32.
56. Byappanahalli MN. 2000. Assessing the persistence and
multiplication of fecal indicator bacteria in Hawaii soil
environment. Ph.D. thesis, University of Hawaii at Manoa,
Honolulu.
57. Brennan FP, Grant J, Botting CH, O’Flaherty V,
Richards KG, Abram F. 2012. Insights into the low-
temperature adaptation and nutritional flexibility of a
soil-persistent Escherichia coli. FEMS Microbiol Ecol 84:
75–85.
58. Fujioka RS, Byappanahalli MN. 2001. Microbial ecology
controls the establishment of fecal bacteria in tropical soil
environment, p. 273–283. In Matsuo T, Hanaki K, Taki-
zawa S, Satoh H (eds), Advances in Water and Wastewater
Treatment Technology: Molecular Technology, Nutrient
Removal, Sludge Reduction, and Environmental Health. Elsev-
ier Science, Amsterdam.
59. Anderson KL, Whitlock JE, Harwood VJ. 2005. Persis-
tence and differential survival of fecal indicator bacteria
in subtropical waters and sediments. Appl Environ Microbiol
71:3041–3048.
60. Davies CM, Long JAH, Donald M, Ashbolt NJ.
1995. Survival of fecal microorganisms in marine and
freshwater sediments. Appl Environ Microbiol 61:1888–
1896.
61. Gerba CP, McLeod JS. 1976. Effects of sediments on the
survival of Escherichia coli in marine waters. Appl Environ
Microbiol 32:114–120.
62. Sayler GS, Nelson J, Justice A, Colwell RR. 1975. Distri-
bution and significance of fecal indicator organisms in the
Upper Chesapeake Bay. Appl Microbiol 30:625–638.
63. Desmarais TR, Solo-Gabriele HM, Palmer CJ. 2002.
Influence of soil on fecal indicator organisms in a tidally
influenced subtropical environment. Appl Environ Micro-
biol 68:1165–1172.
64. Gary HL, Adams JC. 1985. Indicator bacteria in water and
stream sediments near the Snowy Range in southern
Wyoming. Water Air Soil Pollut 25:133–144.
65. An YJ, Kampbell DH, Breidenbach GP. 2002. Escherichia
coli and total coliforms in water and sediments at lake mari-
nas. Environ Pollut 120:771–778.
66. Craig DL, Fallowfield HJ, Cromar NJ. 2004. Use of micro-
cosms to determine persistence of Escherichia coli in recrea-
tional coastal water and sediment and validation with in
situ measurements. J Appl Microbiol 96:922–930.
67. Grimes DJ. 1980. Bacteriological water quality effects of
hydraulically dredging contaminated upper Mississippi
River bottom sediment. Appl Environ Microbiol 39:
782–789.
68. Steets BM, Holden PA. 2003. A mechanistic model of
runoff-associated fecal coliform fate and transport through
a coastal lagoon. Water Res 37:589–608.
69. Jamieson RC, Joy DM, Lee H, Kostaschuk R, Gordon
RJ. 2005. Resuspension of sediment-associated Escherichia
coli in a natural stream. J Environ Qual 34:581–589.
70. Wilkinson J, Kay D, Wyer M, Jenkins A. 2006. Processes
driving the episodic flux of faecal indicator organisms in
streams impacting on recreational and shellfish harvesting
waters. Water Res 40:153–161.
71. Garcia-Armisen T, Servais P. 2007. Respective contribu-
tions of point and non-point sources of E. coli and entero-
cocci in a large urbanized watershed (the Seine River,
France). J Environ Manag 82:512–518.
72. Ferguson C, De Roda Husman AM, Altavilla N, Deere
D, Ashbolt N. 2003. Fate and transport of surface water
pathogens in watersheds. Crit Rev Environ Sci Technol 33:
299–361.
73. Unc A, Goss MJ. 2004. Transport of bacteria from manure
and protection of water resources. J Appl Ecol 25:1–18.
74. Dorner SM, Anderson WB, Slawson RM, Kouwen N,
Huck PM. 2006. Hydrologic modeling of pathogen fate
and transport. Environ Sci Technol 40:4746–4753.
75. Jamieson R, Joy DM, Lee H, Kostaschuk R, Gordon R.
2005. Transport and deposition of sediment-associ-
ated Escherichia coli in natural streams. Water Res 39:
2665–2675.
76. Byappanahalli MN, Whitman RL, Shively DA, Evert
Ting WT, Tseng CC, Nevers MB. 2006. Seasonal persis-
tence and population characteristics of Escherichia coli and
enterococci in deep backshore sand of two freshwater
beaches. J Water Health 4:313–320.
77. Whitman RL, Nevers MB. 2003. Foreshore sand as a
source of Escherichia coli in nearshore water of a Lake Mich-
igan beach. Appl Environ Microbiol 69:5555–5562.
78. Ghinsberg RC, Leibowitz P, Witkin H, Mates A, Sein-
berg Y, Bar DL, Nitzan Y, Rogol M. 1994. Monitoring of
selected bacteria and fungi in sand and seawater along the Tel
Aviv coast. MAP Technical Reports Series 87:65–81.
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-13
79. Halliday E, Gast RJ. 2011. Bacteria in beach sands: an
emerging challenge in protecting coastal water quality
and bather health. Environ Sci Technol 45:370–379.
80. Phillips MC, Solo-Gabriele HM, Piggot AM, Klaus JS,
Zhang YJ. 2011. Relationships between sand and water
quality at recreational beaches. Water Res 45:6763–6769.
81. Yamahara KM, Layton BA, Santoro AE, Boehm AB.
2007. Beach sands along the California coast are diffuse
sources of fecal bacteria to coastal waters. Environ Sci Tech-
nol 41:4515–4521.
82. Fogarty LR, Haack SK, Wolcott MJ, Whitman RL. 2003.
Abundance and characteristics of the recreational water
quality indicator bacteria Escherichia coli and enterococci
in gull faeces. J Appl Microbiol 94:865–878.
83. Kinzelman J, Pond K, Longmaid K, Bagley R. 2004. The
effect of two mechanical beach grooming strategies on
Escherichia coli density in beach sand at a southwestern
Lake Michigan beach. Aquat Ecosyst Health Manage 7:
425–432.
84. Lévesque B, Brousseau P, Simard P, Dewailly E, Meisels
M, Ramsay D, Joly J. 1993. Impact of the ring-billed
gull (Larus delawarensis) on the microbiological quality
of recreational water. Appl Environ Microbiol 59:1228–
1230.
85. Standridge JH, Delfino JJ, Kleppe LB, Butler R. 1979.
Effect of waterfowl (Anas platyrhynchos) on indicator bacte-
ria populations in a recreational lake in Madison, Wiscon-
sin. Appl Environ Microbiol 38:547–550.
86. Wyer MD, Kay D, Jackson GF, Dawson HM, Yeo J, Tan-
guy L. 1995. Indicator organism sources and coastal water
quality: a catchment study on the Island of Jersey. J Appl
Bacteriol 78:290–296.
87. Ge Z, Whitman RL, Nevers MB, Phanikumar MS. 2012.
Wave-induced mass transport affects daily Escherichia coli
fluctuations in nearshore water. Environ Sci Technol 46:
2204–2211.
88. Le Fevre NM, Lewis GD. 2003. The role of resuspension
in enterococci distribution in water at an urban beach.
Water Sci Technol 47:205–210.
89. Gast RJ, Gorrell L, Raubenheimer B, Elgar S. 2011.
Impact of erosion and accretion on the distribution
of enterococci in beach sands. Cont Shelf Res 31:1457–
1461.
90. He LM, He ZL. 2008. Water quality prediction of marine
recreational beaches receiving watershed baseflow and
stormwater runoff in southern California, USA. Water
Res 42:2563–2573.
91. Boehm AB, Shellenbarger GG, Paytan A. 2004. Ground-
water discharge: potential association with fecal indicator
bacteria in the surf zone. Environ Sci Technol 38:
3558–3566.
92. Fong TT, Mansfield LS, Wilson DL, Schwab DJ, Molloy
SL, Rose JB. 2007. Massive microbiological groundwater
contamination associated with a waterborne outbreak in
Lake Erie, South Bass Island, Ohio. Environ Health Perspect
115:856–864.
93. Page RM, Scheidler S, Polat E, Svoboda P, Huggenberger
P. 2012. Faecal indicator bacteria: groundwater dynamics
and transport following precipitation and river water infil-
tration. Water Air Soil Pollut 223:2771–2782.
94. Anderson SA, Turner SJ, Lewis GD. 1997. Enterococci
in the New Zealand environment: implications for water
quality monitoring. Water Sci Technol 35:325–331.
95. Grant SB, Sanders BF, Boehm AB, Redman JA, Kim JH,
Mrše RD, Chu AK, Gouldin M, McGee CD, Gardiner
NA, Jones BH, Svejkovsky J, Leipzig GV, Brown A.
2001. Generation of enterococci bacteria in a coastal
saltwater marsh and its impact on surf zone water quality.
Environ Sci Technol 35:2407–2416.
96. Imamura GJ, Thompson RS, Boehm AB, Jay JA. 2011.
Wrack promotes the persistence of fecal indicator bacteria
in marine sands and seawater. FEMS Microbiol Ecol 77:
40–49.
97. Whitman RL, Shively DA, Pawlik H, Nevers MB, Byap-
panahalli MN. 2003. Occurrence of Escherichia coli and
enterococci in Cladophora (Chlorophyta) in nearshore
water and beach sand of Lake Michigan. Appl Environ
Microbiol 69:4714 – 4719.
98. Muller T, Ulrich A, Ott EM, Muller M. 2001. Identifica-
tion of plant-associated enterococci. J Appl Microbiol 91:
268–278.
99. Badgley BD, Thomas FIM, Harwood VJ. 2010. The
effects of submerged aquatic vegetation on the persistence
of environmental populations of Enterococcus spp. Environ
Microbiol 12:1271–1281.
100. Ott E-M, Muller T, Muller M, Franz CMAP, Ulrich A,
Gabel M, Seyfarth W. 2001. Population dynamics and
antagonistic potential of enterococci colonizing the phyl-
losphere of grasses. J Appl Microbiol 91:54–66.
101. Byappanahalli MN, Sawdey R, Ishii S, Shively DA, Fer-
guson J, Whitman RL, Sadowsky MJ. 2009. Seasonal
stability of Cladophora-associated Salmonella in Lake Mich-
igan watersheds. Water Res 43:806–814.
102. Badgley BD, Nayak BS, Harwood VJ. 2010. The impor-
tance of sediment and submerged aquatic vegetation as
potential habitats for persistent strains of enterococci in a
subtropical watershed. Water Res 44:5857–5866.
103. Badgley BD, Thomas FIM, Harwood VJ. 2011. Quantify-
ing environmental reservoirs of fecal indicator bacteria
associated with sediment and submerged aquatic vegeta-
tion. Environ Microbiol 13:932–942.
104. Whitman RL, Nevers MB, Przybyla-Kelly K, Byappana-
halli MN. 2011. Physical and biological factors influencing
environmental sources of fecal indicator bacteria in surface
water, p. 111–134. In Sadowsky MJ, Whitman RL (eds),
The Fecal Bacteria. ASM Press, Washington, DC.
105. Byappanahalli MN, Ishii S. 2011. Environmental
sources of fecal bacteria, p. 93–110. In Sadowsky MJ, Whit-
man RL (eds), The Fecal Bacteria. ASM Press, Washington,
DC.
106. Gauthier F, Archibald F. 2001. The ecology of fecal indi-
cator bacteria commonly found in pulp and paper mill
water systems. Water Res 35:2207–2218.
107. Beauchamp CJ, Simao-Beaunoir A-M, Beaulieu C, Cha-
lifour F-P. 2006. Confirmation of E. coli among other ther-
motolerant coliform bacteria in paper mill effluents, wood
chips screening rejects and paper sludges. Water Res 40:
2452–2462.
108. Byappanahalli MN, Shively DA, Nevers MB, Sadowsky
MJ, Whitman RL. 2003. Growth and survival of Escheri-
chia coli and enterococci populations in the macro-alga
Cladophora (Chlorophyta). FEMS Microbiol Ecol 46:
203–211.
109. Grimes DJ. 1975. Release of sediment bound fecal coli-
forms by dredging. Appl Microbiol 29:109–111.
110. Crowther J, Wyer MD, Bradford M, Kay D, Francis CA.
2003. Modelling faecal indicator concentrations in large
rural catchments using land use and topographic data.
J Appl Microbiol 94:962–973.
111. Bai S, Lung WS. 2005. Modeling sediment impact on the
transport of fecal bacteria. Water Res 39:5232–5240.
112. Noble RT, Weisberg SB, Leecaster MK, McGee CD,
Dorsey JH, Vainik P, Orozco-Borbón V. 2003. Storm
effects on regional beach water quality along the southern
California shoreline. J Water Health 1:23–31.
113. Parker JK, McIntyre D, Noble RT. 2010. Characterizing
fecal contamination in stormwater runoff in coastal North
Carolina, USA. Water Res 44:4186–4194.
3.4.6-14 ▪ MICROBIAL SOURCE TRACKING
114. Wright ME, Abdelzaher AM, Solo-Gabriele HM, Elmir
S, Fleming LE. 2011. The inter-tidal zone is the pathway
of input of enterococci to a subtropical recreational marine
beach. Water Sci Technol 63:542–549.
115. Ho LC, Litton RM, Grant SB. 2011. Anthropogenic cur-
rents and shoreline water quality in Avalon Bay, California.
Environ Sci Technol 45:2079–2085.
116. Phillips MC, Solo-Gabriele HM, Reniers AJHM, Wang
JD, Kiger RT, Abdel-Mottaleb N. 2011. Pore water trans-
port of enterococci out of beach sediments. Mar Pollut Bull
62:2293–2298.
117. Russell TL, Yamahara KM, Boehm AB. 2012. Mobiliza-
tion and transport of naturally occurring enterococci in
beach sands subject to transient infiltration of seawater.
Environ Sci Technol 46:5988–5996.
118. Badgley BD, Ferguson J, Vanden Heuvel A, Kleinheinz
GT, McDermott CM, Sandrin TR, Kinzelman J, Junion
EA, Byappanahalli MN, Whitman RL, Sadowsky MJ.
2011. Multi-scale temporal and spatial variation in geno-
typic composition of Cladophora-borne Escherichia coli pop-
ulations in Lake Michigan. Water Res 45:721–731.
119. Bermudez M, Hazen TC. 1988. Phenotypic and genotypic
comparison of Escherichia coli from pristine tropical waters.
Appl Environ Microbiol 54:979–983.
120. Whitman RL, Byers SE, Shively DA, Ferguson DM,
Byappanahalli MN. 2005. Occurrence and growth charac-
teristics of Escherichia coli and enterococci within the accu-
mulated fluid of the northern pitcher plant (Sarracenia
purpurea L.). Can J Microbiol 51:1027–1037.
121. Byappanahalli MN, Whitman RL, Shively DA, Ting
WTE, Tseng CC, Nevers MB. 2006. Seasonal persistence
and population characteristics of Escherichia coli and enter-
ococci in deep backshore sand of two freshwater beaches.
J Water Health 4:313–320.
122. Yamahara KM, Walters SP, Boehm AB. 2009. Growth
of enterococci in unaltered, unseeded beach sands
subjected to tidal wetting. Appl Environ Microbiol 75:
1517–1524.
123. American Public Health Association. 2005. Standard
Methods for the Examination of Water and Wastewater, 21st
ed. American Public Health Association, Washington,
DC.
124. Niemala SI, Lee JV, Fricker CR. 2003. A comparison of
the International Standards Organization reference
method for the detection of coliforms and Escherichia
coli in water with a defined substrate procedure. J Appl
Microbiol 95:1285–1292.
125. Prüss A. 1998. Review of epidemiological studies on
health effects from exposure to recreational water.
Int J Epidemiol 27:1–9.
126. Calderon RL, Mood EW, Dufour AP. 1991. Health effects
of swimmers and nonpoint sources of contaminated water.
Int J Environ Health Res 1:21–31.
127. Fleisher JM, Fleming LE, Solo-Gabriele HM, Kish JK,
Sinigalliano CD, Plano L, Elmir SM, Wang JD, Withum
K, Shibata T, Gidley ML, Abdelzaher A, He G, Ortega C,
Zhu X, Wright M, Hollenbeck J, Backer LC. 2010. The
BEACHES Study: health effects and exposures from non-
point source microbial contaminants in subtropical recrea-
tional marine waters. Int J Epidemiol 39:1291–1298.
128. Drozd M, Merrick NN, Sanad YM, Dick LK, Dick WA,
Rajashekara G. 2013. Evaluating the occurrence of host-
specific bacteroidales, general fecal indicators, and bacte-
rial pathogens in a mixed-use watershed. J Environ Qual
42:713–725.
129. Hörman A, Rimhanen-Finne R, Maunula L, Von Bons-
dorff CH, Torvela N, Heikinheimo A, Hänninen ML.
2004. Campylobacter spp., Giardia spp., Cryptosporidium
spp., boroviruses, and indicator organisms in surface water
in Southwestern Finland, 2000–2001. Appl Environ Micro-
biol 70:87–95.
130. Oladeinde A, Bohrmann T, Wong K, Purucker ST, Brad-
shaw K, Brown R, Snyder B, Molina M. 2014. Decay of
fecal indicator bacterial populations and bovine-associated
source-tracking markers in freshly deposited cow pats. Appl
Environ Microbiol 80:110–118.
131. Roslev P, Bastholm S, Iversen N. 2008. Relationship
between fecal indicators in sediment and recreational
waters in a Danish estuary. Water Air Soil Pollut 194:13–21.
132. Bhattarai R, Kalita P, Trask J, Kuhlenschmidt MS. 2011.
Development of a physically-based model for transport of
Cryptosporidium parvum in overland flow. Environ Modell
Softw 26:1289–1297.
133. Desai A, Rifai HS, Petersen TM, Stein R. 2011. Mass bal-
ance and water quality modeling for load allocation of
Escherichia coli in an urban watershed. J Water Res Pl
ASCE 137:412–427.
134. Harmel RD, Karthikeyan R, Gentry T, Srinivasan R.
2010. Effects of agricultural management, land use, and
watershed scale on E. coli concentrations in runoff and
streamflow. T ASABE 53:1833–1841.
135. Pachepsky YA, Sadeghi AM, Bradford SA, Shelton DR,
Guber AK, Dao T. 2006. Transport and fate of manure-
borne pathogens: modeling perspective. Agric Water Manag
86:91–92.
136. Petersen CM, Rifai HS, Stein R. 2009. Bacteria load esti-
mator spreadsheet tool for modeling spatial Escherichia coli
loads to an urban bayou. J Environ Engineer-ASCE 135:
203–217.
137. Riebschleager KJ, Karthikeyan R, Srinivasan R, McKee
K. 2012. Estimating potential E. coli sources in a watershed
using spatially explicit modeling techniques. J Amer Water
Resour Assn 48:745–761.
138. Smith JH, Wickham JD, Norton D, Wade TG, Jones KB.
2001. Utilization of landscape indicators to model poten-
tial pathogen impaired waters. J Amer Water Resour Assn
37:805–814.
139. Surbeck CQ, Jiang SC, Ahn JH, Grant SB. 2006. Flow
fingerprinting fecal pollution and suspended solids in
stormwater runoff from an urban coastal watershed. Environ
Sci Technol 40:4435–4441.
140. Teague A, Karthikeyan R, Babbar-Sebens M, Srinivasan
R, Persyn RA. 2009. Spatially explicit load enrichment
calculation tool to identify potential E. coli sources in
watersheds. T ASABE 52:1109–1120.
141. Vescovi L, Mailhot A, Rousseau AN, Villeneuve JP.
1999. Hydrobiochemical modeling of the Saint-Charles
watershed. 2: water quality. Water Qual Res J Can 34:
335–360.
142. Viau EJ, Goodwin KD, Yamahara KM, Layton BA,
Sassoubre LM, Burns SL, Tong HI, Wong SH, Lu Y,
Boehm AB. 2011. Bacterial pathogens in Hawaiian coastal
streams—associations with fecal indicators, land cover, and
water quality. Water Res 45:3279–3290.
143. Vidon P, Tedesco LP, Wilson J, Campbell MA, Casey
LR, Gray M. 2008. Direct and indirect hydrological con-
trols on E. coli concentration and loading in midwestern
streams. J Environ Qual 37:1761–1768.
144. Zeckoski RW, Benham BL, Shah SB, Wolfe ML,
Brannan KM, Al-Smadi M, Dillaha TA, Mostaghimi S,
Heatwole CD. 2005. Bslc: a tool for bacteria source charac-
terization for watershed management. Appl Eng Agric 21:
879–889.
145. Benham BL, Baffaut C, Zeckoski RW, Mankin KR,
Pachepsky YA, Sadeghi AA, Brannan KM, Soupir ML,
Habersack MJ. 2006. Modeling bacteria fate and trans-
port in watersheds to support TMDLs. T ASABE 49:
987–1002.
 3.4.6. Fecal Indicator Organism Modeling in Environmental Waters ▪ 3.4.6-15
146. Weill S, Mazzia A, Putti M, Paniconi C. 2011. Coupling
water flow and solute transport into a physically-based
surface-subsurface hydrological model. Adv Water Resour
34:128–136.
147. VanderKwaak JE. 1999. Numerical Simulation of Flow and
Chemical Transport in Integrated Surface-Subsurface Hydro-
logic Systems. University of Waterloo, Waterloo.
148. Kolditz O, Bauer S, Bilke L, Böttcher N, Delfs JO,
Fischer T, Görke UJ, Kalbacher T, Kosakowski G,
McDermott CI, Park CH, Radu F, Rink K, Shao H,
Shao HB, Sun F, Sun YY, Singh AK, Taron J, Walther
M, Wang W, Watanabe N, Wu Y, Xie M, Xu W, Zehner
B. 2012. OpenGeoSys: an open-source initiative for
numerical simulation of thermo-hydro-mechanical/chemi-
cal (THM/C) processes in porous media. Environ Earth Sci
67:589–599.
149. Kollet SJ, Maxwell RM. 2006. Integrated surface-
groundwater flow modeling: a freesurface overland flow
boundary condition in a parallel groundwater flow model.
Adv Water Resour 29:945–958.
150. Shen C, Phanikumar MS. 2010. A process-based, distrib-
uted hydrologic model based on a large-scale method for
surface-subsurface coupling. Adv Water Resour 33:1524–
1541.
151. Yeh GT, Shih DS, Cheng JRC. 2011. An integrated
media, integrated processes watershed model. Comput Flu-
ids 45:2–13.
152. Maxwell RM, Putti M, Meyerhoff S, Delfs J-O, Ferguson
IM, Ivanov VY, Kim J, Kolditz O, Kollet SJ, Kumar M,
Lopez S, Niu J, Paniconi C, Park Y-J, Phanikumar MS,
Shen CP, Sudicky EA, Sulis M. 2014. Surface-subsurface
model intercomparison: a first set of benchmark results to
diagnose integrated hydrology and feedbacks. Water Resour
Res 50:1531–1549.
153. Niu J. 2013. Modeling the movement of water, bacteria
and nutrients across heterogeneous landscapes in the great
lakes region using a process-based hydrologic model. Ph.D.
thesis, Michigan State University, East Lansing.
154. Parajuli PB, Mankin KR, Barnes PL. 2009. Source spe-
cific fecal bacteria modeling using soil and water assessment
tool model. Bioresour Technol 100:953–963.
155. Parajuli PB, Douglas-Mankin KR, Barnes PL, Rossi CG.
2009. Fecal bacteria source characterization and sensitivity
analysis of SWAT 2005. T ASABE 52:1847–1858.
156. Cho KH, Pachepsky YA, Kim JH, Guber AK, Shelton
DR, Rowland R. 2010. Release of Escherichia coli from
the bottom sediment in a first-order creek: experiment
and reach-specific modeling. J Hydrol 391:322–332.
157. Guber AK, Karns JS, Pachepsky YA, Sadeghi AM, Van
Kessel JS, Dao TH. 2007. Comparison of release and trans-
port of manure-borne Escherichia coli and enterococci under
grass buffer conditions. Lett Appl Microbiol 44:161–167.
158. Guber AK, Shelton DR, Pachepsky YA, Sadeghi AM,
Sikora LJ. 2006. Rainfall-induced release of fecal coliforms
and other manure constituents: comparison and modeling.
Appl Environ Microbiol 72:7531–7539.
159. Kim JW, Pachepsky YA, Shelton DR, Coppock C. 2010.
Effect of streambed bacteria release on E. coli concentra-
tions: monitoring and modeling with the modified
SWAT. Ecol Model 221:1592–1604.
160. Pachepsky YA, Garzio-Hadzick A, Shelton DR, Hadzick
ZZ, Hill RL. 2011. Survival of E. coli O157:H12 in creek
sediments after inoculation and re-inoculation. Int J Envi-
ron Pollut 46:234–245.
161. Pachepsky YA, Guber AK, Shelton DR, McCarty GW.
2009. Size distributions of manure particles released under
simulated rainfall. J Environ Manag 90:1365–1369.
162. Shelton DR, Pachepsky YA, Sadeghi AM, Stout
WL, Karns JS, Gburek WJ. 2003. Release rates of
manure-borne coliform bacteria from data on leaching
through stony soil. Vadose Zone J 2:34–39.
163. Stout WL, Pachepsky YA, Shelton DR, Sadeghi AM,
Saporito LS, Sharpley AN. 2005. Runoff transport of fae-
cal coliforms and phosphorus released from manure in grass
buffer conditions. Lett Appl Microbiol 41:230–234.
164. Germann PF, Smith MS, Thomas GW. 1987. Kinematic
wave approximation to the transport of Escherichia coli in
the vadose zone. Water Resour Res 23:1281–1287.
165. Cho KH, Pachepsky YA, Kim JH, Kim JW, Park MH.
2012. The modified SWAT model for predicting fecal coli-
forms in the Wachusett Reservoir Watershed, USA. Water
Res 46:4750–4760.
166. Petersen CM, Rifai HS, Villarreal GC, Stein R. 2011.
Modeling Escherichia coli and its sources in an urban
bayou with hydrologic simulation program-FORTRAN.
J Environ Engineer-ASCE 137:487–503.
167. Nekouee N. 2010. Dynamics and Numerical Modeling of
River Plumes in Lakes. Technical Memorandum GLERL-
151. NOAA, Washington, DC.
168. Wendzel A. 2014. Constraining mechanistic models of
indicator bacteria at recreational beaches in Lake Michigan
using easily-measurable environmental variables. M.S. the-
sis, Michigan State University, East Lansing.
169. Schimmelpfennig S, Kirillin G, Engelhardt C, Nützmann
G. 2012. Effects of wind-driven circulation on river intru-
sion in Lake Tegel: modeling study with projection on
transport of pollutants. Environ Fluid Mech 12:321–339.
170. Thupaki P, Phanikumar MS, Schwab DJ, Nevers MB,
Whitman RL. 2013. Evaluating the role of sediment-
bacteria interactions on Escherichia coli concentrations at
beaches in southern Lake Michigan. J Geophys Res C
118:7049–7065.
171. Ge Z, Whitman RL, Nevers MB, Phanikumar MS, Byap-
panahalli MN. 2012. Nearshore hydrodynamics as loading
and forcing factors for Escherichia coli contamination at an
embayed beach. Limnol Oceanogr 57:362–381.
172. Russell TL, Sassoubre LM, Wang D, Masuda S, Chen H,
Soetjipto C, Hassaballah A, Boehm AB. 2013. A coupled
modeling and molecular biology approach to microbial
source tracking at Cowell Beach, Santa Cruz, CA, United
States. Environ Sci Technol 47:10231–10239.
173. Sokolova E, Aström J, Pettersson TJR, Bergstedt O, Her-
mansson M. 2012. Estimation of pathogen concentrations
in a drinking water source using hydrodynamic modelling
and microbial source tracking. J Water Health 10:358–370.
174. Thupaki P, Phanikumar MS, Whitman RL. 2013. Mixing
and transport in the coastal boundary layer of southern
Lake Michigan. J Geophys Res C 118:1–12.
175. Brooks WR, Fienen MN, Corsi SR. 2013. Partial least
squares for efficient models of fecal indicator bacteria on
Great Lakes beaches. J Environ Manag 114:470–475.
176. Cahoon LB, Hales JC, Carey ES, Loucaides S, Rowland
KR, Nearhoof JE. 2006. Shellfishing closures in southwest
Brunswick County, North Carolina: septic tanks vs. storm-
water runoff as fecal coliform sources. J Coast Res 22:
319–327.
177. Edge TA, Hill S, Seto P, Marsalek J. 2010. Library-
dependent and library-independent microbial source track-
ing to identify spatial variation in faecal contamination
sources along a Lake Ontario beach (Ontario, Canada).
Water Sci Technol 62:719–727.
178. Goulding R, Jayasuriya N, Horan E. 2012. A Bayesian
network model to assess the public health risk associated
with wet weather sewer overflows discharging into water-
ways. Water Res 46:4933–4940.
179. Kelsey H, Porter DE, Scott G, Neet M, White D. 2004.
Using geographic information systems and regression anal-
ysis to evaluate relationships between land use and fecal
3.4.6-16 ▪ MICROBIAL SOURCE TRACKING
coliform bacterial pollution. J Exp Mar Biol Ecol 298:
197–209.
180. Wyer MD, Kay D, Watkins J, Davies C, Kay C, Thomas
R, Porter J, Stapleton CM, Moore H. 2010. Evaluating
short-term changes in recreational water quality during a
hydrograph event using a combination of microbial tracers,
environmental microbiology, microbial source tracking
and hydrological techniques: a case study in Southwest
Wales, UK. Water Res 44:4783–4795.
181. U.S. Department of Health and Human Services. 2009.
National shellfish sanitation program guide for the control
of molluscan shellfish, p. 63. FR 38659.
182. Wang J, Deng Z. 2012. Detection and forecasting of oyster
norovirus outbreaks: recent advances and future perspec-
tives. Mar Environ Res 80:62–69.
183. Coulliette AD, Money ES, Serre ML, Noble RT. 2009.
Space/time analysis of fecal pollution and rainfall in an
eastern North Carolina estuary. Environ Sci Technol 43:
3728–3735.
184. Boehm AB, Whitman RL, Nevers MB, Hou D, Weisberg
SB. 2007. Modeling: nowcasting recreational water quality,
p. 179–210. In Wymer LJ (ed), Statistical framework for rec-
reational water quality criteria and monitoring. Wiley, West
Sussex, UK.
185. U.S. EPA. 1999. Action plan for beaches and recreational
waters. EPA/600/R-98/079. U.S. Environmental Protec-
tion Agency. Washington, DC.
186. Nevers MB, Boehm AB. 2011. Modeling fate and trans-
port of fecal bacteria in surface water, p. 165–188. In
Sadowsky MJ, Whitman RL (eds), The Fecal Bacteria.
ASM Press, Washington, DC.
187. Olyphant GA. 2005. Statistical basis for predicting the
need for bacterially induced beach closures: emergence of
a paradigm? Water Res 39:4953–4960.
188. Nevers MB, Whitman RL. 2011. Efficacy of monitoring
and empirical predictive modeling at improving public
health protection at Chicago beaches. Water Res 45:
1659– 1668.
189. David MM, Haggard BE. 2011. Development of
regression-based models to predict fecal bacteria numbers
at select sites within the Illinois River Watershed, Arkansas
and Oklahoma, USA. Water Air Soil Pollut 215:525–547.
190. Kelsey RH, Scott GI, Porter DE, Siewicki TC, Edwards
DG. 2010. Improvements to shellfish harvest area closure
decision making using GIS, remote sensing, and predictive
models. Estuar Coast 5:1–11.
191. Lipp EK, Kurz R, Vincent R, Rodriguez-Palacios C, Far-
rah SR, Rose JB. 2001. The effects of seasonal variability
and weather on microbial fecal pollution and enteric
pathogens in a subtropical estuary. Estuaries 24:266–276.
192. Lee RJ, Morgan OC. 2003. Environmental factors influ-
encing the microbiological contamination of commercially
harvested shellfish. Water Sci Technol 47:65–70.
193. Gonzalez RA, Conn KE, Crosswell JR, Noble RT. 2012.
Application of empirical predictive modeling using con-
ventional and alternative fecal indicator bacteria in eastern
North Carolina waters. Water Res 46:5871–5882.
194. Boehm AB. 2007. Enterococci concentrations in diverse
coastal environments exhibit extreme variability. Environ
Sci Technol 41:8227–8232.
195. Whitman RL, Nevers MB. 2004. Escherichia coli sampling
reliability at a frequently closed Chicago beach: monitoring
and management implications. Environ Sci Technol 38:
4241–4246.
196. Ge Z, Frick WE. 2009. Time-frequency analysis of
beach bacteria variations and its implication for recrea-
tional water quality modeling. Environ Sci Technol 43:
1128–1133.
197. U.S. EPA. 1986. Ambient water quality criteria for bacte-
ria 1986. EPA 440/5-84-002. U.S. EPA, Office of Water
Regulations and Standards, Washington, DC.
198. Sokolova E, Åström J, Pettersson TJR, Bergstedt O, Her-
mansson M. 2012. Decay of bacteroidales genetic markers
in relation to traditional fecal indicators for water quality
modeling of drinking water sources. Environ Sci Technol
46:892–900.
199. Whitman RL, Horvath TG, Goodrich ML, Nevers MB,
Wolcott MJ, Haack SK. 2001. Characterization of E. coli
levels at 63rd Street Beach. Report to the City of Chicago,
Department of the Environment and the Chicago Park
District.
200. Whitman RL, Nevers MB, Korinek GC, Byappanahalli
MN. 2004. Solar and temporal effects on Escherichia coli
concentration at a Great Lakes swimming beach. Appl Envi-
ron Microbiol 70:4276–4285.
201. Byappanahalli MN, Przybyla-Kelly K, Shively DA,
Whitman RL. 2008. Environmental occurrence of the
enterococcal surface protein (esp) gene is an unreliable
indicator of human fecal contamination. Environ Sci Tech-
nol 42:8014–8020.
202. Toledo-Hernandez C, Ryu H, Gonzalez-Nieves J,
Huertas E, Toranzos GA, Santo Domingo J. 2013. Track-
ing the primary sources of fecal pollution in a tropical water-
shed in a one year study. Appl Environ Microbiol 79:
1689–1696.
203. Harwood VJ, Staley C, Badgley BD, Borges K, Korajkic
A. 2014. Microbial source tracking markers for detection
of fecal contamination in environmental waters: relation-
ships to pathogens and human health outcomes. FEMS
Microbiol Rev 38:1–40.
 Risk Assessment Framework
MARYLYNN V. YATES
3.5.1
Formal risk assessment processes inform decisions regarding
potential risks from microorganisms when designing water
projects and have been used for many years. While the process
has been refined over the years and modified for specific pur-
poses, the basic elements of a quantitative risk assessment
process include four components (1):
• Hazard identification: determining whether exposure to a
contaminant, in this case a microorganism, can lead to an
adverse health effect;
• Dose-response assessment: determining the relationship
between the dose of the microorganism to which an indi-
vidual is exposed and the incidence of an adverse health
effect on that individual;
• Exposure assessment: determining the quantity of the
microorganism to which the individual is exposed;
• Risk characterization: determining the type and magni-
tude of the risk to human health.
As discussed in the National Resource Center report (1), a
complete risk analysis requires considering the scientific
information derived from the risk assessment with a risk man-
agement process, in which other considerations (such as
political, technological, societal, and economic considera-
tions) are included (see Fig. 1).
As the use of risk assessment has been used increasingly in
regulatory processes to limit human exposure to pathogenic
microorganisms (and chemicals), the risk assessment frame-
work has been modified to include several important compo-
nents not explicitly present in the original framework. In
2009, the National Research Council (2) described that
framework, as shown in Fig. 2. While the four elements of a
risk assessment are still an integral part of the process, new
components that focus on the design (“problem formula-
tion”) and “evaluation” of the process were included. The
importance of stakeholder involvement throughout the proc-
ess is also highlighted.
PROBLEM FORMULATION
The planning and problem formulation stage of the risk
assessment process was adapted from the U.S. Environmental
Protection Agency’s (EPA) ecological risk assessment process
(3). This stage evolved due to the complexities of ecological
systems, in which multiple sources of contaminants, multiple
doi:10.1128/9781555818821.ch3.5.1
3.5.1-1
types of exposed organisms and numerous stakeholder inter-
ests are commonly involved. According to the EPA (3),
The first step in planning may be to determine if a risk assess-
ment is the best option for supporting the decision. Risk man-
agers and risk assessors both consider the potential value of
conducting a risk assessment to address identified problems.
Their discussion explores what is known about the degree
of risk, what management options are available to mitigate
or prevent it, and the value of conducting a risk assessment
compared with other ways of learning about and addressing
environmental concerns. In some cases, a risk assessment
may add little value to the decision process because manage-
ment alternatives may be available that completely circum-
vent the need for a risk assessment.
Problem formulation is a critical stage in a risk analysis, as
it establishes the foundation for the entire risk assessment. In
this stage, hypotheses about the issue are formed, and objec-
tives for the risk assessment are established. Additionally, a
plan for analyzing data and characterizing the risk is devel-
oped. A thorough formulation of the problem can avoid later
problems such as failing to identify all possible risks, or estab-
lishing endpoints that are either too ambiguous to measure, or
have different meanings to different stakeholders. Spending
time at the beginning formulating well-defined objectives
will also aid in determining the types of expertise that will
be needed for the process.
A diagram of the problem formulation process is shown in
Fig. 3 (4). The steps in the process are as follows (4):
• Define the concern that is the basis for the risk assessment:
what hazard is being addressed and how does it impact
human health (in this case, when exposed via drinking
water)?
• Define clearly and concisely the purpose and objectives of
the risk assessment.
• Understand the history and context of the risk assessment
within the regulatory agency.
• Define the scope of the analysis; this should include:
• Which infectious disease(s) or pathogens are being
addressed?
• Which human populations will be included, and which
will be excluded?
• What health outcomes or endpoints (e.g., infection, ill-
ness, mortality) are being addressed?
3.5.1-2 ▪ MICROBIAL RISK ASSESSMENT
FIGURE 1 Elements of risk assessment and risk management. Redrawn from reference 1, with permission.
doi:10.1128/9781555818821.ch3.5.1.f1
• What unit and routes of exposure are being included?
• For risk assessments in which criteria for acceptable
levels of microorganisms are being set, such as would
be the case for drinking water, what is the target accept-
able risk level? What is the justification for the criteria
(e.g., policy, technical)?
• What specific exposure scenarios will be modeled,
including any spatial and/or temporal aspects?
• Agree on the questions that should be answered by the risk
assessment: these need to be framed within the specific
context, such as policy development, and ensure that
the risk assessment is scientifically defensible. This allows
the risk managers and risk assessors to work together to
ensure that the process is designed properly.
• Develop the conceptual model, which is a graphical repre-
sentation of the scenario being addressed in the risk assess-
ment. The conceptual model and its associated narrative
should:
• Provide a flow chart that illustrates how the risk is
believed to occur
• Outline the models and other tools needed to assess the
risk
• Identify available databases
• Identify default assumptions
• Describe what the risk assessment will or will not be
able to do, and whether it will be qualitative or
quantitative
• Summarize data quality and any data gaps that exist
• Consider interactions between the pathogen, human
host, and environment
• Define key uncertainties that will affect the outcome
• Identify management actions and places where inter-
ventions can occur
• Develop the analysis plan, including strategies to handle
any needs for additional data, obtaining a peer review of
the risk assessment, and other logistical needs.
• Agree on participants and their roles and responsibilities.
• Agree on schedule and deliverables.
• Agree on analytical approaches to ensure that the model
chosen is appropriate for the situation. This should
include information on:
• Tools, including software, mathematical models, and
means of characterizing and handling uncertainty
• Data inventory, including publications, databases,
search engines to be sued, and so on
• Summary of assumptions, including how the assump-
tions contribute to uncertainty and limit the scope of
the risk assessment
• Sources of variability and uncertainty, and how they are
treated in the risk analysis
• Factors and data not included in the risk assessment,
and how they affect the use of the information
• Identified gaps in the knowledge base, such as a lack of
appropriate data, inadequate analytical or statistical
methods.

FIGURE 2 Risk assessment framework schematic. Reprinted from reference 2, with permission. doi:10.1128/9781555818821.ch3.5.1.f2
• Environmental sampling strategies and analysis meth-
ods, including accuracy, precision, and bias in measure-
ments. Especially relevant for analyzing environmental
samples for microorganisms are issues such as the per-
cent recovery of the microorganisms, and whether the
method detects culturable organisms, viable organisms,
nucleic acid, and so on
• Evaluation of management practices: in some situa-
tions, it may be appropriate to identify which compo-
nents in the risk assessment can influence or be
influenced by management actions.
It is important to note that as the problem formulation
proceeds, the statement of concern, statement of purpose,
and questions to be considered may evolve. It is critical for
risk managers to be involved from the beginning, as they
are responsible for ensuring that the problem formulation
documentation is developed in a way that it will address
the problem. Using the information about the scope of
the risk assessment, and considering the available data,
risk assessment tools, and time and resources, the risk asses-
sors must determine whether it is feasible to carry out the
risk assessment. They develop a preliminary conceptual
model and analysis plan. A screening-level risk assessment
may be performed to determine if the questions can be
3.5.1. Risk Assessment Framework ▪ 3.5.1-3
addressed without a formal quantitative risk assessment.
Depending on the specific situation, this may provide
sufficient information for decision making. In situations
that require a formal quantitative risk assessment, and the
appropriate data and resources are available, a detailed con-
ceptual model and analysis plan are developed. The docu-
mentation developed by the risk assessors during problem
formulation can be used to develop options for the risk man-
agers to use when making policy decisions that are needed to
define the scope of the risk assessment. Risk assessments by
their very nature are iterative process, so as new data are
obtained, or health priorities change, problem formulation
may need to be revisited.
CONDUCTING THE RISK ASSESSMENT
During the problem formulation process, a clear idea of the
how the risk assessment process itself will be conducted
should have been developed. In addition, some of the infor-
mation that will be required to conduct the risk assessment
will have been gathered, and needs for additional data iden-
tified. The steps involved in conducting microbial risk assess-
ment will now be described.
3.5.1-4 ▪ MICROBIAL RISK ASSESSMENT
FIGURE 3 Problem formulation process diagram. Reprinted from reference 3, with permission. doi:10.1128/9781555818821.ch3.5.1.f3
Hazard Identification
The first step in the risk assessment process is hazard identifi-
cation, in which the contaminants of concern are identified.
The microorganisms of concern will vary, depending on the
specific situation. For example, if conducting an assessment
to determine the risks associated with use of treated domestic
wastewater for irrigation of crops, one would consider the
microorganisms that could be present in domestic wastewater
(e.g., in human fecal material and urine, on food products
washed in the kitchen sink). One would also have to consider
the effects of wastewater treatment on the pathogens. If the
situation involves calculating risk from swimming in fresh-
water rivers and lakes, one would consider pathogens in
human, livestock, and wildlife feces and urine as hazards.
The ability of the microorganisms to survive, multiply, and
accumulate in the environment would also be important
considerations.
Typically, the outcome from exposure to pathogenic
microorganisms is acute, rather than chronic, disease. Acute
illnesses generally occur after only one or a few exposures to
human pathogenic microorganisms. The length of time
between exposure and the onset of adverse health effects,
known as the incubation period, is characteristic of each
pathogen, and can be from as short as a few hours (e.g., Shi-
gella) to several weeks (e.g., hepatitis A virus).
In many situations, there may be more than one type of
microorganism present, in which case multiple hazards need
to be included in the risk assessment calculations. It is possi-
ble that the presence of multiple pathogens could result in dif-
ferent hazards (e.g., synergistic effects) than if only a single
strain of pathogens were present. Unfortunately, this is an
area in which little information is available to inform risk
assessments. Exposure to a mixture of contaminants can cause
effects that are equal to, less than, or greater than that of the
individual components. Unfortunately, little is known about
the effects on the dose-response relationship of contaminants
when an individual is exposed to more than one contaminant
at a time.
Exposure Assessment
In this step in the risk assessment process, one determines the
magnitude, frequency, and duration of exposure to the patho-
gens of interest. See Chapter 3.5.2 for a complete description
of the exposure assessment process.
Human Health Effects and Dose-Response
Assessment
The dose-response assessment is “the process of characterizing
the relation between the dose of an agent administered or
received and the incidence of an adverse health effect in
the exposed populations and estimating the incidence of
the effect as a function of human exposure to the agent”
(1). The dose-response relationship is typically characterized
using a mathematical model and is a component of the overall
human health effects assessment. Dose-response modeling is
discussed in detail in Chapter 3.5.3. This section focuses on
the human health effects that result from exposure to patho-
genic microorganisms.
Infection
When exposed to pathogenic microorganisms, infection may
occur if the target organ comes into contact with a sufficient
number of organisms, the microorganisms have specific viru-
lence factors, the virulence factors are expressed, and the
defense mechanisms of the host are overcome. In many cases,
infections are asymptomatic, thus the infected individual
does not know that they are infected; this can also be called
a subclinical infection. Asymptomatic infections have
 3.5.1. Risk Assessment Framework ▪ 3.5.1-5

important implications for secondary transmission of disease,
discussed later.
Theoretically, a single organism is all that is required to
cause disease (i.e., there is no threshold infectious dose, as
there is for some chemicals). However, the infectious dose
for microorganisms is typically represented as the ID50—the
dose required to produce an infection in 50% of the exposed
individuals. Determination of ID50 is typically based on
feeding studies, sometimes using human volunteers. In these
studies, groups of individuals are fed different numbers of
organisms, and the numbers of individuals who develop
infection is monitored. The data can then be analyzed to
calculate the dose required to produce infection in 50% of
the individuals. Most microbial dose-response relationships
can be modeled using the exponential or the beta-Poisson
model (5).
Because many dose-response relationships for human
pathogens have been conducted using human volunteers
(see, e.g., [6, 7]) rather than animals, extensive adjustments
to the outcome are not made, unlike the case for chemicals.
For example, health outcomes from microorganisms are typi-
cally seen after a single exposure, so the results are directly
used. In the case of many chemicals, studies must be done
using multiple (typically unrealistically) high-dose exposures
to see measurable effects that can be clearly attributed to the
chemical. The results must then be extrapolated to more real-
istic exposure doses, introducing uncertainty into the result-
ing dose-response model.
However, the number of individuals who can participate
in dose-response determination studies is necessarily very
small (fewer than 100), due to the high cost. In addition,
these studies are typically conducted with healthy individuals
who are not elderly or very young. This raises concerns about
the representativeness of the results when trying to assess the
health impacts for more sensitive subpopulations, such as
individuals who are immune-compromised, elderly, or very
young. As it is known that the dose of an organism required
to produce infection can vary based on the age and health
of the exposed individual, their genetic makeup, as well as
whether they have been exposed to that microorganism pre-
viously, this does introduce uncertainty into the results.
To try to reduce these uncertainties, modifications to the
traditional quantitative microbial risk assessment models
have been made to address some of these data limitations.
For example, Parkin et al. (8) developed a more detailed
model to allow for the incorporation of differences in expo-
sure responses of sensitive subpopulations, rather than assum-
ing that all individuals have the same response to the same
dose of a contaminant.
In some cases, dose-response studies for pathogenic micro-
organism were conducted with animals, rather than humans.
It has generally been found that the ingested dose in animals
from a single exposure presents the same risk as ingesting the
same dose in humans; thus, there is not a need for interspecies
correction (2). For example, Armstrong and Haas (9) found
this to hold true for Legionella, Haas et al. (10) for E. coli
O157:H7, and Rose et al. (11) for Giardia.
The dose-response relationship for microorganisms is typ-
ically established using the response elicited by a single expo-
sure to the pathogen. However, depending on the situation,
an individual may be exposed on multiple occasions. Very lit-
tle is known about the health effects from multiple exposures
to the same pathogen. Therefore, multiple exposures are mod-
eled as independent events (12), although they could result
in a higher (due to sensitization) or lower (due to immune sys-
tem inactivation) dose required to cause an effect. It would be
necessary to conduct feeding studies in which individuals
were exposed to the pathogens on multiple occasions to accu-
rately model the situation.
For some microorganisms, such as hepatitis A virus, it is
believed that infection confers lifelong protection from rein-
fection. For others, such as Giardia, reinfection is possible. In
the case of norovirus, it is believed that genetic factors have a
significant impact on infection (13). The ID50 varies widely
among microorganisms; selected ID50 values are shown in
Table 1.

TABLE 1 ID50 values for selected human pathogens

Microorganism Route of exposure ID50
Adenvirus Inhalation 1.14 plaque-forming units
Campylobacter Oral 890 cfu
Cryptosporidium Oral 121 oocysts
Echovirus Oral 922 plaque-forming units
Entamoeba coli Oral 341 cysts
Enterohemorrhagic E. coli 0157:H7 Oral 31,800 colony-forming units
Giardia duodenalis Oral 34.8 cysts
Influenza Intranasal 945,000
Norovirus Oral 18 virus particles
Poliovirus Oral 1.41 plaque-forming units
Rhinovirus Intranasal 1.81 plaque-forming units
Rotavirus Oral 6.17 plaque-forming units
Salmonella anatum Oral 37,100 colony-forming units
Salmonella meleagridis Oral 16,800 colony-forming units
Shigella Oral 1,480 colony-forming units
Staphylococcus aureus Subcutaneous 9,080,000 colony-forming units
Vibrio cholerae 243 colony-forming units
Sources: http://qmrawiki.canr.msu.edu/index.php?title=Completed_Dose_Response_Models; http://qmrawiki.
canr.msu.edu/index.php?title=Table_of_Recommended_Best-Fit_Parameters; 13
3.5.1-6 ▪ MICROBIAL RISK ASSESSMENT

TABLE 2 Ratio of clinical illness to subclinical TABLE 3 Case-fatality rates for selected pathogens
infection with enteric viruses (from 14) (from 17)

Frequency of Microorganism Case-fatality rate (%)

Virus
illness (%)
Coxsackievirus A2 0.5
Astrovirus (adults) 12.5 Coxsackievirus A16 0.26
Coxsackievirus A16 50 Coxsackie B viruses 0.59–0.94
Coxsackievirus B3 29–96 Echovirus 6 0.29
Coxsackievirus B5 5–40 Echovirus 9 0.27
Echovirus 9 15–60 Hepatitis A virus 0.3
Echovirus 30 50 Poliovirus 1 0.9
Hepatitis A virus (adults) 75 Rotavirus 0.01
Poliovirus 1 0.1–1 Campylobacter 0.1
Rotavirus (adults) 56–60 E. coli O157:H7 0.2
Rotavirus (children) 28 Salmonella 0.1
Shigella 0.2

Disease
Some proportion of individuals who become infected with a
pathogen go on to develop clinical signs and symptoms of that
infection; this is termed disease. Whether a particular indi-
vidual will develop disease is determined by several host fac-
tors, including age, general health status, and immunity. The
extent to which infection by a particular organism will pro-
duce disease is called pathogenicity, which is determined by
the ratio of the number of people developing clinical illness
to the number who were infected. The pathogenicity of
selected enteric viruses is shown in Table 2.
Mortality
Some individuals who are infected with a pathogen will die as
a result; this is termed mortality. The host factors that affect
mortality include age, general health status, and immune sta-
tus. It is well documented that individuals whose immune sys-
tems are weakened in some way are more susceptible not only
to disease but mortality. Gerba et al. (14) reviewed the liter-
ature to determine which groups of individuals are at highest
risk of serious illness and mortality from enteric pathogens
transmitted through food and water. They found that the
very young, the elderly, and pregnant women were at
increased risk from these pathogens. For example, they found
that in developed countries, more than half of documented
deaths from gastroenteritis and hepatitis A occur in the eld-
erly. Nursing home residents have 10 times the risk of dying
from foodborne bacterial gastroenteritis outbreaks than those
in the general population (14). For example, individuals in
nursing homes were found to have a 100 times higher case-
fatality ratio from rotavirus infections than members of the
general population (15). Ryan et al. (16) found that the case-
fatality rate of individuals in a nursing home who were
infected with E. coli O157:H7 was 11.8, compared to a case-
fatality rate of 0.2 in the general population (14). Noteworthy
is that individuals in these groups make up approximately
20% of the U.S. population; this is expected to increase as
more people live to advanced ages. The percent of individuals
with clinical disease who die as a result for selected pathogens
is shown in Table 3.
Secondary Transmission
Secondary transmission is a term that refers to the acquisition
of infection from an individual who acquired the infection
directly from exposure to the contaminated food, water, air,
and so on. This can happen in a number of ways. Not just
individuals who exhibit overt signs of infection can spread
the infection to others: an individual who has a subclinical
infection is capable of transmitting the organism to another.
For many microorganisms, the infected individual sheds the
pathogen in the feces before they know they are infected.
In the case of hepatitis A virus, for example, the virus is
present in the feces for at least one week before clinical man-
ifestations appear (18). In other cases, the pathogen is
excreted in feces for days to weeks after the individual is no
longer experiencing symptoms. In all of these situations,
because the individual is unaware that they are contagious,
they may not take the same precautions that would be taken
if they were experiencing clinical signs and symptoms. In
some disease outbreaks, more individuals are infected as a
result of secondary spread than from exposure to the contami-
nated source water. Table 4 shows the secondary attack rates
for selected enteric viruses.
Risk Characterization
In the risk characterization phase, the information obtained
from the hazard identification, exposure assessment, and
dose-response relationship assessment is combined to develop
a conclusion about risk. In many cases, there are sufficient
data to make a quantitative estimate of the risk. As stated pre-
viously, there are different outcomes, or endpoints, that may
result from exposure to a pathogen: infection, disease, and
mortality. When calculating risk, any of these endpoints
may be chosen.
Two different models have been found to characterize
microbial dose-response data: the beta-Poisson model and
the exponential model. Each microorganism has a character-
istic set of parameters that best describe its dose-response
TABLE 4 Secondary attack rates of selected
enteric viruses (from 19)
Secondary attack
Virus
rate (%)
Coxsackieviruses 76
Echoviruses 40
Hepatitis A virus 78
Noroviruses 30
Poliovirus 90
 3.5.1. Risk Assessment Framework ▪ 3.5.1-7

TABLE 5 Best-fit dose-response parameters for pathogen dosing studies

Pathogen Best fit model Optimized parameters Strain Route of exposure
Campylober jejuni and Beta-Poisson α = 1.44E−01, N50 = 8.9E + 02 A3249 Oral (in milk)
C. coli
E. coli Beta-Poisson α = 1.55E−01, N50 = 2.11E + 06 EIEC 1624 Oral (in milk)
Salmonella anatum Beta-Poisson α = 3.18E−01, N50 = 3.71E + 04 I Oral, with eggnog
Salmonella meleagridis Beta-Poisson α = 3.89E−01, N50 = 1.68E + 04 I Oral, with eggnog
Salmonella serotype Exponential k = 3.97E−06 Salmonella newport Oral
Newport
Shigella Beta-Poisson α = 2.65E−01, N50 = 1.48E + 03 2a (strain 2457 T) Oral (in milk)
Staphylococcus aureus Exponential k + 7.64E−08 Subcutaneous
Vibrio cholerae Beta-poisson α = 2.50E−01, N50 = 2.43E + 02 naba 569B Oral (with NaHCO3)
Adenovirus Exponential k + 6.07E−01 type 4 Inhalation
Echovirus Beta-Poisson α = 1.06E + 00, N50 = 9.22E + 02 12 Oral
Influenza Beta-Poisson α = 5.81E−01, N50 = 9.45E + 05 H1N1,A/California/10/78 Intranasal
attenuated strain,H3N2,
A/Washington/897/80
attenuated strain
Poliovirus Exponential k = 4.91E−01 Type 1, attenuated Oral
Rhinovirus Beta-Poisson α = 2.21E−01, N50 = 1.81E + 00 Type 39 Intranasal
Rotavirus Beta-Poisson α = 2.53E−01, N50 = 6.17E + 00 CJN Oral
Cryptosporidium parvum exponential k = 5.72E−02 TAMU isolate Oral
and C. hominis
Entamoeba coli Beta-poisson α = 1.01E−01, N50 = 3.41E + 02 infected human Oral
Giardia duodenalis Exponential k = 1.99E−02 LEE Oral
Source: http://qmrawiki.canr.msu.edu/index.php?title=Table_of_Recommended_Best-Fit_Parameters

relationship. Table 5 shows best-fit parameters for selected
pathogens; all data were obtained from studies on humans
in which infection or shedding in feces was the endpoint.
The risk associated with exposure to a pathogen whose
dose-response relationship fits the beta-Poisson model can
be calculated using the following equation:
  1=a a
ð2  1Þ
P¼1 1þN
N50
where:
P = probability of infection from a single exposure
N = number of organisms ingested in a single exposure
α, N50 = parameters characterizing the host–pathogen
relationship
For pathogens whose dose-response relationship is charac-
terized by the exponential model, the risk of infection can be
calculated using the following equation:
P ¼ 1  expðkNÞ
where:
P = probability of infection from a single exposure
N = number of organisms ingested in a single exposure
k = parameter characterizing the host–pathogen relationship
The risk of infection from multiple exposures to a patho-
gen can be calculated using the following equation:
Pd ¼ 1  ð1  PÞd
where:
P = probability of infection from a single exposure
d = number of exposures
If an endpoint other than infection is of interest, one can
calculate the risk of illness or mortality using the following
equations:
Risk of illness ¼ PI
Risk of mortality  PIM
where:
P = probability of infection from a single exposure
I = percentage of infections that result in clinical illness
M = percentage of clinical cases that result in mortality
In some situations, it is desirable to compare different
types of risk. For example, one might want to compare the
risk associated with exposures to chemicals and microorgan-
isms in drinking water. Chemical risk is usually expressed as
lifetime risk of cancer resulting from multiple exposures
over many years, while microbial risk is typically an acute
risk resulting from a single exposure. The U.S. Environmental
Protection Agency has based drinking water standards for
carcinogenic chemicals at 10−4 to 10−6 lifetime risk of
developing cancer. To balance those risks with those from
microorganisms, they set treatment requirements for microor-
ganisms using an annual probability of infection of 10−4,
which they state is approximately equivalent to a lifetime
risk of death of 10−5 (20).
A different approach has been taken by the World Health
Organization (WHO). The health effects caused by exposure
to different types of contaminants with different health
effects can more easily be compared by quantifying the bur-
den of disease morbidity and mortality through the use of
disability-adjusted life years (DALYs). Per the WHO, “One
DALY can be thought of as one lost year of ‘healthy’ life.
3.5.1-8 ▪ MICROBIAL RISK ASSESSMENT

The sum of these DALYs across the population, or the burden TABLE 6 Risk-based microbial levels for sites categorized as
of disease, can be thought of as a measurement of the gap severe, high, medium, and low exposures (from 27)
between current health status and an ideal health situation
where the entire population lives to an advanced age, free Exposure Category
of disease and disability.” DALYs are calculated by adding Pathogen
Severe High Medium Low
the number of years of life lost due to disability (for people
who are living with the adverse health effect) to the number E. coli (cfu/100 ml) 126 1,714 4,615 50,000
of years of life lost due to premature deaths across the exposed Cryptosporidium 0.001 0.016 0.033 0.32
population. DALYs therefore account for both acute and (oocysts/liter)
chronic health effects, including morbidity and mortality.
The WHO has determined that a waterborne disease bur- Notes: For each of the exposure classes (grouped by the amount of water
den of 10–6 DALYs per person per year is an acceptable risk ingested), the risk-based levels represent the contaminant concentrations cor-
responding to an annual risk of infection of 10−6.
(21). This disease burden is approximately equivalent to 1
mild diarrheal illness per 1,000 people per year, or a 1 in 10
risk of mild illness over a lifetime (22). Results for specific sce-
narios can then be evaluated in the context of acceptable risk described a method to incorporate person-to-person transmis-
targets. For example, Australia developed guidelines for pot- sion of human pathogenic microorganisms, allowing a calcu-
able reuse (23) that set a tolerable microbial risk of 10−6 lation of the amount of disease that results from secondary
DALYs per person per year (approximately 1 diarrheal illness transmission, rather than just the amount resulting from pri-
per 1,000 people per year). mary exposure to the contaminated water.
In most cases, secondary transmission is not accounted for
when conducting a quantitative risk assessment. This may
lead to a significant underestimation of the health impacts. Examples of Risk Calculations
To try to counter this, risk of infection may be used as the Formal quantitative risk assessments have been conducted for
endpoint of interest, rather than risk of disease. Soller (24) a variety of situations.

TABLE 7 Stormwater exposure assumptions (adapted from 27)

Exposure assumptions
Exposure General description of
Stormwater use Route of exposure Volume Days
classification exposure conditions
ingested (ml) ( per year)
Severe Swimming pool Ingestion of water Ingestion of water 200 9
High Fire fighting Ingestion of water and Firefighter exposure 20 50
spray
Commercial car wash Ingestion of water and Car wash operator 3 250
spray spray exposed 5 days per
week
Medium Home car wash spray Ingestion of water and Once a week car wash for 5 24
spray 6 months
Open space, municipal Ingestion child Frequent contact with 4 130
park drip or spray playground wet surfaces and
irrigation frequent
hand-to-mouth
activity
Home lawn or garden Ingestion after contact Routine indirect 1 90
spray irrigation with plants/grass ingestion via contact
with plants, lawns, etc.
Home lawn or garden Ingestion of irrigated Typical ingestion of 7 50
spray irrigation vegetables and fruit small home garden
seasonal produce
Low Washing machine use Ingestion of sprays Ingestion from 1 load per 0.01 365
day
Toilet flushing Ingestion of aerosol Flushing 3 times per day 0.01 1,100
sprays
Open space, municipal Ingestion with casual Infrequent contact with 0.01 32
park drip or spray contact—picnic, wet grass, picnic tables
irrigation walking pet
Open space, municipal Ingestion with Frequent contact with 2.5 16
park drip or spray high-intensity sports irrigated sports field
irrigation —baseball, soccer

Untreated Graywater for Toilet Flushing
The risk associated with the use of untreated graywater for toi-
let flushing was calculated in a recent report on the use of
graywater and stormwater to augment water supplies (25).
The calculations were made using data obtained from samples
of untreated graywater that were analyzed for Giardia cysts
(26). Giardia was found to be present at concentrations rang-
ing from 0.5 to 1.5 cysts per liter; to make a conservative risk
estimate, the highest concentration was used in the
calculations.
Input data:
Volume of water exposed to from toilet flushing (27):
0.01 ml
Concentration of Giardia (26): 1.5 cysts/liter
Best fit model parameter, r (5): 0.0198
Calculations:
P ¼ 1  expðrNÞ
N = number of microorganisms ingested in a single exposure
¼ 1:5 cysts=L  0:01  ml  1  L=1000 ml
¼ 1:5  105 cysts
P ¼ 1  exp ð0:0198Þð1:5E  5Þ
¼ 2:97  107
In other words, an individual has a risk of 1 in 3,367,003 of
infection by Giardia after a single exposure to graywater from
aerosols created by flushing a toilet.
The annual risk of infection was also calculated, using the
assumption that the number of exposures over the course of a
year is 1,100 (27):
P1100 ¼ 1  ð1  PÞ1100
¼ 1  ð1  2:97  107 Þ1100
¼ 3:27  104
Over the course of a year, an individual has a risk of 1 in
3,061 of infection by Giardia from using untreated graywater
to flush toilets.
Use of Stormwater
The District of Columbia Department of the Environment
(DDOE) has established a framework for applicants to fol-
low when proposing a nonpotable use of harvested storm-
water runoff to comply with site stormwater retention
regulations. They developed a tiered risk assessment method
that requires an evaluation of the public health threat for
microbial and chemical contaminants and consideration of
the potential risk of exposure and related magnitude of
human health impacts with exposure. Four classes of expo-
sure, from low to severe were developed (see Table 6).
The process allows planners to compare typical stormwater
concentrations against tiered risk-based levels, with higher
contaminant concentrations allowed for activities with low
exposure than for those with higher exposure (see Table 7
for examples of activities that could result in exposure).
When stormwater concentrations exceed these screening
levels (i.e., the risk is considered unacceptable), the risk
management process requires treatment or justification for
why treatment is not needed.
3.5.1. Risk Assessment Framework ▪ 3.5.1-9
1 in 1000
1 in 1
million
1 in 1
billion
Crypto Giardia Rotavirus
Infection Illness Mortality
FIGURE 4 Daily risk from exposure to 100 ml reclaimed water.
Data from reference 28. doi:10.1128/9781555818821.ch3.5.1.f4
Ingestion of Stored Reclaimed Water
Rose and Carnahan (28) conducted a risk assessment for a
city in Florida that performed filtration and disinfection of
its secondary-treated wastewater and then stored the water.
The concentrations of several pathogenic microorganisms,
including Giardia, Cryptosporidium, and enteroviruses, were
determined. The risks of infection, illness, and mortality asso-
ciated with accidental ingestion of 100 ml of the stored water
were calculated; these are shown in Fig. 4.
RISK MANAGEMENT
The process of calculating the risk from exposure to a contam-
inant is a scientific process. However, numerous factors must
be considered when determining whether that risk is accept-
able. These include technological feasibility, economics, pol-
itics, and societal factors. The importance of the involvement
of stakeholders throughout the entire process, from problem
formulation through the risk assessment should facilitate
the risk management process. For a more complete descrip-
tion of this process, see EPA (4) and NAS (2).
REFERENCES
1. National Research Council. 1983. Risk Assessment in the
Federal Government: Managing the Process. National Acad-
emy Press, Washington, DC.
2. National Research Council. 2009. Science and Decisions:
advancing Risk Assessment. Committee on Improving Risk Anal-
ysis Approaches Used by the U.S. EPA. National Academy
Press, Washington, DC.
3. U.S. Environmental Protection Agency. 1998. Guidelines
for Ecological Risk Assessment. Risk Assessment Forum,
EPA/630/R-95/002F. U.S. Environmental Protection
Agency, Washington, DC.
4. U.S. Environmental Protection Agency. 2014. Microbio-
logical Risk Assessment (MRA) Tools, Methods, and
Approaches for Water Media, EPA-820-R-14-009. Office
of Science and Technology, Office of Water, U.S. Environ-
mental Protection Agency, Washington, DC.
5. Haas CN, Rose JB, Gerba CP. 2014. Quantitative Microbial
Risk Assessment, 2nd ed. Wiley & Sons, Hoboken, NJ.
6. Dupont HC, Chappell C, Sterling C, Okhuysen P, Rose J,
Jakubowski W. 1995. Infectivity of Cryptosporidium parvum
in healthy volunteers. New England J Med 332:855–859.
7. Ward RL, Bernstein DL, Young EC, Sherwood JR,
Knowlton DR, Schiff GM. 1986. Human rotavirus studies
in volunteers: determination of infectious dose and serolog-
ical response to infection. J Infect Diseases 154:871–880.
3.5.1-10 ▪ MICROBIAL RISK ASSESSMENT
8. Parkin RT, Soller JA, Olivieri AW. 2003. Incorporating
susceptible subpopulations in microbial risk assessment:
pediatric exposures to enteroviruses in river water. J Exposure
Anal Environ Epidemiol 13:161–168.
9. Armstrong TW, Haas CN. 2007. A quantitative microbial
risk assessment model for Legionnaires’ disease: animal
model selection and dose-response modeling. Risk Anal
27:1581–1596.
10. Haas CN, Thayyar-Madabusi A, Rose JB, Gerba CP.
2000. Development of a dose-response relationship for
Escherichia coli O157:H7. Intl J Food Microbiol 56:153–159.
11. Rose JB, Haas CN, Regli S. 1991. Risk assessment and the
control of waterborne giardiasis. Am J Public Hlth 81:709–713.
12. Haas CN. 1996. How to average microbial densities to char-
acterize risk. Water Res 30:1036–1038.
13. Teunis PF, Moe CL, Liu P, Miller SE, Lindesmith L, Baric
RS, Le Pendu J, Calderon RL. 2008. Norwalk virus: how
infectious is it? J Med Virol 80:1468–1476.
14. Gerba CP, Rose JB, Haas CN. 1996. Sensitive populations:
who is at greatest risk? Int J Food Microbiol 30:113–123.
15. Halvorsrud J, Orstavik I. 1980. An epidemic of
rotavirus-associated gastroenteritis in nursing home for the
elderly. Scand J Infect Dis 12:161–164.
16. Ryan CA, Tauxe RV, Hosek GW, Wells JG, Stoesz PA,
McFadden HW, Smith PW, Wright GF, Blake PA. 1986.
Escherichia coli 0157:H7 diarrhea in a nursing home: clinical,
epidemiological, and pathological findings. J Infect Dis 154:
631–638.
17. Gerba CP. 2015. Risk assessment. Chapter 24. In Pepper IL,
Gerba CP, Gentry TJ (eds.), Environmental Microbiology, 3rd
ed. Academic Press, San Diego, CA.
18. Matheny SC, Kingery JE. 2012. Hepatitis A. Am Fam Physi-
cian 86:1027–1034.
19. Gerba CP, Rose JB. 1993. Estimating viral disease risk from
drinking water. In Cothern CR (ed.), Comparative Environ-
mental Risks. Lewis Publishers, Boca Raton, FL.
20. U.S. Environmental Protection Agency. 1992. Use of
microbial risk assessment in setting U.S. drinking water
standards. EPA 814/S-92-001. U.S. EPA Office of Water,
Washington, DC.
21. World Health Organization. 2004. Guidelines for Drinking-
Water Quality, Recommendations. 3rd ed, vol 1. WHO Press,
Geneva, Switzerland.
22. World Health Organization. 2008. Guidelines for Drinking-
Water Quality, Incorporating the First and Second Addenda,
Recommendations, 3rd ed. vol 1. WHO Press, Geneva,
Switzerland.
23. Natural Resource Management Ministerial Council,
Environment Protection and Heritage Council, Austral-
ian Health Ministers’ Conference. 2006. National
Guidelines for Water Recycling: Managing Health and
Environmental Risks. A publication of the Environment
Protection and Heritage Council, the Natural Resource
Management Ministerial Council and the Australian
Health Ministers’ Conference. Editing by Biotext Pty Ltd,
Canberra.
24. Soller JA. 2009. The potential implications of person-to-
person transmission of viral infection for US EPA’s Ground-
water Rule. J Water Health 7:208–223.
25. National Academies of Sciences, Engineering, and Medi-
cine. 2015. Using Graywater and Stormwater to Enhance Local
Water Supplies: An Assessment of Risks, Costs, and Benefits.
National Academies Press, Washington, DC.
26. Birks R, Hills S. 2007. Characterization of indicator organ-
isms and pathogens in domestic greywater for recycling.
Environmental Monitoring and Assessment 129:61–69.
doi:10.1007/s10661-006-9427-y.
27. District Department of Environment. 2013. Stormwater
Management Guidebook. DDOE, Washington, DC.
28. Rose JB, Carnahan RP. 1992. Pathogen removal by full-scale
wastewater treatment. Department of Environmental Regula-
tion, Tampa, FL.
 Exposure Assessment
SUSAN R. PETTERSON AND NICHOLAS J. ASHBOLT
3.5.2
Exposure assessment within microbial risk assessment is the
evaluation of the magnitude and frequency of exposure to
pathogenic microorganisms via the specified exposure path-
way(s). For ingestion and inhalation pathways which are
the focus of this chapter, the magnitude of exposure is most
commonly defined by the mean dose per exposure event
and is quantified as:
Exposure dose ¼ C:q
where:
C is the concentration of pathogens in the exposure medium
q is the amount of material ingested or inhaled per event
The units of C and q depend on the exposure pathway
and may be based on mass for ingestion of soil or food
(mg, g); volume for water consumption (ml, liters), or inhala-
tion (m3) per exposure.
The mean dose is a variable that may take any value ≥0,
however, microorganisms are discrete units and the number
of organisms to which an individual is exposed will always
be a discrete number (i.e., 0, 1, 2, 3…). The most commonly
used dose-response models applied within QMRA relate
the probability of infection to the mean dose (relying on a
Poisson distribution to describe probability of exposure to 0,
1, 2, 3, organisms), and therefore exposure is usually charac-
terized in terms of the mean dose (see chapter 3.5.3). To
quantify the expected number of target organisms to which
a population is exposed, it is necessary to assume a distri-
bution to describe the dispersion of discrete pathogen units
within the exposure medium (e.g., Poisson, negative
binomial).
The overall microbial risk associated with a particular
exposure pathway is not only related to the magnitude of
exposure but also how often (frequency) that exposure is
expected to occur. Within MRA, the probability of infection
associated with different or subsequent exposure events is typ-
ically assumed to be independent. The frequency of exposure
is applied within risk characterization to quantify the overall
probability of one or more infections for a given time interval
(e.g. per year), hence a higher level of exposure may be
tolerated for very rare exposure events, in comparison to daily
exposure pathways. Exposure frequency can vary from daily
(such as unboiled tap water [exposure within 24 h is often
assumed to comprise a single event]) to a few days per week
(e.g., certain food crops) to weekly (e.g., household garden
doi:10.1128/9781555818821.ch3.5.2
3.5.2-1
watering with reclaimed effluent) or even sporadic (e.g., emp-
tying a septic tank).
Following the methodology recommended by WHO (1),
exposure assessment involves the following steps:
1. Define the exposure pathway(s) in terms of pathogen
sources, reduction (or recontamination) barriers, and
mechanisms of exposure, generally as a set of exposure
scenarios.
2. Quantify each model variable using the best available
scientific evidence and an understanding of the expected
variability and uncertainty associated with model
variables.
3. Characterize the exposure by quantifying the magnitude
and frequency of exposure for the range of scenarios to be
considered in the MRA.
DEFINING THE EXPOSURE PATHWAY
Scientifically and statistically, quantifying the concentration
of pathogens (C) in the exposure medium is the greatest
challenge due to limitations with pathogen/indicator enu-
meration methods (typically leading to small and uncertain
data sets) along with the often extreme fluctuations in patho-
gen concentrations (e.g., [2, 3, 4]). In many cases it is not pos-
sible or desirable to quantify the pathogen concentration
directly near the zone of exposure; instead, the pathogen con-
centration is quantified at a point closer to the pathogen
source in the exposure pathway, and the fate and transport
to the point of exposure is subsequently modeled (Fig. 1).
Table 1 includes examples of different approaches for defining
the exposure pathway. Any exposure pathway can be defined
in terms of:
Sources: Environmental media where the pathogen con-
centration can be quantified. For example for MRA of
drinking water the source could be the treated drinking
water (5), untreated source water (6), or even upstream
fecal sources (7, 8).
Barriers or Controls: Environmental (e.g., residence time,
sunlight, overland transport) (9, 10), engineered (drink-
ing water or wastewater treatment barriers) (11), or reg-
ulatory (e.g., withholding periods for crops) barriers or
controls that are expected to lead to a loss or inactivation
of pathogens (12). It is also important to note that
3.5.2-2 ▪ MICROBIAL RISK ASSESSMENT
FIGURE 1 Conceptual components for quantifying exposure. doi:10.1128/9781555818821.ch3.5.2.f1
recontamination from a secondary source (e.g., intrusion
of groundwater to distribution networks, waterfowl on
open ponds, human handling of food) may also occur.
Mechanisms of exposure: The pathway by which human
exposure may result, which depending on the context
may include intentional drinking (unboiled tap water)
(13), unintentional ingestion (e.g., during recreational
swimming) (14), aerosol ingestion (15), and food con-
sumption (16).
Modeling pathogen reduction across barriers (rather than
quantifying pathogen concentration directly in the expo-
sure medium) may be preferred due to the limitations of enu-
meration data or to meet the purpose of the risk assessment
(e.g., to inform barrier or control management).
Limitations with Enumeration Data
• Pathogen enumeration data may be unavailable for the
exposure medium or it may not be possible to measure
the concentration of pathogens in the exposure medium
due to either methodological or sampling limitations.
• Pathogen concentration may be too low in the exposure
media to obtain a consistent detectable concentration:
in the case of drinking water, pathogen concentrations
may be well below the limit of detection and yet still
pose a risk to public health (17, 18). Microbial risks asso-
ciated with drinking water exposures are therefore often
assessed by quantifying the pathogen concentration in
raw water, and modeling reduction across treatment
barriers (19).
• Pathogen enumeration data may be too limited (small
sample size) to describe the full range of fluctuations in
the concentration. For example, pathogen monitoring
data from surface water sources may exhibit a high pro-
portion of zeros even when the water sources are known
to be impacted by fecal sources. This is often due to the
event-driven nature of microbial loading and the
limitations of small monitoring data sets to capture these
events. Modeling the pathogen concentration in fecal
sources, followed by hydrologic modeling of contamina-
tion events, may therefore provide more useful informa-
tion for MRA than relying on monitoring data alone.
Purpose of the Risk Assessment
When the purpose of the assessment is to investigate processes
and drivers associated with microbial risks, it is necessary to
systematically model those environmental processes. Due
to the inherent uncertainties associated with quantifying
exposure, the greatest value of MRA may not be in the final
quantification of risk, but in the exploration of system varia-
bles and risk drivers as a tool to support water safety mana-
gement. In this case, the modeling of pathogen sources and
drivers is of critical value to the risk assessment.
Defining an exposure pathway that fits with the available
data and the purpose of the MRA is critical, and involves the
following steps:
1. Describe the exposure pathway, identifying the sources of
pathogens and how they are transported from the source
to the receptor.
2. Simplify the exposure pathway in the light of the avail-
able data and the purpose of the risk assessment. What
scientific data are available? How would I expect the
pathogen concentration to fluctuate throughout the
system? Where is it possible and reasonable to measure
the pathogen concentration? What factors and environ-
mental processes do I want to consider within the risk
assessment?
3. Define a conceptual model that simplifies the expo-
sure pathway as much as possible to allow quantification
and achieve the goal of the risk assessment. Define the
scope of the exposure assessment by identifying the
sources and processes/barriers that will be considered.
TABLE 1 Examples of different exposure pathways (sources and barriers) for the same exposure scenario (adapted from (1)).
Possible exposure pathways for quantifying C
Exposure Scenario Purpose of the risk
Source Barriers/controls Data considerations
assessment
Consumption of Option 1 Finished Distribution • Concentration is often too low • Evaluate the level of risk
unboiled tap drinking network for detection to the consumer
water from a water • Exposure is often linked to • Evaluate if additional
conventionally fluctuations in source water treatment is required
treated quality, or treatment failures,
municipal supply which is difficult to
characterize with small data
sets
Option 2 Untreated raw Conventional Disinfection Distribution • Concentration is higher in • Evaluate the level of risk to
water treatment network surface water and hence more the consumer
detects • Evaluate if additional
• Fluctuations in surface water treatment is required
concentration may mean that • Investigate the robustness
capturing the full range of of treatment barriers, and
variability may not be possible the impact of failure events
Option 3 Fecal sources Dispersion and Conventional Disinfection Distribution • Fluctuations in pathogen • Identification of the
advection treatment network concentration in the fecal influence of pathogen
sources can be characterized loading events and the
• Pathogen loading events can importance of the
be modeled epidemiology of incidence
• Modeling the combined of infection on surface
influence of different human water quality
and nonhuman sources is a
challenge

3.5.2. Exposure Assessment ▪ 3.5.2-3

Wastewater Option 1 Food crop at • Fluctuations in concentration • Evaluate the level of risk
irrigation of consumption are difficult to detect with to the consumer associated
food crops small sample sizes with the food crop
eaten raw
Option 2 Food crop at Processing and • Quantification (and the need • Evaluation of the level of
harvest handling for treatment/processing) of risk to the consumer
highly contaminated crops is associated with the
possible food crop
• Separation of the influence
of human handling on
potential risk to consumer
(Continued on next page)
TABLE 1 Examples of different exposure pathways (sources and barriers) for the same exposure scenario (adapted from (1)). (Continued )

3.5.2-4 ▪ MICROBIAL RISK ASSESSMENT

Exposure Scenario Possible exposure pathways for quantifying C
Source Barriers/controls Data considerations Purpose of the risk
assessment
Option 3 Irrigation water Contamination Inactivation in Processing • Higher and more consistent • Evaluation of the level of
during the field and handling concentrations of pathogens in risk to the consumer
irrigation wastewater can be quantified • Separation of the influence
of human handling
• Evaluation of irrigation
practices (water quality and
method of application)
on exposure can be
investigated
• Expected inactivation
under different climatic
conditions, and regulatory
crop withholding periods
can be quantified

Swimming in a Option 1 Lake water • Fluctuations in lake water • Evaluate the level of risk
freshwater lake concentration may mean that associated with swimming
capturing the full range of in a surface water
variability may not be possible • Difficult to separate the
influence of other
swimmers on water quality

Option 2 Faecal sources Dispersion and Inactivation • Higher and more consistent
advection concentration of pathogens in
fecal sources can be quantified
• Modeling the combined
influence of different human
and nonhuman sources is a
challenge
• Identify and prioritize
high-risk events
• Relate measured fecal
indicator concentration to
pathogen concentrations

QUANTIFY UNCERTAIN EXPOSURE
VARIABLES
Each component of the exposure pathway must be quantified
to assess the exposure for MRA. This quantification is com-
plicated by both variability inherent to a system and our
uncertainty in describing it.
Variability is inherent to a system and refers to the true
changes in the value of environmental variables over
space and time. Variability complicates quantifying the
model parameters from scientific observations, since on
each occasion that the variable is observed or measured,
the true underlying value may be different. Variability
cannot be reduced with increased information, but can
be better characterized or described. When the value
varies significantly (such as over several orders of magni-
tude as is the case for pathogen concentration in sources)
then the ability to quantify that variability from small
datasets is limited (20). Consideration of variability in
exposure is essential for MRA because risks fluctuate
and are often driven by environmental and anthro-
pogenic events.
Uncertainty refers to the lack of knowledge about the system
and the value of the model parameters (e.g., how repre-
sentative are pathogen samples, what is the pathogen
method recovery and fraction still potentially infec-
tious?). Uncertainty (lack of knowledge) is reduced
with increased data collection and improved under-
standing of the system. Consideration of uncertainty
in exposure is necessary as MRA outcomes are largely
dependent upon the robustness of the exposure
assessment.
Microbial data sets contain unique characteristics that
influence how the data can be used for statistical inference.
Statistical inference is the systematic process of drawing
conclusions from data (in this case, the objective is to draw
conclusions regarding the magnitude and variability of
the pathogen concentration in the environmental media
of interest, removal efficacy of various treatment barriers
and uncertainties in these estimates). Considerations for
undertaking statistical inference based on microbial data
sets include:
Sampling: Microbial data sets are typically small for des-
cribing the range of variability associated with model
inputs. Under what conditions were the samples col-
lected? Do they represent a random sample of the varia-
ble of interest, or were all the samples collected during a
particular set of conditions such as a particular season,
dry weather, or following a particular event?
Reporting: Microbial results are usually reported as a con-
centration (MPN.100 ml−1; PFU.100 ml−1, oocysts.
liter−1), however, these concentration were not directly
observed but estimated from the observations. The actual
observations may be the counted number of organisms,
colonies, or plaques or the presence or absence of PCR
product, culture, or a cytopathic effect in a certain sam-
ple volume. Reported concentrations are then inferred
using standard operating procedures, losing the infor-
mation on method uncertainty present from the original
observations. How does that influence the precision
associated with the reported concentrations?
Data characteristics: It is often desirable to fit a continuous
distribution (normal, log-normal) to a set of microbial
data, however, the data are rarely continuous but may
3.5.2. Exposure Assessment ▪ 3.5.2-5
include discrete (count data often including many
zeros), categorical (from MPN tables), or presence/
absence data types, as well as data with left (less than)
and right (greater than or too numerous to count) cen-
sored outcomes. There is no “correction” for these
types of results; however, substitutions are often made
to enable analyses by simplistic approaches. Common
substitutions include replacing less than values or zero
counts with the detection limit or half the detection
limit, and replacing greater than values with the upper
limit. Whatever approach is selected, the impact of these
replacements on the overall inference will depend on the
individual data set and should be carefully considered
(e.g., [18, 21]).
Methodology: Microbial methods are imperfect, and only a
fraction of organisms present in the original sample will
persist through the methods of sample concentration/
dilution, purification, and final assay. In addition, even
for standard methods, the recovery efficiency can vary
not only between laboratories but also between samples
due to matrix and analyst effects, and considerable error
can be introduced if estimates do not take into account
(or “correct” for) the performance of the method (e.g.,
[22]). When results for quantitative controls are not
available, a pessimistic estimate of recovery should be
considered along with some estimate of the likely
human-infectious fraction.
Which statistic to use: Which statistic should be used in the
MRA—the arithmetic mean, the median, the upper
95th percentile, or the maximum? In essence the appro-
priate statistic depends on the purpose of the MRA.
Some examples of applications of different statistics are
included in Table 2.
For many exposure scenarios, the microbial risks of con-
cern are event-driven (i.e., hazardous events). For example,
loading of pathogens to surface (source) waters may be
dependent on rainfall: overland flow, combined sewer over-
flows, and sewage treatment plant overflows all contribute
to increased pathogen mobilization during rainfall events
(23, 24). At the same time, performance of lakes or reservoirs
for reduction of pathogens can be compromised during high
flow (due to reduced retention time [25]), and treatment effi-
cacy can be reduced by physical and chemical changes in
water quality. Indeed, for well-engineered drinking water
systems, the cause of waterborne outbreaks seem somewhat
equally distributed between source, treatment, and distribu-
tion system failure events (26). Therefore, representative
exposure assessment requires the consideration of both nomi-
nal (or baseline) system behavior and the occurrence (fre-
quency and consequence) of hazardous events that may lead
to fluctuations in exposure.
The following sections outline the key factors driving the
variability associated with the value of each model variable,
and considerations and uncertainties associated with quanti-
fying the variability based on scientific data.
Quantifying Pathogen Concentration in
Environmental Media
Pathogen densities in environmental media (feces, waste-
water, surface waters, and food crops) vary between location
and over time. This variability is driven by many factors high-
lighted for a range of different media in the following sections.
Fecal samples: It is assumed that only infected (or colon-
ized) individuals will excrete pathogens, and therefore
3.5.2-6 ▪ MICROBIAL RISK ASSESSMENT

TABLE 2 Statistics for describing model variables for exposure assessment

Context of the MRA Statistics for describing exposure variables Comments
Evaluate the long-term Mean Average of all values or the The mean is weighted by all outcomes
average risk associated “expected value” including extremely high or low values that
with a particular pathway may lead to peak risk conditions. The
influence of short-term fluctuations is
averaged over the long term, which has
advantages and disadvantages for the
interpretation of risk.
Evaluate the most likely Mode The most likely value to occur Rather than be weighted by peak risk
level of risk Median 50th percentile; equal conditions, the mode and median (not
number of outcomes below technically the most likely, but the middle
and above value) give an indication of the central
location of the risk distribution.
Conservative: given a Upper 95th The value that will be exceeded The upper 95th percentile can be a useful
variable inflow water percentile by 5% of outcomes conservative point estimate.
quality, what level of
treatment is required?
Describe the distribution of Probability density Typically a parametric Rather than rely on a point estimate to
risk (range of values with function (PDF) distribution fitted to available describe a single characteristic of the
associated probability of data (e.g., log-normal, distribution, the full distribution of risk can
occurrence) gamma, beta distribution) be informative to identify the shape of the
distribution including how high extreme
values are and how frequently they occur.
Scenario analysis: evaluate Maximum Boundary value of the variable The maximum risk conditions are useful for
how high the risk could be (concentration) (maximum or minimum exploring the worst case within a scenario
for a particular pathway Minimum (barrier possible) or selected by expert analysis.
(worst case) reduction) opinion based on review of
available data

the concentration of pathogens in fecal samples is
strongly dependent on the epidemiology of infection in
the contributing population; including the incidence
(number of cases of infection or illness/population/
year) or point prevalence ( proportion of population
infected at a specific point in time) of infection; the
size of the population contributing to the fecal sample
(fecal samples representing a homogeneous mix from a
large population will exhibit a more stable pathogen
concentration in comparison to a very small popula-
tion); and the excretion density of pathogens in the feces
(density of pathogens in the feces of an infected indi-
vidual will vary between cases [usually indicative of
severity of infection or illness], and over the course of
an infection). Given the importance of these factors,
an outbreak of disease in a small community or a single
household will lead to an episode of particularly high
pathogen concentration in the fecal material from that
population compared with a more consistent but lower
density in sewage from a large population.
Untreated wastewater: In addition to the factors affecting
fecal samples, the pathogen concentration will vary
due to the magnitude of dilution by household water
usage (which will vary between sites) and additional
flows to sewer from industry or event-driven stormwater
inflows. Wastewater pathogen densities will also vary
over the day, typically exhibiting a peak in the morning,
but dependent on the sampling location in relation to
this “plug-flow” of higher pathogen density.
Treated wastewater: In addition to the factors affecting con-
centration in raw wastewater, the treated wastewater will
be affected by variability in the performance of waste-
water treatment barriers. Events of treatment failure or
process bypass will lead to a peak in pathogen concentra-
tion in treatment wastewater.
Surface waters: Pathogen concentration in surface waters
is affected by the combined influence of all human
and nonhuman fecal sources. These sources may be
point sources (e.g., wastewater release) or diffuse (e.g.,
agricultural land) and may deliver constant (deliberate
discharge) or event driven (rainfall-induced runoff,
treatment barrier failure) loadings. In addition to the
characteristics and location of fecal sources, the patho-
gen concentration will vary due to climatic factors influ-
encing pathogen persistence both between sites and
seasonally for an individual site.
Food crops: The magnitude and variability in pathogen
contamination of food crops depends on the pathogen
sources. When the source of contamination is the irriga-
tion water, then the variability discussed for treated
wastewater and surface waters (combined with climatic
factors influencing pathogen persistence on the food
crops) are relevant. When the source is individual food
handlers, the variability discussed for fecal samples are
relevant.
For assessing the magnitude and variability of pathogen
concentration for any environmental media it is important
to review the sources of pathogens and the events driving
the magnitude and variability in concentration by:
1. Identifying pathogen sources to the environmental
media. Where do the pathogens come from?

2. Identifying the factors driving the pathogen concen-
tration in the sources (including events). What is the
expected concentration of pathogens within each source?
3. Identify factors influencing the fate and transport of
pathogens (including events). How do the pathogens
travel from the source to the point of quantification?
4. Compare of the study site with published information in
the literature to develop an expectation regarding the
magnitude and variability associated with the pathogen
concentration
Using these steps, the monitoring data can be meaning-
fully interpreted with direct reference to the individual study
site. Microbial monitoring data may include fecal indicator
organisms or directly for pathogens.
Fecal Indicator Organisms
Fecal indicator organisms (FIOs), such as Escherichia coli and
enterococci, are generally not human pathogens, but are
selected for enumeration due to their ease of analysis (relative
to pathogens) and their ability to represent the potential pres-
ence of fecal pathogens in environmental media (27). The
relationship between fecal pathogens and FIOs depends on:
• How specifically fecal in origin the particular indicator is,
• How specifically human it is, and
• How environmentally persistent it is in comparison with
human pathogens.
Interpretation of indicator data therefore requires an
understanding of the fecal sources in a sample and how the
indicator may behave in relationship to the human pathogens
of interest. Traditionally, FIOs have only been recommended
to indicate potential fecal contamination (from any fecal
source), so considerable care is needed to interpret where
and when FIO may have value to estimate pathogen presence.
As a consequence, a range of more source-specific fecal indi-
cators have been and continue to be developed, under the
general topic of microbial source tracking (see Section 3.4)
and used to aid in MRA (e.g., 28).
FIO data sets are typically larger than pathogen data sets
(in many situations, the only available microbiological data
for characterization of local water quality will be for FIOs)
and the precision for enumeration is greater than for patho-
gens. FIO data may therefore provide important information
regarding the magnitude and fluctuations in fecal contamina-
tion. Even when pathogen data are available, analysis of FIO
data alongside the pathogen enumeration data may provide
stronger evidence for the characterization of pathogen con-
centration for MRA (e.g., 29).
Pathogen data
In MRA the objective is to interpret pathogen data so as to
gain a quantitative prediction of the viable human infectious
pathogen density in the environmental media of concern.
This interpretation is dependent on the methods associated
with the processing, isolation, and identification of pathogen.
Factors for consideration include:
What characteristic of the target organism was identified?
The characteristic of the organism may be specific mor-
phological features (visual identification under the
microscope), ability to reproduce under a set of specific
conditions (culture-based methods), or the presence of
a certain genetic sequence (molecular methods).
Did the method identify viable organisms? Visual micro-
scopic identification (e.g., for oo/cysts of Cryptosporidium
3.5.2. Exposure Assessment ▪ 3.5.2-7
and Giardia) cannot verify if the organism is human-
infectious, although feature consistent with viability
may be verified (e.g., [30]). Methods based on culture
(ability to reproduce) identify viable organisms, but
may underestimate the total concentration capable of
causing infection in a host (some organisms may be dif-
ficult to culture in the laboratory but still capable of caus-
ing an infection in humans) (31), and molecular
methods generally do not distinguish between viable
and inactivated organisms. Hence combined culture
and molecular methods are increasingly being developed
to seek the best elements of both approaches (e.g., [32]).
Did the method target organisms that are likely to be infec-
tious to humans? For microscopic and culture methods,
some form of confirmation (e.g., molecular probes) is
required to determine if the organism is of a species/strain
that is expected to be infectious to humans. With PCR
detections, virulence factors may be targeted that infer
higher likelihood of their human pathogenicity (e.g.,
33). Also, given the availability of cost-effective partial
or whole genome metagenomic sequencing, there is
now a rapid increase in knowledge of genogroups more
likely to represent human pathogens (e.g., [34, 35]).
Consideration of each of these factors will influence the
extent to which pathogen data is interpreted to represent
human infectious, viable pathogens in the environmental
media examined.
Barrier Performance
The concentration of pathogens across an environmental,
engineered, or regulatory controls is reduced due to physical
removal and/or inactivation. Pathogen reduction across a bar-
rier is most typically quantified on a log scale:
Log10 reduction ¼ Log10 ½Concbefore   Log10 ½Concafter 
When an individual pathogen is assumed to have a bino-
mial probability of either passing or being removed/inacti-
vated across a barrier (36), the relationship between the
probability of passage ( p) and the expected log reduction
for the population is given by:
p ¼ 10Log10 reduction
Barrier efficacy will vary both spatially (between sites)
and temporally (over time between sampling occasions).
While different types of barriers are grouped together
due to their similarity: for example “conventional drinking
water treatment,” “activated sludge,” “UV disinfection,” or
“waste stabilization ponds,” there are often considerable dif-
ferences between study sites, including scale ( pilot/full-scale;
capacity), details of design and operation (geometry, media,
loading rates), water quality characteristics (organic content,
turbidity, pretreatment, temperature), and climate (tempera-
ture, solar radiation, precipitation), that will affect the
microbial reduction performance. While it may be desirable
to generically characterize barriers, variability within catego-
ries is important to capture along with measurement uncer-
tainties (11).
Since barrier performance is dependent on climatic
factors, hydraulic loading, and water quality characteristics,
fluctuations in performance of any barrier are driven by sea-
sonal or event (e.g., rainfall, upstream loading) changes. In
addition, depending on the barrier, performance can also
be affected by system management decisions or failures
(e.g., backwash frequency, chemical dosage rates) which
may also be expected to lead to temporal variations in
3.5.2-8 ▪ MICROBIAL RISK ASSESSMENT
performance. So sampling considerations need to also
account for the sociotechnical system interactions (17).
When assessing the magnitude and variability of pathogen
reduction across a barrier, important considerations include:
1. Identifying the mechanisms of pathogen removal. Why
does the barrier result in a reduction in pathogen concen-
tration (e.g., physical removal, inactivation)?
2. Identify the factors that drive the effectiveness of the
removal mechanism (e.g., for chemical disinfection,
inactivation is dependent on achieving an adequate
concentration-time exposure).
3. Identify the circumstances under which dependent
factors might fluctuate, including short-term ( process
fluctuations and failures, rainfall) and medium-term
(season) events.
4. Compare the study site with published information.
There are many published reviews on barrier perform-
ance under various conditions to compare with site-
specific data.
Quantifying barrier efficacy based on scientific observa-
tions is undertaken using one of two approaches: empirical
or mechanistic.
Empirical: An empirical approach is data driven and relies
on a direct comparison between measurements before
and after the barrier.
The majority of published data in the literature for quan-
tifying the reduction of pathogens and surrogates across engi-
neered treatment barriers rely on an empirical approach;
samples are collected before and after a barrier and the con-
centration is compared on a log scale. The great value of
this approach is that for a given treatment plant, the evalua-
tion of removal efficacy is based on site-specific observations.
The quantified reduction efficacy is only representative for
the conditions during the experiment, and therefore it is
difficult to translate those results to account for behavior dur-
ing different environmental conditions (during an event or
during a different season) or to a different plant or location
with differing hydrodynamics and other conditions.
Data sets are rarely collected for human pathogens; rather,
surrogate organisms (native or spiked) are used to mimic the
behavior of human pathogens. It is important that surrogates
are selected that represent pathogens across the three micro-
bial groups (e.g., E. coli alone is not sufficient to model human
enteric viruses and parasitic protozoa) and address important
factors related to the mechanism of removal (e.g., environ-
mental persistence, size, surface charge).
The variation and relatively low densities of native target
microbes can be overcome by spiking trials or challenge test-
ing with high concentrations of model organisms upstream of
the inflow to a barrier; the concentration is then monitored at
the inflow and outflow of the barrier (e.g., 37). The great
value of spiking trials is that a consistent, high concentration
can be applied, providing for a high log reduction estimate as
well as reducing the influence of concentration fluctuations
on the quantified reduction.
Mechanistic: A mechanistic approach quantifies reduction
based on the assumed mechanism of removal.
When a mechanistic approach is applied, the mechanism
of removal or inactivation/growth of pathogens is explored
and quantified, usually in the laboratory or controlled experi-
ments, and then extrapolated to the full-scale barrier.
The most common application of mechanistic reduction
models in MRA are for the inactivation/growth of microbes
under a range of environmental conditions and in the pres-
ence of a disinfectant. Using this approach the persistence
(survival time) of pathogens and surrogates is explored under
controlled laboratory conditions, when the environmental
conditions relevant to the exposure scenario are known,
and then the laboratory conditions are used to estimate full-
scale inactivation (11), as advised by the U.S. EPA: http://
water.epa.gov/lawsregs/rulesregs/sdwa/mdbp/mdbptg.cfm.
Other examples of mechanistic approaches include:
• Rainfall and runoff models of water catchments for
describing the frequency and magnitude of overland
flow events that mobilize pathogens to surface waters (6).
• Hydrodynamic models of rivers for describing the disper-
sion (dilution) and advection (transport time) of pollu-
tants between an input and a downstream point of
exposure or offtake location (7).
• Hydrodynamic models of reservoirs for describing the var-
iability (in particular potential for short-circuiting) in res-
idence time under the range of inflow (rainfall and runoff )
and stratification conditions (38).
• Hydraulic models of water and wastewater treatment facili-
ties for describing variable residence time within sedimen-
tation basins, storage ponds, and contact chambers (39).
• Hydraulic models of distribution systems for describing the
impact of low-pressure ingress events and the fate and
transport of introduced pathogens (40).
• Hydrogeologic models for describing the fate and trans-
port of waterborne pathogens in the subsurface (41, 42).
Ingestion and Inhalation
The amount of material that is expected to be consumed, and
the frequency of exposure for a specific exposure pathway may
be quantified either using order of magnitude reference values
(a value is selected that is expected to represent the approxi-
mate level of exposure, e.g., 1 ml, 1,000 liters) or directly from
scientific evidence (observations regarding the magnitude
and/or frequency of exposure for a particular population are
used to either quantify a point value or distribution). The
selected approach depends on the availability of scientific
data and the purpose of the risk assessment and may involve
a combination of the two (e.g., Table 3).
NUMERICAL EXAMPLES
Norovirus Exposure via Treated Drinking
Water (Japan)
Masago and coworkers (48) undertook a quantitative risk
assessment of norovirus ingestion from treated drinking water
sourced from a river. The simple exposure pathway is illus-
trated in Fig. 2; The concentration of noroviruses was quan-
tified directly in the treated drinking water and combined
with unboiled water consumption data to assess the exposure.
The driver of the selected exposure pathway was the availabil-
ity of norovirus enumeration data from treated drinking
waters and the desire to assess the public health implications
of the detected levels.
Noroviruses had been enumerated from 98 tap water
samples (arithmetic mean of volume = 303 liters) and gen-
ogroups G1 or G2 were detected in 10 samples using RT-
PCR (Reverse transcription polymerase chain reaction).
The presence/absence data were combined with an
 3.5.2. Exposure Assessment ▪ 3.5.2-9

TABLE 3 Examples of exposure magnitude and frequency assumptions for microbial risk assessment: combination of scientific data and
reference values (from Australian Guidelines for Water Recycling: Table 3.3 Intended uses and associated exposure for recycled water [43])

Route of Volume Frequency/

Activity Comments
exposure (ml) person-year
Garden irrigation Ingestion of 0.1 90 Garden watering estimated to typically occur every second
sprays day during dry months (half year); exposure to aerosols
occurs during watering
Garden irrigation Routine 1 90 Routine exposure results from indirect ingestion via contact
ingestion with plants, lawns, etc.
Accidental 100 1 Infrequent event
ingestion
Municipal irrigation Ingestion 1 50 Frequencies moderate as most people use municipal areas
sparingly (estimate 1/2 to 3 weeks).
People are unlikely to be directly exposed to large amounts of
spray, and therefore exposure is from indirect ingestion via
contact with lawns, etc. Likely to be higher when used to
irrigate facilities such as sports grounds and golf courses
(estimate 1/week)
Food crop consumption Ingestion 5 (lettuce) 7 100 g of lettuce leaves hold 10.8 ml water and cucumbers 0.4
(home grown) 1 (other raw 50 ml at worst case (immediately postwatering). A serve of
produce) lettuce (40 g) might hold 5 ml of recycled water and other
produce might hold up to 1 ml per serve. Calculated
frequencies are based on ABS data.b
Food crop consumption Ingestion 5 (lettuce) 70 100 g of lettuce leaves hold 10.8 ml water and cucumbers 0.4
(commercial) 1 (other raw 140 ml at worst case (immediately postwatering).a A serve of
produce) lettuce (40 g) might hold 5 ml of recycled water and other
produce might hold up to 1 ml per serve. Calculated
frequencies are based on ABS data.c
Toilet flushing Ingestion of 0.01 1100 Frequency based on three uses of home toilet per day. Aerosol
sprays volumes are less than those produced by garden irrigation.
Washing machine use Ingestion of 0.01 100 Assumes one member of the household exposed. Calculated
sprays frequency based on ABS data.d Aerosol volumes are less
than those produced by garden irrigation (machine usually
closed during operation)
Fire fighting Ingestion of water 20 50 Median ingestion for firefighters estimated at 20 ml per fire
and sprays with a maximum number of fires fought within area served
by recycled water of 50 per year.e
Cross-connection of Ingestion 1,000/day 1/1,000 houses Total consumption is assumed to be 2 liters per day, of which
dual-reticulation 1 liter is consumed cold.f Affected individuals may
systems with consume water 365 days per year. A conservative estimate
drinking water mains of 1/1,000 houses has been considered for annual risk of
cross-connection.
ABS = Australian Bureau of Statistics.
a
Reference 44.
b
ABS data show that 12% of households grow lettuce and 35% grow some type of produce (45). They also show that Australians eat leafy vegetables 140 times
per year and eat other vegetables at a similar rate (45). Hence it can be estimated that “other produce” such as tomatoes, carrots in combination are eaten 280
times per year.Watering with recycled water is used to augment rainfall. Assuming that watering occurs for six months of the year, frequency of consumption
of lettuce irrigated with recycled water = 140 × 0.5 × 12% and frequency of consumption of other raw produce = 280 × 0.5 × 35%.
c
Using the same ABS data as in note b, frequency of consumption of lettuce irrigated with recycled water = 140 × 0.5 for lettuce and frequency of consumption
of other raw produce = 280 × 0.5.
d
ABS data show an average of 2.6 people per household (46). The amount of washing is estimated at five loads per week; therefore, the frequency = 5 × 52 ÷ 2.6.
e
Firefighting is an occupational exposure; the exposures were assessed by the Queensland Department of Emergency Services.
f
Reference 47.

assumption regarding the dispersion of viruses in the drinking
water (based on Cryptosporidium dispersion) to estimate
the maximum likelihood parameter values of the log-normal
distribution of norovirus (combined G1 and G2) concentra-
tion in tap water (µ = −8.09, σ =1.71). The unboiled water
consumption was based on a survey conducted in the Nether-
lands where daily water consumption was assumed to follow
a log-normal distribution (µ = −1.88, σ =1.12) (49). These
data were compared with a survey conducted in Japan, which
the authors argued reported similar results (50). The quanti-
fied variables and quantified exposure (by Monte Carlo sim-
ulation) are illustrated in Fig. 3.
The exposure assessment allowed for the public health
implications of the norovirus detection data to be explored.
Interpretation of molecular data is complicated for MRA.
First, the recovery of the method was not taken into consid-
eration, and therefore the concentration of noroviruses was
underestimated; however, recovery can only be incorporated
3.5.2-10 ▪ MICROBIAL RISK ASSESSMENT

FIGURE 2 Schematic of exposure pathway applied for the assessment of norovirus risk via drinking water in Japan.
doi:10.1128/9781555818821.ch3.5.2.f2

into the exposure assessment when quantitative controls
are used during enumeration (see example b). Second, the
molecular method cannot distinguish between infectious
and noninfectious organisms. The concentration is represen-
tative of all genetic material consistent with noroviruses and
is therefore conservative for risk assessment. In this case for
disinfected drinking water, the noroviruses may be reasonably
expected to have been inactivated by the disinfectant. With
an exposure pathway that includes the performance of the
disinfection barrier, the efficacy of the disinfection step could
be included in the assessment (by assessing the C.t. of the
disinfection step, and the susceptibility of norovirus to disin-
fection based on culturable surrogates) to predict the proba-
bility that a detected norovirus may have been inactivated.
Cryptosporidium Oocyst Concentration in Surface
Water (Australia)
Analytical measurements are imperfect, and only a fraction of
the organisms originally present will be detected by the
method. If the imperfect recovery is not accounted for, the
concentration will be underestimated. In the previous
example, analytical recovery was not accounted for in the

FIGURE 3 Quantified model inputs and exposure calculated for the assessment of norovirus risk via drinking water in Japan.
doi:10.1128/9781555818821.ch3.5.2.f3

FIGURE 4 Native versus spiked oocyst counts from a surface water
source in Australia. doi:10.1128/9781555818821.ch3.5.2.f4
concentration predictions for noroviruses. Here is an illustra-
tion of the implications of accounting for recovery for
Cryptosporidium.
Petterson and colleagues (22) reported method recovery
results and statistical analysis for characterizing Cryptospori-
dium and Giardia concentration for a surface water source
FIGURE 5 Maximum likelihood (solid line) and 95th percentile credible interval (dashed lines) for the gamma-distributed Cryptosporidium
oocyst concentration in the surface water source (a) without considering recovery, and (b) correcting each Cryptosporidium count using an inter-
nal ColorSeed recovery control. doi:10.1128/9781555818821.ch3.5.2.f5
3.5.2. Exposure Assessment ▪ 3.5.2-11
for QMRA. Results for Cryptosporidium are summarized
here. Analytical results from 99 assays of 10 liters of environ-
mental water samples collected over 5 years (2000–2004
inclusive) were statistically analyzed to infer the distribution
of concentration of Cryptosporidium in a surface water source
for QMRA. Samples were analyzed using U.S. EPA Method
1623 (51); in addition, each sample received an internal spike
of 100 labeled oocysts (ColorSeed®) as a quantitative recov-
ery control. A summary of the counted number of native and
spiked oocysts is illustrated in Fig. 4. There was no correlation
between the number of counted oocysts and the number of
spiked oocysts recovered.
A negative binomial distribution (a discrete distribution
that directly handles count [0, 1, 2, 3 etc.] data, and allows
for overdispersion of organisms) was fitted to the oocyst
counts (following the approach presented by [19, 49] by
constructing a suitable statistical model and then finding
maximum likelihood estimators of the model parameters.
First, the impact of recovery was ignored in quantifying the
concentration; second, each count was corrected within
the model using the internal recovery result. Uncertainty in
the model parameters was explored using an MCMC (Mar-
kov Chain Monte Carlo procedure) (52).
Figure 5 illustrates the probability density function of the
maximum likelihood gamma distribution (solid line) with
3.5.2-12 ▪ MICROBIAL RISK ASSESSMENT
95% credible interval from MCMC analysis (dashed lines)
describing Cryptosporidium concentration in the surface water
source with and without accounting for recovery. Accounting
for recovery increased the magnitude (mean increased from
0.72 to 1.84 oocysts/liter) and uncertainty (the ratio of
the upper credible interval of uncertainty on the upper
95th percentile: maximum likelihood upper 95th percentile
increased from 1.44 to 2.01) associated with the predicted
concentration.
For the reported data set, a paired internal recovery control
was available for every count; however, this amount of data is
rarely available for incorporating recovery into predicted pro-
tozoa concentration for QMRA. When a smaller number of
measurements are available, a different method is required.
Teunis and colleagues (49), in their original approach to
quantifying Cryptosporidium risks, fitted a beta distribution
to the recovery data and then incorporated recovery within
the Monte Carlo simulation as a random independent varia-
ble. Using this approach, the number of data points will
strongly influence the uncertainty associated with the final
distribution describing Cryptosporidium concentration. This
increase in uncertainty is illustrated in Fig. 6: four different
FIGURE 6 Examples of fitted beta-distributions (inserts) and predicted source water Cryptosporidium oocyst concentration PDFs, including
credible intervals, for various recovery data set sizes. Reprinted from ref 22, with permission. doi:10.1128/9781555818821.ch3.5.2.f6
hypothetical recovery data sets (N = 3, 10, 20, and 50) were
randomly sampled with replacement from the entire recovery
data set, and analyzed by fitting a beta distribution to each.
The Cryptosporidium concentration was then quantified
by Monte Carlo simulation of the uncorrected concentration
and the beta distributed recovery. For the smallest recovery
data set (n = 3), the uncertainty associated with the shape
of the beta distribution had an important effect on the
uncertainty associated with the estimated Cryptosporidium
concentration; the impact of this uncertainty decreased
with increasing data set size.
For the purposes of undertaking a QMRA, system-specific
recovery data should be collected and incorporated into the
source water concentration estimates, so that risks are not
underestimated. Where it is not logistically possible or desir-
able to obtain sample specific recovery control data, adequate
source water concentration estimates can be made from the
unpaired modeling approach utilized here—specifically,
where recovery and count data can be assumed independent
and when at least some 20–30 recovery data points (represen-
tative of the range of site conditions) are available. Particu-
larly when only fewer site-specific recovery data points are

available for analysis, the statistical uncertainty about source
water concentration estimates should be reported along with
data variability statistics. Where no system-specific recovery
data are available, it is recommended that the sensitivity of
QMRA outputs to a range of conservative estimates about
the mean value of the recovery fraction (say, for values from
0.01 to 0.1) be assessed rather than just ignoring the effects
or applying data from another site. Doing so will provide an
assessment of the impacts of method recovery on quantitative
health risk assessments and ensure better informed decisions
about the adequacy of the water supply system to supply safe
water.
Assessment of Risk of Enteric Virus Exposure via
Lettuce Crops
Sensitivity analysis is a powerful tool for assessing and devel-
oping strategies for the management of risks. In most practical
situations, it is not enough to simply characterize the risk; an
understanding of the most important data gaps and drivers of
risk is also needed. The following two examples illustrate the
use of sensitivity analysis in exposure assessment.
Petterson (53) investigated the risks associated with
enteric viruses present in irrigation water, which attach to
the lettuce during spray irrigation and can persist on the let-
tuce crop to consumption. The components of the exposure
assessment are illustrated in Fig. 7. The conceptual pathway
was favored for two reasons: first, data on the occurrence of
viruses on lettuce crops were unavailable, but enumeration
data were available on the concentration of enteric viruses
in secondary treated effluent. Second, the model would allow
for the risk drivers, risk-management options (including
irrigation practices and crop withholding periods), and the
most important sources of uncertainty in the model to be
identified.
The exposure was quantified using the equation:
Exposure ¼ C  v  f  SðtÞ  q
FIGURE 7 Schematic of exposure pathway applied for the assessment of enteric virus risk via consumption of wastewater irrigated lettuce
crops. doi:10.1128/9781555818821.ch3.5.2.f7
3.5.2. Exposure Assessment ▪ 3.5.2-13
where:
Exposure is the mean dose per exposure event
C is the concentration of viruses in the irrigation water
applied to the crop (virus/liter)
v is the volume of irrigation water applied to the crop (L/g)
f is the fraction of applied viruses that successfully attach to
the lettuce leaf
S(t) represents the fraction of viruses remaining infectious at
time (t) following irrigation
q is the quantity of crops consumed per event
The quantification of model components was undertaken
based on scientific evidence; a summary is given in Table 4. A
best estimate and extreme (reasonable worst case) estimate
for each of the model parameters was made. Two measures
of sensitivity were applied for the exposure assessment:
1. Step characteristic was applied to quantify the main
determinants of risk. The step characteristic indicates
the log reduction or increase in the number of organisms
relative to the previous step in the model and is given by:
 
Nk
SCk ¼ Log
Nk1
where Nk is the number of organisms per unit mass/
volume at step k.
2. Factor sensitivity is applied to identify the importance
of uncertainty at each step in the model and is given by
 
Nk extreme
FSk ¼ Log
Nk average
The step characteristic and factor sensitivity were cal-
culated for each component of the process (the volume of
irrigation water applied was kept constant at 1 ml/g, and con-
sumption was assumed to occur at 14 days following last
3.5.2-14 ▪ MICROBIAL RISK ASSESSMENT

TABLE 4 Best and extreme estimates of model parameters applied TABLE 5 Step characteristic and factor sensitivity calculated for
in the exposure assessment of virus exposure via wastewater irrigated the lettuce crop exposure model
lettuce crops
SC (Step FS (Factor
Process step
Model component “Best” estimate “Extreme” estimate characteristic) sensitivity)
Virus occurrence 2.6 vu.l−1a 470,000 vu.l−1b Virus occurrence 5.49
Virus attachment:c f 0.024 0.071 Virus attachment: f −1.6 0.45
Virus inactivation:d S(t) h1 = 2.5 d−1 h1 = 2.0 d−1 Virus inactivation: S(t) −6.2 2.20
Biphasic inactivation h2 = 0.5 d−1 h2 = 0.3 d−1 Consumption per event: q 0.48
S(t) = a.h1 + (1 − a) h2 a = 0.12% a = 0.96%
Consumption per event:e q 100 g 300 g
uncertainty in virus attachment (FS = 0.45) and crop con-
a
b
Californian data set of enteroviruses by cell culture used by (12, 54, 55). sumption (FS = 0.48) had a comparably smaller effect on
Upper limit reviewed by (56). the uncertainty in the calculated exposure.
c
Maximum likelihood (best) and upper 95% credible interval (extreme)
based on modeling undertaken by (53).
The simple deterministic measures of sensitivity calcu-
d
Maximum likelihood (best) and conservative 95% credible interval lated in this example are very informative for understanding
(extreme) based on modeling undertaken by Petterson et al. (57).
e
the drivers in the model in terms of the relative importance
Reference values applied by (53). of the different barriers (step characteristic) and the most
import sources of variability and uncertainty (factor sensitiv-
irrigation). The results are included in Table 5. The SC indi- ity). If additional monitoring or data gathering were to be
cates the importance of each component in the model for undertaken to inform the characterization of exposure,
driving the exposure, and hence quantifies the importance improved understanding of the magnitude of virus concentra-
of virus inactivation relative to attachment in driving the tion in irrigation waters would be much more beneficial than
numbers of infectious viruses consumed. For the current an improved understanding of virus attachment behavior or
model assumptions, virus inactivation is the most important consumer consumption patterns.
barrier for reducing exposure to infectious viruses, identifying
the importance of climatic conditions and crop withholding Assessing Pathogen Risk to Swimmers at Non–Sewage
periods for controlling risk. Impacted Recreational Beaches (United States) (58)
The factor sensitivity quantifies the influence of variabil- At recreational beaches in the United States, a sign may be
ity and/or uncertainty associated with each component of posted warning swimmers when fecal indicator bacterial con-
the model. Virus concentration in the irrigation water is centrations in the waters exceed the U.S. EPA recommended
uncertain and subject to wide fluctuations and therefore criteria. The U.S. EPA fecal indicator bacterial criteria are
was the most important driver of uncertainty in the model based on the relationship reported between fecal indicator
outcomes (FS = 5.49). The second most sensitive factor was concentration and human gastrointestinal illness at recrea-
the quantification of virus survival (FS = 2.20), while the tional beaches impacted by effluent from publically owned

FIGURE 8 Schematic of exposure pathway applied for comparing the exposure to pathogens depending on the fecal source to recreational
waters. doi:10.1128/9781555818821.ch3.5.2.f8

FIGURE 9 Predicted GI illness by reference pathogen (median,
interquartile range, 10th and 90th percentiles, minimum, and max-
imum) for adults following a single accidental ingestion of recrea-
tional water containing fresh fecal contamination at 35 cfu/100 /ml
enterococci contributed by seagulls or primary POTW effluent (T)
total GI risk, (C.j.) C. jejuni risk; (S) Salmonella risk; (C) Cryptospori-
dium risk; (G) Giardia risk; (N) Norovirus risk). Reprinted from ref
58, with permission.
doi:10.1128/9781555818821.ch3.5.2.f9
(sewage) treatment works (POTW). There is concern that
non–sewage impacted beaches may receive a different range
and magnitude of pathogens and may require an alternative
standard to be equally protective of human health. The
high cost and impracticability in undertaking many epide-
miologic studies, along with the lack of any clear relationship
between indicator and health outcome at non–POTW
impacted beaches has prompted the need to find an alterna-
tive way of estimating the conditions under which human
health may be impacted (59).
Schoen and colleagues (58) applied the QMRA frame-
work to predict and prioritize the pathogen risk from nonsew-
age and sewage sources for a water body at the U.S. EPA
recommended water quality limit of total enterococci equal
to 35 cfu/100 ml and to predict when a nonsewage source
of pathogens may dominate the illness risk in a water body
impacted by a mixture of sources.
FIGURE 10 Spearman rank correlation coefficient for exposure parameter inputs to the predicted probability of GI illness from accidental
ingestion of recreation water containing fresh fecal contamination at 35 cfu/100 ml ENT for (a) seagull feces and (b) primary POTW effluent.
Reprinted from ref 58, with permission. doi:10.1128/9781555818821.ch3.5.2.f10
3.5.2. Exposure Assessment ▪ 3.5.2-15
FIGURE 11 Parametric sensitivity analysis of the predicted prob-
ability of gastrointestinal illness (median, 10th, and 90th percen-
tiles) for adults attributable to Campylobacter jejuni from accidental
ingestion of recreational water containing fresh seagull fecal contam-
ination at 35 cfu/100 ml ENT to changes in the assumed fraction of
total C. jejuni strains from seagulls that are human infectious.
Reprinted from ref 58, with permission.
doi:10.1128/9781555818821.ch3.5.2.f11
The exposure pathway and conceptual model applied to
predicting exposure is illustrated in Fig. 8; for full details of
the quantification of model components, the reader is referred
to the original paper. The predicted illness risk by pathogen is
illustrated in Fig. 9, with the median illness risk associated
with human sewage ∼2 orders of magnitude greater than for
gulls, illustrating that a water body at the FIB recreational
water quality limit may present a different risk to swimmers
depending on the source of the fecal contamination.
Sensitivity analysis provided a useful tool for exploring the
interaction between model parameters. First, the authors
undertook a sensitivity analysis of the parameters driving
exposure using the Spearman rank correlation coefficients
of the Monte Carlo simulation. For the risk attributable to
seagulls, the volume of water ingested (V), the density of indi-
cator bacteria in feces, and the density of C. jejuni in feces
were of relatively equivalent importance. The density of
3.5.2-16 ▪ MICROBIAL RISK ASSESSMENT
FIGURE 12 Comparison of median illness risk for adults when
total ENT concentration (at 35 cfu/100 ml) is attributed to a mixture
of primary POTW effluent (sewage) and seagull feces (gulls).
Reprinted from ref 58, with permission.
doi:10.1128/9781555818821.ch3.5.2.f12
Salmonella in the feces and the fraction of human-infectious
pathogens in the feces were relatively less important (Fig. 10).
Due to the lack of data, the authors were specifically con-
cerned to investigate the sensitivity of the calculated risk to
the fraction of human-infectious pathogen strains from gulls.
The results of the parametric sensitivity analysis are illustrated
in Fig. 11. The illness risk associated with gull feces further
decreased as the assumed fraction of the total pathogen strains
from gulls that are human infectious decreases below the best
estimate of 0.2.
The model was also used to explore the relative impor-
tance of different sources when there is a mixture (Fig. 12).
Considering all possible mixtures along the x-axis, the
median predicted probability of GI illness from gulls was
greater than that from sewage only when gulls represent
more than 98% of the fecal indicator load. Therefore, the
dominant source of fecal indicator at a recreational beach
may not be the source of dominant risk. Conversely, little
fresh sewage contamination mixed with nonsewage source
(s) may dominate risk.
While the quantification of model inputs and characteri-
zation of exposure were highly uncertain in this study, the con-
struction of the QMRA model allowed for the scientific data
to be combined and evaluated within a transparent frame-
work. The factors driving the exposure could be explored,
the importance of lack of knowledge (uncertainty) could be
investigated, and the risk associated with different fecal sour-
ces could be theoretically evaluated. This study provides an
important compliment to epidemiologic studies undertaken
in recreational water environments and valuable information
for system understanding and decision making.
REFERENCES
1. WHO. 2016. Application of Quantitative Microbial Risk
Assessment for Water Safety Management. World Health
Organization, Geneva (In preparation).
2. Quintero-Betancourt W, Peele ER, Rose J. 2002. Crypto-
sporidium parvum and Cyclospora cayetanensis: a review of lab-
oratory methods for detection of these waterborne parasites.
J Microbiol Meth 49:209–224.
3. Eyles R, Niyogi D, Townsend C, Benwell G, Weinstein P.
2003. Spatial and temporal patterns of Campylobacter con-
tamination underlying public health risk in the Taieri River,
New Zealand. J Environ Qual 32:1820–1828.
4. Girones R, Ferrûs MA, Alonso JL, Rodriguez-Manzano J,
Calgua B, Corrêa Ade A, Hundesa A, Carratala A, Bofill-
Mas S. 2010. Molecular detecction of pathogens in water—
the pros and cons of molecular techniques. Water Res 44:
4325–4339.
5. Borchardt MA, Spencer SK, Kieke BA, Lambertini E, Loge
FJ. 2012. Viruses in nondisinfected drinking water from
municipal wells and community incidence of acute gastroin-
testinal illness. Environ Health Perspect 120:1272–1279.
6. Signor RS, Roser DJ, Ashbolt NJ, Ball JE. 2005. Quantify-
ing the impact of runoff events on microbiological contam-
inant concentrations entering surface drinking source
waters. J Water Health 3:453–468.
7. Ferguson CM, Croke BF, Beatson PJ, Ashbolt NJ, Deere
DA. 2007. Development of a process-based model to predict
pathogen budgets for the Sydney drinking water catchment.
J Water Health 5:187–208.
8. Ferguson C, Charles K, Deere D. 2009. Quantification of
microbial sources in drinking-water catchments. Crit Rev
Environ Sci Technol 39:1–40.
9. Muirhead RW, Collins RP, Bremer PJ. 2006. Numbers and
transported state of Escherichia coli in runoff direct from fresh
cowpats under simulated rainfall. Lett Appl Microbiol 42:
83–7.
10. Sinton L, Hall C, Braithwaite R. 2007. Sunlight inactiva-
tion of Campylobacter jejuni and Salmonella enterica, com-
pared to Escherichia coli, in seawater and river water. J
Water Health 5:357–365.
11. Hijnen WAM, Medema GJ. 2010. Elimination of Microor-
ganisms by Drinking Water Treatment Processes. A Review.
3rd ed, IWA Publishing, London.
12. Hamilton AJ, Stagnitti F, Premier R, Boland AM, Hale G.
2006. Quantitative microbial risk assessment models for
consumption of raw vegetables irrigated with reclaimed
water. Appl Environ Microbiol 72:3284–90.
13. Mons MN, van der Wielen JM, Blokker EJ, Sinclair MI,
Hulshof KF, Dangendorf F, Hunter PR, Medema GJ.
2007. Estimation of the consumption of cold tap water
for microbiological risk assessment: an overview of studies
and statistical analysis of data. J Water Health 5:151–70.
14. Dufour AP, Evans O, Behymer TD, Cantú R. 2006. Water
ingestion during swimming activities in a pool: a pilot study.
J Water Health 4:425–430.
15. Schoen ME, Ashbolt NJ. 2011. An in-premise model for
Legionella exposure during showering events. Water Res 45:
5826–5836.
16. Hara-Kudo Y, Takatori K. 2011. Contamination level and
ingestion dose of foodborne pathogens associated with infec-
tions. Epidemiol Infection 139:1505–1510.
17. Signor RS, Ashbolt NJ. 2006. Pathogen monitoring offers
questionable protection against drinking-water risks: a
QMRA (quantitative microbial risk analysis) approach to
assess management strategies. Water Sci Technol 54:
261–268.
18. Smeets PW, van Dijk JC, Stanfield G, Rietveld LC,
Medema GJ. 2007. How can the UK statutory Cryptospori-
dium monitoring be used for quantitative risk assessment of
Cryptosporidium in drinking water? J Water Health 5:107–18.
19. Teunis PFM, Havelaar AH. 1999. Cryptosporidium in
drinking waters: evaluation of the ILSI/RSI quantita-
tive risk assessment framework. Report No. 284 550 006.

The Netherlands, National Institute of Public Health and
the Environment (RIVM), Bilthoven, 97.
20. Englehardt JD, Ashbolt NJ, Loewenstine C, Gadzinski E,
Ayenu-Prah AY Jr. 2012. Methods for assessing long-term
mean pathogen count in drinking water and risk manage-
ment implications. J Water Health 10:197–208.
21. Jaidi K, Barbeau B, Carriere A, Desjardins R, Prevost M.
2009. Including operational data in QMRA model:
development and impact of model inputs. J Water Health
7:77–95.
22. Petterson SR, Signor RS, Ashbolt NJ. 2007. Incorporating
method recovery uncertainties in stochastic estimates of raw
water protozoan concentration for QMRA. J Water Health 5:
51–65.
23. Rose JB, Epstein PR, Lipp EK, Sherman BH, Bernard SM,
Patz JA. 2001. Climate variability and change in the United
States: potential impacts on water- and foodborne diseases
caused by microbiologic agents. Environ Health Perspect
109:211–221.
24. Risebro HL, Doria MF, Andersson Y, Medema G, Osborn
K, Schlosser O, Hunter PR. 2007. Fault tree analysis of the
causes of waterborne outbreaks. J Water Health 5:1–18.
25. Hawkins PR, Swanson P, Warnecke M, Shanker SR,
Nicholson C. 2000. Understanding the fate of Cryptospori-
dium and Giardia in storage reservoirs: a legacy of Sydney’s
water contamination incident. J Water Supp Res Technol
AQUA 496:289–306.
26. Yoder J, Roberts V, Craun GF, Hill V, Hicks LA,
Alexander NT, Radke V, Calderon RL, Hlavsa MC,
Beach MJ, Roy SL. 2008. Surveillance for waterborne dis-
ease and outbreaks associated with drinking water and water
not intended for drinking—United States, 2005–2006.
MMWR Surveill Summ 57:39–62.
27. Ashbolt NJ, Grabow WOK, Snozzi M. 2001. Indicators of
microbial water quality, p 289–315. In Fewtrell L, Bartram J.
(eds.), Water Quality: Guidelines, Standards and Health. Risk
Assessment and Management for Water-Related Infectious Dis-
ease. IWA Publishing, London.
28. Staley C, Gordon KV, Schoen ME, Harwood 2012. Per-
formance of two human-associated microbial source tracking
qPCR methods in various Florida water types and implica-
tions for microbial risk assessments. Appl Environ Microbiol
78:7317–7326.
29. Petterson SR, Dumoutier N, Loret JF, Ashbolt NJ. 2009.
Quantitative Bayesian predictions of source water concentra-
tion for QMRA from presence/absence data for E. coli O157:
H7. Water Sci Technol 59:2245–52.
30. Vesey G, Ashbolt N, Fricker E, Deere D, Williams K,
Veal D, Dorsch M. 1998. The use of ribosomal rRNA
targeted oligonucleotide probes for fluorescent labelling
viable Cryptosporidium oocysts. J Appl Microbiol 85:429–440.
31. Rodriguez RA, Gundy PM, Gerba CP. 2008. Comparison of
BGM and PLC/PRC/5 cell lines for total culturable viral assay
of treated sewage. Appl Environ Microbiol 74:2583–2587.
32. Lambertini E, Spencer SK, Bertz PD, Loge FJ, Borchardt
MA. 2010. New mathematical approaches to quantify
human infectious viruses from environmental media using
integrated cell culture-qPCR. J Virol Meth 163:244–252.
33. Bielaszewska M, Mellmann A, Zhang W, Kock R, Fruth A,
Bauwens A, Peters G, Karch H. 2011. Characterisation of
the Escherichia coli strain associated with an outbreak of hae-
molytic uraemic syndrome in Germany, 2011: a microbio-
logical study. Lancet Infect Dis 11:671–676.
34. Casas V, Maloy S. 2011. Role of bacteriophage-encoded
exotoxins in the evolution of bacterial pathogens. Future
Microbiol 6:1461–1473.
35. Ng TF, Marine R, Wang C, Simmonds P, Kapusinszky B,
Bodhidatta L, Oderinde BS, Wommack KE, Delwart E.
2012. High variety of known and new RNA and DNA
3.5.2. Exposure Assessment ▪ 3.5.2-17
viruses of diverse origins in untreated sewage. J Virol 86:
12161–12175.
36. Teunis PFM, Rutjes SA, Westrell T, Husman AMdR.
2009. Characterization of drinking water treatment for virus
risk assessment. Water Res 43:395–404.
37. Petterson SR, Mitchell VG, Davies CM, O’Connor J,
Kaucner C, Roser D, Ashbolt N. 2016. Evaluation of three
full-scale stormwater treatment systems with respect to water
yield, pathogen removal efficacy and human health risk from
faecal pathogens. The Science of the Total Environment 543:
691–702.
38. Brookes JD, Antenucci J, Hipsey M, Burch MD, Ashbolt
NJ, Ferguson C. 2004. Fate and transport of pathogens in
lakes and reservoirs. Environ Int 30:741–759.
39. Smeets PW, Dullemont YJ, Van Gelder PH, Van Dijk JC,
Medema GJ. 2008. Improved methods for modelling drink-
ing water treatment in quantitative microbial risk assess-
ment; a case study of Campylobacter reduction by filtration
and ozonation. J Water Health 6:301–14.
40. Teunis PF, Xu M, Fleming KK, Yang J, Moe CL, Lecheval-
lier MW. 2010. Enteric virus infection risk from intrusion of
sewage into a drinking water distribution network. Environ
Sci Technol 44:8561–6.
41. Schijven JF, de Bruin HAM, Hassanizadeh SM, de Roda
Husman AM. 2003. Bacteriophages and Clostridium spores
as indicator organisms for removal of pathogens by passage
through saturated dune sand. Water Res 37:2186–2194.
42. Schijven JF, Medema GJ, Vogelaar AJ, Hassanizadeh SM.
2000. Removal of bacteriophages MS2 and PRD1, spores of
Clostridium bifermentans and E. coli by deep well injection. J
Contam Hydrol 44:301–327.
43. National Water Quality Strategy 2006. National guidelines
for water recycling: managing health and environmental
risks National Resource Management Ministerial Council.
Environment Protection and Heritage Council, Australian
Health Ministers Conference, Canberra, 389p.
44. Shuval H, Lampert Y, Fattal B. 1997. Development of a risk
assessment approach for evaluating wastewater reuse stand-
ards for agriculture. Water Sci Technol 35:15–20.
45. ABS. 1995. National nutrition survey: foods eaten. Common-
wealth of Australia, Canberra.
46. ABS. 2001. Census of population housing: selected social and
housing characteristics. Commonwealth of Australia, Canberra.
47. WHO. 2004. Guidelines for Drinking-Water Quality. 3rd ed,
vol1. World Health Organization, Geneva.
48. Masago Y, Katayama H, Watanabe T, Haramoto E, Hashi-
moto A, Omura T, Hirata T, Ohgaki S. 2006. Quantitative
risk assessment of noroviruses in drinking water based on
qualitative data in Japan. Environ Sci Technol 40:7428–33.
49. Teunis PFM, Medema GJ, Kruidenier L, Havelaar AH.
1997. Assessment of the risk of infection by Cryptosporidium
or Giardia in drinking water from a surface water source.
Water Res 31:1333–1346.
50. Yano K, Hosaka M, Otaki M, Tanaka A, Lyo T, Tosa K,
Ichikawa H. 2000. The amount of the tap water consump-
tion: from the results of questionaire survey. Procedings 3rd
JSWE Symposium, 159–160, Ichikawa City, Chiba Prefec-
ture, Japan.
51. U.S. EPA. 1999. Method 1623—Cryptosporidium and Giar-
dia in water by Filtration/IMS/IFA. U.S. EPA, Office of
Water, Washington, DC.
52. Gelman A, Carlin JB, Stern HS, Rubin DB. 2003. Bayesian
Data Analysis. Chapman and Hall, Boca Raton, FL.
53. Petterson SR. 2002. Quantitative microbial risk assessment for
evaluating the viral risk from consumption of wastewater irrigated
salad crops. PhD thesis, School of Civil and Environmental
Engineering. Sydney, University of New South Wales.
54. Asano T, Leong LYC, Rigby MG, Sakaji RH. 1992. Eval-
uation of the California wastewater reclamation criteria
3.5.2-18 ▪ MICROBIAL RISK ASSESSMENT
using enteric virus monitoring data. Water Sci Technol 26:
1513–1524.
55. Petterson SR, Ashbolt NJ, Sharma A. 2001. Microbial risks
from wastewater irrigation of salad crops: a screening-level
risk assessment. Water Environ Res 73: 667–672. Errata:
Water Environ Res 74411.
56. Yates MV, Gerba CP. 1998. Microbial considerations
in wastewater reclamation and reuse, p 437–488. In: Asano
T (ed.), Wastewater Reclamation and Reuse, vol. 10. Water
Quality Management Library, Technomic Publishing,
Lancaster, PA.
57. Petterson SR, Teunis PFM, Ashbolt N. 2001. Modeling
virus inactivation on salad crops using microbial count
data. Risk Anal 21:1097–1107.
58. Schoen ME, Ashbolt NJ. 2010. Assessing pathogen risk to
swimmers at non-sewage impacted recreational beaches.
Environ Sci Technol 44:2286–91.
59. Boehm AB, Ashbolt NJ, Colford JM Jr, Dunbar LE, Flem-
ing LE, Gold MA, Hansel JA, Hunter PR, Ichida AM,
McGee CD, Soller JA, Weisberg SB. 2009. A sea change
ahead for recreational water quality criteria. J Water Health
7:9–20.
 Dose-Response Modeling and Use: Challenges and
Uncertainties in Environmental Exposure
MARK H. WEIR
3.5.3
INTRODUCTION
One of the shining examples of quantitative microbial risk
assessment (QMRA) and risk analysis in general is a greater
understanding of the risks incurred due to our voluntary
and involuntary environmental exposure. Environmental
exposures to fecal and non-fecal pathogens offer a unique
set of challenges to the QMRA modeler. One issue that arises
is the amount of exposure that is involved; for example there
is the often used 2 liters of water per day ingestion (1), and the
well accepted ingestion rates from swimming (2), however,
it is impossible if these values will hold for all instances.
Another complication is the issue of primary and secondary
exposures. Taking the example of 2 liters of water per day
ingestion, this is an illustration of primary exposure, where
the host is being exposed to the pathogens from a transport
medium (water) through the primary activity of intentionally
drinking the water. Secondary exposures are slightly more
difficult to develop, the example of ingestion rates during
swimming activities from (2) demonstrate an exposure secon-
dary to the activity being engaged in (secondarily ingest
water during the primary act of swimming. These exposures
also present to the QMRA paradigm a potential difficulty
in the appropriate use of, and the underlying uncertainties
of the dose response models and parameters. This chapter is
intended as a means of introducing dose response models
and modeling to the novice QMRA modeler. The concept
of and derivation of the models is outlined so the user under-
stands the underlying limitations of the models and data. It is
and also intended to highlight not only the appropriate use
of the dose response models and their uncertainty estimates,
but also the inclusion of environmental exposure uncertainty
into the QMRA to again control for uncertainties in the dose
response models.
DOSE RESPONSE MODELS AS YARDSTICKS
The concept of modeling and assessing risk is not a new
one by any stretch of the imagination. At the conceptual
level risk assessment has been in existence since our ancestors
decided whether to hunt larger game, trading a greater reward
for the obviously higher risk. Each person without realizing
it or performing it computationally engages in risk assess-
ment; if choosing to cross the road, not wear a seatbelt or
even as simple as having a single alcoholic drink. This consid-
ered QMRA is still a relatively new field compared to risk
assessment in general and to its modern computational peers.
The QMRA paradigm is a set of intercommunicating
modeling and analysis components. An example of the inter-
connections within the QMRA paradigm can be drawn from
an internal combustion engine. While small errors or effi-
ciency deviations in key components of an engine will not
allow for optimal functioning it does not shut down the
doi:10.1128/9781555818821.ch3.5.3
3.5.3-1
engine (e.g. one or two dirty spark plugs). However, the com-
plete failure of a major component such as a crack in the main
block or loss of oil will shut down the engine. Bringing it back
to QMRA, if the pathogenic hazard cannot be identified,
the exposure cannot be modeled; if there is no dose response
function for the selected pathogen, then the QMRA para-
digm will seize and disallow completion of the risk modeling
since the yardstick of risk is missing. Hazard identification
identifies the pathogen that is being modeled, either through
direct choice or desire to abate or prepare for a suite of symp-
toms or deleterious effects. This component typically informs
the type of environmental matrix(ices) that need to be
modeled, and in some cases informs the most feasible model-
ing framework to use. The hazard identification informs
the QMRA modeler what dose response model to research
(if already optimized) or to optimize and analyze themselves.
While it is relatively difficult to completely derail the other
steps in the QMRA paradigm, there is typically some means
of determining the hazard and modeling its exposure. The
dose response assessment is quite easy to derail, in that since
it is the yardstick of risk (3–5) if an optimized dose response
model is not present or possible the ability to perform a full
QMRA are severely compromised. Without an optimized
dose response model or animal or human data to optimize a
dose response model weak assumed threshold values or other
assumed models are all that are available to the modeler.
DOSE RESPONSE MODELS AND THEIR DATA
One of the important aspects of QMRA modeling is in the
handling and use of model and parameter uncertainties
directly in the QMRA model itself. Considering the rather
young timeline of QMRA in general, the field has been
developed during the modern computational age where
optimization of mathematical models can be undertaken rel-
atively easily, and uncertainty analysis can be presumed to be
a default action. The field of QMRA gained true footing
through the development of mechanistic dose response mod-
els (6). Through this work mechanistic dose response models
were developed for nine pathogens considered waterborne
threats; this demonstrated the differing nature of each model
and also showed how they can be utilized in dose response
modeling, thereby, QMRA modeling. This seminal paper
also showed evidence to support the single hit theory of
pathogen action and disease development. The single hit
theory, works as it sounds; each pathogen (single micro-
organism) has a definable–therefore model-able–likelihood
of initiating an infection. This theory both allows for a
straightforward modeling approach but also places the burden
of model certainty in the quality of the data underlying exper-
imental protocol. These uncertainties add to those inherent
to the QMRA model from other components such as the
exposure model and risk characterization.
3.5.3-2 ▪ MICROBIAL RISK ASSESSMENT
The ability to have a bank of dose response models peer
reviewed and available for use is a very valuable tool for the
QMRA modeler. In recent years there has been a greater
interest in research pertaining to the development and use
of QMRA models (7, 8). Additionally there are online
resources for the connected molder beyond the research
articles traditionally used. Microrisk (http://www.microrisk.
com) is a good example of a grouping of information, exam-
ples and tools for the novice and experienced risk assessor.
The Center for Advancing Microbial Risk Assessment
(CAMRA) developed a wiki to house their large library of
QMRA data and models and burgeoning tools for QMRA
instruction and use (http://qmrawiki.msu.edu). However,
none of these or other online resources can yet claim to be
the complete compendium of QMRA and dose response
data and models. Therefore, it is necessary not only to enact
good scholarship but also to know how, if necessary, to opti-
mize your own dose response model.
For most waterborne pathogens (or those microbiologi-
cally similar to them) and those that will be encountered
in the natural environment, the dose response models have
either already been optimized, or the data is available for
optimization. Should it be necessary to develop your own
optimized dose response model there are some considerations
that need to be taken into account. First a complete literature
review should be undertaken to determine the availabil-
ity of current models and available data. Appropriate, high
quality data is necessary to develop a high quality model
that can be used with confidence. The specific reasons for
these data needs will be overviewed in greater detail during
the description of the dose response model derivation and
optimization.
The first and most vital aspect of data requirements is
the type of data that is being investigated as candidate dose
response model data. As will become much clearer in the
derivation section, the data needs to be based on infectious
units. While at first read discussing units may sound to be a
simple concept, this is vital to the development of the best
candidate model that is physiologically and pathogenically
plausible. Second, in general data that is found in the open
peer reviewed literature is the best and should be the only
option for serious modeling. The open literature allows
for open peer review and discussion among the scientific
community which has greater potential for discovering and
addressing mistakes or even typographical errors in the
articles than the grey literature or internal reports and docu-
ments may not discover. To understand the data needs better
we need to understand where these models come from and
how they are derived.
DOSE RESPONSE MODELING, USE AND
UNCERTAINTIES
Dose Response Model Derivation
In QMRA the dose response assessment is the yardstick of
risk, essentially inserting the vital quantitative risk estimation
into QMRA. A dose response model in essence can be any
function that receives a dose variable (theoretically any
non-negative number) and results in an estimate of the prob-
ability of an adverse effect. Since the result is an estimate of
the probability of an adverse effect and the dose is any non
negative number, we are left with a wide range of potential
functions where; the results are bounded between [0,1] and
is effective from [0,∞]. Therefore the dose response function
can theoretically be one of any infinite number of functions,
independent of functional form or parameterization. Con-
currently a logical expansion using the Bayesian view is that
the dose response models can be any cumulative distribu-
tion function which is bounded as afore mentioned for the
dose response relationship. Therefore there is an obvious
need to narrow the scope of potential and candidate models,
first what needs to be considered is the true applicability of
the models.
As mentioned so long as the functional result of the dose
response model is bounded between [0,1] and can remain
effective for the variable input from [0,∞], it can be con-
sidered a candidate dose response model. While this holds
theoretically, in reality and operationally this is not a suffi-
cient level of detail. What needs to be considered is; what
is being modeled, the limits of this type of system and the
realities of the systems and processes involved. Another lim-
itation to our modeling approach is the need for a model
that can be easily generalized outside of the minutiae of the
system and processes. Therefore an empirical model does
not allow for the degree of flexibility a mechanistic model
does. To illustrate, developing a model specific for inhaled
spores, or ingested oocysts is not as useful as a generalizable
model that can be used for a wide array of pathogens and hosts.
First however, we need to understand what it is that we are
modeling.
To start we need to understand very well what the system
is, this then informs us of the likeliest or sole process(es) that
we must be concerned with. In this case we really have a
system of systems that we need to account for. We need to
account for the host being exposed (specifically a sensitive
location where infection and initial disease can be initiated)
in some fashion (i.e. inhalation, dermal contact or ingestion),
and since we are modeling living pathogens, we need to
account for their stability and likeliness to survive to generate
an infection. Therefore our system is the host’s exposure to
a dose of pathogens of which there are some number of
pathogens surviving to reach a location for development of
an infectious foci. In order to model this combined system
in a generalized form, the likelihood of a dose of pathogens
for each of these subsystems needs to be modeled. Through
modeling the likelihood, the resulting model can be general-
ized beyond specific parameters such as exposure type (inha-
lation vs. ingestion) or pathogen species. The two most
widely used mechanistic dose response models, the exponen-
tial and beta Poisson models will be discussed and used as
examples.
The two most widely used microbial dose response models
were derived using the same fundamental postulation; there
is a d dose of pathogens the host is exposed to, a dose of j
pathogens that are ingested (or inhaled) by the host from
that original average dose d, from the j pathogens available
there are a subset level k pathogens that survive from j to ini-
tiate the infectious foci. Therefore we develop the relatively
simple concept of equation 1 which combines the likelihood
of j organisms being ingested from d organisms, P( j|d); with
the likelihood of k organisms (from j ) surviving to initiate the
infection, P(k|j). Then if we postulate that of k organisms
surviving to initiate an infection, there is a minimum number
which are needed to initiate an infectious foci, we are lead
to equation 2. Taking this a step further and stating that if
kmin is equal to 1 (i.e. cannot have fractions of whole live
pathogens) then equation 3 allows us to model the probability
of illness given a specific dose d of pathogens. It is important
to note that kmin is not intended to be the minimum infective
dose of a pathogen, rather the theoretical minimum of any
infectious agent. Therefore we hold it at one infectious unit
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-3

(CFU, spore, PFU, etc) thus not allowing for negative, zero or derivation describing the very basic mechanisms needed
fractions of infectious units. for infection; transport to the interior of the host’s body
X1 and survival of the pathogen to initiate an infection. This
PðkÞ ¼ p1 ð jjdÞ p2 ðkjjÞ ð1Þ type of model is necessary considering the different dynamic
j¼1 that pathogens present compared to other environmental
hazards. Beyond the fact that pathogens are hampered by
X
1 X
1
a survival rate that can be altered by their environment,
PðdÞ ¼ p1 ðjjdÞp2 ðkjjÞ ð2Þ there is also a need for specific transport, meaning a
k¼kkmin j¼k gastrointestinal pathogen needs to be transported to the
gastrointestinal system of the host for infection and disease
X
1 X
1 X
1
PðdÞ ¼ p1 ð jjdÞ p2 ðkjjÞpðkmin Þ ð3Þ development.
kmin ¼1 k¼kmin j¼k
The other physiologically and pathogenically plausible
model used widely is the beta Poisson model (11, 12);
From this point we can develop the functional forms of that can in essence be considered a special case of the expo-
the individual probabilities ( p, p1 and p2) from equation 3. nential model. The intent of the beta Poisson dose response
Using a simple postulation that for any sample of exposure model is to expand the exponential model beyond one of
media, such as ingested water, organisms are distributed ran- the fundamental assumptions, that the rate r, can be
domly, p1 can be modeled using a discrete random distri- assumed to be constant. The parameter r is optimized using
bution, thus the Poisson distribution (equation 4). animal or human dosing data, but mechanistically remains
assumed to be constant to describe the rate of survival of
p1 ð jÞ ¼ dj ed=j! ð4Þ the pathogen to initiate an infection. Thus the beta Poisson
removes the discrete nature of the parameter r and replaces it
Then using the fundamental assumption that the survival
instead with a distribution of r values, specifically with a fre-
of a single infectious pathogen is independent of other
quency distribution f(r). This derivation was developed first
pathogens present, we can set a constant rate, r, for this sur-
by Furumoto and Mickey (13) from preliminary work by
vival rate. There is a new theory that is attempting to further
Moran (14) who was hampered by limited computational
expand on this assumption of independent action, stating
capabilities, namely integral tables for the function devel-
rather than independent action, quorum sensing plays a key
oped. In (13) it was determined that the beta distribution
role in influencing and regulating pathogenesis (9, 10).
(equation 8) had two primary benefits to describe the rate
While the process of microorganism communication likely
r from the exponential model. This distributional descrip-
has a role to play in the development of the resulting disease
tion of r is used as f(r) in the definite integral in equation
following infection, it is likely a contributing factor along
9 (14). The beta distribution limits the range of probability
with independent action. Until a model is developed for
values, and allows for flexibility where shape can be changed
the combined actions of quorum sensing and independent
effectively and drastically by using just two parameters, this
action and validated as the current mechanistic models are,
benefit allows for optimization of a very wide range of data,
we operationally must continue with only the independent
Figure 1 demonstrates the effect of various parameter values
action assumption. Through the independent action assump-
for α and β.
tion we can model p2 using a binomial distribution, since
there are two possibilities, either the pathogen survives or
does not. Gða þ bÞ a1
f ðrÞ ¼ r ð1  rÞb1 0x1
GðaÞGðbÞ ð8Þ
½ j!  rk  ð1  rÞ jk 
p2 ðkÞ ¼ ð5Þ f ðrÞ ¼ 0 elsewhere
½ð j  kÞ!  k!
Before invoking the final summation we can follow the
first two probabilities to their solution. Equation 6 demon-
strates the probability of a response based on an average
dose d. What is seen in equation 6 is a Poisson summation,
from which we can develop the exponential dose response
model kmin must be assumed to be equal to 1. This means
that there is an absolute minimum of 1 infectious particle for
an infection to develop. When kmin is assumed equal to 1 the
result in equation 7 is the well recognized exponential
dose response model, widely used in QMRA practice and
research.
X1
ðd  rÞ2 edr X1
½dð1  rÞ jk edð1rÞ 
pðdÞ ¼
k¼k
k! j¼k
ð j  kÞ!
min

X1
ðd  rÞk edr
! pðdÞ ¼ ð6Þ
k¼k
k!
min

FIGURE 1 Conceptual diagram of the experimental protocol for

PðdÞ ¼ 1  erd ð7Þ
animal and human dose-response trials, resulting in the development
What we have just described is a physiologically and of the required data for optimization.
pathogenically plausible dose response model; due to the doi:10.1128/9781555818821.ch3.5.3.f1
3.5.3-4 ▪ MICROBIAL RISK ASSESSMENT
ð1
rd
PðdÞ ¼ ð1  e Þ f ðrÞ dr ! PðdÞ
0 protocol such as the following, depicted visual in figure 1:
ð1  
Gða þ bÞ a1
¼ ð1  erd Þ r ð1  rÞb1 dr ð9Þ
GðaÞGðbÞ
0 2. The separation of hosts needs to be random, there cannot
The integral in equation 9 after substitution was shown
that despite the probability of observed response to dose d,
P(d) results in the confluent hypergeometric series (13),
and that for sufficiently high values of β the integral in equa-
tion 9 can be approximated in the form seen in equation 10.
A more descriptive form of the beta Poisson model (equation
12) is one where the median infective dose (N50) is estimated
using the relationship in equation 11. Additionally an opti-
mization of the confluent hypergeometric series is computa-
tionally intensive for the novice modeler.
  a
d response.
PðdÞ ¼ 1  1 þ ð10Þ
b Before discussing the optimization we first need to make
b dose response models to. If you consider the classic examples
N50 ¼ ð11Þ
21=a1
 a
d  21=a  1
PðdÞ ¼ 1  1 þ ð12Þ
N50
Dose Response Model Optimization
The objective of deriving the dose response models is to be
able to use them to describe the likelihood of a response
based on a dose of pathogens. An example of optimized
dose response data can be seen in Box 1. For this to be possible
the models need parameters with values capable of predicting
a likelihood of response for each pathogen and exposure
route they are tested with, this is the true test of a generalized
model. Before briefly discussing the method and mechanics of
optimization, we first must discuss the data that will be used
for the optimization itself and starting of course with the
source of data, the animal model experiments.
The best possible experimental protocol would be one that
can model as close to the acceptable risk levels as possible
for the population and host species that are of most concern,
humans. Volumes of texts can be printed on the subject of
ethical issues of human and animal subjects research for infec-
tious disease research, but we will not delve into that issue
here. Rather let us consider the mathematical constraints of
achieving as clear a picture of the dose response relationship
for these acceptable risk levels. For drinking water we con-
sider an acceptable risk to be 1:10,000 or a probability of
illness of 0.0001. For recreational waters we consider a risk
of 8:1,000 or probability of illness of 0.008 acceptable. This
means that respectively for the drinking water and recrea-
tional acceptable risks to attain complete accuracy of assess-
ing the dose response relationship we would need an animal
experiment where 10,000 and 800 animals respectively
are involved in an illness trial. This is logistically infeasible,
nearly impossible financially and again raises ethical ques-
tions. Therefore the generalized models are limited by a
more realistic number of hosts and experimental protocols.
The experimental protocol details are dependent on the
host species, pathogen and exposure route. Therefore the
laboratory protocols will be different for what is being eval-
uated and can be different between laboratories performing
the same animal model. However for the model optimization
process, all that is required is that the experimental protocol
and resulting data abide by or allow for a computational
1. Hosts (animals or human subjects) are separated into k
groups
be a reason for why some hosts were separated to specific
group
3. In any specific group i within k total groups all hosts in
the group are exposed to a known and quantified dose
(di), within the individual groups i there are a total num-
ber of hosts (Ti) of which there are a number of positive
responses (Pi), positive response being defined as the ini-
tiation of infection, illness or disease; depending on the
response being modeled.
4. The resulting data is quantal, meaning that there are only
two results, the host exhibits either a positive or negative
sure that there is a reasonable set of data to optimize the
of dose response data, there is an increasing trend of like-
lihood of a positive response (Pi) given a greater dose of the
pathogen (di). However the problem with any optimization
routine is that even if there is not a significant trend, the
optimizer can still regularly reach a solution and set of param-
eter(s) for that set of data. Unless there is no trend or the data
is heavily dispersed then the optimizer may still converge at
a solution. Therefore it is important to establish the signi-
ficance of the trend in the dose response data before optimiz-
ing the dose response models to the data set. This is a simple
but vital quality control check so as to develop the most high
quality model possible.
The Cochran-Armitage test of trend can be used in a sim-
ple assessment of the trend of the data. Equation 13 shows
how to determine the critical value that is then compared
to the upper 5th percentile of the Normal distribution (value
of 1.644). Since this is a one tailed test the critical value
(ZCA) needs to be greater than the upper 5th percentile.
Should this test fail then it is strongly advised that this data
is not used to optimize the dose response models since this
is likely suboptimal data that should not be used in dose
response modeling.
Pn
i¼1 ðdi   xÞ pi
zCA ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Pn ffi where :
pð1  pÞ i¼1 ni ðdi  xÞ2
Pn Pn ð13Þ
n d p
and p ¼ Pi¼1
i i i
x ¼ Pi¼1 n n
i¼1 ni i¼1 ni
While there are various options for optimizing a model, a
robust and effective method is the maximum likelihood esti-
mation (MLE) method (6, 8). The entire point of an opti-
mization is to find the parameter(s) that allow for a model
to optimally describe the data. The MLE method requires
the development of a likelihood function that must be maxi-
mized to develop the optimal set of parameters. This protocol
allows for a straightforward method of developing the like-
lihood function that is necessary to use the MLE method.
Since the data is quantal the first obvious step in developing
the likelihood function is to describe the data using the
binomial function (equation 14). The quantal nature of the
data allows us to describe the underlying function of number
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-5

BOX 1

Humans exposed to Vibrio cholera buffered oral exposure

Dose(CFU) Infected (Pi)[#] Not Infected[#] Total (Ti)[#]

10 0 2 2
1,000 3 1 4
10,000 11 2 13
100,000 7 1 8
1,000,000 21 2 23
100,000,000 2 0 2

Estimated parameter cloud from bootstrap uncertainty analysis

(10,000 iterations), ellipse represent the confidence intervals of
the α and N50 parameters of the beta Poisson model.
Optimized beta Poisson model with confidence intervals.

Humans exposed to Giardia duodenalis oral exposure

Dose(CFU) Infected (Pi)[#] Not Infected[#] Total(Ti) [#]

30 0 8 8
300 0 7 7
30,000 0 6 6
125,000 2 1 3
500,000 4 1 5

Histogram of k parameter for exponential model from 10,000

iteration bootstrap uncertainty analysis. Optimized exponential model with confidence intervals.
3.5.3-6 ▪ MICROBIAL RISK ASSESSMENT
of hosts exhibiting a positive responses, f(Pi) using; predicted
probability of response (πi, the dose response model with
parameter(s) Θ and total number of hosts (Ti). Taking
the natural logarithm of this function allows us to improve
the optimization by decreasing most of the dispersion in the
data that leads us to equation 15, which is our likelihood
function.
 
Ti !
f ðPi Þ ¼ ðpi ÞPi ð1  pi ÞðTi Pi Þ ð14Þ
Pi !ðTI  Pi Þ!
X
N
lnðLÞ ¼ ln½f ðPi Þ ! lnðLÞ
i¼1
X
N models’ fit to the data to a χ 2 distribution at confidence level
¼ ½Pi lnðpi Þ þ ðTi  Pi Þlnð1  pi Þ where:
i freedom of n-m, where n is the number of dose groups and
pi ¼ Pðdi ; QÞ
Maximizing a likelihood function is more than appropri-
ate for optimizing a function, however, developing a means
of directly comparing the prediction of a model to the data
is a better option. This best possible value is the observed
probability of response from the raw data being used in the
model. The best means of accomplishing this comparison is
to develop a function that determines the difference between
the predicted and observed probability of response, serving as
a derived statistic describing the fit of the optimized model
to the data. In the case of our quantal dose response data
developing a function for deviance (Y ) from the observed
probability of response can be calculated by taking −2 multi-
plied by the log likelihood ratio (equation 16). The −2 factor
transforms the maximization of a function (the likelihood) to
a minimization of a positive value. From here now though we
need to quantify the fit of the model(s) to the data.
k     
X pi 1  pi
Y ¼ 2 pi ln o þ ðni  pi Þ ln
i¼1
pi 1  poi
pi
where: poi ¼ and pi ¼ Pðdi ; QÞ
ni
Determination of Goodness of Fit and Best
Fitting Model
As described before in the model derivation section, there are
theoretically an infinite number of models and functional
forms that can serve as microbial dose response models. While
we narrowed this range significantly in the need for physio-
logically and pathogenically plausible dose response models,
neither of these models can be chosen as the best option
without sufficient evidence. This is a continual theme in
computational sciences and engineering, where the best
model must be chosen carefully and with sufficient weight
of evidence. Therefore the dose response models must first
prove; their goodness of fit, then their validity of being a
better choice over the alternative model ( prove the benefit
of additional parameters). A third consideration is in the
case of nested models, how to quantify the ability to use
nested models best model to do that with.
These determination can be made immediately after opti-
mization of the models since the criteria for these analysis
do not require the development of confidence bounds for
the parameters. In all cases the chi-squared (χ 2) distribution
will be used to determine the significance of the fits. The
null hypothesis (Ho) for; the goodness of fit analysis is that
the model is a good fit to the data, for the best fitting model
analysis the Ho is that the simpler (one with less parameters)
of the two models is the best choice and for the nested models,
the Ho is that the data comes from a shared distribution and
thus can be described with a combined model. For simplicity
sake we will be considering only the exponential and beta
Poisson models for these examples. Box 2 shows an example
of model optimization with goodness of fit and best fitting
model analysis results.
First it must be determined if the models have a good fit to
the data after being optimized. For this analysis what is needed
is: the value of the optimized (minimized) deviance for both
models, number of parameters in the model and number of
dose groups. This analysis is simply a comparison of the
(α) of 0.05 (or more constrained if desired) and degrees of
m is the number of parameters of the model. It must always
ð15Þ be remembered that the χ 2 distribution is an upper tailed dis-
tribution, therefore, if the minimized deviances of the dose
response models are below the χ 2 critical value then the Ho
is failed to be rejected, or in this case the model is considered
a good fit to the data.
Then the best fitting model is determined using an analy-
sis in which the Ho is that the simpler of the two models is the
best fitting. Therefore this Ho is that the model with less
parameters is the best fitting model. This places an additional
burden of proof required to justify the inclusion of an addi-
tional parameter (comparing exponential and beta Poisson
models). In this case the difference between the deviances
termed as Δ is compared to a χ 2 critical value with the chosen
α (0.05 is used in Box 2) and 1 degree of freedom.
Nested models or pooled data sets are ones where multiple
data sets are combined together and a central model is devel-
oped for this combined data set. The nesting of models have
been shown to be valid across host species and pathogen
strains (12), (15). Consider the previous best fitting model
analysis, but in this case the difference in deviances (improve-
ment of fit) needs to account for all the individual possible
models as well as the pooled model deviance. Let us consider
ð16Þ
a difference between the summed deviances of the individual
model (Yi) and the deviance of the pooled model (Yn) and
term this Δn. Using the Ho in this case that the data sets
have a shared distribution, the Δn is compared to the χ 2 crit-
ical value at a chosen confidence level and k–m degrees of
freedom; where k is the number of parameters for all pooled
models, i.e. two data sets with an exponential model for
one data set with a beta Poisson model for the other will
have a k = 3, and m is the number of parameters in the pooled
model.
Dose Response Model Optimization Uncertainty
Now that there are optimized models, we quantified their fit
to the data and have determined the best fitting model, the
underlying uncertainty needs to address. Model uncertainty
is just as it sounds, the inability to describe with complete
confidence the true underlying risks. All models are limited
by some degree of uncertainty, even a model as simple as an
empirically derived standard curve is limited in confidence
on the underlying data. While human error is very difficult
to quantify in an uncertainty analysis, structure of the models
and their underlying data can be quantified in a relatively
simple means.
There are a number of sources of model uncertainty in the
overall risk assessment. The entire risk model will have uncer-
tainty from the hazard identification and development,
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-7

BOX 2

TABLE B2.1 Example dose-response data

Dose (di) Positive Response (Pi) Negative Response
100,000 9 0
10,000 9 0
1,000 4 5
100 3 6
10 1 7

TABLE B2.2 MLE output for the exponential and beta Poisson models
Model Y α N50 k LD50 or ID50
Exponential 8.4574 NA NA 0.0011 647.2228
Beta Poisson 5.7672 0.5632 307.5282 NA 307.5282

TABLE B2.3 Goodness of fit analysis comparing the deviance (Y) to the χ2 critical value at confidence
interval of 0.05 and degrees of freedom of 4 for exponential and 3 for beta Poisson

Model Y χ2 Critical Value χ2 p-Value Conclusion

Exponential 8.4574 9.4877 0.0762 Exponential model is a good fit
Beta Poisson 5.7672 7.8147 0.1235 Exponential model is a good fit

TABLE B2.4 Best fitting analysis comparing the difference between deviances (Δ) to a χ2 critical value at confidence
interval of 0.05 and 1 degree of freedom

Model Y Δ χ2 Critical Value χ2 p-Value Conclusion

Exponential 8.4574 2.6902 3.8415 0.101 Exponential model is best fit
Beta Poisson 5.7672

FIGURE B2.2 shows a cloud of bootstrapped α and N50 values

FIGURE B2.1 shows the optimized beta Poisson model with from 10,000 bootstrap iterations. The confidence intervals around
95th and 99th confidence intervals around the central model, the could are a replacement for simultaneous confidence intervals
with optimized α and N50 parameters shown in Table B.2. which would be required for these parameters.
3.5.3-8 ▪ MICROBIAL RISK ASSESSMENT
the exposure model, especially as the complexity of the expo-
sure model increases (i.e. simple ingestion rate compared
to a 3-dimensional dispersion model), the dose response
model and combining all of theses models together into a
complete risk model. It is important to understand that
the dose response models are optimized to animals or human
controlled dosing trials. Therefore any uncertainty carried
through from the experimental protocol, such as dose quan-
tification or inoculum preparation, will be carried into the
optimized parameters as well. While all possible forms of
uncertainty cannot be accounted for, there is a simple and
generalizable method that will allow for us to quantify general
dose response parameter uncertainty. The key is to vary the
data point that holds a great amount of influence over the
optimization of the model thus the best fitting parameter
(s). If you think back to the development of the deviance
(equation 16) the observed probability of response was the
data dependent term in the deviance of greatest importance.
Therefore by varying the dependent variable or the observed
probability of response will garner the best results for us.
In essence all uncertainty analyses are iterative systems,
meaning that a computational task is evaluated or executed
a set number of times to determine how and to what degree
the central model or computational data is changing. The
classical example from probability theory is to think of a fair
six-sided die tossed with the results recorded. In essence given
enough throws of the die you can start to determine the prob-
ability of incurring a specific value such as the number 3. You
can take this a level further and develop a model or probabil-
ity distribution of this likelihood to attempt to predict the
likelihood of a 3 showing given past values. The essential con-
cept to take from this parable is that given a greater number of
attempts or iterations the greater your understanding of the
system that you are modeling. Therefore if what we want to
understand is how the underlying distribution or trend of
the data affects the resulting model, we need to computation-
ally test different forms of the dose response data.
The bootstrap is a method of uncertainty analysis that
leverages the underlying distribution of the data to describe
how the data’s uncertainty affects the model that is optimized
to this data. This method of uncertainty analysis is considered
a frequentist approach, as opposed to Bayesian methods. A
Bayesian method leverages assumed or optimized probability
distribution functions to descriptive uncertain variables or
data in model or analysis. The Monte Carlo method is an
example of this type of uncertainty modeling and analysis.
However, what is desired in this case is to not confound and
complicate the current level of uncertainty by using assumed
or optimized probability distribution function. Rather the
underlying distribution will be leveraged to understand the
uncertainty of the optimized models.
The bootstrap method requires that the data is randomly
sampled with replacement. The with replacement part is
very important, especially in the coding of this method, since
the concept of with replacement is, in essence the data is
rearranged. As we discussed previously, the dependent varia-
ble that informs the test of trend and deviance for the opti-
mization of the model, the positive response is the one that
will be randomly sampled with replacement. As can be seen
in Box 3 what this essentially results in is a new set of dose
response data at each iteration. If rather than the positive
responses but the dose is randomly sampled with replacement
then we have varied the dependent variable and cannot see
differences in observed probability of response since we would
not be changing the positive response or the dependent
variable in the data set.
The mechanics of developing a bootstrap for dose response
models is wonderfully simple, so long as the underlying prin-
ciples of sampling with replacement are kept in mind. Given
the computational power of even a simple smart phone, a half
decent laptop computer will have no difficulty in accompli-
shing a 10,000 iteration bootstrap routine. The number of
iterations are a moving target and based on the modeler them-
selves. Typically for testing purposes the number of iterations
are kept purposely low (∼1,000) for debugging and evalua-
tion, the higher the iterations the longer the bootstrap will
take to complete. When the bootstrap is being used for anal-
ysis the number of iterations are increased to improve confi-
dence assessments. The iteration level of 10,000 can
typically be considered a good baseline, since in essence we
are approaching the 10,000 host experimental trial in the
computational world; remember we are developing compu-
tational data via the bootstrap. There is really no reason to
use iterations lower than 10,000 since computation time on
a relatively simple model is no longer a consideration with
even inexpensive modern computers. To far above this value
and you are increasing computation time for not a equally
greater level of confidence assessment.
A bootstrap requires a set of data that can be evaluated in
some way by a function or mathematical construct. So what is
happening in the dose response model sense; we randomly
sample with replacement the dependent variable ( positive
response) from the data, then we use this “new” dose response
data set to optimize the dose response models. This changed
data set will develop a new deviance of fit and new dose
response parameter(s) for each of the models. Therefore as
compared to other uses of the bootstrap method, we are not
interested in the underlying probability distribution, rather,
we are utilizing the underlying distribution of the data. The
bootstrap method allows us a number of benefits for the
increased computation time, depending on the number of
iterations. Think back to the example of the fair die with
an increased number of throws we can start understanding;
what the next likely number is (based on previous values)
what trend does the value of the showing number take if
any and how we can determine the confidence of these
predictions can be estimated. So take this example to dose
response since we are sampling with replacement, we are
essentially developing a larger data set of positive response
values for each dose and total number of hosts in the dose
groups. Therefore we are developing computational data
without processing the original dose response data set through
another model or probability distribution.
At each iteration of the bootstrap the entire data set is
randomly sampled from ith replacement with respect to the
dependent variable (observed probability of response); this
’new’ data set is then used to optimize the dose response
model(s) to develop a new set of optimized parameters and
minimized deviance. If we do this for 10,000 iterations, we
then develop a set of 10,000 potential parameter values
depending on the changed probability of positive response.
From this set of 10,000 bootstrap realizations we can develop
a means of describing the uncertainty of the parameter(s)
in the model(s). This can take the form of confidence inter-
vals around the parameter(s); in the case of the exponen-
tial model a standard confidence interval is possible; in the
case of the beta Poisson, due to the functional relationship
between α and N50 simultaneous confidence intervals are
required; or as can be seen in Box 3 confidence ellipses can
be developed for these parameters. Theoretically in the case
of a data set where all doses had a non-zero probability of
response, then a variability term such as the standard
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-9

BOX 3

TABLE B3.1 Original dose-response data before bootstrap TABLE B3.2. Randomly sampled dose-response data after a
iteration single bootstrap iteration

Dose (di) Positive Response (Pi) Negative Response Dose (di) Positive Response (Pi) Negative Response
1000 0 14 1000 0 14
100000 28 76 100000 15 76
100000 32 84 100000 32 84
10000000 15 15 10000000 28 15
10000000 16 16 10000000 4 16
100000000 8 1 100000000 8 1
1000000000 4 0 1000000000 16 0
1000000000 20 2 1000000000 20 2
*Bold italic font signifies the data randomly sampled and replaced.

FIGURE B3.1 Bootstrapped parameter uncertainty cloud FIGURE B3.2 Dose response model curve for beta Poisson
(10,000 points) for beta Poisson model, showing confidence inter- model, with confidence intervals from 10,000 bootstrap
val ellipses, unless simultaneous confidence intervals are devel- iterations.
oped these ellipses are best estimation of α and N50 uncertainty.

FIGURE B3.4 Dose response model curve for exponential

FIGURE B3.3 Bootstrapped k parameter histogram from model, with confidence intervals from 10,000 bootstrap
10,000 iterations. iterations.
3.5.3-10 ▪ MICROBIAL RISK ASSESSMENT
deviation (σ) of variance of the parameter(s) can be used as
well. Since for all dose levels there was some positive response,
there then is a variable rather than uncertain probability of
response, which can be described with variability on the
dose response parameter(s).
ENVIRONMENTAL MATRICES
Through QMRA we are attempting to model both the envi-
ronment external to the host and internal in the infection
process. The exposure modeling and analysis component,
models the external environment and how human hosts
interact with it. Once the pathogens enter the host’s body
then the dose response through the derivation described
previously becomes semi-independent of the environment
from which they came. There are two primary issues with
this assertion; first the environmental media that is conta-
minated with pathogens may harbor other contaminants
which may limit or enhance our immune system response
(i.e. toxins or buffering agents) and second the type of expo-
sure cannot be accounted for in the dose response model
(i.e. primary and secondary exposures).
Humans are exposed to pathogens through their interac-
tion with each other and the environment. Since this chapter
and book are aimed at environmental exposures we will focus
on them. Exposure assessment and science are complete
scientific fields unto themselves, for which volumes can be
printed and absorbed and therefore, cannot be explored in
this chapter. However, exposure and environmental matrices
are critical to dose response modeling as well. As was discus-
sed in the derivation of the dose response models, each animal
in each dose group is exposed via the best means of exposure
in the laboratory environment. It is important to note that
these means of exposures in the laboratory are not comple-
tely realistic in all cases and need to simulate exposure. An
overview will be presented of the mechanisms used to expose
laboratory animals to pathogens, this this meant as a very
brief overview to introduce the potential error and sources
of uncertainty just from the experimental protocol itself.
Human volunteer exposure trials are much more straight-
forward on the exposure mechanisms needed. Since the
humans are volunteers they freely and willingly ingest the
pathogens in an oral exposure trial, or inhale them in an
aerosol trial.
Independent of the detailed minutia with which the
doses are delivered, it is important to remember that these
are experimental simulations of real host exposure. Let’s start
with an example from aerosol exposure that can be performed
in a number of different ways. There are aerosol masks that
are placed over the face of the host animal. The benefits of
the mask process is to not artificially bypass portions of the
respiratory system, and theoretically it is more likely the
host animal will breath normally and not in a panicked fash-
ion. The mask can be easily secured and sealed (to prevent
leaks) around the face of the host animal. Visualize a hospital
oxygen mask, while being a gross simplification, this is a help-
ful visual to have when thinking of this method. This is the
optimal means of exposure to a bioaerosol, since the entirety
of the respiratory system is being exposed; nose, mouth, tra-
cheobronchial, bronchioles and alveoli. Therefore this is
the closest to a realistic bioaerosol exposure that laboratory
trials have achieved. The main drawback is that realistically
this method is only appropriate for primates and humans,
where a mask can be easily developed or adapted. Aerosol
exposure trials can also take the form of intranasal, intratra-
cheal or intrabronchial exposure. Through these exposure
routes the pathogens are introduced to the host through those
main structures of the respiratory system. Intranasal exposure
is performed by inserting a tube or other means of delivering
a dose of pathogens into the bulk fluid of the nasal cavity
of the host. For intratracheal and intrabronchial exposure
a tube is inserted either through their nasal/oral cavity or
directly through an incision into the trachea and bronchi
(or bronchioles). This is meant to simulate a pathogen that
has been inhaled and minimal to no removal was accom-
plished farther up in the respiratory system.
Going from aerosol experimentation to oral exposures,
these mechanisms are not as straightforward as feeding or
spiking the water of the host organisms. This is another area
where the host organism needs to be manipulated into ingest-
ing the pathogens for the experiment. Oral exposures usually
take the form of a needle (blunt tip) placed in the mouth of
the host animal (typically while being held) inocullum is
then injected into the animal’s mouth with ingestion being
monitored. Another method is to use a feeding tube of sorts
placed into the upper digestive tract thus reducing the need
for ingestion monitoring. And lastly a fairly new method
involving a ball tipped needle, that the animal draws water
(or other liquid media). This shows promise as being the
closest to mimicking true oral ingestion that experimental
protocols have reached.
Through all of these means of exposure the animals are
expected to breath or swallow in a normal manner. This is
obviously not the case for most of the hosts, while maximum
care is taken to ensure that the animals are not unduly harmed
in the exposure process; normal breathing and ingestion
are obviously compromised. The difficulty in accounting
for these sources of error or uncertainty directly in the dose
response model, is significant. Rather these possible sources
of uncertainty need to be grouped together as model parame-
ter uncertainty. As was described previously the uncertainty
analysis performed on the dose response model parameter(s)
varies the dependent variable, of the observed probability
of response. This uncertainty analysis using the bootstrap
method is an essential first step in accounting for these uncer-
tainties, however, these confidence intervals need to be
appropriately used in the QMRA that they are vital to.
Therefore it is highly recommended and not unreasonable
considering the modern computational power readily avail-
able to invoke a Monte Carlo or other uncertainty analy-
sis into the final QMRA. Using the example of the Monte
Carlo method. This is a Bayesian method, therefore, to des-
cribe uncertain variables in the QMRA model with fitted,
known or assumed probability distributions. In the case of the
dose response parameters a triangular distribution has been
successfully used to describe the uncertainty of the dose
response parameters (16, 17). A triangular distribution (equa-
tion 17) is a simple function that can be used readily as an
assumed distribution for sparse data sets. The triangular distri-
bution is bounded by the minimum value (a), maximum
value (b) and values of the distribution informed from the
likeliest value (c).
8
> 0 for x , a
>
>
>
> 
>
> 2ðx aÞ
>
> for a  x  c
< ðb  cÞðc  aÞ
f ðxja; b; cÞ ¼ ð17Þ
>
> 2ðb  xÞ
>
> for c  x  b
>
> ðb  aÞðb  cÞ
>
>
>
:
0 for x . b
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-11
Therefore the median dose response parameter(s) value(s)
can be drawn from the bootstrap realizations as well as an
upper and lower bounds from the confidence intervals. This
will parameterize the triangular distribution and allow for
a depiction of the uncertainty in the dose response parame-
ter(s). Other than using an assumed triangular distribution,
it is reasonable to also fit a probability distribution to the
10,000 bootstrap realizations. However, care must be taken
in this case, some optimizers in available Monte Carlo mod-
eling applications may not be able to handle 10,000 data
inputs and may give erroneous fits or be unable to converge
to a solution. In these cases, using a more computationally
robust algorithm through R or MATLAB (or its open source
equivalents such as Octave) is recommended.
In either case the parameter(s) can then be described using
the probability distribution. This will allow for random sam-
pling of those parameters as an additional layer of uncertainty
control and analysis for these parameter(s). This will not
directly answer the questions raised by the experimental pro-
tocol or the environmental matrix used in the exposure trial.
However, it is a means of detailing the uncertainty in the
parameter(s) that will result from these questions. Further
research is needed to understand and address the underlying
uncertainties that may arise from the experimental protocols.
The advanced model in physiological dependence of the dose
response is a step in this direction, however, is hampered by a
lack of data and significant need for further research.
As touched on earlier primary and secondary exposures
are a source of uncertainty and difficulty in modeling the
exposure and complete QMRA model. The uncertainty is
typically accounted for in the QMRA model; however, there
are concerns in dose response modeling as well, specifically a
twin set of issues, all relating to the experimental protocol.
First all of the methods of exposure in the experimental pro-
tocols are primary exposure, such as direct inhalation or inges-
tion. Secondary exposure trials are very difficult and outside
of human volunteer trials nearly impossible to achieve. Just
reasoning through the mechanics of having a laboratory ani-
mal take part in an activity such as controlled swimming is
nearly impossible, but then to monitor and quantify the exact
dose that was ingested takes this a step further into unachiev-
able. The other complication in secondary exposures is the
additional contaminants and general factors of environmen-
tal media.
Remaining in the water environment for this example
what we see through activities such as boating or swimming
is an ingestion of a smaller quantity of water than through
primary exposure. Added to this that we are talking about
environmental water; additional contaminants such as solids
(comprising turbidity) and other potentially harmful con-
taminants that may weaken the immune system or shelter
the pathogen from initial detection and destruction by the
host’s body may be present. If a pathogen is sheltered by a solid
particle in the water, the host’s body may not be able to detect
the pathogen over the larger foreign object, the particulate
mater. This would then lead to possibly a lower likelihood
of destruction of the pathogen in the digestive system, and
identification and destruction by the host’s immune system.
The effects of additional harmful contaminants is one that
is under research currently, however, there is no information
pertaining to the quantified effects on the dose response rela-
tionship and models. Theoretically a combined exposure
to chemical as well as pathogenic contaminants may lead to
a combined effect on the host. Considering an example of a
chemical contaminant which affects the body leading to a
weakened state, the immune system may not be able to muster
as robust of a defense than compared to being exposed to just a
pathogen. This is not unheard of, opportunistic pathogens
have been known of for some time, pneumonia is a good
example of this. Therefore theoretically a pathogen not typi-
cally considered opportunistic might have a similar foothold,
especially if it has a higher virulence already.
The issue of fecal pollution and exposure is another
issue that needs to be addressed. Gastrointestinal infections
almost always lead to diarrheal disease and effects children
to a greater degree than adults. The global burden of gastroin-
testinal disease for children <5 years old are estimated to be
approximately 1.7 billion children annually (18). There are
a number of means of exposure to fecal pathogens; here again
the concept of primary and secondary exposures comes into
play. Taking Norovirus, an increasing cause of gastroenteritis
(19, 20); this virus can easily cause illness through both pri-
mary and secondary exposures. Humans can be exposed
from drinking contaminated water directly or from swimming
in contaminated water. The interesting thing about Norovi-
rus is the ability for exposed individuals to contaminate other
surfaces, food and water, thus potentially contaminating
others in the group. This wide range of potential exposures
from one virus is something that cannot be accounted for
in the dose response models currently. The level of uncer-
tainty of known dose to the host organism would be too
great for this many exposure routes. Therefore the exposure
model in the QMRA needs to deal with these exposures
and their complexity, since the dose response models are
tied to their single route exposed dose limitation. The con-
cept of advanced dose response modeling is an interesting
option that is being explored in an attempt to use additional
data in dose response experiments (e.g. age of the host), or
develop innovative methods of describing a number of com-
plications that separate theoretical dose response relationship
from the real experienced one.
THE FUTURE OF MICROBIAL DOSE
RESPONSE MODELING
Concept of Advanced Dose Response Modeling
As reviewed previously, dose response models are optimized,
the best values for the parameters, to animal or human dosing
data. This means that the data is developed through con-
trolled laboratory experimentation, which is where all models
can falter or need further development or explanation. The
derivation of the dose response models is based on the devel-
opment of a generic infection from a generic pathogen. This
is stated meaning that the dose response models are mechanis-
tically derived, based on the processes of pathogenesis not
specific to a certain pathogen. This is what separates mecha-
nistically and empirically derived models, the essence of
generalizability, that the models’ derivation are not tied to
or limited by a specific data set, pathogen, exposure route or
other data source.
However, there is a limitation in the data that is applicable
and can be used in the optimization of the dose response mod-
els. While experimentation and experimental protocols are
developed to simulate realistic exposure scenarios, most of
the time, all permutations of an exposure event cannot be
encapsulated into a dosing trial. Experimental protocols are
typically limited by:
• Total host size, meaning the number of hosts in the dosing
trial. As was discussed previously in the example of
modeling the exact acceptable risk, financial, ethical
3.5.3-12 ▪ MICROBIAL RISK ASSESSMENT
and logistic concerns typically countermand the desire to
use an excessively large number of animals in the trials.
• Host animal species and strain. Laboratory strains of ani-
mals in some cases, especially long used inbred strains
have shown differing levels of susceptibility if any. A clas-
sical example of this is the BALB/c mouse, the originating
animals of which were obtained in 1913, with inbreeding
first occurring in 1920. This long standing inbreeding has
resulted in specific decreased resistance to pathogens (21).
This inbreeding issue aside, there are other limitations,
such as the strain of Cryptosporidium that infects humans
is not infectious to mice, which is the likeliest host species
in dose response experimentation. Beyond these limita-
tions we have seen the ability to expand the current mod-
els from one species to another very easily (12, 15, 8).
• Inoculum development: The development of the inocu-
lum can take a number of forms, and is typically dependent
on the pathogen and exposure route that is being tested in
the experiment. In some cases the pathogen is passed
through another host organism, such as a sheep for even-
tual bovine exposure. This is an old means of developing
a known inoculum of infectious pathogens, however, this
method can also result in the development of more viru-
lent pathogens than the original stock solution would
have allowed for. Beyond this complication the ability
to accurately determine the dose of an inoculum can be
difficult and the lack of detail in how the inoculum was
prepared has resulted in the rejection of many candidate
dose response data sets, since their quality could not be
verified.
Considering the limitations and difficulties in conducting
controlled dosing trials, it is relatively simple to see why the
inclusions of other complicating factors are ignored some-
times. However, when the actual infectious process is consid-
ered for the human host, these factors can have a very large
impact in the resulting, infection, disease and consequences
from survival or mortality. If these types of data sets were
developed then either the inclusion of these extraneous fac-
tors, such as age of the host and time post inoculation; were
either not included in the development of new or adaptations
to current dose response models; or they were done so empiri-
cally. This is the impetus of developing advanced mechanistic
dose response, allowing for a generalized form of advanced
dose response models.
Host Age Dependent Advanced
Dose Response Model
The first of the advanced mechanistic dose response models
is based on host age. Host age has been known in clinical
environments to play a significant role in the development
of and severity of some diseases. A classical case is the Influ-
enza pandemic of 1918. The Influenza strain, which resulted
in this pandemic, had a curious means of attack in the
human host. It would replicate in the endothelial cells
throughout the respiratory system, giving it an advantage
over the immune system, allowing for rapid and near immedi-
ate replication upon exposure (22). The healthiest popula-
tion such as those who were being sent off to the war,
would have a higher mortality rate than the those populations
typically considered at a greater risk (the young and elderly).
The 20–40 age group accounted for 60% of death in South
Africa (23) and a mortality rate nearly four times that of
the 41–60 population in Chicago, USA (24). For the war
fighting population, this was likely compounded by the
nature of housing a war fighting force in close living quarters
in temporary or even permanent barracks. The barracks con-
ditions just meant that a wider group could be exposed with
each cough and sneeze before quarantine measure could
take affect for the infected individuals. But the higher mortal-
ity rate is of more interest for this case. The paradoxical result
was seen that given a faster acting and more efficient immune
system, meant that the healthier population was at a disad-
vantage, by being overwhelmed by their own immune system,
allowing for opportunistic infections (such as pneumonias) to
take hold and lead to death (25, 26).
While age dependent dose response data for influenza was
not available we were able to develop a model for Variola
major (V. major causitive agent of smallpox). Smallpox is
known clinically to affect the young and old more severely,
with higher mortality rates and greater likelihood of debilitat-
ing conditions if survived (i.e. blindness). This understanding
leads us obviously to the first of the advanced dose response
models, one that includes the dependency of host age in
the optimization of the dose response parameters; thus result-
ing in a host age dependent dose response models (27). The
framework of developing this first of a new generation of
dose response models is also the guiding framework of the sub-
sequent advanced dose response models.
The entirety of the derivation of the host age advanced
dose response model will not be reviewed here but will be
given more focus than those which base their derivation
on. In essence the framework built for development of this
advanced dose response model is one where the parameter
(s) are made functional forms with host age as the dependent
variable. This is counter to the original form of the the dose
response parameter(s) which are point estimates and are not
dependent on any extraneous information, only dose.
Remember from the optimization of the dose response
parameter(s) that the observed probability of response is a
function of the dose and parameter(s) (Θ). In the case of clas-
sical dose response modeling the parameters are optimized as
static variables to available dose response data. In the case of
age dependent dose response modeling the parameters are
optimized as functions dependent on host age. As can be
seen in the Variola major data in Box 4 there are 4 dose groups
that were isolated by age with each exposed to a set of plaque
forming units (PFU) doses. Each age group is considered a
independent dose response data set and optimized separately,
this leads to the non-linear regression plots seen in Box 4
where the best descriptive regression model is chosen and
used as the age dependent sub model. The age dependent
parameters are then optimized with the combined data set,
to determine the ability to optimize the age dependent dose
response parameters, this developing an optimizable age
dependent dose response model, which is then bootstrapped
to develop confidence intervals for the complete model for all
age groups, show in Figure 2. The age dependent models can
be seen in equations 18 and 20, with the age dependent
parameter models in equations 19, 21 and 22 for the exponen-
tial and beta Poisson models respectively.
PðrÞ ¼ 1  expðk  doseÞ ð18Þ

k ¼ k0  expðk1  ageÞ ð19Þ
  a
dose  21=a  1
PðrÞ ¼ 1  1 þ ð20Þ
N50
a ¼ a0 þ a1 ðageÞ ð21Þ
N50 ¼ N500  expðN501  ageÞ ð22Þ
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-13

BOX 4

TABLE B4.1 Host age-dependent dose-response data for

suckling mice exposed intraperitoneally to Variola major

Mouse
Dose Number Positive Negative
Age
(PFU) of Mice Response Response
(hrs)
0–24 100,000 9 9 0
10,000 9 9 0
1,000 9 94 5
100 9 3 6
10 8 1 7
24–48 1,000,000 11 11 0
100,000 8 8 0
10,000 8 3 5
1,000 8 5 3
100 7 1 6
48–120 2,000,000 9 9 0
200,000 9 7 2
20,000 8 4 4
2,000 7 1 6
120–168 5,000,000 9 8 1
500,000 9 6 3
50,000 8 0 8
5,000 8 0 8

FIGURE B4.1 Age dependent parameter plots, each point rep-

resenting a separate optimization of the dose response data for that
age group; for a) α and b) N50 from the beta Poisson model and c) k
from the exponential model.

The new age dependent parameter models are entered

into the respective dose response models. Then when TABLE B4.2 Assessment of improved fit from key
the dose response models are optimized using these age-dependent parameter models (not all model fit are shown)
functions as the parameters instead of point values,
k0, k1, α0, α1, N50,0, and N50,1, are estimated from Dose-Response
Regression Adjusted
the optimization which are used to calculate k *, α*, Model Conclusion
Model R2
and N *50 Parameter
Beta Poisson a Linear 0.519 Linear is best
k ¼ k0  expðk1  ageÞ Logarithmic 0.180 age-dependent
Inverse 0.147 model for a
Exponential 0.431
a ¼ a0 þ a1 ðageÞ
Beta Poisson N50 Linear 0.485 Exponential is
Logarithmic 0.214 best age
N50 ¼ N500  expðN501 ageÞ Inverse −0.30 dependent
Exponential 0.983 model for N50
FIGURE B4.2 Age dependent dose response model Exponential k Linear 0.337 Exponential is
parameter models. Each parameter from the beta Pois- Logarithmic 0.637 best age
son and exponential models are now models depend- Inverse 0.890 dependent
ent on host age. Exponential 0.976 model for k
3.5.3-14 ▪ MICROBIAL RISK ASSESSMENT

FIGURE 2 Age-dependent dose-response models, for each age group, merged within one large age-dependent model signified with the outer
and inner 99% confidence intervals (27). doi:10.1128/9781555818821.ch3.5.3.f2

The main benefit of this advanced model is by establishing cumulative probability of observing positive responses from
a relatively simple framework for easy adaptation from classi- time t to t + 1 (πi,t). From this a likelihood function can be
cal dose response modeling to advanced dose response mod- developed specific to this cumulative probability distribution
els. Additionally we can now map the effect of a disease or and considering pi,t a positive response probability from time t
pathogen on a greater proportion of the population, since to t + 1 (equation 24). Using this likelihood function a devi-
we are not assuming an age group or other typical host type ance can be calculated (equation 25), which is remarkably
(i.e. average american male). similar to the deviance from traditional dose response models
with the exception of the additional summation which
accounts for time.
Time Post Inoculation Dependent Advanced Dose
Response Model Fðdi ; tt Þ  Fðdi ; tt1 Þ
Based on the framework developed for age dependency, time pi;t ¼ ð23Þ
1  Fðdi ; tt1 Þ
post inoculation (TPI) advanced dose response models,
essentially develop a time line of likely impacts based on Y
t Y
Nd Y Nd 
t Y 
ni;t
the amount of time it has been since the host organism has pi;ti;t ð1  pi;t Þni;t pi;t
p
L¼ Pðpi;t jpi;t Þ ¼
been dosed. Basically the TPI models have time post inocula- pi;t
t¼1 i¼1 t¼1 i¼1
tion as the dependent variable in the dose response parameter
sub models rather than age for the age dependency models. ð24Þ
The original form of this advanced model used the exact Y ¼ 2 lnðLRÞ
same framework as the age dependent advanced model. " ! !#
Thus through optimization of dose response data for each Xt X Nd
pi;t 1  pi;t
time period (typically days) after inoculation of the host ¼ pi;t ln o þ ðni;t  pi;t Þ ln
pi;t 1  poi;t
with; Bacillus anthracis, Francisella tularensis and Mycobacte- t¼1 i¼1
rium tuberculosis; a set of time dependent sub models for the ð25Þ
dose response parameters were developed (28).
This original work was expanded on to develop a more This allows us to use the same coding structure and frame-
generalizable form of this advanced model (28). Using the work as the age dependent dose response model optimization
original probability of time to effect (equation 23) from Hart- code, already developed with minimal adaption and altera-
ley and Seilken (29) to develop a function of probability of a tion. Therefore the optimization is performed just as for
positive response with respect to time post inoculation. This the age dependent model. The time dependent parameter
is a departure from the age dependent advanced model that functions which describe the dose response parameters as
while the raw data is still heavily used, to determine the functions of time post inoculation are optimized and, thereby,
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-15
FIGURE 3 Example of time postinoculation (TPI) dose-response models for Francisella tularensis in the mouse host. Note how as the TPI
increases, the host becomes less susceptible, meaning that more dose is required to have an equitable affect on the host from previous TPI peri-
ods. doi:10.1128/9781555818821.ch3.5.3.f3
optimizing the time dependent parameters. Figure 3 shows an
output from this advanced model, as can be seen as time post
inoculation is increased risk of deleterious effect (in these
cases mortality) increases, which intuitively can be expected
considering that the pathogens have longer to develop an
infection, disease and sufficient body burden for mortality.
There are a number of benefits in developing time post
inoculation models, chiefly among them, most likely is the
use in determining effectiveness of and preparedness for an
outbreak/epidemic. Akin to a timeline, it can be used to
determine nursing and physician rotations to handle the like-
liest time when a large number of infected patients will
present to the hospitals and other medical locations. This
use is currently severely hampered for the current class of
TPI due to the models being for a lethal infection. Therefore
for greatest use these models need to develop for data sets that
do not use mortality as the endpoint.
Physiological and Pathogen Dynamics
While the previous two advanced dose response models were
examples of expansions from traditional dose response mod-
els; they demonstrated the ability to adapt currently accepted
dose response models to host age and TPI factors. These fac-
tors were relatively simple to develop; in that the data was
available for derivation of the factor dependent sub models
for the dose response parameters and inclusion into the
dose response models in such a was as to allow for their opti-
mization with the dose response models. Alternatively effects
from a larger issue, host physiology and pathogen kinetics
need to be modeled under a different framework.
In essence this type of model is analogous to physiologi-
cally base pharmacokinetic (PBPK) models. PBPK models
have helped revolutionize risk assessment for chemical and
radiological exposures. Through PBPK modeling the metab-
olism and processing of hazardous compounds from exposure
to excretion are modeled, thereby, allowing to more com-
pletely estimating the risk to the exposed population. PBPK
models have been shown to also aid in interspecies extrapola-
tion (30, 31) and an improved means of assessing differences
in certain sensitive populations such as the elderly (32)
and children (33). The theory is that physiologically based
microbial dose response models would have a similar effect
on QMRA, expanding our ability to model more realistic
exposures.
There are currently two main schools of thought in the
development of these physiologically based advanced dose
response models; but interestingly working completely
separately, both focused on inhalation of Bacillus antrhacis
(B. anthracis, causative agent of Anthrax). An impressive
amount of data has been developed, and postulated potential
modeling approaches (34–36) using the rabbit lung as a basis
for their model. An alternative theory has been postulated
and a complete model has been outlined (37) and how to
combine the two essential systems; transport through the res-
piratory system and pathogen kinetics for the human host;
with a preference for known human lung transport dynamics.
The general thought behind this type of model is to take the
dose response models from utilizing exposed dose to deliver
and effective doses. Delivered dose is defined as that dose
which is transported with the bulk fluid (air in the case of
inhalation, water and food for gastrointestinal) to a location
of potential infection development. Effective dose is defined
as that amount of pathogens that reach a location conducive
for infection development and progresses through its patho-
genesis processes to initiate that infection, thus potentially
becoming an increased dose than the delivered dose.
Further details on this new generation of dose response
models can be found in the establishing works (5, 34, 37).
They will not be delved into greater detail here, as neither
approach is appropriate for use in QMRA modeling in their
present form. The main limitation is the inability to validate
a complete model. The approach being developed by (34) is
hampered by a lack of details in the further infection dynam-
ics and data for these dynamics in rabbits. The approach used
in (37) has a similar limitation where the model cannot be
validated for human, a greater limitation than (34), as for
the rabbit model this required data can be developed. Even
though this type of modeling is not ready for use in QMRA
3.5.3-16 ▪ MICROBIAL RISK ASSESSMENT
modeling, this type of dose response model show promise for
greater descriptive power and ability to model the real patho-
genesis rather than a much simplified one.
Limitations and Benefits of Advanced Dose
Response Models
The physiologically dependent model has two basic limi-
tations to its use. First, given the high risk of an inhalation
exposure to B. anthracis and unfortunately considering the
experiences of malicious spore release, a response strategy is
quite well in hand then for its desire for use may be severely
limited. The second limitation is that these models need to
have a specific set of animal experiments, which must be per-
formed to validate these models. The greatest value of the
alternative approach taken by (34) is in the concurrent
experiments being performed. This framework is presented
to advise where the future of dose response modeling may
and in specific instances is trending. The main benefits of
developing these physiologically dependent dose response
models is that once the framework has been established, it
can be relatively easily adapted to additional exposure routes
and pathogens. Additionally just as PBPK models for chemi-
cals and radiological agents advanced our understanding of
the dynamics of their toxic effects and resulting risks, physio-
logically dependent models may have the same effect in
QMRA. Thus potentially answering some of the uncertainty
questions which arise given large exposure events, and the
un-equitable results considering everyone’s baseline
exposure.
The advance models for age and time dependencies are
good examples of utilizing a simple modeling framework to
account for additional factors in infectious disease. Develop-
ing these types of advanced dose response models is a logical
expansion of dose response modeling in the future, to develop
more accurate, descriptive and expandable QMRA models.
The data being from the original animal experimentation tri-
als makes these models more acceptable for use and likely to
have validation data sets available or producible. Given these
two advanced models we can now quantify the differences
in response for a spread of age groups. This allows us now to
have a model which can describe children specifically instead
of using one meant for adults and trying to interpolate to a
younger host group, with different sensitivities. The time
dependent models will allow us the potential to determine
almost a schedule to outbreaks. Being able to predict a patient
load to local hospitals and medical care centers, will allow us
to respond to natural and malicious incidences much better;
with respect to staffing and supplies management.
The future of dose response modeling can be thought of as
an uncertain one, considering the specialized nature and costs
associated with the science. However with the wealth of his-
torical data from years of previous studies there is continual
potential for enlightening current models with more data,
and develop new optimized models.
REFERENCES
1. U.S. EPA. 2011. Exposure Factors Handbook 2011 Edition
(Final). U.S. Environmental Protection Agency, Washing-
ton DC, EPA/600/R-09/052F.
2. Dufour AP, Evans O, Behymer TD, Cantu R. 2006. Water
ingestion during swimming activities in a pool: a pilot study.
J Water Health 4(4):425–430.
3. Rose JB. 1997. Environmental ecology of cryptosporidium
and public health implications. Annu Rev Public Health
18:135–161.
4. Buchanan RL, Whiting RC. 1998. Risk assessment: a
means of linking HACCP plans and public health. J Food
Prot 61(11):1531–1534.
5. Weir MH, Haas CN. 2011. A model for in-vivo delivered
dose estimation for inhaled bacillus anthracis spores in
humans with interspecies extrapolation. Environ Sci Technol
45(13):5828–5833.
6. Haas CN. 1983. Estimation of Risk Due to Low Doses of
Microorganisms: A Comparison of Alternative Methodolo-
gies. Am Jounral Epidemiol 118(47):573–582.
7. Teunis PF. 1997. Infectious gastro-enteritis - opportunities
for dose response modeling. National Instiute of Public
Health and the Environment, Bilthoven, The Netherlands,
284 550 003.
8. Haas CN, Rose JB, Gerba CP. 2014. Quantitative Microbial
Risk Assessment, Second. John Wiley and Sons.
9. DeKievit TR, Iglewski BH. 2000. Bacterial quroum sensing
in pathogenic relationships. Infect Immun 68(9):4839–4849.
10. Antunes LCM, Ferreira RBR, Buckner MMC, Finlay BB.
2012. Quorum sensing in bacterial virulence. Microbiology
156:2271–2282.
11. Teunis PF, Nagelkerke NJD, Haas CN. 1999. Dose
response for infectious gastroenteritis. Risk Anal 19(6):
1251–1260.
12. Bartrand TA, Weir MH, Haas CN. 2008. Dose-response
models for inhalation of bacillus anthracis spores: interspe-
cies comparisons. Risk Anal 28(4):1115–1124.
13. Furumoto WA, Mickey RA. 1967. Mathematical model for
the infectivity-dilution curve of tobacoo mosaic virus. Virol-
ogy 32:216–223.
14. Moran PAP. 1954. The dilution assay of viruses. J Hyg
(Lond) 52:189–193.
15. Teske SS, Huang Y, Tamrakar SB, Bartrand TA, Weir
MH, Haas CN. 2011. Animal and human dose-response
models for brucella species. Risk Anal 31(10):1576–1596.
16. Weir MH, Razzollini MTP, Rose JB, Masago Y. 2011.
Water reclamation redesign for reducing cryptosporidium
risks at a recreational spray park using stochastic models.
Water Res 45(19):6505–6514.
17. Coulliette AD, Enger KS, Weir MH, Rose JB. 2013. Risk
reduction assessment of waterborne Salmonella and Vibrio
by a chlorine contact disinfectant point-of-use device. Int J
Hyg Environ Health 216(3):355–361.
18. Fischer Walker CL, Rudan I, Liu L, Nair H, Theodoratou
E, Bhutta ZA, O'Brien KL, Campbell H, Black RE. 2013.
Global burden of childhood pneumonia and diarrhoea. The
Lancet 381(9875):1405–1416.
19. deWitt MAS, Koopmans MPG, van Duynhoven YTHP.
2003. Risk factors for norovirus, apporo-like viruses, and
group a rotavirus gastroenteritis. Emerg Infect Dis 9(12):
1563–1570.
20. Lopman BA, Vennema H, Kohli E, Pothier P, Sanchez A,
Negredo A, Buesa J, Schreier E, Reacher M, Brown D,
Gray J, Iturriza M, Gallimore C, Bottiger B, Hedlund K-
O, Torven M, von Bonsdorff C-H, Manula L, Poljsak-Pri-
jatlj M, Zimsek J, Reuter G, Szucs G, Melegh B, Svennson
L, von Dujnhoven Y, Koopmans MPG. 2004. Increase in
Viral Gastroenteritis Outbreaks in Europe and Epidemic
Spread of New Norovirus Variant. The Lancet 363:682–688.
21. Tamrakar SB, Haas CN. 2008. Dose-response model for
Burkholderia pseudomallei (melioidosis). J Appl Microbiol
105:1361–1371.
22. Taubenberger JK, Hultin JV, Morens DM. 2007. Discov-
ery and characterization of the 1918 Pandemic Influenza
virus in historical context. Antivir Theory 12(4 pt B):581–
591.
 3.5.3. Dose-Response Modeling and Use: Challenges and Uncertainties in Environmental Exposure ▪ 3.5.3-17
23. Katzenellenbogen JM. 1988. The 1918 influenze epidemic
in Mamre. South Afr J Med 74:362–364.
24. Van Hartesveldt FR. 1992. In Pandemic of Influenza: The
Urban Impact in the Western World. Edwin Mellen Press, Lew-
iston, NY, pp. 121–144.
25. LeCount E. 1919. The pathologic anatomy of influenzal
bron-chopneumonia. J Am Medial Assoc 72:650–652.
26. Wolbach S. 1919. Comments on the pathology and bacteri-
ology of fatal influenza cases, as observed at camp Devens,
Mass. Johns Hopkins Hosp Bull 30:104–109.
27. Weir MH, Haas CN. 2009. Quantification of the effects of
age on the dose response of variola major in suckling mice.
Hum Ecol Risk Assess Int J 15 (6):1245–1256.
28. Huang Y, Bartrand TA, Haas CN, Weir MH. 2009.
Incorporating time postinoculation into a dose-response
model of yersinia pestis in mice. J Appl Microbiol 107:
727–735.
29. Hartley HO, Sielken RL. 1977. A biometrics invited paper.
estimation of ‘Safe Doses’ in carcinogenic experiments. Bio-
metrics 33:1–30.
30. Nestorov I. 2007. Whole body physiologically based phar-
macokinetic models. Drug Metab Toxicol 3 (2):235–244.
31. Maruyama W, Hirano S, Kobayashi T, Aoki Y. 2006.
Quantitative risk analysis of particulate matter in air inter-
species extrapolation with bioassay and mathematical mod-
els. Inhal Toxicol 18 (13):1013–1023.
32. Clewell HJ, Gentry PR, Convington TR, Sarangapani R,
Teeguarden JG. 2004. Evaluation of the Potential Impacts
of Age-and-Gender-Specific Pharmacokinetic Differences
in Tissue Dosimetry. Toxicol Sci 79 (2):381–393.
33. Bjorkman S. 2005. Prediction of drug disposition in infants
and children by means of Physiologically Based Pharmacoki-
netic (PBPK) Modelling: theophylline and midazolam as
model drugs. Br J Clin Pharmacol 59:691–704.
34. Gutting BW, Nichols TL, Channel SR, Gearhart JM,
Andrews GA, Berger AE, Mackie RS, Watson BJ, Taft
SC, Overheim KA, Sherwood RL. 2012. Inahaltional
anthrax (Ames aerosol) in naive and vaccinated New Zealand
rabbits: characterizing the spread of bacteria from lung deposi-
tion to bacteremia. Front Cell Infect Microbiol 2(87):12.
35. Gutting BW, Channel SR, Berger AE, Gearhart JM,
Andrews GA, Sherwood RL, Nichols TL. 2008. Mathe-
matically Modeling Inhalational Antrhax - Successive
Advances in Modeling Applied to Updated Data Help in Esti-
mating Risks from Inahaltional Anthrax. Microbe 3(2):78.
36. Gutting BW, Gaske KS, Schilling AS, Slaterbeck AF,
Sobota L, Mackie RS, Buhr TL. 2005. Differential suscept-
ibility of macrophage cell lines to bacillus anthracis - Vollum
1B. Toxicol In Vitro 19(2):221–229.
37. Weir MH. 2009. Development of Physiologically Based
Pathogen Transport and Kinetics Model for Inhalation of
Bacillus anthracis Spores. Drexel University, Philadelphia, PA.
Microbial Ecology
VOLUME EDITOR: ROBERT V. MILLER
SECTION EDITORS: LARRY J. FORNEY, ROBERT H. FINDLAY, BRIAN P. HEDLUND, AND
JULIAN R. MARCHESI
4.1 THEORY
4.1.1 Phylogenomic Networks of Microbial
Genome Evolution / 4.1.1-1
TAL DAGAN, OVIDIU POPA, THORSTEN
KLÖSGES, AND GIDDY LANDAN
4.1.2 Evolutionary Ecology of Microorganisms:
From the Tamed to the Wild / 4.1.2-1
JAY T. LENNON AND VINCENT J. DENEF
4.2 AQUATIC ENVIRONMENTS
4.2.1 The Microbial Ecology of Benthic
Environments / 4.2.1-1
ROBERT H. FINDLAY AND TOM J. BATTIN
4.2.2 Heterotrophic Planktonic Microbes: Viruses,
Bacteria, Archaea, and Protozoa / 4.2.2-1
JED A. FUHRMAN AND DAVID A. CARON
4.2.3 Aquatic Biofilms: Development, Cultivation,
Analyses, and Applications / 4.2.3-1
JOHN R. LAWRENCE, THOMAS R. NEU,
ARMELLE PAULE, DARREN R. KORBER, AND
GIDEON M. WOLFAARDT
4.3 EXTREME ENVIRONMENTS
4.3.1 The Microbiology of Extremely Acidic
Environments / 4.3.1-1
D. BARRIE JOHNSON AND ANGELES
AGUILERA
4.3.2 Life in High Salinity Environments / 4.3.2-1
AHARON OREN
4.3.3 Microbial Life in Extreme Low-Biomass
Environments: A Molecular
Approach / 4.3.3-1
KASTHURI VENKATESWARAN, MYRON T. LA
DUC, PARAG VAISHAMPAYAN, AND JAMES
A. SPRY
4.3.4 Life in High-Temperature
Environments / 4.3.4-1
BRIAN P. HEDLUND, SCOTT C. THOMAS,
JEREMY A. DODSWORTH, AND CHUANLUN
L. ZHANG
4.4 ANIMAL-GUT MICROBIOMES
4.4.1 Invertebrate Gut Associations / 4.4.1-1
DANIELE DAFFONCHIO, ALBERTO ALMA,
GUIDO FAVIA, LUCIANO SACCHI, AND
CLAUDIO BANDI
4.4.2 Studying the Mammalian Intestinal
Microbiome Using Animal
Models / 4.4.2-1
FLOOR HUGENHOLTZ, JING ZHANG,
PAUL W. O’TOOLE, AND HAUKE SMIDT
4.4.3 Animal Gut Microbiomes / 4.4.3-1
RICHARD J. ELLIS AND CHRISTOPHER S.
MCSWEENEY
 Phylogenomic Networks of Microbial Genome Evolution
TAL DAGAN, OVIDIU POPA, THORSTEN KLÖSGES, AND GIDDY LANDAN
4.1.1
Methods for the study of microbial phylogenetics can be
roughly classified by the breadth of genomic information
that is included in the analysis, ranging from a single marker
gene to the analysis of whole genomes. The most commonly
used marker gene for microbial phylogeny is the 16S rDNA
gene that encodes for an rRNA component of the prokarty-
otic small ribosomal subunit. Using the 16S gene as a marker
supplies an efficient and informative representation of micro-
bial phylogeny in many cases because it is universally distrib-
uted across the three domains of life and its gene sequence is
relatively conserved so that it can be compared between dis-
tantly related organisms (1, 2). However, the use of a single
conserved marker gene for microbial phylogenetics is very
much limited for the study of closely related strains as micro-
bial species having similar 16S rRNA gene sequences may
encode for very different proteomes (3). For example the
genomes of different Escherichia coli strains that are almost
identical in their 16S rRNA gene may have only 40% genes
in common (4, 5). The discrepancy between the phyloge-
netic signal observed in marker genes and overall genomic
properties stems from the very different processes that under-
lie their evolution. While the 16S rRNA is an essential gene
that evolves mostly by descent with modification, whole
microbial genomes evolve by lateral (or horizontal) gene
transfer (LGT or HGT) (6–8). Accumulating evidence shows
that LGT plays a major role in prokaryote genome evolution
(9–12), affecting virtually all genes (13–16). In this chapter,
we review the currently recognized mechanisms for gene
transfer among microbial species and the existing methods
for whole genome phylogenetic reconstruction.
LGT MECHANISMS
The common mechanisms for LGT include transformation,
conjugation, and transduction. In addition, more taxa-
specific mechanisms include gene transfer agents (GTAs),
cytoplasmic bridges, nanotubes, and outer membrane vesicles
(OMVs).
Transformation
The concept of transformation, albeit without a supporting
biological mechanism, has been known since 1928 when
Griffith demonstrated that it is possible to transform an atte-
nuated and nonencapsulated Pneumococcus (type R) into a
fully encapsulated and virulent Pneumococcus cells (type S)
doi:10.1128/9781555818821.ch4.1.1
4.1.1-1
(17). In his experiment, Griffith injected mice with a small
amount of living type R culture together with a large amount
of heat-killed type S culture. The injected mice frequently
succumbed to infection, and their blood samples yielded
pure cultures of S-type Pneumococcus. Only in 1944 did Avery
et al. (18) reveal that the transforming material in Griffith’s
experiment was DNA. In their experiment they centrifuged
the heat-killed type S culture. They then incubate the
R-type culture separately with each of the fractions and
injected it to mice. They found that only one fraction trans-
formed the R-type culture into a virulent S-type Pneumococ-
cus as in Griffith’s experiment—the one that contained
deoxyribonucleotide acids. From this result they concluded
that DNA was the transforming material of Pneumococcus
type R.
Transformation involves the uptake of naked DNA from
the environment (Fig. 1a). The uptake of raw DNA in trans-
formation is enabled during a competence state that involves
20–50 proteins, including the type IV pillus and type II secre-
tion system proteins (20, 21). Experiments in DNA uptake of
fluorescence-labeled DNA in Helicobacter pylori revealed that
the length of imported DNA fragments is on average 10 kb,
and these are imported at a mean velocity of 1,230 bp per sec-
ond (22).
In some species, an effective transformation requires the
presence of uptake signal sequences (USSs; called also
DUS, DNA uptake signal). These are specific DNA motifs,
about 10 bp long, that are encoded within the recipient
genome in a frequency that is much above that expected by
chance (23). The most commonly known examples for spe-
cific uptake motifs are those of Haemophilus influenza—USS
50 -AAGTGCGGT-30 (24), and Neisseria meningitides—
DUS 50 -GCCGTCTGAA-30 (25). Environmental DNA
molecules bearing the USS motif are recognized by specific
receptors at the cell surface. In N. meningitidis specific recog-
nition of DNA is enabled by a specific DNA binding site at
the minor pillin protein (ComP) that is a component of the
type IV pili system (26).
Natural competence and transformation may depend on
environmental cues that are intercepted by the quorum sens-
ing system. In Vibrio cholerae, the regulation of competence
genes has been shown to be partially dependent on the pres-
ence of chitin, which is typically found in zooplankton-rich
environments (27, 28). The sensing of chitin degradation
products in V. cholerae initiates a regulatory pathway that
4.1.1-2 ▪ THEORY

Transformation
(a)

Conjugation

(b)

(c)
Transduction

(d)
Gene transfer agents

FIGURE 1 Lateral gene transfer mechanisms mechanisms. (a) Transfer of raw DNA during transformation. (b) Plasmid transfer during con-
jugation. (c) DNA transfer by phages during transduction. (d) LGT by gene transfer agents (GTAs). (adapted from 19)
doi:10.1128/9781555818821.ch4.1.1.f1

leads to the expression of competence genes and transforma-
tion by DNA uptake (29). DNA molecules that are imported
into the cytoplasm via transformation may be integrated into
the recipient chromosomes, usually via homologous recom-
bination (20, 21, 30, 31). As the success of homologous
recombination depends on the nucleotide sequence similar-
ity between the imported DNA fragment and the recipient
chromosome, it is expected that most gene acquisition by
transformation will occur between closely related species,
having similar genomes. Transformed fragments encoding
site-specific recombination mechanisms (e.g., transposons)
are an exception as they can be integrated into the recipient
chromosome without relaying on donor-recipient sequence
similarity. An experiment where naturally transformable cells
of Acinetobacter baylyi were exposed to DNA from various
integron-carrying genera showed that the integrase activity
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-3

could facilitate gene acquisition in A. baylyi from distantly The percolation of an acquired DNA within the population
related species, such as Salmonella enterica (32). can be extremely fast in species where the cells are arranged
in chains, such as Bacillus subtilis. Tracking the spread of a
Conjugation green fluorescence protein–labeled ICE under the micro-
Gene recombination as a result of direct DNA transfer scope showed that in 43 (81%) out of 53 cases a recipient
between cells was discovered by Lederberg and Tatum in cell turned into a donor and transconjugated the ICE to the
1946 (33, 34). In their experiment, they used two E. coli next cell in line, often within 30 min (40).
knockout strains that were deficient in their ability to synthe- DNA transfer by conjugation is governed by a complex
size essential nucleic and amino acids as well as biotin. One regulatory network that may be activated or repressed in
strain could not synthesize biotin, phenylalanine, and cys- response to various environmental stimuli. For example, in
teine. The other strain was deficient in its ability to produce Enterecoccus faecalis, the conjugative transfer of pCP10 plas-
threonine, leucine, and thiamin. Neither strain could survive mid is induced by a pheromone that is expressed and secreted
when plated on minimal medium that lacked those essential by recipient cells. The pheromone, cCP10, binds to the pro-
components. Tatum and Lederberg mixed the two auxotro- tein that is a receptor for the pheromone encoded by the
phic strains, incubated them for 24 to 48 h, and plated pCP10 plasmid and PrgZ activates the plasmid replication
them on minimal medium. To their surprise, they observed and conjugative transfer (41, 42). In E. coli, the transfer of
viable colonies of prototroph. Moreover, they were able to SXT ICE is induced by an SOS response to DNA damage.
characterize additional strains that were lacking the ability Under normal conditions the SXT excision is repressed by
to synthesize only one essential organic compound (e.g., thia- the SetR protein that is encoded within the SXT. During
mineless, leucineless). These results led them to conclude an SOS response, SetR is cleaved, which leads to the allevia-
that the two strains recombined to produce multiple new tion of expression suppression on the SXT genes that are then
strains that encode different combinations of the genes in expressed and mediate the transfer of this ICE (43).
the two ancestral strains. To rule out the possibility that
this was a result of transformation or any other transfer via Cytoplasmic Bridges and Nanotubes
the media, they centrifuged one strain and mixed the filtered An archaeal mating system where DNA is transferred via
media (lacking cells) with the other strain. This experiment cytoplasmic bridges was documented in the haloarchaeon
did not yield viable colonies of prototrophs. Their final con- Haloferax volcanii by Mevarech and colleagues in the mid-
clusion thus stated that bacteria are able of recombination 1980s (44, 45) (Fig. 2a).
that is mediated by an unknown mechanism (34). Early experiments revealed that a plasmid encoding for
Direct DNA transfer between cells is termed conjugation selectable properties could be transferred between H. volcanii
(Fig. 1b). This DNA transfer mechanism is mediated by a pro- cells. Like in conjugation, the transfer required that the cells
teinaceous cell-to-cell junction, forming a tunnel that con- be viable and in physical contact, yet unlike conjugation, the
nects the donor and recipient cells through which the DNA transfer was bidirectional, where the cells could func-
DNA is transferred (20, 21). DNA transferred during conju- tion either as a donor or a recipient (44). Scanning electron
gation typically comprises conjugative plasmids—mobile cir-
cular DNA elements—that encode the proteins required for
the conjugation tunnel formation and DNA replication.
The plasmids are transferred in a “copy-paste” mechanism
where at the end of the transfer both donor and recipient har-
bor an identical copy of the plasmid (35). Mobile plasmids
encoding only the DNA replication machinery can be also
transferred by conjugation when the proteins required for
the tunnel are encoded in the chromosomes or coresident
plasmids (35). Additional DNA elements that are transferred
by conjugation are integrative conjugative elements (ICEs).
These are mobile genetic elements whose genetic repertoire
includes the genes required for conjugation, excision from
the donor genome, and integration into the recipient genome
(36, 37). In contrast to conjugative plasmids, ICEs are usually
transferred to the recipient in a “cut and paste” mechanism
(36, 38). Genetic elements transferred by conjugation may
be integrated into the recipient chromosomes. Plasmids are
commonly integrated by homologous recombination that
may entail insertion sequences or other sequences conserved
between plasmid and recipient chromosomes that carry the
minimal sequence similarity required for homologous recom-
bination (21). The integration of ICEs is usually facilitated by
integrase-mediated site-specific recombination (36).
The success rate of gene acquisition by conjugative plas-
mids was estimated in E. coli under laboratory conditions
(39). Transferred DNA was labeled by a yellow fluorescence
protein, and the odds of successful integration of plasmid FIGURE 2 Intercellular DNA transfer. (a) Intercellular bridges
DNA into the recipient genome were measured by time-lapse connecting Halobacterium volcani cells (adapted from 45). (b) Nano-
microscopy. This experiment revealed that labeled plasmid tubes connecting Staphylococcus aureus (PY79) and Escherichia coli
DNA was integrated into the chromosome in 96% of the (MG1655) cells (adapted from 46).
recipients and inherited by the next generation (39). doi:10.1128/9781555818821.ch4.1.1.f2
4.1.1-4 ▪ THEORY
microscopy revealed intercellular bridges between the cells
where the membrane and envelope of the connected cells
fuse to form what seems to be a cytoplasmic continuity.
The bridge length was estimated to reach up to 2 μm and their
diameter was up to 0.1 μm (45). A recent experiment where
H. volcanii was mated with the closely related species H. med-
iterranei revealed a high frequency of interspecific recombina-
tion between the two strains (47). Genome sequencing of
several hybrids that resulted from this mating revealed multi-
ple DNA transfers from H. volcanii to H. mediterranei of
fragments ranging from 301 kb to 530 kb. The acquired frag-
ments recombined frequently into the recipient genome at
loci that shared nearly 100% identical nucleotide sequence
similarity between the donor and recipient, such as the
genomic neighborhood of tRNA genes (47). This suggests
that DNA acquired by Haloarchaea via the cytoplasmic
bridges is integrated into the recipient genome by homolo-
gous recombination.
A similar transfer mechanism was recently discovered in
the eubacterial domain and was named nanotubes (46)
(Fig. 2b). These are tubular protrusions composed of mem-
brane components that can bridge between neighboring cells
and conduct the transfer of DNA and proteins. The nano-
tubes are between 30 and 130 nm wide and up to 1 μm
long. The tube dimension is correlated with the distance
between the connected cells. Proximal cells are commonly
connected by several small nanotubes, while thicker tubes
connect distant cells. The rate of transfer via the nanotubes
correlates negatively with the size of the transferred sub-
stance. Cellular interconnection mediated by nanotubes is
not species specific. However, morphology and diameter of
the tubes seem to depend on characteristics of the connected
cells (46).
Transduction
Transduction is DNA acquisition following a phage infection
(48) (Fig. 1c). This DNA transfer mechanism was discovered
by Zinder and Lederberg during a study of recombination in
Salmonella strains (48). In their study, they found that unlike
in previous recombination experiments performed with
E. coli showing bidirectional exchange of genetic material,
recombination in Salmonella was unidirectional, where a
single strain (Salmonella typhimurium LT22) was extremely
successful in transferring marker phenotypes to other Salmo-
nella strains. Later on they demonstrated that no physical con-
tact was required for the recombination using a U-shaped tube
in which the donor and recipient were cultured in common
medium but at opposite ends that were separated by a filter
dense enough to block the passage of cells. That approach
revealed small particles that were suspected as the recombina-
tion agents, later identified as bacteriophage PLT-22 (see [49]
for a detailed perspective).
Temperate (or lysogenic) phages multiply via the lyso-
genic cycle, which is established by an integration of the
phage genome into the host chromosomes, creating a pro-
phage within the host genome. The phage typically remains
dormant within the host and is replicated with the host until
the lytic cycle is induced. In the lytic cycle new phages are
produced using the host metabolism and are released during
the host cell lysis. The excision of phage DNA from the
host genome and the production of phages may be accompa-
nied by packing of host DNA into the phages, which can then
transfer it to the next host (50). Specialized transduction
occurs when the phage integrases cleave, in addition to the
prophage, bacterial genes that are encoded at the prophage
flanking regions. These are packed with the phage DNA
into the phages. Generalized transduction occurs when ran-
dom bacterial DNA is packed into the phages (51, 52).
The frequency of gene acquisition via transduction within
pure cultures of marine bacteria has been estimated to range
between 1.33 × 10−7 and 5.13 × 10−9 transduced cells per
plaque-forming unit (PFU). Gene acquisition was observed
also in colonies that did not form a plaque; however, these
occurred at a significantly lower frequency ranging between
6.8 × 10−10 and 2.6 × 10−11 transducants per colony-forming
unit (CFU). Applying the same approach on a mixed
marine microbial population yielded similar ranges between
1.58 × 10−8 and 3.7 × 10−8 transduced cells per PFU (53).
Nevertheless, applying more sensitive methods to detect
transferred marker gene sequences yielded frequencies that
are up to five orders of magnitude higher (54). A recent study
of transduction rates of three E. coli phages, including P1,
T4, and EC10, in freshwater microbial communities revealed
gene acquisition frequencies that ranged between 0.3 × 10−3
and 8 × 10−3 transducants/PFU. A total of 20% from among
the transduced cells were confirmed to be viable. This study
further revealed the occurrence of gene acquisition in non–
plaque-forming Enterobacteriaceae recipients demonstrating
that the transfer range of certain phages can be broader
than their host range (54).
GTAs
GTAs are phage-like DNA vehicles that are produced
by donor cells and released to the environment (55–58)
(Fig. 1d). GTAs were described in 1974 by Marrs, who
observed small phage-like particles that were released into
the medium by Rhodopseudomonas capsulatus strains and could
mediate the transfer of antibiotic resistance genes to nonresis-
tant R. capsulatus strains. As the newly discovered particles
were significantly smaller than any known phage and did
not induce plaque formation or cell lysis, they were recognized
as a novel gene transfer mechanism (55). The proteins
required for GTA synthesis are encoded within an operon
of 15–17 genes. Several genes resemble typical phage genes,
hence GTAs might be related to phages (57). The DNA
stored in GTAs is imported into the recipient in a generalized
transduction process mediated by the cellular RecA recombi-
nation system (59). The mechanism of DNA packing and
capsule release from the cell is still unknown. GTA systems
have been documented in the Deltaproteobacteria Desulfovi-
brio desulfuricans (60) and the Euryarchaeota Methanococcus
voltae (61).
OMVs
Many microbial species secrete OMVs that are used for vari-
ous extracellular functions as well as intercellular communi-
cation (62) (Fig. 3). Earlier reports implying a role of
OMVs in DNA transfer were published by Kahn and col-
leagues, who focused on the study of DNA transfer mecha-
nisms in Haemophilus influenzae (63, 66). Earlier studies of
H. influenzae showed that this bacterium is naturally compe-
tent and is capable of DNA acquisition (65), however, it was
not clear how the transferred DNA is protected from restric-
tion enzymes during the transfer from donors to recipients. By
using electron microscopy Kahn et al. (66) discovered small
membrane vesicles that were attached to the recipient cell
membrane. Later studies revealed that double-stranded
DNA is transferred within the membrane vesicles that protect
it during the transfer (63). These OMVs were named transfor-
masomes (63).
Accumulating evidence suggests that OMVs in various
microbial species play an important role in lateral gene
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-5
FIGURE 3 Outer membrane vesicles. (a) A Pseudoalteromonas marina cell in the process of releasing OMVs by budding (adapted from 64).
(b) an OMV released by the archaebacterium Thermococcus gammatolerans. doi:10.1128/9781555818821.ch4.1.1.f3
transfer. For example, Dorward et al. (67) centrifuged a sam-
ple of Neisseria gonorrhoeae and mixed the fraction containing
OMVs (which they called membrane blubs) with a set of
potential recipients. The result of this experiment showed
that a plasmid carrying a marker encoding for streptomycin
resistance was transferred and conferred antibiotic resistance
to the recipients at an average success rate of 6.5 × 10E-3
transformants per CFU. Transferred DNA was, however,
rarely integrated into the chromosome (67). Gene transfer
by OMVs was reported also for enteric bacteria where
OMVs mediated the transfer of genes encoding for green flu-
orescence protein, virulence factors, and antibiotic resistance
from Escherichia coli O157:H7 to Salmonella enterica serovar
Enteritidis and E. coli JM109 (68, 69). The transformation
success rate of antibiotic resistance genes encapsulated in
OMVs between various strains of the Gammaproteobacte-
rium Acinetobacter baumanii has been shown to be close to
100%. Furthermore, the transformation was shown to be
strictly dependent on the integrity of the OMVs (70). Recent
studies revealed that OMVs may compose either one mem-
brane layer or a double membrane layer (64, 71); the latter
type is much rarer and makes up about 0.1% of the total
OMVs (71). The DNA capacity of OMVs has been estimated
as ∼20 kb in the marine alphaproteobacterium Ahrensia kie-
lensis (64) and ∼600 bp in the arctic gammaproteobacterium
Shewanella vesiculosa M7T (71). DNA bearing OMVs have
been reported also in the archaeal domain. The hyperthermo-
philic archaeon Thermococcus kodakaraensis produces OMVs
that can carry a shuttle plasmid and transform plasmid-
lacking cells (72) (Fig. 3).
MICROBIAL PHYLOGENOMICS
The evolutionary history of a species is most commonly
depicted as a bifurcating phylogenetic tree comprising
nodes and branches. The external nodes in the tree corre-
spond to contemporary species while the internal nodes
correspond to ancestral species. The branches represent ver-
tical inheritance, linking ancestors with their descendants.
The accumulation of fully sequenced genomes since the
early 2000s has enabled the practice of phylogenomics,
that is, the study of phylogenetic relationships at the whole
genome level (73). The evolutionary reconstruction of gene
phylogenies from many genomes at once allows a more
accurate reconstruction of evolutionary events such as
gene loss, gene gain, and gene duplication (74). However,
the widespread occurrence of LGT means that a tree model
that takes only vertical inheritance into account fits only a
fraction of the bacterial genomic repertoire (75–77). The
most natural generalization and alternative to trees are net-
works (78–80).
Networks
A network (or a graph) is a mathematical model of pairwise
relations among entities. The entities (vertices or nodes) in
the network are linked by edges representing the connections
or interactions between these entities. In a coauthorship net-
work, for example, the vertices signify scientists and the edges
represent common scientific publications to the scientists
they connect with (81). In an aviation network, airports
around the world are connected by direct flights from one
connected airport to the other (82). Network approaches
are common in almost all fields of science, including social
sciences, cell biology, ecology, and statistical physics. The
network model supplies an abstract representation of an
entire system, enabling research on the unifying principles
that could explain complex relations among entities. Hence
the most basic issues in networks science are structural (83).
The network properties and connection pattern can teach
us about the topology, dynamics, and development of the
modeled system (83–85).
The information in unweighted networks is limited to
whether the vertices are connected (Fig. 4a). Vertex connec-
tivity (or the degree of a vertex) is the number of vertices con-
nected to the vertex. In a weighted network the edges can
also have a certain weight that signifies the strength of the
connection between the vertices. Vertex connectivity in a
weighted network is calculated as the total edge weight of
4.1.1-6 ▪ THEORY

(a) i ii
a 3 i i a b c d e
j a b c d e j
b
a 0 1 0 0 1 a 0 3 0 0 2
1
b 1 0 1 1 1 b 3 0 1 2 1
2 1
2
c c 0 1 0 0 0 c 0 1 0 0 0
d 0 1 0 0 1 d 0 2 0 0 4
e 4 e 1 1 0 1 0 e 2 1 0 4 0
d Ci 2 4 1 2 3 Wi 5 7 1 6 7

(b) i ii
a 3 i i a b c d e OUT
b j a b c d e OUT j
a 0 1 0 0 1 2 a 0 3 0 0 2 5
1
b 0 0 1 1 0 2 b 0 0 1 2 0 3
2 1
2 c 0 0 0 0 0 0 c 0 0 0 0 0 0
c
3 d 0 0 0 0 1 1 d 0 0 0 0 1 1
e e 0 1 0 1 0 2 e 0 1 0 3 0 4
d IN 0 2 1 2 1 IN 0 4 1 5 3
1
Betweenness
(c) Centrality
3 3 high
2 a 3
b f
g

c
5
e
3 h 1
d
3 i
3
Module A Module B

low

FIGURE 4 An introduction to networks. (a) A network composed of vertices (circles) and edges (lines). (i) An unweighted network of n
vertices can be fully defined by a matrix, A = [aij ]n*n, with aij = 1 if an edge is connecting between vertex i and vertex j, and aji = 0 otherwise.
Vertex centrality (Ci) is calculated as the sum of vertices linked to the vertex. (ii) A weighted matrix representation of the network. Cells of
connected vertices i and j contain the edge weight linking the vertices. Vertex connectivity (wi) is the sum of edge weights of the edges con-
nected to the vertex. (b) A directed network comprising vertices and directed edges. (i) In the matrix representation of an unweighted directed
network of n vertices, aij = 1 if a directed edge is pointing from vertex i to vertex j, and aji = 1 if a directed edge is pointing from vertex j to vertex i.
Vertex IN degree is the sum of vertices connected to the vertex. Vertex OUT degree is the number of vertices to which the vertex is connected.
(ii) A matrix representation of a weighted directed network. Cells of edges directed from vertex i to vertex j contain the edge weight. Vertex IN
degree is the sum of edges connected to the vertex. Vertex OUT degree is the sum of edges connecting the vertex to other vertices. (c) Network
modularity describes the vertices’ clustering behavior, and vertex degree measures the vertex connectivity within the network. Vertices those are
more densely connected to each other than with the rest of the network may be clustered into modules (Module A and B, gray and green clouds).
The degree centrality is presented near the vertices (Ca = 2, Cb = 3, etc.). Vertex color corresponds to the betweenness centrality that is calcu-
lated by the frequency of the vertex occurrence along shortest paths between nodes in the network. Vertices having higher betweenness central-
ity are usually connected to two or more distinct modules. (Adapted with modification from 86.) doi:10.1128/9781555818821.ch4.1.1.f4

edges connecting to the vertex (85) (Fig. 4a). For example,
edge weight in the coauthorship network is the number of
publications coauthored by the two scientists linked by the
edge (81). The connectivity of a scientist in this network is
the number of edges connected to it, representing the number
of her or his coauthors. A comparison of vertex connectivity
in coauthor networks reconstructed for different scientific
disciplines reveals stark differences in the magnitude of
collaborative research projects, depending on the scientific
field. For example, the mean coauthors per scientist in the
biomedical studies (18.1 ± 1.3) is much higher than that of
physicists (9.7 ± 2) (81). In directed networks the edges are
oriented from one vertex to another (Fig. 4b). Directed net-
works can be either unweighted or weighted. Vertex connec-
tivity in a directed network is calculated depending on the
edge direction. The out and in degrees of any given vertex
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-7
are defined as the number of edges that are directed from or
into the vertex respectively (85, 87–89) (Fig. 4b). For exam-
ple, in a directed social network reconstructed from tele-
phone calls among individuals, the edges signify a phone
call between the two individuals that they connect, while
the edge direction defines the calling individual and the
receiving individual (89, 90). In the phone calls network,
the edge weight is the number of phone calls from one indi-
vidual to another. Vertex connectivity in directed networks
is calculated separately for outgoing and incoming edges. In
the directed phone call network, vertex out and in degrees
correspond to the number people whom the individual called
and the number of people who called the individual, respec-
tively (89).
Directed networks of biological systems include mainly
models of metabolic pathways (e.g., 88, 91) and regulation
schemes (e.g., 92–94). In a directed network of metabolic
pathways, the vertices represent chemical compounds (meta-
bolites) while the edges signify chemical reactions catalyzed
by the corresponding enzyme(s). The edges are directed
from the substrate to the product of the enzymatic reaction
(88). Comparing the structural properties of metabolic net-
works reconstructed for organisms from the three domains
of life revealed similar substrate in and out connectivity distri-
butions, as well as other network properties. Those results sug-
gest that the main principles governing metabolic pathway
organization are common among all living organisms (88).
Regulation networks have been used to study different regula-
tory mechanisms of gene expression. In a transcriptional reg-
ulation network, the vertices represent genes and the edges
are directed from the regulating gene (i.e., transcription fac-
tor) to the regulated gene (92). A transcriptional regulation
network reconstructed for E. coli revealed that most of the
genes in this gammaproteobacterium are regulated by only
one or two regulators with an average of three regulators per
gene. The loose connectivity as described by the degree distri-
bution in this network suggests that large groups of coregu-
lated genes are rare (92).
Networks are highly efficient as tools for visualizing
information. Modeling complex systems using a networks
approach supplies an abstract visual representation of the sys-
tem (87) enabling our brain (the most powerful of all known
computers) to look for patterns in the data. Ordering the ver-
tices in the network according to vertex or edge properties can
assist in the search for visual patterns that can then be formu-
lated as hypotheses regarding the modeled system and be
tested statistically. For example, the network of Facebook
user connections comprising 500 million people intercon-
nected via the social network is incomprehensible. However,
distributing the vertices in the network according to the geo-
graphical coordinates of user address reveals a clear user con-
nectivity pattern resembling the globe continents (http://
www.facebook.com/note.php?note_id=469716398919).
The clear geographical structure of human pairwise connec-
tions conducted via the web suggests that human relations
are primarily initiated by meetings in the real world.
Phylogenetic Networks
Networks are commonly used in phylogenetic research for the
reconstruction of evolutionary processes that are non–tree-
like in nature, including species hybridization, gene recombi-
nation, genome fusions, and LGT (78–80, 95–97). Network
applications can be also used for tree-like gene phylogenies
(i.e., genes that did not evolve by LGT) to analyze conflict-
ing phylogenetic signals stemming from biases in the data
or model misspecification. Similarly to phylogenetic trees,
phylogenetic networks can be reconstructed from various
data types including molecular sequences, evolutionary dis-
tances, presence/absence data, and trees (98).
Phylogenies of closely related strains are typically compli-
cated by multiple recombination events, producing contra-
dicting phylogenetic signals. The presence of phylogenetic
signals stemming from recombination can be tested statisti-
cally by measuring the genealogical correlation of adjacent
sites using the pairwise homoplasy index (PHI or Φw) (99).
The PHI test is applied onto multiple sequence alignment
data. The test statistic (Φw) is calculated from pairwise com-
parisons of adjacent sites. The level of pairwise site incompat-
ibility is averaged along a predefined number of sites (e.g., 100
alignment sites). Significance of the observed statistic is cal-
culated using a permutation test. The null hypothesis is that
permutations of a fully compatible (i.e., tree-like) alignment
will have small impact on the resulting Φw values. In contrast,
the observed PHI in sequences undergoing recombination is
expected to depend on the order of alignment positions
because close positions are expected to have more similar
genealogies than distant positions (99). A significant pres-
ence of recombination signals in the data suggests that any
phylogenetic analysis of the sequences using a bifurcating
phylogenetic tree will miss a considerable portion of the
genealogical information.
Phylogenetic split networks present an alternative to phy-
logenetic trees for the study of sequences where recombina-
tion have occurred. These networks are reconstructed from
all bipartitions of a set of taxa as implied by the underlying
data (100–102). The splits are classified as compatible if
they correspond to the branching pattern of a phylogenetic
tree, and incompatible if they do not (101). A phylogenetic
split network includes both compatible and incompatible
splits, hence it can be used to depict and analyze multiple evo-
lutionary scenarios, not only those that are represented by a
single phylogenetic tree (101). A phylogenetic reconstruc-
tion of a split network can reveal conflicting phylogenetic sig-
nals resulting from recombination events during microbial
evolution. Such networks can be reconstructed for specific
gene families, from concatenated alignments of core gene
families, or from global alignments of whole genome sequen-
ces (Fig. 5).
Splits networks are commonly used in evolutionary stud-
ies of closely related strains and may be used also as supporting
tool for multilocus sequence typing analysis (e.g., 104–107).
For example, a splits network reconstructed for E. coli strains
sampled from among human symbionts and environmental
strains reveals that recombination events between the two
groups are rare (103). In their study of H. pylori strains from
eastern Asia, Kawai and colleagues (108) used a splits net-
work to analyze and visualize recombination events occurring
along protein domains. Their analysis revealed differential
protein sequence variants that occur in the different ecotypes,
probably as a result of multiple recombination events in differ-
ent protein domains (108).
Phylogenomic networks are a special type of phylogenetic
networks that are reconstructed from the analysis of whole
genomes. The vertices in a phylogenomic network corre-
spond to fully sequenced genomes that are linked by edges
representing evolutionary relationship reconstructed from
whole-genome comparisons (86). Current applications in
the literature include genomes from the three domains of
life (e.g., 109, 110), or prokaryotes only (e.g., 111–113) as
well as genomes of plasmids (e.g., 110, 114, 115) and bacter-
iophages (e.g., 116). Phylogenomic networks can be divided
into two main types: gene-sharing and LGT networks.
4.1.1-8 ▪ THEORY

FIGURE 5 Splits networks. (a) An illustration of the simplest splits graph that is not a tree. The graph “contains” the splits in trees T1 and T2
but not T3 (adapted from 101). (b) A splits network Escherichia genomes reconstructed from a concatenated alignment of 1,910 universal
single copy (USC) genes. The zoom on the left highlights the two major splits connecting E. fergusonii with the pathogenic and C-I group
(red) or S. typhi and E. albertii clade (green) (adapted with modification from 103). doi:10.1128/9781555818821.ch4.1.1.f5
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-9
Phylogenomic Networks of Shared Genes
Networks of shared genes are reconstructed from the pres-
ence/absence pattern of all orthologous protein families dis-
tributed across the genomes in the network (e.g., 15, 109,
110, 113, 116). The vertices in the network signify genomes
(or strains) while the edges correspond to gene sharing
between the genomes they connect. A typical gene sharing
network reconstruction procedure includes the following
steps: (a) selecting the genomes to be included in the net-
work, (b) sorting all proteins encoded in the selected
genomes into protein families, and (c) calculating the num-
ber of shared genes for each pair of genomes as the number
of protein families present in both genomes. In the simplest
form of this type of network, the edge weight corresponds
to the number of shared protein families between the
genomes it connects (e.g., 110, 113) (Fig. 6a). The shared
genes network can be fully defined by a matrix A = [aij ]n*n,
where n is the number of genomes in the network, and aij is
calculated as the number of common proteins families to
genomes i and j. Genomes that share one or more proteins
are thus connected by an edge. Because genome size may
vary considerably among the strains included in the network,
defining the edge weight as the frequency of shared genes can
yield bias in the network where large genomes appear to be
highly connected. An alternative edge weight definition
that is used in gene sharing networks is thus the proportion
of shared protein families from the total genome sizes of the
connected vertices (e.g., 15, 109, 116). A graphical represen-
tation of a gene-sharing network can reveal various structural
properties of the network. For example, a network recon-
structed from both eukaryotic and prokaryotic genomes
reveals a strong phylogenetic structure within the network
with a clear distinction between the three domains of life
(109). Phylogenomic shared-genes networks of microbial
genomes reveal strong connections between closely related
species (15, 109) (e.g., Fig. 6a) as well as abundant gene shar-
ing across taxonomic groups that is characteristic of evolution
by LGT (15, 110, 113). Gene-sharing networks in the liter-
ature are typically reconstructed from complete genomes of
known taxonomic classification (15, 109, 113). Neverthe-
less, there are also examples for networks comprising genomes
of plasmids (115, 119) or bacteriophages (116) or even envi-
ronmental metagenomes (110).
In an example study, Lima-Mendez and colleagues (116)
tested the applicability of networks approach to bacterio-
phage classification using a phylogenomic shared genes net-
work. Phages are traditionally classified by their host range
and genetic material (i.e., RNA or DNA in single- or double-
stranded formation), however, the accumulating number of
complete phage genomes enable a classification by gene con-
tent (116). The network was reconstructed from 306 bacter-
iophage genomes encoding a total of 19,537 genes that were
clustered by sequence similarity into 8,576 protein families.
The resulting phage network reveals that clusters of similar
genomes in terms of gene content largely overlap the
accepted classification according to phage type and form. In
this example the authors exemplify the usefulness of networks
approach for a systematic classification of phages from
genomic data (116).
The network of shared microbial transposases is an
example of a phylogenetic network reconstructed from gene
presence/absence data of a specific gene family (120). Trans-
posases—the most abundant enzymes in nature (121)—
catalyze DNA transposition within and between genomes
by a “cut-and-paste” or “copy-paste” mechanism (122). An
analysis of sequence divergence patterns among transposases
showed that these enzymes are transferred more frequently by
LGT than by vertical inheritance (123). Thus the distribu-
tion of shared transposases among microbial genomes is
expected to correlate with LGT. In the microbial transposases
network, the vertices are species and the edges correspond to
transposase families shared between the genomes that they
connect (120). The shared transposases network reveals
that most of the interactions are between closely related spe-
cies living in the same environment. However, interhabitat
connections are also quite common in the network supplying
evidence for prokaryotic mobility across habitats, either at
present or in the past (120).
Halary and colleagues used a phylogenomic network of
shared genes to study the evolution of genetic diversity across
different DNA vehicles—chromosomes, plasmids, and
phages (110). Their network comprises 111 genomes of euk-
aryotes and prokaryotes, as well as several thousands of phage
and plasmid protein sequences, many of the latter obtained
from metagenomic data sets. Using a network of shared genes
among the different DNA vehicles revealed that the distribu-
tion of most protein families in the network was limited to a
specific type of DNA vehicle. This result suggests that
although different genetic worlds are interconnected, there
exist clear boundaries between the different DNA vehicles.
However, the network also contained a large connected com-
ponent where chromosomes, plasmids, and phages are highly
interconnected. Frequent links between bacterial chromo-
somes and plasmids in that component indicate that LGT
by conjugation is highly prevalent in natural habitats (110).
Shared gene content among fully sequenced genomes can
be studied also using split networks (124). Using the exten-
sive set of tools developed for split networks reconstruction
(125) enables the analysis and depiction of conflicting phylo-
genetic signals within gene sharing data. Split networks of
shared gene content among prokaryotes can reveal insights
into the most ancient splits among microbial genomes
(124). The splits in this type of networks are reconstructed
from the presence/absence pattern of protein families across
fully sequenced microbial genomes. Each protein family
defines a partitioning of the genomes into those that encode
for that protein and those that do not. Such a split network
reconstructed from 22 archaebacterial and 169 eubacterial
genomes revealed an ancient divide within microbial life
between archaebacteria and eubacteria and an interdomain
root position (124).
Phylogenomic LGT Networks from Shared Genes
Phylogenomic LGT networks have been developed to study
the lateral component in microbial evolution and are recon-
structed from LGT events that are inferred from genomic data
(112, 113, 126, 127). Networks of laterally shared genes
(LSGs) are a special case of shared genes networks. These
are designed specifically to study gene distribution patterns
resulting from LGT during prokaryotic evolution. The verti-
ces in the network are the external and internal nodes of a
reference species phylogenetic tree. Edges in the network cor-
respond to putative gene transfer events between the nodes
they connect (111, 113, 127) (Fig. 6b). LGT inference in cur-
rent applications of LSG networks is based on mapping gene
gain and loss events within each protein family onto the refer-
ence tree nodes. A gene gain event can be either a gene birth
(e.g., by gene duplication, see [128] for review) or a gene
acquisition via LGT. The underlying assumption in the infer-
ence of LGT is that genuine gene birth events are much more
rare than LGT events. Hence, in protein families were N > 1
4.1.1-10 ▪ THEORY

FIGURE 6 (See legend on following page.)

 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-11

gain events were inferred, only one of the gains is a gene birth
and the remaining N−1 gain events are gene acquisitions by
LGT. In the LGS network, nodes in the reference tree are
connected if there is at least one protein family that is shared
between the nodes via a putative LGT event. Edge weight in
the LSG network corresponds to the number of LSG gains
between the connected nodes (111, 112).
Two different LSG network reconstruction methods are
documented in the literature. Gene gain and loss events in
the “net of life” network (111) are inferred by a parsimo-
nious algorithm for ancestral gene content reconstruction.
In the minimal lateral network (MLN) approach (112)
gene gain and loss events are reconstructed by the ancestral
genome size criterion. The application of phylogenomic
LSG network including both gene inheritance and gene
acquisition by LGT enables an inference of the cumula-
tive impact of LGT during microbial evolution. An MLN
reconstructed from 181 fully sequenced microbial genomes
revealed that, on average, 81% ± 15% of the proteins in
each genome are affected by LGT at some time during evo-
lution (112).
Phylogenomic LGT Networks from Trees
Phylogenomic LGT networks have been reconstructed from
LGT events detected in gene phylogenies as well (126,
127). As in the LSG network, the phylogenomic LGT net-
work reconstruction requires a species tree that is considered
as a reference for distinction between vertical inheritance and
LGT. For the network reconstruction, a phylogenetic tree is
reconstructed for each protein family. Branches (splits) in
the protein family tree that are found in disagreement with
the reference species tree are considered LGT events and
are included in the network (126, 127).
The LGT network reconstructed by Beiko et al. (126) is a
summary of all LGT events inferred from 22,432 phylogenies
of orthologous protein families encoded in 144 prokaryote
genomes. The nodes in the network correspond to 21 higher
taxonomic groups of microbes (e.g., Cyanobacteria, Euryarch-
aeota, Bacilli). Edges in the network correspond to LGT
events between members of the groups and are weighted by
the number of laterally transferred genes (126). The network
comprises a total edge weight of 1,398 LGT events. The heav-
iest edges in the network connect the vertices of Alphapro-
teobacteria, Betaproteobacteria, and Gammaproteobacteria.
The sum of the edge weights linking these three groups corre-
sponds to 56% of the transfers in the network, indicating that
LGT is frequent during the evolution of species in these
classes (126). However, the current sampling density of com-
pletely sequenced microbial genomes is strongly biased
towards Proteobacteria, many of them are human pathogens.
Hence the high frequency of LGTs observed within this phy-
lum could be attributed to their overrepresentation in the data
(51 (37%) of the species in the network).
LGT inference methods that include the identification of
the donor and recipient in the gene transfer event enable the
reconstruction of directed phylogenomic networks. Popa
et al. (127) described a directed network of LGT (dLGT)
comprising 32,027 recent LGT events reconstructed from
657 fully sequenced microbial genomes. The vertices in
this network are contemporary and ancestral microbial spe-
cies (as in the LSG network). Edges in the dLGT network
correspond to one or more recent LGT events between the
species they connect and are directed from the donor to
the recipient. The edge weight is the number of genes that
were laterally transferred between the connected genomes
(127) (Fig. 6c). The nodes in the dLGT network are
arranged by the density of their connections. Highly con-
nected species, having frequent recent LGTs between
them, are placed close together in the graph (Fig. 6c). Spe-
cies from the same taxonomic group are colored by the same
node color. The resulting network shows that vertices that
are close together in the graph often have the same color
(e.g., the clusters of Enterobacteriales or Xanthomonadales
in Fig. 6c). Hence most of the recent LGT events within
the dLGT network are among closely related species. Using
a dLGT networks approach enables coupling of information
regarding LGT events and cellular properties of donors and
recipients. The dLGT network reconstructed by Popa et al.
(127) revealed that DNA repair mechanisms could be
involved in DNA integration into the recipient genome dur-
ing an LGT event, enabling gene acquisition from distantly
related donors.

FIGURE 6 Phylogenomic networks reconstructed from gammaproteobacterial genomes. (a) A matrix representation of a phylogenomic
shared genes network. Protein families were reconstructed under the constraint of 30% (left) and 70% (right) amino acid identities. The species
are sorted by an alphabetical order of the order and genus. The color scale of cell aij in the matrix indicates the number of shared protein families
between genomes i and j. The matrix representation of the phylogenomic shared genes network reconstructed from gammaproteobacterial
genomes clearly shows groups of highly connected species having many genes in common. These groups usually comprise closely related species.
Examples are 14 Shewanella species (Alteromonadales order) at the top left corner, and 6 Xanthomonas species (Xanthomonadales order) at the
bottom right corner of the matrix, intraconnected species corresponding to (top to bottom) 12 Escherichia species, 7 Salmonella species, 6 Shigella
species, and 12 Yersinia species, which have many genes in common. Applying a higher protein similarity cutoff (right) yields a shared genes
network of conserved genes only. The network shows a clear phylogenetic signal with most genes shared among closely related species. (b) A
phylogenomic network of laterally shared genes reconstructed by the minimal lateral network (MLN) approach. Vertical edges (tree branches)
are indicated in gray, with both the width and the shading of the edge shown proportional to the number of inferred vertically inherited genes
along the edge (see scale on the left). The lateral network is indicated by edges that do not map onto the vertical component, with the number of
genes per edge indicated in color (see scale on the right). Edges of weight <10 are excluded (adapted from 113). (c) A dLGT network. The nodes
correspond to contemporary or ancestral species that are connected by directed edges of LGT. The edges point from the LGT donor to the
recipient. Node color corresponds to species taxonomic classification (see legend at the bottom). A cluster of connected bacilli (marked
with a star) is enlarged to exemplify the network underlying data. Species names are shown next to the nodes. Gene identifier and protein
annotation of detected recent gene transfers are noted next to the corresponding edge. The lateral acquisition of genes for sucrose utilization
in Pediococcus pentosaceus from Lactobacillus plantarum has been suggested before (117). Oenococcus oeni (strain PSU-1)—which is associated
with malolactic fermentation in wine—is connected with three different Lactobacillus species as donor and recipient. Bon et al. (118) showed
recently that gene acquisition from various donors, especially lactic acid bacilli, contributes to genome plasticity in this species and suggested
LGT as a mechanism to enhance O. oeni tolerance for harsh wine conditions. doi:10.1128/9781555818821.ch4.1.1.f6
4.1.1-12 ▪ THEORY

FIGURE 7 Network reconstruction workflow. Bioinformatic pipelines for the construction of phylogenomic networks from genomic data.
Alternative combinations of the steps in the analysis yield different phylogenomic networks. For example, pipelines 1 and 2 can be used to
construct phylogenomic networks from self-made protein families. The two pipelines differ in the species reference tree used for the LGT infer-
ence that is reconstructed wither from USC genes ( pipeline 1) or 16S rRNA genes ( pipeline 2). Pipelines 3 and 4 demonstrate the use of ready-
made clusters. Pipelines 4, 5 result in a dLGT using a USC or 16S species reference tree. (1) Genomic database preparation. Collect microbial
genome/proteome sequence data for the strains of interest. These could be either self-sequenced or publicly available genomes (e.g., NCBI
RefSeq). Draft genome sequences including annotated contigs could be used as well although these should be used with caution as missing
proteins due to misassembly and/or misannotation can bias the results. (2) Construction of orthologous protein families. Phylogenomic net-
works are reconstructed within the protein family framework. Clusters of orthologous genes are reconstructed from an all-versus-all genome
comparison by sequence similarity. In the following description, the term genome is related to a multifasta format containing all protein sequen-
ces encoded within the genome of a specific strain (i.e., sequenced from a pure culture). (2.1) Genomic comparison. Orthologous genes are
identified by the reciprocal best BLAST hit procedure (rBBH) (135). There are several approaches to perform this stage; here we describe
two of those. The first approach includes a pairwise BLAST of all genomes. This includes the construction of a BLAST database for each
genome using formatdb and running the blast command for all genome pairs where one genome constitutes the query and the other genome
serves as the database, and the other way around. In the second approach, all genomes are grouped to construct a single BLAST search database.
The total proteins are then BLASTed on that single database either one by one or genome by genome. (2.2) Inference of orthologous genes.
Orthologous sequences are defined as pairs of proteins encoded in two different genomes that reciprocally found each other as the best BLAST
hit (BBH). In the first BLAST approach this is a straightforward application. In the second approach one has to make sure that the BBH is
indeed the best hit from among all hits of the sequence within the same genome. Pairs of putative orthologs have to be sorted by the significance
of their rBBH result. Using the commonly used threshold only those BLAST results having an E-value <1E-10 are included. In principle, the
clustering into protein families (step 2.3) can be performed using the E-value or the percentage (%) sequence identity as extracted directly from
the reciprocal BBH procedure. However, to remove biases in the analysis as a result of local sequence similarity it is advisable to sort out putative
orthologs whose global sequence similarity is low. For this purpose, sequence pairs are aligned with a global alignment algorithm (e.g., needle
that is included in the EMBOSS package; 136) and the proportion of identical amino acids is calculated (e.g., using a PERL script). The com-
monly used threshold for orthologous proteins is global identical amino acids ≥30% (e.g., 15). (2.3) Clusters of orthologous protein families.
The resulting pairs of orthologous genes are summarized into a protein similarity matrix that is used for the clustering into protein families. The
basic similarity matrix is defined as M = [mij ]n*n, where n as the total number of genes in the genomic database, and mij is calculated as the
sequence similarity between rBBH gene pairs as measured either from a global pairwise sequence alignment or the BLAST results (see step
2.2). Matrix cells of gene pairs that were not found as rBBH are set to 0. The protein similarity matrix is the input for the clustering software.
The Markov clustering algorithm—MCL (137)—is commonly used in such applications (e.g., 109, Dagan:2007ju, Halary:2010bq). A typical
output of a clustering algorithm will include an arbitrary protein family number and the list of member protein sequences. Singletons will appear
in the output as single-member clusters. Multiple paralogs are often clustered into the same protein family due to the limitation of sequence
similarity information for the distinction between orthologs and paralogs. Hence a protein family may include more than one representative
gene for certain genomes (i.e., strains). (2.4) Ready-made protein families. As an alternative to the full construction of protein families, one
could also use any of the ready-for-use clusters in a number of public databases (e.g., OMA (138) or IMG 139). (3) Phylogenomic network
of shared genes. Clusters of orthologous genes supply the required information for the construction of a shared genes network. The network
is defined by a matrix M = [mij ]n*n, where n is the number of genomes in the data set, and mij is calculated as the number of common proteins
families to genomes i and j. A visual presentation of the matrix is enabled by various tools, for example using image function in MATLAB, or
(See legend on following page)
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-13

Structural Properties of Phylogenomic Networks
Structural properties of networks can be analyzed and under-
stood using an extensive set of tools developed over the years
(83–85). Node connectivity, for example, is a measure that
quantifies the extent to which a node is central within the
network (85). A similar measure, vertex centrality, quantifies
the frequency in which the vertex occurs along the shortest
path between any vertex pair in the network. The overall dis-
tribution of vertex centrality is commonly used to test for
internal structure within the network. A distribution that is
different from that of a random network indicates that verti-
ces in the network have a preferential attachment resulting
from the evolutionary history of the network (87).
Vertex connectivity in phylogenomic LSG networks can
serve as a measure for the frequency in which the species
donates or acquires genes by LGT. The genomes of the plan-
comycetes Rhodopirellula baltica str. SH1 (Pirellula sp.) and the
alphaproteobacteria Bradyrhizubium japonicum, for exam-
ple, are highly connected within the LSG network (hub
genomes) (111). These two species harbor a relatively big
proteome, R. baltica with 7,325 proteins and B. japonicum
with 8,317 proteins. Genome size and the frequency of
acquired genes are positively correlated (129), hence species
having large genomes are expected to be highly connected in
phylogenomic networks of LGT. In the dLGT network,
genome size correlates positively with both in and out vertex

FIGURE 7 (Continued.) heatmap function in R. (4) Reference species tree. The reference tree can be reconstructed from a single marker gene
(e.g., 16S rRNA) or a group of genes. (4.1) Species tree from 16S rRNA sequences. Sequences of the 16S rRNA should be extracted from the
genome (annotated genomes in NCBI contain a specific file for rRNA genes only) and stored in a multifasta file. For closely related species
having similar 16S rRNA it is advisable to repeat this step for the 5S and 23S genes. Their inclusion in the phylogenetic tree will help to improve
the splits resolution. (4.2) Species tree from universal single copy (USC) genes. A second approach for the reconstruction of a species tree is to
use the phylogenetic information in all USC genes. These are protein families (from step 2.3) that are encoded in all genomes in the dataset as a
single copy. The inclusion of single copy genes is aimed to reduce the risk for orthology misspecification. Furthermore, single copy genes are less
attributed to LGT (16) that may bias the vertical phylogenetic signal. Each of the protein families has to be formatted into a multifasta file
containing all protein sequences for the genome. (4.3) Multiple sequence alignment. Progressive alignment algorithms can be applied using
ClustalO (140) or MUSCLE (141). Another possibility is to use the fast Fourier transformation alignment approach using MAFFT (142),
which has been shown to produce somewhat better alignments in comparison to other approaches (143). When multiple genes are used for
the tree reconstruction, these should be aligned first and then concatenated into a single alignment (e.g., a concatenated alignment of all
rRNA genes or all USC protein families). It is advisable to test for multiple sequence alignment quality of each protein family using, for example,
the GUIDANCE tool (144). (4.4) Species phylogenetic tree. The species phylogenetic tree is reconstructed from a marker gene alignment (e.
g., 16S rRNA) or a concatenated alignment (e.g., rRNA genes or USC genes). The reconstruction can be performed using the neighbor-joining
approach (NJ) (145) applying the PHYLIP package (146) programs dnadist or protdist and then neighbor. Alternatively, the maximum-
likelihood approach may be applied using, for example, PHYML (147). It is advisable to calculate bootstrap support in addition. When
USC genes are used for the reconstruction, it is worthwhile to calculate in addition a consensus tree that will assist in the detection of weak
phylogenetic signal due to recombination and LGT (e.g., 148). For this purpose one has to reconstruct all gene trees and find the consensus
topology using, for example, the PHYLIP program consense. (5) Species SplitsTree. The collection of USC genes can be used to reconstruct a
species SplitsTree that is helpful for the study of recombination events (Fig. 6). The SplitsTree can be reconstructed either from the USC genes
concatenated alignment or from the total USC gene trees using SplitsTree software (125). (6) Minimal lateral network. The MLN is recon-
structed from the clusters of orthologous genes (step 3.3) and the reference species tree (step 4). The MLN approach is based on three assump-
tions (15): (i) All gene trees are perfectly compatible with the same reference tree. (ii) Gene loss is unpenalized. (iii) All within-genome
duplications for each gene family are assumed to have occurred subsequent to the last divergence for each lineage. For the MLN reconstruction,
the presence and absence patterns (PAPs) of protein families are superimposed on the reference tree to infer laterally shared genes. The PAPs
can be defined as a matrix P = [ pij ]n*m, where n is the number of genomes in the data set, m is the number of protein families, and pij is set to 1 if
genome i encodes for at least one copy of protein family j or 0 otherwise. Gene gain and loss events are inferred using a parsimonious algorithm
(15) and the MLN is reconstructed by connecting the gain events within protein families having more than a single gene origin (112). (7)
Phylogenomic LGT network. The construction of a phylogenomic LGT network from trees requires the reconstruction of gene trees for all
protein families using one of the multiple sequence alignment algorithm (see step 4.3) and phylogenetic tree reconstruction approach (see
step 4.4). Protein families that are represented in <4 genomes may be excluded from the analysis. All resulting trees are compared to the species
reference tree and branches (i.e., splits) that are found in significant disagreement with the reference tree are considered as LGT events (see 126
for details). The network can be presented by summarizing the discordant splits (i.e., LGT events) between the main taxonomic groups (126), or
similarly to the MLN where discordant splits are superimposed on the reference tree. The interpretation of discordant branches of paralogs may
be problematic when using this approach. Multiple gene copies can be pruned out according to their rBBH results (see 124), which results in a
considerable loss of data. Alternatively, the network can be reconstructed from only single copy families only. (8) Inference of directed LGT
(dLGT) events. The reconstruction of a dLGT network includes the identification of the recipient and donor in the gene transfer event (127).
(8.1) Recent LGT event. The detection of recently acquired genes is performed by comparing the percent of G and C nucleotides (%GC) in all
genes to the distribution of the gene-%GC for the total genome (see (127) for details). Those genes whose %GC is significantly different from
that of the genome are considered as putative laterally transferred genes. (8.2) Donor inference. The identification of donors in the LGT event
includes the reconstruction of a phylogenetic tree for each detected gene acquisition and a comparison of the gene tree to the reference tree. The
main stages in the tree reconstruction include the identification of orthologs using rBBH procedure (step 2.1), a multiple sequence alignment of
the putative acquired gene with the orthologs using ClustalW (149), and a phylogenetic tree reconstruction using PHYML (147). Putative
donors are defined as genes (i.e., species) that branch with the acquired gene in the gene-tree but not in the reference species tree. Putative
acquisitions of closely related orthologs (as defined by the species reference tree) are grouped into a single acquisition event that occurred at
the ancestral strain, hence internal nodes in the reference tree (i.e., hypothetical taxonomic units) may also result as putative donors and recip-
ients (see 127 for details). (8.3) Directed LGT network. The directed network of LGT is defined as an asymmetric matrix M = [mij ]n*n, where n
is the number of nodes in the reference tree (both internal and external) and mij is the number of detected laterally transferred genes from tree
node i (the donor) to tree node j (the recipient). The network can be presented similarly to the MLN where detected LGT events are super-
imposed on the reference tree or by using Cytoscape (150). doi:10.1128/9781555818821.ch4.1.1.f7
4.1.1-14 ▪ THEORY
degree (rIN = 0.38, rOUT = 0.39) indicating that species with
large genomes are not only frequent recipients but also fre-
quent donors (127). In the phylogenomic gene-sharing
network among different DNA carriers, plasmids have signif-
icantly higher centrality than do phages (110). This result
suggests that LGT in nature is more frequently mediated by
conjugation than by transduction (110). Edge weight distri-
bution in weighted networks can also supply information
regarding link patterns in the network. The edge weight dis-
tribution in the LSG and dLGT networks is linear in a log-log
scale, indicating that the majority of LGT events are of one
or few genes while bulk transfers of many genes are rare
(111–113, 127).
Another measure of interest is the diameter of a network,
which quantifies the mean shortest path length between any
two vertices in the network (85). In the aviation network, for
example, this is the average number of flights that one needs
to book to travel from any city to any other city in the world
(82). Networks having a small diameter are designated “small
world” networks (83–85, 130). The human society is an
example for such a network; the median of distances between
any given pair of humans measured by mutual acquaintances
is only 5.5 (130). The diameter of the LSG network meas-
ured by the mean shortest path between any genomes pair
ranges between two and five nodes, indicating that they
form a small world network (111, 112). This implies that a
gene can be transferred between any two random species
by no more than five LGT events via intermediate recipi-
ents/donors. This could be the reason for the rapid percola-
tion of antibiotic resistance genes (131) within pathogenic
populations.
Networks can also display community structure (132).
A network that includes groups of vertices that are densely
connected within the group but sparsely connected with ver-
tices from other groups is said to have an internal community
structure (85, 132–134). Communities are the functional
building blocks of the network and could supply information
about its evolutionary history (133, 134). An example is the
network of protein–protein interactions within the cell. In
this network, proteins (vertices) that were found to interact
are linked by an edge. The protein–protein interaction net-
work has a significant community structure. Proteins that
function in the same cellular process form communities of
densely interacting proteins while proteins from different cel-
lular processes interact sparsely (132).
The phylogenomic networks of shared genes among pro-
karyotes have a clear community structure that largely corre-
sponds to the taxonomic classification of the connected
species (112). In Proteobacteria, the community structure
within a network comprising 329 genomes reveals a deep
split, one that was not detected by common phylogentic
methods, between alpha-, delta-, and dpsilon-proteobacteria
in one group and beta- and gamma-proteobacteria in the
other group (113). Communities in the network of shared
genes among DNA carriers are strictly homogeneous with
regard to plasmids and phages. This indicates that these two
gene vehicles rarely carry the same genes (110). Community
structure within the dLGT network reveals groups of species
that are connected by LGT events much more than with spe-
cies outside the group. Most of the communities in this net-
work comprise species from the same taxonomic group,
hence the majority of recent LGT events occur between
closely related donors and recipients. The rare communities
that group together distantly related species are evidence for
frequent LGT within a common habitat or via a common
phage (127).
CONCLUDING REMARKS
The different types of phylogenomic networks reconstructed
from whole microbial genomes offer slightly different insights
into microbial genome evolution (Fig. 7). Networks capture a
substantial component of genome evolution, which is not
tree-like in nature. Therefore, in biological systems where
reticulated evolutionary events are common, phylogenomic
networks offer a general computational approach that is
more biologically realistic and evolutionarily more accurate.
The prevalence of LGT during microbial and viral evolution
makes phylogenomic networks an essential tool in the study
of these systems.
The networks approach enables studying of several geno-
mic and species characteristics in parallel, such as evolution-
ary relatedness, common habitats, shared gene content, and
common metabolic pathways. The rapid advance of new
sequencing technologies will deliver a genome sample density
that was previously unthinkable. It is clear that there is abun-
dant interspecific gene recombination among prokaryotic
genomes in nature. Phylogenomic networks will enable the
mathematical modeling of evolutionary processes and the
investigation of cellular mechanisms that drive microbial
genome evolution.
We acknowledge support from the European Research Council (grant
no. 281357, EVOLATERAL).
REFERENCES
1. Olsen GJ, Woese CR. 1993. Ribosomal RNA: a key to
phylogeny. FASEB J 7:113–123.
2. Woese CR, Kandler O, Wheelis ML. 1990. Towards a nat-
ural system of organisms: proposal for the domains Archaea,
Bacteria, and Eucarya. Proc Natl Acad Sci USA 87:
4576–4579.
3. Doolittle WF, Zhaxybayeva O. 2009. On the origin of pro-
karyotic species. Genome Res 19:744–756.
4. Welch RA, Burland V, Plunkett G, Redford P, Roesch P,
Rasko D, Buckles EL, Liou S-R, Boutin A, Hackett J,
Stroud D, Mayhew GF, Rose DJ, Zhou S, Schwartz
DC, Perna NT, Mobley HLT, Donnenberg MS, Blattner
FR. 2002. Extensive mosaic structure revealed by the com-
plete genome sequence of uropathogenic Escherichia coli.
Proc Natl Acad Sci USA 99:17020–17024.
5. Beauregard-Racine J, Bicep C, Schliep K, Lopez P,
Lapointe F-J, Bapteste E. 2011. Of woods and webs: possi-
ble alternatives to the tree of life for studying genomic fluid-
ity in E. coli. Biol Direct 6:39.
6. Anderson NG. 1970. Evolutionary significance of virus
infection. Nature 227:1346–1347.
7. Dickerson RE. 1980. Evolution and gene transfer in purple
photosynthetic bacteria. Nature 283:210–212.
8. Cohen SN, Chang AC. 1973. Recircularization and auton-
omous replication of a sheared R-factor DNA segment in
Escherichia coli transformants. Proc Natl Acad Sci USA 70:
1293–1297.
9. Syvanen M. 1984. The evolutionary implications of
mobile genetic elements. Annu Rev Genet 18:271–293.
10. Lawrence JG, Ochman H. 1998. Molecular archaeology of
the Escherichia coli genome. Proc Natl Acad Sci USA 95:
9413–9417.
11. Martin W. 1999. Mosaic bacterial chromosomes: a chal-
lenge en route to a tree of genomes. BioEssays 21:99–104.
12. Doolittle WF. 1999. Phylogenetic classification and the
universal tree. Science 284:2124–2128.
13. Mirkin BG, Fenner TI, Galperin MY, Koonin EV. 2003.
Algorithms for computing parsimonious evolutionary
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-15
scenarios for genome evolution, the last universal common
ancestor and dominance of horizontal gene transfer in the
evolution of prokaryotes. BMC Evol Biol 3:1–34.
14. Kunin V, Ouzounis CA. 2003. The balance of driving
forces during genome evolution in prokaryotes. Genome
Res 13:1589–1594.
15. Dagan T, Martin W. 2007. Ancestral genome sizes specify
the minimum rate of lateral gene transfer during prokaryote
evolution. Proc Natl Acad Sci USA 104:870–875.
16. Sorek R, Zhu Y, Creevey CJ, Francino MP, Bork P,
Rubin EM. 2007. Genome-wide experimental determina-
tion of barriers to horizontal gene transfer. Science 318:
1449–1452.
17. Griffith F. 1928. The significance of pneumococcal types.
J Hyg (Lond) 27:113–159.
18. Avery OT, Macleod CM, McCarty M. 1944. Studies of the
chemical nature of the substance inducing transformation
of pneumococcal types. J Exp Med 79:137–158.
19. Popa O, Dagan T. 2011. Trends and barriers to lateral
gene transfer in prokaryotes. Curr Opin Microbiol 14:
615–623.
20. Chen I, Christie PJ, Dubnau D. 2005. The ins and outs of
DNA transfer in bacteria. Science 310:1456–1460.
21. Thomas CM, Nielsen KM. 2005. Mechanisms of, and bar-
riers to, horizontal gene transfer between bacteria. Nature
Rev Micro 3:711–721.
22. Stingl K, Muller S, Scheidgen-Kleyboldt G, Clausen M,
Maier B. 2010. Composite system mediates two-step
DNA uptake into Helicobacter pylori. Proc Natl Acad Sci
USA 107:1184–1189.
23. Smith HO, Tomb JF, Dougherty BA, Fleischmann RD,
Venter JC. 1995. Frequency and distribution of DNA
uptake signal sequences in the Haemophilus influenzae Rd
genome. Science 269:538–540.
24. Danner DB, Deich RA, Sisco KL, Smith HO. 1980. An
eleven-base-pair sequence determines the specificity of
DNA uptake in Haemophilus transformation. Gene 11:
331–318.
25. Goodman SD, Scocca JJ. 1988. Identification and arrange-
ment of the DNA sequence recognized in specific transfor-
mation of Neisseria gonorrhoeae. Proc Natl Acad Sci USA
85:6982–6986.
26. Cehovin A, Simpson PJ, McDowell MA, Brown DR,
Noschese R, Pallett M, Brady J, Baldwin GS, Lea SM,
Matthews SJ. 2013. Specific DNA recognition mediated
by a type IV pilin. Proc Natl Acad Sci USA 110:
3065–3070.
27. Meibom KL, Blokesch M, Dolganov NA, Wu C-Y,
Schoolnik GK. 2005. Chitin induces natural competence
in Vibrio cholerae. Science 310:1824–1827.
28. Scrudato Lo M, Blokesch M. 2012. The regulatory net-
work of natural competence and transformation of Vibrio
cholerae. PLoS Genet 8:e1002778.
29. Blokesch M. 2012. A quorum sensing-mediated switch
contributes to natural transformation of Vibrio cholerae.
Mob Genet Elements 2:224–227.
30. Wang Y, Goodman SD, Redfield RJ, Chen C. 2002. Nat-
ural transformation and DNA uptake signal sequences in
Actinobacillus actinomycetemcomitans. J Bacteriol 184:
3442–3449.
31. Snyder LAS, McGowan S, Rogers M, Duro E, O’Farrell
E, Saunders NJ. 2007. The repertoire of minimal mobile
elements in the Neisseria species and evidence that these
are involved in horizontal gene transfer in other bacteria.
Mol Biol Evol 24:2802–2815.
32. Domingues S, Harms K, Fricke WF, Johnsen PJ, da Silva
GJ, Nielsen KM. 2012. Natural transformation facilitates
transfer of transposons, integrons and gene cassettes
between bacterial species. PLoS Pathog 8:e1002837.
33. Lederberg J, Tatum EL. 1946. Gene recombination in
Escherichia coli. Nature 158:558.
34. Tatum EL, Lederberg J. 1947. Gene recombination in the
bacterium Escherichia coli. J Bacteriol 53:673–684.
35. Smillie C, Garcillán-Barcia MP, Francia MV, Rocha
EPC, la Cruz de F. 2010. Mobility of plasmids. Microbiol
Mol Biol Rev 74:434–452.
36. Wozniak RAF, Waldor MK. 2010. Integrative and con-
jugative elements: mosaic mobile genetic elements ena-
bling dynamic lateral gene flow. Nature Rev Microbiol 8:
552–563.
37. Grohmann E, Muth G, Espinosa M. 2003. Conjugative
plasmid transfer in gram-positive bacteria. Microbiol Mol
Biol Rev 67:277–301.
38. Salyers AA, Shoemaker NB, Stevens AM, Li LY. 1995.
Conjugative transposons: an unusual and diverse set of
integrated gene transfer elements. Microbiol Rev 59:
579–590.
39. Babic A, Lindner AB, Vulic M, Stewart EJ, Radman M.
2008. Direct visualization of horizontal gene transfer. Sci-
ence 319:1533–1536.
40. Babic A, Berkmen MB, Lee CA, Grossman AD. 2011.
Efficient gene transfer in bacterial cell chains. mBio 2:
e00027-11.
41. Chandler JR, Hirt H, Dunny GM. 2005. A paracrine
peptide sex pheromone also acts as an autocrine signal to
induce plasmid transfer and virulence factor expression in
vivo. Proc Natl Acad Sci USA 102:15617–15622.
42. Dunny GM. 2007. The peptide pheromone-inducible con-
jugation system of Enterococcus faecalis plasmid pCF10: cell-
cell signalling, gene transfer, complexity and evolution. Phil
Trans Roy Soc B 362:1185–1193.
43. Beaber JW, Hochhut B, Waldor MK. 2004. SOS response
promotes horizontal dissemination of antibiotic resistance
genes. Nature 427:72–74.
44. Mevarech M, Werczberger R. 1985. Genetic transfer in
Halobacterium volcanii. J Bacteriol 162:461–462.
45. Rosenshine I, Tchelet R, Mevarech M. 1989. The mech-
anism of DNA transfer in the mating system of an archae-
bacterium. Science 245:1387–1389.
46. Dubey GP, Ben-Yehuda S. 2011. Intercellular nanotubes
mediate bacterial communication. Cell 144:590–600.
47. Naor A, Lapierre P, Mevarech M, Papke RT, Gophna U.
2012. Low species barriers in halophilic archaea and
the formation of recombinant hybrids. Curr Biol 22:
1444–1448.
48. Zinder ND, Lederberg J. 1952. Genetic exchange in Sal-
monella. J Bacteriol 64:679–699.
49. Zinder ND. 1992. Forty years ago: the discovery of bacterial
transduction. Genetics 192:291–294.
50. Weinbauer MG. 2004. Ecology of prokaryotic viruses.
FEMS Microbiol Rev 28:127–181.
51. Burke J, Schneider D, Westpheling J. 2001. Generalized
transduction in Streptomyces coelicolor. Proc Natl Acad Sci
USA 98:6289–6294.
52. Canchaya C, Fournous G, Chibani-Chennoufi S, Dill-
mann M-L, Brüssow H. 2003. Phage as agents of lateral
gene transfer. Curr Opin Microbiol 6:417–424.
53. Jiang SC, Paul JH. 1998. Gene transfer by transduction in
the marine environment. Appl Env Microbiol 64:
2780–2787.
54. Kenzaka T, Tani K, Nasu M. 2010. High-frequency
phage-mediated gene transfer in freshwater environments
determined at single-cell level. ISME J 4:648–659.
55. Marrs B. 1974. Genetic recombination in Rhodopseudo-
monas capsulata. Proc Natl Acad Sci USA 71:971– 973.
56. Solioz M, Yen HC, Marris B. 1975. Release and uptake
of gene transfer agent by Rhodopseudomonas capsulata.
J Bacteriol 123:651–657.
4.1.1-16 ▪ THEORY
57. Lang AS, Beatty JT. 2000. Genetic analysis of a bac-
terial genetic exchange element: the gene transfer agent
of Rhodobacter capsulatus. Proc Natl Acad Sci USA 97:
859–864.
58. Lang AS, Zhaxybayeva O, Beatty JT. 2012. Gene transfer
agents: phage-like elements of genetic exchange. Nature
Rev Microbiol 10:472–482.
59. Genthner FJ, Wall JD. 1984. Isolation of a recombination-
deficient mutant of Rhodopseudomonas capsulata. J Bacteriol
160:971–975.
60. Rapp BJ, Wall JD. 1987. Genetic transfer in Desulfovi-
brio desulfuricans. Proc Natl Acad Sci USA 84:9128–9130.
61. Bertani G. 1999. Transduction-like gene transfer in the
methanogen Methanococcus voltae. J Bacteriol 181:
2992–3002.
62. Berleman J, Auer M. 2012. The role of bacterial outer
membrane vesicles for intra- and interspecies delivery.
Env Microbiol 15:347–354.
63. Kahn ME, Barany F, Smith HO. 1983. Transformasomes:
specialized membranous structures that protect DNA dur-
ing Haemophilus transformation. Proc Natl Acad Sci USA
80:6927–6931.
64. Hagemann S, Stöger L, Kappelmann M, Hassl I, Ellinger
A, Velimirov B. 2013. DNA-bearing membrane vesicles
produced by Ahrensia kielensis and Pseudoalteromonas mar-
ina. J Basic Microbiol 53:1–11.
65. Herriott RM, Meyer EM, Vogt M. 1970. Defined non-
growth media for stage II development of competence in
Haemophilus influenzae. J Bacteriol 101:517–524.
66. Kahn ME, Maul G, Goodgal SH. 1982. Possible mecha-
nism for donor DNA binding and transport in Haemophi-
lus. Proc Natl Acad Sci USA 79:6370–6374.
67. Dorward DW, Garon CF, Judd RC. 1989. Export and
intercellular transfer of DNA via membrane blebs of Neis-
seria gonorrhoeae. J Bacteriol 171:2499–2505.
68. Kolling GL, Matthews KR. 1999. Export of virulence
genes and Shiga toxin by membrane vesicles of Escherichia
coli O157:H7. Appl Env Microbiol 65:1843–1848.
69. Yaron S, Kolling GL, Simon L, Matthews KR. 2000.
Vesicle-mediated transfer of virulence genes from Escheri-
chia coli O157:H7 to other enteric nacteria. Appl Env Micro-
biol 66:4414–4420.
70. Rumbo C, Fernandez-Moreira E, Merino M, Poza M,
Mendez JA, Soares NC, Mosquera A, Chaves F, Bou G.
2011. Horizontal transfer of the OXA-24 carbapenemase
gene via outer membrane vesicles: a new mechanism of dis-
semination of carbapenem resistance genes in Acinetobacter
baumannii. Antimicrob Agents Chemother 55:3084–3090.
71. Perez-Cruz C, Carrion O, Delgado L, Martinez G,
Lopez-Iglesias C, Mercade E. 2013. New type of outer
membrane vesicle produced by the Gram-negative bacte-
rium Shewanella vesiculosa M7T: implications for DNA
content. Appl Env Micro 79:1874–1881.
72. Gaudin M, Gauliard E, Schouten S, Houel-Renault L,
Lenormand P, Marguet E, Forterre P. 2012. Hyperther-
mophilic archaea produce membrane vesicles that can
transfer DNA. Environ Microbiol Rep 5:109–116.
73. Eisen JA. 1998. Phylogenomics: improving functional pre-
dictions for uncharacterized genes by evolutionary analysis.
Genome Res 8:163–167.
74. Eisen JA, Fraser CM. 2003. Phylogenomics: intersection
of evolution and genomics. Science 300:1706–1707.
75. Gogarten JP, Doolittle WF, Lawrence JG. 2002. Prokary-
otic evolution in light of gene transfer. Mol Biol Evol 19:
2226–2238.
76. Philippe H, Douady CJ. 2003. Horizontal gene transfer
and phylogenetics. Curr Opin Microbiol 6:498–505.
77. Dagan T, Martin W. 2006. The tree of one percent.
Genome Biol 7:118.
78. Dagan T, Martin W. 2009. Getting a better picture of
microbial evolution en route to a network of genomes.
Phil Trans Roy Soc B 364:2187–2196.
79. Huson DH, Scornavacca C. 2011. A survey of combinato-
rial methods for phylogenetic networks. Genome Biol Evol
3:23–35.
80. Bapteste E, O’Malley MA, Beiko RG, Ereshefsky M,
Gogarten JP, Franklin-Hall L, Lapointe F-J, Dupré J,
Dagan T, Boucher Y, Martin W. 2009. Prokaryotic evolu-
tion and the tree of life are two different things. Biol Direct
4:34.
81. Newman ME. 2001. The structure of scientific colla-
boration networks. Proc Natl Acad Sci USA 98:404–409.
82. Guimera R, Mossa S, Turtschi A, Amaral LN. 2005. The
worldwide air transportation network: anomalous central-
ity, community structure, and cities’ global roles. Proc
Natl Acad Sci USA 102:7794–7799.
83. Strogatz SH. 2001. Exploring complex networks. Nature
410:268–276.
84. Barabasi AL. 2002. Linked: How Everything Is Connected to
Everything Else and What it Means. Perseus Publishing, Cam-
bridge, MA.
85. Newman M. 2009. Networks: An Introduction. Oxford Uni-
versity Press, Oxford.
86. Dagan T. 2011. Phylogenomic networks. Trends Microbiol
19:483–491.
87. Barabasi AL, Albert R, Jeong H. 2000. Scale-free charac-
teristics of random networks: the topology of the world-
wide web. Physica A 281:69–77.
88. Jeong H, Tombor B, Albert R, Oltvai ZN, Barabasi AL.
2000. The large-scale organization of metabolic networks.
Nature 407:651–654.
89. Palla G, Barabási A-L, Vicsek T. 2007. Quantifying social
group evolution. Nature 446:664–667.
90. Miritello G, Moro E, Lara R. 2011. Dynamical strength
of social ties in information spreading. Phys Rev E 83:
045102.
91. Pál C, Papp B, Lercher MJ. 2005. Adaptive evolution of
bacterial metabolic networks by horizontal gene transfer.
Nat Genet 37:1372–1375.
92. Thieffry D, Huerta AM, Pérez Rueda E, Collado Vides J.
1998. From specific gene regulation to genomic networks: a
global analysis of transcriptional regulation in Escherichia
coli. BioEssays 20:433–440.
93. Shen-Orr SS, Milo R, Mangan S, Alon U. 2002. Network
motifs in the transcriptional regulation network of Escher-
ichia coli. Nat Genet 31:64–68.
94. Tsang JS, Ebert MS, van Oudenaarden A. 2010.
Genome-wide dissection of microRNA functionsand
cotargeting networks using gene set signatures. Mol Cell
38:140–153.
95. Sneath P. 1975. Cladistic representation of reticulate evo-
lution. Syst Zool 24:360–368.
96. Koonin EV, Wolf YI. 2008. Genomics of bacteria and
archaea: the emerging dynamic view of the prokaryotic
world. Nuc Acids Res 36:6688–6719.
97. Swithers KS, Gogarten JP, Fournier GP. 2009. Trees in
the web of life. J Biol 8:54.
98. Huson DH, Rupp R, Scornavacca C. 2010. Phylogenetic
Networks: Concepts, Algorithms and Applications. Cam-
bridge University Press, Cambridge.
99. Bruen TC, Philippe H, Bryant D. 2006. A simple and
robust statistical test for detecting the presence of recombi-
nation. Genetics 172:2665–2681.
100. Bandelt H-J, Dress AW. 1992. A canonical decomposi-
tion theory for metrics on a finite set. Adv Math 92:47–105.
101. Bryant D, Moulton V. 2004. Neighbor-net: an agglomer-
ative method for the construction of phylogenetic net-
works. Mol Biol Evol 21:255–265.
 4.1.1. Phylogenomic Networks of Microbial Genome Evolution ▪ 4.1.1-17
102. Dress AWM, Huson DH. 2004. Constructing splits graphs.
IEEE/ACM Trans Comput Biol Bioinform 1:109–115.
103. Luo C, Walk ST, Gordon DM, Feldgarden M, Tiedje JM,
Konstantinidis KT. 2011. Genome sequencing of environ-
mental Escherichia coli expands understanding of the ecol-
ogy and speciation of the model bacterial species. Proc
Natl Acad Sci USA 108:7200–7205.
104. Van Berkum P, Elia P, Song Q, Eardly BD. 2012. Devel-
opment and application of a multilocus sequence analysis
method for the identification of genotypes within genus
Bradyrhizobium and for establishing nodule occupancy of
soybean (Glycine max L. Merr). Mol Plant-Microbe Interact
25:321–330.
105. Perchec-Merien A-M, Lewis GD. 2012. Naturalized
Escherichia colifrom New Zealand wetland and stream envi-
ronments. FEMS Microbiol Ecol 83:494–503.
106. Jolley KA, Hill DMC, Bratcher HB, Harrison OB, Fea-
vers IM, Parkhill J, Maiden MCJ. 2012. Resolution of a
meningococcal disease outbreak from whole-genome
sequence data with rapid Web-based analysis methods.
J Clin Microbiol 50:3046–3053.
107. Maiden MCJ, van Rensburg MJJ, Bray JE, Earle SG, Ford
SA, Jolley KA, McCarthy ND. 2013. MLST revisited: the
gene-by-gene approach to bacterial genomics. Nat Rev
Microbiol 11:728–736.
108. Kawai M, Furuta Y, Yahara K, Tsuru T, Oshima K, Handa
N, Takahashi N, Yoshida M, Azuma T, Hattori M,
Uchiyama I, Kobayashi I. 2011. Evolution in an oncogenic
bacterial species with extreme genome plasticity: Helico-
bacter pylori East Asian genomes. BMC Microbiol 11:104.
109. Kunin V. 2005. Measuring genome conservation across
taxa: divided strains and united kingdoms. Nucleic Acids
Res 33:616–621.
110. Halary S, Leigh JW, Cheaib B, Lopez P, Bapteste E. 2010.
Network analyses structure genetic diversity in independ-
ent genetic worlds. Proc Nat Acad Sci USA 107:127–132.
111. Kunin V, Goldovsky L, Darzentas N, Ouzounis CA.
2005. The net of life: reconstructing the microbial phyloge-
netic network. Genome Res 15:954–959.
112. Dagan T, Artzy-Randrup Y, Martin W. 2008. Modular
networks and cumulative impact of lateral transfer in pro-
karyote genome evolution. Proc Nat Acad Sci USA 105:
10039–10044.
113. Kloesges T, Popa O, Martin W, Dagan T. 2011. Networks
of gene sharing among 329 proteobacterial genomes reveal
differences in lateral gene transfer frequency at different
phylogenetic depths. Mol Biol Evol 28:1057–1074.
114. Brilli M, Mengoni A, Fondi M, Bazzicalupo M, Liò P, Fani
R. 2008. Analysis of plasmid genes by phylogenetic profil-
ing and visualization of homology relationships using
Blast2Network. BMC Bioinform 9:551.
115. Fondi M, Fani R. 2010. The horizontal flow of the plasmid
resistome: clues from inter-generic similarity networks. Env
Microbiol 12:3228–3242.
116. Lima-Mendez G, van Helden J, Toussaint A, Leplae R.
2008. Reticulate representation of evolutionary and func-
tional relationships between phage genomes. Mol Biol
Evol 25:762–777.
117. Naumoff DG, Livshits VA. 2001. Molecular structure of
the Lactobacillus plantarum sucrose utilization locus: com-
parison with Pediococcus pentosaceus. Mol Biol 35:15–22.
118. Bon E, Delaherche A, Bilhère E, de Daruvar A, Lonvaud-
Funel A, Le Marrec C. 2009. Oenococcus oeni genome plas-
ticity is associated with fitness. Appl Env Microbiol 75:
2079–2090.
119. Fondi M, Bacci G, Brilli M, Papaleo CM, Mengoni A,
Vaneechoutte M, Dijkshoorn L, Fani R. 2010. Exploring
the evolutionary dynamics of plasmids: the Acinetobacter
pan-plasmidome. BMC Evol Biol 10:59.
120. Hooper SD, Mavromatis K, Kyrpides NC. 2009. Micro-
bial co-habitation and lateral gene transfer: what transpo-
sases can tell us. Genome Biol 10:R45.
121. Aziz RK, Breitbart M, Edwards RA. 2010. Transposases
are the most abundant, most ubiquitous genes in nature.
Nucleic Acids Res 38:4207–4217.
122. Curcio MJ, Derbyshire KM. 2003. The outs and ins of
transposition: from mu to kangaroo. Nat Rev Mol Cell Biol
4:865–877.
123. Wagner A. 2006. Periodic extinctions of transposable ele-
ments in bacterial lineages: evidence from intragenomic var-
iation in multiple genomes. Mol Biol Evol 23:723–733.
124. Dagan T, Roettger M, Bryant D, Martin W. 2010.
Genome networks root the tree of life between prokaryotic
domains. Genome Biol Evol 2:379–392.
125. Huson DH, Bryant D. 2006. Application of phylogenetic
networks in evolutionary studies. Mol Biol Evol 23:
254–267.
126. Beiko RG, Harlow TJ, Ragan MA. 2005. Highways of
gene sharing in prokaryotes. Proc Natl Acad Sci USA
102:14332–14337.
127. Popa O, Hazkani-Covo E, Landan G, Martin W, Dagan T.
2011. Directed networks reveal genomic barriers and DNA
repair bypasses to lateral gene transfer among prokaryotes.
Genome Res 21:599–609.
128. Kaessmann H. 2010. Origins, evolution, and phenotypic
impact of new genes. Genome Res 20:1313–1326.
129. Nakamura Y, Itoh T, Matsuda H, Gojobori T. 2004.
Biased biological functions of horizontally transferred genes
in prokaryotic genomes. Nat Genet 36:760–766.
130. Milgram S. 1967. The small world problem. Psychol Today
2:60–67.
131. Croucher NJ, Harris SR, Fraser C, Quail MA, Burton J,
van der Linden M, McGee L, Gottberg von A, Song JH,
Ko KS, Pichon B, Baker S, Parry CM, Lambertsen LM,
Shahinas D, Pillai DR, Mitchell TJ, Dougan G, Tomasz
A, Klugman KP, Parkhill J, Hanage WP, Bentley SD.
2011. Rapid pneumococcal evolution in response to clini-
cal interventions. Science 331:430–434.
132. Girvan M, Newman MEJ. 2002. Community structure
in social and biological networks. Proc Natl Acad Sci USA
99:7821–7826.
133. Palla G, Derényi I, Farkas I, Vicsek T. 2005. Uncovering
the overlapping community structure of complex networks
in nature and society. Nature 435:814–818.
134. Newman ME. 2003. The structure and function of com-
plex networks. SIAM Ref 45:167–256.
135. Tatusov RL, Koonin EV, Lipman DJ. 1997. A genomic
perspective on protein families. Science 278:631–637.
136. Rice P, Longden I, Bleasby A. 2000. EMBOSS: the Euro-
pean molecular biology open software suite. Trends Genet
16:276–277.
137. Enright AJ, Van Dongen S, Ouzounis CA. 2002. An effi-
cient algorithm for large-scale detection of protein families.
Nucleid Acids Res 30:1575–1584.
138. Altenhoff AM, Schneider A, Gonnet GH, Dessimoz C.
2011. OMA 2011: orthology inference among 1000 com-
plete genomes. Nucleic Acids Res 39:D289–D294.
139. Markowitz VM, Chen IMA, Palaniappan K, Chu K,
Szeto E, Grechkin Y, Ratner A, Jacob B, Huang J, Wil-
liams P, Huntemann M, Anderson I, Mavromatis K, Iva-
nova NN, Kyrpides NC. 2011. IMG: the integrated
microbial genomes database and comparative analysis sys-
tem. Nucleic Acids Res 40:D115–D122.
140. Sievers F, Wilm A, Dineen D, Gibson TJ, Karplus K, Li
W, Lopez R, McWilliam H, Remmert M, ding JSO,
Thompson JD, Higgins DG. 2011. Fast, scalable genera-
tion of high-quality protein multiple sequence alignments
using Clustal Omega. Mol Sys Biol 7:1–6.
4.1.1-18 ▪ THEORY
141. Edgar RC. 2004. MUSCLE: multiple sequence alignment
with high accuracy and high throughput. Nucleic Acids Res
32:1792–1797.
142. Katoh K, Standley DM. 2013. MAFFT multiple sequence
alignment software version 7: improvements in perform-
ance and usability. Mol Biol Evol 30:772–780.
143. Thompson JD, Linard B, Lecompte O, Poch O. 2011. A
comprehensive benchmark study of multiple sequence
alignment methods: current challenges and future perspec-
tives. PLoS One 6:e18093.
144. Penn O, Privman E, Landan G, Graur D, Pupko T. 2010.
An alignment confidence score capturing robustness to
guide tree uncertainty. Mol Biol Evol 27:1759–1767.
145. Saitou N, Nei M. 1987. The neighbor-joining method: a
new method for reconstructing phylogenetic trees. Mol
Biol Evol 4:406–425.
146. Felsenstein J. 2005. PHYLIP (Phylogeny Inference Pack-
age), version 3.6.
147. Guindon S, Dufayard J-F, Lefort V, Anisimova M,
Hordijk W, Gascuel O. 2010. New algorithms and
methods to estimate maximum-likelihood phylogenies:
assessing the performance of PhyML 3.0. System Biol 59:
307–321.
148. Dagan T, Roettger M, Stucken K, Landan G, Koch R,
Major P, Gould SB, Goremykin VV, Rippka R, Tandeau
de Marsac N, Gugger M, Lockhart PJ, Allen JF, Brune I,
Maus I, Puhler A, Martin WF. 2013. Genomes of stigone-
matalean cyanobacteria (Subsection V) and the evolution
of oxygenic photosynthesis from prokaryotes to plastids.
Genome Biol Evol 5:31–44.
149. Thompson JD, Higgins DG, Gibson TJ. 1994. CLUS-
TALW: improving the sensitivity of progressive multiple
sequence alignment through sequence weighting, position-
specific gap penalties and weight matrix choice. Nucleic
Acids Res 22:4673–4680.
150. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT,
Ramage D, Amin N, Schwikowski B, Ideker T. 2003.
Cytoscape: a software environment for integrated models
of biomolecular interaction networks. Genome Res 13:
2498–2504.
 Evolutionary Ecology of Microorganisms: From the
Tamed to the Wild
JAY T. LENNON AND VINCENT J. DENEF
4.1.2
OVERVIEW: INTERPLAY BETWEEN
ECOLOGICAL AND EVOLUTIONARY
PROCESSES
An overarching goal of biology is to understand how evolu-
tionary and ecological processes generate and maintain bio-
diversity. Despite this seemingly unified goal, historically
the fields of evolutionary biology and ecology have largely
advanced separately. Although evolutionary biologists inter-
ested in biodiversity tend to focus on the mechanisms con-
trolling rates of evolution and how this influences the
phylogenetic relationship among species, ecologists attempt
to explain the distribution and abundance of taxa based on
interactions among species and their environment. Recently,
a more concerted effort has been made to integrate some
of the theoretical and empirical approaches from the fields
of ecology and evolutionary biology. This has been moti-
vated in part by the growing evidence that evolution can hap-
pen on “rapid” or contemporary time scales (1). When this
occurs, evolutionary changes can select for functional traits
and behaviors of species in ways that influence ecological
processes, such as population dynamics, the outcome of
species interactions, and even ecosystem functioning (2–5).
Ultimately eco-evolutionary feedbacks can alter system
dynamics in ways that cannot be predicted based on ecologi-
cal principles alone (6) (Fig. 1). As such, it may be inappro-
priate to ignore evolutionary processes when attempting to
understand ecological phenomena in natural and managed
ecosystems.
Evolutionary ecology is a broad discipline that covers a
wide range of topics, including life history theory, sexual
selection, sociobiology, and coevolution, which are addressed
in greater detail elsewhere (7–9). In this chapter, we highlight
questions and approaches that are relevant to studying the
evolutionary ecology of microorganisms, with a focus on rapid
evolution. Because there is no single right way for conducting
research on the evolutionary ecology of microorganisms, we
provide an overview of some of the commonly used methods
in experimental evolution, along with studies that track evo-
lution in the wild using sequencing-based approaches. We
emphasize some of the major processes that are thought to
influence the strength of eco-evolutionary dynamics, provide
an overview of methods used to quantify the relative impor-
tance of ecology and evolution, and showcase the importance
of considering evolution in a community context and how
doi:10.1128/9781555818821.ch4.1.2
4.1.2-1
this may influence the dynamics and stability of microbial
systems under novel environmental conditions.
WHY STUDY THE EVOLUTIONARY ECOLOGY
OF MICROORGANISMS?
Whereas textbooks dealing with evolution and ecology tend
to highlight macroscopic organisms (e.g., insects, plants, and
fish) there are a number of important reasons why scientists
should consider the evolutionary ecology of microbes:
1. Microorganisms are diverse: Microorganisms make up the
vast majority of the planet’s biodiversity. Owing to recent
advances in sequencing technology, we now know that
most phyla in the tree of life are comprised of microbial
taxa. At local scales, the richness (a primary compo-
nent of α-diversity) of microbial taxa within a given a
given habitat (e.g., soils, ocean, gut) can be quite high.
It is not uncommon to recover thousands of bacterial
“species” from a single sample (10). In addition, there
is high compositional turnover (i.e., β-diversity) of
microbial communities in both time and space (11,
12). By convention, most scientists study the diversity
of bacterial and archaeal communities using operational
taxonomic units (OTUs), which are based on compara-
tive analysis of 16S rRNA gene sequences. Populations
whose 16S rRNA sequences are >97% similar are con-
sidered to be members of the same taxon. Although
this similarity cutoff correlates well with DNA-DNA
reassociation kinetics used to define microbial species
(13, 14), it underestimates the extensive microdiversity
commonly found within various groups of microorgan-
isms (15–17). Collectively, the standing genetic and
phenotypic variation found in microbial communities
provides a plethora of materials for ecological and evolu-
tionary processes to act on.
2. Microbes have high evolutionary potential: Owing to their
large population sizes and short generation times, micro-
organisms have the potential to evolve much faster than
plants and animals. In addition, microbes tend to live in
close proximity with one another (e.g., biofilms), which
allows them to share resources and by-products and estab-
lish coevolved, syntrophic interactions (18). In theory,
the lack of sexual reproduction should dramatically
reduce rates of evolution in species with finite population
4.1.2-2 ▪ THEORY
sizes, since sex, through recombination, accelerates
the rate at which multiple favorable mutations emerge
within a genome (19). However, homologous recombi-
nation occurs within and between microbial populations
(20, 21) and microorganisms can acquire novel sources
of genetic information through horizontal gene transfer
(22). Even at low frequencies, the vast population size
of microbes ensures that such mechanisms, in combina-
tion with mutation, generate large reservoirs of diversity
for evolutionary processes to act on.
3. Microbes are model systems for studying evolutionary ecology:
Compared with “macrobial” systems, microorganisms
have unique features that can readily be harnessed for
studying evolutionary ecology (23). With microorgan-
isms, one does not typically need to be concerned about
small population sizes, which can be important when
making inferences about evolutionary processes. More-
over, many of the taxa that are used in laboratory settings
have fairly short generation times, which is a requisite
for studying evolution in action. Although great progress
has been made in evolutionary ecology by studying
model organisms (e.g., Escherichia coli, Pseudomonas aeru-
ginosa, and Bacillus subtilis), an increasing number of
microorganisms can be isolated from the natural envi-
ronment and maintained under laboratory conditions
(24, 25). In some cases, these microbes are amenable
to genetic manipulation, which means that scientists
can explore the genetic underpinnings of phenotypic
traits using molecular tools such as recombineering.
In microbial systems where genetic manipulations are
not feasible, scientists are taking advantage of advances
in genomics, transcriptomics, proteomics, and metabo-
lomics to explore the eco-evolutionary complexities of
microbial communities (26–28). Together, these features
allow evolutionary ecologists to explore gene–gene inter-
actions (e.g., epistasis) along with fitness trade-offs that
tend to influence the strength of natural selection in
different environments.
A TRAITS-BASED APPROACH TO THE
EVOLUTIONARY ECOLOGY OF
MICROORGANISMS
One of the most important criteria for studying evolutionary
ecology is the ability to identify and quantify changes in
the functional traits of a focal population. Functional traits
can be defined as morphological, behavioral, or physiolo-
gical properties that influence the fitness of an individual
under a given set of conditions (29). These properties have
a genetic basis and are passed down from one generation to
the next (i.e., they are heritable). Measuring traits can be
fairly straightforward for some biologists. For instance, in
the textbook example of Darwin’s finches, the relative fre-
quency of beak sizes changes over time as a function of pre-
cipitation variability and the resulting distribution of seed
sizes (30). In principle, similar approaches can be applied to
microorganisms.
Quantifying traits that are under natural selection can be
challenging when studying microorganisms. Often the mor-
phological characteristics among divergent taxa, observed
using standard microscopy, appear identical. Other traits,
such as metabolic functions, can be measured under labora-
tory conditions, but the vast majority of microorganisms
are difficult to cultivate from natural environments. Conse-
quently, there are hurdles to studying the evolutionary
FIGURE 1 Conceptual diagram depicting feedbacks between eco-
logical and evolutionary processes. Within the domain of ecological
processes, there are interacting hierarchical levels of organization
(individuals, populations, communities, and ecosystems), which
can affect microevolutionary processes (i.e., anagenesis) and macro-
evolutionary processes (cladogenesis). Reciprocally, evolutionary
processes can affect ecological processes. The strength of these feed-
backs is influenced by the time scale at which ecological and evolu-
tionary processes take place and by factors such as mutation rates,
genetic drift, gene flow/disperal, and the diversity of a biological com-
munity. Adapted from (8), with permission.
doi: 10.1128/9781555818821.ch4.1.2.f1
ecology for most of the life on our planet. However, there is
a growing set of tools that can be used for studying microbial
traits. For example, it is now possible to visualize traits of
individuals, such as the capacity for nitrogen fixation, using
high-resolution nano-scale secondary ion mass spectrometry
(nanoSIMS) (31) or single-cell resource quotas using Raman
microspectroscopy (32). Similarly, the chemotactic behavior
of bacteria in relation to resource patches can be observed
using a combination of microfluidics and advanced image
analysis (33).
Genotypic features (a.k.a. genotypic traits) provide a
novel opportunity and potentially transformative way of char-
acterizing microbial traits (34). Although it is well established
that genetic information does not always translate directly
into an observable phenotype, the presence or absence of,
for example, a nifH gene will help predict whether an organ-
ism has the capacity to carry out nitrogen fixation. One of
the most commonly used high-throughput methods to date
involves marker gene analysis of the small subunit rRNA
gene. This type of approach characterizes the phylogenetic
diversity of a microbial sample in a cost-effective way. In
some cases phylogenetic gene markers can be a good proxy
for functional traits, but this is determined by the degree to
which a trait of interest is phylogenetically conserved (35).
Recent studies suggest that phylogenetic conservation in Bac-
teria and Archaea depends on trait complexity, with simpler
traits (e.g., glucose utilization) being more phylogenetically
dispersed than complex traits (e.g., methanogenesis) (36).
We are no longer restricted to making inferences about
microbial traits based on a single gene. For example, whole
genomes are now being used to gain eco-evolutionary insight
into the lifestyles of cultivated organisms (37). Furthermore,
we are increasingly able to identify relevant genotypic traits
using cultivation-independent approaches that rely on
gene inventories and their expression patterns derived from
nucleic acids (DNA and RNA) and proteins extracted from
environmental samples (27). For example, using techniques
such as single-cell genome amplification (38) or shotgun
metagenomics (39), it is now possible to reconstruct the
entire genomes of representative taxa directly from the
environment without cultivation. As the availability of
(near-complete) genome sequences continues to increase,
 4.1.2. Evolutionary Ecology of Microorganisms: From the Tamed to the Wild ▪ 4.1.2-3
we may eventually be able to revert to single marker genes as
reliable predictors of genotypic traits (40).
Nevertheless, it is still a challenge to link these genotypic
traits with the phenotypic traits on which natural selection
acts. One promising approach for identifying the genotypic
traits that underpin ecological differentiation, and thus
the phenotypic traits that affect fitness, combines genomic
and transcriptomic/proteomic analyses of closely related
populations sampled in their natural environments to detect
signatures of directional selection (41). These signatures refer
to evidence of positive selection, expression of population-
specific genes, and differential expression of shared genes
when two populations co-occur in the same environment.
Initial applications of such approaches have confirmed
laboratory-based findings regarding the important role that
the evolution of gene expression has in the early stages of eco-
logical differentiation (41). Extending these approaches to
time series analyses of either laboratory isolates or in situ pop-
ulations may help elucidate the microevolutionary under-
pinnings of fitness differences for microorganisms under
different environmental conditions.
EVIDENCE OF RAPID EVOLUTION IN
MICROBES: FROM THE LAB AND INTO
THE WILD
Experimental Evolution
Not long after the publication of On the Origin of the Species,
scientists began to design evolution experiments with micro-
organisms. In the 1870s, William Dallinger conducted selec-
tion experiments in which he challenged protozoa to
increasing temperatures (42), and by the middle of the 20th
century scientists were conducting studies that explored the
rapid evolution of virus resistance by bacteria (43, 44). Since
then, methods and approaches used to study the evolutionary
ecology of microbial populations have been refined. Argu-
ably, a new era of experimental evolution was initiated
by Richard Lenski and colleagues in the late 1980s. One
FIGURE 2 Relationship between phenotypic and genotypic change over time. Data originate from competing and evaluating fitness differ-
ences between ancestral and evolved E. coli lineages. While fitness increases saturate over time, fixed genetic changes continue to increase lin-
early over time. This pattern highlights some of the difficulties when trying to translate genotypic traits to phenotypic traits. Adapted from (48),
with permission. doi: 10.1128/9781555818821.ch4.1.2.f2
of the ongoing long-term experiments involves the semi-
continuous culturing of replicate (n = 12) E. coli populations.
Conceptually, the experiment is fairly straightforward: 1%
of cells from a culture are transferred into fresh medium
(glucose-supplemented minimal broth) on a daily basis. A
critical feature of most experimental evolution trials is the
ability to keep populations from different time points in
suspended animation. This is typically achieved by storing
cells (either single colony isolates or mixed populations)
in a cryoprotectant (e.g., glycerol or dimethyl sulfoxide) at
−80°C. The cyropreserved cells can then be resurrected
and used to make comparisons among ancestral and derived
lineages. For example, one might examine how traits such
as cell size, colony morphology, or the ability to use different
substrates changes over time (45, 46). Scientists can also use
this “fossil record” of cryopreserved isolates to ask questions
about how historical contingencies set the stage for the evo-
lution of novel traits (47). Experimental evolution trials
allow one to identify the genetic basis for neutral and adaptive
evolutionary change. For example, it is now possible to rese-
quence whole genomes of derived isolates from a long-term
experiment and identify mutations that arise compared
with an ancestral reference strain (48) (Fig. 2). This approach
can help reveal whether phenotypic changes are controlled
by mutations in structural genes or regulatory genes (the
latter is often found to be true). Transcriptomics is another
tool that is providing new insight into how populations
phenotypically evolve, for example, along environmental
gradients (49). Collectively, experimental evolution trials
allow one to estimate rates of neutral and adaptive evolution
within an experimental unit. Furthermore, because experi-
mental evolution trials are fairly easy to replicate, one can
assess the degree to which strains diverge, converge, or evolve
in parallel across experimental units (50).
It is common in experimental evolution studies to quan-
tify the relative fitness of a derived population to the ances-
tral population. This is typically achieved by conducting
head-to-head experiments in which two populations are
mixed and allowed to compete for a given amount of time.
4.1.2-4 ▪ THEORY
Growth rates can be estimated as [ln(Ntf )/ln(Nt0)]/t where ln
(Ntf ) is the natural log of cell densities of a given population
at the end of a competition experiment that runs for time t
and ln(Nt0) is the natural log of cell densities of a given pop-
ulation at the beginning of the experiment. From this, relative
fitness can be estimated as the ratio of growth rates for the
derived and ancestral population, respectively. However,
when two strains are mixed, it can be difficult to differentiate
the competing cell lines. This complication to estimating rel-
ative fitness can be overcome through the use of a marker
gene that provides a means of selecting or distinguishing dif-
ferent populations. For example, one could select for neutral
markers, such as lactose utilization, and then enumerate via
plating with and without lactose (51). Other strategies might
involve insertion of green fluorescent protein or selection for
antibiotic resistance (52), but researchers must be aware of
how the associated fitness costs of a marker could potentially
confound inferences that would be made about evolutionary
trajectories. Another strategy is to compete ancestral and
derived cell lines against a third-party “tester” strain (53,
54), but scientists must be comfortable with the assumption
that the tester strain interacts with the ancestral and derived
strains in ecologically similar ways.
Over time, the traditional approach to experimental
evolution with E. coli (55) has expanded to accommodate dif-
ferent taxa, environmental conditions, and species interac-
tions. In addition to batch cultures, microbiologists can set
up experimental evolution trials using continuous cultures
or chemostats. One benefit of using chemostats is that there
is a constant inflow of fresh media, which means that microbes
do not experience fluctuations in physiological conditions
that are typical of a batch culture environment. Second,
by altering medium composition or environmental con-
ditions, researchers have the ability to closely control the
growth-limiting factor (e.g., nitrogen, phosphorus, light) of
a population in a chemostat. Third, mathematical theory
has been developed and applied to microbes in the chemostat
environment (56), which allows researchers to identify key
parameters, such as resource uptake or predation defense,
that are under selection (57). Although chemostats are ideal
for studying evolution of planktonic microorganisms, contin-
uous culture techniques have also been developed for study-
ing biofilm-forming strains (58). Other creative variations
have been used to study evolution in environments that
deviate from the assumptions of spatial homogeneity in the
chemostats. For example, through the use of liquid-handling
robotics on 96-well plates, researchers have been able to
simulate eco-evolutionary dynamics that occur when species
move among patches in heterogeneous landscapes (59).
Evolutionary Ecology in the Wild
Laboratory-based studies have contributed immensely to
our basic understanding of microbial evolutionary ecology.
However, it is not clear whether the processes contributing
to, for example, the rise to dominance of specific genetic
variants are similar under laboratory and natural conditions.
For example, a recent study demonstrated that the adaptive
diversification of Pseudomonas fluorescens was greatly reduced
via interactions with a diverse soil microbial community (60).
These types of evolutionary dynamics are highly dependent
on the population–genetic environment (e.g., the impor-
tance of genetic drift owing to effective population size) along
with other chemical, physical, and biological processes,
which are almost certainly more variable and less controllable
in nature than in the laboratory.
In particular, gene flow represents major challenges when
studying evolutionary ecology of free-living microorganisms
in nature. Immigration may be less of a concern when study-
ing relatively “closed” environments, such as acid mine drain-
age (AMD) ecosystems, which are biogeographically isolated
from other source populations that are adapted to such unique
conditions (e.g., low pH and high metal concentrations).
After reconstructing the genome of a bacterial population
in one of these AMD sites, researchers were able to track
the accumulation of fixed single nucleotide polymorphisms
(SNPs) over time (Fig. 3). From this, they were able to esti-
mate an evolutionary rate of 1.3 × 10−9 substitutions per
nucleotide per generation, which is similar to rates reported
in many lab experiments (61). Using the AMD as a model sys-
tem, researchers were able to reconstruct the timeline of
recent divergence events and demonstrate the rise of domi-
nance for mutations in different lineages resulting from posi-
tive selection and drift. Similar patterns of periods of positive
selection alternating with periods of drift were observed in a
study tracking the evolution of Pseudomonas aeruginosa in
the lungs of cystic fibrosis patients (62).
When studying evolution in the wild, just as in laboratory
studies, researchers need to determine the relative importance
of genetic drift and positive selection on the rise in domi-
nance of particular variants. This can be accomplished by cal-
culating dN/dS ratios for the genes affected by mutations
(63). The dN/dS ratio, which can be applied to specific loci
or entire genomes, calculates the number of nonsynonymous
mutations across all available nonsynonymous sites relative
to the number of synonymous mutations across all available
synonymous sites. It is becoming more common to estimate
dN/dS ratios using metagenomic data from environmental
samples (64, 65). Care has to be taken, however, when inter-
preting the dN/dS ratio for a population because the metric
makes assumptions that are only valid for comparisons
between more distantly diverged organisms (66). Methods
are available to correctly assess the directionality ( positive,
negative, neutral) of selection (67, 68), but microbiologists
must be aware that high error rates associated with different
sequencing technologies will be misinterpreted as mutations
(69). Nevertheless, the dN/dS ratio can provide clear insight
into the relative importance of evolutionary processes in
some instances. For example, tight population bottlenecks
between insect generations resulted in strong effects of
genetic drift on a bacterial endosymbiont (Buchnera), which
resulted in rapid reductive genome evolution (70).
It is well established from the study of microbial isolates
that homologous recombination and lateral gene transfer
are important processes that influence microbial divergence
(20, 22, 71). Through the use of environmental genomics
(metagenomics), it has been shown that these evolutionary
processes are also important for the generation of population-
level diversity. In particular, metagenomic studies are starting
to answer outstanding questions regarding the relative impor-
tance of recombination and mutation (21, 72, 73). To fully
document the nature and rate of introduction of new genes
into genomes over time, it is critical to reconstruct (nearly)
complete population genomic data sets for each time point.
Two recent time-series analyses of the gut colonization of
preterm infants show the potential of metagenomics to track
the varying gene content of closely related microorganisms
and relate it to the varying abundances of these strains over
time (74, 75). Comprehensively tracking the flow of genes
in and out of populations remains an unmet challenge how-
ever, which may be aided by emerging longer-read DNA
sequencing technology.
 4.1.2. Evolutionary Ecology of Microorganisms: From the Tamed to the Wild ▪ 4.1.2-5

FIGURE 3 Determining rates of evolutionary in the wild. (a) Samples were collected from one location in the AMD system (C75) and de
novo sequence assembly of sequencing reads led to the reconstruction of a genome for the dominant Leptospirillum group II at the site (type III).
(b) Read recruitment of all 13 sequence data sets generated from C75 samples over 5 years to the type III reference genome allowed for the
identification of additional fixed mutations and estimation of the nucleotide substitution rate. Lower frequency mutations could be observed
in each of the data sets as well, but only fixed variants are included for rate calculations. Adapted from (61), with permission.
doi: 10.1128/9781555818821.ch4.1.2.f3

SPATIAL SCALE AND THE EVOLUTIONARY
ECOLOGY OF MICROBES
The example of the AMD system (Fig. 3) is unique because
we can assume that the immigration and establishment of
novel genotypes from similar ecosystems is rare. In more
open natural systems, spatial processes are critical for under-
standing the evolutionary ecology of microorganisms. The
movement of individuals and the resulting gene flow between
subpopulations can have strong effects on allele frequencies
and the evolutionary trajectory of the local and metapopula-
tion (i.e., the collection of geographically separated but inter-
acting populations of a species). Specifically, reductions in
gene flow increase divergence between isolated populations,
which in turn can lead to speciation via selection or drift.
Migration (or dispersal) is also an important ecological proc-
ess that can influence the assembly of communities (76). For
example, dispersal limitation may contribute to high levels of
β-diversity (i.e., high compositional turnover among sites),
while high rates of dispersal can create “mass effects” that
allow for the persistence of competitively inferior species in
a local community (77).
Owing to their small size, it is assumed that microbes
can be carried long distances via passive mechanisms or
through close association with larger host organisms.
Through the use of analog microspheres, it has been shown
that microbial-sized particles can be transported up to 2 km
within days depending on weather conditions (78). In other
studies, it is estimated that a bacterial cell in the atmosphere
has a residence time of 2–15 days (79), which in some cases
can lead to the continental-scale dispersion of microorgan-
isms (80). These high dispersal rates could result in the
cosmopolitan distribution of microbial populations. How-
ever, multiple lines of evidence suggest that this is not entirely
the case. Using multilocus sequencing of hyperthermophilic
Archaea, it was shown that a Sulfolobus sp. had high FST
values (a population genetic index that quantifies the var-
iance in allele frequencies between populations), consistent
4.1.2-6 ▪ THEORY
FIGURE 4 Pairwise sequence divergence of Sulfolobus populations isolated from a global survey of hot springs ecosystems scales positively
with geographic distance providing evidence against the view of panmicitic microbial distributions. Adapted from (81), with permission.
doi: 10.1128/9781555818821.ch4.1.2.f4
with the view that there was minimal mixing among hot
spring environments spanning a 6000 km sample gradient
(81). Pairwise genetic divergences estimated from Sulfolobus
isolates were positively correlated with geographic distance,
providing further evidence that not all microorganisms
have panmictic distributions (Fig. 4).
A classic way to examine the spatial patterns of bio-
diversity for entire communities is through the construction
of species–area relationships. These relationships describe
diversity with the power function: S = cA z, where S is species
richness, A is area, and c and z are constants. When S and A
are plotted on a log–log scale, the slope, z, can be used to
quantify the rate at which new species are encountered
with increasing sampling area. When estimated for microbes,
z-values tend to be much lower than they are for macroscopic
organisms (e.g., plants and animals), but significantly greater
than zero (82). It has been hypothesized that these patterns
arise from dispersal limitation, but they could also be
attributed to other factors, including the fact that environ-
mental heterogeneity tends to scale positively with geo-
graphic distance (82). In a recent meta-analysis, geographic
distance was found to have a significant effect on microbial
composition in half of the studies. Approximately 10% of
the observed variance could be uniquely attributed to geo-
graphic distance while ∼25% was uniquely attributed to
measured environmental factors and ∼15% to combined
effects (83).
If microorganisms experience dispersal limitation in
patchy environments, we should expect to find evidence
for local adaptation in at least some microbial populations.
Local adaptation occurs when the performance or fitness of
an individual is higher in its “home” versus “away” environ-
ment. Evidence for local adaptation is often obtained from
transplant studies and suggests that the strength of selection
caused by local conditions exceeds the strength of gene
flow. To test for local adaptation, heterotrophic soil bacteria
were isolated from multiple sites within a 1 ha old-growth
forest and cultured in soil medium derived from local and
distant sites (84). When the authors focused on fast-growing
isolates, they found that bacteria had the highest fitness on
locally derived medium and fitness decayed exponentially
on media derived from more distant sites (Fig. 5). Such
findings led to the conclusion that edaphic heterogeneity
and limited dispersal, relative to evolutionary rates, created
complex fitness landscapes for bacteria at relatively small
spatial scales. Microorganisms may also show signs of local
adaptation to the types of organisms with which they interact.
For example, many bacteria have to contend with the selec-
tive pressures caused by predation and parasitism. It is known
that many bacterial populations can evolve resistance to
phage, but less is understood about how this evolutionary
adaptation plays out over larger spatial scales. Such questions
form the basis of the geographic mosaic theory of coevolution
(85), which has been addressed using laboratory experiments
(86), and also natural communities. For example, bacteria
and phage were isolated from 25 cm × 25 cm grids for two
soil samples that were separated by 100 m (87). The bacterial
isolates were then challenged with co-occurring and geo-
graphically distant phage populations. On average, phages
were 9% more infective on their local bacterial hosts. Phage
fitness diminished when challenged with bacteria that were
only centimeters away, suggesting that viruses may be ahead
of bacteria in the coevolutionary arms race and that biotic
eco-evolutionary interactions are not always swamped out
by rampant dispersal.
TEMPORAL SCALE AND THE EVOLUTIONARY
ECOLOGY OF MICROBES
We have pointed out in this chapter that microorganisms
attain large population sizes, can have short doubling times,
and in some cases can exchange genes with distantly related
taxa. Combined, these characteristics set the stage for evolu-
tion to occur on ecologically relevant or rapid time scales.
 4.1.2. Evolutionary Ecology of Microorganisms: From the Tamed to the Wild ▪ 4.1.2-7
FIGURE 5 Evidence for local adaptation demonstrating the
distance decay for the relative fitness of soil bacteria grown on
resources from different geographic locations. Adapted from (84),
with permission. doi: 10.1128/9781555818821.ch4.1.2.f5
Perhaps the best evidence of this comes from the study of
E. coli in batch culture. After being inoculated into fresh
medium, E. coli enters exponential growth phase within
hours. During this time, cells grow at their maximum poten-
tial and rapidly deplete resources. As a result, per capita
growth rates decline and E. coli enters a stationary phase,
which is followed shortly thereafter (2–5 days) by a death
phase, in which population densities decline by about an
order of magnitude. Intriguingly, cell densities can remain
fairly constant after the death phase for extended periods
of time (years), due partly to a phenomenon referred to as
growth advantage in stationary phase (GASP) (Fig. 6)
(88). Although the aggregate population appears relatively
stable, bacteria are extremely dynamic during periods of
prolonged starvation. Ecologically, this can be attributed to
the fact that some individuals die and release their cellular
constituents back into the environment, while other individ-
uals assimilate this material along with other metabolic
by-products for growth and reproduction. Evolutionarily it
has been shown that cannibalistic subpopulations are variants
that arise and invade the system in a negative frequency-
dependent manner (89). GASP-related research has led
to the prevailing view that starvation is not only a strong
selective agent but also alters the rates of de novo mutation
either through methyl-directed mismatch repair or the SOS
response, which activates error-prone polymerases (e.g.,
PolIV and PolV) (88, 90). The GASP phenomenon demon-
strates that starvation stress is a proximal cue that leads to the
accumulation of beneficial mutation (88), while also provid-
ing an explanation for the persistence of the population under
resource-limited conditions (Fig. 6).
Microorganisms can contend with unfavorable conditions
(including starvation) by hedging their bets and entering a
reversible state of reduced metabolic activity or dormancy
(Fig. 7). Dormancy has evolved many different times in the
tree of life and is a functional trait that allows genotypes or
FIGURE 6 Some bacteria can rapidly evolve in response to starva-
tion. The upper panel shows a typical growth curve of E. coli. When
populations deplete resources, they enter stationary phase followed
by a death phase. Subsequently, E. coli (and other types of bacteria)
can enter growth advantage in stationary phase (GASP), where
novel starvation-resistant mutants evolve and invade a system as
depicted by the colored curves in the top panel (adapted from
(88), with permission) and the conceptual model in the lower panel.
doi: 10.1128/9781555818821.ch4.1.2.f6
even entire populations to avoid extinction. For example, via-
ble microorganisms have been retrieved from ancient materi-
als (e.g., permafrost and amber) that in some cases are
hundreds of millions of years old (91). The resurrection of
populations from so-called seed banks has obvious evolution-
ary implications, but is also important for maintenance of
biodiversity and community functioning (92). There are a
variety of ways to estimate dormancy in microbial commun-
ities. Some taxa produce spores, cysts, or akinetes when
they enter inactive states, but these morphological traits are
not reliable indicators for dormancy for all microorganisms.
Single-cell assays based on fluorescent in situ hybridization
or the uptake of tetrazolium stains can be useful for estimating
the activity of microbial cells (93). Recently, inferences about
the metabolic activity of bacteria have been made by examin-
ing the 16S region of ribosomal RNA genes (rDNA) and
FIGURE 7 When challenged with conditions that are suboptimal
for growth and reproduction, some microorganisms enter a reversible
state of reduced metabolic activity or dormancy. The size of the active
population is determined by the net reproductive rates, losses due to
mortality, and losses due to dormancy. The size of the dormant pop-
ulation is determined by the rate at which active individuals transi-
tion into dormancy, the mortality rate during dormancy, and
resuscitation from dormancy. This bet-hedging strategy is important
for the maintenance of microbial biodiversity. Adapted from (94),
with permission. doi: 10.1128/9781555818821.ch4.1.2.f7
4.1.2-8 ▪ THEORY
ribosomal RNA (rRNA) (94). Justification for this approach
is as follows: in general, rDNA is a stable molecule that is
widely used to infer the presence (and potential activity) of
a population. In contrast, rRNA is an ephemeral molecule
that is only produced by growing cells, which require ribo-
somes for protein synthesis (95). As such, rRNA has been
used for identifying active taxa in complex microbial
communities (e.g., (96)). Although RNA:DNA is strongly
correlated with microbial growth rates in lab settings (97),
concerns have been raised about applying this technique to
broad ranges of taxa (98). An alternate approach is to focus
on genes (e.g., toxin-antitoxin modules or resuscitation-
promoting factors) that are directly involved in the transi-
tions between active and dormant metabolic states (92).
Last, epigenetic processes can also affect the temporal
scale of eco-evolutionary processes by allowing organisms to
rapidly respond to environmental signals and pass this
response on to their offspring (99). “Epigenetic processes”
refers to nongenetic mechanisms (i.e., not directly related
to differences in nucleotide sequence) that cause variability
in gene expression that can result in phenotypic variation
subject to natural selection. While a variety of systems are
referred to as epigenetic, the best studied one is based on
DNA methylation and interactions with histone proteins.
Histone–DNA interactions condense DNA and render these
stretches of the DNA unavailable for transcription, effec-
tively shutting down gene expression. Epigenetic marks
(methylations) are accrued during an organism’s life in res-
ponse to environmental or developmental cues, are reversi-
ble, and, importantly, are heritable. Although epigenetic
studies have mostly focused on eukaryotes, the mechanism
is relevant in bacteria as well (100, 101) and genome-wide
determination of methylation patterns can readily be per-
formed using sequencing approaches (102). Although this
area is relatively new, and the implications on evolutionary
and ecological processes are still unclear, there is evidence
that phenotypic variation between bacterial subpopulations
of the same species can be caused by heritable variability in
DNA methylation patterns. Methylation plays an important
role as a signal for a variety of bacterial cellular processes; for
example, repair enzymes use them to differentiate the original
(methylated) template DNA strand versus the newly (tempo-
rarily unmethylated) copied strand during replication. Main-
taining stretches of DNA in the hemi- or unmethylated state
beyond the replication phase can affect gene expression and
has been shown to be the mechanism for several phase-
variable phenotypes, including the expression of pili in uro-
pathogenic E. coli (101). Creating subpopulations that
diverge in the expression of phase-variable genes that affect
important functional traits can be seen as another example
of bet hedging. The ability to transmit a fitness-affecting
phenotype acquired through epigenetic modifications can
influence the evolution of a lineage in multiple ways and is
another mechanism to keep in mind when determining the
impacts of eco-evolutionary dynamics on microbial systems.
ECO-EVOLUTIONARY FEEDBACKS IN
MICROBIAL SYSTEMS
We have emphasized that microbial communities are taxo-
nomically, phylogenetically, and metabolically diverse. We
have also shown that microorganisms have the capacity to
evolve on ecologically relevant time scales. Together, these
features set the stage for eco-evolutionary feedbacks. From
the ecological side of the feedback, it is well established
that species interactions (e.g., competition, parasitism, or
mutualisms) can affect evolutionary processes such as adapta-
tion and speciation (Fig. 1). From the evolutionary side, evo-
lutionary changes (e.g., selection for traits) can modify
population dynamics, species interactions, and even ecosys-
tem processes (103). Over the past decade, evidence has
been accumulating that eco-evolutionary feedbacks are
important for understanding plant and animal dynamics (3,
104, 105). In the last section of this chapter, we highlight
examples where eco-evolutionary feedbacks are important
for understanding how microbes interact with each other
and their hosts.
Feedbacks Involving Antagonistic Interactions
Antagonistic interactions between predators and prey or hosts
and parasites can often give rise to eco-evolutionary feed-
backs. In microbial systems, clear evidence of this can be
found when studying bacteria–phage dynamics. Lytic phages
can reduce the population size of sensitive bacteria by orders
of magnitude within a short period of time. This strong
top-down force creates strong selective pressure for phage
resistance. Bacteria have evolved various ways of resisting
phage attacks, including the modification of surface recep-
tors, DNA restriction-modification systems, and clustered
regularly interspaced short palindromic repeat immunity
(106). It is generally assumed that the benefits afforded by
the specialization of phage resistance come at a cost (54,
107). For example, the loss or configuration change of a recep-
tor molecule that interferes with phage attachment to the cell
surface can also reduce rates of resource uptake (108).
The fitness costs associated with predator defense traits are
critical for understanding microbial population dynamics
involving eco-evolutionary feedbacks. For example, the cost
of resistance establishes a trade-off between phage defense
and resource competition that allows for coexistence of bacte-
rial variants and the ancestral phage population. Both models
and empirical evidence indicate that microbial population
dynamics are highly sensitive to this type of trade-off. In a
chemostat study of a eukaryotic alga (Chlorella) and a preda-
tory rotifer (Brachionus), it was shown that periodic selection
on resource acquisition and predator defense led to surprising
population dynamics (109). Specifically, the authors anti-
cipated relatively fast cycles where peaks in predator abun-
dances tracked peaks in prey abundances by one quarter of a
cycle as predicted by general ecological theory. Instead,
they found that the cycles were much slower. Moreover, the
predator and prey densities were almost exactly out of phase
(109). Subsequently, it was put forth that trophic interactions
can be masked by rapid evolution caused by antagonistic spe-
cies interactions giving rise to “cryptic” population dynamics
(6). These types of controlled studies may help explain the
absence of classic predator–prey cycles between bacteria
and phages in natural systems when analyzing data at a coarse
phylogenetic resolution (110).
Rapid evolution caused by antagonistic species interac-
tions can also affect ecosystem processes. Phages are highly
abundant in the open ocean and can account for a substan-
tial fraction of bacterial mortality. Indirectly, phages are
thought to increase the concentration of carbon and nutrient
in the oceans by reducing microbial population sizes. In
addition, phage lysis events are directly responsible for releas-
ing labile resources into the environment, which can affect
global biogeochemical cycles through a process known as
the viral shunt (111). How might rapid evolutionary change
affect the viral shunt? This question was explored in a
chemostat experiment with Synechococcus, a marine picocya-
nobacterium, and an infectious phage (112). Synechococcus
 4.1.2. Evolutionary Ecology of Microorganisms: From the Tamed to the Wild ▪ 4.1.2-9
population densities plummeted after the initial phage
attack, which led to significant increases in phosphorus and
alterations of the elemental stoichiometry of microbial
biomass. However, these effects of phage on nutrient
cycling diminished with time because of the evolution of
phage-resistant bacteria. These laboratory results with envi-
ronmental isolates suggest that rapid evolution may be impor-
tant when attempting to understand and model the impacts
of viruses on microbial food webs.
Feedbacks Involving Mutualistic Interactions
Although historically overlooked, mutualistic interactions
can be important drivers of eco-evolutionary dynamics.
Many microbial taxa engage in mutualistic interactions,
either with other microbes, or with plants and animals. These
mutualisms range from relatively loose associations to obligate
endosymbioses. In the case of endosymbionts and their hosts,
coevolution and codifferentiation ( parallel evolutionary
paths of symbionts and hosts) have been occurring for mil-
lions of years (113). But how dynamic are these interactions
on ecological time scales?
Growing evidence suggests that many microbial-based
mutualisms have the potential to evolve rapidly. For exam-
ple, experimental evolution trials were conducted with a
sulfate-reducing bacterium (Desufovibrio vulgaris) and a
methanogenic archaeon (Methanococcus maripaludis), two
isolates that had no known history of interaction (114).
Although both populations could be grown in pure culture,
the authors attempted to establish an obligate syntrophic
mutualism by growing the strains in lactate medium in rep-
licate (n = 24) coculture. Initially, growth of the cocultures
was unstable, leading to the extinction of one of the popu-
lations. In only 300 generations, the evolved cocultures grew
up to 80% faster and produced 30% more biomass, which
was the result of evolution by both partners. The stability
of this novel mutualism, however, was challenged by muta-
tions that gave rise to more antagonistic variants. In addi-
tion, the stability of the mutualism was influenced by the
heterogeneity of the environment. Specifically, contribu-
tions of the methanogenic partner to the performance of
the community were greater in heterogeneous environments
(nonshaken flaks) than homogeneous environments
(shaken flasks), presumably due to the increased exchange
of substrates among mutualists. This study uniquely demon-
strates the power of controlled experiments to investigate
how metabolism, habitat features, and behaviors such as
cheating might influence the development and stability of
microbial mutualisms.
Microorganisms can also readily establish mutualistic
relationships with plant and animal populations. This has
become an important topic of research given concerns about
the accelerating rate of global change. Some species may be
able to persist in novel or changing environment through
ecological strategies such as phenotypic plasticity, behavioral
modifications, or migration to more favorable habitats. A
second strategy is for the plant or animal population to adapt
to new conditions, but it remains unclear whether macro-
scopic organisms have the capacity to evolve at a fast enough
pace to keep up with environmental change (115). A third
strategy is for plant and animal populations to “outsource”
adaptive traits to symbiotic microorganisms. This concept
has been articulated in the hologenome theory of evolution
(116). The major tenets of this theory are that (a) all plants
and animals establish symbiotic relationships with microbes,
(b) symbiotic microorganisms can be vertically transmitted
via different mechanisms, (c) microbe–host interactions
affect the fitness of the “holobiont,” and (d) genetic varia-
tion of the holobiont can be enhanced through the rapid
recruitment of microorganisms from diverse communities.
The hologenome theory of evolution was initially developed
to help explain a coral-bleaching phenomenon. Specifically,
because corals lack adaptive immunity, it was hypothesized
that they could recruit beneficial microorganisms from
the marine environment to prevent infection from patho-
genic bacteria (117). Since then, some of these ideas have
been tested in other systems as well. For example, when chal-
lenged by drought stress for multiple generations, reciprocal
transplant experiments revealed that plant fitness was
strongly affected by the rapid shifts in soil microbial com-
munities (118). The hologenome theory of evolution may
also have important implications for understanding macroe-
volutionary processes. For example, within just a few gener-
ations, diet-driven shifts in the composition of commensal
bacteria altered the mating preferences of Drosophila mela-
nogaster, which could lead to prezygotic reproductive isola-
tion (119). In addition, it was recently shown that hybrid
lethality among closely related wasp species (Nasonia sp.)
was due to negative epistasis (i.e., mismatched gene–gene
interactions) between the host genome and the gut micro-
biome (120).
CONCLUSION
Over the past 50 years, biologists’ views regarding the inter-
play between ecology and evolutionary biology have dramat-
ically changed (105). For example, Dobzhansky famously
stated that “Nothing in biology makes sense except in the
light of evolution,” while Peter and Rosemary Grant retorted
with “Nothing in evolutionary biology makes sense except in
the light of ecology.” More recently, it seems we have arrived
at the notion that “Nothing in evolution or ecology makes
sense except in the light of the other” (121). We argue that
microbiologists are uniquely poised to make advances to the
field of evolutionary ecology. In fact, major advances have
already been made owing in large part to the amenability of
microbial systems to laboratory-based study. While it is con-
ceivable that all of these findings are relevant to in situ con-
ditions, laboratory experiments deviate from real-world
systems in temporal and spatial scale, and in the level of com-
plexity of ecological interactions. In this chapter, we have
highlighted (a) that evolutionary rates and processes are sim-
ilar in the laboratory and in the wild; (b) that in laboratory
settings, ecological and evolutionary processes occur on sim-
ilar time scales, and both need to be taken into account to
explain experimental observations; (c) what is currently
known regarding temporal and spatial processes that may
impact in situ eco-evolutionary feedbacks; and (d) some
examples where eco-evolutionary feedbacks have been shown
to be relevant in the wild.
Similar to plant and animal ecologists and evolutionists,
we are only at starting to answer the question of how relevant
eco-evolutionary feedbacks are in understanding community
structure and functional stability. As summarized in the first
section of this chapter, the nature of microbial systems may
give us the chance to acquire insights much faster, contribu-
ting not only to our own field’s progress but also to the under-
standing of universal eco-evolutionary principles, applying to
all forms of life.
This research was supported in part by the National Science Founda-
tion (1442246 to J.T.L.) and the U.S. Army Research Office
(W911NF-14-1-0411 to J.T.L.).
4.1.2-10 ▪ THEORY
REFERENCES
1. Hendry AP, Kinnison MT. 1999. Perspective: the pace of
modern life: Measuring rates of contemporary microevolu-
tion. Evolution 53:1637–1653.
2. Thompson JN. 1998. Rapid evolution as an ecological
process. Trends Ecol Evol 13:329–332.
3. Fussmann GF, Loreau M, Abrams PA. 2007. Eco-
evolutionary dynamics of communities and ecosystems.
Funct Ecol 21:465–477.
4. Bassar RD, Marshall MC, Lopez-Sepulcre A, Zandona E,
Auer SK, Travis J, Pringle CM, Flecker AS, Thomas SA,
Fraser DF, Reznick DN. 2010. Local adaptation in Trini-
dadian guppies alters ecosystem processes. Proc Natl Acad
Sci USA 107:3616–3621.
5. Matthews B, Narwani A, Hausch S, Nonaka E, Peter H,
Yamamichi M, Sullam KE, Bird KC, Thomas MK,
Hanley TC, Turner CB. 2011. Toward an integration of
evolutionary biology and ecosystem science. Ecol Lett 14:
690–701.
6. Yoshida T, Ellner SP, Jones LE, Bohannan BJM, Lenski
RE, Hairston NG. 2007. Cryptic population dynamics:
Rapid evolution masks trophic interactions. PLoS Biol 5:
1868–1879.
7. Fox CW, Roff DA, Fairbairn DJ. 2001. Evolutionary Ecol-
ogy: Concepts and Case Studies.Oxford University Press,
Oxford.
8. Mayhew PJ. 2006. Discovering Evolutionary Ecology: Bring-
ing Together Ecology and Evolution, vol. 232. Oxford Univer-
sity Press, Oxford.
9. Pianka ER. 2011. Evolutionary Ecology, 7th ed. (ebook).
PiankaEric R.
10. Fierer N, Lennon JT. 2011. The generation and mainte-
nance of diversity in microbial communities. Am J Botany
98:439–448.
11. Martiny JBH, Bohannan BJM, Brown JH, Colwell RK,
Fuhrman JA, Green JL, Horner-Devine MC, Kane M,
Krumins JA, Kuske CR, Morin PJ, Naeem S, Ovreas L,
Reysenbach AL, Smith VH, Staley JT. 2006. Microbial
biogeography: putting microorganisms on the map. Nat
Rev Microbiol 4:102–112.
12. Shade A, Caporaso JG, Handelsman J, Knight R, Fierer
N. 2013. A meta-analysis of changes in bacterial and arch-
aeal communities with time. ISME J 7:1493–1506.
13. Amann RI, Lin CH, Key R, Montgomery L, Stahl DA.
1992. Diversity among Fibrobacter isolates—towards a phy-
logenetic classification. System Appl Microbiol 15:23–31.
14. Goris J, Konstantinidis KT, Klappenbach JA, Coenye T,
Vandamme P, Tiedje JM. 2007. DNA-DNA hybridization
values and their relationship to whole-genome sequence
similarities. Int J System Evol Microbiol 57:81–91.
15. Ramette A, Tiedje JM. 2007. Multiscale responses of
microbial life to spatial distance and environmental heter-
ogeneity in a patchy ecosystem. Proc Natl Acad Sci USA
104:2761–2766.
16. Hunt DE, David LA, Gevers D, Preheim SP, Alm EJ,
Polz MF. 2008. Resource partitioning and sympatric differ-
entiation among closely related bacterioplankton. Science
320:1081–1085.
17. Dethlefsen L, McFall-Ngai M, Relman DA. 2007. An
ecological and evolutionary perspective on human-
microbe mutualism and disease. Nature 449:811–818.
18. Hillesland KL, Stahl DA. 2010. Rapid evolution of stabil-
ity and productivity at the origin of a microbial mutualism.
Proc Natl Aca Sci USA 107:2124–2129.
19. Otto SP. 2009. The evolutionary enigma of sex. Am Nat
174:S1–S14.
20. Vos M, Didelot X. 2008. A comparison of homologous
recombination rates in bacteria and archaea. ISME J 3:
199–208.
21. Eppley JM, Tyson GW, Getz WM, Banfield JF. 2007.
Genetic exchange across a species boundary in the archaeal
genus Ferroplasma. Genetics 177:407–416.
22. Ochman H, Lawrence JG, Groisman EA. 2000. Lateral
gene transfer and the nature of bacterial innovation. Nature
405:299–304.
23. Elena SF, Lenski RE. 2003. Evolution experiments with
microorganisms: the dynamics and genetic bases of adapta-
tion. Nat Rev Genet 4:457–469.
24. Kaeberlein T, Lewis K, Epstein S. 2002. Isolating “uncul-
tivable” microorganisms in pure culture in a simulated
natural environment. Science 296:1127.
25. Giovannoni S, Stingl U. 2007. The importance of cultur-
ing bacterioplankton in the “omics” age. Nat Rev Microbiol
5:820–826.
26. Raes J, Bork P. 2008. Molecular eco-systems biology:
towards an understanding of community function. Nat
Rev Microbiol 6:693–699.
27. DeLong EF. 2009. The microbial ocean from genomes to
biomes. Nature 459:200–206.
28. Denef VJ, Mueller RS, Banfield JF. 2010. AMD biofilms:
using model communities to study microbial evolution and
ecological complexity in nature. ISME J 4:599–610.
29. Lennon JT, Aanderud ZT, Lehmkuhl BK, Schoolmaster
DR Jr, 2012. Mapping the niche space of soil microorgan-
isms using taxonomy and traits. Ecology 93:1867–1879.
30. Grant PR. 1999. Ecology and evolution of Darwin’s finches.
Princeton University Press, Princeton, NJ.
31. Foster RA, Kuypers MMM, Vagner T, Paerl RW,
Musat N, Zehr JP. 2011. Nitrogen fixation and transfer
in open ocean diatom-cyanobacterial symbioses. ISME J
5:1484–1493.
32. Hall EK, Singer GA, Poelzl M, Haemmerle I, Schwarz C,
Daims H, Maixner F, Battin TJ. 2011. Looking inside
the box: using Raman microspectroscopy to deconstruct
microbial biomass stoichiometry one cell at a time. ISME
J 5:196–208.
33. Seymour JR, Ahmed T, Durham WM, Stocker R. 2010.
Chemotactic response of marine bacteria to the extracellu-
lar products of Synechococcus and Prochlorococcus. Aquat
Microbial Ecol 59:161–168.
34. Green JL, Bohannan BJM, Whitaker RJ. 2008. Microbial
biogeography: from taxonomy to traits. Science 320:
1039–1043.
35. Philippot L, Andersson SGE, Battin TJ, Prosser JI, Schi-
mel JP, Whitman WB, Hallin S. 2010. The ecological
coherence of high bacterial taxonomic ranks. Nat Rev
Microbiol 8:523–529.
36. Martiny AC, Treseder K, Pusch G. 2013. Phylogenetic
conservatism of functional traits in microorganisms.
ISME J 7:830–838.
37. Livermore JA, Emrich SJ, Tan J, Jones SE. 2013. Fresh-
water bacterial lifestyles inferred from comparative
genomics. Environ Microbiol 16:746–758.
38. Stepanauskas R, Sieracki ME. 2007. Matching
phylogeny and metabolism in the uncultured marine bacte-
ria, one cell at a time. Proc Natl Acad Sci USA 104:
9052–9057.
39. di Rienzi SC, Sharon I, Wrighton KC, Omry K, Hug LA,
Thomas BC, Goodrich JK, Bell JT, Spector TD, Banfield
JF, Ley RE. 2013. The human gut and groundwater harbor
non-photosynthetic bacteria belonging to a new candidate
phylum sibling to Cyanobacteria. eLife 2:e01102.
40. Langille MGI, Zaneveld J, Caporaso JG, McDonald D,
Knights D, Reyes JA, Clemente JC, Burkepile DE,
Thurber RLV, Knight R, Beiko RG, Huttenhower C.
2013. Predictive functional profiling of microbial com-
munities using 16S rRNA marker gene sequences. Nat
Biotechnol 31:814.
 4.1.2. Evolutionary Ecology of Microorganisms: From the Tamed to the Wild ▪ 4.1.2-11
41. Denef VJ, Kalnejais LH, Mueller RS, Wilmes P, Baker
BJ, Thomas BC, Verberkmoes NC, Hettich RL, Banfield
JF. 2010. Proteogenomic basis for ecological diver-
gence of closely related bacteria in natural acidophilic
microbial communities. Proc Natl Acad Sci USA 107:
2383–2390.
42. Lenski RE. 2011. Evolution in action: a 50,000-generation
salute to Charles Darwin. Microbe 6:30–33.
43. Luria SE, Delbruck M. 1943. Mutations of bacteria from
virus sensitivity to virus resistance. Genetics 28:491–511.
44. Lederberg J, Lederber E. 1952. Replica plating and indi-
rect selection of bacterial mutants. J Bacteriol 63:399–406.
45. Elena SF, Cooper VS, Lenski RE. 1996. Punctuated
evolution caused by selection of rare beneficial mutations.
Science 272:1802–1804.
46. Riley MS, Cooper VS, Lenski RE, Forney LJ, Marsh TL.
2001. Rapid phenotypic change and diversification of a soil
bacterium during 1000 generations of experimental evolu-
tion. Microbiol SGM 147:995–1006.
47. Blount ZD, Borland CZ, Lenski RE. 2008. Historical con-
tingency and the evolution of a key innovation in an exper-
imental population of Escherichia coli. Proc Natl Acad Sci
105:7899–7906.
48. Barrick JE, Yu DS, Yoon SH, Jeong H, Oh TK,
Schneider D, Lenski RE, Kim JF. 2009. Genome evolu-
tion and adaptation in a long-term experiment with Escher-
ichia coli. Nature 461:1243–1274.
49. Puentes-Teliez PE, Hansen MA, Sorensen SJ, van Elsas
JD. 2013. Adaptation and heterogeneity of Escherichia
coli MC1000 growing in complex environments. Appl
Environ Microbiol 79:1008–1017.
50. Reid SD, Herbelin CJ, Bumbaugh AC, Selander RK,
Whittam TS. 2000. Parallel evolution of virulence in
pathogenic Escherichia coli. Nature 406:64–67.
51. Levin BR, Stewart FM, Chao L. 1977. Resource-limited
growth, competition, and predation—a model and experi-
mental studies with bacteria and bacteriophage. Am Nat
111:3–24.
52. Finkel SE, Kolter R. 1999. Evolution of microbial diversity
during prolonged starvation. Proc Natl Acad Sci USA 96:
4023–4027.
53. Gagneux S, Long CD, Small PM, Van T, Schoolnik GK,
Bohannan BJM. 2006. The competitive cost of antibiotic
resistance in Mycobacterium tuberculosis. Science 312:
1944–1946.
54. Lennon JT. 2007. Is there a cost of viral resistance in
marine cyanobacteria? ISME J 1:300–312.
55. Lenski RE, Rose MR, Simpson SC, Tadler SC. 1991.
Long-term experimental evolution in Escherichia coli 1:
adaptation and divergence during 2,000 generations L.
Am Nat 138:1315–1341.
56. Smith HL, Waltman P. 1995. The Theory of the Chemostat:
Dynamics of Microbial Competition. Cambridge University
Press, New York.
57. Bohannan BJM, Lenski RE. 2000. The relative impor-
tance of competition and predation varies with productiv-
ity in a model community. Am Nat 156:329–340.
58. Traverse CC, Mayo-Smith LM, Poltak SR, Cooper VS.
2013. Tangled bank of experimentally evolved Burkholderia
biofilms reflects selection during chronic infections. Proc
Natl Acad Sci USA 110:E250–E259.
59. Kerr B, Neuhauser C, Bohannan BJM, Dean AM. 2006.
Local migration promotes competitive restraint in a host-
pathogen “tragedy of the commons.” Nature 442:75–78.
60. Gomez P, Buckling A. 2013. Real-time microbial adaptive
diversification in soil. Ecol Lett 16:650–655.
61. Denef VJ, Banfield JF. 2012. In situ evolutionary rate
measurements show ecological success of recently emerged
bacterial hybrids. Science 336:462–466.
62. Yang L, Jelsbak L, Marvig RL, Damkiaer S, Workman
CT, Rau MH, Hansen SK, Folkesson A, Johansen HK,
Ciofu O, Hoiby N, Sommer MOA, Molin S. 2011. Evolu-
tionary dynamics of bacteria in a human host environment.
Proc Natl Acad Sci 108:7481–7486.
63. Nei M, Gojobori T. 1986. Simple methods for estimating
the numbers of synonymous and nonsynonymous nucleo-
tide substitutions. Mol Biol Evol 3:418–426.
64. Konstantinidis KT, Braff J, Karl DM, DeLong EF. 2009.
Comparative metagenomic analysis of a microbial com-
munity residing at a depth of 4,000 meters at Station
ALOHA in the North Pacific Subtropical Gyre. Appl
Environ Microbiol 75:5345–5355.
65. Tai V, Poon AFY, Paulsen IT, Palenik B. 2011. Selection
in coastal Synechococcus (Cyanobacteria) populations eval-
uated from environmental metagenomes. PLoS ONE 6:
e24249.
66. Kryazhimskiy S, Plotkin JB. 2008. The population genet-
ics of dN/dS. PLoS Genetics 4:e1000304.
67. McDonald JH, Kreitman M. 1991. Adaptive protein evo-
lution at the Adh locus in Drosophila. Nature 351:652–654.
68. Simmons SL, DiBartolo G, Denef VJ, Goltsman DSA,
Thelen MP, Banfield JF. 2008. Population genomic anal-
ysis of strain variation in Leptospirillum group II bacteria
involved in acid mine drainage formation. PLoS Biol 6:
e177.
69. Johnson PLF, Slatkin M. 2008. Accounting for bias from
sequencing error in population genetic estimates. Mol Biol
Evol 25:199–206.
70. Moran NA, McLaughlin HJ, Sorek R. 2009. The dynam-
ics and time scale of ongoing genomic erosion in symbiotic
bacteria. Science 323:379–382.
71. Cooper TF. 2007. Recombination speeds adaptation by
reducing competition between beneficial mutations in
populations of Escherichia coli. PLoS Biol 5:e225.
72. Konstantinidis KT, Delong EF. 2008. Genomic patterns of
recombination, clonal divergence and environment in
marine microbial populations. ISME J 2:1052–1065.
73. Johnson PLF, Slatkin M. 2009. Inference of microbial
recombination rates from metagenomic data. PLoS Genetics
5:e1000674.
74. Morowitz MJ, Denef VJ, Costello EK, Thomas BC,
Poroyko V, Relman DA, Banfield JF. 2011. Strain-
resolved community genomic analysis of gut microbial col-
onization in a premature infant. Proc Natl Acad Sci USA
108:1128–1133.
75. Sharon I, Morowitz MJ, Thomas BC, Costello EK, Rel-
man DA, Banfield JF. 2013. Time series community
genomics analysis reveals rapid shifts in bacterial species,
strains, and phage during infant gut colonization. Genome
Res 23:111–120.
76. Nemergut DR, Schmidt SK, Fukami T, O’Neill SP, Bilin-
ski TM, Stanish LF, Knelman JE, Darcy JL, Lynch RC,
Wickey P, Ferrenberg S. 2013. Patterns and processes
of microbial community assembly. Microbiol Mol Biol Rev
77:342–356.
77. Leibold MA, Holyoak M, Mouquet N, Amarasekare P,
Chase JM, Hoopes MF, Holt RD, Shurin JB, Law R,
Tilman D, Loreau M, Gonzalez A. 2004. The metacom-
munity concept: a framework for multi-scale community
ecology. Ecol Lett 7:601–613.
78. Warner NJ, Allen MF, Macmahon JA. 1987. Dispersal
agents of vesicular-arbuscular mycorrhizal fungi in a dis-
turbed arid ecosystem. Mycologia 79:721–730.
79. Burrows SM, Butler T, Joeckel P, Tost H, Kerkweg A,
Poeschl U, Lawrence MG. 2009. Bacteria in the global
atmosphere—Part 2: Modeling of emissions and transport
between different ecosystems. Atmos Chem Phys 9:
9281–9297.
4.1.2-12 ▪ THEORY
80. Yamaguchi N, Ichijo T, Sakotani A, Baba T, Nasu M.
2012. Global dispersion of bacterial cells on Asian dust.
Scientific Rep 2:525.
81. Whitaker RJ, Grogan DW, Taylor JW. 2003. Geographic
barriers isolate endemic populations of hyperthermophilic
archaea. Science 301:976–978.
82. Horner-Devine MC, Lage M, Hughes JB, Bohannan
BJM. 2004. A taxa-area relationship for bacteria. Nature
432:750–753.
83. Hanson CA, Fuhrman JA, Horner-Devine MC, Martiny
JBH. 2012. Beyond biogeographic patterns: processes
shaping the microbial landscape. Nat Rev Microbiol 10:
497–506.
84. Belotte D, Curien JB, Maclean RC, Bell G. 2003.
An experimental test of local adaptation in soil bacteria.
Evolution 57:27–36.
85. Thompson JN. 2005. The Geographic Mosaic of Coevolu-
tion. University of Chicago Press, Chicago.
86. Forde SE, Thompson JN, Holt RD, Bohannan BJM.
2008. Coevolution drives temporal changes in fitness and
diversity across environments in a bacteria-bacteriophage
interaction. Evolution 62:1830–1839.
87. Vos M, Birkett PJ, Birch E, Griffiths RI, Buckling A.
2009. Local adaptation of bacteriophages to their bacterial
hosts in soil. Science 325:833–833.
88. Finkel SE. 2006. Long-term survival during stationary
phase: evolution and the GASP phenotype. Nat Rev Micro-
biol 4:113–120.
89. Vulic M, Kolter R. 2001. Evolutionary cheating in Escher-
ichia coli stationary phase cultures. Genetics 158:519–526.
90. Yeiser B, Pepper ED, Goodman MF, Finkel SE. 2002.
SOS-induced DNA polymerases enhance long-term sur-
vival and evolutionary fitness. Proc Natl Acad Sci USA
99:8737–8741.
91. Johnson SS, Hebsgaard MB, Christensen TR, Maste-
panov M, Nielsen R, Munch K, Brand T, Gilbert
MTP, Zuber MT, Bunce M, Ronn R, Gilichinsky D,
Froese D, Willerslev E. 2007. Ancient bacteria show
evidence of DNA repair. Proc Natl Acad Sci USA 104:
14401–14405.
92. Lennon JT, Jones SE. 2011. Microbial seed banks: the
ecological and evolutionary implications of dormancy.
Nat Rev Microbiol 9:119–130.
93. del Giorgio PA, Gasol JM. 2008. Physiological structure
and single-cell activity in marine bacterioplankton,
p. 243–298. In Kirchman DL (ed), Microbial Ecology of
the Oceans, John Wiley & Sons, New York.
94. Jones SE, Lennon JT. 2010. Dormancy contributes to the
maintenance of microbial diversity. Proc Natl Aca Sci USA
107:5881–5886.
95. Flardh K, Cohen PS, Kjelleberg S. 1992. Ribosomes exist
in large excess over the apparent demand for protein syn-
thesis during carbon starvation in marine Vibrio sp
strain-CCUG-15956. J Bacteriol 174:6780–6788.
96. Campbell BJ, Yu L, Heidelberg JF, Kirchman DL. 2011.
Activity of abundant and rare bacteria in a coastal ocean.
Proc Natl Acad Sci USA 108:12776–12781.
97. Kemp PF, Lee S, Laroche J. 1993. Estimating the growth
rate of slowly growing marine bacteria from RNA conent.
Appl Environ Microbiol 59:2594–2601.
98. Blazewicz SJ, Barnard RL, Daly RA, Firestone MK. 2013.
Evaluating rRNA as an indicator of microbial activity in
environmental communities: limitations and uses. ISME
J 7:2061–2068.
99. Pfennig DW, Servedio MR. 2013. The role of transgenera-
tional epigenetic inheritance in diversification and specia-
tion. Non-genet Inherit 1:17–26.
100. Casadesus J, Low D. 2006. Epigenetic gene regulation in
the bacterial world. Microbiol Mol Biol Rev 70:830–856.
101. Casadesus J, Low D. 2013. Programmed heterogeneity:
epigenetic mechanisms in bacteria. J Biol Chem 17:
13929–13935.
102. Bendall ML, Luong K, Wetmore KM, Blow M, Korlach J,
Deutschbauer A, Malmstrom RR. 2013. Exploring the
roles of DNA methylation in the metal-reducing bacterium
Shewanella oneidensis MR-1. J Bacteriol 195:4966–4974.
103. Haloin JR, Strauss SY. 2008. Interplay between ecological
communities and evolution. Ann NY Acad Sci 1133:
87–125.
104. Hairston NG, Ellner SP, Geber MA, Yoshida T, Fox JA.
2005. Rapid evolution and the convergence of ecological
and evolutionary time. Ecol Lett 8:1114–1127.
105. Schoener TW. 2011. The newest synthesis: understand-
ing the interplay of evolutionary and ecological dynamics.
Science 331:426–429.
106. Labrie SJ, Samson JE, Moineau S. 2010. Bacteriophage
resistance mechanisms. Nat Rev Microbiol 8:317–327.
107. Jessup CM, Forde SE. 2006. Ecology and evolution in
microbial systems: the generation and maintenance of
diversity in phage-host interactions. Res Microbiol 159:
382–389.
108. Bohannan BJM, Travisano M, Lenski RE. 1999. Epistatic
interactions can lower the cost of resistance to multiple
consumers. Evolution 53:292–295.
109. Yoshida T, Jones LE, Ellner SP, Fussmann GF, Hairston
NG. 2003. Rapid evolution drives ecological dynamics in a
predator-prey system. Nature 424:303–306.
110. Rodriguez-Brito B, Li L, Wegley L, Furlan M, Angly F,
Breitbart M, Buchanan J, Desnues C, Dinsdale E,
Edwards R, Felts B, Haynes M, Liu H, Lipson D, Mahaffy
J, Martin-Cuadrado AB, Mira A, Nulton J, Pasic L, Ray-
hawk S, Rodriguez-Mueller J, Rodriguez-Valera F, Sala-
mon P, Srinagesh S, Thingstad TF, Tran T, Thurber
RV, Willner D, Youle M, Rohwer F. 2010. Viral and
microbial community dynamics in four aquatic environ-
ments. ISME J 4:739–751.
111. Weitz JS, Wilhelm SW. 2012. Ocean viruses and their
effects on microbial communities and biogeochemical
cycles. F1000 Biol Rep 4:17.
112. Lennon JT, Martiny JBH. 2008. Rapid evolution buffers
ecosystem impacts of viruses in a microbial food web. Ecol
Lett 11:1178–1188.
113. Moran NA, McCutcheon JP, Nakabachi A. 2008.
Genomics and evolution of heritable bacterial symbionts.
Annu Rev Genet 42:165–190.
114. Hillesland KL, Stahl DA. 2010. Rapid evolution of stabil-
ity and productivity at the origin of a microbial mutualism.
Proc Natl Acad Sci 107:2124–2129.
115. Hoffmann AA, Sgro CM. 2011. Climate change and evo-
lutionary adapatation. Nature 470:479–485.
116. Zilber-Rosenber I, Rosenberg E. 2008. Role of microorgan-
isms in the evolution ofanimals and plants: the hologenome
theoryof evolution. FEMS Microbiol Rev 32:723–735.
117. Reshef L, Koren O, Loya Y., Zilber-Rosenberg I, Rosen-
berg E. 2006. The coral probiotic hypothesis. Environ
Microbiol 8:2068–2073.
118. Lau JA, Lennon JT. 2012. Rapid responses of soil micro-
organisms improve plant fitness in novel environments.
Proc Natl Acad Sci 109:14058–14062.
119. Sharon G, Segal D, Ringo JM, Hefetz A, Zilber-Rosenber
I, Rosenberg E. 2010. Commensal bacteria play a role in
mating preference of Drosophila melanogaster. Proc Natl
Acad Sci 107:20051–20056.
120. Brucker RM, Bordenstein SR. 2013. The hologenomic
basis of speciation: gut bacteria cause hybrid lethality in
the genus Nasonia. Science 341:667–669.
121. Pelletier F, Garant D, Hendry AP. 2009. Eco-evolutionary
dynamics. Phil Trans R Soc B 364:1483–1489.
 The Microbial Ecology of Benthic Environments
ROBERT H. FINDLAY AND TOM J. BATTIN
4.2.1
The benthic realm of life carpets the sedimentary surface of
all aquatic ecosystems including the oceans, lakes, rivers,
and streams (Fig. 1). In shallow aquatic ecosystems, microbial
abundance and biomass are highest at the sediment water
interface, often reaching 1010 prokaryotic cells per gram of
substrate. This high abundance coupled with high rates of
metabolism and elemental transformations and fluxes results
in these communities being critical with regard to many
ecosystem functions. For habitats in which light penetrates
to the benthic zone, photosynthetic organisms, usually a
mixture of eukaryotic algae and cyanobacteria, significantly
contribute to, and at times dominate, microbial life. In addi-
tion to strongly influencing elemental fluxes across the boun-
dary between aquatic and sedimentary habitats, benthic
microbial communities can serve as important trophic link-
ages for macroscopic aquatic organisms.
In this chapter we explore major questions in the ecology
of benthic microbial life and current methodological
approaches used in addressing these questions. We restrict
this discussion to freshwater and shallow marine systems,
because a recent review has examined the microbial ecology
of the dark ocean seafloor (1).
DEFINITIONS AND CONCEPTS
Abundance—The number of organisms in a sample or habitat
Alpha diversity—Originally introduced by Whittaker (6) as
the “richness in species of a particular stand or community”;
it has now viewed as “within community” diversity with
diversity being defined by both species richness and relative
abundance.
Archaea—organisms that are classified in the domain
Archaea.
Autotroph—an organism that uses carbon dioxide as its source
of structural carbon.
Bacteria—organisms that are classified in the domain
Bacteria.
Benthic—the environment that begins at the interface
between water and the bottom of streams, lakes, estuaries,
and oceans. These habitats are typically composed of sedi-
ments but may include rocks and are inhabited by a wide vari-
ety of heterotrophic microorganisms and macroorganisms and
chemolithotrophic prokaryotes. In areas were light reaches
the bottom, phototrophic organisms may also be present.
doi:10.1128/9781555818821.ch4.2.1
4.2.1-1
Benthic boundary layer—the layer of water directly above the
bottom of streams, lakes, estuaries, and oceans caused by
the friction between the bottom and the water. As
a concept it is the zone where solute transport is limited to
molecular diffusion, as such it is sometimes referred to as
the diffusive boundary layer. It is thinned by increasing water
velocity and can be disrupted by a number of physical and
biological processes. See book by Boudreau and Jorgensen
(4) for a comprehensive review.
Beta diversity—the “among communities” component of
gamma diversity.
Biomass—The mass of organisms in a sample or habitat; often
communicated in grams carbon.
Bioturbation—all transport processes carried out by animals
that directly or indirectly affect sediment matrices (sensu Kris-
tensen et al. (3)).
Chemolithotroph—an organism that uses reduced inorganic
molecules as its energy source.
Chemoorganotroph—an organism that uses reduced organic
carbon as its energy source.
Diversity—A complex concept comprised of both species
richness and their relative abundance.
Exploitation—species-species interactions that have a nega-
tive effect on one species while benefiting the other; can
include both parasitism and predation.
Fermentation—the incomplete oxidation of organic matter.
Fick’s first law (of diffusion)—as a concept, this law indicates
that molecular flux is directly proportional to the steepness
of the gradient. The mathematical form of the law, in its dis-
crete form, states that:
 
DC
J ¼ D
Dx
where J is diffusive flux, D is the molecular diffusive constant,
ΔC is the change in concentration, and Δx is distance
Free-energy state change—as a concept, this parameter
describes the likelihood of a chemical reaction proceeding,
with positive free-energy state changes being endergonic
and therefore unlikely to proceed and negative free-energy
state changes being exergonic and likely to proceed. The
standard free-energy state change for a chemical reaction
4.2.1-2 ▪ AQUATIC ENVIRONMENTS
FIGURE 1. Typical benthic ecosystems ranging from (a) headwater streams (b) to coastal waters. Copious microbial communities also
including significant amounts of algae that can carpet gravel streams (c), whereas microbial communities in the sandy habitats of coastal waters
often have lower microbial biomass (d). doi:10.1128/9781555818821.ch4.2.1.f1
can be calculated from the free energy of formation (Gof ) of
the reactants and products using:
DGo ¼ SDGof (products)  SDGof (reactants)
where ΔGo is standard free-energy change of a reaction at 1
atm pressure and 1 M concentrations. The abbreviation
ΔGo is used for free-energy change under conditions specified
and ΔGo0 for free-energy change under standard conditions at
pH 7. See Appendix 1 Brock Biology of Microorganisms (2)
for an excellent discussion on the calculations of ΔGo; also
referred to as Gibb’s free-energy change.
Gamma diversity—The diversity of an extended region, which
is viewed as being decomposable into alpha and beta diversity.
The past decade has seen considerable debate as to whether
the decomposition should be multiplicative or additive (5).
Heterotroph—an organism that uses preformed organic carbon
as its source of structural carbon, also heterotrophic or hetero-
trophy when applied to metabolism.
Microeukaryotes—microorganisms that are classified in the
domain Eukarya.
Microorganisms—for the purpose of this chapter, we use the
term microorganisms to indicate all single-celled forms of
life, including bacteria, archaea, and eukaryotes.
Mutualism—species-species interactions that benefits both
species.
Phototroph—an organism that uses light as its energy source
for the production of ATP and reducing power.
Prokaryotes—while this term is considered outdated with
respect to taxonomy, it is useful when referring to the
species present in the domains Bacteria and Archaea and
we use this term to refer collectively to these two groups of
microorganisms.
Relative abundance—The contribution of a species (or OTU)
to a community.
Richness—The number of species (or operational taxonomic
units, OTUs) in a sample or habitat
Roughness—a characteristic of the bottom of streams, lakes,
estuaries, and oceans that is influenced by sediment grain
size, the morphology of the bed (e.g., the presence of ripples),
and bioturbation. The presence or absence of roughness mod-
ulates the effects of water flow on exchange across the
sediment-water interface.
Syntrophy—a metabolic mutualism where one species uses the
waste product produced by the other, and in so doing allowing
both metabolic pathways to be energetically feasible.
Tortuosity—is the property of being tortuous or twisted, hav-
ing many curves. In benthic ecology it is commonly used to
describe the diffusive path through the interstitial space of
sediments.
MAJOR QUESTIONS IN THE MICROBIAL
ECOLOGY OF BENTHIC ENVIRONMENTS
Most research questions in microbial ecology are comparable
to those studied in the ecology of plants and animals. What
abiotic factors control the distribution and abundance of
organisms? How do organisms assemble into communities?
What are the emergent properties of ecosystems? The physical
structure of benthic environments, at the microorganism’s
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-3
appropriate spatial and temporal scales, provides the specific-
ity of these questions in benthic microbial ecology. Sedimen-
tary surfaces and interstitial spaces are major physical
templates that contribute to the establishment of chemical
and biological gradients and may ultimately support high bio-
diversity and ecosystem functions. No doubt, these questions
are fundamental to climate change and other pressing envi-
ronmental issues, but they are also a useful starting point to
investigate prokaryotic life in benthic environments. Includ-
ing bacteria and archaea in such questions adds a variety of
interesting twists. For example, until recently microbial
ecologists were unable to write the list of taxa occurring in
these environments. In fact, the species concept for prokar-
yotes is different than for sexually reproducing eukaryotic
organisms; furthermore, horizontal gene transfer is a critical
component of both the evolution and maintenance of pro-
karyotic species.
Beginning with Zuckerkandl and Pauling’s statement that
the origin of an organism was written in its genes (7), micro-
bial ecologists and evolutionary biologists have used gene
sequences as a useful tool to identify and study the phylogeny
of microorganisms. As technological advances have permit-
ted ever-faster sequencing of genes from organisms and subse-
quently genes from communities, it has become possible for
microbial ecologists to begin to identify the organisms in a
given environment. However, microbial ecologists are likely
a century behind their colleagues who study plant and animal
ecology. As complete genome sequences of multiple isolates
within a species have been developed, it has become clear
that all of the members of a bacterial species may share as little
as 8% or as much as 56% of their collective genome (these
shared genes are referred to as the core genome) and that
the number of genes found in at least one isolate of that
bacterial species can exceed 25,000 (referred to as the pan
genome). For some species, research has identified an
extended core or character genome that consists of genes
shared by all sequenced members except when excluded by
the gene-calling software or sequencing errors, or lost in a
few isolates as a result of recent genome reduction events.
The extended core often doubles the number of shared genes
over the core genome sensu stricto. Interestingly, genes found
within the pan genome may be critical for strains to adapt to
environmental conditions. A well-explored example is Pro-
chlorococcus, in which high light and low light adaptations,
as well as temperature adaptations within these ecotypes are
reflected in differences within the pan genome (8). Horizon-
tal gene transfer is seen as important to both the development
of the pan genome as well as the maintenance of core
genomes. The latter typically involves homologous recombi-
nation and is easily conceptualized within the earliest exper-
imental evidence of natural competence (9) and is seen as a
counterbalance to stochastic genetic processes, such as muta-
tion and genetic drift (10, 11). A variety of mechanisms and
mobile genetic elements can facilitate the transfer of genomic
information ranging in size from a single gene to genetic
islands and integrating conjugative elements, which may
encode a complex function ( pathogenicity, symbiosis, meta-
bolic, resistance, fitness, etc.) including necessary regulatory
elements. It is against this backdrop of our ever-expanding
understanding of evolution and speciation that microbial
ecologists investigate the distribution and abundance of
microbial life.
Possible Advantages of Benthic Life
Why live on or in sediments, or in some cases, attached to
rocks or other hard surfaces? The attached mode of microbial
life as biofilms (see chapter 4.2.3 for further discussions of bio-
films) is thought to be an ancient and integral characteristic
of microorganisms, which has enabled them to grow even
in adverse environments (12). Fossilized stromatolites (a
complex accreting sedimentary biofilm) and their associated
microfossils are widely regarded as the first fossil evidence of
life (13). Not only do they indicate that microbial life flour-
ished ∼3.5 billion years ago, they clearly show that prokar-
yotes had firmly established a benthic lifestyle early in the
evolution of life. The earliest evidence of life on Earth is sta-
ble isotope ratios of iron and carbon within sedimentary rocks
formed 3.8 billion years ago. These also suggest a sedimentary
lifestyle (14). Stromatolites and other microbial mats became
widespread during the Archean and increased in complexity,
indicating that prokaryotes occupied an increasingly wide
range of sedimentary environments throughout the eons
(15). The restriction of modern stromatolites to environ-
ments where conditions exclude macrofauna suggests a gen-
eral susceptibility of concentrated prokaryotic communities
to predation and suggests a role of macrofauna in controlling
the structure of benthic microbial communities (16). This
hypothesis is supported by the general disappearance of
benthic microbial mats during the Neoproterozoic-Cambrian
(a period some authors are referring to as the “Cambrian sub-
strate revolution”) (e.g., 17), concurrent with the first fossil
evidence of bioturbation (18, 19) and the reemergence of
microbial mats following the end Permian and Late Devonian
mass extinction (20). As biofilms can easily contain 1013
cells/g dry mass and sediments typically contain between
108 to 1010 cells/g dry mass, current predators of benthic
prokaryotes are confronted with a food resource that is diluted
at least 1,000-fold with inorganic particles and that must
either be scraped or digested from sediment particles. This
is not to say that benthic microorganisms do not form bio-
films; they do, they are just smaller and form on sediment
grain surfaces.
For environments where movement of water is a key deter-
minant, the simplistic answer to our question is that the
attached mode of life avoids erosion and export. The physical
structure of benthic biofilms increases spatial heterogeneity
with several notable consequences (Fig. 2). For instance, gra-
dients of electron donors and acceptors establish at the fine
scale, allowing development of environmental conditions
not found within the overlying water. Biofilms also allow
for colocation of organisms, facilitating mutualisms depend-
ent on juxtaposition of interacting metabolisms. The sedi-
mentary milieu increases this complexity and the potential
variation in microhabitats. Surfaces in water accumulate
organic matter and as such have higher concentrations of
organic nutrients than the surrounding water. The surface
area of sediments is inversely proportional to grain size and
can approach 10 m2/g dry mass of sediment. The large surface
area plus gravitational settling causes the benthic zone to be
the repository for organic matter (and for contaminants) in
most shallow aquatic systems. This leads to strong differences
in organic matter concentration in these systems. For exam-
ple, acetate, one of the most common dissolved organic com-
pounds in aquatic systems, rarely exceeds µM concentrations
in water (21, 22), while concentrations in excess of 1 mM are
often observed for sediment pore water (23). In systems where
the delivery of reduced carbon exceeds the delivery of oxygen,
anaerobic conditions can develop, allowing the growth of
microbes using alternative electron acceptors. These anaero-
bic microorganisms typically produce reduced inorganic com-
pounds that diffuse back toward the oxic microhabitats where
chemolithotrophic microorganisms are favored.
4.2.1-4 ▪ AQUATIC ENVIRONMENTS
FIGURE 2. Photographs of benthic microbial communities. (a) An initial state of a benthic community showing the various forms of micro-
bial organisms. (b) A filamentous streamer with a pronounced head and a long tail; streamers tend to develop in turbulent flow. (c) Long stream-
ers (extending up to 10 cm and more) can harbor various microorganisms and algae. All photographs from epifluorescence microscopy using
SYTOX (green) and autofluorescence (red) of algae. Courtesy of Iris Hödl. doi:10.1128/9781555818821.ch4.2.1.f2
Factors Controlling the Distribution and Abundance
of Microorganisms
Classically, Atlas and Bartha (24) list 10 abiotic deter-
minants of microbial distribution and abundance. These are
temperature, radiation, redox potential (Eh), pH, pressure
(hydrostatic and osmotic), water activity, movement of air
and water, magnetism, inorganic nutrients, and organic
nutrients. Brock (2), McArthur (25), and Kirchman (26)
use a less formal presentation but describe many of the same
determinants as controls on the distribution of microorgan-
isms. In this chapter, we focus on five abiotic factors as well
as on biotic interactions as important determinants of benthic
microbial communities.
Flow
The flow of water over the sediment can produce various
effects on benthic microbial communities. The most obvious
of these, flow velocity, can sort sediments based on grain size.
For water carrying sediment, reduction of flow velocity typi-
cally induces deposition, with the largest and/or most dense
grains settling first, thereby shaping the physical structure of
the benthic environment. Unidirectional flows can wash
microbes into and out of a habitat. Tidal flows, which in
some habitats can equal those observed in fast-flowing rivers
and streams, are bidirectional and can remove and return
microbes. Habitats that are typically viewed as lacking flowing
water (lakes and coastal embayments) can be strongly affected
by wind and waves (27), which can develop rotational flows
sufficient to episodically resuspend bottom sediments. In the
absence of advective flow, solute supply across the benthic
boundary layer is limited by molecular diffusion, although
small-eddy diffusion may dominate solute transport above
the benthic boundary layer. Interactions of flow with rough-
ness of the benthic zone and even well-developed biofilms
produce areas of turbulence enhancing exchange across the
sediment–water interface and if sufficiently robust may even
lead to advective pore water exchange (28, 29). A structure
as simple as a single seagrass blade can enhance exchange
across the sediment–water interface, producing food web
effects that cascade through bacteria to harpacticoid cope-
pods (30). Flow velocity and related shear stress and boundary
layer phenomena shape the biomass of benthic biofilms
and also their physical structures. This in turn can affect
hydrodynamics above biofilms and can even contribute to
the transient storage of water and solutes (31, 32). Studies
from stream ecosystems also suggest that the flow environ-
ment above the benthic zone may influence dispersal of
microbial cells, community assembly, and biodiversity pat-
terns (33, 34).
Disturbance
Disturbance can be defined as any process that adversely
affects single organisms, populations, or entire communi-
ties. Some of the strongest physical disturbances of benthic
microbial communities are observed in streams and rivers
where elevated shear stress can scour the sediments and
abrade microbial biomass. However, such disturbances can
take place in any environment where sufficient energy is
present to move sediment producing particle-to-particle col-
lisions which can damage, kill, or detach microorganisms.
Holmes et al. (35) found that sediment chlorophyll a val-
ues (as a measure of benthic algae biomass) dropped from
∼400 mg/m2 to near zero following a flash flood. Direct
counts of bacteria indicated that bacterial abundance
decreased from ∼9 × 107 to ∼3 × 107/ml of sediment in the
same flood. Findlay et al. (36) showed that periodic abrasion
in marine microcosms led to a decrease in total microbial bio-
mass of 31% and preferential removal of microeukaryotes and
aerobic bacteria. Cell abrasion and export are proximate con-
sequences of physical disturbance but can also be beneficial
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-5
for the system as a whole. In fact, biomass loss can trigger com-
munity succession and even increase specific production of
biofilms (37). Predators and parasites also cause disturbances
and we discuss these under the “Biotic Interactions” section.
Seasons
Seasonality induces changes in environmental controls on
benthic microbial life. Such effects are most pronounced in
shallow aquatic ecosystems where the benthic zone is poorly
buffered from external drivers. In the temperate biome, tem-
peratures of river and stream water, lake surface water, and
near-shore marine environments may vary 20–30°C between
winter lows and summer highs. Rates of photosynthesis in
microalgae, diatoms in particular, appear to be unaffected
by low temperatures and if sufficient light is present will rap-
idly accumulate biomass. Poikilothermic heterotrophs, how-
ever, show a direct correlation between temperature and
metabolic activity, and many aquatic metazoans will stop
feeding below 4°C (a biotic disturbance—see below). In
environments where flow does not vary seasonally, these fac-
tors lead to a pattern of change in sediment community
structure where phototrophic microeukaryotes and aerobic
bacteria dominate during cold-water months, and anaerobic
bacteria are most important during warm-water months
(38). In systems were flow varies with season (as in many tem-
perate streams), it is possible to have four seasons—high flow/
cool water, low flow/warm water, low flow/cool water, and
high flow/cold water. Water temperatures typically lag air
temperatures by about a month, but by mid-spring, water tem-
peratures begin to warm and no longer limit grazer activity
(cool water period) although flows remain high. With leaf
out evapotranspiration increases, and by mid-summer flows
decrease and water temperatures rise (low flow/warm water).
During autumn temperatures decrease, but for many biomes
river flows remain low (low flow/cool water) and do not begin
to increase until leaf fall and the onset of winter rains. By win-
ter water temperatures decrease sufficiently to limit grazer
activity and the resident microbial community experiences
the high flow/cold water season. For many areas heavy early
spring rain and/or snowmelt produce the highest flows and
very cold water temperatures. As above, cold water favors
accumulation of algal biomass and disturbances (abiotic
and biotic) remove it (39). In systems where there is insuffi-
cient water movement to overcome thermal resistance, the
water column will stratify, allowing for different seasonal pat-
terns between sediments above and below the thermocline. If
labile carbon supplies exceed available oxygen supplies, sedi-
ments (and the overlying water column) will become anoxic
with the concomitant shifts in microbial community struc-
tures. Smoot and Findlay (40), working in a temperate reser-
voir, found seasonal variation in epilimnetic sediments (those
above the seasonal thermocline) similar to those observed for
temperate streams. Hypolimnetic sediments (those below the
seasonal thermocline) were distinct in terms of community
structure, but they also showed seasonal variations similar to
those observed in temperate streams and epilimnetic sedi-
ments. Abiotic and biotic disturbances were less frequent in
hypolimnetic sediments. This was posited as the proximal
cause of the greater biomass observed in the hypolimnion ver-
sus epilimnion sediments.
UV- and Visible Radiation
Shallow benthic environments are particularly prone to UV
radiation. Notably, UV-B (280–310 nm) is a biologically
active component of UV radiation that affects biomolecules.
UV radiation can shape the structure of benthic communities
and affect their productivity and contributions to elemental
cycling (41–43). Microbial life in benthic ecosystems has
developed various strategies over evolutionary times to
escape, to some extent, the negative effects of UV radiation.
For instance, in stromatolites, an ancient form of benthic
microbial life, laminated, organo-sedimentary structures may
shade and thereby protect cyanobacteria from UV radiation.
Garcia-Pichel and Bebout (44) demonstrated that UV radia-
tion can penetrate into microbial mats with UV radiation
intensities in the upper layers even exceeding those in the
overlying water. High intensities of UV radiation in these
mats were attributed to scattering by minerals. To escape
this pressure, cyanobacteria migrate vertically, gathering in
deeper layers of the mat during elevated UV-exposure (see
Stal [45] for review). Surface-attached microorganisms can
produce copious amounts of extracellular polymeric substan-
ces, which form the biofilm matrix (see chapter 4.2.3)
and may protect enclosed cells from UV radiation (46, 47).
Finally, as a distinct feature of sedimentary ecosystems, gentle
currents (see flow section above) can episodically resuspend
fine sediments, thereby increasing turbidity above the sedi-
ment. This may further attenuate the effects of UV radiation
on benthic microbial life (48).
Photosynthetically active radiation is frequently the limit-
ing determinant for benthic phototrophs, and its presence or
absence within a habitat can have dramatic effects on sedi-
mentary microbial communities. Given sufficient light, pho-
totrophs can contribute to and at times dominate total
sediment microbial biomass. For all but deep-sea sediments,
these organisms are a major source of polyunsaturated fatty
acids, an essential nutrient for most metazoans. They are
also an autochthonous source of dissolved organic matter,
much of which is highly labile and, as such, holds the poten-
tial to induce a priming effect in adjacent heterotrophic
microorganisms (see the next two sections, “Organic Matter
and Nutrients” and “Biotic Interactions”). They also contrib-
ute to the sediment particulate organic matter pool on their
death. In streams, biome and stream order strongly influence
the concentration of light reaching benthic habitats. In forest
biomes, for instance, a canopy often shades low-order streams,
thus significantly limiting autochthonous primary produc-
tion and greatly increasing the importance of terrestrial detri-
tus. In larger rivers, as well as lakes and marine environments,
water column processes such as scattering and absorption
attenuate the light.
Organic Matter and Nutrients
Microbial life in the benthic zone is sustained from particulate
organic matter (POM), dissolved organic matter (DOM), and
nutrients from the overlying water column and from the
sedimentary environment itself. Portions of these originate
in terrestrial environments—the relative contribution of ter-
rigenous detritus and nutrients are dependent, at least in part,
on the habitat. Bacteria, archaea, fungi, and other fungal-like
microeukaryotes are considered the primary decomposers of
particulate detritus, yet they acquire carbon and energy
from dissolved compounds. All are challenged to acquire
soluble decomposition products of POM before these materi-
als diffuse from the site of production. As such, even aerobic
heterotrophs likely contribute to pore water DOM. Once fer-
mentation becomes the energetically favored metabolism,
incomplete oxidation products become dominant, and in
depositional environments, there is typically a net flux of
DOM from the sediments to the overlying water (29). Signif-
icant organic carbon (OC) is often sequestered at depth and
under such conditions burial efficiency (buried OC:deposited
4.2.1-6 ▪ AQUATIC ENVIRONMENTS
OC) appears to be controlled by carbon lability, oxygen expo-
sure, bioturbation, and for marine but not freshwater systems
absorption onto mineral surfaces (49–51).
For erosional environments or those dominated by hard
surfaces, the water column can become depleted of labile
DOM and nutrients; however, these can still be regenerated
and mobilized from the various niches among chemical gra-
dients within the sediments. Furthermore, the extracellular
matrix formed by numerous attached microbial communities
(see chapter 4.2.3) can buffer microorganisms against fluctu-
ations in the external supply (52). For instance, Battin et al.
(31) showed that close spatial proximity of algae and micro-
bial heterotrophs supports internal carbon cycling in certain
stream biofilms, thus at least partly disconnecting these com-
munities from external sources.
DOM constitutes a prime energy source to microbial het-
erotrophs in aquatic ecosystems and its diversity is massive, as
increasingly revealed by ultra-high-resolution mass spectrom-
etry (53). For instance, Mosher et al. (54) found a total of
5,672 unique low-molecular-weight chemical formulae
(250–1,000 Da) in forested headwaters. Benthic microorgan-
isms metabolize certain functional groups within this DOM
pool and produce new compounds, while others remain unaf-
fected by microbial activity. Experimenting with bioreactors
mimicking the streambed and contained heterotrophic
communities, Kim et al. (55) showed that microbial metabo-
lism modifies DOM to lower molecular weight molecules, and
oxygen-rich molecules are selectively biodegraded. Work on
benthic biofilms in streamside flumes further suggests a link
between the spatial heterogeneity of streambed topography,
microbial biodiversity patterns, and removal of DOM from
the stream water by benthic microorganisms (56, 57). These
authors showed that the diversity of DOM compounds
removed from the stream water is higher when the physical
heterogeneity of the stream bed is elevated—as induced, for
instance, by flow and sediment sorting. The potential corre-
lations among roughness, enhanced solute exchange (of
both O2 and DOM), and enhanced degradation is intriguing.
What becomes clear from these studies is that the inter-
relationship among DOM molecular composition and the
structure and function of heterotrophic microorganisms
inhabiting benthic environments is complex (58) and is an
important area of study. Studies on DOM increasingly out-
pace those on particulate organic matter in sediments, which
is surprising given the importance of photodetritus, for
instance, for benthic microorganisms and even for higher tro-
phic levels (59). The recognition of DOM as the major repo-
sitory of organic carbon and as the intermediary to the global
carbon cycle has hastened this dichotomy.
Biotic Interactions
The close spatial proximity in benthic communities offers
multiple opportunities for microorganisms to interact. Here
microorganisms exhibit all forms of biotic interactions,
including exploitation, competition, and mutualism as they
are typically prescribed to animals and plants. Indeed, many
of the first experiments demonstrating the effects of a partic-
ular interaction were conducted using microorganisms (e.g.,
predation and competition; 60). If there is contrast to be
drawn between microorganisms and animals or plants in
terms of biotic interactions, it is the apparent ease and fre-
quency with which microorganisms enter into cooperative
interactions, either with one another or with plants and
animals. In this section, we focus on two interaction types,
one exploitative ( predation/parasitism) and one potentially
mutualistic.
For shallow marine systems, it has been long held that
there is insufficient bacterial biomass to support the meta-
bolic needs of macrofauna and that benthic algae provided
sustenance for these organisms (61). This is in spite of studies
that showed that marine invertebrates can grow and repro-
duce on a diet consisting solely of marine bacteria (62) and
that sediment-ingesting macrofauna remove a significant frac-
tion of bacteria present in sediments. Recent studies suggest
that bacteria and benthic algae are ingested by deposit-
feeding macrofauna at similar rates (63), which is consistent
with recent findings for freshwater gizzard shad Dorosoma
cepedianum (64). Smoot and Findlay (65) went on to demon-
strate that gizzard shad obtained sufficient living microbial
biomass from sediments through selective feeding on low-
density particles to support its caloric requirements. Under-
standing the role of bacteria within benthic food webs
remains one of the largest unaddressed challenges within
microbial ecology and will require an effort similar to that
directed toward resolving the role of bacteria in marine pela-
gic food webs (see chapter 4.2.2).
One key step in understanding the role of bacteria in
marine pelagic food webs was the recognition of the role of
viruses in promotion of bacterial secondary production (see
chapter 4.2.2). While benthic food webs remain under-
studied, the role of viruses within them is in its infancy. Envi-
ronments with high linear flows (streams and rivers) where
viruses would be rapidly washed from local habitat will likely
show a diminished role for viral exploitation of bacteria with
the possible exception of biofilms where matrix effects
may allow for either direct cell-to-cell transmission or reten-
tion of mature virions. However, environments with either
low flow or bidirectional flows hold the potential for high
impacts of viral activity on benthic food webs. Helton et al.
(66) recently contrasted the current understanding of the
role of viruses within marine and freshwater benthic com-
munities as
While no direct observations of viral-induced changes in
benthic microbial communities have been reported, a hand-
ful of studies measuring benthic viral production rates indi-
cate that up to 61% of co-occurring bacterial host cells are
lost to viral lysis each hour in coastal sediments (67–69)….
However, direct observations of infected cells and measure-
ments of viral decay rates in freshwater sediments contradict
these estimates of substantial viral impact and instead argue
for limited levels of active viral infection and production
within these benthic ecosystems (70–72).
It is not clear why the role of viruses would differ so greatly
between marine and freshwater systems; what is clear is that
the role of viruses is understudied in benthic systems, and
their role in structuring benthic microbial systems has yet to
be resolved.
Priming, the enhanced degradation of refractory organic
matter in the presence of labile refractory matter, has recently
been suggested as an important phenomenon in aquatic sys-
tems (73). If correct, this enhanced degradation of detrital
carbon will have important consequences for our understand-
ing of carbon degradation in aquatic habitats and watershed
carbon budgets (74). Two recent studies suggest that priming
may be a widespread occurrence in freshwater habitats. The
first, using 13C-labeled leaf leachate, showed, on average, a
10% increase in stream water DOM decomposition when
labeled leaf leachate was added at trace levels to experimental
bioreactors (75). In the second, using a periphyton-detrital
complex showed both a significant increase in detrital Typha
decomposition and transfer of 13C-CO2 from phytoplankton
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-7
to fungi and bacteria was shown in the presence of light but
not during the dark or in the presence of a photosynthesis
inhibitor (76). It remains to be determined if this interaction
driven by release of DOC (and potentially mutualistic) or
alteration of local environment by algal photosynthesis
(increases in O2 concentration and pH). It does suggest
that phototrophic–heterotrophic interactions throughout
benthic habitats deserve further study.
Elemental Cycles
Element cycling not only serves as an important integrator of
several abiotic determinants of bacterial distribution and
abundance but also links the ecology of benthic environ-
ments to large ecological scales and associated environmental
challenges. As recently reviewed by Falkowski et al. (77) and
Schlesinger et al. (78) the biogeochemical cycles that we
often study, learn, and teach an element at a time, are in
actuality highly linked global processes. Of the dozen ele-
ments that typically make up the majority of the dry weight
of a bacterial cell (O, H, C, N, P, S, Na, K, Mg, Ca, Fe,
Mn) O, H, C, N, S, Fe, and Mn have significant biogeochem-
ical cycles where living organisms, as the basis of their metab-
olism, conduct coupled redox half-reactions where one
element is oxidized and another is reduced. This metabolic
activity drives global biogeochemical cycles and sets the aver-
age redox conditions of the oceans and atmosphere, as well as
the specific redox conditions within benthic habitats. Not all
redox half-reactions are created equal, and therefore not all
coupled redox reactions provide the same energy yield. The
impact of this inequality was first formalized by Berner (79)
and is now often referred to as Berner’s law of competitive
energy yield. Simply stated, the metabolism that yields the
highest free energy will dominate a system.
Life in the Sediment: Standard State Free Energy
Changes, Fick’s First Law, and Tortuosity
Berner (79), writing in his seminal book Early Diagenesis: A
Theoretical Approach, stated “Catabolic processes generally
follow a definite succession in sediments depending upon
the nature of the oxidizing agent… The reason why such a
succession occurs is generally explained in terms of the meta-
bolic free energy for each reaction.” Aerobic decomposing
occurs first, as oxygen produced the highest energy yield, until
the energy source or oxygen is exhausted (Fig. 3a). If reduced
carbon remains, decomposing proceeds through alternative
electron acceptors, in order of their free energy yield, as
each electron acceptor is successively depleted (Table 1).
This relationship is often referred to as the free energy compe-
tition hypothesis. This hypothesis yields the idealized distri-
bution of heterotrophic metabolic types within sediments.
The progressive depletion of electron acceptor species is char-
acterized by an increasingly negative reduction potential or
Eh. However, heterotrophic decomposition is not the only
metabolism energetically favored within sediments. Many
fermentations (incomplete oxidation) have standard state
free energy changes of around −250 kJ/mol e−, which is sig-
nificantly higher than the energy captured by the production
of 2 ATPs (e.g., −63.6 kJ/mol e−; Hexose → 2 ethanol +
2CO2 + 2 ATP). In addition, many chemolithotrophs oxi-
dize the reduced end products of anaerobic heterotrophic
decomposition and these reactions can be highly exergonic
(e.g., H2S + 2O2 → SO2− 4 + 2H − 399.1 kJ/S atom oxidized)
+
and can account for significant oxidant demand whenever
carbon supply exceeds the ability of the environment to sup-
ply oxygen. Oxygen supply is a function of concentration,
demand, and water movement. When flows are turbulent
(and under unique situations such as losing reaches), water
can be forced into sediments carrying with it dissolved oxi-
dants. During laminar flow, the water column remains well
mixed except for the diffusive boundary layer and delivery
is via molecular diffusion and follows Fisk’s first law and is
strongly influenced by water velocity (Fig. 3b). For benthic
environments that reside within the photic zone, both
benthic carbon stores and sediment oxygen concentrations
can be influenced by the activity of phototrophic microor-
ganisms (Fig. 3c). Finally, standard conditions do not prevail
within sediments; the metabolic activity of organisms can
often decrease reactant or product concentrations to the
µmolar or 10−4 atm range, significantly altering the free
energy change of some bacterial metabolisms. Combined,
all of the above provides the bulk vertical distribution of met-
abolic types within sediments (Fig. 3d). However, sediments
also restrict diffusion, a property typically referred to as tor-
tuosity, and can be seen in the significantly lower molecular
diffusion constants for sediments versus water. This property
allows for the development of local depletions of the most
electropositive oxidants, allowing anaerobic metabolisms to
flourish in otherwise aerobic sediments. Coupling these
microhabitats with juxtaposition (Fig. 3e) and the ability of
one species to use the products of another allow for syntrophic
species interaction as an emergent property of sediment sys-
tems (Fig. 3f ).
Elemental Cycles and the Distribution of
Microorganisms in Sediment
Viewed from the perspective of electron acceptors, the high-
est free energy yields are obtained with O, followed by Mn,
Fe, N, S, and finally C. However, organic carbon is not the
only electron donor available to microorganisms, chemoli-
thotrophic microorganisms compete with heterotrophs for
electron acceptors, and their metabolic activity (hydrogen
sulfide oxidizers and nitrifiers, to name but a few) is included
within the energy hierarchy of favored metabolism. To rea-
sonably predict the distribution of different metabolic types
within sediments also requires consideration of fermentation
and the concepts of microenvironments and bioturbation
(80). Add the recognition that free energy yields are highly
dependent on reactant and product concentrations and that
bacterial and archaeal metabolic activity can drastically alter
local metabolite concentrations, and the distribution of
many key syntrophic mutualisms become apparent. When
examined in terms of standard Gibbs free energy change
some syntrophic mutualisms appear thermodynamically
unfavorable. Examples include the classically described
inhibition of Syntrophomonas wolfei by H2 (81) and the
role of sulfate-reducing bacteria in anaerobic methane oxi-
dation (82). In both cases, consumption of H2 reducing its
concentration to ≤1 Pa (a decrease of more than 1,000-
fold compared to standard conditions) allows for energeti-
cally favorable metabolisms for both syntrophs. As microbial
ecologists continue to investigate the role of interspecies
electron transfer, we will likely continue to discover hitherto
undescribed metabolic pathways (83).
Macrofauna, in addition to feeding on bacteria, also dis-
turb sediment via bioturbation. This translocation of sedi-
ment and water can have profound effects on sediment
microbial communities and therefore biogeochemical cycles,
the most obvious being the production of oxic zones at depth
within the sediments. In shallow or intertidal sediments
with a healthy macrofauna community, apparent sediment
4.2.1-8 ▪ AQUATIC ENVIRONMENTS

FIGURE 3. The supply of carbon and oxygen to sediment strongly influences the distribution of bacterial and archaeal metabolic types. (a)
Experimentally determined relationship between carbon flux to the benthos and the supply of O2 influenced by water velocity. The curved line
represents maximum O2 flux rate for 15°C seawater at O2 saturation—for any site with a flow velocity × carbon flux rate that falls below or to the
right of the curve, oxygen supply will exceed demand, and for those sites that fall above or to the left of the curve oxygen demand will exceed
supply. In these cases, aerobic decomposition will be limited (redrawn from 104). (b) Effect on water velocity on benthic boundary layer. Curves
represent O2 concentration of water and sediment as determined by microelectrode with the embedded numbers representing water velocity
(cm s−1). Horizontal lines represent variation among multiple measurements. The vertical portion of each line indicates the well-mixed portion
of the water column and the thickness of the benthic boundary layer can be determined as the distance between the sediment–water interface
and the bottom of the well-mixed layer (redrawn from 110, with permission). (c) The effect of benthic photosynthetic activity on sediment O2
concentration. Open bar represent photosynthetic rate and circles O2 concentration during early afternoon and sundown. Note the subsurface
O2 maximum (redrawn from 109). (d) Idealized distribution of bacterial and archaeal metabolic types within sediment. Light crosshatching
indicates the suboxic zone and dark crosshatching anaerobic zone (redrawn from 80, with permission). (e) Drawing from a photomicrograph
of a mix microcolony in a depression on the surface of sand grain (from 188, with permission). (f ) Scanning electron micrograph of a syntrophic
coculture of Desulfovibrio G20 and Syntrophomonas wolfei. Fermentation of butyrate to acetate and H2 is energetically inhibited at standard tem-
peratures and concentrations. Desulfovibrio G20 cannot oxidize butyrate alone (although some sulfate-reducing bacteria can), however, via a
tightly coupled mutualistic interaction Desulfovibrio G20 and S. wolfei successful oxidize butyrate. In general, syntrophy is currently thought to
proceed through hydrogen or formate production and can also involve methanogenic archaea (189). doi:10.1128/9781555818821.ch4.2.1.f3
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-9
TABLE 1 Standard-state free energy change of some bacterial metabolisms
Reaction
CH2 O þ O2 ! CO2 þ H2 O
5 CH2 O þ 4NO 
2 ! 2N2 þ 4HCO3 þ CO2 þ 3H2 O
CH2 O þ 3CO2 þ H2 O þ 2MnO2 ! 2Mn2þ þ 4HCO 3
CH2 O þ 7CO2 þ 4Fe(OH)3 ! 4Fe2þ þ 4HCO
3 þ 3H2 O

2 CH2 O þ SO2
4 ! H2 S þ 2HCO3
2CH2 O ! CH4 þ CO2
Source: Berner (79).
diffusion coefficients for oxygen are typically two to three
times higher than the molecular diffusion coefficient (84,
85). Single-species mesocosm studies with conservative trac-
ers can show apparent sediment diffusion coefficients
approaching 40 times greater than the molecular diffusion
coefficient (86) (see Riisgard and Larsen [87] for a review
of types of burrowing benthic animals and common methods
for measuring pumping rates). Effects are not always direct, as
some species can elevate rates of chemolithotrophy at the
expense of aerobic heterotrophy, apparently by increasing
vertical movement of sulfide (88). Another common indirect
effect of macrofauna bioturbation is to produce surface rough-
ness that can interact with flow, enhancing solute transport
into sediments.
Biogeography
Currently within microbial ecology, one of the most compel-
ling questions is “do microorganisms exhibit biogeographical
patterns?” (89). This question is obviously complicated by the
complex issue of what constitutes a microbial species. We
posit that soil microbial communities will be most likely to
show biogeographical patterns while pelagic marine microor-
ganisms will be the least likely to show biogeographical
patterns, with microorganisms inhabiting sediments inter-
mediate in their likelihood to show these patterns. In fact,
soil microorganisms likely have lower rates of dispersal while
pelagic marine the highest rates, which should allow for high
colonization potentials within marine habitats and low colo-
nization potential within soil habitats.
Spatial heterogeneity in physical, chemical, and biologi-
cal conditions in sedimentary ecosystems increases niche
diversification with consequences for spatial heterogeneity
(i.e., beta diversity) of microbial taxa from a fine scale
(<meter) to multi-kilometer scales. For instance, working
with experimental streams, Besemer et al. (34) found differ-
ences in microbial community composition even at the scale
of 1-m-long geomorphological features as they occurred in
natural streams. Here gradients in flow velocity and related
shear stress over these geomorphological features created local
hydrodynamic microhabitats that harbored different micro-
bial communities. At an intermediate level, Wilhelm et al.
(90) found a spatial shift in microbial community composi-
tion of benthic biofilms dwelling in high-elevation streams
across the European Alps. Fierer et al. (91) found a similar
pattern for bacteria inhabiting fine benthic organic matter.
At an even larger scale, Zinger et al. (92) showed clear spatial
differences between benthic microbial communities in
coastal water and in the deep sea across the globe. Collec-
tively, these findings open new perspectives for microbial bio-
geography of the sedimentary realm.
These recent studies complement findings from terrestrial
and pelagic habitats. Currently, the data for soil microbes is
ΔGo0 (kJ mol−1 of CH2O)
−475
−488
−349
−114
−77
−58
equivocal, with some researchers finding little support for bio-
geographic patterns among soil bacteria (93, 94) and others
finding that a combination of edaphic and spatial patterns
explain the patterns of bacterial distributions (95). Data
from marine systems are similarly equivocal. At one extreme,
a study utilizing pyrosequencing of the V6 region of the small
subunit ribosomal rRNA gene indicated that 78% and 70%
operational taxonomic units were unique to the Southern
Ocean and Arctic Ocean, respectively, indicating that
many marine pelagic bacteria exhibit biogeographic patterns
(96). The ubiquitous, dominant pelagic marine heterotro-
phic bacterial clade, SAR 11, exemplifies the opposite
extreme. Phylotypes in this clade (SAR 11 most likely com-
prises a monophyletic family) are globally distributed with
the distribution of each phylotype correlated to temperature
and latitude. In addition, the genomes of an Artic/Antarctic
phylotype (designated P3.2) are strikingly similar, suggesting
that for these organisms, exchange of individuals overwhelms
the processes of speciation and extinction, or, in other words,
that the polar SAR 11 populations are operating as a metapo-
pulation (97).
ANALYSES OF BENTHIC MICROBIAL
COMMUNITIES
Sampling
As with any ecological study, a key to experimental design is
obtaining a representative sample of the population, com-
munity, or ecosystem being studied. The general topic of sam-
pling is covered in Section 2.6 of this manual, so we limit our
comments to those aspects that are particular to benthic hab-
itats, including patchiness, disturbance during sampling, and
the use of mesocosms. In fact, as mentioned already, sedimen-
tary ecosystems are highly heterogeneous in space and time.
Continuous reworking and remolding of the sediments by
currents and burrowing organisms, for instance, induce a
high degree of physical, chemical, and biological complexity
that can change from the millimeter to the kilometer scale.
This fact needs to be taken into account for sampling design
of sedimentary ecosystems.
Any sampling protocol should be linked to the ecological
question being addressed with regard to both spatial and tem-
poral scales. At the centimeter to kilometer scale, Federle
et al. (98) provide a comprehensive analysis of spatial varia-
tion in microbial community structure in an estuarine mud-
flat. Similarly, MacGregor et al. (99) provide a similar study
at the micro-vertical scale. For environments that experience
disturbance, temporal scales become important with the time
for communities to return to predisturbance state being the
relevant timeframe. Disturbance during sampling can affect
microbial communities in a variety of ways, including altera-
tion of in situ activities or loss of biomass (either total or of
4.2.1-10 ▪ AQUATIC ENVIRONMENTS

specific populations). As discussed in the section “Elemental
Cycles,” benthic microbial communities can be ordered by
prevailing redox conditions, and as shown by MacGregor
et al. (99) and many others, the distances between different
metabolic types can be as little as 100 µm. Any sample proc-
essing that disrupts the redox gradient can dramatically alter
the metabolic activity of the community (100). In many sys-
tems where flow does not disrupt sediment surfaces, a fine
layer of low-density particles is associated with the sedi-
ment–water interface (referred to as “fluff” in field jargon).
These fluff layers can contain high microbial biomass, can
be areas of high microbial activity, and may show a commun-
ity structure that is significantly different from the underlying
sediments. Any sampling of sediments that produces a “bow-
wave” as the sampler approaches the sediment is likely to dis-
rupt surficial sediments and significantly under sample the
associated microorganisms (see Schonfeld [101] for a recent
discussion). Mesocosms are a useful tool in ecological studies
but must be applied with care (102, 103). We use a variety of
both short- and long-term mesocosms in our work (bioreac-
tors, flow-through chambers, artificial streams; Fig. 4) and
strongly recommend that all researchers availing themselves
of these useful tools but always compare the communities
present within the mesocosms to those found in the ecosys-
tem being studies (36).
Sediments
When possible, sediments should be sampled by hand. This
can be accomplished using push cores or sampling rings (see
Findlay and Watling [104] and Findlay et al. [105] for
descriptions of both approaches). Scuba-assisted divers can
sample sediments that cannot be reached by wading. Sedi-
ments should be visually examined to ensure that surficial
sediments are intact, and any samples where the surficial sedi-
ment has been disrupted should be discarded. Where logistics
or expense preclude hand-sampling, ship-board indirect sam-
pling will be required, but the same diligence in collection of
undisturbed sediments is required. Depending on the assays to
be performed, samples can be stored on ice, frozen in the field
or on arrival in laboratory, or in some cases preserved with for-
malin (106).
Hard Surfaces
Microbial biomass can be harvested from hard surfaces using
physical agents alone or in combination with chemicals. For
instance, sterile blades or small brushes are generally sufficient
to quantitatively collect biomass from stones and rocks.
Scrapings can be collected in a clean tray or petri dish and
transferred into vials. The area scraped can be photographed
and determined by image analysis. If the surface is too tex-
tured or the biomass must be removed from sediments, the
action of tetrasodium pyrophosphate and mild sonication is
efficient in removing biomass. The detergent tetrasodium
pyrophosphate improves the detachment of microbial cells
from surfaces and furthermore prevents cell aggregation in
the sample after detachment. This may be important for
some single-cell analyses, including cell counts using epifluor-
escence microscopy and flow cytometry. Biomass intended for
molecular analyses is best frozen immediately in liquid nitro-
gen pending macromolecular extraction.

FIGURE 4. Meso- and microcosms typically used to study benthic microbial communities. (a) A header tank ensures constant pressure and
computerized valves regulate the water flow in 32 3-m-long flumes. (b) Large (40 m long) streamside flumes are typically used to study the effect
of the flow environment on benthic microbial life and the effect of benthic microorganisms on ecosystem processes. (c) Bioreactors serve study
how benthic microorganisms transform dissolved organic or how bioturbation affect sedimentary microbial communities. Courtesy of Mia
Bengtsson and Iris Hödl. doi:10.1128/9781555818821.ch4.2.1.f4
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-11
Measurement of Activity
Activity is classically defined as a rate measurement or a unit
mass change per unit time. These values are occasionally bio-
mass corrected. With the advent of modern molecular meth-
ods, the analysis of RNA—both rRNA and mRNA—has
been viewed as representing the “active” community. How-
ever, these RNA-based measurements do not yield a rate
(for further discussion of “the active community” see the sec-
tion “Identifying the Active Community”). Activity meas-
urements can be divided into two broad categories, those
returning a rate of metabolic activity (i.e., rate of oxygen con-
sumption, carbon dioxide production) and those returning an
estimation of bacterial growth.
Classical
Classic rate measurements based on substrate or product con-
centrations require a minimum of two measurements sepa-
rated by time. Intact cores are incubated under in situ
conditions (with temperature likely being the most critical)
modified to prevent competing or confounding processes
(for example, O2 consumption and CO2 production measure-
ments would be conducted in the dark to preclude photosyn-
thesis) with time periods chosen to avoid changes in rate due
to changes in substrate or product concentrations. In systems
where flow is an important environmental determinant, pro-
visions should be made to mimic in situ flows. Substrate or
product concentrations can be measured by titration (micro-
Winkler, Gran, etc.), electrodes (if available), or analytical
instrumentation. Careful application of these techniques
can yield near unity for labile carbon supply, O2 consump-
tion, and CO2 production (104, 107). Some rate measure-
ments are more easily accomplished by use of radiolabeled
tracers (e.g., sulfate reduction; 108). The development of
microelectrodes provided an alternative methodology that
utilizes diffusion gradients for rate determination (109) and
has the added benefit of providing accurate determination
of the vertical distribution of metabolic activity. These stud-
ies also provide convincing evidence of the importance
of flow as a determinant of microbial distribution and
abundance, as well as the need to mimic in situ flows when
conducting rate measurements (110). The increasing avail-
ability of isotope ratio mass spectrometers has spurred the
development of stable isotope–based rate measurements.
Using methane oxidation as an example, it is now possible
to determine the rate of methane oxidation by tracking the
conversion of 13C-CH4 to 13C-CO2 (111) or calculate rates
of methane production, oxidation, and transport over an
annual cycle using isotope profiles and a one-dimensional
diffusion–reaction model (112). The development of field
deployable cavity ring down spectrometers, some with
13
C-CH4 and 13C-CO2 capacity (commercially available
from Picarro) portends increased ease and accuracy of in situ
activity measurements (for recent example, see 113).
For samples amenable to microscopic analysis, bacterial
growth rates were traditionally calculated from the frequency
of dividing cells (114). Assumptions as to the mass of visual-
ized cells provided an estimate of biomass production. Adap-
tation of flow cytometry, image analysis, and artificial neural
networks has provided a degree of automation for this proce-
dure, reducing analysis time and tedium (115, 116). These
techniques typically use fluorescent stains to aid in the visual-
ization of bacteria. One staining procedure produces a ratio of
living-to-dead cells by using two different nucleic acid stains
to distinguish live bacteria with intact plasma membranes
from dead bacteria with compromised membranes (a number
of commercial firms provide kits for this analysis). While this
measurement is technically not a rate measurement it is some-
times, possibly erroneously (if environmental bacteria can be
alive but not active), referred to as percent active bacteria.
Measurement of the electron transport system (ETS) activity
has also been used to determine the percent active bacteria
within a population (117). In this technique, ETS acti-
vity reduces 2-( p-iodophenyl)-3-( p-nitrophenyl)-5-phenyl
tetrazolium chloride (INT) to INT-formazan (118) or 5-
cyano-2,3-tolyl-tetrazolium chloride (CTC) to CTC-forma-
zan. It should be noted that significant technical difficulties
have been identified when applying this technique to sedi-
ments (119, 120). A similar technique, using radiolabeled
substrates and microautoradiography, also yields an estimate
of respiring cells (121–123) although this approach seems
to be sensitive to the substrates used with the highest percen-
tages found with mixtures of amino acids (124). In contrast
to tetrazolium-based techniques, microautoradiographic
approaches, especially when coupled with fluorescent in situ
hybridization approaches, are still commonly applied, espe-
cially in the study of sewage treatment (125). Their applica-
tion in sediments will likely pose significant technical
challenges associated with the introduction of labeled sub-
strates without disturbance (see the next section for further
discussion).
Biochemical
All of the biochemical components of cells can be useful
in investigating the ecology of microorganisms, with the
synthesis of macromolecules being particularly useful for
measurement of growth rates. Three main groups of macromo-
lecules—DNA, protein, and lipids—form the basis for most
approaches. In each case, the concentration of the macromo-
lecule in the typical bacterial cell, the number of monomers
typically incorporated into each macromolecule, and the
rate of monomer incorporation (determined using a radiola-
beled tracer) are used to determine bacterial growth rate.
Production can be determined by estimating the average
mass of a bacterium. Estimates based on DNA synthesis use
3
H-thymidine as the precursor (126, 127), those based on
protein synthesis use 3H-leucine as the precursor (128), and
those based on lipid synthesis use 14C-actetate as the precur-
sor (129). Each method has its particular challenges when
applied within benthic environments. As the rates of protein
and lipid synthesis can change very rapidly with changes in
bacterial metabolic status, these assays are strongly affected
by sediment disruption. Extreme care must be used when
introducing precursors and incubation times should be lim-
ited to minutes. Mixing, while desirable in that it produces
a uniform label distribution, disrupts sediment, which can
increase measured growth rates 10-fold (129). This sensitivity
to disruptions extends to substrate introduction by sediment
perfusion (130). While rates of DNA synthesis seem unal-
tered by mixing, at least during the incubation period (typi-
cally 2 h), thymidine is catabolized during this period,
which does have strong consequences for estimates of bacte-
rial production (131). As with classic rate measurements,
the increasing availability of isotope-ratio mass spectrometers
has led to the development of stable isotope–based growth
measurements. Wegener et al. (132), using a combination
of deuterated water (D2O), 13C-DIC, and 13C-acetate incor-
poration into lipid, were able to calculate growth rates from
subsurface marine sediments of 4,600 × 106 and 0.18 × 106
cells/cm/y, respectively. These rates should not be considered
in situ rates for the reasons outlined above, as incubations were
carried out as slurries and for 21 days.
4.2.1-12 ▪ AQUATIC ENVIRONMENTS
Molecular
The strong correlation between growth rate and cell rRNA
content (first expressed as RNA/DNA ratio; 133) is also use-
ful as a measure of bacterial growth rates and can be used at the
community (134), population (135), and even the single cell
level (136). This relationship provides a compelling link
between microbial ecology and ecological stoichiometry,
which is the study of how the balance of energy and elements
affects and is affected by organisms and their interactions in
ecosystems (137, 138).
Measurement of Abundance and Biomass
Abundance and biomass are key characteristics of any micro-
bial population or community and can be determined by a
number of classical and biochemical techniques.
Direct Microscopic Analysis
The total number of microbial cells per unit volume or
mass, often referred to as direct counts, can be counted using
well-established procedures (139). These typically include
preserving the sample, most commonly with formaldehyde,
removing the bacteria from the sediment and staining with
a fluorochrome. The two most often used fluorochromes
are 3,6-bis[dimethylamino]acridinium chloride (acridine
orange [AO]) and 40 ,6-diamidino-2-phenylindole (DAPI).
Although AO was the first introduced and still has efficacy
in specific applications, DAPI is the more widely used, with
new fluorochromes being developed continuously (e.g.,
YOYO-1, SYBR II, SYTO-13). Cells are collected on a filter
and observed under an epifluorescence microscope. Chal-
lenges encountered with this approach, besides analyst
fatigue, include interference from particle masking and differ-
entiating detrital particles and bacteria. Semi and full auto-
mation using either image capture and analysis or flow
cytometry reduce some of these difficulties (140, 141).
Biochemical
A variety of the biochemical constituents of cells have been
used as measures of total microbial biomass, including mur-
amic acid, ATP, and phospholipid phosphate. Of these, the
phospholipid phosphate assay is the most commonly applied.
Samples are extracted with a modified Bligh-Dyer extraction
(142), the phospholipid recovered in the organic solvent
phase, phospholipid oxidized to CO2 and ortho-phosphate,
and the resulting ortho-phosphate detected after formation
of colored complex (143–145). Strengths of this approach
include that the lipid extraction and phosphate determina-
tions are typically quantitative and phospholipids are found
in all cells. In most environments phospholipids have a
short half-life in detrital pools and the technique is sensitive
enough to be applied to several hundred milligrams of
sediment. In many environments estimates of bacterial abun-
dance made using direct counts and phospholipid phosphate
analysis are highly correlated (105). The weakness of this
approach is that it requires access to solvent-handing equip-
ment and the use of conversion factors to relate moles of
recovered phosphate to cell counts or biomass.
Molecular
For the most part, molecular methods are currently not suit-
able for the measurement of abundance or biomass because
DNA is well preserved in many sedimentary systems (146,
147). Furthermore, it is difficult, if not impossible, to recover
quantitatively DNA and rRNA from sediments (148). The
rRNA content of cells also varies greatly with cell growth rate.
Measurement of Diversity and Community Structure
Species diversity is a complex metric typically viewed as
including two components—species richness, or the number
of species present within a sample or community, and species
evenness, or the relative abundance, one to another, of the
species present. Community structure is a metric of even
greater complexity because it considers taxonomic diversity,
as well as biotic interactions from mutualisms to food webs.
Each of the three major methodological approaches com-
monly applied to benthic microbial ecology offer insight
into at least two of these characteristics of biological com-
munities and ecosystems.
Direct Microscopic Analysis
Direct observation of microbial cells, when coupled with
gene-specific dyes, either through fluorescent in situ hybrid-
ization (FISH) or in one of its variants, can provide quantifi-
cation of types of microorganisms along with a level of
taxonomic resolution (see 149, 150 for convenient over-
views). The basic procedures involve fixing and permeabiliz-
ing the cells, hybridization with a fluorescently tagged
oligonucleotide (often referred to as the probe), washing
excess probe from the sample, and viewing the sample using
microscopy or flow cytometry. A wide number of fluorophores
are available and it is possible to use more than one probe at a
time. Since the introduction of this technique, there have
been steady improvements (with the corresponding prolifera-
tion of acronyms). One of the most useful is catalyzed reporter
deposition or CARD-FISH, which uses horseradish peroxi-
dase–labeled probes to provide signal amplification overcom-
ing low signal strength of conventional FISH when applied to
bacteria with low rRNA content. Other advancements
include coupling FISH with other technologies to amplify
genomic DNA and using “decorator” probes to locate the tar-
geted cell. A detailed description of these approaches is
beyond the scope of this chapter and interested investigators
are directed to chapter 2.2.1 within this manual and a recent
review (151). The greatest strengths of this approach are that
it allows the investigator to see the targeted organisms within
their environment changed only by the fixation, permeabili-
zation, and hybridization steps; the high degree of taxonomic
information obtained; and the quantitative delivery of the
relative abundance (evenness) of the organisms detected.
The greatest weaknesses are the same as those encountered
with direct count determination of total abundance (similarly
ameliorated by semi or full automation) and that detection is
sequence-specific, which makes detection of unknowns
difficult.
Biochemical
Phospholipid fatty acids are commonly used to provide a
measure of microbial community structure of benthic com-
munities. Total lipid, extracted as for biomass (see the pre-
vious “Biochemical” section), is separated into fractions
using a silicic acid column (typically neutral, glyco-, and
phospholipid), the phospholipid fatty acids are converted to
fatty acid methyl esters (FAMEs), and individual FAMEs
are quantified using gas chromatography (152). The resulting
fatty acid profiles provide a fingerprint of the community,
which typically correlate well with those provided by molec-
ular fingerprinting techniques (153; see the next section on
“Molecular” for discussion of molecular fingerprinting tech-
niques). The usefulness of the technique derives from the
development of species fatty acid profiles as a phenotypic
character (widely accepted by the International Journal of
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-13
Systematic and Evolutionary Microbiology as a diagnostic char-
acteristic), the short detrital half-life of phospholipids, and
the quantitative nature of the assay. The major weakness of
the approach, along with those mentioned in the previous
“Biochemical” section, is the relatively low taxonomic infor-
mation returned. While it is quite easy to calculate the ratio of
bacterial biomass to microeukaryotic biomass, it is not possi-
ble to confirm the presence of a specific genus or species
(although if a genus contains a marker fatty acid, it is possible
to exclude their presence in a sample). Although directed to
soil microbial ecologists, we recommend a recent review by
Frostegard et al. (154) for a frank discussion on the
strengths and pitfalls associated with the application of
the phospholipid fatty acid analysis. It is possible to expand
the approach to include Archaea as their phospholipids
remain after FAME conversion and can be separated from
the FAMEs using a second silicic acid column. Quantifica-
tion can be either by high-temperature gas chromatography
(155) or high-performance liquid chromatography/atmos-
pheric pressure chemical ionization mass spectrometry
(156). A more detailed fingerprint of the microeukaryotic
community is also possible, especially with regard to differ-
entiating green plants and fungi (which can be problematic
using FAME analysis), by collecting the neutral lipid frac-
tion and processing it for sterols (see 157–159 as recent
examples).
Molecular
No area of methods development within microbial ecology
has seen greater activity or growth in the past few decades
than the application of molecular methods for the determina-
tion of microbial diversity and community structure. Several
chapters (2.3.2, 2.3.3, 2.4.1–2.4.5) within this manual are
devoted to this subject and investigators interested in apply-
ing these techniques are directed to these. Here we provide
an overview of current perspectives and direct investigators
to recent relevant reviews. For example, see Fig. 1 in Su
et al. (160) for a particularly comprehensive overview of
culture-independent methods used to assess microbial diver-
sity and community structure in environmental samples. Sev-
eral genomic fingerprinting techniques have been developed
and widely applied, most notably denaturing gradient gel
electrophoresis and terminal restriction fragment length poly-
morphism. These share DNA extraction, purification, and
PCR amplification but differ in the treatment and separation
of the resultant products. Patterns of change among profiles
are quite consistent with differences in determinants of
microbial distribution and abundance, indicating a high util-
ity for documenting differences in structure among commun-
ities; however, fragments vary in taxonomic specificity (161).
Most researchers analyze data at the presence/absence level
(nonparametric multidimensional scaling is the most com-
monly applied statistical technique), although it is possible
to quantify the amount of PCR product present with the
gels. For the genomic fingerprinting techniques, it is impossi-
ble to avoid PCR biases completely (160), although it is pos-
sible with some hybridization-based techniques (e.g., DNA
microarray, FISH) to obtain sufficient signal strength to avoid
PCR amplification. Production of 16S rRNA gene libraries is
the other traditional molecular method that is commonly
applied. While some researchers advocate for production
clone libraries without PCR amplification (162), most accept
the biases induced by PCR amplification for the convenience
of clone library construction with amplified gene sequences.
High-throughput pyrosequencing or Illumina sequencing
(e.g., see 163) now allows for the rapid and relatively
inexpensive sequencing of vast amounts of DNA. Two major
approaches within microbial ecology—16S rRNA gene
amplicon analysis and metagenomics/metatranscriptomics—
rely on this approach. In 16S rRNA gene amplicon analysis,
primers for one or more hypervariable regions, known as V1
through V9, are used to amplify gene fragments from the
DNA extracted from a microbial system and yields estimates
of diversity based on the sequences discerned. Although
this is a powerful technique that has yielded remarkable dis-
coveries, it is not without its biases. Lee et al. (164) con-
structed a “microbial system” from 20 plasmids, each
containing a unique rRNA gene cloned from 20 different bac-
teria. Purified plasmid DNA was mixed and the V3–V4 and
V6 hypervariable regions were sequenced using the Roche
454 GS FLX Titanium platform. After comparing expected
and observed relative abundance of the 20 “species,” they
concluded,
Commonly used quality control protocols resulted in the
formation of OTUs with >1% abundance composed entirely
of erroneous sequences, while over-aggressive clustering
approaches obfuscated real, expected OTUs. The pyrose-
quencing process itself did not appear to impose significant
biases on overall community structure estimates, although
the detection limit for rare taxa may be affected by PCR
amplicon size and quality control approach employed.
Meanwhile, PCR biases associated with the initial amplicon
generation [the actual temple that is pyrosequenced] may
impose greater distortions in the observed community
structure.
Metagenomics/metatranscriptomics are used to investi-
gate the functional and expressed genetic potential of micro-
bial systems. As with 16S rRNA gene amplicon analysis,
DNA is extracted from a microbial system; however, in meta-
genomics analysis DNA is sheared to produce to templates,
the templates are clonally amplified and sequenced by synthe-
sis. Bioinformatics tools are used to compile sequences into
genes, operons, and individual genomes. Metatranscriptom-
ics uses a similar approach except that RNA is extracted
and reverse transcribed to cDNA and templates are produced.
The resultant sequences are analyzed and interpreted as
microbial gene expression by that microbial system, and this
is often interpreted as representing the active community
(see “Identifying the Active Community”). Interestingly,
these approaches can yield a large number of noncoding
RNA sequences, many of which may be involved in gene reg-
ulation (165, 166).
The application of high-throughput sequences is evolv-
ing quickly, and researchers should stay abreast of current
developments. For example, recent refinements in quality
control procedures and clustering techniques have greatly
reduced errors for pyrosequencing estimates of species rich-
ness (167, 168). As of this writing, comparable methodology
has yet to be developed for metagenomic approaches, render-
ing them less reliable for species richness estimations; how-
ever, they have proven more accurate for the determination
of relative abundance than amplicon sequencing (169,
170). Refinements of competing systems are continually
being made and can greatly influence the relative accuracy
(171, 172).
Methods Specific to Topics within 2.0
Methods for Characterizing DOM
DOM in water is a complex mixture of thousands of com-
pounds, few of which are easily categorized as lipids, carbohy-
drates, or proteins. These compounds originate from living
4.2.1-14 ▪ AQUATIC ENVIRONMENTS
and decaying organisms and can be modified by physical,
chemical, and biotic processes. Characterization of this
organic matter pool is a daunting challenge. Total DOM is
most often quantified as dissolved organic carbon using a total
organic carbon analyzer. Determinations of the chemical
nature or quality of DOM can utilize fluorescence spectro-
scopy (173), nuclear magnetic resonance (174), or mass spec-
tral analysis (175). One of the most promising approaches in
this field involves combining a suite of analytical methods to
provide molecular-level characterization (176).
Methods for Food Web Analysis
The combining of analytical approaches is leading to advan-
ces in studying the role of microorganisms in benthic food
webs. In particular, combining fatty acid analysis with
compound-specific stable isotope analysis is proving highly
useful. In this context, fatty acid analysis relies on the pres-
ence of specific fatty acids within the aqueous phase of the
digestive tract or the storage lipid of a consumer to indicate
its food source (64, 65, 177). Carbon stable isotope analysis
reveals the source of carbon and energy used by an organism.
When combined in fatty acid stable isotope probing, these
techniques offer a degree of specificity previously unattain-
able. For example, Sanseverino et al. (178) were able to dem-
onstrate that biogenic methane, converted to bacterial
biomass by methane-oxidizing bacteria, was an important car-
bon source for the tropical aquatic food web, including several
fishes. They found marker fatty acids for methane-oxidizing
bacteria within the total fatty acids of the fishes and showed
that these fatty acids were highly depleted in 13C (biogenetic
methane is the most 13C depleted carbon typically found in
the environment).
Stable isotope ratios of nitrogen are commonly used to
determine the trophic position of organisms within food
webs. One difficulty associated with this approach is that it
is often difficult to establish the δ15N of the trophic base
(179). Compound-specific, stable isotope analysis of amino
acids (180) provides methodology for simultaneously deter-
mining the trophic base of a food web and an organism’s tro-
phic position (181). The Findlay laboratory is currently
investigating the efficacy of this approach within freshwater
food webs (182).
Identifying the Active Community
Identifying the “active community” has been informally
defined within microbial ecology with a variety of techniques.
Bacteria with sufficient rRNA to be detected by traditional
FISH are often considered active because of the correlation
between rRNA content and growth rate. Similarly, the
organisms contributing mRNA to a metatranscriptomics
analysis are also considered active. Stable isotope probing
(SIP) provides a methodology to determine which microor-
ganisms present within an ecosystem are utilizing a particular
food resource and are therefore active in performing an eco-
system process. Two classes of macromolecules are commonly
used in this approach—phospholipid fatty acids (PLFAs) and
nucleic acids, although sterols are quite likely also amenable
to this approach. As DNA- and RNA-SIP are covered in
this manual, we focus here on PLFA-SIP. In PLFA-SIP, a
13
C-labeled substrate is introduced into a sample with the
least disruption possible (see “Elemental cycles and the distri-
bution of microorganisms in sediment”) and incubated (times
range from minutes to days depending on substrate). Lipids
are extracted, processed as if for community structure analysis
(see previous section “Biochemical”), and the increased
presence of 13C determined for individual fatty acids using
compound-specific, stable isotope analysis (183).
INTERPLAY AMONG METHODS: CLASSICAL,
BIOCHEMICAL, MOLECULAR
Microbial ecology has long been considered to be a science
limited by the available methodology and as such highly vul-
nerable to the “law of the instrument”. This law has been
attributed to several authors and has a number of formula-
tions, but for our purposes it may be stated: “When the only
tool you have is a hammer, all your problems appear as nails.”
On the other hand, master craftsmen espouse the philosophy
“the right tool for the right job.” It is difficult to argue that
microbial ecology has a resplendent toolbox as most environ-
mentally relevant organisms are as yet uncultured and are
unidentifiable by direct observation, forcing microbial ecolo-
gists to rely on indirect methods to determine structure and
function of microbial communities. However, the advent of
molecule sequence–based methods have partly transformed
the study of microbial communities and for the first time
provide the ability to describe communities in terms of the
species present. This is the first step in developing a mechanis-
tic understanding of microbial communities. While many of
the initial concepts within ecology were established in the
17th and 18th centuries by microscopist Antonie van Leeu-
wenhoek and botanist Richard Bradley, it wasn’t until the
late 19th century, when natural history was transformed
into a quantitative science, that ecology was born. Of the
available classical, biochemical, and molecular methods
available for application to the significant questions within
microbial ecology, none return both species identity and pop-
ulation size for all organisms within a community. As such,
they provide complimentary data and researchers should con-
sider the strengths and weakness of each as they are applied.
Microbial ecologists should remember that most, if not all,
experimental designs will benefit by the judicious application
of a suite of analytical techniques.
CURRENT TOPICS OF INTEREST
We provide a noncomprehensive list of current topic of inter-
est within the general field of benthic microbial ecology,
likely biased by our personal research and teaching interests.
The order of presentation should be considered chaotic.
• A comprehensive understanding of “What is a bacterial
species?”—few, if any, benthic bacteria have been studied
to the extent that their core and pan genomes are well
understood. Likely candidates for this type of analysis
should be present among sulfate- and iron-reducing
bacteria.
• A rapidly emerging field is microbial biogeography, and
there are a number of questions as it relates to benthic
microbial ecology. What is the biogeography of benthic
bacteria and archaea? Are patterns similar to the patterns
beginning to emerge in the marine environment? Are
there any universally distributed benthic bacteria and
archaea? How substantial are the differences in patterns
among stream, lake, and shallow marine habitats? Can
we write laws that predict alpha, beta, and gamma diver-
sity patterns?
• Another question challenging all of ecology, with benthic
microbial ecology making significant contributions, con-
cerns biodiversity and its relationship to functional redun-
dancy. As an example of a potential “prokaryotic” wrinkle
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-15
in this overall discussion, it is conceivable that genetic
variation within the pan genome is a mechanism by which
stream heterotrophs adapt to local variations in DOM
molecular composition.
• There is a concerted effort to explore the general applic-
ability of ecological theory (typically generated with
data produced by studying macroorganisms) to the ecology
of microorganisms. For all involved in this process, we rec-
ommend a small book by Andrews (184)—among its
substantial insights is the ever amusing quote “To oversim-
plify, one could say that macroecology consists of phenom-
ena in search of mechanistic explanation, whereas
microbial ecology is experimentation in search of theory.”
See the Theoretical Ecology section of this book for a
complete discussion of this topic.
• As previously stated, the interrelationship among DOM
molecular composition and the structure and func-
tion of heterotrophic microorganisms inhabiting benthic
environments is a complex and an important area of
study.
• Understanding the role of microorganisms within benthic
food webs remains one of the largest unaddressed chal-
lenges within microbial ecology and will require an effort
similar to that directed toward resolving the role of micro-
organisms in marine pelagic food webs.
• The role of viruses is understudied in benthic systems, and
their role in structuring benthic microbial systems and in
controlling elemental fluxes (viral shunt) has yet to be
resolved.
• The interactions between phototrophic and heterotro-
phic microorganisms throughout benthic ecosystems
deserve further study. In virtually every study conducted
by Findlay and co-workers where seasonality impacted
benthic microbial community structure, the major com-
ponent of variation was the balance between eukaryotes
and bacteria. Phototrophic microeukaryotes can account
for as much as 90% of total microbial biomass or be virtu-
ally absent from the community (e.g., 38).
• The role of archaea in benthic systems. With the discovery
of aerobic, mesophilic Thaumarchaeota and Euryarch-
aeota came the recognition that archaea were not rele-
gated to extreme environments and the need to discover
their ecological significance (185, 186). Ammonia-oxi-
dizing archaea provide the most obvious example, to
date, of unexpected ecological significance of a previously
unknown lineage of archaea (187). One difficulty in the
study of many archaea is the high molecular weight of their
lipids, which necessitates liquid chromatography (com-
pared with gas chromatography for most bacterial lipids)
and as a result few labs can quantify archaeal biomass in
benthic systems.
REFERENCES
1. Orcutt BN, Sylvan, JB, Knab, NJ, Edwards, KJ. 2011.
Microbial ecology of the dark ocean above, at, and below
the seafloor. Microbiol Mol Biol Rev 75:361–422. doi:
10.1128/MMBR.00039-10.
2. Madigan MT, Martinko JM, Stahl DA, Clark DP. 2012.
Brock Biology of Microorganisms, 13th ed. Benjamin Cum-
mings, Boston, MA.
3. Kristensen E, Penha-Lopes G, Delefosse M, Valdemarsen
T, Quintana CO, Banta GT. 2012. What is bioturbation?
The need for a precise definition for fauna in aquatic
sciences. Mar Ecol Prog Ser 446:285–302. doi:
10.3354/meps 09506.
4. Boudreau BP, Jorgensen BB. 2001. The Benthic Boundary
Layer: Transport Processes and Biogeochemistry. Oxford Uni-
versity Press, New York.
5. Chao A, Chiu C-H, Hsieh TC. 2012. Proposing a resolu-
tion to debates on diversity partitioning. Ecology 93:
2037–2051.
6. Whittaker RH. 1960. Vegetation of the Siskiyou Moun-
tains. Or Calif Ecol Monogr 30:279–338.
7. Zuckerkandl E, Pauling L. 1965. Molecules as documents
of evolutionary history. J Theor Biol 8:357–366.
8. Kettler GC, Martiny AC, Huang K, Zucker J, Coleman
ML, Rodrigue S, Chen F, Lapidus A, Ferriera S, Johnson
J, Steglich C, Church GM, Richardson P, Chisholm SW.
2007. Patterns and implications of gene gain and loss in the
evolution of Prochlorococcus. PLoS Genet 3:2515–2528.
doi: 10.1371/journal.pgen.0030231
9. Griffith F. 1928. The significance of pneumococcal types.
Epidemiol Infect 27:113–159. doi: 10.1017/S0022172400
031879
10. Redfield R. 1988. Evolution of bacterial transformation—
is sex with dead cells ever better than no sex at all. Genetics
119:213–221.
11. Koonin EV, Wolf YI. 2012. Evolution of microbes and
viruses: a paradigm shift in evolutionary biology? Front
Cell Infect Microbiol 2. doi: 10.3389/fcimb.2012.00119.
12. Hall-Stoodley L, Costerton J, Stoodley P. 2004. Bacterial
biofilms: from the natural environment to infectious dis-
eases. Nat Rev Microbiol 2:95–108. doi: 10.1038/nrmicro
821.
13. Schopf JW, Kudryavtsev AB. 2012. Biogenicity of Earth’s
earliest fossils: a resolution of the controversy. Gondwana
Res 22:761–771. doi: 10.1016/j.gr.2012.07.003.
14. Craddock PR, Dauphas N. 2011. Iron isotopic composi-
tions of geological reference materials and chondrites.
Geostand Geoanalyt Res 35:101–123. doi: 10.1111/
j.1751-908X.2010.00085.x.
15. Altermann W. 2008. Accretion, trapping and binding of
sediment in archean stromatolites—Morphological expres-
sion of the antiquity of life. Space Sci Rev 135:55–79. doi:
10.1007/s11214-007-9292-1.
16. Meysman FJR, Middelburg JJ, Heip CHR. 2006. Biotur-
bation: a fresh look at Darwin’s last idea. Trends Ecol Evol
21:688–695. doi: 10.1016/j.tree.2006.08.002.
17. Bottjer DJ. 2010. The Cambrian substrate revolution and
early evolution of the phyla. J Earth Sci 21:21–24.
18. McIlroy D, Logan G. 1999. The impact of bioturbation on
infaunal ecology and evolution during the Proterozoic-
Cambrian transition. Palaios 14:58–72. doi: 10.2307/
3515361.
19. Rogov V, Marusin V, Bykova N, Goy Y, Nagovitsin K,
Kochnev B, Karlova G, Grazhdankin D. 2012. The oldest
evidence of bioturbation on Earth. Geology 40:395–398.
doi: 10.1130/G32807.1.
20. Mata SA, Bottjer DJ. 2012. Microbes and mass extinc-
tions: paleoenvironmental distribution of microbialites
during times of biotic crisis. Geobiology 10:3–24. doi:
10.1111/j.1472-4669.2011.00305.x.
21. Yang X-H, Lee C, Scranton MI. 1993. Determination of
nanomolar concentrations of individual dissolved low
molecular weight amines and organic acids in seawater.
Anal Chem 65:572–570.
22. Wu H, Green M, Scranton MI. 1997. Acetate cycling in
the water column and surface sediment of Long Island
Sound following a bloom. Limnol Oceanogr 42:705–713.
23. Albert DB, Martens CS. 1997. Determination of low-
molecular-weight organic acid concentrations in seawater
and pore-water samples via HPLC. Mar Chem 56:27–37.
4.2.1-16 ▪ AQUATIC ENVIRONMENTS
24. Atlas RM, Bartha R. 1993. Microbial Ecology: Fundamen-
tals and Applications, 3rd ed. Benjamin/Cummings, Red-
wood City, CA.
25. McArthur JV. 2006. Microbial Ecology: An Evolutionary
Approach. Elsevier, Amsterdam.
26. Kirchman DL. 2012. Processes in Microbial Ecology. Oxford
University Press, New York.
27. Reimers CE, Stecher H III, Taghonb GL, Fullerb CM,
Huettel M, Ruschc A, Ryckelyncka N, Wild C. 2004.
In situ measurements of advective solute transport in per-
meable shelf sands. Cont Shelf Res 24:183–201.
28. Cook PLM, Wenzhofer F, Glud RN, Janssen F, Huettel
M. 2007. Benthic solute exchange and carbon mineraliza-
tion in two shallow subtidal sandy sediments: effect of
advective pore-water exchange. Limnol Oceanogr 52:
1943–1963.
29. Chipman L, Podgorski D, Green S, Kostk J, Cooper W,
Huettela M. 2010. Decomposition of plankton-derived
dissolved organic matter in permeable coastal sediments.
Limnol Oceanogr 55:857–871.
30. Thistle D, Reidnauer JA, Findlay RH, Waldo R. 1984.
An experimental investigation of enhanced harpacticoid
(Copepoda) abundances around isolated seagrass shoots.
Oecologia 63:295–299.
31. Battin T, Kaplan L, Newbold J, Hendricks S. 2003. A
mixing model analysis of stream solute dynamics and the
contribution of a hyporheic zone to ecosystem function.
Freshw Biol 48:995–1014. doi: 10.1046/j.1365-2427.
2003.01062.x.
32. Bottacin-Busolin A, Singer G, Zaramella M, Battin TJ,
Marion A. 2009. Effects of streambed morphology and bio-
film growth on the transient storage of solutes. Environ Sci
Technol 43:7337–7342. doi: 10.1021/es900852w.
33. Woodcock S, Besemer K, Battin TJ, Curtis TP, Sloan
WT. 2013. Modelling the effects of dispersal mechanisms
and hydrodynamic regimes upon the structure of microbial
communities within fluvial biofilms. Environ Microbiol 15:
1216–1225. doi: 10.1111/1462-2920.12055.
34. Besemer K, Singer G, Hoedl I, Battin TJ. 2009. Bacterial
community composition of stream biofilms in spatially
variable-flow environments. Appl Environ Microbiol 75:
7189–7195. doi: 10.1128/AEM.01284-09.
35. Holmes R, Fisher S, Grimm N, Harper B. 1998. The
impact of flash floods on microbial distribution and bio-
geochemistry in the parafluvial zone of a desert stream.
Freshw Biol 40:641–654. doi: 10.1046/j.1365-2427.
1998.00362.x.
36. Findlay RH, Trexler MB, Guckert JB, White DC. 1990.
Laboratory study of disturbance in marine sediments:
response of a microbial community. Mar Ecol Prog Ser 62:
121–133.
37. Feminella J, Hawkins C. 1995. Interactions between
stream herbivores and periphyton: a quantitative analysis
of past experiments. J N Am Benthol Soc 14:465–509.
doi: 10.2307/1467536.
38. Findlay RH, Watling L. 1998. Seasonal variation in the
structure of a marine benthic microbial community. Microb
Ecol 36:23–30.
39. Sutton S, Findlay R. 2003. Sedimentary microbial com-
munity dynamics in a regulated stream: East Fork of the Lit-
tle Miami River, Ohio. Environ Microbiol 5:256–266. doi:
10.1046/j.1462-2920.2003.00396.x.
40. Smoot JC, Findlay RH. 2001. Spatial and seasonal varia-
tion in a freshwater reservoir sedimentary microbial com-
munity as determined by phospholipid fatty acid analysis.
Microb Ecol 42:350–358.
41. Vinebrooke R, Leavitt P. 1996. Effects of ultraviolet radi-
ation on periphyton in an alpine lake. Limnol Oceanogr 41:
1035–1040.
42. Vinebrooke R, Leavitt P. 1999. Differential responses of
littoral communities to ultraviolet radiation in an alpine
lake. Ecology 80:223–237. doi: 10.1890/0012-9658
(1999)080[0223:DROLCT]2.0.CO;2.
43. Stal LJ. 2010. Microphytobenthos as a biogeomorphologi-
cal force in intertidal sediment stabilization. Ecol Eng 36:
236–245. doi: 10.1016/j.ecoleng.2008.12.032.
44. Garcia-Pichel F, Bebout B. 1996. Penetration of ultravio-
let radiation into shallow water sediments: high exposure
for photosynthetic communities. Mar Ecol Prog Ser 131:
257–262. doi: 10.3354/meps131257.
45. Stal L. 2012. Cyanobacterial mats and stromatolites, p.
65–125. In Whitton BA (ed), Ecology of Cyanobacteria II.
Springer, New York.
46. Elasri M, Miller R. 1999. Study of the response of a biofilm
bacterial community to UV radiation. Appl Environ Micro-
biol 65:2025–2031.
47. Dupraz C, Visscher P. 2005. Microbial lithification in
marine stromatolites and hypersaline mats. Trends Microbiol
13:429–438. doi: 10.1016/j.tim.2005.07.008.
48. Li Y, Huang Y, Wang X. 2012. Suspended sediment as an
attenuating factor of ultraviolet-B radiation effects on the
growth and DNA damage of Chlorolla sp. Adv Mat Res
518:5165–5171.
49. Hulthe G, Hulth S, Hall POJ. 1998. Effect of oxygen on
degradation rate of refractory and labile organic matter in
continental margin sediments. Geochim Cosmochim Acta
62:1319–1328.
50. Emerson S, Hedges J. 2003. 6.11—Sediment diagene-
sis and benthic flux, p. 293–319. In Holland HD, Tur-
ekian KK (eds) Treatise on Geochemistry. Pergamon,
Oxford.
51. Sobek S, Durisch-Kaiser E, Zurbruegg R, Wongfun N,
Wessels M, Pasche N, Wehrli B. 2009. Organic carbon
burial efficiency in lake sediments controlled by oxygen
exposure time and sediment source. Limnol Oceanogr 54:
2243–2254. doi: 10.4319/lo.2009.54.6.2243.
52. Freeman C, Lock M. 1995. The biofilm polysaccharide
matrix—a buffer against changing organic substrate supply.
Limnol Oceanogr 40:273–278.
53. Kim S, Simpson A, Kujawinski E, Freitas M, Hatcher P.
2003. High resolution electrospray ionization mass spec-
trometry and 2D solution NMR for the analysis of DOM
extracted by C-18 solid phase disk. Org Geochem 34:
1325–1335. doi: 10.1016/S0146-6380(03)00101-3.
54. Mosher JJ, Klein GC, Marshall AG, Findlay RH. 2010.
Geological influences of natural organic matter structure
in stream waters by FT-ICR-MS. Org Geochem 41:
1177–1188.
55. Kim S, Kaplan L, Hatcher P. 2006. Biodegradable dis-
solved organic matter in a temperate and a tropical stream
determined from ultra-high resolution mass spectrometry.
Limnol Oceanogr 51:1054–1063.
56. Singer GA, Besemer K, Schmitt-Kopplin P, Hödl I-A,
Battin TJ. 2010. Physical heterogeneity increases biofilm
resource use and its molecular diversity in stream meso-
cosms. PLoS One 5:e9988. doi: 10.1371/ journal.
pone.0009988.
57. Singer G, Besemer K, Hochedlinger G, Chlup AK, Battin
TJ. 2011. Monomeric carbohydrate uptake and structure-
function coupling in stream biofilms. Aquat Microb Ecol
62:71–83.
58. Singer GA, Fasching C, Wilhelm L, Niggemann J,
Steier P, Dittmar T, Battin TJ. 2012. Biogeochemically
diverse organic matter in Alpine glaciers and its down-
stream fate. Nat Geosci 5:710–714. doi: 10.1038/
NGEO1581.
59. Hunter WR, Jamieson A, Huvenne VAI, Witte U. 2013.
Sediment community responses to marine vs. terrigenous
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-17
organic matter in a submarine canyon. Biogeosciences 10:
67–80. doi: 10.5194/bg-10-67-2013.
60. Gause GF. 1934. Experimental analysis of Vito Volterra’s
mathematical theory of the struggle for existence. Science
79:16–17. doi: 10.1126/science.79.2036.16-a.
61. Lopez GP, Levinton JS. 1987. Ecology of deposit-
feeding animals in marine sediments. Q Rev Biol 62:
235–260.
62. Zobel CE, Felthons CB. 1938. Bacteria as food for certain
marine invertebrates. J Mar Res 1:312–327.
63. Pascal P, Dupuy C, Mallet C, Richard P, Niquil N.
2008. Bacterivory by benthic organisms in sediment:
quantification using ISN-enriched bacteria. J Exp Mar
Biol Ecol 355:18–26. doi: 10.1016/j.jembe.2007.11.012.
ER.
64. Smoot JC, Findlay RH. 2010a. Microbes as food for
sediment-ingesting detritivores: low-density particles
confer a nutritional advantage. Aquat Microb Ecol 59:
103–109.
65. Smoot JC, Findlay RH. 2010b. Caloric needs of detriti-
vorous gizzard shad Dorosoma cepedianum are met with
sediment bacterial and algal biomas.s. Aquat Biol 8:
105–114.
66. Helton RR, Wang K, Kan J, Powell DH, Wommack KE.
2012. Interannual dynamics of viriobenthos abundance
and morphological diversity in Chesapeake Bay sediments.
FEMS Microbiol Ecol 79:474–486. doi: 10.1111/j.1574-
6941.2011.01238.x.
67. Hewson I, Fuhrman J. 2003. Viriobenthos production and
virioplankton sorptive scavenging by suspended sediment
particles in coastal and pelagic waters. Microb Ecol 46:
337–347. doi: 10.1007/s00248-002-1041-0.
68. Mei M, Danovaro R. 2004. Virus production and life
strategies in aquatic sediments. Limnol Oceanogr 49:
459–470.
69. Siem-Jorgensen M, Glud RN, Middelboe M. 2008. Viral
dynamics in a coastal sediment: seasonal pattern, control-
ling factors and relations to the pelagic-benthic coupling.
Mar Biol Res 4:165–U20. doi: 10.1080/1745100080
1888718.
70. Fischer U, Wieltschnig C, Kirschner A, Velimirov B.
2003. Does virus-induced lysis contribute significantly
to bacterial mortality in the oxygenated sediment layer
of shallow oxbow lakes? Appl Environ Microbiol 69:
5281–5289. doi: 10.1128/AEM.69.9.5281-5289.2003.
71. Filippini M, Buesing N, Gessner MO. 2008. Temporal
dynamics of freshwater bacterio- and virioplankton along
a littoral-pelagic gradient. Freshw Biol 53:1114–1125.
doi: 10.1111/j.1365-2427.2007.01886.x.
72. Ram ASP, Sabart M, Latour D, Sime-Ngando T. 2009.
Low effect of viruses on bacteria in deep anoxic water
and sediment of a productive freshwater reservoir. Aquat
Microb Ecol 55:255–265. doi: 10.3354/ame01300.
73. Guenet B, Danger M, Abbadie L, Lacroix G. 2010. Pri-
ming effect: bridging the gap between terrestrial and
aquatic ecology. Ecology 91:2850–2861. doi: 10.1890/
09-1968.1.
74. McCallister SL, del Giorgio PA. 2012. Evidence for the
respiration of ancient terrestrial organic C in northern tem-
perate lakes and streams. Proc Natl Acad Sci USA 109:
16963–16968. doi: 10.1073/pnas.1207305109.
75. Wiegner TN, Kaplan LA, Ziegler SE, Findlay RH. 2015.
Consumption of terrestrial dissolved organic carbon by
stream microorganisms. Aquat Microb Ecol 75:225–237.
doi: 10.3354/ame01761.
76. Kuehn KA, Francoeur SN, Findlay RH, Neely RK.
2014. Priming in the microbial landscape: periphytic algal
stimulation of litter-associated microbial decomposers.
Ecology 95:749–762. doi: 10.1890/13-0430.1.
77. Falkowski PG, Fenchel T, Delong EF. 2008. The micro-
bial engines that drive Earth’s biogeochemical cycles. Sci-
ence 320:1034–1039. doi: 10.1126/science.1153213.
78. Schlesinger WH, Cole JJ, Finzi AC, Holland EA. 2011.
Introduction to coupled biogeochemical cycles. Front
Ecol Environ 9:5–8. doi: 10.1890/090235.
79. Berner RA. 1980. Early Diagenesis: A Theoretical Approach.
Princeton University Press, Princeton, NJ
80. Kristensen E. 2000. Organic matter diagenesis at the oxic/
anoxic interface in coastal marine sediments, with empha-
sis on the role of burrowing animals. Hydrobiologia 426:
1–24. doi: 10.1023/A:1003980226194.
81. McInerney M, Bryant M, Pfennig N. 1979. Anaerobic
bacterium that degrades fatty-acids in syntrophic associa-
tion with methanogens. Arch Microbiol 122:129–135.
doi: 10.1007/BF00411351.
82. Boetius A, Ravenschlag K, Schubert C, Rickert D, Wid-
del F, Gieseke A, Amann R, Jorgensen B, Witte U,
Pfannkuche O. 2000. A marine microbial consortium
apparently mediating anaerobic oxidation of methane.
Nature 407:623–626. doi: 10.1038/35036572.
83. Stams AJM, Plugge CM. 2009. Electron transfer in syntro-
phic communities of anaerobic bacteria and archaea. Nat
Rev Microbiol 7:568–577. doi: 10.1038/nrmicro2166.
84. Andersen F, Helder W. 1987. Comparison of oxygen
microgradients, oxygen flux rates and electron-transport
system activity in coastal marine-sediments. Mar Ecol
Prog Ser 37:259–264. doi: 10.3354/meps037259.
85. Hofman P, Dejong S, Wagenvoort E, Sandee A. 1991.
Apparent sediment diffusion-coefficients for oxygen and
oxygen-consumption rates measured with microelectrodes
and bell jars—applications to oxygen budgets in Estuarine
intertidal sediments (Oosterschelde, SW Netherlands).
Mar Ecol Prog Ser 69:261–272. doi: 10.3354/meps 069261.
86. Hedman JE, Gunnarsson JS, Samuelsson G, Gilbert F.
2011. Particle reworking and solute transport by the
sediment-living polychaetes Marenzelleria neglecta and
Hediste diversicolor. J Exp Mar Biol Ecol 407:294–301. doi:
10.1016/j.jembe.2011.06.026.
87. Riisgard H, Larsen P. 2005. Water pumping and analysis of
flow in burrowing zoobenthos: an overview. Aquat Ecol 39:
237–258. doi: 10.1007/s10452-004-1916-x.
88. Kristensen E, Hansen T, Delefosse M, Banta GT,
Quintana CO. 2011. Contrasting effects of the poly-
chaetes Marenzelleria viridis and Nereis diversicolor on
benthic metabolism and solute transport in sandy coastal
sediment. Mar Ecol Prog Ser 425:125–139. doi: 10.3354/
meps09007.
89. Martiny J, Bohannan B, Brown J, Colwell R, Fuhrman J,
Green J, Horner-Devine M, Kane M, Krumins J, Kuske
C, Morin P, Naeem S, Ovreas L, Reysenbach A, Smith
V, Staley J. 2006. Microbial biogeography: putting micro-
organisms on the map. Nat Rev Microbiol 4:102–112. doi:
10.1038/nrmicro1341.
90. Wilhelm L, Singer GA, Fasching C, Battin TJ, Besemer
K. 2013. Microbial biodiversity in glacier-fed streams.
ISME J 7:1651–1660. doi: 10.1038/ismej.2013.44.
91. Fierer N, Morse JL, Berthrong ST, Bernhardt ES, Jack-
son RB. 2007. Environmental controls on the landscape-
scale biogeography of stream bacterial communities. Ecol-
ogy 88:2162–2173. doi: 10.1890/06-1746.1.
92. Zinger L, Amaral-Zettler LA, Fuhrman JA, Horner-
Devine MC, Huse SM, Welch DBM, Martiny JBH, Sogin
M, Boetius A, Ramette A. 2011. Global patterns of bacte-
rial beta-diversity in seafloor and seawater ecosystems. PLoS
One 6:e24570. doi: 10.1371/journal.pone.0024570.
93. Fierer N, Jackson R. 2006. The diversity and biogeography
of soil bacterial communities. Proc Natl Acad Sci USA 103:
626–631. doi: 10.1073/pnas.0507535103.
4.2.1-18 ▪ AQUATIC ENVIRONMENTS
94. Fierer N, McCain CM, Meir P, Zimmermann M, Rapp
JM, Silman MR, Knight R. 2011. Microbes do not follow
the elevational diversity patterns of plants and animals.
Ecology 92:797–804. doi: 10.1890/10-1170.1.
95. Drenovsky RE, Steenwerth KL, Jackson LE, Scow KM.
2010. Land use and climatic factors structure regional pat-
terns in soil microbial communities. Glob Ecol Biogeogr 19:
27–39. doi: 10.1111/j.1466-8238.2009.00486.x.
96. Ghiglione J, Galand PE, Pommier T, Pedros-Alio
C, Maas EW, Bakker K, Bertilson S, Kirchman DL, Love-
joy C, Yager PL, Murray AE. 2012. Pole-to-pole biogeog-
raphy of surface and deep marine bacterial communities.
Proc Natl Acad Sci USA 109:17633–17638. doi:
10.1073/pnas.1208160109.
97. Brown MV, Lauro FM, DeMaere MZ, Muir L, Wilkins D,
Thomas T, Riddle MJ, Fuhrman JA, Andrews-Pfann-
koch C, Hoffman JM, McQuaid JB, Allen A, Rintoul
SR, Cavicchioli R. 2012. Global biogeography of SAR11
marine bacteria. Mol Syst Biol 8:595. doi: 10.1038/
msb.2012.28.
98. Federle T, Hulllar M, Livingston R, Meeter D, White D.
1983. Spatial-distribution of biochemical parameters indi-
cating biomass and community composition of microbial
assemblies in estuarine mud flat sediments. Appl Environ
Microbiol 45:58–63.
99. MacGregor B, Moser D, Baker B, Alm E, Maurer M,
Nealson K, Stahl D. 2001. Seasonal and spatial variability
in Lake Michigan sediment small-subunit rRNA concen-
trations. Appl Environ Microbiol 67:3908–3922. doi:
10.1128/AEM.67.9.3908-3922.2001.
100. Findlay RH, White DC. 1984. In situ determination of
metabolic activity in aquatic environments. Microbiol Sci
1:90–95.
101. Schonfeld J. 2012. History and development of methods in
recent benthic foraminiferal studies. J Micropalaeontol 31:
53–72. doi: 10.1144/0262-821X11-008.
102. Spivak AC, Vanni MJ, Mette EM. 2011. Moving on up:
can results from simple aquatic mesocosm experiments be
applied across broad spatial scales? Freshw Biol 56:
279–291. doi: 10.1111/j.1365-2427.2010.02495.x.
103. Petersen J, Cornwell J, Kemp W. 1999. Implicit scaling in
the design of experimental aquatic ecosystems. Oikos 85:
3–18. doi: 10.2307/3546786.
104. Findlay RH, Watling L. 1997. Prediction of benthic
impact for salmon net-pens based on the balance of benthic
oxygen supply and demand. Mar Ecol Prog Ser 155:
147–157.
105. Findlay RH, Yeates C, Hullar MAJ, Stahl DA, Kaplan
LA. 2008. Biome level biogeography of streambed micro-
biota. Appl Environ Microbiol 74:3014–3021.
106. Federle T, White D. 1982. Preservation of estuarine sedi-
ments for lipid analysis of biomass and community structure
of microbiota. Appl Environ Microbiol 44:1166–1169.
107. Kaplan LA, Wiegner TN, Newbold JD, Ostrom PH,
Gandhi H. 2008. Untangling the complex issue of dis-
solved organic carbon uptake: a stable isotope approach.
Freshw Biol 53:855–864. doi: 10.1111/j.1365-2427.
2007.01941.x.
108. King G. 2001. Radiotracer assays (S-35) of sulfate reduc-
tion rates in marine and freshwater sediments. Methods
Microbiol 30:489–500. doi: 10.1016/S0580-9517(01)
30059-4.
109. Revsbech N, Jorgensen B, Blackburn T, Cohen Y. 1983.
Microelectrode studies of the photosynthesis and O-2, H2
s, and pH profiles of a microbial mat. Limnol Oceanogr 28:
1062–1074.
110. Jorgensen B, Marais D. 1990. The diffusive boundary-layer
of sediments—oxygen microgradients over a microbial mat.
Limnol Oceanogr 35:1343–1355.
111. Beal EJ, Claire MW, House CH. 2011. High rates of anae-
robic methanotrophy at low sulfate concentrations with
implications for past and present methane levels. Geobiol-
ogy 9:131–139. doi: 10.1111/j.1472-4669.2010.00267.x.
112. Shoemaker JK, Schrag DP. 2010. Subsurface characteriza-
tion of methane production and oxidation from a
New Hampshire wetland. Geobiology 8:234–243. doi:
10.1111/j.1472-4669.2010.00239.x.
113. Roberts HM, Shiller AM. 2015. Determination of dis-
solved methane in natural waters using headspace analysis
with cavity ring-down spectroscopy. Anal Chim Acta
856:68–73. doi: 10.1016/j.aca.2014.10.058.
114. Hagstrom A, Larsson U, Horstedt P, Normark S. 1979.
Frequency of dividing cells, a new approach to the determi-
nation of bacterial-growth rates in aquatic environments.
Appl Environ Microbiol 37:805–812.
115. Blackburn N, Hagstrom A, Wikner J, Cuadros-Hansson
R, Bjornsen P. 1998. Rapid determination of bacterial
abundance, biovolume, morphology, and growth by neural
network-based image analysis. Appl Environ Microbiol 64:
3246–3255.
116. Zeder M, Ellrott A, Amann R. 2011. Automated sample
area definition for high-throughput microscopy. Cytometry
A 79A:306–310. doi: 10.1002/cyto.a.21034.
117. Jones J, Simon B. 1979. Measurement of electron-trans-
port system activity in freshwater benthic and planktonic
samples. J Appl Bacteriol 46:305–315. doi: 10.1111/j.
1365-2672.1979.tb00825.x.
118. Tabor P, Neihof R. 1982. Improved method for determina-
tion of respiring individual microorganisms in natural
waters. Appl Environ Microbiol 43:1249–1255.
119. Merlin G, Lissolo T, Morel V, Rossel D, Tarrdellas J.
1995. Precautions for routine use of INT-reductase activity
for measuring biological activities in soil and sediments.
Environ Toxicol Water Qual 10:185–192.
120. Nielsen J, Christensen D, Kloppenborg M, Nielsen P.
2003. Quantification of cell-specific substrate uptake by
probe-defined bacteria under in situ conditions by microau-
toradiography and fluorescence in situ hybridization. Envi-
ron Microbiol 5:202–211. doi: 10.1046/j.1462-2920.
2003.00402.x.
121. Brock T, Brock M. 1968. Measurement of steady-state
growth rates of a Thermophilic Alga directly in nature.
J Bacteriol 95:811.
122. Fuhrman J, Azam F. 1982. Thymidine incorporation as a
measure of heterotrophic bacterioplankton production in
marine surface waters—evaluation and field results. Mar
Biol 66:109–120. doi: 10.1007/BF00397184.
123. Tabor P, Neihof R. 1982. Improved micro-auto-radio-
graphic method to determine individual microorganisms
active in substrate uptake in natural-waters. Appl Environ
Microbiol 44:945–953.
124. Carman K. 1990. Radioactive labeling of a natural assem-
blage of marine sedimentary bacteria and microalgae for
trophic studies—an autoradiographic study. Microb Ecol
19:279–290. doi: 10.1007/BF02017172.
125. Nielsen J, de Muro M, Nielsen P. 2003. Evaluation
of the redox dye 5-cyano-2,3-tolyl-tetrazolium chloride
for activity studies by simultaneous use of microauto-
radiography and fluorescence in situ hybridization. Appl
Environ Microbiol 69:641–643. doi: 10.1128/AEM.69.1.
641-643.2003.
126. Fuhrman J, Azam F. 1980. Bacterioplankton secondary
production estimates for coastal waters of British Columbia,
Antarctica, and California. Appl Environ Microbiol 39:
1085–1095.
127. Moriarty D, Pollard P. 1981. DNA-synthesis as a measure
of bacterial productivity in seagrass sediments. Mar Ecol
Prog Ser 5:151–156. doi: 10.3354/meps005151.
 4.2.1. The Microbial Ecology of Benthic Environments ▪ 4.2.1-19
128. Kirchman D, Knees E, Hodson R. 1985. Leucine incorpo-
ration and its potential as a measure of protein synthesis by
bacteria in natural aquatic systems. Appl Environ Microbiol
49:599–607.
129. Findlay RH, Pollard PC, Moriarty DJW, White DC.
1985. Quantitative determination of microbial activity
and community nutritional status in estuarine sediments:
evidence of a disturbance artifact. Can J Microbiol 31:
494–498.
130. Dobbs F, Guckert J, Carman K. 1989. Comparison
of 3 techniques for administering radiolabeled sub-
strates to sediments for trophic studies—incorporation
by microbes. Microb Ecol 17:237–250. doi: 10.1007/
BF02012837.
131. Carman K, Dobbs F, Guckert J. 1989. Comparison of 3
techniques for administering radiolabeled substrates to sedi-
ments for trophic studies—uptake of label by harpacticoid
copepods. Mar Biol 102:119–125. doi: 10.1007/BF00
391329.
132. Wegener G, Bausch M, Holler T, Nguyen Manh Thang,
Mollar XP, Kellermann MY, Hinrichs K, Boetius A.
2012. Assessing sub-seafloor microbial activity by com-
bined stable isotope probing with deuterated water and
13C-bicarbonate. Environ Microbiol 14:1517–1527. doi:
10.1111/j.1462-2920.2012.02739.x.
133. Dortch Q, Roberts TL, Clayton JR, Ahmed SI. 1983.
RNA DNA ratios and DNA concentrations as indicators
of growth-rate and biomass in planktonic marine organ-
isms. Mar Ecol Prog Ser 13:61–71. doi: 10.3354/meps
013061.
134. Fernandes C, De Souza M, Nair S, Bharathi P. 2005.
Response of sedimentary nucleic acids to benthic disturb-
ance in the Central Indian Basin. Mar Georesour Geotech-
nol 23:289–297. doi: 10.1080/ 10641190500446607.
135. Hall EK, Maixner F, Franklin O, Daims H, Richter A,
Battin T. 2011. Linking microbial and ecosystem ecology
using ecological stoichiometry: a synthesis of conceptual
and empirical approaches. Ecosystems (NY) 14:261–273.
doi: 10.1007/s10021-010-9408-4.
136. Moran XAG, Ducklow HW, Erickson M. 2011. Single-
cell physiological structure and growth rates of heterotro-
phic bacteria in a temperate estuary (Waquoit Bay, Massa-
chusetts). Limnol Oceanogr 56:37–48. doi: 10.4319/lo.
2011.56.1.0037.
137. Sterner RW, Elser JJ. 2002. Ecological Stoichiometry: the
Biology of Elements from Molecules to the Biosphere. Princeton
University Press, Princeton, NJ.
138. Franklin O, Hall EK, Kaiser C, Battin TJ, Richter A.
2011. Optimization of biomass composition explains
microbial growth-stoichiometry relationships. Am Nat
177:E29–E42. doi: 10.1086/657684.
139. Kepner R, Pratt J. 1993. Effects of sediments on estimates
of bacterial density. Trans Am Microsc Soc 112:316–330.
doi: 10.2307/3226566.
140. Sunamura M, Maruyama A, Tsuji T, Kurane R. 2003.
Spectral imaging detection and counting of microbial cells
in marine sediment. J Microbiol Meth 53:57–65. doi:
10.1016/S0167-7012(02)00224-5.
141. Amalfitano S, Fazi S, Puddu A. 2009. Flow cytometric
analysis of benthic prokaryotes attached to sediment
particles. J Microbiol Meth 79:246–249. doi: 10.1016/j.
mimet.2009.09.005.
142. Bligh E, Dyer W. 1959. A rapid method of total lipid
extraction and purification. Can J Biochem Physiol 37:
911–917.
143. White D, Davis W, Nickles J, King J, Bobbie R. 1979.
Determination of the sedimentary microbial biomass by
extractable lipid phosphate. Oecologia 40:51–62. doi:
10.1007/BF00388810.
144. Findlay RH, King GM, Watling L. 1989. Efficacy of phos-
pholipid analysis in determining microbial biomass in sedi-
ments. Appl Environ Microbiol 55:2888–2893.
145. Frostegard A, Tunlid A, Baath E. 1991. Microbial biomass
measured as total lipid phosphate in soils of different
organic content. J Microbiol Meth 14:151–163. doi:
10.1016/0167-7012(91)90018-L.
146. Dell’Anno A, Bompadre S, Danovaro R. 2002. Quantifi-
cation, base composition, and fate of extracellular DNA in
marine sediments. Limnol Oceanogr 47:899–905.
147. Corinaldesi C, Barucca M, Luna GM, Dell’Anno A.
2011. Preservation, origin and genetic imprint of extra-
cellular DNA in permanently anoxic deep-sea sediments.
Mol Ecol 20:642–654. doi: 10.1111/j.1365-294X.2010.
04958.x.
148. Mumy KL, Findlay RH. 2004. Convenient determination
of DNA extraction efficiency using an external DNA
recovery standard and quantitative competitive PCR.
J Microbiol Methods 57:259–268.
149. Amann R, Fuchs BM. 2008. Single-cell identification
in microbial communities by improved fluorescence in
situ hybridization techniques. Nat Rev Microbiol 6:
339–348. doi: 10.1038/nrmicro1888.
150. Wagner M, Horn M, Daims H. 2003. Fluorescence in
situ hybridisation for the identification and characterisa-
tion of prokaryotes. Curr Opin Microbiol 6:302–309. doi:
10.1016/S1369-5274(03)00054-7.
151. Wagner M, Haider S. 2012. New trends in fluorescence in
situ hybridization for identification and functional analyses
of microbes. Curr Opin Biotechnol 23:96–102. doi:
10.1016/j.copbio.2011.10.010.
152. Findlay RH. 2004. Determination of microbial community
structure using phospholipid fatty acid profiles. In: Kowal-
chuk GA, De Bruijn FJ, Head IM, Akkermans ADL, Van
Elsas JD (eds), Molecular Microbial Ecology Manual, 2nd
ed. Kluwer Academic, Dordrecht, The Netherlands.
4.08:983–1004.
153. Hazen TC, Dubinsky EA, DeSantis TZ, Andersen GL,
Piceno YM, Singh N, Jansson JK, Probst A, Borglin
SE, Fortney JL, Stringfellow WT, Bill M, Conrad ME,
Tom LM, Chavarria KL, Alusi TR, Lamendella R, Joy-
ner DC, Spier C, Baelum J, Auer M, Zemla ML, Chakra-
borty R, Sonnenthal EL, D’haeseleer P, Holman HN,
Osman S, Lu Z, Van Nostrand JD, Deng Y, Zhou J,
Mason OU. 2010. Deep-sea oil plume enriches indigenous
oil-degrading bacteria. Science 330:204–208. doi:
10.1126/science.1195979.
154. Frostegard A, Tunlid A, Baath E. 2011. Use and misuse of
PLFA measurements in soils. Soil Biol Biochem 43:
1621–1625. doi: 10.1016/j.soilbio.2010.11.021.
155. Nichols P, Sahw P, Mancuso C, Frazmann P. 1993.
Analysis of archaeal phospholipid-derived diraether and
tetraether lipids by high-temperature capillary gas-chroma-
tography. J Microbiol Methods 18:1–9. doi: 10.1016/
0167-7012(93)90066-Q.
156. Hopmans E, Schouten S, Pancost R, van der Meer M,
Damste J. 2000. Analysis of intact tetraether lipids in
archaeal cell material and sediments by high performance
liquid chromatography/atmospheric pressure chemical
ionization mass spectrometry. Rapid Commun Mass Spec-
trom 14:585–589. doi: 10.1002/(SICI)1097-0231
(20000415)14:7<585:AID-RCM913>3.3.CO;2-E.
157. Guedes MJ, Pereira R, Duarte K, Rocha-Santos TAP,
Antunes SC, Goncalves F, Duarte AC, Freitas AC.
2011. Sterols and fatty acid biomarkers as indicators
of changes in soil microbial communities in a uranium
mine area. J Environ Sci Health A Tox Hazard Subst
Environ Eng 46:659–668. doi: 10.1080/10934529.2011.
563171.
4.2.1-20 ▪ AQUATIC ENVIRONMENTS
158. Allen MA, Neilan BA, Burns BP, Jahnke LL, Summons
RE. 2010. Lipid biomarkers in Hamelin Pool microbial
mats and stromatolites. Org Geochem 41:1207–1218. doi:
10.1016/j.orggeochem.2010.07.007.
159. Jungblut AD, Allen MA, Burns BP, Neilan BA. 2009.
Lipid biomarker analysis of cyanobacteria-dominated
microbial mats in meltwater ponds on the McMurdo
Ice Shelf, Antarctica. Org Geochem 40:258–269. doi:
10.1016/j.orggeochem.2008.10.002.
160. Su C, Lei L, Duan Y, Zhang K, Yang J. 2012. Culture-
independent methods for studying environmental microor-
ganisms: methods, application, and perspective. Appl
Microbiol Biotechnol 93:993–1003. doi: 10.1007/s00253-
011-3800-7.
161. Liu W, Marsh T, Cheng H, Forney L. 1997. Characteriza-
tion of microbial diversity by determining terminal restric-
tion fragment length polymorphisms of genes encoding
16S rRNA. Appl Environ Microbiol 63:4516–4522.
162. DeLong E. 2002. Microbial population genomics
and ecology. Curr Opin Microbiol 5:520–524. doi:
10.1016/S1369-5274(02)00353-3.
163. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D,
Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer
M, Gormley N, Gilbert JA, Smith G, Knight R. 2012.
Ultra-high-throughput microbial community analysis on
the Illumina HiSeq and MiSeq platforms. ISME J 6:
1621–1624. doi: 10.1038/ismej.2012.8.
164. Lee CK, Herbold CW, Polson SW, Wommack KE, Wil-
liamson SJ, McDonald IR, Cary SC. 2012. Groundtruth-
ing next-gen sequencing for microbial ecology-biases and
errors in community structure estimates from PCR ampli-
con pyrosequencing. PLoS One 7:e44224. doi: 10.1371/
journal.pone.0044224.
165. Gierga G, Voss B, Hess WR. 2012. Non-coding RNAs
in marine Synechococcus and their regulation under
environmentally relevant stress conditions. ISME J 6:
1544–1557. doi: 10.1038/ismej.2011.215.
166. Shi Y, Tyson GW, DeLong EF. 2009. Metatranscriptomics
reveals unique microbial small RNAs in the ocean’s water
column. Nature 459:266–269. doi: 10.1038/nature08055.
167. Kunin V, Engelbrektson A, Ochman H, Hugenholtz P.
2010. Wrinkles in the rare biosphere: pyrosequencing
errors can lead to artificial inflation of diversity estimates.
Environ Microbiol 12:118–123. doi: 10.1111/j.1462-
2920.2009.02051.x.
168. Huse SM, Welch DM, Morrison HG, Sogin ML. 2010.
Ironing out the wrinkles in the rare biosphere through
improved OTU clustering. Environ Microbiol 12:
1889–1898. doi: 10.1111/j.1462-2920.2010.02193.x.
169. Shakya M, Quince C, Campbell JH, Yang ZK, Schadt
CW, Podar M. 2013. Comparative metagenomic and
rRNA microbial diversity characterization using archaeal
and bacterial synthetic communities. Environ Microbiol
15:1882–1899. doi: 10.1111/1462-2920.12086.
170. O’Sullivan DM, Laver T, Temisak S, Redshaw N, Harris
KA, Foy CA, Studholme DJ, Huggett JF. 2014. Assessing
the accuracy of quantitative molecular microbial profiling.
Int J Mol Sci 15:21476–21491. doi: 10.3390/ijms15112
1476.
171. Mosher JJ, Bernberg EL, Shevchenko O, Kan J, Kaplan
LA. 2013. Efficacy of a 3rd generation high- throughput
sequencing platform for analyses of 16S rRNA genes
from environmental samples. J Microbiol Methods 95:
175–181. doi: 10.1016/j.mimet.2013.08.009.
172. Mosher JJ, Bowman B, Bernberg EL, Shevchenko O, Kan
J, Korlach J, Kaplan LA. 2014. Improved performance of
the PacSio SMRT technology for 16S rDNA sequencing.
J Microbiol Meth 104:59–60. doi: 10.1016/j.mimet.2014.
06.012.
173. Cory RM, Miller MP, McKnight DM, Guerard JJ, Miller
PL. 2010. Effect of instrument-specific response on the
analysis of fulvic acid fluorescence spectra. Limnol Oceanogr
Meth 8:67–78.
174. Simpson AJ, Simpson MJ, Soong R. 2012. Nuclear mag-
netic resonance spectroscopy and its key role in environ-
mental research. Environ Sci Technol 46:11488–11496.
doi: 10.1021/es302154w.
175. D’Andrilli J, Chanton JP, Glaser PH, Cooper WT. 2010.
Characterization of dissolved organic matter in northern
peatland soil porewaters by ultra high resolution mass spec-
trometry. Org Geochem 41:791–799. doi: 10.1016/j.
orggeochem.2010.05.009.
176. Zhong J, Sleighter RL, Salmon E, McKee GA, Hatcher
PG. 2011. Combining advanced NMR techniques with
ultrahigh resolution mass spectrometry: a new strategy for
molecular scale characterization of macromolecular compo-
nents of soil and sedimentary organic matter. Org Geochem
42:903–916. doi: 10.1016/j.orggeochem.2011. 04.007.
177. Wyatt ASJ, Waite AM, Humphries S. 2012. Stable iso-
tope analysis reveals community-level variation in fish
trophodynamics across a fringing coral reef. Coral Reefs
31:1029–1044. doi: 10.1007/s00338-012-0923-y.
178. Sanseverino AM, Bastviken D, Sundh I, Pickova J,
Enrich-Prast A. 2012. Methane carbon supports aquatic
food webs to the fish level. PLoS One 7:e42723. doi:
10.1371/journal.pone.0042723.
179. Vander Zanden M, Rasmussen J. 2001. Variation in delta
N-15 and delta C-13 trophic fractionation: implications for
aquatic food web studies. Limnol Oceanogr 46:2061–2066.
180. Macko S, Uhle M, Engel M, Andrusevich V. 1997. Stable
nitrogen isotope analysis of amino acid enantiomers by gas
chromatography combustion/isotope ratio mass spectrome-
try. Anal Chem 69:926–929. doi: 10.1021/ac960 956l.
181. Popp BN, Graham BS, Olson RJ, Hannides CCS,
Lott MJ, López-Ibarra GA, Galván-Magaña F, Fry B.
2007. Insight into the trophic ecology of yellowfin tuna,
Thunnus albacares, from compound-specific nitrogen iso-
tope analysis of proteinaceous amino acids, pp. 173–190.
In Dawson TE, Siegwolf RTW (eds), Stable Isotopes as
Indicators of Ecological Change, Elsevier, Terrestrial Ecol-
ogy Series, San Diego, CA.
182. Ghosh P. 2014. Compound specific isotope analysis of
amino acids in freshwater ecosystems: insights and applica-
tions. PhD dissertation. University of Alabama.
183. Boschker H, de Brouwer J, Cappenberg T. 1999. The con-
tribution of macrophyte-derived organic matter to microbial
biomass in salt-marsh sediments: stable carbon isotope anal-
ysis of microbial biomarkers. Limnol Oceanogr 44:309–319.
184. Andrews JH. 1991. Comparative Ecology of Microorganisms
and Macroorganisms. Springer, New York.
185. Könneke M, Bernhard A, de la Torre J, Walker C,
Waterbury J, Stahl D. 2005. Isolation of an autotrophic
ammonia-oxidizing marine archaeon. Nature 437:543–
546. doi: 10.1038/nature03911.
186. Brochier-Armanet C, Boussau B, Gribaldo S, Forterre P.
2008. Mesophilic crenarchaeota: proposal for a third arch-
aeal phylum, the Thaumarchaeota. Nat Rev Microbiol 6:
245–252. doi: 10.1038/nrmicro1852.
187. Jarrell KF, Walters AD, Bochiwal C, Borgia JM, Dickin-
son T, Chong JPJ. 2011. Major players on the microbial
stage: why archaea are important. Microbiology (UK) 157:
919–936. doi: 10.1099/mic.0.047837-0.
188. Meadows P, Anderson J. 1968. Micro-organisms attached
to marine sand grains. J Mar Biol Assoc UK 48:161.
189. Sieber JR, McInerney MJ, Gunsalus RP. 2012. Genomic
insights into syntrophy: the Paradigm for anaerobic meta-
bolic cooperation. Annu Rev Microbiol 66:429–452. doi:
10.1146/annurev-micro-090110-102844.
 Heterotrophic Planktonic Microbes: Virus,
Bacteria, Archaea, and Protozoa
JED A. FUHRMAN AND DAVID A. CARON
4.2.2
BACKGROUND AND HISTORICAL
DEVELOPMENT
Marine Microbial Ecology into the 1970s
Although heterotrophic marine microorganisms in the open
sea have been studied since the late 1800s and early 1900s
by pioneers like Fischer, Haeckel, and Calkins (1–3), the
contribution of these diminutive species to the food webs
of oceanographic systems was not fully recognized until
nearly a century later. Attention was drawn to the larger
and more conspicuous taxa of photosynthetic protists (e.g.,
diatoms and many dinoflagellates) early in the history of
biological oceanography. In contrast, little was known of
the abundances or activities of bacteria in the ocean until
the past 50 years, and most early studies of marine hetero-
trophic protists (the protozoa) focused on morphological
descriptions and natural history of larger species of these
taxa rather than their functional roles in marine food
webs. Improvements in microscopy methods used to observe
bacteria in seawater during the 1930s to 1950s indicated
that bacterial abundances were several orders of magni-
tude greater than previously believed. However, confusion
remained during regarding the ecological significance of
marine bacteria because these early abundance estimates
were often hundreds of times greater than counts made by
cultivation techniques (4).
During the 1960s and 1970s, however, the metabolic
activity of aquatic bacterial assemblages was demonstrated
by the uptake of radioactive organic compounds in marine
and freshwater samples (e.g., [5, 6] and others). These studies
demonstrated that organic compounds were readily turned
over by microorganisms in aquatic ecosystems, and that
microbial communities appeared to be quite dynamic. Dur-
ing that same period, abundant and diverse assemblages of
protozoa were demonstrated from a wide array of aquatic eco-
systems using more quantitative approaches for the collec-
tion, observation, and enumeration of eukaryotes (7–10).
Together these observations implied the presence of an
active and complex microbial community that might be
responsible for much of the metabolic activity in marine
ecosystems. This hypothesis stood in contrast to broad oce-
anographic models of the time that included the bacteria
only as a sink for nonliving organic matter on the sea floor
and largely ignored the potentially important roles of heter-
otrophic protists (11).
doi:10.1128/9781555818821.ch4.2.2
4.2.2-1
The Microbial Loop Revolution
Pomeroy’s prescient analysis, “The ocean’s food web, a chang-
ing paradigm” (12) had a significant impact toward trans-
forming the field. This publication is often cited by many
microbial ecologists as a turning point in our understanding
of the structure and function of marine ecosystems. In it, Pom-
eroy pointed to lines of evidence that were starting to emerge
showing that the smallest members of the food web, including
heterotrophic bacteria, cyanobacteria, and small protists
(algae and protozoa <20 µm), were probably responsible for
a large fraction of important system activities such as overall
respiration, photosynthesis, and organic matter turnover
(i.e., the ingestion of food particles by protozoa or the uptake
of dissolved substances by bacteria). Early studies of bacterial
utilization of dissolved organic matter (e.g., 1), and subse-
quent studies of community metabolism into the 1980s
(e.g., 14) indicated that this process accounted for a surpris-
ingly high percentage of total organic matter turnover. Sim-
ilarly, Fenchel and Jorgensen (15) pointed out that
approximately 10–30% of primary productivity might be
released as dissolved organic matter, which was then pre-
sumed to be taken up by bacteria who respired a portion
and passed the rest on to the next trophic level, composed
largely of protozoa. In this scenario, protozoa constituted a
mechanism by which bacterial biomass reentered the “classi-
cal” food web including the metazoan zooplankton and nek-
ton, and also served as a means of remineralizing some of the
bacterial biomass back to inorganic nutrients and carbon
dioxide for subsequent utilization by primary producers.
Although these conclusions were later found to be generally
correct, the concepts were a radical departure from the estab-
lished general biological oceanographic thinking of the time.
The introductions of epifluorescence microscopy of fluo-
rescently stained cells (16) and polycarbonate filters to posi-
tion the cells all in a single optical plane (17) were significant
technological advancements that allowed much more accu-
rate estimates of the total number of microorganisms present
in natural water samples. Using this method, it was deter-
mined that bacteria are typically present at abundances of 1
million cells per ml in near-surface seawater. This number is
surprisingly constant around the world, with most variation
falling within a factor of 10 worldwide. Appreciation of the
importance of medium to large protozoa (i.e., >20 µm) in
oceanic food webs was made possible in the late 1960s and
early 1970s largely through the pioneering work of Beers
4.2.2-2 ▪ AQUATIC ENVIRONMENTS
et al. (7, 18). However, as with marine bacteria, epifluores-
cence microscopy facilitated the observation of small single-
celled eukaryotes during the 1970s and 1980s and enabled
easy discrimination of protists without chloroplasts ( proto-
zoa) from those with chloroplasts (algae) based on the auto-
fluorescence of photosynthetic pigments (19–21). The
development and refinement of this approach for eukaryotic
microorganisms was instrumental in establishing the standing
stocks of small protozoa, which typically occur at abundances
of tens to thousands per ml in most marine ecosystems. The
development of microscopy approaches for larger protozoa
such as heterotrophic dinoflagellates and ciliates (10, 22)
has been equally important in documenting abundances of
these taxa that range up to tens per ml in much of the world’s
oceans.
Estimates of the overall biomass of various microbial
assemblages, and techniques to measure rate processes (e.g.,
rates of growth, substrate uptake, prey consumption) began
to appear in the late 1970s and 1980s, and the refinement
of these estimates and measurements continue to the present
day. Early attempts to measure bacterial growth rates in sea-
water involved “indirect” methods such as relating the fre-
quency of dividing cells to rates of division in cultured
strains (23). Isotope-uptake based approaches, specifically
the incorporation of radioactively labeled thymidine into
DNA (24, 25) and/or the incorporation of leucine into pro-
tein (26) have become the most commonly used methods
(see “Estimating ‘Bacterial’ Biomass and ‘Bacterial Produc-
tion’”). These methods have indicated that bacterial dou-
bling times can be on the order of one day in coastal
temperate waters. Combined with estimates of bacterial bio-
mass, these results led to the conclusion that bacteria must
be consuming a substantial proportion—on the order of
50%—of the total system primary productivity. A similar
conclusion was reached using direct estimation of microbial
respiration by careful measurements of oxygen concentration
changes (micro-Winkler method) in seawater that had been
prefiltered through 5 µm pore filters to remove animals and
many of the protists (14).
During this same period, small protozoa ( primarily flagel-
lates and ciliates) were gaining recognition as important
consumers of bacteria in the marine plankton and benthos
(27–29). An increasing volume of experimental work dem-
onstrated a dominant role for small, bacterivorous protozoa
as a mechanism for removing bacterial production and
repackaging bacteria into larger particles that might be con-
sumed by metazoan zooplankton. Also, it became recognized
around this time that a significant fraction of the phytoplank-
ton biomass and production was consumed directly by herbiv-
orous protozoa rather than by metazoan zooplankton such as
copepods (30, 31). Consequently, heterotrophic protists were
acknowledged as an important food source for a variety of
metazoan zooplankton, and numerous experimental studies
subsequently demonstrated this trophic connection (32, 33).
These observations were synthesized in a second bench-
mark paper (34). The latter publication marked the begin-
ning of the widespread recognition and use of the term
“microbial loop” in marine planktonic systems, a concept
that emphasizes the remarkable importance of the tiniest
organisms as well as dissolved organic matter as an intermedi-
ate in material and energy transfer in aquatic ecosystems. An
updated illustration of this basic concept is shown in Fig. 1.
Definitions and Concepts
Abundance: the number of individuals in a sample or a
population.
Algae: protists that exhibit phototrophic nutrition. Like
protozoa, algae span a wide size range (<1 to >200
µm), and have generally been referred to as phototrophic
pico-, nano-, or microplankton.
Amensalism: interaction where members of a species
inflicts harm to another species without any costs or ben-
efits received by the other.
Autotroph: an organism that uses carbon dioxide as its
source of structural carbon.
Biomass: the mass of living organisms within a population,
community, or ecosystem.
Chemolithotroph: an organism that uses reduced inorganic
molecules as its energy source.
Chemoorganotroph: an organism that uses reduced organic
carbon as its energy source.
Commensalism: interaction between species where one
benefits from the other but the other is not affected.
Competition: species-species interactions that have a nega-
tive effect on both species.
Cyanobacteria: Prokaryotic photosynthetic organisms that
contain chlorophyll a and generate oxygen during photo-
synthesis. The free-living ancestors of primary chloroplasts.
Exploitation: species-species interactions that have a nega-
tive effect on one species while benefiting the other—
can include both parasitism and predation.
Heterotroph: an organism that uses preformed organic car-
bon as its source of structural carbon, also heterotrophic
or heterotrophy when applied to metabolism.
Mixotroph: any of a number of types of organisms that com-
bine (in one organism) multiple metabolic types as
described above. For example, a protist that consumes bac-
teria as prey (heterotroph) but also contains functioning
chloroplasts (phototroph) will often be referred to as a
mixotroph. Similarly an archaeon that oxidizes ammonia
for energy (chemolithotroph) but uses amino acidsto build
proteins (heterotroph) could be considered a mixotroph.
Mutualism: interaction between species where both benefit
from each other.
Phototroph: an organism that uses light as its energy source
for production of ATP (or to produce proton gradients in
the case of rhodoposin-based phototropy) and some-
times also reducing power from water (in cyanobacteria).
Phytoplankton: the photoautotrophic component of the
plankton including cyanobacteria and a large number
of eukaryotic phyla that contain chloroplasts.
Protists: eukaryotic species that can exist as a single cell
other than a spore, gamete, or zygote (although there
are many that form colonies).
Protozoa: protists that exhibit heterotrophic nutrition. Pro-
tozoa span a wide size range (≈2 to >200 µm), and have
generally been referred to as nano- or microzooplankton.
Relative abundance (and the related term, evenness): the
contribution of each species or operational taxonomic
unit to a community.
Species diversity: a complex concept composed of species
richness and relative abundance.
Species richness: the number of different species or opera-
tional taxonomic units present in a sample, habitat, or
environment.
Stoichiometry: Studies that involve calculation of the rela-
tive quantities of elements or compounds, for example,
C:N:P ratios.
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-3

FIGURE 1 An early vision of the “microbial loop” and its connections to the classical grazing food chain via dissolved organic matter (DOM)
flux and particulate trophic transfer, with viruses included as a side loop. Modified from (34). Large gray arrows indicate the flow of organic
carbon and energy into higher trophic levels of the food web, with recognition of the important roles for heterotrophic microbes (bacteria
and protozoa) in this process. Large stippled arrows indicate the production of DOM via excretion and trophic interactions (not all groups
are represented). Thin, dotted arrows indicate mineralization of major nutrients contained in organic matter respired by consumers. White
arrows indicate bacteria lysis by viruses and DOM released by that process. doi:10.1128/9781555818821.ch4.2.2.f1

Symbiosis: the intimate living together of two kinds of
organisms, especially if such an association is of mutual
advantage—too vague to be of use in quantitative
descriptions of population interactions but very useful
in indicating a close association among organisms.
Syntrophy: a metabolic mutualism where one species uses
the waste product produced by the other, and in so doing
allows both metabolic pathways to be energetically
feasible.
Zooplankton: Planktonic eukaryotes that consume other
plankton. Includes single-celled organisms ( protozoa
or protists) and metazoans, and some that are planktonic
only as larvae.
Estimating “Bacterial” Biomass and “Bacterial
Production”: Definitions and Methods
Aquatic microbiologists tend to use the term “bacteria” with a
lowercase “b” to describe organisms that appear to be prokary-
otic by microscopy—that is, organisms with no membrane-
bound nucleus. They include members of the taxonomic
domains Bacteria and Archaea (see “Bacteria and Archaea”).
Organisms within and between these domains differ in many
biochemical and genetic aspects, but they tend to look similar
by traditional epifluorescence microscopy. Special methods,
such as different versions of fluorescence in situ hybridization
(FISH) are required to distinguish individual members of
these domains microscopically (35–37).
The term “bacterial production” here refers to heterotro-
phic production of biomass by bacteria. It is meant to include
production of nonphotosynthetic bacterial biomass based on
the heterotrophic consumption of preformed organic matter
(i.e., organic matter in various forms that has been produced
primarily by phytoplankton).
Bacterial biomass is usually determined by converting
direct counts of bacteria using an estimate of the amount of
carbon per cell. Direct counts are most commonly done by
epifluorescence microscopy with stains such as acridine
orange, 40 ,6-diamidino-2-phenylindole or SYBR green I
(17, 38, 39). Special procedures are usually applied for sedi-
ment samples and samples containing large numbers of bacte-
ria attached to particles (40). SYBR green I also permits direct
visualization and counts of viruses in the same preparation.
Increasingly, direct bacterial counts in seawater samples
have been performed by flow cytometry of fluorochrome-
stained cells (41, 42), a method that allows separate counts
of cyanobacteria, such as Synechococcus and Prochlorococcus,
which have unique fluorescent signatures due to their photo-
synthetic pigments, and which can sometimes make up a sub-
stantial fraction of the total number of bacteria (43). Flow
cytometry is rapid and has a statistical advantage in that it typ-
ically observes thousands of prokaryotic and minute photo-
synthetic eukaryotic cells rather than the hundreds counted
microscopically; drawbacks include the cost of the instrument
and the fact that cells attached to each other or to other par-
ticles are counted as one. Bacterial carbon per cell has been
estimated in a variety of ways, most commonly from a deter-
mination of cell volume and carbon density per unit volume.
These numbers are difficult to obtain accurately for native
marine bacteria, which are very small, typically 0.5 μm in
4.2.2-4 ▪ AQUATIC ENVIRONMENTS

diameter (range is about 0.2–1 μm for free-living unicells).
Published estimates of bacterial carbon per cell vary widely
and probably constitute the greatest uncertainty with estimat-
ing bacterial biomass in natural samples. Typical estimates of
the carbon content of a bacterium range from 7 to 50 fg C
(1 fg is 10−15 g), with most open ocean estimates near 10–
20 fg C per cell and coastal ones about double that (44).
Thus, in a typical mesotrophic ocean environment with
109 bacteria per liter, and an average per cell C content
of 15 fg, bacterial biomass would be approximately 109 ×
1.5 × 10−14 = 1.5 × 10−5 g C per liter, or 15 μg C per liter.
Bacterial production is most often measured by incorpora-
tion of tritiated thymidine into DNA (24, 45) or tritiated leu-
cine into protein (26, 46). Thymidine and leucine are
intracellular precursors of DNA and protein, respectively, so
incorporation of these precursors can be used to estimate
the total rates of synthesis of the macromolecules. DNA is syn-
thesized for cell division and protein synthesized roughly in
proportion to total biomass, so measuring their rates of synthe-
sis is presumed to track production. Both methods have been
calibrated on the basis of theoretical considerations as well as
purely empirical approaches, and both methods yield similar
results (44). Leucine has a lower detection limit, so it is pre-
ferred for slower rates. The results of thymidine incorporation
are most often presented as cells produced per liter per hour,
which can be converted to a carbon production rate via esti-
mates of carbon per cell as noted. Leucine incorporation may
be used to calculate cell production as well as biomass produc-
tion directly, because protein is a major biomass constituent.
Geographic and Temporal Distributions of
Microbial Biomass and Activity
Bacterial abundances by epifluorescence microscopy show
that bacteria are present in most marine euphotic zone
environments at approximately 106 cells per ml. Samples
from around the world rarely vary by more than threefold
from this typical value (i.e., rarely <3 × 105 or >3 × 106),
which is extraordinary compared to phytoplankton and zoo-
plankton, which may vary by several orders of magnitude over
the same spatial scales. However, despite this remarkable gen-
eral predictability, there is significant variation across both
space and time. More nutrient-rich, eutrophic environments
tend to have more bacteria (sometimes >107 per ml; 47), and
oligotrophic open ocean environments have less (summar-
ized below). Although bacterial assemblages in warm temper-
ate coastal waters may have doubling times as short as 1 day,
this is at the rapid end of the spectrum of in situ growth rates.
Bacterial assemblages in the open sea, especially in oligotro-
phic environments, have average generation times typically
of a week or perhaps more (see Table 1). These abundances
generally apply to the euphotic zone, and bacteria in colder,
darker waters have substantially lower abundances and slower
growth rates than those of surface waters (48). Benthic bacte-
ria also exhibit fairly constant abundances across wide geo-
graphic ranges, but benthic bacteria occur at much higher
densities because of the organic-rich and particle-laden
nature of the environment (an average of ≈109 per ml fluid
volume is typical; 49). This value is three orders of magnitude
greater than abundances in the water column, implying that
the abundances of benthic bacterial assemblages are regulated
by a different set of parameters than planktonic assemblages.
Bacterial biomass and productivity vary temporally on a
number of scales ranging from diel (50) to seasonal (51) or
interannual (52). On time scales of hours, bacterial abun-
dance and production have been shown to often peak in
the middle of the day and be low in the middle of the night
(50, 53–55). This pattern has been interpreted as a tight cou-
pling between the production of labile organic compounds
via photosynthesis and bacterial growth on one hand, and

TABLE 1 Bacterioplankton properties in relation to phytoplankton in the open sea, as compiled by Ducklow (44)
Property N Atlantica Eq Pac-Sprb Eq Pac-Fallc Sub N Pacd Arabiane Hawaiif Bermudag Ross Seah
Euphotic zone m 50 120 120 80 74 175 140 45
Biomass (mg C m−2)
Bacteria 1000 1200 1,467 1,142 1,448 1,500 1,317 217
Phytoplankton 4,500 1,700 1,940 1,274 1,248 447 573 11450
B:P 0.2 0.7 0.75 0.9 1.2 3.6 2.7 0.02
Production (mg C m−2 d−1)
Bacteria 275 285 176 56 257 Nd 70 5.5
Phytoplankton 1,083 1,083 1,548 629 1,165 486i 465 1248
B:P 0.25 0.26 0.11 0.09 0.22 Nd 0.18 0.04
Growth rates (d−1)
Bacteria 0.3 0.13 0.12 0.05 0.18 Nd 0.05 0.25
Phytoplankton 0.3 0.64 0.8 0.5 0.93 1.1 0.81 0.11
B:P 1 0.2 0.15 0.1 0.19 Nd 0.06 2.3
Notes: All bacterial biomass estimates based on 20 fg C per cell. Data may overestimate heterotrophic bacterial biomass as a consequence of lower C per cell or
interference by Prochlorococcus and Archaea. Production estimated from 3,000 g C per mole leucine incorporated.
a
Eastern North Atlantic spring phytoplankton bloom, 47 N 20 W, May 1989, n = 13.
b
Equatorial Pacific, 0 N 140 W, March–April 1992, n = 8.
c
Equatorial Pacific 0 N, 140 W, September–October 1992, n = 19.
d
Subarctic North Pacific, 45 N.
e
Northwest Arabian Sea 10–20 N, 165 E, January–December 1995, n = 21.
f
Hawaii Ocean Time Series (HOT); 1995–1997; n = 21 (http://hahana.soest.hawaii.edu/hot/hot_jgofs.html).
g
Bermuda Atlantic Time Series (BATS); 1991–1998, n = 106 paired comparisons. The ratios are means of the ratios, not ratios of the means. BP calculated
from thymidine incorporation (1.6 × 1018 cells per mole incorporated).
h
Ross Sea, Antarctica, 76 S 180 W, 1994–1997.
i
1989–1996; n = 64.
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-5
bacterial mortality via grazers or viral lysis on the other hand.
It is consistent with some measurements of extremely rapid
turnover, sometimes several times per hour in rich coastal
waters, of labile dissolved organic compounds such as dis-
solved free amino acids (56) and might also indicate greater
predation pressure by protozoa during the night. On longer
time scales of weeks to months, bacteria show distinct sea-
sonal patterns. For example, in temperate coastal waters, bac-
terial biomass and production increase considerably in
summer months compared with winter. However, bacteria
do not typically show a significant increase during early
“spring” blooms in temperate waters when water is still very
cold (51). It has been hypothesized that this phenomenon
is the result of the suppression of the rate of substrate uptake
by temperate bacterial assemblages at low temperature (57).
However, while temperature probably has the effect of setting
a limit on maximal growth rates (as for phytoplankton [58]
and protozoa [59]), temperature alone does not appear to be
the main factor controlling growth of marine bacteria under
most circumstances (see “Light, Temperature, and Pressure”).
It would be overly simplistic to think that all the bacteria
and archaea in a sample or habitat have the same level of
activity per cell, but it is also easy to think of measured activ-
ities as characteristic of all members of a microbial assem-
blage. So the question arises: are most of the cells active at
a similar level, or are some hyperactive while others are com-
pletely dead or moribund? This question has been addressed
several ways, including microautoradiography, selective stain-
ing, “direct viable counts” (where nutrients are added to see
what part of the community grows), and in situ hybridization
(e.g., 60). Based on these contrasting approaches, it appears
that a continuum of activity exists within bacterial assemb-
lages from truly dead (cannot be revived) to extremely active.
A reasonable interpretation of the existing data is that under
typical conditions, a small percentage of the marine bacterial
cells, perhaps 10–20%, are generally inactive or dead; the plu-
rality or majority of cells, perhaps 25–75%, are intact and
have some moderate level of activity; and a small percentage,
perhaps 5–20%, are highly active. It is useful to consider this
spectrum, conceptually and numerically, when modeling
microbial processes.
Comparisons of bacterial and phytoplankton biomass
within planktonic ecosystems show that these are positively
correlated across broad scales. Analyses of marine and fresh-
water samples from several studies (61, 62) have shown that
bacterial abundance increases with chlorophyll concentra-
tion, at least at the level of a log-log relationship. Similarly,
bacterial abundances and the abundances of small protozoa
correlate over broad spatial and temporal scales (63). These
relationships are sensible in that on the largest scale, primary
production is the source of organic material that fuels hetero-
trophic bacterial activity, and bacteria constitute the prey of
many small protozoa. Individual data sets also have sometimes
shown strong correlations between bacterial abundance and
chlorophyll (e.g., 45), but variability in this relationship
over short temporal or spatial scales is to be expected. It would
presumably be a consequence of rapid, short-term changes in
the rate of substrate supply as well as the normal, oscillatory
nature of predator-prey relationships between bacteria and
their consumers.
Interestingly, the extrapolation of the positive log-log rela-
tionship between bacterial and phytoplankton biomass
to environments with very low chlorophyll concentrations
(e.g., ultra-oligotrophic oceans) indicates that bacterial bio-
mass may exceed phytoplankton biomass in these situations.
While this conclusion is obvious for the deep sea, where there
is no photosynthesis, this situation is also often true in oligo-
trophic surface waters. Measurements made in oligotrophic
waters bear out the high bacterial contribution to total
biomass (64, 65). Moreover, Cho and Azam (66) confirmed
a linear relationship between the log of chlorophyll and log
of bacterial abundance, but only at chlorophyll concen-
trations above approximately 0.5 μg per liter. Below that
concentration, bacterial abundance did not correlate signifi-
cantly with chlorophyll. It should be noted that subsequent
analyses have revealed that early epifluorescence measure-
ments of bacterial biomass included the common cyanobacte-
rium Prochlorococcus, which can make up to 20% of total
bacterial numbers (67). Nonetheless, heterotrophic bacterial
biomass is a major fraction of the living biomass of all plank-
tonic ecosystems.
The geographical and temporal distributions of marine
protozoa are much more varied than those of the bacteria.
As an all-inclusive group, protozoa generally occur in plank-
tonic ecosystems at abundances ranging from 10 s to 1,000 s
per ml. Abundances in benthic ecosystems can be one to
three orders of magnitude higher, commensurate with the
higher abundances of bacteria in those ecosystems. However,
it is important to remember that like the term “‘bacteria,” the
term “protozoa” is a rather artificial conglomeration of evolu-
tionarily and ecologically divergent taxa (see “The Changing
and Complex World of Eukaryote Phylogeny”). Thus, the
abundances of specific lineages of bacteria or protozoa may
show spatial (or temporal) variability that is considerably
greater than the variability characteristic of these overarching
groupings.
The Changing and Complex World of
Eukaryote Phylogeny
Not that long ago, textbooks still divided eukaryotic organ-
isms into four major kingdoms (Animalia, Plantae, Fungi,
and Protista) while prokaryotic organisms were placed into
a single kingdom, the Monera (68). Within this scheme, pro-
tists (eukaryotic organisms that can exist as single cells) were
divided into two subkingdoms (algae and protozoa) based on
their basic nutritional mode, a carryover from the historical
distinction between single cells with “animal-like” or “plant-
like” nutrition. This distinction presupposed a basic evolu-
tionary divergence among protists into species that retained
a heterotrophic, phagocytotic mode of life ( protozoa) and
those that abandoned phagocytosis for a photosynthetic
mode of life (algae). Moreover, the presence/absence of chlor-
oplasts was a feature that could be easily distinguished by early
microscopists.
The five-kingdom classification system of Whittaker was
recognized as an improvement over previous classification
schemes, but it posed a number of problems relating to
protists. For example, the distinction between single-celled
and multicellular eukaryotes was somewhat arbitrary. More
important, the division of protists based on whether they
were heterotrophic or photosynthetic was clearly not an
appropriate feature if the classification was to recapitulate
evolutionary relationships. We now know that chloroplast
acquisition and loss has occurred several times in the bio-
logical history of our planet (69), giving rise to some closely
related protistan taxa that differ largely in the presence or
absence of a chloroplast. Further complicating the matter,
within many protistan lineages there are species that possess
chloroplasts and carry out photosynthesis ( phototrophy)
but also possess the ability to ingest and digest prey (hetero-
trophy; 70–72). Some heterotrophic protists even ingest
4.2.2-6 ▪ AQUATIC ENVIRONMENTS
phytoplankton prey and retain the chloroplasts of their prey in
a functional state for a limited amount of time (kleptidoplas-
tidy;73). Various forms and degrees of mixotrophy (mixed
phototrophic and heterotrophic nutrition) are common
among a number of algal/protozoan lineages (74–77). Under
Whittaker’s scheme, phytoplankton ecologists studying a lin-
eage of microalgae might have had little familiarity with
closely related heterotrophic species, while protozoologists
studying a particular protozoan group might have known little
about closely related photosynthetic species.
One might expect, given these caveats, that the terms
“algae” and “protozoa” are no longer used. In fact, the term
“protozoa” is still commonly used (especially by ecologists)
to recognize those eukaryotic species that exist as single cells,
and whose nutrition is dependent on the uptake of preformed
organic substances ( primarily via prey ingestion), while pro-
tists possessing chloroplasts are still commonly called “algae.”
Similarly, although the term “protist” has been abandoned as
a kingdom designation, it is still widely employed to describe
eukaryotic species that are capable of existence as single cells
(i.e., algae and protozoa). The term “phagotrophic protist”
has also gained popularity in recent years because it recognizes
that many protistan species are capable of phagocytosis even
though they may also possess their own chloroplasts and thus
are technically “algae.”
Despite the shortcomings of Whittaker’s scheme, it domi-
nated the hierarchical organization of life for approximately a
quarter century. During the past few decades, however, this
system has given way to a new organizational scheme that rec-
ognizes three domains of life (Archaea, Bacteria, Eukarya [or
Eucarya]; 78; Fig. 2, upper panel), and is based on what is pres-
ently believed to reflect a more realistic view of the evolution-
ary distances that have developed between organisms in the
≈4 billion years that life has existed on our planet. Within
the Eukarya of Woese’s scheme, hypotheses regarding the
phylogeny of “protists” have changed continuously and rap-
idly during the past two decades, reflecting new insights
into eukaryote evolution provided largely by DNA sequence
information (Fig. 2, lower panel; from [79]).
The former protistan phyla of Whittaker’s system have now
been dispersed among candidate “supergroups” within the
domain Eukarya to better reflect hypothesized phylogenetic
relationships. For example, the dinoflagellates (which encom-
pass phototrophic, heterotrophic, and mixotrophic species)
now form a single group and have been placed together with
the ciliates and apicomplexans (sporozoans) in the monophy-
letic Alveolata (Fig. 2, lower panel). On the other hand,
eukaryotic, heterotrophic, single-celled species falling within
the general description “protozoa” are now widely distributed
among a number of protistan lineages. In short, nutrition has
been demoted as a phylogenetic character, and other charac-
ters ( presumably more indicative of evolutionary relatedness)
have ascended to address some long-standing contradictions,
although the debate over the relationships among some line-
ages is still very active at the present time; 79).
Individual protozoan cells range in size from less than 2 µm
to greater than 1 cm in diameter (>4 orders of magnitude)
(80, 81) with some colonial radiolaria forming cylindrical
gelatinous structures a centimeter in diameter and more
than a meter in length (82, 83). Because they constitute
such a large size range of organisms, protozoa are often divided
into size classes that very crudely correlate with their general
nutritional preferences. A commonly used convention is that
of Sieburth et al. (84), which groups planktonic microorgan-
isms into order-of-magnitude size classes (0.2–2.0 µm =
picoplankton; 2.0–20 µm = nanoplankton; 20–200 µm =
microplankton; 0.2–2 mm = mesoplankton). Most protozoa
fall into the nanoplankton or microplankton size classes.
Modeling microbial trophodynamics using this convention
assumes that protozoa in one size category generally consume
prey one order of magnitude smaller in size (34, 85).
Although this approach misses much of the detail and diver-
sity of the trophic activities of individual protozoan taxa, it is
a necessary, practical compromise for examining community-
scale flows of energy and elements. It also provides a useful
mechanism for summarizing and comparing the abundances
and biomasses of protozoa from different environments and
to other microbial assemblages. Protozoan abundance, sum-
marized in this way, has been shown to contribute signifi-
cantly to the living biomass of planktonic ecosystems
throughout the world ocean (Fig. 3).
Estimates of protozoan biomass, such as those depicted in
Fig. 3, typically do not include the contribution of mixotro-
phic phytoflagellates to heterotrophy. There is presently no
easy way to determine the abundances of small phagotrophic
phytoflagellates in natural samples, so these species are typi-
cally counted as phytoplankton unless specific methods are
employed to identify the algae as consumers, such as the use
of fluorescently labeled particles (87–90) or through the
examination of food vacuole contents (91). On average
these species appear to constitute a modest percentage of
the phytoplankton assemblage (typically <25%), although
they may at times dominate the phototroph assemblages of
natural plankton communities. It is important to recognize
that their inclusion as functional heterotrophs, rather than
phototrophs, could significantly shift the relative contribu-
tions of phototrophic and heterotrophic microbial biomass
to total biomass within microbial assemblages, and the flow
of energy within plankton communities (92).
Heterotrophic protists that harbor photosynthetic pro-
tists, or their chloroplasts, within their cytoplasm constitute
another complexity for estimating the contribution of pro-
tozoa to total microbial biomass. When bulk water samples
are analyzed, the contribution of chlorophyll contained
within those protozoa is generally assumed to come from
free-living phytoplankton. However, studies have shown
that chloroplast-bearing ciliates can contribute up to half
the total biomass of planktonic ciliates in ecosystems, and
chloroplast-retaining ciliates can sporadically dominate the
chlorophyll and primary production of some planktonic eco-
systems (93–96). The environmental conditions promoting
the success of these ciliates are poorly known. Similarly,
many species of planktonic foraminifera, polycystine radio-
laria, and acantharia harbor large numbers (thousands per
protozoan) of endosymbiotic algae within their cytoplasm
(97, 98). Caron et al. (99) have demonstrated that primary
productivity within these species can contribute significantly
to total primary productivity in oceanic ecosystems and can be
very important locally in the convergences of Langmuir circu-
lation cells (100).
BACTERIA AND ARCHAEA
“Culturable” versus “Nonculturable” Cells
Most conventional cultivation methods can grow only 1% or
less of the bacteria that can be visualized by direct microscopy
techniques (e.g., 4). This is true even though most can be
shown to be active by techniques such as microautoradiogra-
phy (25). These readily cultivable organisms appear to repre-
sent a group of fast-growing so-called weeds that are adapted
to take advantage of rapid growth in rare, organically enriched
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-7

FIGURE 2 The three domains of life (upper left) as proposed by Woese et al. (78), and a recent overview of modifications that have been
proposed by Adl et al. (79) to higher-level phylogentic groups within the eukaryotic component of the tree (lower right). Domains figure from
Woese et al. (78), eukaryotic tree figure from Adl et al. (79). doi:10.1128/9781555818821.ch4.2.2.f2

environments. This strategy contrasts with the numerically
dominant bacteria that are adapted specifically for growth
in the dilute nutrient conditions that characterize the vast
majority of the volume of the water column.
The most common taxa readily cultured from seawater
with standard nutrient broth media include the gamma pro-
teobacterial genera Vibrio, Alteromonas, Pseudoalteromonas,
Marinomonas, Oceanospirillum, Shewanella (usually isolated
from surfaces such as shellfish and sediments), the alpha pro-
teobacterial genera Roseobacter, Sphingomonas, members of the
family Flavobacteriaceae, and Planctomycetes, as summarized
in Giovannoni and Rappé (101) and Fuhrman and Hagstrom
(102). The cyanobacteria Synechococcus and Prochlorococcus
are also now readily culturable, but on low-nutrient inorganic
media targeting photosynthetic forms, as opposed to organic
media used to cultivate the others listed above.
Because of the low percentage of marine bacteria that can
be grown in standard media, organisms that until recently
were called “nonculturable” make up the large majority of
bacteria in the plankton. Only during the past ∼20–25 years
have molecular biological methods based on 16S rRNA gene
sequences been available to identify these organisms, and
these powerful techniques have opened up a large area for
exploration (see next section). Similar but more recent stud-
ies use 18S rRNA sequences for characterizing protistan
diversity, as will be noted below.
Molecular Phylogeny and Metagenomics:
Field Applications
Modern phylogeny of microorganisms is based primarily
on genetic sequences, the most well-studied gene being the
4.2.2-8 ▪ AQUATIC ENVIRONMENTS

FIGURE 3 (a, b) Plankton biomass in the Arabian Sea during the 1995 southwest monsoon (a) and intermonsoon period (b). Areas of the
boxes indicate the relative magnitudes of the biomass in each category. Categories within the dashed boxes in a, b, c are composed of protozoa.
Arrows indicate the direction of energy/material flow in the food web, thicker arrows depicting greater flow. Redrawn from (86).
(c) Depth-integrated biomass (mg/m2) in the upper 100 m of the Sargasso Sea near Bermuda, and in the upper 200 m of the equatorial Pacific
at 175°E. The width of the bars indicates the biomass in each size category. Heterotrophs have been separated by size class, while phytoplankton
have not. Size classes delineated by the dotted box are comprised of protozoa. Redrawn from (65). doi:10.1128/9781555818821.ch4.2.2.f3

small subunit ribosomal RNA gene (16S rRNA in Bacteria
and Archaea, and its larger homolog 18S rRNA in eukar-
yotes). This molecule is strongly conserved over evolutionary
time, so this single molecule has been used for constructing
phylogenetic trees of all living organisms (http://tolweb.org/
tree/phylogeny.html). Analysis of 16S/18S rRNA gene
sequences has been used to evaluate deep evolutionary rela-
tionships among organisms and was instrumental in point-
ing out that Archaea, Bacteria, and Eukarya should be
considered different Domains of equivalent phylogenetic
rank, above kingdoms (78). However, there are sufficient
differences in 16S/18S rRNA gene sequences to demonstrate
relationships (or make distinctions) at the genus or sometimes
species level.
The first phylogenetic studies based on 16S/18S rRNA
genes used sequences derived from cultures. However, one
does not need cultures to obtain rRNA gene sequences (or
any other sequences, for that matter). An idea developed in
the lab of Norman Pace in the mid-1980s involved extraction
of DNA directly from natural samples, and then cloning and
sequencing of the DNA as a means of assaying the microbes
present in the samples (103, 104). The original protocols
called for cloning by creating what are called “phage libraries”
from the natural DNA, but since 1986, PCR has been applied
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-9
extensively for cloning and related studies. The target
sequence can be almost instantly “identified” to its closest
phylogenetic neighbor by what have come to be standard
online sequence comparisons.
Beyond the study of targeted genes (like the 16S rRNA
gene), shotgun metagenomic studies have examined the
entire genetic repertoire of the microbes in a given sample.
The metagenome is the collective genome of all organisms
in the sample. Initially these studies extracted DNA from
all organisms in a sample (usually prefiltered through a filter
approximately 1 µm to remove most organisms larger than
bacteria), sheared it to produce fragments, and cloned
them into standard vectors, either as small (thousands of
bases) or large (to hundreds of thousands of bases) inserts,
that is, fragments of DNA from the environment now
cloned into the vectors in a form suitable for sequencing.
The best known early marine study of the former type is
the Global Ocean Survey, with initial results published by
Venter et al. (105), which generated more than a billion
bases of DNA sequence and reported 1.2 million previously
unsequenced genes, estimated to come from at least 1,800
different genomic species cumulatively in the many samples
they analyzed.
With the advent of next-generation sequencing that gen-
erates millions or more sequences in a run (known by a variety
of acronyms including 454, Illumina, SOLiD, etc.), cloning
of genes has largely been replaced by clone-free sequencing.
The extent and power of such sequencing has recently
been demonstrated by the ability to construct essentially
the entire genome of an uncultivated marine Group II Eur-
yarchaeon that constituted only approximately 2% of a
microbial community sample, using SOLiD sequencing,
made possible by high coverage and the use of mate pair
sequencing of ∼3,000 bp fragments (106) (this length is
needed to span repeats and highly conserved genes that oth-
erwise make genome construction difficult).
For 16S rRNA gene studies, “tag sequencing” pioneered
by Sogin et al. (107) consists of amplifying a suitable sized
part of the gene with broadly conserved primers (choice is
important—few are truly universal for the groups intended),
often “barcoded” to allow multiple samples to be combined
into a single run. The amplified products are then sequenced,
and sequences processed en masse. Many thousands of partial
SSU rRNA gene sequences per sample are economically ana-
lyzed this way, though they are usually short (currently a few
hundred bases each, depending on the sequencing platform).
This way of analyzing the composition of microbial commun-
ities is now standard. These approaches provide so much
information even about very rare sequences, that the results
have led to the important concept of the “rare biosphere,”
organisms that may be active or dormant and constituting a
very small proportion of the community (e.g., often much
less than 0.1%), but are potentially important for dispersion,
adaptation to changing conditions, or even critical activities
like nitrogen fixation or vitamin production (108). However,
due to the potential for artifacts like error sequences, this
approach requires rigorous application of quality filtering
and clustering algorithms to avoid erroneous taxa and overes-
timation of species richness (109, 110).
Next-generation sequencing approaches have also allowed
work to begin metagenomics studies of microbial eukaryotes
(111). These studies are still constrained by the much larger
genomes of eukaryotes, and therefore the difficulties of
obtaining sufficient sequences to reconstruct the genome of
a particular species. Along with advances in sequencing abil-
ity, studies of eukaryotic metagenomics have been facilitated
by advances in single-cell isolation which, when coupled
with high-throughput sequencing approaches, reduce the tre-
mendous complexity present in natural, complex eukaryotic
communities to a manageable task (112). Such single-cell
techniques are also quite valuable in studies of bacteria and
archaea, though the amplification technique tends to be
very uneven and typically generates less than half the genome
of each isolated cell regardless of domain (113, 114).
Additionally, similar to DNA, mRNA is amenable to
extraction and sequencing although greater care must be
taken during extraction and purification as RNA shows a
greater susceptibility to degradation during processing. Copy-
ing of mRNA by reverse transcription of RNA into cDNA,
followed by DNA sequencing, has allowed insights into the
metatranscriptomes of environmental samples. Metatran-
scriptomic studies provide information on gene expression
in an ecosystem, and thereby indicates “activity” of the micro-
bial community rather than simply “potential” represented by
the genomic DNA present in the sample, with many applica-
tions, from showing which processes are being carried out
by which organisms to fine-scaled diel studies (115–118).
Nevertheless, given variations in the lifetimes of different
transcripts and protein molecules, the transcriptome may
not be fully representative of the current activity of an
organism.
Limitations of these analyses include sequencing errors,
PCR mismatches or biases, clustering and bioinformatics
challenges, and chimeras generated during PCR. Also, the
phylogenetic resolution of short sequences is limited, given
the high conservation of rRNA sequences. Even with clone-
free shotgun metagenomics, there may be biases, such as non-
random losses of DNA during extraction and preparation,
or biases (e.g., from G + C content or secondary structure)
in the sequencing procedures. Determination of species
diversity by these approaches provides an example of these
limitations. Shakya et al. (119) working with synthetic com-
munities ( purified genomic DNA from 16 Archaea represent-
ing 3 phyla and 48 Bacteria representing 16 phyla, remixed to
simulate an environmental DNA extract) applied both meta-
genomic analysis (454 and Illumina platforms) and PCR
amplification followed by 454 sequencing of 16S rRNA genes
to determine both species richness and relative abundance.
They found that PCR amplification/454 sequencing of 16S
rRNA genes yielded an accurate measure of species richness
( providing that appropriate data processing was applied)
but that the relative abundance of up to 94% of the species
(depending on domain and variable region amplified) was
over- or underestimated by at least 1.5-fold (values ranged
from not detected to 10.3-fold overestimation). In contrast,
both metagenomic approaches yielded relative abundances
that were within the authors’ 1.5-fold accuracy cutoff for
∼50 of the species. However, they concluded that addressing
richness overestimation in metagenomic analyses, that is,
distinguishing rare but real OTUs from experimental and
computational artifacts, awaits further computational and
classification improvements. More recently, Parada et al.
(120) used mock communities composed of 16S rRNA
clones from 27 common marine taxa (from nine Bacterial
and two Archaeal phyla) to show that small differences in
PCR primers (and different clustering methods) can yield
large differences in apparent relative abundances of reported
taxa. However, one primer pair and informatics pipeline
they tested, using a particular version of 515F-926R (V4–
V5), provided accurate estimates of relative clone abundance
(r 2 = 0.95) when comparing observed versus expected clone
abundance.
4.2.2-10 ▪ AQUATIC ENVIRONMENTS
Culture-Independent Diversity Studies
The first groups to be identified using cultivation-independ-
ent molecular techniques from the marine plankton (121)
were the bacterial SAR11 cluster (a group of closely related
gene sequences, or phylotypes) and marine picoplanktonic
unicellular cyanobacteria Synechococcus and Prochlorococcus).
Of these, the SAR11 cluster was completely unknown, but
the cyanobacteria had previously been recognized by their
unique pigment fluorescence; Waterbury et al. (122) and
Johnson et al. (123) used epifluorescence to observe Synecho-
coccus, and Chisholm et al. (124) discovered Prochlorococcus
using flow cytometry. These cyanobacteria were later isolated
and grown in phytoplankton culture media. These two groups
are generally common in the euphotic zone, with SAR11 typ-
ically comprising one third of the planktonic bacteria (125),
and the cyanobacteria common everywhere but polar waters.
Probably the biggest surprise to come from the application
of molecular tools was the discovery by Fuhrman et al. (126)
of abundant archaea in the deep sea. The archaea were found
to be in a unique phylogenetic cluster that was only distantly
related to any previously known archaea, but the “closest rel-
atives” (not really close at all) were extreme thermophiles. A
subsequent study also found archaea to be present in near-
surface coastal waters, albeit relatively rare (<2% relative
abundance). This study used a PCR technique specifically
targeted archaea, and found “marine Crenarchaea” plus a sec-
ond group belonging to the phylum Euryarchaea (127).
Up until this time, all known Archaea were thought to be
“extremophiles”—adapted for either very high temperatures
(thermophiles), extremely salty conditions (halophiles), or
strictly anaerobic environments (methanogens). Yet these
organisms were present in cold or cool water at ordinary salin-
ity and high oxygen concentrations. Fluorescent in situ
hybridization (FISH) measurements from deep-sea samples
have since showed that the archaea may make up appro-
ximately 40% of the total countable prokaryotes, with the
percentage reaching to 60% at 200 m depth in the Mediterra-
nean (35, 36). An extensive time series of FISH measure-
ments near Hawaii confirmed that the archaea are indeed
very abundant throughout the year from below the photic
zone to at least 4,000 m, and typically constituting 30–40%
of the total prokaryotes present in waters deeper than a few
hundred meters (128), with similar results found elsewhere
(129, 130) (Fig. 4). They have been reported from many pla-
ces, including the Atlantic, Pacific, and Southern Oceans,
and are dynamic components of the plankton, the most com-
mon type by far being the “marine Crenarcheaea” (128, 129,
131). While the marine archaea have been reported to be
dominated by a few major “phylotypes” (132), they also
have been shown to have a great deal of microdiversity within
these phylotypes, suggesting there are many kinds of close rel-
atives coexisting (133). Interestingly, this group of archaea
may likely be the most abundant kind of organism on Earth,
given the huge volume of the deep sea and their high abun-
dance there (134). It has recently been proposed that “marine
Crenarchaea” be elevated to their own major phylum outside
the Crenarchaeota, called the Thaumarchaeota, on the basis
of deep phylogenetic branching and fundamental differences
between them and the Crenarchaeota (135). The Thau-
marchaeota possess the uniquely archaeal membrane lipid
crenarchaeol, and are now also known to be abundant in soils.
The physiology of the Thaumarchaeota has been an
intriguing area of study. Initially, an autoradiography-FISH
combination approach demonstrated that they take up
amino acids (130, 136). But subsequent data showed that
crenarchaeol was derived largely from CO2, suggesting auto-
trophic metabolism (137, 138). An experiment showing
uptake of 13C-labeled bicarbonate into these archaeal lipids
(139) directly pointed to autotrophy in this group. Chemoli-
thoautotrophy was first hinted at by Venter et al. (105), whose
metagenomic analysis showed an apparent archaeal scaffold
that contained genes suggestive of ammonia oxidation. Fur-
ther evidence came from Schleper et al. (140), who detected
several ammonia oxidation genes in order directly adjacent to
a Thaumarchaeota 16S rRNA gene in a soil-derived metage-
nomic clone. The issue was directly resolved when Konneke
et al. (141) isolated a related marine archaeon from sediment
of a marine aquarium, and this organism, Candidatus Nitroso-
pumilus maritimus, was found to have a chemoautotrophic
metabolism, quantitatively oxidize ammonia to nitrite, and
contain an archaeal ammonia monooxygenase gene (amoA)
and interestingly did not grow heterotrophically. Cultivation
experiments showed this organism has a high affinity for
ammonium, allowing it to outcompete bacterial nitrifiers at
low (submicromolar) concentrations as are typical in the
sea (142). The complete genome of the organism has shown
novel adaptations for nitrification and autotrophy (143).
Even if the Thaumarchaea are primarily chemolitho-
autotrophs, field data suggest a level of mixotrophy, in that
some organic substrates are being incorporated into biomass.
A stable isotope study using cells collected at 670 m depth
off Hawaii estimated that about 80% of the carbon incor-
porated into archaea-specific lipids came from inorganic sour-
ces and about 20% from organic compounds (144, 145).
Interestingly, it has also been claimed that the genomes
of deep-sea (>2,000 m depth) members of the Thaumarch-
aea, as well as those living in equatorial waters, rarely contain
the amoA gene and thus may be primarily chemoorgano-
trophs (48).
Major bacterial groups that have been documented
from seawater using 16S rRNA characterization include
some that are also known from culture (e.g., Alteromonas,
Roseobacter), and several that are phylogenetically distant
from standard cultures. The most common groups are, in
rough order of their relative abundance in clone libraries
from most to least abundant: SAR11 (relatives of Pelagibac-
ter ubique), Roseobacter, SAR86, cyanobacteria, SAR116,
SAR202, SAR234, and Marine Group A. The SAR designa-
tion followed by a number is an arbitrary sequential clone
identifier from Sargasso Sea cloning studies done by the Gio-
vannoni lab that did most of the early systematic cataloging of
clones. Summaries of the data and phylogenetic relationships
of these groups can be found in Fuhrman and Hagstrom (102)
and Giovannoni et al. (101).
Molecular Genetic Discoveries in Bacterial and
Archaeal Marine Biology
As described earlier, metagenomics is an extension of the
ideas used in the 16S rRNA cloning studies, in that all genes
from the native microorganisms are separated and cloned
without having cultivated the organisms. These methods
early on started to find unexpected and very interesting
results. The best examples involve unexpected marine photo-
trophy, the first of which is discovery of a nonchlorophyll
photosynthetic bacterial pigment, called proteorhodopsin
(146). The gene was found on a large environmentally
derived fragment of DNA that also had a gene coding for
16S rRNA from the so-called SAR86 group (one of
the groups common in seawater). This pigment can act as a
light-driven proton pump, thought to permit cells to generate
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-11

FIGURE 4 Distribution of Bacteria, Euryarchaea, and marine Crenarchaea (recently renamed Thaumarchaea), along a transect in the North
Atlantic Ocean, as measured by CARD FISH and expressed as % of total bacteria + archaea counts via DAPI stained epifluorescence, from
Teira et al. (130). Top panel shows station locations that are shown on the top of the lower three panels, with the distance in km reported along
the transect from lower to higher station numbers. doi:10.1128/9781555818821.ch4.2.2.f4
4.2.2-12 ▪ AQUATIC ENVIRONMENTS
ATP from sunlight. Interestingly, different versions of this pig-
ment are found at different depths, apparently “tuning” the
absorption to match the ambient wavelengths of light as
they change with depth (147). The proteorhodopsin genes
are very widespread and diverse, perhaps in half or more of
all marine bacteria including SAR11 and Euryarchaeota
(105, 148–150). However, almost all of several cultured
organisms with proteorhodopsin studied to date do not show
a growth benefit from light (150), with the exception of a
member of the Flavobacteria, Dokdonia sp. strain MED134,
that grows faster in the light only under moderately low
nutrient conditions (151), and a marine Vibrio has been
shown to survive starvation longer in the light than in dark-
ness (152). These observation may explain the wide phyloge-
netic distribution and high abundance of this gene in that
proteorhodopsin may often assist long-term survival of bacte-
ria under extreme energy-limiting conditions, yet not have
much effect on growth during more energy-replete conditions.
Interestingly, a recent report suggests the growth benefit from
light in Dokdonia MED134 is from enhanced uptake of its
required growth factor vitamin B1, which is transported by a
tonB-dependent transporter powered by a proton gradient
(153). This pigment might also have sensory or other roles
not yet well examined (150).
The second kind of phototrophy found by molecular
genetic and fluorescence techniques to be unexpectedly
important in marine plankton is anoxygenic aerobic bacterial
phototrophy, based on the pigment bacteriochlorophyll a.
Culturable aerobic anoxygenic photosynthetic (AAP) bacte-
ria have been known from seawater for several years (154) and
are widespread (155). Initial recent reports of direct counts
claimed they represent about 11%, of the total bacterial com-
munity in the euphotic zone (156), but those early counts did
not correct for the presence of other bacteria, and such correc-
tion often yields estimates averaging closer to 2% (157). Direct
measurements show the bacteriochlorophyll pigment is rela-
tively rare (158). However studies in various ocean locations,
including ocean gyres, report that these AAP bacteria
can indeed, on occasion, make up a quarter of the total prokar-
yotes (159, 160). A recent energetic modeling of AAP and
proteorhodopsin-containing bacteria (161) has suggested
that proteorhodopsin-containing ones can gain approximately
0.2% as much energy from sunlight as Synechococcus (common
marine cyanobacterium) and AAP bacteria can gain approxi-
mately 1.3% as much energy from sunlight as Synechococcus.
They concluded the AAP bacteria may gain energy sufficient
to meet maintenance costs, but proteorhodopsin-containing
bacteria were not expected to do so except at high light inten-
sities and with large numbers of proteorhodopsin molecules per
cell. They concluded the ease and low cost of maintaining
PR-based phototrophy (a few genes required) may explain
the high incidence of proteorhodopsin genes.
Other metagenomic studies have yielded interesting
insights about organisms and processes that would otherwise
be difficult or impossible to establish. For example, the
sequence of an archaeal gene fragment isolated from seawater
revealed extensive evidence of genetic exchange with other
types of archaea and even bacteria (162). Such genetic
exchange was also quite evident from the complete gene
sequences of different strains of Prochlorococcus and Synecho-
coccus, which was attributed in part to virus-mediated gene
flow (163–166).
Dilution Cultures of Bacteria and Archaea
Some recent cultivation techniques that permit growth of
“typical” oligotrophic marine bacteria in pure culture are
based on the simple concept that bacteria living in seawater
on dissolved organic matter might best be grown in the labo-
ratory in ordinary filtered seawater. This conclusion follows
from similar techniques to grow mixed “seawater cultures”
of marine bacteria (167), but the seawater inoculum in this
case is diluted so that only one or a few bacteria are added
to the initial culture vessel (168). A rapid throughput version
of this method has been used to cultivate members of bacte-
rial groups thought to be conventionally uncultivable, such as
the SAR11 clade (169, 170), albeit often at very low den-
sities. Such cultures permit focused studies of the properties
of those organisms, including genomic analysis (169, 170).
This valuable work provides insight into the roles and activ-
ities of some of the most common bacterial phylotypes
observed in the ocean (171), showing for example that culti-
vated SAR11 requires a source of reduced sulfur, as well as
providing templates aiding the interpretation of metage-
nomic analysis (172).
PROTOZOA
Diversity and Distribution
Heterotrophic protists have been identified historically from
their morphological features that are apparent at the level
of the light or electron microscope. Features of cell size, shape,
type, and pattern of flagellation/ciliation, skeletal structures,
characteristics of the nucleus, and other cellular structures
and organelles have been used to differentiate the many thou-
sands of described species. A description of the extent of this
diversity of form and function is well beyond the scope of this
book (see [79] for more information). However, much of this
diversity can be reduced to three basic body plans that dictate
the broadest ecological roles of these cells; amoeboid, flagel-
lated, and ciliated forms (Fig. 5). Among the lineages possess-
ing one of these three forms only the last group, the ciliated
protists, constitute a monophyletic group within the domain
Eukarya.
One of the simplest body plans for protozoa is the amoe-
boid cell, exemplified by the gymnamoebae, or “naked”
amoebae. Motile, nonphotosynthetic cells lacking flagella
or cilia occur as life stages in a number of phylogenetically
diverse taxa, but for many free-living species of protozoa
this form constitutes the only life stage. Motility is largely
confined to movement along surfaces by means of pseudopo-
dia that can take on a variety of (species-specific) shapes. Sig-
nificant abundances of amoebae are largely confined to
benthic and epibiotic environments (174) and to suspended
particulate material where they can occasionally be highly
enriched (27). Most amoebae consume bacteria and other
minute prokaryotes and eukaryotes.
Several heterotrophic protistan groups possess complex
amoeboid body plans, most notably the foraminifera, polycys-
tine and phaeodarian radiolaria, and the acantharia. Plank-
tonic forms are predominantly oceanic in their distributions
(although a large number of benthic species of foraminifera
exist). These species are heterotrophic, and many are visible
to the naked eye (individual cells can be ≥1 cm, gelatinous
colonies can form ribbon-like structures ≥1 m in length).
They feed on a wide variety of bacterial, protistan, and meta-
zoan prey using pseudopodial networks, have rather long,
complex life cycles for individual cells (weeks to months),
and are extensively used in studies of paleoclimatological
reconstruction (82, 175). In addition, many of these species
possess intracellular symbiotic, usually eukaryotic, algae.
The widespread occurrence of algal endosymbiosis among
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-13

FIGURE 5 Body plans and size ranges of protozoa. These micrographs depict amoeboid (a–g), flagellated (h–j), and ciliated (k–m) forms of
protozoa. From Caron et al. (173). Markers bars are 5 (i), 10 (h), 20 (c, j, l), 30 (a), 50 (k, m), 100 (b, f ), 500 (d), and 1,000 (e, g) µm.
doi:10.1128/9781555818821.ch4.2.2.f5

these protozoa in oceanic pelagic environments implies that
strong selective forces appear to give rise to these associations
(97, 98).
Flagellated forms of protozoa exist across many protistan
lineages. Eukaryotic flagella come in various sizes, numbers,
and forms (e.g., with or without tiny hairs) that are character-
istic of the different lineages. Many flagellate species are appa-
rently purely phototrophic or heterotrophic ( phagotrophic),
but a significant (still poorly known) fraction of flagellates are
mixotrophic, combining both phototrophy and phagotrophy
(70, 88). Flagella in phagotrophic flagellates are employed for
motility and prey capture. Most free-living flagellates possess
one to four flagella (typically one or two) that can be many
times the length of the cell itself. Although there is tremen-
dous species diversity among flagellated protozoa, many of
these species have broadly overlapping ecologies. As a group,
flagellates are the most numerically abundant protozoa of
both benthic and pelagic ecosystems, and they are fundamen-
tally important as consumers of bacteria, cyanobacteria,
and other eukaryotes. Flagellated protozoa within the nano-
plankton size class (2–20 µm) are often counted as a single
assemblage in plankton studies because morphological details
apparent using light microscope are insufficient to distinguish
among the many species. This assemblage has been variously
referred to by a variety of names and acronyms including het-
erotrophic nanoplankton (HNAN, HN), heterotrophic flag-
ellates (Hflags), heterotrophic nanoflagellates (HNF), and
microflagellates. Heterotrophic nanoplankton is the most
accurate term for most methodologies employed to count
these cells because it does not require visualization of flagella
(which are often lost from these small cells during preparation
for microscopy). The confusion over terminology is partly a
consequence of the different methodologies that have been
used to count these species, and partly a consequence of the
fact that these species were first studied prior to the wide
acceptance of the size convention of Sieburth et al. (84).
Heterotrophic flagellates in the microplanktonic size class
(20–200 µm, mostly dinoflagellates) are important consum-
ers of phytoplankton in pelagic ecosystems. Many of these
species are capable of the production of large pseudopodial
nets, the pallium, that can engulf prey ( particularly diatoms)
significantly larger than the diameter of the dinoflagellate
theca (176). This behavior and its ecological significance for
energy flow in plankton communities have been recognized
4.2.2-14 ▪ AQUATIC ENVIRONMENTS
only within the past few decades (177). These species are
often abundant in waters where diatoms dominate (178,
179). In such situations, heterotrophic dinoflagellate biomass
can be similar to that of ciliated protozoa (180).
Ciliates are generally the most recognizable form of pro-
tozoa to nonspecialists. Nonetheless, ciliates are a diverse
group morphologically as well as ecologically (181, 182).
The degree of ciliature on these species can range from uni-
formly ciliated to totally devoid of cilia during most of their
life cycle, to ciliature restricted to specific regions of the
cell. In addition, cilia can fuse to form various complex
structures (e.g., ciliary membranes or veils, cirri) that assist
in locomotion, food capture, or attachment. Ciliates are abun-
dant and ecologically important species in both benthic (8)
and pelagic (183) ecosystems and can consume a variety of
prokaryotic and eukaryotic prey. Together with the heterotro-
phic dinoflagellates, these species are the dominant consum-
ers of phytoplankton in many pelagic ecosystems (184), and
as a consequence they form an important trophic link to met-
azoan zooplankton (185, 186).
Diversity and Biogeography of Protists: From
Morphology to DNA Sequences
Until relatively recently, the biodiversity of marine protistan
assemblages was not generally considered a controversial
topic. It was generally accepted that, while all species of pro-
tists certainly have not been identified, representatives of
most types of algae and protozoa had been observed and
described, if not actually brought into culture in the labora-
tory. Direct sequencing of 18S rRNA genes from environ-
mental samples (as described in “Molecular Phylogeny and
Metagenomics: Field Applications”) has changed that view.
Initial forays into environmental DNA indicated a much
greater diversity of protists than previously documented using
classical approaches of culture and microscopy (81, 187–
191). Numerous publications over the past decades have
expanded these findings (see [192] for a recent global analy-
sis), which are highly analogous to discoveries in marine pro-
karyote research as described already.
Previously uncharacterized protistan diversity has now
been documented at virtually every level of eukaryotic
organismal classification. Some of these findings could
have been expected, but some have been very unexpected.
For example, a much greater diversity than noted previously
has been observed among small (<10 µm) protists in plank-
tonic ecosystems. These species generally possess few distinc-
tive morphological features, and one could expect that many
cryptic species might be present among these small morpho-
types. The molecular “discovery” of this eukaryotic diversity
has stimulated progress on the isolation and description
of new species and genera of minute algae and protozoa
(193–197).
Analyses of environmental samples have also indicated
the unanticipated existence of novel 18S rRNA gene sequen-
ces that imply the presence of novel lineages of eukaryotes in
natural protistan communities (198–200). These are sequen-
ces for which there are apparently no known, described, or
cultured species. The degree to which these sequences dif-
fer from sequences of known sequenced eukaryotes implies
that some of these lineages may be distinct at the level of
phylum (201).
These findings have raised basic questions, and some
debate, concerning the true diversity and biogeography of
protistan assemblages in natural ecosystems (202–204), the
validity of the many novel phylotypes or cryptic species that
are being documented (205), the potential importance and
significance of the many rare taxa that characterize these
communities (206), and the implications of these findings
for the ecological/biogeochemical roles that protists play
in aquatic ecosystems. For example, it was postulated and
subsequently confirmed that some novel alveolate lineages
reported from marine ecosystems make up a suite of parasitic
protozoa whose ecological importance may have been signifi-
cantly underestimated in the past (207). Deciphering the
identity and significance of these many unknown phylotypes
will constitute a significant effort for protistologists in the
future (80, 85).
Life Histories and Ecological Strategies of Protozoa
Abundance, prey type, and life histories all vary tremendously
among marine protozoa. The smallest species (e.g., many flag-
ellates) tend to be the most abundant and widely distributed
in the world ocean. Indeed, many of these species may be
globally distributed (208). Most nanoplanktonic flagellates
have potentially rapid rates of grazing and growth. Under
optimal conditions these species can divide by binary fission
several times a day, and thus dramatic increases in their pop-
ulations can take place in response to favorable conditions
within a few days (209). Many of these species are capable
of surviving for limited periods of time without food and
have developed a variety of physiological or life cycle strat-
egies to cope with these events (209). However, in contrast
to some bacteria that may remain viable through long periods
of starvation (210), protozoa will expire or encyst in response
to low food abundance, and thus “boom-and-bust” popula-
tion changes are characteristic of small flagellates.
On the other end of the size spectrum from the rather
ubiquitous nanoflagellate species, many of the largest proto-
zoan species ( polycystine and phaeodarian radiolaria, plank-
tonic foraminifera) are exclusively oceanic (i.e., do not
survive in most coastal environments) and/or have specific
latitudinal and depth distributions. Thus, the abundances
of these latter species may range from undetectable to max-
imal abundances of >105 individuals/m3. Dramatic changes
in the abundance of these protozoan taxa also can be related
to changes in prey abundance, physical/behavioral aggrega-
tion, or to periodicity of life cycle events. For example, the
planktonic foraminifer Hastigerina pelagica reproduces on a
lunar cycle and thus abundances (and life stages) of this spe-
cies in oceanic waters can vary considerably over the course
of a month (211). In general, life cycles for the large amoe-
boid protozoa are lengthy and complex (for single-celled
organisms) with life spans unknown for many species
(attempts to culture them in the lab have so far been unsuc-
cessful) but estimated to be on the order of months to per-
haps years (175).
The abundance and activities of microplanktonic hetero-
trophic protists (mostly ciliates and heterotrophic dinoflagel-
lates) tend to be somewhat intermediate to those of
nanoplanktonic flagellates and the larger amoeboid forms.
These species are present in the majority of marine ecosys-
tems and collectively play an important role in the control
of phytoplankton biomass (and probably the abundance of
nanoplanktonic protozoa, although there is little informa-
tion on this topic) in waters throughout the world ocean.
Like small flagellates, ciliates and dinoflagellates reproduce
primarily by binary fission, but their maximal growth rates
are typically slower (one division a day is typical) and their
life cycles often include sexual phases that allow for genetic
recombination.
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-15

VIRUSES in upper left). Each type of virus has a fixed morphology

Viral Abundance and General Properties
Viruses are simple biological agents, typically 20–200 nm in
diameter, composed of a nucleic acid genome in a protein
coat, that infect cells and “commandeer” the cell’s machinery
to make more viruses, which are released into the environ-
ment when the host cell lyses or bursts. A particular virus is
thought to be capable of infecting only a narrow range of hosts
(usually one species, sometimes a genus, rarely broader).
Although some early studies had isolated viruses from the
sea, there were no data prior to the 1980s showing such viruses
were very abundant, and more significantly, no evidence that
infection was occurring in any important part of the plankton
community. It wasn’t until the late 1980s that electron micro-
scopy with suitable concentration methods showed that
viruses are extremely abundant, similar to or even higher
than bacterial abundance (212, 213). Interestingly, the
most suitable transmission electron microscopy (TEM)
approach, used by Bergh et al. (described in detail by [214])
was actually similar to a direct centrifugation method devel-
oped in 1949 (215).
TEM studies of viruses permit high-resolution images and
observation of viral morphology, showing features such as
head diameter, tails, sheaths, and tail fibers (Fig. 6, collage
(unlike bacteria that are potentially more plastic in their
appearance) and hence a coarse measure of viral diversity
has been possible by cataloging morphologies. Studies that
have done so have found dozens or more different morpholo-
gies of marine viruses in a given sample (217–219). A large
proportion resemble bacteriophages, which are viruses that
infect bacteria.
The availability of brightly fluorescent nucleic acid stains
and high-porosity fine pore size (0.02 μm) filters made from
Al2O3 have made it easy to accurately count viruses by epi-
fluorescence microscopy (38, 220, 221). Epifluorescence per-
mits abundance estimates but does not allow observation of
viral morphology, as viruses are below the resolution limit
of light microscopy; viruses are visible only as sources of light
(like stars in the night sky, see Fig. 6, lower micrograph). An
extension of manual epifluorescence counts is the use of flow
cytometry to count viruses, now commonly employed in some
labs (222).
Virus abundance has been found to be closely related to
bacterial abundance, with a virus:bacteria ratio typically
10–30:1. A typical oceanic profile of bacterial and viral abun-
dance is shown in (Fig. 6, panel on right). This tight ratio
and the strong correlation to bacterial abundance, in relation
to weaker correlations to chlorophyll, have been cited as

FIGURE 6 Marine viruses (upper left collage) viewed by transmission electron microscopy. These are cyanophages in the groups myoviridae
(right and bottom), siphoviridae (left), and podoviridae (top). Scale bars represent 100 nm. From Sullivan (216). Epifluorescence micrograph
(lower left) of SYBR green I stained viruses (small fluorescing objects) and bacteria + archaea (large fluorescing objects). Depth distribution of
virus and bacteria + archaea abundances, obtained by epifluorescence microscopy of SYBR Green stained cells, in the central San Pedro Basin,
California (11 August 2000). doi:10.1128/9781555818821.ch4.2.2.f6
4.2.2-16 ▪ AQUATIC ENVIRONMENTS
evidence that most marine viruses infect bacteria rather than
eukaryotic phytoplankton (217, 218, 223, 224), though there
are also many important viruses of phytoplankton and other
organisms, with significant impacts on primary productivity,
blooms, and ecosystem function (225–227).
The first demonstration of the activity of marine viruses
was presented by Proctor and Fuhrman (228), who showed
by electron microscopic examination that viruses were
actually infecting marine bacteria and cyanobacteria at a
measurable rate. Because only the final portion of the virus
life cycle is visible by electron microscopy (when the viruses
are assembled and ready to lyse the host cell), only a very small
fraction of the infected cells can be counted as infected at any
given time. Proctor and Fuhrman (228) reported that only a
few percent of the total bacteria from coastal waters and the
Sargasso Sea were visibly infected, but they interpreted the
data with a model that implied the actual fraction of the total
community that was infected was much higher. They con-
cluded that the total fraction of bacterial mortality attribut-
able to viruses is roughly 10–40%. Subsequent studies have
used refined versions of that same general approach and
model (229, 230), and numerous studies have used a variety
of alternative approaches to estimate viral activity—all yield-
ing basically the same general conclusion (although the
model parameters need to be adjusted for cyanobacteria,
which on average seem to be infected less than bacteria).
These approaches include (a) calculation of virus turnover
from decay estimates (231), (b) measurement of viral DNA
synthesis by incorporation of tritiated thymidine (232), (c)
observation of effects of added viruses (233, 234), (d) estima-
tion of bacterial mortality in the absence of protists, (e) use of
fluorescent viruses to estimate production by an approach
analogous to isotope dilution (38), and (f ) dilution of viruses
in filtered samples to estimate viral production (235). The
overall consensus of these studies is that the initial estimates
were basically correct, that is, that viruses are responsible
for about 10–40% of the bacterial mortality in most marine
systems (reviewed by [218, 223, 226, 236, 237]). However,
there is still some disagreement about the higher estimates,
and it is likely that 40% mortality from viruses alone is not
typical for most marine systems.
Most of the foregoing work has focused on the viral infec-
tion of bacteria, primarily thought to be heterotrophic,
although the original report by Proctor and Fuhrman (228)
also noted the occurrence of cyanobacterial infection. Sev-
eral subsequent studies focused on phytoplankton, including
the potential effect of viruses on the termination of phyto-
plankton blooms of Emiliania and Phaeocystis, and the likely
impact this might have on release of climate-active gases
such as dimethyl sulfide (225, 227, 238–247). Although
details are beyond the scope of this chapter, viruses are
thought to infect virtually all marine organisms, with poten-
tially significant impacts from zooplankton to whales (226).
Detailed studies of viruses infecting cyanobacteria such as
Synechococcus have shown differences in viral host specific-
ity, particularly toward coastal and oceanic host strains,
and some occasional high virus abundances (to 105/ml), as
measured by most probable number (MPN) cultivation tech-
niques, in the Gulf of Mexico near Texas (248, 249). Other
virus cultivation studies with Prochlorococcus and Synecho-
coccus in oligotrophic waters of the Sargasso Sea showed
an interesting pattern of cross-infection between these gen-
era by some virus types (suggesting gene flow among these
organisms), but generally low MPN estimates of abundance,
to 103/ml, even when cyanobacterial abundance was near
105/ml (216, 250).
Overall, the consensus emerging from direct comparisons
of viral-mediated mortality and grazer-mediated (i.e., micro-
zooplankton) mortality have indicated that viral lysis of bac-
teria constitutes a significant fraction of total mortality of
this assemblage, while the mortality of phytoplankton
appears to be dominated by microzooplankton in most instan-
ces (251, 252).
As mentioned, viruses have significant morphological
diversity as observed in TEM studies. It is also possible to
investigate aspects of their genetic diversity. In early work
on this topic, Wommack et al. (253) and Steward et al.
(254) observed the diversity of viral genome lengths in a field
sample by pulsed field gel electrophoresis, with viral genomes
ranging 25 kb to >300 kb in length. Field results show that
the viral community composition is dynamic in space and
time, with clear changes in the banding patterns over seasons
and locations in Chesapeake Bay (253), between ocean
basins and subsequent to dinoflagellate blooms (254), and
with depths to 500 m at one location (255).
Although viruses do not all share a set of core genes
that allows a universal viral phylogeny (comparable to SSU
rRNA gene in cellular organisms), genetic diversity among
a single group of viruses can be examined by sequence analysis
of shared genes within the group. An example is the g20 gene
in T4-like cyanophages, that has shown extremely high diver-
sity even among very closely related viruses and has demon-
strated geographic and seasonal variation (256). A second
example is g23, which is found broadly in diverse and wide-
spread T4-like phages (257) and can have seasonally repeat-
ing patterns (258) as well as short-term rapid dynamics that
correlate to those in bacteria (259).
Viral Metagenomics
As with cellular organisms, metagenomics can be used to
examine viral diversity and genetics without the many restric-
tions of cultivation. Viruses can be collected by selective fil-
tration and concentrated by tangential flow filtration or
flocculated with iron chloride (260, 261), then their collec-
tive metagenome can be extracted, linker amplified, and ana-
lyzed by sequencing (262, 263). Viral metagenomic studies
are particularly challenging because the large majority of
sequences have no annotated matches in any databases, but
so far results from marine samples around the world have
shown extremely high diversity and variations with depth,
location, and time, presumably with highly dispersed types
selected by local conditions (262–266). Because metage-
nomes are best interpreted when there are representative cul-
tures available, the best matches of viral metagenomes
originally tended to be to the few viral isolates infecting truly
common marine bacteria, like cyanophage infecting Synecho-
coccus and Prochlorococcus (267). However, the development
of dilution-to-extinction cultures representing common het-
erotrophic marine taxa, like Peligibacter (a member of the
SAR11 clade) and SAR116, has allowed isolation from sea-
water of viruses infecting these common organisms, and these
isolates have indeed been found to be highly abundant in
marine viral metagenomes (268, 269). New approaches to
interpret the results include clustering the proteins independ-
ent of known proteins to compare samples to each other and
try to find environmental factors driving viral community
changes (270). Such analysis of a large global data set (Tara
Oceans expedition) has shown that extensive sampling has
come close to reaching the total diversity in tropical and tem-
perate waters of such viral protein clusters (which essentially
represent various viral protein motifs, not all viral protein
types), and that viruses appear to be directionally dispersed
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-17
“downstream” in major ocean currents, as one may expect
from first principles but also supporting the “seed bank”
hypothesis of viral biogeography (271).
Viruses and Host Diversity
Viruses are themselves thought to be instrumental in driving
increased diversity of their microbial hosts, via a hypothesized
process often called “kill the winner.” This is because viral
infection is host-specific and density-dependent, the latter
because viruses diffuse from host to host, so an abundant
host is more likely to pass on infection than a rare one.
This means that if an organism becomes abundant and
blooms, winning the competition for resources, it becomes
more susceptible to a viral epidemic. This would benefit the
rarer organisms and thus help foster diversity (223, 272,
273). However, bloom scenarios as described above involve
systems far from steady state, and the formal theory of Thing-
stad and Lignell (274) has interesting steady-state solutions
where several viruses infect several hosts stably over time
(via trade-offs between growth rates and viral susceptibility),
which may occur at the strain or species level (274). There is
some experimental evidence that viruses have effects on nat-
ural marine microbial community composition weaker than
the kill the winner (bloom version) hypothesis would sug-
gest (275, 276), so something resembling the steady-state
coexistence described in the model may in fact be common.
There also appear to be processes that foster coexistence
between viruses and hosts, but the mechanisms are largely
speculative (223, 277, 278).
Viruses may also be directly involved in host genetic diver-
sity because they can be the agents of genetic exchange
between microorganisms (223, 273). This often involves
the viral lifestyle known as lysogeny, whereby viruses survive
within host cells as DNA only, integrated into the host chro-
mosome and being reproduced each time the host divides. A
host harboring such a genome is called a lysogen because
under conditions of stress to the host cell, a genetic switch
may cause the viral genome to initiate the lytic process, pro-
ducing many progeny viruses and bursting from the host.
Lysogeny is a very common property, occurring in a signifi-
cant part of the bacterial community (279, 280), although
the incidence of induction of the lytic phase in nature is appa-
rently low (281). Overall, lysogeny is poorly understood, but
thought to have both positive and negative impacts on the
microbial community (282). Recent results suggest that ben-
efits of being lysogenic in highly seasonal polar seas leads to
fundamental difference between polar and other marine viral
communities (283).
Viruses and the Microbial Loop
As part of the food web, viruses occupy a unique position.
They infect host cells that are mostly thought to be heterotro-
phic bacteria, and by doing so they typically burst the hosts to
release progeny viruses and cellular debris. But what is the fate
of this material? Viruses themselves do not last indefinitely,
and a simple steady-state assumption implies that from each
burst of viruses (typically 20–100 per lytic event), only one
successfully infects another cell. The rest are inactivated
and broken down by sunlight (UV and visible exposure)
and enzymatic attack (284) or consumed by minute phagotro-
phic protists (285), thus reentering the food web as substrate
for bacteria or food for protistan consumers. Experiments in
controlled laboratory systems and field studies with radioac-
tively labeled viral lysis products have supported the conclu-
sion that most of the organic matter released by the viral
infection is either taken up by bacteria or respired (286, 287).
Modeling this process as part of the microbial loop shows
that viral lysis represent a sort of side loop that has the net
effect of remineralizing a significant amount of the carbon
and nutrients that enter the bacteria-protist part of the micro-
bial loop (Fig. 7). A theoretical numerical steady-state model
comparing a system with no viral activity to one where viruses
are responsible for 50% of bacterial mortality showed that the
system with viruses had 33% more bacterial production and
respiration than the virus-free system, implying that the
viruses had the effect of permitting the bacteria to process
more of the primary production than they would otherwise
(223). Although 50% is a high number unlikely to be com-
mon in the sea, this model nevertheless illustrates that viruses
can reduce the amount of energy reaching higher trophic lev-
els. The implication is that viruses lead to increased bacterial
activity at the expense of the larger organisms.
MAJOR ENVIRONMENTAL CONTROLS
Light, Temperature, and Pressure
Temperature has an important potential influence on bio-
chemical reactions and therefore on biological processes
in general. Most ocean waters fall in the range of –2°C to
30°C, with obvious exceptions in hydrothermally heated
areas. Temperature has long been known to be a regulating
factor for the growth of heterotrophic microbes.
In temperate waters, it has been established that microbial
activity is generally much higher in warm summer waters than
in winter (55). The relationship is not simple, however,
because multiple factors act at the same time. Some contro-
versy still exists regarding the highest and lowest extremes
for marine bacterial growth, although there is broad agree-
ment that bacteria grow >100°C at hydrothermal vents and
<−5°C in sea ice brines. Pomeroy et al. (57) noted the inter-
esting observation that bacteria seem particularly inhibited
near the freezing point of seawater (ca. –2.2°C), compared
to eukaryotic phytoplankton. This effect results in polar
spring phytoplankton blooms that accumulate organic car-
bon in advance of the response of the bacterial community
and development of the microbial loop, and perhaps lead to
enhanced benthic-pelagic coupling (288).
The relationship between temperature and the growth rate
of marine phytoplankton was described broadly in a now
FIGURE 7 Modification of the microbial loop concept that incor-
porates the functional role of viruses. Export can be via predation or
sinking. From Fuhrman (223).
doi:10.1128/9781555818821.ch4.2.2.f7
4.2.2-18 ▪ AQUATIC ENVIRONMENTS
classic paper by Eppley (58) and later Goldman and Carpen-
ter (289). Temperature was shown to exert a strong and direct
effect on the maximal growth rates of these species, with max-
imal intrinsic growth rates at 0°C generally <1 division/day
while growth rates at 30°C may be >4/day. These relation-
ships indicate the maximal rate that might be attained by
phytoplankton at these temperatures, but they do not take
other factors into account (nutrients, light). While warm
temperature ostensibly allows more rapid growth, it creates
hydrographic conditions that typically give rise to nutrient
limitation of algal growth. Thus, some of the coldest waters
in nature witness some of the most massive phytoplankton
blooms (290) while warm oceanic gyres represent some of
the most oligotrophic areas of the ocean. The multiple and
often conflicting effects of temperature on the growth of nat-
ural phytoplankton assemblages limit the accuracy of present
predictions regarding how primary producers will respond to
climate change (291).
An analysis of the effect of temperature on the growth of
heterotrophic protists has indicated that the response is qual-
itatively similar to that of phytoplankton, but with a twist.
The growth rates of at least some protozoa can exceed those
of phytoplankton at warmer environmental temperatures,
but the opposite effect is apparent at very low environmental
temperature. This differential effect of temperature on the
growth of phototrophic and heterotrophic protists was based
on a large meta-analysis of published protistan growth (59).
That analysis demonstrated that the maximal growth rates
attained by phototrophic protists could exceed the maximal
growth rates attained by heterotrophic protists (all other
potential growth-limiting factors not considered). Therefore
protozoan growth rates may be constrained to a greater degree
at low environmental temperature than rates for phytoplank-
ton. If so, then phytoplankton blooms may get a head start on
grazers during spring in polar ecosystems. This scenario is con-
sistent with information on seasonal biomass changes and
microzooplankton herbivory in the Ross Sea, Antarctica
(292, 293) but there are still too few data to fully vet this
hypothesis.
The importance of high pressure on bacterial growth
gained considerable attention in the late 1960s, when the
deep sea submersible Alvin was accidentally lost overboard
with its hatch open. While no lives were lost in this accident,
some workmen’s lunches sank to the bottom (∼1,500 m)
inside the submersible. Alvin was recovered after 10 months
and, interestingly, there was a waterlogged lunch containing
apples, bologna sandwiches, and broken vacuum bottles
with broth that all appeared hardly degraded and tasted palat-
able. Yet when placed in a refrigerator on the ship, these items
degraded relatively quickly. Initially, pressure was thought to
be the preserving factor, as the sea floor temperature was sim-
ilar to the refrigerator temperature and the only major differ-
ence would be pressure (294). Following this observation,
a series of experiments to measure degradation of various
organic materials left in the deep sea for extended periods
indicated that degradation was typically significantly reduced,
implying that pressure reduces the degradation rates (295).
Nonetheless, changes in the protozoan community of natural
detrital material sinking to the deep ocean floor indicate that
the microbial community can respond relatively quickly in
some situations (296). Deep sea microbiology has advanced
considerably, yet it is still difficult to interpret results with
respect to actual in situ rates of naturally occurring organic
matter. Deep sea bacteria adapted to high pressures have
been isolated that are barophilic (also called piezophilic),
meaning that they prefer high pressures and have reduced
activity at lower pressures, whereas others are barotolerant
( piezotolerant), tolerating but not preferring high pressures,
for example, see (297, 298).
Few data are available on barotolerant/barophilic marine
protozoa. Protozoa certainly exist and grow at great oceanic
depths, but measurements of in situ growth rates for these spe-
cies do not yet exist. Measurable protozoan numbers have
been documented in the deep-sea sediments for more than
30 years (299, 300) and viable protozoa have occasionally
been cultured from these environments (301–305), but
very few direct measurements of the activities of these species
in situ have been reported (306). A few protozoa have been
isolated that will grow at high pressure (302, 303, 307), and
protists that appear to be unique to the deep ocean have
been observed either directly or through the analysis of
DNA sequences (308–311), but possibly the best direct evi-
dence that protozoan activity takes place at the high pressures
characteristics of the deep sea are experimental and observa-
tional work noting the stimulatory effect that detrital deposi-
tion has on some components of the protozoan community
(296, 312). These observations indicate a diverse and active
protozoan fauna of the deep ocean, although their biogeo-
chemical significance is largely uncharacterized at this time.
Dissolved and Particulate Organic Matter
Bacteria and archaea are thought to be by far the most impor-
tant organisms with respect to the processing of dissolved
organic matter (DOM) and nonliving particulate organic
matter (POM; also called detritus) in the ocean. While there
may be some uptake of DOM by protists, particularly for
growth factors needed in trace amounts such as vitamins
(313), the bulk of this material is probably utilized by bacteria
and archaea (314), including the smallest cyanobacterium
Prochlorococcus (315, 316), which thus may be considered a
mixotroph. Due to their small size, bacteria have extremely
high surface:volume ratios and, combined with their over-
whelmingly high abundance, an extremely high integrated
surface area. Protozoa tend to obtain the organic materials
that they require for growth from their prey rather than
through the uptake of DOM. Overall, protozoa tend to be
sources of dissolved and detrital organic substances, through
the excretion of unassimilated prey biomass in expelled
food vacuoles.
Particulate organic matter is not directly available as sub-
strate to bacteria. These substances must first be reduced to
small molecules that can be transported into the cell. This
is accomplished by the production of extracellular enzymes
(note that few if any large polymers are directly taken up by
bacteria, with the possible exception of DNA). Hydrolytic
enzymes produced by bacteria (and almost certainly archaea)
break down polymers like proteins, polysaccharides, and
nucleic acids. As with DOM, POM is composed of a complex
mixture of compounds that vary in their susceptibility to bac-
terial degradation and utilization.
Particulate material in the water column serves not only as
bacterial substrate but also as substratum. POM occurs in the
water column across a huge size spectrum from micrometers
up to some detrital aggregates more than 1 m in diameter
(317). Much of this particulate material is in a constant state
of flux, with colloidal material constantly coalescing and
aggregating to form new or larger particles (318) as microbial
degradation acts simultaneously to remineralize this material.
Detrital particles that attain macroscopic size either by direct
formation (317, 319, 320) or via accretion and aggregation
(321) and are often called marine snow or macroaggregates.
Marine snow particles are readily colonized by bacteria and
 4.2.2. Heterotrophic Planktonic Microbes: Virus, Bacteria, Archaea, and Protozoa ▪ 4.2.2-19
many other microorganisms, and their abundances on these
aggregates can be orders of magnitude greater than abundan-
ces in an equivalent volume of the water surrounding the
aggregates (320, 322–324), and their diversity and highly
concentrated abundances on detrital aggregates implies that
there may be stronger trophic coupling and efficient energy/
elemental cycling relative to these processes among free-
living microorganisms in the surrounding waters. This specu-
lation has resulted in a number of formalizations of the poten-
tial impact of aggregate microbiology on biogeochemical
processes in the ocean (325, 326). Most definitions of bio-
films include aggregated particulate and these particles are
often studied within the biofilm conceptual model (see chap-
ter by Lawrence et al. in this section).
Inorganic Nutrients
Macronutrients (N, P)
Virtually all heterotrophic organisms in the ocean contribute
to the pool of available macronutrients via the excretion of
metabolic wastes, and protozoa contribute significantly to
this process. Macronutrients are often in excess within proto-
zoan prey relative to the consumers’ needs because much of
the prey carbon is respired to produce energy for metabolism
and growth. “Excess” nutrients are eliminated as either
organic compounds, or more commonly as ammonium (for
N) and phosphate (for P). The importance of the match
(or mismatch) between predator and prey stoichiometry has
been demonstrated experimentally and repeatedly in marine
and freshwater ecosystems and has been elevated to a funda-
mental ecological principle governing elemental and energy
flow in aquatic food webs (327). Based on this reasoning,
Caron et al. (328) concluded that bacterivorous protozoa
play an important role in nutrient remineralization and
release in the ocean. Bacterial biomass is typically rich in N
and P relative to other organisms, so bacterial predators expe-
rience large excesses of these elements relative to their growth
needs. Moreover, bacterivorous protozoa should be dispropor-
tionately important in nutrient remineralization among
single-celled eukaryotes because, as the smallest protozoa in
aquatic ecosystems, they have high weight-specific metabolic
rates (329). Thus, both stoichiometric and allometric rela-
tionships implicate small, bacterivorous protozoa as impor-
tant sources of remineralized nutrients in the plankton.
The reasoning also holds true for bacterial utilization of
organic compounds, and it has led to some interesting
findings and conclusions regarding the role of bacteria in
nutrient cycling in the ocean. Bacteria are important sources
remineralized nutrients when they are consuming N- or
P-rich substrate. Contrary to this traditional view, however,
heterotrophic bacteria are often strong competitors for inor-
ganic nutrients under many situations, largely because bacte-
rial biomass is richer in N and P than much of the organic
matter they consume in nature, so they require additional N
and P to produce their biomass (summarized by [330]). Exper-
imental studies have demonstrated the uptake of ammonium
by bacteria other than cyanobacteria using stable isotope 15N
(331), and also the short-lived radioisotope 13N (332), often
finding that bacteria may be responsible for one third of the N
or P uptake. Additionally, there is evidence that bacteria in
some oceans, like the Sargasso Sea, are limited by phosphorus
(333, 334). Growth limitation of bacteria by phosphorus
appears to be a common phenomenon of freshwater ecosys-
tems (335, 336).
N2 fixation is restricted to prokaryotes, and in the ocean
water column it was thought for many years to be done
primarily by Trichodesmium, a warm-water colonial cyanobac-
terium that blooms sporadically, and Richelia, a cyanobacterial
symbiont that lives within certain diatoms (337). Some uni-
cellular cyanobacteria also contribute significantly to global
N2 fixation (338), and molecular biological data suggest
that a variety of other bacteria—not phytoplankton—may
also be fixing nitrogen in seawater (339). An exciting recent
discovery in this field is that a bacterial nitrogen fixer origi-
nally known only from nifH sequences, UCYN-A, has had
its genome sequenced by parallel pyrosequencing, and was
found be a cyanobacterium (Candidatus Atelocyanobacterium
thalassa) lacking the apparatus to generate oxygen, as well as
other pathways (340). Further study showed it to be symbiotic
with a marine alga (341).
Micronutrients (Trace Metals, Growth Factors)
Bacterial growth may often be limited by macronutrients as
described, and sometimes by trace nutrients, particularly
iron, that occur at very low concentrations in seawater
(342, 343). The importance of Fe as an element limiting pri-
mary production in some oceanic regions has come under
close scrutiny and extensive experimental investigation in
recent years. Many bacteria have special uptake mechanisms
to utilize extremely low concentrations of iron, including
siderophores (released compounds that bind iron and are
then specifically taken up). Even if bacteria are capable of
growth at low Fe concentrations, growth may be less efficient.
Kirchman et al. (344) found that Fe limitation strongly
reduced bacterial growth efficiency in a cultured marine bac-
terium (ca. 50% efficiency Fe replete, but <10% when Fe
deplete). Thus the effects may be complex. Marine N2 fixa-
tion also may be limited by the availability of Fe because
nitrogen fixers have high Fe requirements.
Protozoa tend to be a source rather than a sink for Fe in the
ocean, analogous to their role in macronutrient remineraliza-
tion. Protozoan grazing activity has been shown to release
iron from bacterial biomass thereby relieving Fe limitation
in co-occurring phytoplankton (345), a mechanistic demon-
stration of the process using protozoa growing on Fe-replete
bacteria. However, it has also been reported that Fe-limited
bacteria may contain insufficient Fe for the bacterivorous pro-
tozoa that consume them, thus leading to Fe limitation in the
protozoan as well (346).
Oxygen
Hypoxic and anoxic regions of the ocean are expected to sig-
nificantly expand in the coming decades as a consequence of
changes in ocean stratification resulting from global climate
change (347). As oxygen becomes depleted, bacteria have
the capability to utilize a series of alternate electron acceptors.
Compositional changes in the bacterial assemblage accom-
pany these changes in metabolic activity, but are beyond
the scope of this chapter. Many of the studies on benthic
microbiology have focused on redox reactions that occur
there. Oxygen is usually consumed by sediment microbial
activity much faster than the rate at which it can be replaced
by diffusion, so most sediments, especially organically rich
ones, are anaerobic below a few mm (fine-grained muds) or
cm (coarser sediments) from the sediment/water interface,
except where animals ventilate the benthos through tubes.
When oxygen is absent, bacteria and archaea use alternate
electron acceptors, such as nitrate, nitrite, oxidized Fe or
Mn, organic matter, or sulfate. The use of nitrate or nitrite
as an electron acceptor, with concomitant production of N2
gas (denitrification) results in net loss of biologically available
nitrogen. Bacteria that perform this reaction are typically
4.2.2-20 ▪ AQUATIC ENVIRONMENTS

“facultative” organisms that can switch from oxygen to nitrate during incubations, relative to control trea