Handbook of SPR

Handbook of Surface Plasmon Resonance
To Femke and Nick, Margot and Niels

Handbook of Surface Plasmon
Resonance
Edited by
R.B.M. Schasfoort and Anna J. Tudos
University of Twente, Enschede, The Netherlands
ISBN: 978-0-85404-267-8
A catalogue record for this book is available from the British Library
r The Royal Society of Chemistry 2008
All rights reserved
Apart from fair dealing for the purposes of research for non-commercial purposes or for
private study, criticism or review, as permitted under the Copyright, Designs and Patents
Act 1988 and the Copyright and Related Rights Regulations 2003, this publication may not
be reproduced, stored or transmitted, in any form or by any means, without the prior
permission in writing of The Royal Society of Chemistry, or in the case of reproduction in
accordance with the terms of licences issued by the Copyright Licensing Agency in the UK,
or in accordance with the terms of the licences issued by the appropriate Reproduction
Rights Organization outside the UK. Enquiries concerning reproduction outside the terms
stated here should be sent to The Royal Society of Chemistry at the address printed on this
page.
Published by The Royal Society of Chemistry,
Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, UK
Registered Charity Number 207890
For further information see our web site at www.rsc.org
Foreword
Make no bones about it, I love surface plasmon resonance (SPR)-based bio-
sensor technology. After spending three years trying to measure binding con-
stants using analytical afnity chromatography, I immediately saw the benets
of SPR the rst time I sat down in front of a Biacore in 1991. Even today, no
comparable technology exists to characterize molecular interactions in real time
without labeling in an automated and robust fashion. But as the technology has
expanded over the years, I nd that there are three general attitudes towards
SPR. There are the nay-sayers who hate the technology. There are long-time
users who think they are experts. And there are the users who recognize they do
not know everything about SPR but are eager to improve their skills.
Ever since the rst commercially viable instrument was unleashed in 1990 by
the biosensor group at Pharmacia (which was spun out into a separate com-
pany called Biacore in 1996, only to be acquired recently by General Electric,
which previously bought Amersham who at one time had merged with
Pharmacia, so in fact now the biosensor group has come full circle, even
though they have always shared the same cafeteria in Uppsala, Sweden), there
have been critics of SPR technology. So much so that in 2003 I created a
character called Dr. Evil Pessimist, who represents a composite of the
various detractors of SPR. Dr. Pessimist rants and rages about problems he
has with the technology, including nonspecic binding, instrument drift, mass
transport and avidity effects. He argues that since SPR uses a surface the rate
constants we measure will never reect solution-based binding constants. In
fact, much of his resentment of the technology stems from the fact that his
experiments fail or his data never t a simple model.
It has been my experience that there are two primary causes of this SPRa-
phobia: poor-quality reagents and/or poor experimental design. Perhaps the
molecules Dr. Pessimist is studying do not in fact interact or the preparations of
samples are not active to begin with. Dont shoot the messenger. Dr. Pessimist
asserts that his proteins are of high quality because they are a single band on
an SDS page gel. He fails to realize that this is not evidence of an active
preparation or a conformationally homogeneous sample. I think biosensor
experiments are akin to protein crystallography. No structural biologist I know
would attempt to crystallize an impure, half-denatured preparation of protein
v
that has precipitated at the bottom of an Eppendor tube. The sad thing is
that garbage into a biosensor will often give complex responses that users
misinterpret as some interesting binding event.
I have found that when experiments are designed appropriately with good-
quality reagents and data are processed and analyzed properly, binding
responses can be routinely t to a simple interaction model. However, unlike
Dr. Pessimist, I do not expect to obtain perfect binding responses when I set up
experiments on a new interaction. I realize that obtaining high-quality data is
an iterative process. In my research group, we usually set up a trial experiment
to verify that the binding partners actually interact. Then we will often try
different coupling chemistries, surface densities, and/or buer conditions to
optimize surface activity.
And when it comes to the number one complaint about SPR technology (that
the surface will automatically change the thermodynamics of the system), what
Dr. Pessimist fails to realize is that most biosensor experiments do not use a at
surface. Instead, the surface is coated with a dextran layer, which suspends
the molecule in solution. We and others have shown with numerous systems
that when experiments are performed properly, binding constants (including
thermodynamic parameters) measured with SPR do in fact match those
obtained from solution-based measurements.
However, I agree with Dr. Pessimist in one regard. Since 1991, I have read
every paper that reported using a commercial SPR biosensor and Dr. Rebecca
Rich and I have composed a yearly review of the literature since 1998. This is
becoming a fairly daunting task since more than 1000 research papers are
published annually. More, unfortunately, is not always better. We nd that the
data in most biosensor articles are not worth the paper they are printed on. For
example, about half the time authors even fail to present gures showing the
binding responses and yet they expect us to believe the rate constants they
report for their interactions. Without a visual inspection of the data, we have
no idea if the experiments were run properly. And oftentimes, even when data
are presented, it is clear that the investigators do not know how to utilize the
technology properly. Also, while a fundamental dogma of science is to replicate
and randomize samples, less than 3% of published biosensor data include
replicate injections even within a single experiment. An overlay of replicate
injections demonstrates the stability of the reagents and multiple independent
experiments yield an average and standard deviation for the reported binding
constants, yet this attention to detail in a biosensor experiment is more rare
than nding a four-leaf clover in the outeld at Fenway Park. In addition, less
than 5% of the authors who report kinetic constants include an overlay of the
binding response with the tted model. And nally, even from a brief glance
through the literature, it is apparent that the majority of investigators do not
understand that the shape of the response prole should be an exponential in
both the association and dissociation phases (maybe many users do not even
understand what an exponential is). It is no wonder that scientists outside the
biosensor use community think SPR technology does not work. I would think
the same thing if all I had to rely on was the published data.
vi Foreword
You might ask yourself, how did it get to this point? I often wonder if all
scientists are so poorly educated in basic scientic technique (which could
actually explain why we havent found a cure for the common cold). I place the
blame on the kit mentality that was introduced with molecular biology back
in the early 1990s, back when we were listening to our Walkmans while typing
on our IBM 286 personal computers. Nowadays you can buy a kit to clone,
mutate, express and purify a protein. Well, the kit mentality continued when
these same investigators got access to commercially available biosensor tech-
nology. Since these instruments are so easy to use, anyone can walk up to the
machine, chuck in their proteins, collect some response, t the data and publish
the results, believing that the results must be correct because they came out of
this very expensive machine. Unfortunately, it actually takes some skill and
know-how to set up, execute and analyze a biosensor experiment properly.
This leads me to the next group of biosensor users that give the technology a
black eye. These people are the ones who have been using instruments for a
long time and think they are experts. I call them SPiRts. SPiRts are even
more threatening than Pessimists because their complacency often leads them
to perpetuate poor experimental technique. A common SPiRt mistake pub-
lished in the literature is the use of multivalent analytes in solution (e.g.
monoclonal antibodies or GST fusion proteins), which can produce avidity
effects. All too often, SPiRts present elaborate biological justications for the
shape of their unusual binding proles when in fact the responses are simply
indicative of poor reagent quality and/or inadequate experimental optimization
or data processing. Even worse, SPiRts use complex models to describe their
poor-quality data. It seems that the latest fad of these model surfers is to apply
a conformational change mechanism. My data t a conformational change
model, which must mean there is a conformational change, right? Wrong!
To set the record straight, in 1994 my colleague and software engineer
extraordinaire, Tom Morton (who I refer to as SoftEE), developed the numeri-
cal integration approach to data analysis that allows one to apply any
interaction model. Before then, we were in the caveman days of linear trans-
formation and, believe me, you dont want to go back there. We were the rst
to show that a change in conformation that stabilized a bound complex would
in fact produce a change in response even though there was no additional
change in mass. However, in the intervening 13 years I have never needed to
apply this model to describe the responses obtained from more than 1000
systems I have examined. The reason I am reluctant to use this model is that
typically a data set that ts a conformational change model can be equally well
described by other models such as those for ligand and/or analyte heterogene-
ity. Even more alarmingly, the rates for the supposed conformational changes
measured on the biosensor are extremely slow, often with half-lives of 2060
minutes if you take the time to calculate them. These rates do not make
biological sense to me. A quick search of the classical conformational change
literature shows that re-organizational events which occur during binding
happen on a nanosecond to millisecond time-scale. The hot new trend with
the SPiRts is to t their biosensor data with a conformational change model
vii Foreword
and then present crystal structure data of unbound and bound complexes and
say See, this change in conformation proves it. But an objective viewer would
disagree. The fact that you see a change in conformation in the structure
actually may not relate to the complex binding response you are measuring on
the biosensor. Dont be fooled by these sleight-of-hand arguments. (What
would help conrm the conformational change suggested by SPR would
actually be to use a time-resolved structural method such as circular dichroism
or uorescence resonance energy transfer and demonstrate that the time-
dependent changes are the same.) The cause of the complex binding response
on the biosensor is actually more likely due to surface aggregation, nonspecic
binding, molecular crowding, avidity effects or sample heterogeneity.
This brings me to my favorite SPR users, who I refer to as SPiRits. SPiRits
are new users or those having some familiarity of biosensor technology who
have a deep desire to learn more about its features, applications and potential.
They are the ones who are participating in our yearly benchmark studies, which
are geared toward calibrating users experimental technique. They are willing to
put in the effort to troubleshoot their systems and want to improve the quality
of the data and not just settle for whatever the machine spits out. SPiRits will
be the users who develop novel applications and implement new technologies in
the future.
We need SPiRits because the number and types of SPR instruments are
exploding. An Internet search reveals more than 20 companies developing
SPR-based biosensor systems. Lately, biosensor advances have occurred on
two fronts. First, many of the recently released instruments (and others
currently under development) are dedicated to specic applications ranging
from small-molecule drug discovery to the characterization of complex mix-
tures in the clinical and food sciences. Cornings Epic plate-based system is an
example of targeting the technology for screening applications. Second, we
have seen a push to increase the throughput of biosensor analyses. In the past
few years, the launches of BioRads ProteOn XPR36 and Biacores A100 have
dramatically impacted the biosensor eld since they allow for parallel process-
ing of multiple analytes over multiple targets simultaneously. Array-based
platforms represent the next wave in biosensor development. Biacores Flexchip
and instruments being developed by GWC Technologies, Lumera, IBIS Tech-
nologies, Genoptics and Maven open up the possibility of characterizing
hundreds to thousands of interactions at one time. But not surprisingly, these
array formats come with their own sets of challenges. The methods used for
spotting DNA may not be optimal for producing protein arrays. Clearly, a lot
of work remains to be done before protein array systems meet their full
potential.
As biosensor applications expand and new instruments are released, the
technologys user base also increases. I worry that higher-throughput systems
may allow more users simply to generate more bad data faster. So, we clearly
need to improve the skill level of both novice and seasoned users.
This book is a great resource to obtain the fundamental knowledge of
biosensor technology, and also discover recent developments in both
viii Foreword
instrumentation and applications. But in order to turn professional, remember
that the biosensor is just a tool. Use it wisely. Be skeptical, but keep an open
mind. Know when to say when (not all systems are amenable to biosensor
analysis). Go forth and become a good ShePaRd of my favorite technology.
David G. Myszka
University of Utah
ix Foreword
Preface
Lectori salutem,
Editors, authors often claim that completion of their book was a giant,
lonesome and tedious task. Often they add, their mission was once in a
lifetime. As they say in Germany with a sense of humor, Einstein
macht noch kein Haus
1
meaning that the cooperation of people is in the
heart of big achievements. So it has been with our book: all the authors had to
nd time in their busy life among other important engagements for
timely writing activities, for which we cannot say often enough how grateful
we are.
This Handbook of Surface Plasmon Resonance is the product of an
intensive interaction process and is intended for a wide audience: scientists and
students intending to use the technology, the wider public interested in
SPR as a phenomenon and its application, but also providers of (parts of) the
technology. Although the book as a whole covers many aspects of the technology
at present spanning a bridge between theory, instrumentation and applications,
the chapters are written so as to be comprehensible individually as well. It is
hoped that the readers of this book will share our enthusiasm for biomolecular
interaction analysis based on SPR technology. We also hope that we have
succeeded in revealing the potential of SPR by showing highly exciting and
unique opportunities for unraveling the functional relationships of complex
biological processes.
Special thanks are also due to the members of the Biochip Group of the
MESA+ Institute for Nanotechnology of the University of Twente who have
contributed to the book: Stefan Schlautmann and Hans de Boer for technical
support and some of the drawings. In addition, we thank Geert Besselink
Bianca Beusink, Angelique Lokate, Dietrich Kohlheyer, Ganesh Krishna-
moorthy, Dawid Zalewski, Remco Verdoold, Mayke van der Ploeg and Bjorn
Harink for their input. This devoted team provided the warm and inspiring
atmosphere of the Biochip Group during the two-year period from the birth of
the idea to completion of the manuscript.
1
Literally: even Einstein could not build a house and the German meaning of ein stein is one stone.
x
The editors would also like to thank Annie Jacob of the Royal Society of
Chemistry for her clear guidance and enduring patience throughout the
editorial process. Wout van Bennekom is acknowledged for nal reading of
several chapters.
Richard Schasfoort and Anna Tudos
Enschede and Amsterdam
xi Preface
Contents
Chapter 1 Introduction to Surface Plasmon Resonance
Anna J. Tudos and Richard B.M. Schasfoort
1.1 What is Surface Plasmon Resonance? 1
1.1.1 A Simple Experiment 1
1.1.2 From Dip to Real-time Measurement 2
1.2 How to Construct an SPR Assay? 3
1.2.1 The Steps of an Assay 4
1.2.2 Calibration Curve 6
1.2.3 Determination of Kinetic Parameters 7
1.2.4 Basics of Instrumentation 8
1.3 History of SPR Biosensors 9
1.3.1 Early History of SPR Biosensors 9
1.3.2 History of SPR Biosensors After 1990 11
1.4 How to Read This Book 12
1.5 Questions 13
References 13
Chapter 2 Physics of Surface Plasmon Resonance
Rob P.H. Kooyman
2.1 Introduction 15
2.2 The Evanescent Wave 16
2.3 Surface Plasmons 17
2.3.1 Surface Plasmon Dispersion Equations,
Resonance 17
2.3.2 Excitation of Surface Plasmons 19
2.3.3 Surface Plasmon Properties 21
2.3.4 Choice of Experimental Parameters 26
2.4 Analysis of Multi-layered Systems 27
2.5 SPR Spectroscopy 28
2.5.1 Enhancement of Fluorescence and Absorbance 28
2.5.2 SPR and Metal Nanoparticles 29
2.6 Concluding Remarks 31
xiii
2.7 Questions 32
2.8 Symbols 32
References 33
Chapter 3 SPR Instrumentation
Richard B.M. Schasfoort and Alan McWhirter
3.1 Introduction 35
3.1.1 From Surface Plasmon to SPR Signal 35
3.2 SPR Optics 38
3.2.1 Fan-shaped Beam 39
3.2.2 Fixed Angle 40
3.2.3 Angle Scanning 40
3.2.4 Grating Coupler 41
3.2.5 Other Optical Systems 42
3.2.6 SPR Imaging Instruments 43
3.2.7 General Optical Requirements
for SPR Instruments 44
3.3 SPR Liquid Handling Systems 45
3.3.1 Flow Cell Systems 45
3.3.2 Cuvette Systems 48
3.4 SPR Instruments: State of the Art 50
3.4.1 Examples of Fan-shaped Beam
SPR Instruments 50
3.4.2 Examples of Fixed-angle SPR Instruments 54
3.4.3 Examples of Angle Scanning SPR Instruments 59
3.4.4 Examples of Grating Coupler SPR Instruments 60
3.4.5 Examples of Other SPR Instruments 60
3.4.6 Examples of SPR Imaging Instruments 63
3.5 Protein Interaction Analysis Systems of Biacore 69
3.5.1 Introduction 69
3.5.2 Biacore T100 70
3.5.3 Biacore A100 73
3.5.4 FLEXChip 75
3.6 Conclusion 77
3.7 Questions 78
References 78
Chapter 4 Kinetic Models Describing Biomolecular Interactions
at Surfaces
Damien Hall
4.1 Introduction 81
4.1.1 Terminology of Adsorption 82
4.1.2 Optical Quantication of Adsorption
at an Interface 86
xiv Contents
4.2 Dening Factors of the Adsorption Event 87
4.2.1 Mass Transfer 90
4.2.2 Adsorption Mechanisms 98
4.2.2.1 Idealized Partition Processes 99
4.2.2.2 Eect of Competing Reactions 101
4.2.2.3 Surface Functions for Dierent
Modes of Adsorption 102
4.3 Summary and Conclusions 115
4.4 Questions 116
4.5 Symbols 118
4.6 Acknowledgements 119
References 120
Chapter 5 Kinetic and Thermodynamic Analysis of LigandReceptor
Interactions: SPR Applications in Drug Development
Nico J. de Mol and Marcel J.E. Fischer
5.1 Introduction 123
5.2 Anity and Kinetics of a Transport-limited
Bimolecular Interaction at the Sensor Surface 125
5.2.1 Anity Constants Derived from Equilibrium
SPR Signals 126
5.2.1.1 Correction for Depletion of Free
Analyte Concentration in the Cuvette 128
5.2.2 Anity Constants and Rate Constants
Derived from Kinetic Analysis 129
5.2.2.1 k
obs
Kinetic Analysis 130
5.2.2.2 Global Kinetic Analysis with a
Simple Bimolecular Binding Model 131
5.3 Detecting Mass Transport Limitation: A Practical
Approach 135
5.3.1 Eect of Viscosity Change on the Association
Phase 135
5.3.2 Transport Limitation in the Dissociation
Phase 137
5.3.2.1 Rebinding Model for Transport-
limited Dissociation 137
5.3.2.2 Competing Ligand to Prevent
Rebinding During
Dissociation 139
5.3.2.3 Experimental Procedure to Assay
High O-rates 141
5.3.3 Quantitative Considerations on Mass
Transport Limitation 142
5.3.3.1 Flow or Cuvette? 143
xv Contents
5.4 Global Kinetic Analysis of Complex Binding
Models 143
5.4.1 Global Kinetic Analysis Including Mass
Transport and a Conformational Change 144
5.4.2 Unusual Kinetics: Intermolecular Bivalent
Binding to the Sensor Surface 147
5.4.3 Global Kinetic Analysis: Concluding Remarks 152
5.5 Anity in Solution Versus Anity at the Surface 154
5.6 Thermodynamic vant Ho Analysis Using SPR Data 158
5.6.1 vant Ho Thermodynamic Analysis 158
5.6.2 Comparison of SPR Thermodynamics
with Calorimetry 161
5.6.3 Transition State Analysis Using Eyring Plots 163
5.7 SPR Applications in Pharma Research: Concluding
Remarks and Future Perspectives 165
5.8 Questions 167
5.9 Symbols 168
References 169
Chapter 6 Surface Chemistry in SPR Technology
Erk T. Gedig
6.1.1 General Aspects of Surfaces for Biomolecular
Interaction Analysis 174
6.1.2 Selection of the Optimal Surface 177
6.2 Adhesion Linking Layers for Gold, Glass and
Plastics 181
6.2.1 Adhesion Linking Layers for Metal Surfaces 182
6.2.2 Adhesion Linking Layers for Inorganic
Dielectrics 182
6.2.3 Adhesion Linking Layers for Plastics 183
6.3 Bioinert Matrices 183
6.3.1 Non-specic Adsorption of Biomolecules 183
6.3.2 Bioinert Hydrogels 185
6.4 Choosing the Optimal Nanoarchitecture 187
6.4.1 Two-dimensional Surfaces 189
6.4.2 Three-dimensional Hydrogels 191
6.5 Coupling Procedures for Ligand Immobilization 194
6.5.1 Adsorptive Immobilization 195
6.5.2 Preconcentration Methods Prior to Covalent
Immobilization 195
6.5.2.1 Electrostatic Preconcentration 195
6.5.2.2 Dry Immobilization 197
xvi Contents
6.5.3 Covalent Activation Chemistries 199
6.5.3.1 Amine Coupling via Reactive Esters 199
6.5.3.2 Amine Coupling Through Reductive
Amination 202
6.5.3.3 Thiol Coupling 203
6.5.3.4 Immobilization of Aldehydes
Through Hydrazide Groups 207
6.5.3.5 Coupling Through Epoxy Groups 208
6.5.4 Electrostatic Methods 210
6.5.5 Directed Immobilization 212
6.5.6 Immobilization of Membrane Proteins 213
6.6 Conclusions and Outlook 216
6.7 Questions 217
References 218
Chapter 7 Measurement of the Analysis Cycle: Scanning SPR
Microarray Imaging of Autoimmune Diseases
Richard B.M. Schasfoort, Angelique M.C. Lokate, J. Bianca
Beusink, Ger J.M. Pruijn and Gerard H.M. Engbers
7.2 The Analysis Cycle 222
7.3 Buer Solutions for Measuring the Analysis
Cycle 224
7.3.1 Baseline Buer 224
7.3.2 Regeneration Solution 225
7.4 SPR-based Immunoassays 225
7.4.1 Direct Assay 226
7.4.2 Competition Assay 226
7.4.3 Inhibition Assay 227
7.4.4 Sandwich Assay 228
7.5 Detection of Multiplex Analysis Cycles Using
Scanning SPR Imaging 228
7.5.1 Dynamic Range of Scanning SPR Imaging 230
7.5.2 Liquid Handling Procedures 232
7.5.3 Determination of the Limit of Detection Using
Multiplex Analysis Cycles 232
7.6 Monitoring of Autoantibodies in Serum
of Rheumatoid Arthritis Patients 235
7.6.1 Experimental Conditions for Serum
Measurements 235
7.6.1.1 Serum Samples 235
7.6.1.2 SPR Microarray Interaction Studies 236
7.6.2 Results and Discussion of Monitoring Analysis
Cycles for Autoantibody Screening 236
xvii Contents
7.7 Features and Benets of Monitoring Analysis Cycles
with SPR Imaging 241
7.8 Conclusion 242
7.9 Questions 243
7.10 Acknowledgement 243
References 243
Chapter 8 Advanced Methods for SPR Imaging Biosensing
Alastair W. Wark, Hye Jin Lee and Robert M. Corn
8.2 Advances in SPRI Instrumentation and Surface
Chemistry 247
8.3 Surface Enzymatic Transformations for Enhanced
SPRI Biosensing 254
8.3.1 Measuring Surface Enzyme Kinetics 254
8.3.2 RNase HAmplied Detection of DNA 257
8.3.3 Fabrication of RNA Microarrays with
RNA-DNA Surface Ligation Chemistry 259
8.4 Nanoparticle-amplied SPRI Biosensing 260
8.4.1 Single Nucleotide Polymorphism Genotyping 262
8.4.2 MicroRNA Detection 264
8.5 Summary and Outlook 269
8.6 Questions 270
References 271
Chapter 9 Surface Plasmon Fluorescence Techniques
for Bioafnity Studies
Wolfgang Knoll, Amal Kasry, Jing Liu, Thomas Neumann,
Lifang Niu, Hyeyoung Park, Harald Paulsen,
Rudolf Robelek, Danfeng Yao and Fang Yu
9.2 Surface Plasmon Fluorescence Spectroscopy (SPFS) 278
9.3 Interface Kinetics Based on the Langmuir Adsorption
Model 282
9.3.1 Mass Transport-controlled Kinetics 284
9.3.2 Interaction-controlled Kinetics 284
9.3.3 Equilibrium Analysis 286
9.4 Applications of the Kinetic Model 286
9.4.1 Surface Hybridization Reactions
of Oligonucleotides 286
9.4.2 Protein Binding Studies 298
9.5 Novel Approaches to SPFS 300
xviii Contents
9.5.1 Grating Coupling for SPFS 301
9.5.2 Long-range Surface Plasmons for SPFS 303
9.5.3 Fluorescence Imaging and Color Multiplexing 306
9.6 Conclusions 309
9.7 Questions 309
References 310
Chapter 10 SPR Imaging for Clinical Diagnostics
Elain Fu, Timothy Chinowsky, Kjell Nelson and Paul Yager
10.2 Achieving Miniaturization and Low Cost 314
10.2.1 Tuning in SPR Imager Design 315
10.2.2 Compact Design of Additional SPR Imager
Optical Elements 315
10.3 Optimizing Imager Performance 317
10.3.1 Refractive Index Resolution 317
10.3.2 Lateral Resolution Over the Field of View 320
10.4 Robust Operation 321
10.4.1 Effects of Temperature Fluctuations 321
10.4.2 Strategies to Alleviate the Effects
of Temperature Fluctuations 322
10.4.3 Bulk RI Compensation 323
10.5 SPR Imaging Assays 324
10.5.1 Microuidic Immunoassay Design for Small
Molecule Analytes 324
10.5.2 Assay Compatibility with Complex
Samples 327
10.5.3 Assay Implementation 329
10.6 Conclusion 329
10.7 Questions 330
References 330
Chapter 11 The Benets and Scope of Surface Plasmon Resonance-based
Biosensors in Food Analysis
Alan McWhirter and Lennart Wahlstrom
11.1.1 Why Analyze Food? 334
11.1.2 Food Analysis Steps and SPR Assay
Formats 334
11.2 Biacore Q and Qex Kits the Workhorse of Food
Analysis 338
11.2.1 Qex Kits for Screening Veterinary Drug
Residues 339
xix Contents
11.2.2 Qex Kits for Quantifying Vitamin Content 343
11.2.3 AOAC Certication of Qex Kits 344
11.3 Examples of Applications for SPR-based Biosensors
in Food Analysis 345
11.3.1 Quantifying Antibiotics in Honey 345
11.3.2 Screening for Veterinary Drug Residues 346
11.3.3 Milk Testing 347
11.3.4 Detecting Antibodies to Salmonella in Meat 350
11.3.5 Genetically Modied Organisms 350
11.4 Conclusions 351
11.5 Questions 352
References 352
Chapter 12 Future Trends in SPR Technology
Richard B.M. Schasfoort and Peter Schuck
12.2 Trends in SPR Instrumentation 355
12.2.1 SPR Imaging 356
12.2.2 Hyphenation SPR Technology 356
12.2.2.1 SPRMS 357
12.2.2.2 Other Hyphenated SPR
Techniques 359
12.2.3 Nanoparticle SPR 360
12.3 Trends in Fluidics 361
12.3.1 Microarray Spotting on Gold 362
12.3.1.1 DNA Coding Technology 364
12.3.2 Prospects for SPR-based Point of Care
Devices 366
12.3.3 Implementation of Lab-on-a-Chip Devices
for SPR Systems 367
12.3.3.1 Pumping Liquids Using
Electroosmotic Flow in Micro-
uidic Devices with Gold Layers 367
12.3.4 Lab-on-a-Chip Implementation Using Free
Flow Electrophoresis and SPR Imaging
for Proteomics-on-a-Chip 369
12.3.5 Digital Microuidics 373
12.3.5.1 Cell Diagnosis and Monoclonal
Antibody Screening Using SPR
Imaging and Digital Microuidics 375
12.4 Trends in Sensor Surfaces 376
12.4.1 Smart Polymer Brushes 376
12.4.2 Photoactivation of Surfaces
for Immobilization 378
12.4.3 Gradient Chemistries 380
xx Contents
12.5 Trends in Measuring Reliable Kinetic Parameters 381
12.5.1 Introduction 381
12.5.2 The Model for Distribution Analysis of Rate
and Equilibrium Constants 383
12.5.3 Examples of the Distribution Analysis
Method 385
12.5.4 Conclusions and Perspectives of the
Distribution Analysis Model 388
12.6 Final Comments 391
12.7 Questions 391
References 392
Subject Index 395
xxi Contents
CHAPTER 1
Introduction to Surface Plasmon
Resonance
ANNA J. TUDOS
a
AND RICHARD B.M. SCHASFOORT
b
a
Shell Global Solutions International BV, P.O. Box 38000
1030 BN Amsterdam, The Netherlands;
b
Biochip Group, MESA+ Institute
for Nanotechnology, Biomedical Technology Institute (BMTI), Faculty of
Science and Engineering, University of Twente, P.O. Box 217,
7500 AE Enschede, The Netherlands
1.1 What is Surface Plasmon Resonance?
Since its rst observation by Wood in 1902 [1,2], the physical phenomenon of
surface plasmon resonance (SPR) has found its way into practical applications in
sensitive detectors, capable of detecting sub-monomolecular coverage. What is
surface plasmon resonance? Wood observed a pattern of anomalous dark and
light bands in the reected light, when he shone polarized light on a mirror with a
diraction grating on its surface. Physical interpretation of the phenomenon was
initiated by Lord Rayleigh [3], and further rened by Fano [4], but a complete
explanation of the phenomenon was not possible until 1968, when Otto [5] and in
the same year Kretschmann and Raether [6] reported the excitation of surface
plasmons. Application of SPR-based sensors to biomolecular interaction mon-
itoring was rst demonstrated in 1983 by Liedberg et al. [7]. A historical overview
of the use of the phenomenon for biosensor applications is given in Section 1.3 of
this chapter. To understand the excitation of surface plasmons, let us start with a
simple experiment.
1.1.1 A Simple Experiment
Consider the experimental set-up depicted in Figure 1.1. When polarized light is
shone through a prism on a sensor chip with a thin metal lm on top, the light
will be reected by the metal lm acting as a mirror. On changing the angle of
incidence, and monitoring the intensity of the reected light, the intensity of the
1
reected light passes through a minimum (Figure 1.1, line A). At this angle of
incidence, the light will excite surface plasmons, inducing surface plasmon
resonance, causing a dip in the intensity of the reected light. Photons of
p-polarized light can interact with the free electrons of the metal layer, inducing
a wave-like oscillation of the free electrons and thereby reducing the reected
light intensity.
The angle at which the maximum loss of the reected light intensity occurs is
called resonance angle or SPR angle. The SPR angle is dependent on the optical
characteristics of the system, e.g. on the refractive indices of the media at both
sides of the metal, usually gold. While the refractive index at the prism side is
not changing, the refractive index in the immediate vicinity of the metal surface
will change when accumulated mass (e.g. proteins) adsorb on it. Hence the
surface plasmon resonance conditions are changing and the shift of the SPR
angle is suited to provide information on the kinetics of e.g. protein adsorption
on the surface.
1.1.2 From Dip to Real-time Measurement
Surface plasmon resonance is an excellent method to monitor changes of the
refractive index in the near vicinity of the metal surface. When the refractive
index changes, the angle at which the intensity minimum is observed will shift
as indicated in Figure 1.2, where (A) depicts the original plot of reected light
intensity vs. incident angle and (B) indicates the plot after the change in
refractive index. Surface plasmon resonance is not only suited to measure the
dierence between these two states, but can also monitor the change in time, if
one follows in time the shift of the resonance angle at which the dip is observed.
A

Angle ()
B
Intensity of
reflected light
(%)
Figure 1.1 Schematic experimental set-up of surface plasmon resonance excitation. A

sensor chip with a gold coating is placed on a hemisphere (or prism).
Polarized light shines from the light source (star) on the sensor chip.
Reected light intensity is measured in the detector (disk). At a certain
angle of incidence (j), excitation of surface plasmons occurs, resulting in a
dip in the intensity of the reected light (A). A change in refractive index at
the surface of the gold lm will cause an angle shift from A to B.
2 Chapter 1
Figure 1.2 depicts the shift of the dip in time, a so-called sensorgram. If this
change is due to a biomolecular interaction, the kinetics of the interaction can
be studied in real time.
SPR sensors investigate only a very limited vicinity or xed volume at the
metal surface. The penetration depth of the electromagnetic eld (so-called
evanescent eld) at which a signal is observed typically does not exceed a few
hundred nanometers, decaying exponentially with the distance from the metal
layer at the sensor surface. The penetration depth of the evanescent eld is a
function of the wavelength of the incident light, as explained in Chapter 2.
SPR sensors lack intrinsic selectivity: all refractive index changes in the evanes-
cent eld will be reected in a change of the signal. These changes can be due to
refractive index dierence of the medium, e.g. a change in the buer composition
or concentration; also, adsorption of material on the sensor surface can cause
refractive index changes. The amount of adsorbed species can be determined after
injection of the original baseline buer, as shown in Figure 1.2. To permit selective
detection at an SPR sensor, its surface needs to be modied with ligands suited for
selective capturing of the target compounds but which are not prone to adsorbing
any other components present in the sample or buer media.
1.2 How to Construct an SPR Assay?
Now we have a basic understanding of the surface plasmon resonance signal and
how to measure it in time. We know that the sensor surface needs to be modied
to allow selective capturing and thus selective measurement of a target compound.
Angle
A
B
Time (s)
Figure 1.2 A sensorgram: the angle at which the dip is observed vs. time. First, no
change occurs at the sensor and a baseline is measured with the dip at SPR
angle (A). After injection of the sample (arrow) biomolecules will adsorb on
the surface resulting in a change in refractive index and a shift of the SPR
angle to position B. The adsorptiondesorption process can be followed in
real time and the amount of adsorbed species can be determined.
3 Introduction to Surface Plasmon Resonance
In the following, we are going to learn more about an SPR measurement. First, the
steps of an SPR assay will be discussed from immobilization through analysis to
regeneration in a measurement cycle. Next, we get acquainted with a typical
calibration curve, followed by examples of assay formats. Finally, a short outlook
is provided on the basics of the instrumentation.
1.2.1 The Steps of an Assay
In the simplest case of an SPR measurement, a target component or analyte is
captured by the capturing element or so-called ligand (Figure 1.3). The ligand is
permanently immobilized on the sensor surface previous to the measurement.
Various sensor surfaces with immobilized ligands are commercially available,
and many more can be custom-made, as explained in Chapters 6 and 7.
In the simplest case, the event of capturing the analyte by the ligand gives rise
to a measurable signal, this is called direct detection. Figure 1.4 shows the
sensor signal step-by-step in the measurement cycle with direct detection.
Each measurement starts with conditioning the sensor surface with a suitable
buer solution (1). It is of vital relevance to have a reliable baseline before the
capturing event starts. At this point, the sensor surface contains the active
ligands, ready to capture the target analytes. On injecting the solution containing
Glass
Gold 50 nm
Flow of sample
with analyte
Bound ligand Y
Figure 1.3 Schematic representation of direct detection: the analyte is captured by the
ligands (Y) immobilized on the sensor surface. Accumulation of the
analyte results in a refractive index change in the evanescent eld shifting
the SPR angle. Here the ligand is immobilized in a hydrogel.
4 Chapter 1
the analytes (2), they are captured on the surface. Also other components of the
sample might adhere to the sensor surface; without a suitable selection of the
ligand, this adherence will be non-specic, and thus easy to break. At this step,
adsorption kinetics of the analyte molecule can be determined in a real-time
measurement. Next, buer is injected on to the sensor and the non-specically
bound components are ushed o (3). As indicated in the gure, the accumulated
mass can be obtained from the SPR response (DR). Also in this step, dissociation
of the analyte starts, enabling the kinetics of the dissociation process to be
studied. Finally, a regeneration solution is injected, which breaks the specic
binding between analyte and ligand (4). If properly anchored to the sensor
surface, the ligands remain on the sensor, whereas the target analytes are
quantitatively removed. It is vital in order to perform multiple tests with the
same sensor chip to use a regeneration solution which leaves the activity of the
ligands intact, as the analysis cycle is required to take place repeatedly for
hundreds, sometimes even thousands of times. Again, buer is injected to
condition the surface for the next analysis cycle. If the regeneration is incomplete,
remaining accumulated mass causes the baseline level to be increased.
Often SPR measurements are carried out to determine the kinetics of a
binding process. For realistic results it is vital to prevent immobilization from
Step:
Time
Injection
R
SPR-
dip
shift
t
1
2. association 1. baseline 3. dissociation 4. regeneration 1. baseline
Figure 1.4 Sensorgram showing the steps of an analysis cycle: 1, buer is in contact
with the sensor (baseline step); 2, continuous injection of sample solution
(association step); 3, injection of buer (dissociation step); DR indicates
the measured response due to the bound target compound; 4, removal of
bound species from the surface during injection of regeneration solution
(regeneration step) followed by a new analysis cycle. A bulk refractive
index shift can be observed at t
1
. See also page 222.
changing the ligand in a way that would inuence its strength or anity
towards the target component. In addition, kinetic experiments can provide
information on the thermodynamics, e.g. on the binding energy of processes. A
description of the kinetic theory can be found in Chapter 4 and examples of
kinetic studies in Chapter 5.
1.2.2 Calibration Curve
Apart from kinetic and thermodynamic studies, SPR measurements can also be
used for the determination of the concentration of the analyte in a sample
(quantitative analysis). In this case, rst dierent concentrations of the analyte
are applied in separate analysis cycles. The sensorgrams measured at dierent
concentrations give an overlay plot similar to that depicted in Figure 1.5, with the
plateaus of the association step increasing at increasing analyte concentration [8].
A calibration curve can be constructed by plotting the response (DR) after a
certain time interval (t
1
) versus concentration.
When analyzing samples with an unknown concentration of the analyte,
usually multiple dilutions are made, for example 10, 100 and 1000 times, or for
more accurate determinations serial dilutions by a factor of 2. If the concentration
time
t
1
dR/dt
t
0
R
S
P
R
-
d
i
p

s
h
i
f
t

Step: 1. baseline 2. association 3. dissociation 4. regeneration 1. baseline
Injection
Figure 1.5 Typical overlay plot of sensorgrams from serial diluted analyte concen-
trations. Just after injection at t
0
a sample specic binding of the analyte
occurs and mass transport to the surface is rate limiting and linearly
dependent on the concentration. From the slopes of a positive control
(dR/dt), the concentration of an unknown sample can be determined.
During the association phase the number of unbound ligand molecules
decreases and dissociation takes place. The o-rate constant or dissocia-
tion constant (k
d
) can be determined after injecting dissociation buer at
t
1
. See for more details chapter 4 and 5 of this book.
6 Chapter 1
of the analyte in the sample is very high, the undiluted sample will yield results on
the upper plateau range of the calibration curve. The diluted solutions, however,
might yield points along the lower, concentration-dependent sections of the
calibration curve and the concentration of the analyte can be determined.
As mentioned above, SPR sensing means detection of refractive index
changes at the sensor surface, which in practice translates to the amount of
mass deposited at the sensor surface. Direct detection is only possible if the
capturing event of the analyte brings about measurable refractive index
changes. This is easier to achieve if the molecular weight of the analyte is high
(i.e. around 1000 Da or higher). However, for small molecules to produce a
measurable refractive index change, large numbers would be required, making
the analysis intrinsically less sensitive. If the analyte is a small molecule
(MWo1000 Da), often direct detection is not viable.
Detection of small molecules can be carried out using a dierent strategy.
Most often, small molecules are detected in a sandwich, competition or inhi-
bition assay format. In all assay formats, not only the lower detectable
concentration is limited, but also the physical number of immobilized elements
on the sensor surface, which provides a maximum limiting value. Discussion of
the dierent assay formats can be found in Chapter 7 and other methods for
concentration determination are described in Chapters 4 and 5.
1.2.3 Determination of Kinetic Parameters
The most prominent benet of direct detection using SPR biosensor technology
is the determination of kinetics of (bio)molecular interactions. Reaction rate
and equilibrium constants of interactions can be determined, e.g. the interac-
tion A+B-AB can be followed in real time with SPR technology, where A is
the analyte and B is the ligand immobilized on the sensor surface.
Table 1.1 contains the most relevant kinetic parameters, the association and
dissociation constants, for the simplest case A+B-AB. The association
constant is the reaction rate of complex (AB) formation, giving the number of
complexes formed per time at unit concentration of A and B. As soon as the
complex AB is formed, its dissociation can commence. The dissociation
rate constant describing this process expresses the number of AB complexes
Table 1.1 Denitions of the most relevant kinetic parameters: the association
and dissociation constants.
Association rate constant, k
a
Dissociation rate constant, k
d
Denition A+B-AB AB-A+B
Description Reaction rate of AB
formation: number of AB
complexes formed per
unit time at unit
concentration of A and B
Dissociation rate of AB:
number of AB complexes
dissociating per unit time
Units l mol
1
s
1
s
1
Typical range 10
3
10
7
10
1
5 10
6
dissociating per unit time. Note that the unit dimensions for the association and
dissociation rates are dierent and can vary with the stoichiometry of the
complex. The typical range of the association and dissociation constant shows
large variations and is dependent on, among other things, the temperature.
When association of A and B starts, no product is yet present at the sensing
surface. At this point, the rate of the association reaction is highest and that of
the dissociation reaction is lowest. As the process progresses, more and more of
the AB complex is produced, enhancing the rate of dissociation. Due to
decreasing A and B concentration, the rate of association might decrease.
Equilibrium is reached when the rates of the association and dissociation
reactions are equal; the denitions and unit dimensions are given in Table 1.2.
As can be seen, the equilibrium association and dissociation constants, which
represent the anity of an interaction, have a reciprocal relationship with each
other. The eect of parameters such as temperature is described in later chapters.
The rate constants (Table 1.1) and equilibrium constants (Table 1.2) of
(bio)molecular interactions provide information on the strength of association
and the tendency of dissociation. Various aspects of kinetics, models and
calculation of anity constants are described in Chapters 4, 5 and 9.
1.2.4 Basics of Instrumentation
Studying biomolecular interactions using SPR does not require a detailed
understanding of the physical phenomena. It is sucient to know that SPR-
based instruments use an optical method to measure the refractive index near a
sensor surface (within B200 nm to the surface). SPR instruments comprise three
essential units integrated in one system: optical unit, liquid handling unit and the
sensor surface. The features of the sensor chip have a vital inuence on the
quality of the interaction measurement. The sensor chip forms a physical barrier
between the optical unit (dry section) and the ow cell (wet section).
SPR instrumentation can be congured in various ways to measure the shift of
the SPR-dip. In general, three dierent optical systems (Chapter 2) are used to
excite surface plasmons: systems with prisms, gratings and optical waveguides.
Most widespread are instruments with a prism coupler, also called Kretschmann
conguration [9]. In this conguration, which is shown in Figure 1.1, a prism
couples p-polarized light into the sensor coated with a thin metal lm. The light is
Table 1.2 Denition of the equilibrium association and dissociation constants.
Equilibrium association
constant, K
A
Equilibrium dissociation
constant, K
D
Denition [AB]/[A][B] k
a
/k
d
[A][B]/[AB] k
d
/k
a
Description Anity to association:
high K
A
, high anity to
associate
Stability of AB: high K
D
,
low stability of AB
Unit l mol
1
mol l
1
Typical range 10
5
10
12
10
5
10
12
8 Chapter 1
reected on to a detector, measuring its intensity, using a photodiode or a camera.
In instruments with a grating coupler [10], light is reected at the lower refractive
index substrate. In practice, this means that light travels through the liquid before
photons generate surface plasmon waves as in ellipsometric instruments [11].
Besides the grating couplers, some instruments apply optical waveguide couplers
[12] or measure the SPR wavelength shift as a result of the biomolecular
interaction process (see Chapter 2 and ref. [13]).
All congurations share the same intrinsic phenomenon: the direct, label-free
and real-time measurement of refractive index changes at the sensor surface.
SPR sensors oer the capability of measuring low levels of chemical and
biological compounds near the sensor surface. Sensing of a biomolecular
binding event occurs when biomolecules accumulate at the sensor surface
and change the refractive index by replacing the background electrolyte.
Protein molecules have a higher refractive index than water molecules
(Dn E10
1
). The sensitivity of most SPR instruments is in the range Dn E10
5
or 1 pg mm
2
of proteinous material. Often in real-time biosensing absolute
values are not a prerequisite, only the change is monitored as a result of
biospecic interaction at the sensor surface. A detailed description of commer-
cial instruments is given in Chapter 3.
1.3 History of SPR Biosensors
The term biosensor was introduced around 1975, relating to exploiting trans-
ducer principles for the direct detection of biomolecules at surfaces. Currently the
most prominent example of a biosensor is the glucose sensor, reporting glucose
concentration as an electronic signal, e.g. based on a selective, enzymatic process.
Some argued that all small devices capable of reporting parameters of the human
body were biosensors (e.g. ion-sensitive eld-eect transistors (ISFETs) measur-
ing pH). But then, a thermometer recording fever should also be called biosensor.
According to the present denition, in biosensors the recognition element
(ligand) of the sensor or the analyte should originate from a biological source.
Biosensors are analytical devices comprised of a biological element (tissue,
microorganism, organelle, cell receptor, enzyme, antibody) and a physicochemi-
cal transducer. Specic interaction between the target analyte and the biological
material produces a physico-chemical change detected by the transducer. The
transducer then yields an analog electronic signal proportional to the amount
(concentration) of a specic analyte or group of analytes.
1.3.1 Early History of SPR Biosensors
Application of SPR-based sensors to biomolecular interaction monitoring
was rst demonstrated in 1983 by Lundstroms pursuit towards physical
methods for label-free, real-time detection of biomolecules [7]. The intrinsic
properties of the molecules, e.g. mass, refractive index and/or charge distri-
bution [14], were probed using ellipsometry, refractometry, surface plasmon
resonance, photothermic detection methods and others. At the National
Defense Research Laboratory of Sweden, proteinprotein interactions were
monitored in real time, label-free, using ellipsometry. Most importantly, the
refractive index change at a light-reecting surface was the operating trans-
ducer mechanism. Although successful in the detection of refractive index
change due to the binding of biomolecules on optical transducer surfaces, a
disadvantage of the ellipsometer is that light passes through the bulk of the
sample solution, hence light-absorbing or particle-containing samples cannot
easily be measured.
Among other research laboratories in the same period, the University of
Twente (The Netherlands) was active in the search for nding new transduction
principles for measuring immunochemical reactions at eld eect transistor
devices (ImmunoFET) [15] and at surfaces with an optical read-out (immuno-
chemical optical biosensor, IMOB). Optical transducer principles [16] including
ellipsometry, surface plasmon resonance and interferometric principles (Mach
Zehnder) showed promise for direct transduction of the biomolecular binding
event. Successful measurements of immunochemical reactions using SPR were
carried out as early as in the mid-1980s [17].
Pharmacia Biosensor AB chose SPR as their platform technology for direct
sensing of biomolecular interactions. The Kretschmann conguration oered
advantages in freedom of design of the liquid handling system. Coming from
the higher refractive index medium (the prism), light does not pass through the
liquid, but is reected at the sensor surface covered with a thin metal layer.
Gold was chosen as the best inert metal lm required for surface plasmon
resonance, although from a physical point of view silver provides a better SPR
eect (see Chapter 2).
Studies on the surface chemistry led to modication of the gold with a self-
assembling layer of long-chain thiols to which a hydrogel could be attached.
Carboxylated dextran was immobilized at the surface, which provides a subst-
rate for ecient covalent immobilization of biomolecules, in addition to a
favorable environment for most biomolecular interactions. The thickness of the
dextran hydrogel of 100 nm is perfectly compatible with the ca. 200 nm
evanescent eld (see Figure 1.3). The reliable production of these high-quality
sensor chips was unequivocally the basis for the successful launch of SPR
instruments.
Techniques were developed to etch silica to form a casting mold for the
manufacture of microuidic ow channels. Also, development proceeded on
optogels for use between the prism in the optical unit of the instrument and the
sensor chip. The optogel ensures optical contact with the prism, allowing simple
replacement of the sensor chip. These eorts in research and development relied
on the combination of three unrelated elds: optics, microuidics and surface
chemistry, and resulted in the successful development of the instrumental
concept of biomolecular interaction analysis (BIA).
10 Chapter 1
1.3.2 History of SPR Biosensors After 1990
In 1990, Pharmacia Biosensor AB launched the rst commercial SPR product,
the Biacore instrument [18]. The instrument was the most advanced, sensitive,
accurate, reliable, reproducible direct biosensor technique and SPR became
(and still is) the golden standard of transducer principles for measuring real-
time biomolecular interactions. Since the early 1990s, producers have been
struggling to meet the standards set by Biacore. Fisons Instruments
1
[19] made
serious attempts to compete with Biacores technology; their cuvette-based
IAsys instrument uses evanescent eld-based technology, essentially not SPR,
for the study of biomolecular interactions.
The Biacore 2000 instrument was introduced in 1994 with improved detec-
tion and a dierent ow system so that the sample could interact at four spots
on the sensor. Data of the reference spot could be used for signal correction.
With the introduction of Biacore 2000 it also became possible to monitor
directly interactions of small molecule analytes reacting with immobilized
protein ligands [20].
In 1995, the cuvette based SPR system of IBIS Technologies was launched.
The instrument was compatible with the Biacore sensor chip. In 1997, the IBIS II,
a two-channel cuvette-based SPR instrument with autosampler operation, was
introduced [21]. Following the merger with the sensor chip coating company
Ssens BV in 1999, the development of an SPR imaging instrument was initiated at
IBIS Technologies. In 2007, the development of the IBIS-iSPR instrument, with
the scanning angle principle, resulted in the required reliability and accuracy for
microarray imaging of multiple biomolecular interactions (4500). The potency
of the instrument is demonstrated in Chapter 7.
Biacore X, a two-spot instrument introduced in 1996, was followed by the
Biacore 3000 in 1998. The latter was later extended with recovery tools to
improve interfacing with mass spectrometry [22]. Biacore Q was introduced for
the food analysis market in 2000 (Chapter 11). Positioned for small molecule
analysis and drug discovery, the introduction of the Biacore S51 marked a
technology shift in terms of detection, ow cell design and sample capacity: the
area of the detected spot was reduced from 1 to 0.01 mm
2
and the number of
spots was increased from four to six. In 2004, a high-end instrument was
introduced with four channels each with ve sensor spots (Biacore A100).
Combining the ow cell of the Biacore S51 instrument and the performance of
the four-channel Biacore 3000, this instrument has 20 in-line sensors to monitor
biomolecular interactions in the ow cells. The technology is not suitable,
however, to image the surface. In Chapter 3, other Biacore instruments (T100
and X100) are described. In order to measure up to 400 interactions simulta-
neously, in 2005 Biacore acquired the grating coupler SPR system of HTS
Biosystems, co-developed with Applied Biosystems (8500 Anity Analyzer),
which was capable of imaging the sensor surface. After restyling, this product
(named Flexchip) was launched in 2006 [10].
1
Later Anity Sensors.
Although it is impossible to describe accurately the history of the develop-
ments of the 25 companies producing SPR (related) instruments (see Chapter 3),
it is justied to treat the history along the Biacore product line. During the
years following the introduction of the rst SPR instrument, detection sensi-
tivity has improved by roughly 20-fold. The range of anity and kinetic data
that can be determined has been extended at least 100-fold as a consequence of
the increased sensitivity and due to improvements in data analysis. The amount
of independent sensor surfaces grew from four channels in 1990 (Biacore) to at
least 500 in the new IBIS SPR imaging instrument. The carboxymethylated
dextran surface introduced in 1990 [23], still the rst choice for many applica-
tions, has been complemented with a range of other surfaces. Systems for
dedicated applications have been introduced by various manufacturers as
complements to all-purpose research instrumentation [24]. A good gauge of
the success of biosensor technology is that more than 1000 publications each
year include data collected from commercial biosensors. In the paper entitled
Survey of the 2005 commercial optical biosensor literature, Rich and Myszka
[25] gave an outstanding overview of the SPR literature, including practical
lessons in performing and interpreting biomolecular interaction analysis ex-
periments. The majority of the publications (985) in 2005 employed Biacore
technology (87%), indicating the relevance of Biacores technology in the
market. Anity Sensors was the runner-up company with 40 publications
(B4%), Eco Chemie/Windsor Scientic (distributor) totaled 18 publications.
which was essentially from the same technology provider (originally IBIS
Technologies), Texas Instruments scored 17 publications in 2005 and 60
publications (B6%) were attributed to 13 other companies.
With the introduction of a number of new SPR instruments (Chapter 3) and
a series of novel sensor surfaces and chemistries, the impact of SPR biosensors
on molecular interaction studies will continue to grow. With improved exper-
imental design, including SPR imaging instruments and advanced data analysis
methods, high-quality data for the determination of kinetic parameters of
biomolecular interaction phenomena can be obtained. These data promise
additional insights into the mechanisms of molecular binding events, which will
be important for functionregulatory protein interaction studies in order to
unravel the exciting processes in living species.
1.4 How to Read This Book
Although most chapters can be read as stand-alone literature on dierent
aspects of SPR technology, this handbook aims to provide the reader with a
total coverage of the basics of the technique and applications and the most
relevant developments at the time of reviewing. The book starts with a
description of the physics of surface plasmons and SPR in its original form
and some novel applications, for example, nanoparticle SPR.
The description of SPR instrumentation and a survey of currently available
commercial products from 25 companies follows in Chapter 3. An introduction
on how to obtain kinetic information from SPR measurements can be found in
12 Chapter 1
Chapter 4, followed by Chapter 5 illustrating kinetic and thermodynamic
analysis of ligandreceptor interactions, probing the validity of this approach
in pharmaceutical applications. Chapter 6 brings the reader closer to the
surface architecture and chemical design strategies of SPR sensors. An
in-depth treatise on the analysis cycle and modern assay architecture, including
SPR microarray imaging, is provided in Chapter 7, followed by advanced
methods for SPR imaging biosensing in Chapter 8.
Specic application areas are highlighted in the last few chapters of the book,
revealing Surface Plasmon Fluorescence Techniques (Chapter 9) and the future
of medical applications at the point of patient care (Chapter 10) and for
food safety (Chapter 11). Finally, Chapter 12 gives an outlook on future trends
in SPR technology, including lab-on-a-chip microuidics and trends for
measuring reliable kinetic parameters.
1.5 Questions
1. SPR technology for direct and label-free detection of biomolecular inter-
actions dominates anity biosensor technologies to a great extent and it is
expected that in 2007 more than 1000 papers regarding SPR results will be
published. What are the technical reasons for the success of SPR?
2. In SPR, the intrinsic refractive index of a protein which accumulates on
the sensor surface is measured. Explain how we can distinguish between
the refractive index of the buer and that of the adsorbed protein.
3. Why should we express the sensitivity of an SPR instrument in accumu-
lated mass per square surface and not in moles per liter?
4. Consider the monophasic reversible interaction A+B"AB, where A is
the analyte and B is the immobilized ligand. The sample is injected and
shows a higher background electrolyte refractive index. Draw the sensor-
gram of two analysis cycles of injection of a sample with the second
analysis cycle a two times diluted sample. Consider 100% regeneration
after each analysis cycle.
5. The response DR gives us an indication of the amount of accumulated
mass per unit surface area. How can we determine the concentration of an
analyte in solution from these responses?
6. The study of the rate constants of biomolecular interactions is an impor-
tant feature of surface plasmon resonance biosensors. Why?
References
1. R.W. Wood, Philos. Mag., 1902, 4, 396402.
2. R.W. Wood, Philos. Mag., 1912, 23, 310317.
3. Lord Rayleigh, Proc. R. Soc. London, Ser. A, 1907, 79, 399.
4. U. Fano, J. Opt. Soc. Am., 1941, 31, 213222.
5. A. Otto, Z. Phys., 1968, 216, 398410.
6. E. Kretschmann and H. Reather, Z. Naturforsch., Teil A, 1968, 23, 21352136.
7. B. Liedberg, C. Nylander and I. Lundstrom, Sens. Actuators, 1983, 4,
299304.
8. S. Geib, G. Sandoz, K. Mabrouk, A. Matavel, P. Marchot, T. Hoshi,
M. Villaz, M. Ronjat, R. Miquelis, C. Leveque and M. de Waard,
Biochem. J., 2002, 364, 285292.
9. E. Kretschmann, Z. Phys., 1971, 241, 313324.
10. D. Wassaf, G. Kuang, K. Kopacz, Q.L. Wu, Q. Nguyen, M. Toews,
J. Cosic, J. Jacques, S. Wiltshire, J. Lambert, C.C. Pazmany, S. Hogan,
R.C. Ladner, A.E. Nixon and D.J. Sexton, Anal. Biochem., 2006, 351,
241253.
11. M. Pe rez-Moralesa, J.M. Pedrosab, E. Mun oza, M.T. Mart n-Romeroa,
D. Mo biusc and L. Camachoa, Thin Solid Films, 2005, 488, 247253.
12. J. Ctyrokya, J. Homola and M. Skalskya, Opt. Quantum Electron., 1997,
29, 301311.
13. C.R. Yonzon, E. Jeoung, S. Zou, G.M. Mrksich and R.P. van Duyne,
J. Am. Chem. Soc., 2004, 126, 1266912676.
14. Z. Salamon, H.A. Macleod and G. Tollin, Biochim. Biophys. Acta, 1997,
1331, 117129a.
15. R.B.M. Schasfoort, R.P.H. Kooyman, P. Bergveld and J. Greve, Biosens.
Bioelectron., 1990, 5, 103125.
16. J. Homola, Surface Plasmon Resonance Based Sensors. Springer Series on
Chemical Sensors and Biosensors, Series Ed. O.S. Wolfbeis, Vol. 4, Springer,
Berlin, 2006.
17. R.P.H. Kooyman, H. Kolkman, J. Van Gent and J. Greve, Anal. Chim.
Acta, 1988, 213, 3545.
18. U. Jonsson, L. Fagerstam, B. Ivarsson, B. Johnsson, R. Karlsson,
K. Lundh, S. Lofas, M. Malmqvist, BioTechniques, 1991, 11, 620622, 624.
19. L.A. Chtcheglova, M. Vogel, H.J. Gruber, G. Dietler and A. Haeberli,
Biopolymers, 2006, 83, 6982.
20. R. Karlsson and R. Stahlberg, Anal. Biochem., 1995, 228, 274280.
21. T. Wink, J. de Beer, W.E. Hennink, A. Bult and W.P. van Bennekom, Anal
Chem., 1999, 71, 801805.
22. D. Nedelkov, A. Rasooly and R.W. Nelson, Int. J. Food Microbiol., 2000,
60, 113.
23. B. Johnsson, S. Lofas and G. Lindquist, Anal. Biochem., 1991, 198, 268277.
24. R.L. Rich and D.G. Myszka, J. Mol. Recognit., 2006, 19, 478534.
25. R.L. Rich and D.G Myszka, Anal. Biochem., 2007, 361, 16.
14 Chapter 1
CHAPTER 2
Physics of Surface Plasmon
Resonance
ROB P.H. KOOYMAN
Biophysical Engineering Group, Faculty of Science and Technology,
University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
2.1 Introduction
In the last two decades, surface plasmon resonance (SPR) has evolved from a
fairly esoteric physical phenomenon to an optical tool that is widely used in
physical, chemical and biological investigations where the characterization of
an interface is of interest. Recently, the eld of SPR nano-optics has been added
where metallic structures on a nanoscale can be designed such that they can
perform certain optical functions. This chapter will be mainly concerned with
the more conventional, well-understood SPR theory used in sensor applications
and it will touch upon some of the newer developments relevant for this area.
Essential for the generation of surface plasmons (SPs) is the presence of free
electrons at the interface of two materials in practice this almost always
implies that one of these materials is a metal where free conduction electrons
are abundant. This condition follows naturally from the analysis of a metal-
dielectric interface by Maxwells equations. From this analysis, the picture
emerges that surface plasmons can be considered as propagating electron density
waves occurring at the interface between metal and dielectric. Alternatively,
surface plasmons can be viewed as electromagnetic waves strongly bound to
this interface; it is found that the surface plasmon eld intensity at the interface
can be made very high, which is the main reason why SPR is such a powerful
tool for many types of interface studies.
Experimental research on SPs started with electron beam excitation; in 1968,
optical excitation was demonstrated by Otto [1] and Kretschmann and Raether [2].
This last approach turned out to be much more versatile, so in this chapter the
focus will be on the optics of SPR. The following is by no means intended as an
15
in-depth treatment of surface plasmons, rather it is an attempt to provide a low-
threshold introduction to the physics of SPR for those who are actually
involved in SPR work and want to understand a bit more than measuring
the shift of the SPR dip.
2.2 The Evanescent Wave
Before we discuss SPs in more detail, it may be appropriate to provide a mathe-
matical description of the evanescent wave, which is so central in the concept of
SPR sensing. This is conveniently done by considering the phenomenon of total
internal reection.
An electromagnetic plane wave that propagates in a medium with refractive
index n can mathematically be described by an electric eld E:
E E
0
exp jot jk r E
0
exp jot jk
x
x jk
y
y jk
z
z
_ _
2:1
where E
0
is the amplitude of the electric eld, o is the angular frequency, k is
the wavevector, r (x,y,z) is the position vector and j O1. Note that eq.
(2.1) only represents a traveling wave if the exponent is complex.
In the present context, we will mainly be concerned with the wavevector k: its
direction is parallel to that of the wave propagation; its magnitude is given by
k
k
2
x
k
2
y
k
2
z
_
n
2p
l
n
o
c
2:2
where l and c are the wavelength and propagation velocity in vacuum,
respectively.
Next we consider the refraction of such a wave at an interface between two
media 1 and 2 with refractive indices n
1
and n
2
, respectively (see Figure 2.1).
Without loss of generality, we can choose the direction of the light beam such
that k
z
0 and our problem becomes essentially two-dimensional. From
elementary physics we know that for this situation Snells law holds:
n
1
sin a n
2
sin b 2:3a
or, equivalently,
k
x
1
k
x
2
k
x
2:3b
By using eqs. (2.2) and (2.3b), we can nd an expression for the component of
the wavevector k
y
perpendicular to the interface
1
:
k
2
y
2
n
2
1
2p
l
_ _
2
n
2
2
n
2
1
sin
2
a
_ _
2:4
Now, let us assume that n
1
4n
2
. From eq. (2.4), it is seen that for sin a 4n
2
/n
1
the right part is negative, and, consequently, k
y
is purely imaginary. Returning to
1
Note that the direction y is in this chapter always perpendicular to the surface.
16 Chapter 2
eq. (2.1), we conclude that for this case in medium 2 there is only a traveling wave
parallel to the interface:
E
2
E
0
e
k
y
2
y
exp jot jk
x
x 2:5
with the amplitude of the electric eld exponentially decaying along the y-direction
with a characteristic distance 1/k
y2
1/jk
y2
. For obvious reasons, this eld in
medium 2 is denoted as the evanescent eld. Eq. (2.4) can be used to calculate its
penetration depth, which is of the order of half a wavelength. This explains the
interface sensitivity of the evanescent eld: only close to the interface is an
electromagnetic eld present; therefore, only a changing dielectric property (e.g.
a changing refractive index) in the vicinity of the interface will inuence this eld.
We will see that also in SPR an evanescent eld is generated.
2.3 Surface Plasmons
2.3.1 Surface Plasmon Dispersion Equations, Resonance
There are several approaches that all result in the dispersion relation for an SP,
that is, a relation between the angular frequency o and the wavevector k. In his
last standard treatise on SPs, Raether [3] calculated the SP dispersion relation
from rst principles, viz. Maxwells equations. A particularly elegant approach
was suggested by Cardona [4] and we will adopt it here. For reasons that will
become clear in the course of Section 2.3.2, we will only discuss p-polarized
2
Figure 2.1 Refraction of light at an incident angle a, at an interface of two materials
with refractive indices n
1
and n
2
. Definition of axis system and quantities.
2
p-Polarized light has its electric eld vector in the plane of incidence.
17 Physics of Surface Plasmon Resonance
light interacting with an interface. For any interface between two media, the
complex reection coefcient r
p
for p-polarized incident light electric eld is
described by Fresnels equations (see, e.g., ref. [5] for a derivation on the basis
of Maxwells equations):
r
p

E
i
E
r
r
p
e
jj
tana b
tana b
e
jj
2:6a
where E
i
and E
r
are the incident and reected electric elds, respectively, and
the angles a and b are dened as shown in Figure 2.1.
3
Of course, the angles a and b are again related by Snells law [eq. (2.3)]; in
addition, a phase change j of the reected eld relative to the incident eld
occurs, depending on the refractive indices of the materials involved.
For the reectance, dened as the ratio of the reected intensities, the
following relation holds:
R
p
r
p
2
2:6b
Now, following Cardona [4], two special cases exist: if a+bp/2, then the
denominator of eq. (2.6a) becomes very large and thus R
p
becomes zero. This
situation describes the Brewster angle, where there is no reection for
p-polarized light. The other special case occurs when abp/2: we see from
eqs. (2.6a) and (2.6b) that R
p
becomes innite: there is a nite E
r
for a very
small E
i
. This circumstance corresponds to resonance. From this relation
between a and b we can deduce the dispersion relation if ab p/2, then
cosa sinb and tana k
1x
/k
1y
n
2
/n
1
. For the components of the wave-
vector k(k
x
, k
y
), we can write
k
2
x
k
2
1
k
2
y
1
k
2
1
k
2
x
e
1
e
2
2:7
k
x

o
c
e
1
e
2
e
1
e
2
_
and k
yi

o
c
e
2
i
e
1
e
2
2:8
where e
1
and e
2
are the dielectric constants
4
of materials 1 and 2, respectively,
and i 1 or 2. Equation (2.8) is the sought SPR dispersion equation for an
interface between two half-innite media.
Next, we investigate the case where medium 2 is a metal. This medium then
contains a large number of free electrons and the consequence is that at an angular
frequency ooo
p
its dielectric constant e
2
will be negative (see, e.g., ref. [5]):
e
2
o 1
o
2
p
o
2
2:9a
o
p

4pn
e
e
2
=m
e
_
2:9b
3
Note that in Figure 2.1 the direction y (instead of z) is perpendicular to the surface.
4
Dielectric constant and refractive index are related by e n
2
.
18 Chapter 2
where o
p
is the so-called plasma frequency, n
e
is the free electron density and
e and m
e
are the electron charge and mass, respectively.
Generally, this implies that for ooo
p
no electromagnetic eld can propagate
in a metal [cf. eqs. (2.1) and (2.2)]. More specifically, provided that e
2
4e
1
, we
nd for the interface that k
yi
is imaginary, whereas k
x
remains real. Thus an
electromagnetic wave exists, propagating strictly along the interface, with
evanescent tails extending into both sides of the interface [cf. eq. (2.5)]. To
get a feeling for the quantities involved, it is instructive to calculate penetration
depths for a real case, on the basis of eq. (2.8). We take l 700 nm, thus
o2.69 10
15
s
1
and a gold/water interface. At this wavelength e
gold
E16
and e
water
E1.77. We calculate for the penetration depths 1/k
y,water
238 nm
and 1/k
y,gold
26 nm.
Now all ingredients are available to appreciate the use of SPR in sensor
applications. Let us assume that we have a situation where molecules X are
allowed to adsorb to the water/metal interface. We can view this as a process
where water molecules are replaced by molecules X. Because, generally,
e
X
ae
water
, the average dielectric constant close to the interface will change.
Equation (2.8) then describes the concomitant change of the wavevector k
x
.
Because the SP eld is evanescent in the direction perpendicular to the inter-
face, a change of the dielectric constant e
2
is only detectable in SP character-
istics if this change occurs within the penetration depth of the SP eld: an SPR
sensor will only be sensitive to molecular processes (binding, adsorption, etc.)
that occur at a distance to the metal surface that is roughly half the wavelength
of the used light.
2.3.2 Excitation of Surface Plasmons
By substitution of eqs. (2.9a) and (2.9b) into eq. (2.8), we obtain a graphical
representation of the SPR dispersion relation as shown in Figure 2.2 (line I). In
the same gure, the dispersion relation for normal light is depicted (line a).
We immediately see that, apart from the origin, there is no point where the SPR
curve and the light curve intersect, implying that in this geometry normal
light cannot simultaneously provide the correct wavevector and angular
frequency to excite a surface plasmon.
One way to circumvent this problem is to introduce a second interface, as
depicted in the inset of Figure 2.2. Here a thin metal layer (dielectric constant e
m
)
is sandwiched between two dielectric materials 1 and 3 with different dielectric
constants e
1
and e
3
, with e
1
4 e
3.
By applying Fresnels equations to the two
interfaces, more complicated dispersion equations are found than eq. (2.8);
however the essential physics remains unchanged. We now nd two dispersion
equations for k
x
, one for each interface, and we see that the line representing the
dispersion relation for normal light in medium 1 (line b) intersects the SP
dispersion line for the metal/medium 3 interface. This indicates that light
incident from medium 1 can excite SPs: by proper adjustment of the incoming
angle a (Figure 2.2, inset), we can tune the incoming wavevector k
x
kn
1
sina to
match the wavevector necessary for SP excitation. In this way, any k
x
between
the two lines, labeled a and b in Figure 2.2, can be set. As an example, one such
line, labeled c, is indicated. This so-called attenuated total reection (ATR)
technique was rst demonstrated
5
by Kretschmann and Raether [2] and has
since then almost become the standard technique for SP excitation.
Another way of providing a wavevector appropriate for SP excitation is the
use of a metal layer on which a periodic structure is prepared [6] as illustrated in
Figure 2.3.
When light with wavevector k
x
2p/l.n
i
siny falls on such a structure, this
acts as a diffraction grating and diffraction orders m0, 1, 2, . . . are
generated in the reected light (see, e.g., ref. [7]). The generated wavevector
k
x,net
parallel to the interface can be written as
k
x;net
k
x
m
2p
L
2:10
Figure 2.2 Dispersion relation for surface plasmons. Curves I and II represent the SP
dispersion for the interfaces e
3
/e
m
and e
1
/e
m
, respectively. The lines a and b
are the dispersion relations for normal light in medium e
3
and e
3
,
respectively, which are dependent on the angle of incidence a in the
experimental setup as indicated in the inset. By varying a, any line c
between the lines a and b can be realized.
5
In fact, Otto was the very rst to demonstrate this in a somewhat less versatile form.
20 Chapter 2
where L is the periodicity of the grating. Again, the wavevector k
x,net
can be
tuned to the SPR wavevector, given by eq. (2.8),
6
by changing the incident
angle.
Up to now the required polarization direction of the incoming light remained
unmentioned. As already pointed out, SPs are conductivity uctuations
brought about by collective surface charge density oscillations. These charge
density waves have to be excited by an external electric eld. Only an electric
eld with a component perpendicular to the interface can induce a surface
charge density; only p-polarized light has a perpendicular electric eld
component.
2.3.3 Surface Plasmon Properties
With SPs, a number of specic properties are associated that are particularly
relevant to sensor applications: (1) the eld enhancement, (2) the phase jump of
the reected eld upon SP excitation and (3) the SP coherence length.
Field enhancement. A calculation of the electric eld transmission coefcient
on the basis of Fresnels equations for the interface reveals that the electric eld
at the low index side of the metal can be much larger than that at the other side
of the metal layer.
In Figure 2.4, the intensity enhancement is depicted as a function of the angle
of incidence of incoming light for a number of different thicknesses of a gold
layer. It is found that very close to the SPR angle the intensity can be enhanced
by a factor of more than 30. This circumstance accounts for much of the
Figure 2.3 Schematic view of a grating coupler. By diffraction of an incident light
beam, the grating produces k
x
values larger than that of the incident light.
By adjusting the incident angle a the wavevector can be tuned to k
x
required to produce a surface plasmon.
6
The dispersion equation eq. (2.8) is hardly affected if the metal surface has a shallow corrugation.
remarkable sensitivity that the SPR condition has for a changing dielectric
environment.
Phase jump. As already mentioned in Section 2.3.1 and expressed in eq. (2.6),
a reection event at an interface is generally accompanied by a phase jump of
the reected eld. This is illustrated in Figure 2.5a for a prismgoldwater
system.
For comparison, the conventional SPR dip is shown in Figure 2.5b for the
same layer system. We see that around the SPR dip the phase of the reected
electric eld undergoes a relatively large change. The significance of this phe-
nomenon for sensing purposes is more clear when we plot the reectance and
phase changes as a function of incident angle for a certain change in dielectric
constant of the water. This is depicted in Figure 2.5c and d. In the following
rough calculation, we assume that both the change in reection coefcient DR
and the phase change Dj are proportional to De. From Figure 2.5c we estimate
that DR/De E30, whereas from Figure 2.5d we nd that Dj/De E250.
Experimentally, a minimum DRE10
3
can be measured, whereas a mini-
mum Dj E 10
3
is feasible, using interferometric techniques. The conclu-
sion is that on the basis of reectance measurements a minimum DeE4 10
5
can be detected, whereas a phase measurement provides a sensitivity of
De E4 10
6
. In view of the very approximate character of this calculation,
Figure 2.4 Field enhancement for various values of the thickness of the gold layer.
Wavelength of the excitation light is 700 nm; the low-index side of the
metal layer consists of water (e
3
1.77).
22 Chapter 2
the absolute values found are of limited validity; however, the nding that a
phase measurement provides an order of magnitude better sensitivity is a hard
conclusion and, indeed, this was demonstrated by Nikitin and co-workers [8,9].
The only drawback of this approach seems to be the much more complicated
experimental setup.
SP coherence length. Generally, the metals dielectric constant e
2
is complex
and this circumstance results in a complex propagation constant
k
x
k
x
0 0
+jk
x
0 0
[cf. eq. (2.8)], where k
x
0 0
and k
x
0 0
are real and imaginary parts,
respectively. For a surface plasmon, traveling along the interface with wave-
vector k
x
, this implies that the eld intensity decays with a characteristic
distance 1/2k
x
0 0
. For gold and silver, the standard metals in sensor applications,
Figure 2.5 Comparison of the angle-dependent phase changes (a, c) and reectance
changes (b, d) for variation of the dielectric constant at the low-index side
of the metal layer. A gold layer is used, SPs are excited at l 700 nm.
(c) and (d) depict the differential phase and reectance, respectively, for a
change in the mediums dielectric constant of 0.01.
the imaginary part of the dielectric constant increases with decreasing wave-
length and the SP propagation length decreases accordingly.
This is illustrated in Figure 2.6: here a layer system was prepared where a
30nm SiO
2
strip was deposited on a 50nm silver layer. For a series of wave-
lengths the angle of incidence was chosen such that SPs were excited in the area
Figure 2.6 SPR response to a dielectric step at several wavelengths. For each wave-
length the light angle of incidence is set such that outside the strip
(extending from 0 to 125 mm) the interrogating k
x
is resonant with the
surface plasmon wavevector. The surroundings of the strip has dielectric
constant e
3
1.
24 Chapter 2
outside the strip and for each wavelength the whole area was illuminated with a
collimated light beam under a constant angle of incidence. Because of the con-
trast in dielectric constant between the strip and its surroundings (air). The SP
resonance condition is not fullled in the area below the strip and we see the
decaying SP (increasing reectance) at the left edge of the strip. Beyond the right
edge of the strip the SPR condition is again fullled and the SP resonance builds
up. The gure nicely demonstrates that with decreasing wavelength the SP
propagation length becomes shorter: the blurring on the left side of the strip
becomes less prominent for shorter wavelengths. It turns out that in the wave-
length range 50000nm the propagation length varies between o10 and 40 mm.
For a quantitative description of the ndings depicted in Figure 2.6, we have
to analyze the interference between the several elds that are present in the layer.
In Figure 2.7, the layer system and the elds involved are indicated: the
resonant SP eld, the non-resonant SP eld and the external exciting eld with
amplitudes E
1
, E
2
, E
3
, respectively. For a resonant SP that enters the SiO
2
strip-
covered layer, the total eld reaching the detector can be written as [10,11]
E
tot
x E
1
E
2
e
j k
0
1
jk
00
1
_ _
x
E
2
e
jk
0
0
x
E
3
e
jk
0
0
x
2:11
where k
0
1
is the wavevector corresponding to resonance in the covered area and
k
0
0
is the wavevector that excites SPs in the uncovered area.
Dening AE
1
E
2
and BE
2
E
3
, the resulting intensity at the detector
becomes
I
tot
x B
2
A
2
e
2k
00
1
x
2ABe
k
00
1
x
cos k
0
1
k
0
0
_ _
x 2:12
When a non-resonant SP leaves the covered area, the resonant SP builds up and
the intensity at the detector decreases accordingly [11]:
I
tot
x B A 1 e
k
00
0
x
_ _ _ _
2
2:13
From Figure 2.6, we see that this model gives a very accurate description of the
experiments.
Figure 2.7 Definition of wave vectors and elds for the system consisting of a
dielectric strip on top of a metal layer.
Both this model and the experiments indicate that a plasmon needs roughly
four times the propagation length L
x
for a full decay or for a full build-up; this
propagation length can be loosely dened as
L
x

1
2k
00
x
2:14
This implies that SPs with mutual distances significantly larger than L
x
are
independent. This is a very important conclusion because it is the fundament of
surface plasmon microscopy [12,13], with its many applications in SPR imaging
and SPR multisensing: on a substrate we can dene areas that in an SPR
experiment will behave mutually independently, provided that these areas are
significantly larger than L
x
2
. For SPs on gold, excited at l 632 nm, L
x
E7 mm
and on a total sensor area of 1 cm
2
more than 10
4
independent sensor patches
that each have an area of somewhat smaller than 100 100 mm
2
can in principle
be dened, of which the optical responses can be simultaneously read out by
using an imaging system. As practical aspects are outside the scope of this
chapter, the interested reader should consult Chapter 7 for more details.
2.3.4 Choice of Experimental Parameters
It is impossible to dene a general set of optimum SPR parameters, for
instance, optimal spatial resolution in an SPR microscopy/imaging setup
requires values of the experimental parameters other than those to obtain
maximum sensitivity for dielectric changes. Therefore, this section provides
only some general guidelines, based on consideration of the properties of the
metal layer.
To obtain maximum sensitivity, it is advantageous to maximize the steepness
of the reectance as a function of the angle of incidence, because this allows for a
more accurate determination of the angle of minimum reectance (cf. Figure 2.8).
This implies optimization of the reectance minimum R
min
and minimizing the
width of the resonance curve. R
min
can be made very close to zero by selecting
the appropriate thickness of the metal layer; as can be seen in Figure 2.8,
optimum thicknesses are somewhat dependent on the applied wavelength and
are between 40 and 50nm. The width of the resonance curve is mainly deter-
mined by the complex value of the metals dielectric constant. Generally, a large
(negative) real part, together with a small imaginary part, results in narrow
resonance curves. In practice only two options are available for the choice of the
metal layer: gold or silver. As seen in Figure 2.8, silver has the better SPR
characteristics in view of the larger real part of its dielectric constant; however, it
is chemically less inert. In Figure 2.8, it is also seen that the use of higher
excitation wavelengths has an appreciable effect on the width of the resonance
curve. This is one of the reasons why (near-) infrared SPR experiments are
attracting attention [14,15]. However, it should be realized that narrowing the
reectance curve necessarily implies increasing the SPR propagation length [eq.
(2.13)], which can be a disadvantage in certain SPR imaging applications. For a
26 Chapter 2
gold layer, it can be calculated that an increase in wavelength from 450 to
1500 nm results in a change in the propagation length from 100 to almost 1 mm.
Finally, it should be mentioned that an increase in wavelength results in an
increase in the penetration depth 1/k
y
[cf. eqs. (2.4) and (2.5)], with the conse-
quence that the reectance minimum will become more sensitive to dielectric
changes relatively far from the metal/dielectric interface; hence the surface-
sensitive character of SPR becomes less prominent. This implies that for detec-
tion of the growth of thin layers the optimum choice of wavelength will be
different from that in a situation where a more bulk-like change in refractive
index has to be detected [33].
2.4 Analysis of Multi-layered Systems
In most SPR-based sensor applications, the system of interest consists of a gold
or silver layer on which one or more thin layers are deposited in an aqueous
environment. Often the desired parameters are the thicknesses of the several
layers, which can be converted into surface concentrations of the layer-composing
molecules (cf. Chapters 4 and 5).
One way to obtain these parameters is a repeated application of the Fresnel
equation [eq. (2.6a)]. The following relation holds for a system consisting of N
Figure 2.8 The SPR dip for 46 nm of silver (dashed) and gold (solid) with water
on the low-index side, for several excitation wavelengths. Dielectric data
for the metal layer obtained from refs. [3] and [33].
layers with dielectric constants and thicknesses e
i
and d
i
, respectively, placed
between a prism with dielectric constant e
p
and a medium (e.g. water) with
dielectric constant e
w
(see. e.g., ref. [5]):
r
p
a
M
11
M
12
k
y;w
e
w
_ _
k
y;p
e
p
M
21
M
22
k
y;w
e
w
_ _
M
11
M
12
k
y;w
e
w
_ _
k
y;p
e
p
M
21
M
22
k
y;w
e
w
_ _
2:15a
where M is the so-called transfer matrix:
M M
1
M
2
. . . M
N
2:15b
with
M
i

cos k
y;i
d
i
_ _
je
i
k
y;i
sin k
y;i
d
i
_ _
jk
y;i
e
i
sin k
y;i
d
i
_ _
cos k
y;i
d
i
_ _
_ _
2:15c
The angular dependence of r
p
is contained in the wavevectors k
y,i
, perpen-
dicular to the layer system; these can be calculated using eq. (2.4). The
reectance can be obtained by application of eq. (2.6b).
Provided all thicknesses d
i
and dielectric constants e
i
are known, eq. (2.15)
gives an accurate description of the SPR experiment. Of course, in practice one
is concerned with the inverse problem and a priori knowledge, such as the
dielectric constant and dimensions of the molecules composing a certain layer,
is required for a satisfactory analysis of experimental SPR results. This
precludes an unambiguous analysis of more than two or three layers.
Another, more intuitive, approach is the introduction of an effective dielec-
tric constant, e
e
. Here, the actual multi-layered system is replaced by a two-
layer system, where e
1
in eq. (2.8) is replaced by the effective dielectric constant
e
e
, given by the average of all dielectric constants in the layer system, weighted
by the penetration depth y
0
of the SPR evanescent eld [16,17]:
e
eff

2
y
0
_
N
0
eye
2y=y
0
dy 2:16
Of course, with the use of this equation. we face the same problems as those
when we use eq. (2.15).
2.5 SPR Spectroscopy
2.5.1 Enhancement of Fluorescence and Absorbance
Up to now we have only considered, apart from the metal layer, transparent
layers, i.e. layers that are characterized by a positive, real dielectric constant.
When one or more layers contain a light-absorbing compound, SPs can boost
the uorescence intensity from a thin layer more than 40-fold [18]. This effect is
due solely to the large eld enhancement that occurs on the low index side of
28 Chapter 2
the metal layer when an SP resonance condition is established (cf. Figure 2.4).
More information on this phenomenon can be found in Chapter 9. Another,
more subtle, feature of the interaction between SPs and light-absorbing mol-
ecules is the increased sensitivity for the detection of absorbances in thin layers.
The addition of a light-absorbing layer results in two effects on the SPR
angular-dependent curve, which are quantitatively described by eq. (2.15): (1)
the SPR dip shifts to a larger angle of incidence and (2) the value of the
reectance minimum increases. This last effect can readily be understood as
light absorption necessarily results in a decreased reectance. In addition, the
SPR eld enhancement on the low-index side of the metal layer will result in
increased sensitivity for dielectric changes [19] and therefore also for changes in
the absorbance. The rst effect is a consequence of the KramersKronig
relation (see, e.g., ref. [20]), which in the present context can be expressed as
the statement that any change in the imaginary part of the dielectric constant
will be accompanied by a change in the real part; in SPR it is mainly the real
part of the dielectric prole on top of the metal layer that determines the
angular position of the SPR dip. It has been demonstrated [21] that an SPR-
assisted monolayer absorbance measurement can result in a 40-fold reectance
increase as compared with a metal-lacking ATR system. In addition, it is
possible to extract unambiguously the thickness and the dielectric constant of
an absorbing layer from a single SPR experiment [22].
2.5.2 SPR and Metal Nanoparticles
SPR phenomena are not restricted to planar multilayers as discussed so far; it
turns out that for metal particles with dimensions much smaller than the
wavelength of the interacting light, SP effects can be much more prominent.
Generally, the net electric eld E
tot
around a dielectric particle is composed of
the superposition of an external applied eld E
0
and the induced (dipole) eld in
the particle. For a polarizable spherical particle with radius r
m
and dielectric
constant e, placed in a medium with dielectric constant e
1
, the following
expression is found (see, e.g., ref. [5]) for the eld gain G:
Go
E
tot
o
E
0
o
E1
eo e
1
eo 2e
1
r
m
r r
m
_ _
3
2:17
It is seen that G can reach enormous values for e close to 2e
1
; in a normal
dielectric medium where e
1
4 0, this condition points to the use of a metal,
where e can be negative; additionally, the imaginary part of e should be as small
as possible. It turns out that this condition corresponds to the excitation of a
surface plasmon in the metallic nanoparticle [23]. Particularly in the eld of
Raman spectroscopy this can result in enormous sensitivity enhancements
7
(for
a review, see ref. [24]).
7
Conventional Raman spectroscopy suffers from a very low scattering efciency which can be 12
orders of magnitude lower than that of uorescence.
Now consider a Raman-active molecule near a metal nanoparticle. The
detected Raman intensity I(o,o
sc
) can be expressed as
I
Raman
o; o
sc
gE
2
exc
E
2
sc
gG
2
oG
2
o
sc
I
0
oI
0;sc
o
sc
2:18
where E
exc
is the total excitation electric eld to which the molecule is exposed
and E
sc
is the total Raman-scattered eld. The constant g is an experimental
constant that is unimportant in the present discussion.
By choosing an angular frequency o that excites surface plasmons in the
metal (usually gold or silver) and detecting scattering frequencies o
sc
not too
far from the excitation frequency, both the excitation and the scattered eld are
enhanced by the presence of the metal particle. By substituting eq. (2.17) into
eq. (2.18), we see that the distance dependence of the net Raman scattering
intensity changes with the power 12 of the moleculenanoparticle distance!
Indeed, it has been found experimentally that a surface-enhanced Raman
spectroscopy (SERS) experiment can result in experimental Raman signals that
are enhanced 10
12
10
14
times compared with those obtained from non-surface-
enhanced experiments.
It should be added that apart from this SPR enhancement mechanism,
another chemical enhancement effect is operational, which accounts for a
10100-fold amplication of the bare Raman signal [24].
It has been demonstrated [25,26] that SERS is able to detect single molecules.
Together with its very high molecular specicity, this offers great promise as a
detection tool for very low concentrations of biomolecules, such as DNA
strands or proteins.
In principle, these eld enhancements should also be important in the
detection of uorescent molecules near a metal nanoparticle. However, the
nearby presence of a metal layer leads to additional non-radiative decay paths
of the electronic excited states of a nearby molecule, with the net result that in
many cases the uorescence will be largely quenched.
So far, metal nanoparticles were considered as surface plasmon-assisted eld
ampliers. However, these particles can also be exploited as intrinsic refractive
index sensors, analogous to the more familiar planar SPR experiments (for
reviews, see refs. [27] and [28]). The physical basis of this application is the light
extinction (absorption and scattering), which is heavily dependent on the
nanoparticles dielectric constant, size and geometry and also on the dielectric
constant e
1
of the surrounding medium. Mie theory gives a reasonably adequate
description of the extinction coefcient A
ext
and for spherical particles with
diameter less than about 20 nm the following expression is found [23]:
A
ext

18pN
p
Ve
3=2
1
l
Ime
Ree 2e
1

2
Ime
2
2:19
where N
p
is the number of nanoparticles, each of which has a volume V, and l
is the wavelength of the applied light.
30 Chapter 2
Again we see the pronounced inuence of the occurrence of SPs: at Re(e) 2e
1
we nd a maximum in the extinction coefcient, which can reach large values
for low values of Im(e). Hence also in this situation we are led to the use of gold
or silver as a metal nanoparticle.
More sophisticated models (see, e.g., ref. [29]) also account for the size and
shape of the nanoparticles and computer programs are available in the public
domain that can predict the extinction spectrum of nanoparticles of any shape
[30], by modeling the particle as a series of dipoles placed in an oscillating
electric eld. However, the main features of nanoparticle extinction remain
contained in eq. (2.19).
For one nanoparticle with a diameter around 25 nm, excited close to its SP
resonance, eq. (2.19) results in A
ext
of the order of 10
16
m
2
, which corresponds
to the more familiar molar extinction coefcient in the order of 10
9
l mol
1
cm
1
. This value, which indeed was observed experimentally [29], is
more than three orders of magnitude larger than that of strong light-absorbing
organic dye molecules, allowing for relatively simple optical detection and
characterization of individual nanoparticles [31,32]. In another series of exper-
iments, the shift of the extinction maximum as a function of the refractive index
of the surrounding medium was investigated [28]. It was found experimentally
that for silver nanoparticles the spectrum could shift as much as 20 nm for a
change of 0.1 in the refractive index. Because molecules that adsorb to a
nanoparticle change the refractive index around the particle, it is obvious that
this, analogous to conventional SP resonance, can be used as a sensor principle.
Indeed, it has been demonstrated experimentally that the full coverage of a
silver nanoparticle with low molecular mass molecules resulted in a spectrum
shift of approximately 40 nm. The full coverage corresponded to only 4 10
4
molecules. Together with the single particle detection capability, this promises
enormous sensitivity, allowing for near single molecule detection [32].
2.6 Concluding Remarks
The phenomenon of SPR is one of the many examples where an interesting
physical phenomenon leads to applications that are highly important to both
applied science and society. In a planar SPR system, it is particularly the
combination of eld enhancement and relatively short coherence length that
allows for a unique sensor concept that provides both multiplexing capabilities
and very high sensitivity. The general physical picture is well understood;
however, some areas are still in vivid scientic debate (SERS, optics of
nanoparticles).
From a technological point of view, the emerging eld of nanotechnology
will enable us to exploit to its full potential the SP phenomena of tailored
nanoparticles. It is the authors rm conviction that merging of (bio-)nano-
technology and SP phenomena of nanoparticles will ultimately lead to sensor
concepts and sensor realizations that will really be important in numerous
aspects of society, varying from food safety monitoring and high-throughput
screening to early in vivo detection of tumor growth.
2.7 Questions
1. Derive eq. (2.4) from eqs. (2.2) and (2.3).
2. Calculate for a gold/water interface at l 700 nm (
gold
16;
water
1:770) the angular shift of the SPR dip when
water
increases to
1.775. For the light-incoupling we use a semi-circular glass piece with
refractive index n
glass
1.5.
3. Estimate the effective dielectric constant for the following system, when
this interface is probed with a wavelength l 700 nm (for dielectric
constants of gold and water, see previous question)
gold
water
The squares in the gure represent cubes of protein molecules in an
aqueous environment, adsorbed to the gold surface. Each protein mole-
cule has a volume of 5*5*5 nm
3
and a dielectric constant
protein
2.30.
The average distance between the edges of the cubes is 7 nm.
4. A particular SPR application could be the detection of micro-organisms
in, e.g., waste water, by detecting changes in bulk refractive index. Which
SPR excitation wavelength region would be more favourable, the blue/UV
or the infrared region?
5. Express the extinction coefcient A
ext
(eq. 2.16) in units M
1
cm
1
.
2.8 Symbols
A A E
1
E
2
A
ext
extinction coefcient
B B E
2
E
3
const experimental constant in surface-enhanced Raman spectroscopy
c propagation velocity in vacuum
d
i
lm or layer thickness
E electric eld strength
e charge of electron
E
0
amplitude of electric eld
E
0
external applied eld E
0
(Section 2.5.2)
E
exc
total excited electric eld (Raman)
E
i
incident electric eld
E
r
reected electric eld
32 Chapter 2
E
sc
total scattered eld (Raman)
E
tot
net electric eld around nanoparticles
G eld gain of nanoparticle SPR
I intensity
I
Raman
Raman intensity near an SPR nanoparticle
k wavevector
k
y
y component of wavevector
k
x, net
x component of wave vector
L
x
propagation length of a full decay or build-up
m
e
mass of electron
m diffraction order
N number of layers in system
n
e
free electron density
n refractive index
N
p
number of nanoparticles
n
i
refractive index of material i
r position vector
r
m
radius of polarizable spherical (nano)particle
r
p
reection coefcient (complex)
R
p
reectance
V volume of nanoparticles
y
0
penetration depth of SPR evanescent eld
a incident angle of light
b refraction angle of light
e dielectric constant
e
0
dielectric constant of medium
e
e
effective dielectric constant
e
p
dielectric constant of prism
e
w
dielectric constant of water
k
0
wavevector excited in uncovered area
k
1
wavevector corresponding to resonance of strip-covered area
l wavelength in vacuum
L periodicity of grating
j phase change of the reected eld relative to the incident eld
o angular frequency
o
p
plasma frequency
o
sc
scattering frequency
References
1. A. Otto, Z. Phys., 1968, 216, 398.
2. E. Kretschmann and H. Raether, Z. Naturforsch., 1968, 230, 2135.
3. H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on
Gratings, Springer, Berlin, 1988.
4. M. Cardona, Am. J. Phys., 1971, 39, 1277.
5. J.R. Reitz, F.J. Milford and R.W. Christy, Foundations of Electromagnetic
Theory, Addison-Wesley, New York, 1993.
6. D.C. Cullen, R.G. Brown, C.R. Lowe, Biosensors, 1987/88, 3, 211.
7. E. Hecht, Optics, Addison-Wesley, New York, 2002.
8. A.N. Grigorenko, P.I. Nikitin and A.V. Kabashin, Appl. Phys. Lett., 1999,
75, 3917.
9. A.V. Kabashin and P.I. Nikitin, Opt. Commun., 1998, 150, 5.
10. B. Rothenha usler, W. Knoll, J. Opt. Soc. Am. B, 1988, 5, 1401.
11. C.E.H. Berger, R.P.H. Kooyman and J. Greve, Opt. Commun., 1999, 167,
183.
12. E. Yeatman and E. Ash, Electron. Lett., 1987, 23, 1091.
13. B. Rothenha usler and W. Knoll, Nature, 1988, 332, 615.
14. B.P. Nelson, A.G. Frutos, J.M. Brockman and R.M. Corn, Anal. Chem.,
1999, 71, 3928.
15. S. Patskovsky, A.V. Kabashin, M. Meunier and J.H.T. Luong, J. Opt. Soc.
Am. A., 2003, 20, 1644.
16. K. Tiefentahler, W. Lukosz, J. Opt. Soc. Am. B, 1989, 6, 209.
17. R.G. Heideman, Optical waveguide based evanescent eld immunosensors,
PhD Thesis, University of Twente, Enschede, 1993.
18. T. Liebermann and W. Knoll, Colloids Surf. A, 2000, 171, 115.
19. H. Kogelnik, in Topics in Applied Physics, vol 7, ed. T. Tamir, Springer,
Berlin, 1975.
20. R.W. Ditchburn, Light, Blackie, London, 1963.
21. S. Wang, S. Boussaad and N.J. Tao, Rev. Sci. Instrum., 2001, 72, 3055.
22. P.S. Vukusic, J.R. Sambles and J.D. Wright, J. Mater. Chem., 1992, 2,
1105.
23. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters, Springer,
Berlin, 1995.
24. K. Kneipp, H. Kneipp, I. Itzkan, R.R. Dasari and M.S. Feld, J. Phys.
Condensed Matter, 2002, 14, R597.
25. S. Nie and S.R. Emory, Science, 1997, 275, 1102.
26. K. Kneipp, Y. Wang, H. Kneipp, L.T. Perelman, I. Itzkan, R. Dasari and
M.S. Feld, Phys. Rev. Lett., 1997, 78, 1667.
27. C.R. Yonzon, D.A. Stuart, X. Zhang, A.D. McFarland, C.L. Haynes and
R.P. Van Duyne, Talanta, 2005, 67, 438.
28. C.L. Haynes, R.P. Van Duyne, J. Phys. Chem. B, 2001, 105, 5599.
29. S. Link, M.A. El-Sayed, J. Phys. Chem. B, 1999, 103, 8410.
30. http://www.astro.princeton.edu/Bdraine/DDSCAT.6.1.html
31. S. Schultz, D.R. Smith, J.J. Mock and D.A. Schultz, Proc. Natl. Acad. Sci.
USA, 2000, 97, 996.
32. A.J. Haes, D.A. Stuart, S. Nie and R.P. Van Duyne, J. Fluor., 2004, 14,
355.
33. K. Johansen, H. Arwin, I. Lundstro m and B. Liedberg, Rev. Sci. Instrum.,
2000, 71, 3530.
34 Chapter 2
CHAPTER 3
SPR Instrumentation
RICHARD B.M. SCHASFOORT
a
AND
ALAN MCWHIRTER
b
a
Biochip Group, MESA+ Institute for Nanotechnology, Biomedical
Technology Institute (BMTI), Faculty of Science and Engineering, University
of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands;
b
General
Electric Healthcare, Biacore AB, Rapsgatan 7, SE-754 50 Uppsala, Sweden
1
3.1 Introduction
General aspects of surface plasmon resonance (SPR) instrumentation are
described in this chapter, providing the reader with an insight into the develop-
ment of SPR technology to its current state. SPR instruments comprise three
essential units integrated in one system: optics, liquid handling unit and the sensor
chip. Instruments may dier optically, in the liquid handling system or in the
degree of automation, all inuencing their performance. Additionally, the quality
and features of sensor chips determine the quality of the measurement of bio-
molecular interactions; however sensor surfaces are treated in detail in Chapter 6.
In this chapter, rst general aspects of the dierent SPR optical systems are
discussed, followed by aspects of liquid handling and operational automation,
concluded by a discussion of various SPR instruments. The latest instruments
of Biacore, the largest company in the SPR market with a complete SPR
product line, are discussed separately in Section 3.5; however, the Biacore Q,
dedicated to food analysis, is treated in Chapter 11 on food analysis.
3.1.1 From Surface Plasmon to SPR Signal
Real-time and label-free biomolecular interaction sensing is unthinkable at
present without the surface plasmon resonance principle [1,2]. Currently,
instruments on the market dier in performance based on dierences in their
1
Author of the section on Biacore instruments (Secton 3.5).
35
optical systems and their degree of development and automation. SPR optics,
liquid handlings and the sensor chip are integrated in the SPR instrument as
depicted schematically in Figure 3.1. As indicated, the sensor chip forms a
physical barrier between the optical system (dry section) and the liquid handling
unit e.g. ow cell (wet section).
How is the refractive index measured at the surface? SPR causes an intensity
decrease or dip in the reected light at the sensor surface, as discussed in Chapter 2.
In the left section of Figure 3.2, the reectivity, i.e. the ratio of incoming
and reected light vs. angle of incidence passes through a minimum depending
on the angle of incidence of the p-polarized light. The angle at which the
minimum is found depends on the refractive index of the medium in the
immediate vicinity of the sensor surface. On adsorption or binding of mole-
cules, the refractive index at the sensor surface will change, causing a shift of
the reectivity curve. This shift can be observed in real time, as shown in the
sensorgram in the right section of Figure 3.2.
A sensorgram can be obtained in at least three ways: (1) reectivity shift
versus time, (2) angle shift versus time or (3) wavelength shift versus time. In
Figure 3.3, a sensorgram of analyte binding versus time is presented, together
with a rotated view of an SPR dip. While the SPR angle is the most
sensor
chip
SPRoptics
liquid handling
system
2
3
1
Figure 3.1 Schematic view of the three main units of an SPR system: 1, SPR optics; 2,
liquid handling unit; 3, sensor chip. The biomolecular interaction takes
place on the wet side of the sensor.
36 Chapter 3
representative parameter of the biomolecular interaction process in time, one
should follow the angle of the SPR minimum (dip). As the dip shifts in the left
section from A to B, the SPR dip is followed in the right section of Figure 3.3.
The angle shift as a function of time is shown called the sensorgram.
Time (s) Angle in m
A B
X
X
1
X
11
Reflectivity
in %
Figure 3.2 Shift of the SPR dip from A to B. A derivative parameter of the shift of the
SPR curve is the reectivity (y-axis) measured at a xed angle at the left ank
of the SPR dip (see arrow showing reectivity shift of X
1
to X
11
). In the
sensorgram on the right the reectivity shift as a function of time is shown.
Time (s) Reflectivity
Angle in m
A
B
X
X
1
X
11
Figure 3.3 Rotated presentation of the SPR dip (left section) that forms directly the
pointer of the sensorgram (right section). The angle shift of the SPR dip is
determined (left, A to B), followed by plotting the angle of the SPR
minimum in the sensorgram vs. time (right). Here the SPR dip minimum of
the initial curve (A) shifts with time towards a larger angle (B).
37 SPR Instrumentation
Usually, the SPR curve is measured or tted and the angle of the SPR dip is
determined, followed by plotting the angle of the SPR minimum in the
sensorgram (right part of Figure 3.3). A sensorgram can be obtained from a
monochromatic light source using either the reectivity change at xed angle
(y-axis in %R) or the SPR dip shift (y-axis in RU or millidegrees), as illustrated
in Figures 3.2 and 3.3, respectively.
However, if a sensorgram of an equal experiment is either plotted as
reectivity (Figure 3.2) or as SPR angle (Figure 3.3) on the y-axis versus time,
the two sensorgrams are not equal and large deviations of the presented binding
curve may occur! A reectivity change cannot be exchanged with a shift of the
SPR dip. From theory as described in Chapter 2, the amount of biomolecules
accumulated in the evanescent eld results an (almost) linear shift of the SPR
dip. Therefore, only real-time measurement of the SPR dip as indicated in
Figure 3.3 is the preferred mode of operation [3]. Reectivity change is only a
derivative parameter of this shift. Therefore, optimally the SPR dip shift should
be detected continuously to follow changes at the surface, e.g. a biomolecular
interaction event [4]. The angle shift can be expressed in dierent units in the
various instruments, for example response units (RU, Biacore [5]) or in milli-
degrees (IBIS and Eco Chemie systems [6]), which directly depict the SPR angle
shift, and also in percentage reectivity (%R) (GWC Technologies system,
Genoptics, K-MAC [7,8]) or wavelength shifts (Thermo system, Granity
Pharmaceuticals [9]). Because of a near-linear relationship between the amount
of surface-bound protein and the SPR signal, the sensorgram of the SPR dip
shift provides quantitative, real-time data.
Errors in the position of the starting angle on the initial SPR curve can result
in a highly unreliable sensorgram. Importantly, a bulk refractive index shift or
background response will be generated if the refractive index of the medium
changes, e.g. due to dierences in solvent. Reectivity measurements should be
checked with the measurement of the angle- (or wavelength-) resolved SPR
curve. Moreover, the shape of SPR dip can reveal the quality of the sensor
surface. For example, if inhomogeneities or agglutinates (particles) are present
on the surface, the dip becomes shallower and broader as a result of spatial
dierences in resonance conditions within a sensing area. Therefore, checking
the SPR dip is a prerequisite for good SPR measurement practice. However,
most of the (cost-eective) instruments described in this chapter only support
reectivity measurements.
3.2 SPR Optics
SPR instrumentation can be congured in various ways to measure the shift of the
SPR dip. In general, three dierent optical systems are used to excite surface
plasmons: systems with prisms, gratings and optical waveguides, as explained
in Chapter 2 and more extensively by Homola [10]. Most widespread are
instruments with a prism coupler, also called the Kretschmann conguration
[11]. In this conguration, a prism couples p-polarized light into the SP lm and
38 Chapter 3
reects the light on to a light intensity detecting device, e.g. a photodiode or even a
camera. This conguration can be further divided into three subgroups: fan-
shaped beam, xed-angle and angle scanning SPR instruments, as explained in
the following sections. In instruments with a grating-coupler [12], light is reected
at the lower refractive index substrate. In practice, this means that light travels
through the liquid before photons generate surface plasmon waves as similar set-
up to ellipsometer instruments [13]. Besides the grating couplers, some instru-
ments apply optical waveguide couplers [14] or apply the SPR wavelength shift as
a result of the biomolecular interaction process [15]. In the following sections,
basic features and characteristics of the dierent optical SPR systems are treated.
In the nal section, SPR imaging instruments will be discussed [16].
3.2.1 Fan-shaped Beam
In a fan-shaped beam instrument, a converging or diverging beam of
p-polarized light is coupled in the higher refractive index medium using a
cylindrical or triangular prism. In a converging beam fan-shaped instrument,
the beam is focused on to an innitely narrow line on the sensor chip. A
photodiode array is used to detect the reected diverging beam with the SPR
dip. A line on the sensor can be imaged on a camera by measuring, e.g., 420
sensor spots. This principle has been successfully applied in Biacore instru-
ments since 1990 Figure 3.4. Other instruments in this category often use a
diverging fan-shaped beam, resulting in a not spatially dened location of the
SPR dip on the sensor chip: the position of the minimum walks over the
sensor chip while the biomolecular interaction process proceeds.
Figure 3.4 In a fan-shaped beam instrument the SPR dip can be followed in real time
without moving parts. Here the reected beam is drawn reciprocal meaning
that black corresponds with high intensity while white has low intensity.
3.2.2 Fixed Angle
Fixed-angle SPR instruments (Figure 3.5) measure the reectivity on the left ank
of an SPR curve resulting from the shift of the SPR dip, as shown by the vertical
line between X
1
and X
11
in Figure 3.2. In many instruments the angle can be
adjusted to nd the inection point of the SPR curve in order to obtain the
maximum reectivity change. Sometimes a stepper motor-operated angle positi-
oner (or mirror) is applied to automate the setting of the (optimal) starting angle.
An inection point can be determined from the second derivative of the SPR
curve, which gives the user the most sensitive starting angle or highest slope of the
reectivity shift versus angle shift. However, two dierently treated spots usually
require dierent optimal starting angles and therefore the reectivities cannot be
quantitatively compared without correction for the shape of the SPR curve.
3.2.3 Angle Scanning
The SPR curve can be scanned fast (see the double arrow in Figure 3.6) in a
so-called angle scanning SPR instrument. An angle-controlled mirror is applied to
follow the position of the SPR dip in real time. The surface of the sensor chip is at
a certain angle in full resonance with the advantage that the area of the light beam
is averaging the reectivity and therefore potential defects of homogenous coated
sensor chips. To acquire the data of a full scan very fast, an angle-controlled
mirror (or angle scanner) is preferred. Instruments of Eco Chemie (SPRINGLE
and ESPRIT), for example, scan the full SPR curve 76 times per second. In this
instrument, averaging of the acquired dip positions during the time interval is
applied, resulting in an improvement of the signal-to-noise ratio by the square
root of the number of averages. For example, an interval time of 1 s with 76 scans
Figure 3.5 In a xed-angle SPR instrument the reectivity is detected at a set angle.
The angle position can be set manually or using an angle stepper motor.
40 Chapter 3
averaged results in an improvement of the signal-to-noise ratio of O76 or a factor
better than 8. The recently introduced SPR imaging instrument of IBIS Tech-
nologies (see Section 3.2.6 and Chapter 7) also applies such a scanning principle,
with additional, angle-controlled scanning features and curve-tting routines,
while measuring the angle shift of multiple regions of interests [17].
3.2.4 Grating Coupler
Although grating coupler instruments (Figure 3.7) are regarded as SPR instru-
ments, they are quite dierent. Whereas in prism-coupled SPR optical cells the
light never passes through the sample, in grating coupler instruments the light
passes through the sample solution in the ow cell, resulting in decreased
stability of the sensor signal in comparison with the Kretschmann
Figure 3.6 In angle scanning SPR instruments an angle scanner scans the sensor
surface in a small angle window (e.g. 51) very fast. At a certain moment
the total surface, not just part of it, is in full resonance, resulting improved
sensitivity.
Figure 3.7 In a grating coupler, the SPR phenomenon is exploited dierently, as the
light beam travels through the ow cell.
conguration. To avoid unwanted internal reection eects, ow cells of
grating coupler SPR devices have an increased height. Therefore, the sample
volume in grating coupler instruments is normally larger than that in instru-
ments using the Kretschmann conguration, where the height of the ow cell
always exceeds the evanescent eld of the reected light. The attractive feature
of grating coupler SPR devices is the option of disposable gratings, which can be
mass produced by injection molding replication techniques. The optical con-
guration of grating coupler SPR instruments is the same as for ellipsometry
platform instruments (e.g. Nanolm; see section 3.4.2).
3.2.5 Other Optical Systems
Some other instruments measure the refractive index change in the evanescent
eld of the optical device by modulation of the coupling wavelength as shown
in Figure 3.8. A spectrophotometer is applied as detector to follow the shift of
the plasmon wavelength. A class of instruments uses modulation of the phase
or polarization [14]. This phenomenon is exploited in SPR-like resonant
mirror instruments [15]. The principle of resonant mirror measurements, as
depicted later in Figure 3.26, is that two phases of the reected light are
measured and, due to relative phase dierences of the two modes (TE and TM
components), a signal can be obtained as a result of refractive index changes in
the evanescent eld. Although plasmons are not generated and the enhance-
ment of the evanescent eld does not occur, multiple reections in the resonant
mirror contribute to a high-sensitivity measurement of refractive index changes
in the evanescent eld. Because the system utilizes a less inert oxide layer than
gold, more uncontrolled drift behavior can be observed than in conventional
Figure 3.8 A wavelength interrogation-based SPR device applies a polychromatic
light source to follow the wavelength-dependent SPR shift, i.e. the color of
the light changes as a result of the biomolecular interaction.
42 Chapter 3
gold-based SPR systems. Because of this unreliable drift, the sensitivity of this
instrument is lower than in other real SPR systems.
Interferometric read-outs of optical waveguides also use the refractive index
change as a result of the biomolecular interaction event at the surface of the
device. An evanescent eld is formed when reection occurs at the medium of
higher refractive index. Although physically SPR-like instruments including
interferometers might be more sensitive than SPR-based instruments, they lack
the inertness of the gold surface. The oxides with the required refractive indices
often show uncontrolled and unreliable baseline drifts including memory eects
from previous sample/salt additions. In the literature, various examples are
described [10]; however, only a few have been commercialized [18].
3.2.6 SPR Imaging Instruments
In SPR imaging instruments, the sensor surface is optically imaged by a
camera. Although some SPR instrument manufacturers use cameras for meas-
uring the SPR shift, a prerequisite for classication as an SPR imaging
instrument is that a microscopic view of the SPR sensor surface is generated.
Any inhomogeneities, surface coating defects, errors of spotting, including e.g.
missing spots, and eects of drying of the surface can be directly observed in the
microscopic reectivity image of the SPR sensor surface. If, for example, air
bubbles adsorb on the surface or appear in the ow system, it can be seen
directly on the SPR image monitor. Prior to performing the analysis cycle,
immobilization of capturing entities or ligands is usually carried out o-line by
spotting selective ligands in a microarray format.
2
In 1988, Knoll
3
et al.
described SPR microscopy [19], and later Kooyman
4
et al. established the
principles of SPR imaging in reectivity mode as well. Berger et al. showed that
in total 16 dierent sensor spots could be imaged [20] in a two-dimensional
array of antibodyantigen sensor surfaces in real time.
A serious drawback of many SPR imaging instruments is that they measure
reectivity as a parameter derived from the refractive index in the evanescent
eld. For accurate kinetic measurements of biomolecular interactions, only the
shift of the SPR dip of the microarray spots is reliable, because an optimal xed
angle position based on the inection point of the SPR dip can only be found
for one spot and not for hundreds of spots. Using the scanning angle principle,
the recently introduced IBIS microarray imaging instrument can follow the real
shift of the SPR dip of all the regions of interests simultaneously even when
spots deviate 42000 millidegrees in resonance angle, as shown in Chapter 7.
The in-house instrument of Granity Pharmaceuticals images the wavelength
shifts of the biomolecular interactions [21]. An example of a grating coupler
imaging SPR instrument is the FLEXChip of Biacore (Section 3.5).
2
A microarray contains a number of predened selective regions or spots regularly ordered in rows
and colums.
3
Author of Chapter 9 in this volume.
4
Author of Chapter 2 in this volume.
3.2.7 General Optical Requirements for SPR Instruments
The dynamic range of an instrument in the Kretschmann conguration for
studying biomolecular interactions is determined by its range of angles. The
larger the dynamic range, the smaller the angle resolution of an instrument
should be in order to maintain the required sensitivity. Therefore, an instrument
is usually built to zoom in to a small angle window (e.g. 2000 millidegrees) to
measure tiny angle shifts, e.g. of the order of millidegrees. However, if the angle
range is very small, the SPR dip might be outside the available range of angles.
A way to overcome this shortcoming is to apply the wavelength mode (see
Figure 3.8), with the disadvantage that the penetration depth of the evanescent
eld and the propagation length of the plasmon are not constant for dierent
frequencies of light. The penetration depth of the evanescent eld, and also the
lateral resolution of SPR, are dependent on the wavelength of the incident light
(see also Chapter 2). The lateral resolution at 680 nm wavelength incident light is
about 10mm [17]. If dierent wavelengths are used to excite SPR, the instrument
needs a complex algorithm to compensate for the variable volume of the
evanescent eld. Hence most instrument manufacturers choose a single wave-
length window for SPR excitation. In every instrument the polarity of the excited
light needs to be kept perpendicular to the surface (p-polarized light). This
requires trimming of the polarization lter at full resonance condition to obtain
an optimal reectance minimum. In some instruments, e.g. GWC [22], not only
p-polarized light (perpendicular to the surface, but also s-polarized light (in plane
with the surface) is needed to calibrate the intensity of the light source.
An additional experimental requirement is to provide refractive index
matching between the coated gold surface of the sensor chip and the optical
element (prism, hemisphere, etc.). Biacore solved this problem by applying an
optogel
5
to coat the prism and ensure refractive index matching. In other
instruments, an optical matching oil is used; however, it is a tedious task to
replace a sensor chip and clean the prism between measurements. Other
manufacturers have solved the optical matching problem by applying costly,
disposable, gold-covered prisms.
At present, gold on sensor chip surfaces is the gold standard in SPR.
Although from a physical point of view silver is better, gold provides more
chemical inertness. Attempts to protect the thin, non-inert silver layer have so far
failed and the performance of these sensors usually decreases rapidly to an
unacceptable level. Protection of the silver with, e.g., deposited oxides leads to
unwanted memory eects and drifts of the sensor signal. For this reason,
currently all manufacturers apply the physically second-best metal for SPR: gold.
The refractive index of aqueous liquids is highly temperature dependent: at
least 14 millidegree shift (or 115 Biacore resonance units) per degree Celsius
is observed. Therefore, temperature control of the instrument is essential.
Temperature stabilization (better than 0.01 1C) is a prerequisite for reliable
measurements in the millidegree range or in refractive index ranges below 10
4
.
5
Proprietary product of Biacore.
44 Chapter 3
Moreover, reference measurements are essential to reject common mode eects,
e.g. temperature drift or fast temperature jumps and also a bulk refractive index
shift, but cannot compensate for the temperature eect on the anity constant
of biomolecular interactions.
In order to determine accurately the kinetics of biomolecular interactions [23],
the shift of the SPR dip is the parameter that reects best the amount of
accumulated mass at the sensor surface, usually expressed as pgmm
2
. In Section
3.1.1 we pointed out that the measurement of reectivity change at a xed angle
position of the SPR curve is an inaccurate parameter and is not appropriate for
quantitative determination of the refractive index change in the evanescent eld of
the sensor surface. Many cost-eective instruments on the market apply the
sensorgram reectivity (in %R), including most of the SPR imaging instruments.
The quality of the optical components and optical alignment of the beam
including lenses, noise of both light source and photodetector or camera all
contribute to the quality of the measurements. Finally, the software including
the algorithms for averaging, subtracting or eliminating raw data points and
the performance of hardware contribute to the quality of the SPR instrument in
terms of sensitivity, repeatability, accuracy and robustness.
3.3 SPR Liquid Handling Systems
In combination with the optical unit and sensor chips, the liquid handling system
forms a vital part of SPR measurement systems. As shown in Figure 3.1, the
bottleneck remains the integration of these parts in one instrument provided that
the quality of the weakest part will be the limiting factor for the overall quality of
the system. For instance, high-quality optics can never compensate for a bad-quality
sensor chip. Additionally, the way in which the sample is exposed to the sensor
surface determines the kinetic proles in terms of rate constants, mass transport
limitation, stagnant layer and diusion gradient and depletion at the surface. Three
main liquid handling systems can be identied: ow cells, cuvettes and microuidic
(bio)chips. The rst two systems are described in this section. An outlook on SPR
microuidic biochips or lab-on-a-chip devices is given in Chapter 12.
In most instruments, ow systems are applied with various degrees of
automation ranging from simple to highly automated cartridges. Samples can
be transported using syringes or peristaltic pumps, either with or without
pneumatic valves and sample loops. Also, cuvettes can be operated automat-
ically using a liquid handler. In cuvette systems, dierent mixing methods are
applied to stir the sample solution. The cuvette can be described as a multi-
parameter, controllable batch reactor, in which binding events take place at the
bottom or sensor surface of the cuvette.
3.3.1 Flow Cell Systems
In SPR instruments, liquids are transported into the ow cell to study (bio)
molecular interactions [24] at the sensor surface. Peristaltic or syringe pumps
are applied to pump the liquid along the sensor chip surface. Some instruments
apply automatic injection systems, denitely improving performance. In ow
cells with laminar ow the sample is sometimes separated from the buer by an
air bubble; however, air bubbles may adhere and aect the surface properties of
the sensor chips if hydrophobic coatings are used.
Flow cells are formed by pressing a micromachined device with preformed
microuidic channels against the sensor surface. In 1990 Biacore introduced a
exible microuidics system for its SPR technology based around integrated micro
uidics cartridges (IFC) technology (Figure 3.9). The IFC allows analyte to pass
over the sensor surface in a continuous, pulse-free and controlled ow, maintaining
constant analyte concentrations at the sensor surface. A liquid handling system is
used to automate the biospecic interaction analysis procedures.
Three major ow cell congurations are generally applied in SPR systems
(Figure 3.10). Most commonly used are the planar ow cells. Conned wall-jet
Liquid Handling
IFC channels IFC
Flow cells, formed by contact of the IFC
on the sensor surface
Sensor surface
Glass slide
Prism
Minaturized
system
Low volume of
reagents
Intergrated and
automated liquid
handling
Buffer
Sample
Valve
Figure 3.9 The integrated microuidic cartridge (IFC) of the Biacore 3000 instru-
ment. Flow cells are formed when a microuidic cartridge is pressed
against a sensor surface (top). Pneumatic valves are used to guide the
sample in the injection loop followed by owing the sample over the
sensor surface.
46 Chapter 3
ow cells and hydrodynamic addressing ow cells are less common. A planar
ow cell contains a simple inlet and outlet and a single broad channel through
which the sample ows and interacts with the sensor surface. Owing to the
small dimensions of the chamber
6
and the resulting low Reynolds numbers, the
ow is laminar [25]. Usually, the ow cell is inserted in the SPR instrument after
the sensor surface has been prepared by the user; in some congurations also
the immobilization cycle can be performed in a closed loop. The pneumatic
system of a microuidic cartridge in the Biacore instruments (Figure 3.9) opens
and closes valves in user-dened protocols to control the ow in the ow cell.
Sample loops can be manufactured at a very close distance to the sensor
surface, minimizing dead volumes and dispersion, creating clearly dened
injection plugs. Small deviations are related to the dead volumes (o1 ml) and
to dispersion between ow cells, if e.g. four parallel ow cells are used.
Wall-jet ow cells are mainly applied when high mass transport to the sensor
surface is required at very low ows. In a conned wall-jet ow cell, the direction
of the jet is radial along the sensor surface. Wall jet cells are often used to
monitor fast surface processes [26]. Polar coordinates (circular approach) are
applied to describe the radial gradient of ow velocity. As the ow decreases
with increase in the radius of the wall-jet ow cell, the eect of ow rate on the
biomolecular interaction process can be studied in one experiment on uniformly
coated sensor discs using SPR imaging instruments.
Hydrodynamic addressing can be used for the simultaneous measurement of
multiple interactions in a single ow cell and is shown later in Figures 3.35 and
3.36 [27]. By adjusting the relative ow at the two inlets (one for the immo-
bilized partner and one for buer), liquids can be directed to dierent address-
able detection spots. The ow cell design allows rapid and ecient switching of
ow between buer and sample solutions and the transverse arrangement of the
detection spots ensures simultaneous access of sample to all spots. As there is
no lag time between interactions, highly accurate reference subtraction allows
the measurement of very rapid kinetics. By immobilizing several ligands in one

Figure 3.10 Top and side views of three ow cells used in SPR instruments. The
hydrodynamic addressing ow cell is discussed in Section 3.5.
6
The smaller the sample volume, the better.
ow cell, comparative binding properties can be directly examined under
identical experimental conditions.
3.3.2 Cuvette Systems
Cuvette liquid handling systems consist of an open container lled manually or
automatically by a liquid handling robot or liquid handler (Figure 3.11). The
cuvette must be mounted leak-free on the sensor chip and a correct congu-
ration of the optics (from the bottom) is required to apply the open container or
cuvette. The biomolecular interaction takes place on the replaceable sensor chip
at the bottom of the cuvette. If the liquid is not mixed, sensorgrams are
unreliable due to inhomogeneous, uncontrolled molecular transport to the
surface. Therefore, cuvette systems are usually equipped with a mixing system.
In contrast with ow cells, cuvette systems are not prone to clogging, hence
liquid samples with solid particles can be measured in cuvette systems, provided
that the sensor surface is not damaged by impinging particles. Examples of such
samples are fermentation media, blood plasma, cell cultures and food products.
Because the sample stays in the cuvette during the entire biomolecular inter-
action process or analysis cycle, the undiluted sample can be recovered almost
completely.
In principle, a sample volume in the range of 25 ml is sucient to study
biomolecular interactions for up to an hour. A disadvantage of the cuvette
system is the open architecture, allowing uncontrolled evaporation of the
Figure 3.11 Cuvette of the ESPRIT, Eco Chemie Netherlands, system with two
containers, two drain connections for each container and two inlet
connections. The cuvette can be automatically lled and drained while
mixing takes place with two needles inserted in the containers connected
to two syringe pumps.
48 Chapter 3
sample solution. When low-concentration samples are used, depletion of
the analyte occurs and kinetic models should compensate for this eect, as
described in Chapter 5. Ideally, an SPR instrument would be compatible with
both cuvette or ow cell, allowing the user to choose for the best performance
and quality of the biomolecular interaction process of interest.
Various mixing systems can be used to agitate the sample in the cuvette,
including stirrers [28], vibrating acoustic plates [30] and aspirate/dispense mixing
[6]. In principle, every device can be applied that controllably and eectively
induces turbulence to homogenize the sample solution and to minimize the
stagnant layer at the sensor surface. For example, in early resonant mirror
instruments the sample solution was stirred by a high-speed rotating stirrer [29].
A disadvantage of this system is that mass transport to the surface is not
eectively controlled even at fast rotation speeds. Later, a vibrating plunger/
piston was developed, showing better mass transport characteristics [30].
Mixing can also be achieved by a syringe pump with a controllable automatic
aspirate/dispense mixing needle.
7
In these instruments, a syringe pump is
constantly aspirating and dispensing the sample or buer in the cuvette during
real-time measurements to obtain reproducible hydrodynamic conditions at
the sensor surface. During a mixing cycle of set volume and speed, rst part of
the cuvette volume is aspirated. During the dispense action, a free wall-jet of the
sample solution can be forced to ow in the diusion layer of the surface. As a
result, the mass transport can be dramatically increased. Relevant hydrody-
namic parameters of an aspirate/dispense mixing system are sample volume,
mixing volume, which should be part of the sample volume, speed or frequency
of mixing, diameter of the nozzle (inner diameter of the needle), distance and
position of the nozzle to the sensor surface, diameter of the cuvette, viscosity
and temperature of the sample solution.
The cuvette needs to be drained after the interaction process, but the surface
should not become dry, therefore hydrophilic coatings are used to prevent the
sensor surface from being exposed to air. If the hydrogel dries out, however,
irreproducible eects will occur.
Physical transport determines how fast the sample molecules are transported
to the surface. Mass transport limitations arise when the concentration of the
analyte at the sensor surface is lower than the sample bulk concentration (for a
detailed description, see Chapter 4). In aspirate/dispense mixing systems, mass
transport to the surface is greatly increased and the diusion-controlled stag-
nant layer at the surface is strongly reduced. The process can be described by
the dynamic free wall-jet principle rst published by Glaubert [31]. As various
parameters aect the biomolecular interaction process, the reproducibility of
the measurement can be maintained best in automated operation at constant
hydrodynamic parameters. The reader is referred to Chapter 5 for details on
kinetic measurements and kinetic evaluation software for biomolecular inter-
action parameters, and a practical example is given in Chapter 9.
7
The systems of IBIS, SPRINGLE and ESPRIT of Eco Chemie and the instruments derived from
the PLASMOON system of Biotul, e.g. by Plasmonic Biosensor.
3.4 SPR Instruments: State of the Art
Currently, SPR instruments are available from various manufacturers. In
Table 3.1 and the following sections, a short description is given of commercially
available SPR instruments, sorted by their optical conguration, describing the
main features and benets of the instruments.
In addition, other instruments can be identied that detect biomolecular
interactions in real time and label free which are not directly based on SPR.
Information on these products and manufacturers is compiled in Table 3.2. It
should be noted that this list might not be complete but it provides an
illustration on the range of products and manufacturers in the label-free
biosensor area.
3.4.1 Examples of Fan-shaped Beam SPR Instruments
Biacore
8
(Uppsala, Sweden) dominates the SPR market with more than 90%
of installed products and 87% of the publications in 2005 [39] and dictates
the standards serving as reference for all other instrument manufacturers
(Figure 3.12). The principle of fan-shaped SPR instruments has been used by
Biacore since the launch of the rst product in 1990. In Biacore instruments a
light-emitting diode (l 760 nm) is used and a convergent light beam reects
at an exact position at the sensor surface. A photodiode array is used to
determine accurately the SPR dip position. Extending the fan of light into a
wedge and using a two-dimensional detector allows detection along a line on
the sensor surface. The benet of using a fan of light and a linear array detector
is that no moving parts are required to carry out SPR assays. A computer
algorithm calculates the exact position of the minimum to a fraction of the size
of a single diode. The accuracy of the optical system corresponds to about
0.1 millidegree shift of the SPR angle. Biacore applies an optogel to optically
match the sensor chip with the cylindrical prism. The latest instruments have 20
sensor spots in four ow cells. With dierent probes immobilized at dened
spots along a line, simultaneous measurement of dierent biomolecular inter-
actions is realized. Biacore applies several types of ow cells connected to a
microuidic cartridge.
The BI-SPR of Biosensing Instruments (BI) (Tempe, AZ, USA) uses an
innovative method to detect the SPR angle, key to the high performance of the
instrument (Figure 3.13). A shift of the surface plasmon resonance angle results
in a shift in location of the SPR dip on the detector surface. The two ow
channels can be used in conjunction with two valves for simultaneous meas-
urement of two samples. Alternatively, one channel can be used for back-
ground subtraction. The system provides a quick and easy setup with various
cell modules for DNA sequencing, proteinprotein interaction, ligand/receptor
recognition, drug development applications and the optional EC cell module
(not included with the system) for electrochemical SPR measurements.
8
Acquired by GE Healthcare in 2006.
50 Chapter 3
T
a
b
l
e
3
.
1
O
v
e
r
v
i
e
w
o
f
S
P
R
i
n
s
t
r
u
m
e
n
t
s
a
n
d
t
h
e
i
r
m
a
n
u
f
a
c
t
u
r
e
r
s
[
3
2
,
3
3
]
.
T
y
p
e
S
P
R
s
y
s
t
e
m
M
a
n
u
f
a
c
t
u
r
e
r
U
R
L
a
d
d
r
e
s
s
F
a
n
s
h
a
p
e
d
,
c
o
n
v
e
r
g
i
n
g
B
i
a
c
o
r
e
p
r
o
d
u
c
t
l
i
n
e
a
B
i
a
c
o
r
e
w
w
w
.
b
i
a
c
o
r
e
.
c
o
m
F
a
n
s
h
a
p
e
d
,
c
o
n
v
e
r
g
i
n
g
B
I
-
S
P
R
1
0
0
0
B
i
o
s
e
n
s
i
n
g
I
n
s
t
r
u
m
e
n
t
w
w
w
.
b
i
o
s
e
n
s
i
n
g
u
s
a
.
c
o
m
F
a
n
s
h
a
p
e
d
,
d
i
v
e
r
g
i
n
g
S
e
n
s
i
Q
N
o
m
a
d
i
c
s
w
w
w
.
n
o
m
a
d
i
c
s
.
c
o
m
/
F
a
n
s
h
a
p
e
d
,
c
o
n
v
e
r
g
i
n
g
S
P
R
-
2
0
D
K
K
-
T
O
A
w
w
w
.
d
k
k
t
o
a
.
n
e
t
/
F
a
n
s
h
a
p
e
d
,
d
i
v
e
r
g
i
n
g
S
R
7
0
0
0
R
e
i
c
h
e
r
t
w
w
w
.
r
e
i
c
h
e
r
t
a
i
.
c
o
m
/
F
a
n
s
h
a
p
e
d
,
d
i
v
e
r
g
i
n
g
P
l
a
s
m
o
n
i
c
P
l
a
s
m
o
n
i
c
B
i
o
s
e
n
s
o
r
w
w
w
.
p
l
a
s
m
o
n
i
c
.
d
e
/
F
i
x
e
d
a
n
g
l
e
S
P
T
M
R
e
s
o
n
a
n
t
P
r
o
b
e
s
w
w
w
.
r
e
s
o
n
a
n
t
-
p
r
o
b
e
s
.
d
e
/
F
i
x
e
d
a
n
g
l
e
S
P
R
6
7
0
M
M
o
r
i
t
e
x
w
w
w
.
m
o
r
i
t
e
x
.
c
o
m
F
i
x
e
d
a
n
g
l
e
M
u
l
t
i
s
k
o
p
O
p
t
r
e
l
G
B
R
w
w
w
.
o
p
t
r
e
l
.
d
e
F
i
x
e
d
a
n
g
l
e
B
I
O
S
U
P
L
A
R
-
3
2
1
A
n
a
l
y
t
i
c
a
l
m
-
S
y
s
t
e
m
s
w
w
w
.
b
i
o
s
u
p
l
a
r
.
d
e
F
i
x
e
d
a
n
g
l
e
S
e
n
s
i
a
b
-
S
P
R
S
e
n
s
i
a
w
w
w
.
s
e
n
s
i
a
.
e
s
F
i
x
e
d
a
n
g
l
e
+
i
m
a
g
i
n
g
S
P
R
L
A
B
a
n
d
S
P
R
i
K
-
M
A
C
w
w
w
.
k
-
m
a
c
.
c
o
.
k
r
/
F
i
x
e
d
a
n
g
l
e
+
i
m
a
g
i
n
g
N
a
n
o
l
m
E
P
3
N
a
n
o
l
m
S
u
r
f
a
c
e
A
n
a
l
y
s
i
s
w
w
w
.
n
a
n
o
l
m
.
d
e
/
A
n
g
l
e
s
c
a
n
n
i
n
g
S
P
R
I
N
G
L
E
/
E
S
P
R
I
T
E
c
o
C
h
e
m
i
e
w
w
w
.
e
c
o
c
h
e
m
i
e
.
n
l
G
r
a
t
i
n
g
c
o
u
p
l
e
r
F
L
E
X
C
h
i
p
B
i
a
c
o
r
e
w
w
w
.
b
i
a
c
o
r
e
.
c
o
m
W
a
v
e
l
e
n
g
t
h
S
P
R
1
0
0
T
h
e
r
m
o
E
l
e
c
t
r
o
n
C
o
r
p
.
w
w
w
.
t
h
e
r
m
o
.
c
o
m
R
e
s
o
n
a
n
t
m
i
r
r
o
r
I
A
s
y
s
N
e
o
S
e
n
s
o
r
s
w
w
w
.
n
e
o
s
e
n
s
o
r
s
.
c
o
m
/
I
m
a
g
i
n
g
S
P
R
i
m
a
g
e
r
I
I
G
W
C
T
e
c
h
n
o
l
o
g
i
e
s
w
w
w
.
g
w
c
t
e
c
h
n
o
l
o
g
i
e
s
.
c
o
m
I
m
a
g
i
n
g
S
P
R
i
-
P
l
e
x
G
e
n
O
p
t
i
c
s
/
H
o
r
i
b
a
J
o
b
i
n
Y
v
o
n
w
w
w
.
g
e
n
o
p
t
i
c
s
-
s
p
r
.
c
o
m
/
I
m
a
g
i
n
g
P
r
o
t
e
O
n
X
P
R
3
6
B
i
o
-
R
a
d
L
a
b
o
r
a
t
o
r
i
e
s
w
w
w
.
b
i
o
-
r
a
d
.
c
o
m
I
m
a
g
i
n
g
M
u
l
t
i
S
P
R
i
n
t
e
r
T
o
y
o
b
o
w
w
w
.
t
o
y
o
b
o
.
c
o
.
j
p
I
m
a
g
i
n
g
P
r
o
t
e
o
m
i
c
P
r
o
c
e
s
s
o
r
L
u
m
e
r
a
w
w
w
.
l
u
m
e
r
a
.
c
o
m
/
I
m
a
g
i
n
g
/
w
a
v
e
l
e
n
g
t
h
P
l
a
s
m
o
n
I
m
a
g
e
r
G
r
a
n
i
t
y
T
e
c
h
n
o
l
o
g
i
e
s
w
w
w
.
g
r
a
n
i
t
y
.
c
o
m
I
m
a
g
i
n
g
/
s
c
a
n
n
i
n
g
I
B
I
S
i
S
P
R
I
B
I
S
T
e
c
h
n
o
l
o
g
i
e
s
w
w
w
.
i
b
i
s
-
s
p
r
.
n
l
a
B
i
a
c
o
r
e
p
r
o
d
u
c
t
l
i
n
e
,
i
n
c
l
u
d
i
n
g
:
B
i
a
l
i
t
e
,
B
i
a
c
o
r
e
J
,
-
X
,
-
1
0
0
0
,
-
2
0
0
0
,
-
3
0
0
0
,
-
C
,
-
S
5
1
,
-
Q
,
-
A
1
0
0
,
-
T
1
0
0
,
-
X
1
0
0
.
Nomadics (Dallas, TX, USA) markets SPR technology in a miniaturized
SPR sensing device named SensiQ, in collaboration with Texas Instruments
(TI) (Figure 3.14). TIs SPREETA SPR sensor, coupled with a high-resolution
electronic interface and ow equipment in the SensiQ device, allows analysis
using Windows-based computers. SensiQ is a dual-channel, semi-automated
SPR system which utilizes advanced microuidics and proven surface attach-
ment methodologies. Concepts can be evaluated quickly with minimal or no
hardware and software development.
The Model SPR-20 instrument of DKK-TOA (Takadanobaba, Tokyo,
Japan) performs measurements using a 101 convergent beam directed into
the prism and a CCD camera as the detector (Figure 3.15). The SPR angle can
be adjusted manually between 35 and 851. Standard equipment includes both
ow-type and cuvette-type analysis cells. Measurements at two sensor spots in a
ow cell can be performed without the need to move the mobile stage because
the convergent beam has sucient angle resolution. The instrument is suited for
both gas- and liquid-phase samples and has a resolution of 3 millidegrees. The
software associated with this instrument is run on a standard Apple Macintosh
personal computer.
The Reichert (New York, USA) SR7000 DC surface plasmon resonance dual-
channel spectrometer uses a fan-shaped divergent beam, unlike in Biacore
instruments, where a convergent fan shaped beam is used (Figure 3.16). The
advantage of this setup is that no moving parts are required. However, SPR does
not occur at a xed position on the sensor but walks along the surface while
the biomolecular interaction is proceeding and therefore the SPR path is moving,
Table 3.2 Non-SPR instruments for real time and label-free monitoring of
biomolecular interactions.
Manufacturer and product Working principle URL address
Axela Biosensor: dotLab
System
Grating-based light
diraction
www.axelabiosensors.com
Fareld Scientic (UK):
AnaLight
Dual-polarization
interferometry
www.fareld-scientic.com
Silicon Kinetics: nanoporous
silicon (npSi)
Optical biosensor
substrate
www.siliconkinetics.com
SRU Biosystems (Woburn,
MA, USA): BIND system
Nanostructured
optical grating
www.srubiosystems.com
Corning: Epic system, high-
throughput optical
biosensor
Dynamic mass
redistribution
(DMR) eect
www.corning.com/
lifesciences
ForteBio: Octet system Bio-layer
interferometry
(BLI) optical ber
biosensor
www.fortebio.com
Maven Biotechnologies:
LFIRE
Ellipsometry in
Kretschmann
conguration
www.mavenbiotech.com
Akubio: acoustic biosensor Acoustic eect www.akubio.com
52 Chapter 3
limiting the use of small ow cells. The Peltier temperature-controlled instrument
is connected to a liquid handler (Endurance, Spark, The Netherlands).
Plasmonic Biosensor (Wallenfels, Germany)
9
acquired autosampler tech-
nology from Biotul (Munich, Germany) and developed a new, patented optical
detection system with disposable prisms (Figure 3.17). The optical system
generates surface plasmon waves by a defocused laser beam which reects at
the surface of a gold-covered prism. With this optical cell, adhesion phenomena
can be detected by using the SPR eect. The Plasmonic instrument allows the
real-time analysis of unpuried samples contained in an array of eight cuvettes
in parallel. The sample handling is carried out with a patented, removable,
accurately positioned pipette tip system with no cross-contamination.
Figure 3.12 Part of the SPR product line of Biacore: Biacore J, X, 3000, C, S51 and
X100. The other Biacore systems [A100, T100, Q (see Chapter 11) and
FLEXChip instrument, Section 3.5] are not shown here but separately
highlighted in Section 3.5.
9
Formerly Jandratek GmbH.
3.4.2 Examples of Fixed-angle SPR Instruments
The SPTM instrument of Resonant Probes (Goslar, Germany) is a simple SPR
spectrometer. The setup consists of a high-precision goniometer
10
, a versatile
and accessible sample holder, a laser and a detection unit based on lock-in
amplication. The lock-in detection makes the system insensitive to stray light,
even though the setup is completely open. SPTM is ideally suited for research
environments. The coupling angle, y
c
, is determined with an accuracy of 0.011.
Moritex (Myutron, Tokyo, Japan) has a complete product line of bioscience-
related laboratory instruments including an SPR platform
11
. The Moritex SPR
Figure 3.13 The BI-SPR of Biosensing Instruments.
Figure 3.14 Nomadics SensiQ with SPREETA chip. A divergent fan-shaped beam
reects at the sensing surface.
10
To determine a precise angular position.
11
Moritex Corporation acquired SPR technology from Nippon Laser and Electronics Lab in 2004.
54 Chapter 3
Figure 3.16 The SR7000 DC instrument of Reichert (right) connected to an auto-
sampler (middle) controlled by syringe pumps (left).
Figure 3.15 Photograph and scheme of the DKK-TOA SPR-20 instrument. The
angle position can be set in a large range.
Figure 3.17 The SPR instrument of Plasmonic Biosensor.
670M system operates with two sensing spots and employs a ow system using
a peristaltic pump. The system is currently under redesign before market
introduction.
Optrel GBR (Kleinmachnow, Germany) was founded in 1996 and focuses
on the design of instrumentation for the investigation of interfaces and thin
lms. The Multiskop of Optrel GBR is a unique, single, modular system
which incorporates surface plasmon spectroscopy (SPS) and ellipsometry in a
single instrument (Figure 3.18). The modular design permits many dierent
congurations.
The BIOSUPLAR series of Analytical m-Systems/Mivitec (Sinzing, Germany)
were introduced to the market at the end of the 1990s. At present, the third
generation is in production (BIOSUPLAR-321). The device has two optical
channels which can be used independently or as reference and sensing channels
for dierential measurements. The exible setup allows measurements of angle
dependence of the reected light, monitoring of the time dependence of the
resonance angle and measurements of light intensity at a xed angle. The open
system can be used for setting up an SPR experiment. Moreover, the
basic conguration allows practical training and tutorials for new SPR users
(Figure 3.19).
The b-SPR research platform of Sensia (Madrid, Spain) uses the Kretschmann
conguration to achieve total internal reection resonance conditions
(Figure 3.20). Polarized light from a laser diode reects at the gold surface of
two ow cells, with a volume of 300 nl each, suited for the simultaneous
measurement of two biomolecular interactions. The platform operates with a
sample and reference cell. An advantage is that the technology is simple from the
users point of view; however, alignment of the beamsplitter and guiding
reected light from the small ow cell make the system complicated from a
technology point of view. A prism and a multi-photodiode are located on
Figure 3.18 The Multiskop of Optrel GBR in a horizontal (left) and vertical (right)
optical setup. The instrument can be applied either in an SPR Kretschmann
conguration or in an ellipsometric setup.
56 Chapter 3
a concentric rotary stage with an angular resolution of 0.011. The b-test SPR
system has a refractive index detection limit of 10
5
.
K-MAC (Korea Materials and Analysis Corp., Daejeon, Korea) is a venture
company for analytical instruments. In 2003, K-MAC started the development
of SPR instruments with the SpectraBio2000, recently renamed SPR
LAB
, a
motor-operated, xed-angle instrument (Figure 3.21). A second instrument
(SPRi ) is now available based on SPR imaging technology for protein analysis
in drug discovery and disease diagnosis applications.
Figure 3.19 The Bio-Suplar is a compact instrument.
Figure 3.20 The b-SPR instrument of Sensia.
SPR
LAB
is an angle scanning instrument with gold-covered prisms; its liquid
handling system combines a microuidic ow cell with a precision syringe
pump. The instrument is characterized by a wide dynamic range combined with
high sensitivity to detect mass changes on the gold surface. It is suited to work
with strong acids and bases, organic solvents including DMSO and CCl
4
. The
stepper motor-operated angle position instrument can be tuned to follow the
angle shift as a function of time.
The recently launched SPRi imaging system is a reectivity-based SPR
imaging instrument with manual setting of the angle of incidence, designed
for rapid monitoring of the biochip sensor, an array of various biomolecules,
such as proteins, recombinant proteins, cells or other microorganisms. SPR
imaging is used for high-throughput analysis of bioanity interactions by
fabricating DNA/protein arrays on biochip gold surfaces. The company applies
a range of sensor chips, including prism coated gold slides.
Nanolm Surface Analysis (Go ttingen, Germany) developed the Ellipsome-
try Platform EP
3
with an additional SPR experimental option in the classical
Kretschmann conguration (Figure 3.22). The SPR cell is specially designed for
kinetic measurements of biomolecular interactions as an alternative for ellipso-
metric measurements on transparent substrates, e.g. glass slides (OptiSlides).
The sensitivity of the OptiSlide ellipsometric measurements, attained without
Figure 3.21 The SPR
LAB
(A) and the SPR imaging SPRi system (B) of K-MAC.
58 Chapter 3
the gold layer, is one order of magnitude less than that of conventional SPR.
However, the lateral resolution is at least 10 times better, permitting the label-
free study of the quality of microarray spots. The Nanolm EP
3
measures the
ellipsometric parameters Psi and Delta in addition to the reected light inten-
sity in the SPR mode of operation. Psi is analogous to the reected light
intensity in classical SPR, whereas the phase signal Delta gives additional
information on, e.g., surface roughness, exceeding the limitations of standard
SPR. The combination of SPR and imaging ellipsometry allows for quality
control of microarrays, as it measures surface morphology through 2D thick-
ness maps and 3D proles of the inner layers of stacked multilayers, in addition
to microarray spots with improved lateral resolution. The most frequent
application of SPR imaging ellipsometry is marker-free parallel kinetics
recording of up to 100 spots.
3.4.3 Examples of Angle Scanning SPR Instruments
The SPR system of EcoChemie (Utrecht, the Netherlands) applies angle
scanning optics. The minimum of the SPR dip is detected as fast as 76 Hz,
hence it is fair to say that this instrument detects the true SPR dip continuously.
The initial design of the instrument
12
has proven its reliability and exibility for
a decade. Redesigned by EcoChemie, the SPRINGLE is sold as a single-beam
instrument whereas the ESPRIT system operates with two channels and has
automated sample transfer with an autosampler (Figure 3.23). The SPR angle
resolution and noise levels are better than 0.1 millidegrees, similar to those of
Figure 3.22 The Ellipsometry Platform EP
3
including SPR option of Nanolm.
12
Previously sold by IBIS Technologies as IBIS I and IBIS II instruments.
the best high-end SPR instruments on the market. The cuvette-based instrument
applies aspirating-dispensing mixing through a syringe, whereas the interaction
process is followed in real time. EcoChemie sells an adaptor for using a Biacore
sensor chip with the instruments. A unique feature is the integrated AUTOLAB
electrochemical measurement workstation to perform electrochemical SPR
(E-SPR) measurements. In this application, the gold surface is used as an
electrode connected to the potentiostat output of the AUTOLAB workstation.
Real-time SPR can be combined with electrochemical measurements in one
experiment. The prices of both EcoChemie systems make them competitive
single- and double-beam type instruments. In the literature [38], Windsor
Scientic, which is a former distributor of Eco Chemie and IBIS Technologies,
is incorrectly denoted as the instrument manufacturer of IBIS instruments.
3.4.4 Examples of Grating Coupler SPR Instruments
The FLEXChip system
13
of Biacore (GE Healthcare) provides an open, array-
based platform for kinetic screening, allowing the simultaneous study of the
interactions of hundreds of proteins and peptides against a single sample
(Figure 3.24). The FLEXChip system utilizes a grating coupler for SPR
analysis and is also an imaging apparatus. For more details, see Section 3.5.4.
3.4.5 Examples of Other SPR Instruments
The SPR 100 module of Thermo Electron Corp. (Waltham, MA, USA) is based
on wavelength-specic SPR detection (Figure 3.25). An added Fourier trans-
form infrared (FT-IR) spectrometer module provides fully edged infrared
spectroscopic capabilities providing chemical information on the bound
Figure 3.23 The double-beam ESPRIT instrument (left) and the single-beam
SPRINGLE SPR scanning instrument of EcoChemie.
13
Originally from HTS Biosystems (Hopkinton, MA, USA).
60 Chapter 3
species. FT-IR-SPR measures the wavenumber corresponding to the minimum
reectivity, resulting in signicant sensitivity advantages over the traditional
angle shift-based SPR technology. The SPR 100 captures quantitative data
for an extraordinarily diverse range of applications. The unexceeded broad
dynamic range is due to the combination of broad spectral range with broad
incident light angles from B40 to 701. Measurements can be made in the liquid
or gas phase in a single instrument. The module was specically designed for
use with Thermo Electrons FT-IR spectrometers. At present, GWC Technol-
ogies (Madison, WI, USA) markets the SPR 100 and also the SPRimager II
instrument as described in the SPR imaging section (Section 3.4.6).
In the mid-1990s, the IAsys system of NeoSensors. (Sedgeeld, UK) was the
cuvette-based alternative to Biacore instruments and deserves special attention
here (Figure 3.26). It is still the second label free biosensor, with 40 publications
in 2005. The cuvette was mixed with a stirrer, later a vibrating acoustic plate,
Figure 3.24 Photograph including optical setup of Biacores FLEXChip instrument.
Figure 3.25 The SPR 100 module of Thermo Electron Corp. This instrument uses the
wavelength interrogation SPR principle (see Figure 3.8).
and operated with a liquid handler to inject the sample automatically into the
cuvette. Although the resonant mirror technology of the IAsys system
14
is not
aimed at SPR, it provides an alternative technology to measure refractive index
changes in the evanescent eld at the sensor surface (see Section 3.2.5). The
IAsys system is a user-friendly instrument with elaborate sensor surface chem-
istry; uidics (cuvettes) and applications are similar to those of SPR biosensors.
Sample well
Fibrin monomers
Vibro-stirrer
Fibrinogen layer
Ligand attachment surface
Resonant mirror
Low index resonant layer
Resonated
reflected light
Prism
Incident
laser light
Figure 3.26 The IAsys system of NeoSensors (from Ref. [30]). The resonant mirror
principle of the IAsys applies detection in the evanescent eld but is not
SPR.
14
Originally developed by Fisons (UK) and acquired by Thermo Instruments and Anity Sensors
(Cambridge, UK), now NeoSensors (Sedgeeld, UK), which is a division of the Fareld group.
62 Chapter 3
3.4.6 Examples of SPR Imaging Instruments
The SPRimager II of GWC Technologies (Madison, WI, USA), originally
developed in the group of Robert Corn
15
, contributor of Chapter 8 of this book
[34], captures data on the entire sensor surface simultaneously with a CCD
camera (Figure 3.27). A xed angle is set manually at the left angle ank of the
SPR dip. As described in Section 3.1.1, the optimal angle in terms of maximal
reectivity is in the inection point of the SPR dip. However, this angle can
only be set for a reference spot and not for all dierent spots simultaneously.
Under appropriate conditions, the SPR response is a linear function of the
surface coverage provided D%Rr10%. The instrument can be used to study
biomolecular interactions at dierent spots, for example in a 5 5 microarray.
The instrument provides various manually operated control features to check
the quality of the interactions. The degree of automation is limited and a single
peristaltic pump is used to pump the sample through a vertically positioned,
gold-covered sensor mounted in a ow cell. The instrument uses disposable
sensors of SF10 glass. These high refractive index substrates are not compatible
with standard optical glass substrates (e.g. K5 or BK7), hence the prism is also
made of SF10. Refractive index matching oil is used between the disposable
sensor and the prism. GWC Technologies also markets the Thermo Electron
product the SPR100 (see Section 3.4.5).
Figure 3.27 The SPRimager II of GWC Technologies. The SPR angle should be set
manually using a spindle.
15
Present aliation: University of California, Irvine, CA, USA.
GenOptics (Orsay, France) is a laboratory instrumentation provider of high-
performance systems with advanced optical imaging (SPRi) to quantify and
monitor biomolecular interactions. GenOptics instruments use the Kretschmann
conguration to excite surface plasmons and are equipped with a rotating mirror
for scanning precisely the angle of incidence in the SPR reectivity dip. This scan
allows the selection of the best reectivity performance of the sensor chip to
monitor proteinprotein interactions in real time. A broad monochromatic
polarized light illuminates the whole functionalized area of the SPRi-Biochip,
which is combined with a detection chamber. Information can be quickly
obtained from the interaction process while monitoring reectivity variations
against time. GenOptics currently markets two protein-array platforms: the
SPRi-Plex for large-scale, automated screening and SPRi-Lab+ with an open
conguration for biomolecular interaction experiments and application develop-
ment (Figure 3.28). Horiba Jobin Yvon (Longjumeau, France) obtained
exclusive, world-wide distribution rights for both the SPRi-Lab+ detector and
SPRi-Plex.
The ProteOn XPR36 protein interaction array system of BioRad Laboratories
(Hercules, CA, USA) is a SPR imaging biosensor with a multi-channel module
and interaction array sensor chip for analysis of up to 36 protein interactions in a
single injection step (Figure 3.29). The capability of the ProteOn XPR36 system
to generate rapidly a 6 6 interaction array between six ligands and six analytes
greatly increases the throughput, exibility and versatility of experimental design
for a wide range of biomolecular interaction studies. A901 mechanical switch
allows the placement of six ow lines perpendicular to each other on the sensor
surface. Prior to the analysis cycle, six dierent ligands can be immobilized on the
sensor surface. Then perpendicular to this immobilized ligands in lines, the user
can inject the analyte and acquire kinetic data in so-called one-shot kinetics of six
biomolecular interactions in six analyte dilutions in a single run.
Toyobo (Osaka, Japan) was founded in 1882 to manufacture high-tech
materials and currently oers products for the life sciences, including biosensor
equipment. Toyobos MultiSPRinter is an array-based SPR sensor platform
distributed in Japan. This system provides a complete detection system includ-
ing a spotter. The platform has been used to monitor DNA hybridization and
kinetics for protein binding on an array of related double-stranded DNAs. In
collaboration with RIKEN (www.rikenresearch.riken.jp), Toyobe launched the
so called Photo-linker Chip for the MultiSPRinter for fast microarray produc-
tion. The photo-linker chip is able to provide up to 96 low molecular weight
spots on a microarray.
The LFIRE from Maven Biotechnologies (Pasadena, CA, USA) allows pre-
cise, real-time measurements of specic interactions between molecular entities in
a microarray or well-plate format (Figure 3.30). These molecular entities can be
proteins, nucleic acids, lipids, small molecules such as drugs or steroids or even
whole cells. LFIREis based on ellipsometry, a technique that measures changes
in the polarization of light upon reection from the interface between materials.
However, the optical setup is dierent from the Nanolm conguration shown in
Section 3.4.2, but similar to the Kretschmann conguration commonly used in
64 Chapter 3
SPR instruments. As a real-time system, it monitors reactions as they happen,
providing kinetic information on biomolecular interactions.
The Proteomic Processor of Lumera (Bothell, WA, USA) uses SPR microscopy
in which a beam of light is directed on to a spot of a microarray (Figure 3.31). The
PEEK
tubing
Waste
Injection valve
Syringe pump
Buffer
Mirror
Optical
system 2
Optical
system 1
Flow
cell
SPRi
Biochip
Light
source
CCD
Camera
Polarizer
Figure 3.28 The optical setup of the GenOptics SPRi-Plex and SPRi-Lab+ system.
By changing the mirror angle an optimal reectivity contrast can be
obtained as shown in the image with 460 spots.
Figure 3.31 The Proteomic Processor of Lumera. A scalable beam is applied and the
reectivity can be monitored for thousands of spots simultaneously.
Figure 3.30 The LFIRE of Maven BioTechnologies is not an SPR imaging instru-
ment but an ellipsometry imaging system in attenuated total reection
conguration. Biomolecules in the evanescent eld will change the
polarization of the reected light.
A B A B
Figure 3.29 The ProteOn
TM
XPR36 of BioRad with crisscross microuidics. (A)
Ligands can be immobilized in 6 lanes. (B) The analyte can be passed
perpendicular in 6 lanes over the surface and detection at the cross
sections take place.
66 Chapter 3
reected light in SPR microscopy changes due to a chemical binding event. The
proprietary optics of the instrument allows simultaneous addressing thousands of
spots in the reectivity mode. The Proteomic Processor extends the power of SPR
to the analysis of high-density microarrays by use of a novel optical design that
includes a microelectromechanical systems (MEMS) mirror for rapid scanning the
microarray. The light of a diode laser is precisely oriented onto spots of an array
surface. Scanning at 60Hz, the MEMS mirror optical design permits the simulta-
neous interrogation and analysis of high-density microarrays. The light beam is
scalable to larger surface areas while maintaining uniform beam intensity. Array
spot intensity is recorded by a high-denition CCD camera. The device utilizes the
Kretschmann conguration and gold-covered, high refractive index glass slides.
Lumera combines protein arrays with high-throughput surface plasmon resonance
through its proprietary Heterodimer Protein Technology (HPT) peptide tag system.
Granity Pharmaceuticals GmbH (Heidelberg, Germany) has developed the
Plasmon Imager for discovery of small molecular hit and lead compounds. The
platform is used to nd novel small molecules for drug discovery and chemical
genomics approaches. Granity has developed high density chemical micro-
arrays for fragment screening consisting of small molecules immobilized on to
gold chips in combination with high density spotting [79,80]. High-throughput,
label-free fragment screening and screening of displayed-fragment microarrays
on the proprietary SPR platform, allows Granity to rapidly identify novel
drug fragments and lead-like molecules as ligands for a biomolecular target.
Synthesized fragment libraries are stored in microtiter plates until spotting
on to the sensor elds by using a customized high-throughput and high-
precision pintool spotting robot. In Granitys setup, a wavelength shift that
corresponds to the increase in mass concentration on the chip surface during
binding between the target proteins and the immobilized chemical substances is
recorded. Array screening is performed by owing protein solutions over the
chemical microarray, followed by the binding of the target by end-point
measurements within 3 hours. Granitys Plasmon Imager allows the parallel
readout for up to 9612 sensor elds per array. The measured SPR shift can be
visualized in false-colored 2D plots to obtain anity ngerprints, such as those
given in Figure 3.32. The Plasmon Imager instrument has not been marketed,
but is used in projects for partners of Granity.
In the 1990s, the IBIS I and II systems of IBIS Technologies
16
(Hengelo, The
Netherlands) were introduced to the market. In 2007, IBIS Technologies
launched a new advanced imaging SPR instrument with a patent pending for
angle scanning imaging technology. The system can be categorized as a new
combination of angle scanning SPR (see Section 3.4.3) and an SPR imaging
instrument (see Section 3.4.6). The IBIS iSPR instrument (Figure 3.33) allows
simultaneous monitoring of multiple binding interactions on microarrays spotted
on the sensor surface. The instrument is equipped with a standard planar ow
cell but can be used with a cuvette or with a conned wall-jet ow cell as well.
16
Later redesigned by Eco Chemie (Utrecht, The Netherlands) as ESPRIT and SPRINGLE
instruments, respectively (see Section 3.4.3).
A liquid handler operates the instrument automatically using various vials or
microtiter plates (96 or 384 wells), allowing extended stand-alone operation. The
Peltier element-controlled sensor compartment is thermostated to better than
0.011C and the sample compartment can be cooled separately.
The IBIS iSPR system combines the detection of the SPR dip with imaging of
the entire sensor surface. In so-called dynamic scan operation, the reectivity
change which is normally applied in xed-angle imaging instruments is converted
into a real SPR dip angle shift and can be used for real-time detection and also
calculation of kinetic parameters of binding events at user-dened spots on a
microarray, even when spots have SPR dip dierences above 3000 millidegrees.
Images of the entire microarray can be observed at once from the microscopic
view on the monitor. Contrasts of the image for specic regions of interest
(ROIs) can be set by the SPR angle or digitally improved by software settings. In
the software the SPR dip shifts of reference and control spots can be subtracted
from each other. In addition, the user can dene hundreds of regions of interest,
adding enormous exibility to the system. In a technical sense, the IBIS iSPR
system combines the best of both worlds: real-time imaging of the entire sensor
surface is combined with an SPR dip scan (instead of reectivity) of hundreds of
biomolecular interactions (4500). Further, a whole range of (patent-protected)
surface chemistries can be delivered by IBIS Technologies. In Chapter 7, results
are presented of direct serum detection of autoantibodies from 50 rheumatoid
arthritis patients.
Figure 3.32 The in-house built Plasmon Imager of Granity Pharmaceuticals applies
the wavelength interrogation principle (see Section 3.2.5) and a high-
density microarray with thousands of spots will be imaged.
68 Chapter 3
3.5 Protein Interaction Analysis Systems of Biacore
3.5.1 Introduction
Biacore is the market leader in the development and production of SPR-based
analytical instruments since their introduction to the market in 1990. The data
generated from real-time protein interaction studies using these systems have
provided scientists with an unprecedented wealth of information on protein
function, within areas as diverse as learning how specic protein domains
contribute to biological function and the ability to make informed judgments
on the potential of specic proteins as targets in drug development. The Biacore
product line includes the Bialite, Biacore J, -X, -1000, -2000, -3000, -C, -S51,
Q, -A100, -T100, -X100 and FLEXChip. Three Biacore systems, the Biacore
A100, T100 and FLEXChip, each designed to address a specic need within life
science research and the pharmaceutical industry, are reviewed here.
SPR-based arrays are now available, with the emphasis on information-rich
data rather than throughput, delivering information on association, dissociation
and strength of interaction while analyzing interactions in parallel. Two inter-
action array systems are available: FLEXChip, for simultaneous proling of
up to 400 protein interactions in a single run, and Biacore A100, which delivers
Figure 3.33 The SPR imaging instrument IBIS-iSPR of IBIS Technologies. The left
computer screen shows the sensorgram and quality of the SPR dip, and
the screen on the right shows the microscopic image of the surface. A
region of interest (ROI) can be selected in size and position on the surface
(red squares) and the user can inspect the quality of the surface before a
large series of samples are automatically passed over the surface using the
autosampler. The autosampler operation can be inspected through the
instrument window.
high-quality data on up to 3800 interactions in 24 hours. These systems, together
with the Biacore T100, are reviewed in Table 3.3 (see also Figure 3.34A).
3.5.2 Biacore T100
The Biacore T100 is a highly automated system for comprehensive protein
interaction analysis from early drug discovery through drug development to
quality control. In addition to providing detailed information on kinetics and
anity, software support allows interactions to be thermodynamically charac-
terized. Both the sample compartment and ow cell system of the instrument
are temperature-controlled, a feature that extends the potential use of the
instrument to include the analysis of interactions at physiological temperatures
and above.
Although the key features of Biacores protein interaction analysis systems
have remained consistent over the years label-free, detailed characterization
of protein interactions in real time they have continued to develop not only in
terms of kinetic resolution, ease of use, automation and regulatory compliance,
but also in the depth of information revealed about individual interactions.
Table 3.3 Overviewof the Biacore A100, T100 and FLEXChip instrumentation.
Instrument Application area Main features
Biacore
T100
Research to quality
control
Automatic analysis of up to 384 samples
per run over the broadest kinetic range
Analysis and evaluation of interactions
involving low-MW compounds
(o1000 Da)
Integrated buer degassing improves
robustness when studying interactions
at elevated temperatures
Integrated sample cooling for
temperature-sensitive samples
Automated recovery of interaction
partners for MS analysis
Biacore
A100
Protein interaction
analysis
Analysis of up to 3800 interactions in 24
hours
Unique ow system for parallel analysis
increased sample throughput and
multiplexed analyses
Robust data and long unattended run
times
Software tools designed for array
applications
FLEXChip Comparative proling:
map, rank and select
Simultaneous proling of up to 400
protein interactions
Ecient comparison and ranking
hundreds of interactions
Sample recirculation for measuring slow
association events
70 Chapter 3
For example, in addition to characterizing kinetic proles, it is now also quick
and simple to characterize the underlying thermodynamic principles that drive
interactions, and the possibility of integrating Biacore T100 into an analysis
and identication workow together with mass spectrometry and hence protein
identication has rmly placed the system in the eld of functional proteomics.
Figure 3.34 The Biacore instruments as described in this section. (A) Biacore T100;
(B) Biacore A100. The FLEXChip instrument is shown in Figure 3.24.
The design of the ow cell system in the Biacore T100 creates optimal
conditions for accurate reference subtraction. Four ow cells allow single,
paired or serial runs, and paired, on-chip ow cell connections mean that the
void volume between ow cells is as small as possible. Kinetic rate constants
may be measured over a broad range, from the fastest on-rates encountered in
biological systems to the slowest o-rates on-rates from 10
3
to 10
7
l mol
1
s
1
(and higher for macromolecular analytes) and o-rates from 10
5
to 0.5 s
1
may be condently measured. Samples may be maintained from 4 to 45 1C in a
temperature-controlled compartment, allowing unattended analysis of temper-
ature-sensitive samples. Assuming an analysis cycle of 7 minutes, up to 384
samples may be processed during 48 hours of unattended operation in a single
run. Finally, dedicated software supports kinetic evaluation of low molecular
weight compound interactions involving binding partners with molecular
weight as low as 100 Da.
In addition to a cooling compartment, the entire ow cell system is temper-
ature controlled and is monitored with an integrated buer degasser, which
eliminates the appearance of air bubbles in the ow system, making possible the
analysis of samples at elevated temperatures. It is therefore not only possible to
analyze temperature-sensitive samples, but also the interactions themselves may
be characterized at temperatures ranging from 4 to 45 1C. Consequently,
interactions may be studied at physiological temperatures, permitting the
behavior of therapeutics in vivo to be predicted more condently. This is an
important advantage of the Biacore T100 over typical benchtop assays in
applications such as the characterization of monoclonal antibodies as biother-
apeutics, where interaction proles with target proteins may dier radically at
ambient and physiological temperatures.
A further consequence of the ability to accurately control temperature within
the entire ow cell system is the possibility to derive transition state thermo-
dynamic data from kinetic proles. Kinetic proling imparts information about
the rate of complex association and dissociation (in addition to revealing the
anity of an interaction), but to understand why the interaction proceeds at
these rates, it is necessary to dene the thermodynamics of the system. Fully
understanding molecular recognition by being able to predict binding energe-
tics through thermodynamic analysis may provide the basis for structure-based
molecular design of drugs and engineered antibodies.
Protein interaction analysis on the Biacore T100 in combination with mass
spectrometry (MS) provides the possibility of identifying proteins on the basis
of functional binding criteria. In a typical experiment, molecules are isolated
from a complex matrix based on their ability to bind specically an interaction
partner immobilized on a sensor surface. The bound material then can be
recovered in a non-destructive manner by using a recovery solution that
completely dissociates the bound analyte without damaging the immobilized
interaction partner on the sensor surface and thus allowing MS analysis of the
sample. Several examples of how SPR systems and matrix-assisted laser des-
orption/ionization time-of-ight MS (MALDI-TOF MS) have been combined
to capture (Biacore) and identify (MS) proteins from complex matrices may be
72 Chapter 3
found in the literature [35,36]. Although it has been possible to perform
SPR-MS on earlier systems from Biacore, the process has been rened and
optimized for Biacore T100, with the entire recovery process dened in the
Method Builder software in a predened template.
Characteristics of buers such as ion content, salt concentration and pH are
all variables that can radically aect the prole of an interaction. Buer
scouting is a novel feature of the Biacore T100 and is intended to help the user
rapidly nd the optimal buer conditions to suit a specic interaction; up to
four dierent buers can be tested at one time. In addition, this exibility
allows the user to study micro-environmental eects on binding properties in
mechanistic and stability studies and to dene the kinetic properties of protein
samples in varied biochemical or biophysical environments. This information
may be crucial when selecting candidates intended for use in the complex and
variable environment of clinical treatment.
3.5.3 Biacore A100
The Biacore A100 is a protein interaction analysis system delivering high-quality
kinetic, anity, concentration and specicity data. It is designed primarily for
applications requiring high sample throughput using small panels of proteins,
but can also be used for detailed characterization studies, allowing integration
into multiple phases of project pipelines. The open, exible format permits
multiplexed assays that create new possibilities for faster data acquisition. The
system oers enhanced productivity in key areas such as antibody selection,
biotherapeutic and low molecular weight drug development, immunogenicity
studies and proteomics. An optional package is available for work in regulated
environments.
Hydrodynamic addressing (HA; see Figure 3.35 and Section 3.3.1) is a
process by which multiple interactants may be immobilized on detection spots
in a single ow cell, allowing simultaneous analysis of interactions. As there is
no lag time between interactions, highly accurate reference subtraction allows
the measurement of very rapid kinetics. Further, by immobilizing several
interactants in one ow cell, comparative binding properties may be directly
examined under optimal experimental conditions. By adjusting the relative ow
at the two inlets (one for the ligand which should be immobilized and the other
for buer), liquid can be directed to dierent addressable detection spots. The
ow cell design allows rapid and ecient switching of ow between buer and
sample solutions and the transverse arrangement of the detection spots ensures
that access of sample to all spots is simultaneous. Although the detection spots
are addressed separately during immobilization, the injected sample ows over
all spots simultaneously (see Figure 3.35).
The ow cell conguration of the Biacore A100 enables up to 3800 interac-
tions to be monitored in a 24 hour run, with a selectable conguration either for
maximum number of samples or for maximum information per sample (see
Figure 3.36). The capacity for parallel processing and the high quality of kinetic
data make the Biacore A100 an attractive option for applications such as
interaction proteomics, drug discovery and biotherapeutics development,
where it is vital to be able to handle many samples and have condence in
the data. In biotherapeutics development, for example, the development of
monoclonal antibodies is a complex, time-consuming and thus expensive
business involving the generation, maintenance and screening of thousands
of hybridoma clones. Condent early identication of hybridomas to produce
the best candidate antibodies is a critical step in successful, cost-ecient
development.
In drug discovery programs, the Biacore A100 can provide information-rich
data, allowing the identication of high-quality lead compounds that are
crucial for progress. Direct binding analysis, oering comprehensive charac-
terization of critical selectivity and kinetic properties, provides data to guide
Figure 3.35 Schematic view of the ow system in the Biacore A100. Planar view; the
four injection ports (I1I4) allow unique interaction conditions and
parallel analysis at ve measurement spots in each ow cell. Top: side
view of ow cell with sensor surface and measurement spots on the
bottom and ow cell cartridge (gray) on top.
74 Chapter 3
key decisions. The identication of highly selective compounds against complex
therapeutic targets may benet strongly from a comprehensive panel approach
in order to eliminate potential target-dependent artifacts that may result in
false-positive or false-negative leads. Furthermore, the use of multiple control
targets to identify non-specic protein binding, binding to recombinant protein
tags and possible expression-system artifacts, should be generally applicable to
almost any drug discovery program.
3.5.4 FLEXChip
The development of SPR-based biosensors has been geared to delivering data of
the highest possible quality on a limited number of interactions. There are several
reasons, however, to support the design and production of systems with greatly
increased capacities for sample throughput. Perhaps the most pressing call for a
commercially available protein interaction array is from the proteomics com-
munity; with a bewildering amount of novel proteins at hand since the comple-
tion of the human genome project, any technology that helps explain their
functions is welcome. Additionally, many applications such as antibody screen-
ing, hit selection in drug development programs, peptide epitope mapping
Figure 3.36 Conguration of detection spots in the Biacore A100 system optimized
for (top) sample throughput with parallel analysis of ve components in
four samples and (bottom) number of analyses per sample with a
maximum of 20 components per sample.
and even on-line quality control/safety testing during food production are all
activities that could benet from the increased sample throughput on an array.
The development of protein arrays is more complex than the DNA counter-
part. Proteins are more dicult to handle, as post-translational modications,
vital for functionality, are seldom preserved in the course of the amplication
steps required to obtain reagents in sucient quantity and purity. Consideration
must also be given to variables such as immobilization conditions (see also
Chapter 6), orientation and the possibility that the immobilization process may
impede or conceal the very binding site of interest. Further complications
include the desirability of immobilizing proteins eciently and at precise con-
centrations (important for obtaining meaningful association rates) across the
entire array. Proteins are also less discriminatory in their choice of binding
partners than DNA and so non-specic adsorption to both the sensor surface
and to other proteins in a complex mixture, such as clinical samples or
hybridoma supernatants, may complicate the interpretation of results from a
multiplexed array.
The FLEXChip (see Figure 3.24) is an SPR-based array for the simultaneous
proling of up to 400 protein interactions. The ow cell in FLEXChip is a
single, broad channel through which a single sample is injected, interacting
simultaneously with all the spots on the array. After a panel of interacting
partners have been externally spotted on the sensor surface according to
experimental requirements, a gasketed window with an inlet and an outlet
valve is positioned and hermetically sealed over the array to form the ow cell,
which is then inserted into the FLEXChip apparatus. FLEXChip uses a variant
of SPR known as grating-coupled SPR, in which incident light from above
strikes the entire array, instead of the light from the prism as in the true
Kretschmann operated SPR imaging systems.
It is important to remember that the information sought from functional
protein arrays goes beyond mere detection and that, consequently, technologies
delivering this information should not be judged solely in terms of throughput.
A technology that proles entire interactions and delivers data on the associ-
ation rate constant (k
a
), dissociation rate constant (k
d
) and anity constant
(K
D
) delivers information about the function of a restricted and usually selected
population of proteins in a cell or a group of antibodies or peptides. Protein
interaction arrays are therefore, of necessity, smaller than DNA arrays.
Whereas the proteome itself is nite, the range of protein functions is
practically boundless. Cancer specialists, for example, may be interested in
identifying binding partners of proteins uniquely expressed in cancer but will be
able to answer a more explicit question if they are able to make informed
judgments on the consequences of these interaction patterns and ask how they
aect disease progression and options for therapy. Functional protein arrays
may well prove to be the link between the vast repository of data that has
emerged from proteomics initiatives and the world of clinical and biological
research.
An excellent example illustrating the power of the FLEXChip has been
demonstrated, in which its scope for parallel interaction analysis was fully
76 Chapter 3
exploited to integrate secondary screening with a fully automated phage display
process [12]. Phage display can generate a tremendous number of potential
protein therapeutics, such as antibody Fab fragments. Advances in the auto-
mation of the selection (panning a phage display library against an immobilized
target molecule) and primary screening (e.g. ELISA) processes are largely
responsible for the discovery of a great number of unique antibodies. However,
it then becomes a challenge to identify the most promising leads from a large
number of candidates.
Typically, candidates from a phage display process are selected after several
rounds of selection and amplication followed by a primary screen, but the
adoption of such a strategy, which aims to reduce demands on the secondary
screen, decreases the hit diversity and increases the risk of missing potential
candidates. An alternative strategy is to limit the number of rounds of selection
so that a more diverse collection of binders makes it through to the secondary
screen. Although having such a diverse collection of binders in terms of kinetic
proles increases the likelihood of the discovery of candidates with the desired
functional activities, it places unique demands on the performance of any
secondary screening process.
Wassaf et al. [12] incorporated FLEXChip as a secondary screen into a highly
automated selection and screening process in a scheme to nd potent human
antibody inhibitors to a human serine protease (tissue kallikrein 1, hK1), involved
in inammation. The process was based on a limited number of selection and
amplication rounds followed by interaction analysis on hundreds of candidates.
This case study showed how an automated procedure, including the secondary
screen, was developed to select, purify and concentrate Fabs prior to rapid
characterization using FLEXChip. The ability to monitor simultaneously hun-
dreds of interactions in a secondary screen may signal a signicant improvement
in the speed at which potential therapeutic candidates can be identied.
As candidates are fed through the development pipeline, the need for protein
interaction analysis continues, for example, for further selection during opti-
mization, for monitoring immunogenic responses and even for batch release
testing. The need to fulll regulatory requirements such as GxP compliance also
increases further downstream. These needs can be met with systems such as the
Biacore T100 and Biacore A100. FLEXChip is the lter between the vast
numbers of candidates generated in a phage display process and ensuring that
only the best candidates are advanced along the drug discovery pipeline.
3.6 Conclusion
Instruments contain at least three integrated components: (1) SPR instrumental
optics, (2) a liquid handling system and (3) the sensor chip. The quality of each of
these components reects the overall performance of the SPR instrument. In this
chapter, a short description has been given of SPR products from 25 companies.
Gold is still the gold standard for generating the SPR phenomenon in almost all
commercial instruments available on the market. The commercial availability of
sensor surfaces essentially contribute to accurate and reliable results (see Chapter 6).
Although in the past only Biacore has dominated the market (490%), new players
can be identied. However, as explained in this chapter, not all instruments from
these manufacturers will generate reliable quantitative kinetic data for kinetic
evaluation of rate and anity constants, but should be regarded as instruments
which are able to show qualitatively binding of the analyte with the immobilized
ligand. The degree of automation which also contributes to accurate and reliable
kinetic data retrieval diers from totally manual to highly automated and can
greatly enhance the performance of the SPR instrument.
Typically, SPR dip shifts are monitored in the reectivity mode or angle shift
mode. Instruments are categorized in six sections according to their optical
conguration. The liquid handling systems mainly comprise ow cells or
cuvettes. Special attention has been paid to three Biacore instruments, T100,
A100 and FLEXChip, and their application to protein studies, in Section 3.5.
As indicated in this timely chapter, the market is now more open than ever
before and competition between companies is taking place on several aspects of
the SPR systems. The customers of instruments will prot further from this
competition, oering more exibility, innovation and cost eectiveness.
3.7 Questions
1. An instrument for detection of biomolecular interactions can be con-
sidered as a total analysis system that consists of three main technology
parts. Which three parts are essential for such a biosensor system?
2. A divergent beam SPR instrument shows walking of the SPR dip over
the sensor surface. Explain this eect.
3. In convergent beam instruments, the exact SPR angle is detected without
moving parts. A camera is needed to measure, e.g., 20 sensor spots in line.
How can one make an image of the sensor surface?
4. In SPR imaging instruments, a parallel beam of light will bring a homo-
geneous surface in full resonance. How can we make an image of the
surface and follow the SPR angle of each region of interest during the
biomolecular interaction process in order to calculate the SPR angle
position of each spot? How many sensor patches can then be obtained
from a sensor surface of, e.g., 5 mm square?
5. Consider an angle scanning instrument. Draw the reectivity curve as a
function of time if the scanner rst moves in forward direction and passes
the SPR dip and then shifts backwards with the same speed as in forward
direction. What happens to this curve after the SPR dip has shifted?
References
1. B. Liedberg, C. Nylander and I. Lundstrom, Biosens. Bioelectron., 1995,
10, iix.
2. R. Karlsson, J. Mol. Recognit., 2004, 17, 151161.
3. E. Stenberg, B. Persson, H. Roos and C. Urbaniczky, J. Colloid Interface
Sci., 1991, 143, 513526.
78 Chapter 3
4. M. Malmqvist, Nature, 1993, 361, 186187.
5. B. Johnsson, S. Lofas and G. Lindqvist, Anal. Biochem., 1991, 198, 268277.
6. T. Wink, S.J. van Zuilen, A. Bult and W.P. van Bennekom, Anal. Chem.,
1998, 70, 827832.
1999, 71, 39283934.
8. B.P. Nelson, T.E. Grimsrud, M.R. Liles, R.M. Goodman and R.M. Corn,
Anal. Chem., 2001, 73, 17.
9. L.-M. Zhang and D. Uttamchandani, Electron. Lett., 1988, 24, 14691470.
10. J. Homola, Surface Plasmon Resonance Based Sensors. Springer Series on
Chemical Sensors and Biosensors, Vol. 4, Series Editor O.S. Wolfbeis,
Springer, Berlin, 2006.
11. E. Kretschmann, Z. Phys., 1971, 241, 313324.
12. D. Wassaf, G. Kuang, K. Kopacz, Q.L. Wu, Q. Nguyen, M. Toews, J. Cosic,
J. Jacques, S. Wiltshire, J. Lambert, C.C. Pazmany, S. Hogan, R.C. Ladner,
A.E. Nixon and D.J. Sexton, Anal. Biochem., 2006, 351, 241253.
13. M. Pe rez-Moralesa, J.M. Pedrosab, E. Mun oza, M.T. Mart n-Romeroa,
D. Mo biusc and L. Camachoa, Thin Solid Films, 2005, 488(1-2), 247253.
14. J. Ctyrokya, J. Homola and M. Skalskya, Opt. Quantum Electron., 1997,
29, 301311.
15. C.R. Yonzon, E. Jeoung, S. Zou, G.M. Mrksich and R.P. Van Duyne,
J. Am. Chem. Soc., 2004, 126, 1266912676.
16. J. Homola, S.S. Yee and G. Gauglitz, Sens. Actuators B, 1999, 54, 315.
17. A.M.C. Lokate, J.B. Beusink, G.A.J. Besselink, G.J.M. Pruijn and R.B.M.
Schasfoort, J. Am. Chem. Soc., 2007; DOI: 101021/ja075103x.
18. P.A. Lowe, T.J.H.A. Clark, R.J. Davies, P.R. Edwards, T. Kinning and
D. Yeung, J. Mol. Recogn., 1998, 11, 194199.
20. C.E.H. Berger, T.A.M. Beumer, R.P.H. Kooyman and J. Greve, Anal.
Chem., 1998, 70, 703706.
21. S. Dickopf, M. Frank, H.D. Junker, S. Maier, G. Metz, H. Ottleben, H. Rau,
N. Schellhaas, K. Schmidt, R. Sekul, C. Vanier, D. Vetter, J. Czech,
M. Lorenz, H. Matter, M. Schudok, H. Schreuder, D.W. Will and
H.P. Nestler, Anal. Biochem., 2004, 335, 5057.
22. J.M. Brockman, B.P. Nelson and R.M. Corn, Annu. Rev. Phys. Chem.,
2000, 51, 4163.
23. P. Schuck, Annu. Rev. Biophys. Biomol. Struct., 1997, 26, 541566.
24. T.A. Morton, D.G. Myszka and I.M. Chaiken, Anal. Biochem, 1995, 227,
176185.
25. H.A. Stone, A.D. Stroock and A. Ajdari, Annu. Rev. Fluid Mech., 2004, 36,
381411.
26. D.E. Davey, D.E. Mulcahy and G.R. OConnell, Electroanalysis, 2005, 5,
581588.
27. C.L. Baird and D.G. Myszka, J. Mol. Recognit., 2001, 14, 261268.
28. Cush, et al., Biosens. Bioelectron., 1993, 8, 347364.
29. H.J. Watts, D. Yeung and H. Parkes, Anal. Chem., 1995, 67, 42834289.
30. L.A. Chtcheglova, M. Vogel, H.J. Gruber, G. Dietler and A. Haeberli,
Biopolymers, 2006, 83, 6982.
31. M.B. Glaubert, J. Fluid. Mech., 1956, 1, 625643.
32. Internet references, March 2007: www.biacore.com, www.biosensingusa.com,
www.nomadics.com, www.dkktoa.net/, www.reichertai.com/, www.plasmonic.
de, www.moritex.com, www.biosuplar.de, www.sensia.es, www.k-mac.co.kr/,
www.nanolm.de/, www.bio-rad.com, www.ecochemie. nl, www.thermo.com,
www.neosensors.com, www.gwctechnologies.com, www.genoptics-spr.com,
www.mavenbiotech.com, www.lumera.com/, www.granity.com, www.ibis-
spr.nl.
33. A. Marquart, Surface Plasmon Resonance, SPR pages http://www.
sprpages.nl.
34. H. Jin Lee, T.T. Goodrich and R.M. Corn, Anal. Chem., 2001, 73, 55255531.
35. E.A. Smith and R.M. Corn, Appl. Spectroscopy, 2003, 57, 320A332A.
36. D. Nedelkov and R.W. Nelson, Am. J. Kidney Dis., 2001, 38, 481487.
37. A. Zhukov, M. Schurenberg, O
. Jansson, D. Areskoug and J. Buijs,

J. Biomol. Tech., 2004, 15, 112119.
38. R.L. Rich and D.G. Myszka, J. Mol. Recognit., 2006, 19(6), 478534.
39. R.L. Rich and D.G. Myszka, Anal. Biochem., 2007, 361, 16.
80 Chapter 3
CHAPTER 4
Kinetic Models Describing
Biomolecular Interactions at
Surfaces
DAMIEN HALL
1,a,b,c
a
University Chemical Laboratory, University of Cambridge, Lenseld Road,
Cambridge CB2 1EW, UK;
b
Institute for Protein Research, Osaka
University, 3-2 Yamadaoka, Suita, Osaka 565-0871, JAPAN;
c
Graduate
School of Bioscience and Biotechnology, Tokyo Institute of Technology,
B-39, 5249 Nagatsuta, Midori-ku, Yokohama 226-8501, JAPAN
4.1 Introduction
To say that the adsorption of molecules from solution to a surface is an
important phenomenon in biochemistry is to make a major understatement.
Indeed, so much of the fundamental chemistry of life occurs at interfacial
regions [112] (Table 4.1) that the biochemists or biophysicists who take their
subject seriously are required to have both an adequate understanding of how
these processes occur and how they might be measured experimentally. The
major thrust of this volume is concerned with interrogating the adsorption of
biologically important molecules to surfaces by the use of optical biosensor
technology. In this chapter, we discuss adsorption events from the perspective
of monitoring a measurement signal which provides information, in real time,
on the extent of solute adsorption to a surface. As discussion of time-dependent
adsorption phenomena necessitates the use of kinetic models, we spend our
time reviewing kinetic models that describe a wide range of adsorption behavior.
We begin, however, by introducing the terminology of adsorption so that we
might have the necessary language with which to conduct a discussion of the
subject.
1
Present address: Hall Laboratory. Institute for Basic Medical Sciences. Tsukuba University. 1-1-1
Tennoudai, Tsukuba, Ibaraki, 305-8575, JAPAN.
81
4.1.1 Terminology of Adsorption
In any solution-based adsorption experiment, one wishes to monitor the
adsorption of a certain molecule in solution, termed the solute (or alternatively
analyte), to the surface of a solid phase. The solid-phase adsorptive surface is
composed of a number of sites that exert an attractive force for the solute. Once
adsorbed to the surface the solute is termed the adsorbate. However, the term
solute and adsorbate refer to any molecule that will bind to the surface
specically or non-specically. Adsorption experiments are generally preceded
by the preparation of a suitable adsorptive surface that is composed of a
number of binding sites (alternatively known as matrix sites or immobilized
ligand binding sites). The type of adsorptive surface that is prepared is neces-
sarily dependent upon the system that is being studied however we might make
a distinction as to the general type of surface in relation to a specic solute as
constituting either a distinct array of binding sites or an eective continuum of
binding sites (Figure 4.1). As a general rule, we may consider the adsorbing
Table 4.1 Role of Adsorption in Fundamental Life Processes.
Biological Function Role of Adsorption
Intra/Extra-Cellular
Tracking
First step in transfer of individual molecules and molecular
complexes between various compartmentalized structures of
the cell [1]. Also involved in transport to and from the cell e.g.
pinocytosis and exocytosis [2].
Cell Signalling Cell to cell communication is important both for single cell [3]
and multi-cellular organisms [4]. Communication during such
processes as tissue development is mediated by adsorption of
specic signalling molecules to receptors located at the cell
membrane surface [5].
Olfactory/Taste
Senses
Initial step in sensory pathways of taste and smell involve
adsorption of specic ligands to specic taste or smell
receptors located on the cell surface [6,7].
Neuronal Impulse
Transmission
Transmission of nerve impulses across synapses involves release
of neurotransmitters from one side of the synaptic cleft and
their subsequent adsorption to receptors on the opposing face
of the synaptic cleft [8].
DNA Transcription Transcription of DNA and translation of RNA involve
adsorption of specic protein complexes to the one
dimensional adsorptive surface of the nucleic acid polymer
chain [9].
RNA Translation
Immune Response The rst step in the defence against invading foreign agents
involves adsorption of antibodies to the invading agent. The
foreign agent constitutes an adsorptive surface composed of
matrix sites termed epitopes that are specically recognised
by various immunoglobulin molecules [10].
Articial Implant
Technology
Material introduced into the body e.g. articial joints, contact
lenses, slow release drug reservoir etc. should have minimal
potential to act as an adsorptive surface. Reduction of non-
specic adsorption to various implant devices is an active area
of research [11,12].
82 Chapter 4
(A)
(B)
z
Type 1
Radius = r
1
Radius = r
2
Type 2
y
x
Figure 4.1 (A) The radius r of idealized spherical solute molecules (shown in green)
may be small (r
1
) or large (r
2
) in relation to the distance between the
adsorption sites on the adsorptive surface. (B) Diagram indicating a two-
dimensional surface (surface plane). Colors indicate change in potential
energy associated with the normal position of the center of the solute
molecule from a distance r to r+Dz away from the surface plane. Regions
of lowest potential (blue) indicate position of adsorption sites. For
adsorption of type 1 solute this two-dimensional adsorptive surface would
constitute a distinct array. For adsorption of type 2 solute the surface
would constitute a near continuum array. (C) Diagram indicating a three-
dimensional adsorptive phase positioned at the interfacial boundary of the
solid and liquid phases. Regions of low potential energy for solute
(adsorption sites) are indicated by blue spheres. The supporting phase
(often a polymer gel) is shown by red rods. As with the surface plane
shown in (B), this surface phase may exist as either a distinct array or a
near continuum array depending upon the dimensions of the adsorbing
solute.
83 Kinetic Models Describing Biomolecular Interactions at Surfaces
surface as a distinct array of sites if an adsorbate molecule, once bound, does
not interfere with the subsequent adsorption of solute to any neighboring sites.
Once the adsorptive surface has been constructed, an adsorption experiment
may be conducted by exposing this surface to a solution containing solute.
Once introduced to the surface the solute will begin to adsorb
2
a process
which will continue until an equilibrium position is reached between the
concentration of solute and the concentration of adsorbate. At this point, the
process of adsorbate desorption may be studied by replacing the liquid phase
with a solution not containing solute and recording the dissociation of adsorb-
ate until an equilibrium ratio between solute and adsorbate is achieved. These
basic adsorption and desorption experiments can be performed at dierent
concentrations of added solute/initial adsorbate to examine the eect of
concentration on the time-dependent (Figure 4.2a) or time-independent
(Figure 4.2b) adsorption/desorption processes. Evaluation of the concentration
dependence of the kinetic and equilibrium situations allows for the evaluation
of the rate and equilibrium constants that dene the behavior of the interaction
under the particular conditions chosen for the study. These experiments may be
repeated under dierent conditions in which a single environmental variable
(e.g. temperature, pressure or composition of the buer) is systematically
altered to examine how the rate and equilibrium constants that dene the
adsorption event respond to the imposed change.
Until relatively recently with the experimental interrogation of solute
adsorption from solution was made indirectly, using measurement strategies
(C)
z
y
x
Figure 4.1 Continued
2
The particular forces that drive adsorption are outside the scope of this discussion.
84 Chapter 4
Figure 4.2 (A) Generalized prole of a time-dependent adsorption experiment:
normalized concentration of adsorbate vs. time t. The rst arrow shows
the point of introduction of the solute to the adsorptive surface. In the
association phase solute will continue to adsorb to the surface until an
equilibrium is attained. The second arrow shows the point at which the
liquid phase is replaced by buer not containing solute. In this dissoci-
ation phase, adsorbate is allowed to dissociate from the surface until a
new equilibrium position is reached. The ve lines represent experiments
carried out at ve dierent concentrations. (B) General prole of a time
independent adsorption experiment: the time independent values attained
at the nish of each of the kinetic adsorption experiments are plotted
against the equilibrium concentration of solute in the liquid phase.
based on dierence measurements of the amount of solute depleted from the
liquid phase [13]. Such indirect measurement strategies placed a limit on both
the type of adsorption processes that could be studied and the temporal
resolution with which that process could be resolved. The recent development
of new techniques [14,15] allows for direct measurement of the amount of
adsorbed solute and has aorded a time resolution that is almost contempo-
raneous with the adsorption event. Of these new techniques, those based on
evanescent wave optical biosensor technology [1618] have become the most
popular, mainly due to the successful commercialization of the technology.
4.1.2 Optical Quantication of Adsorption at an Interface
Evanescent wave optical biosensor technology is reliant upon the physics of the
total internal reection (TIR) of light. Light shone at an interface
3
below a
certain critical angle will enter the second medium and be refracted by it to an
extent governed by the refractive indices of the two mediums [19]. At incident
angles greater than the critical angle, the light will not enter the second optical
medium but will instead be totally internally reected back into the rst
medium. The electric and magnetic elds associated with the reected ray of
light, however, do not end abruptly at the point of reection but will penetrate
somewhat into the second medium. Interaction of these decaying electric and
magnetic elds (termed evanescent elds) with the second medium beyond the
interfacial plane will subtly change the properties of the reected beam. The
degree of non-absorptive interaction of the electrical component of the eva-
nescent light with the matter at the interfacial region will be determined by the
electrical polarizability of the material at the interface. Thus, optical biosensor
technology can be taken to reect changes in refractive index at the interfacial
layer.
4
The eld strength associated with the evanescent light at the interface
decays exponentially with distance normal to the surface, z, making the
observed signal, S, proportional to refractive index changes very close to the
surface [eq. (4.1)].
S /
_
N
0
Dnz expz=sdz 4:1
The parameter s is the eld strength decay constant and has units of distance.
For planar surfaces it is typically of the order of 0.30.5 times the wavelength of
the light used in the TIR experiment [19]. For a wide range of substances
5
the
change in refractive index in response to varying the weight concentration of
3
An interface existing between two optically transparent mediums.
4
For optically transparent materials, the macroscopic quantity related to the electrical polariz-
ability is the refractive index.
5
The approximate dn/dc values for proteins and nucleic acids are 0.18 and 0.16 ml g
1
, respec-
tively. The values for dierent types of carbohydrate range from 0.10 to 0.18 ml g
1
.
86 Chapter 4
component i, c
i
, can be approached as a linear function as the derivative dn/dc
i
is approximately constant [eq. (4.2)].
nc
i
nc
i
0
dn
dc
i
_ _
c
i
4:2
Substitution of eq. (4.2) into eq. (4.1) allows us to relate the change in signal, DS,
for situations in which initially no component i is present (c
i
0), with changes in
the weight concentration of component i at the interfacial layer [eq. (4.3)].
DS /
dn
dc
i
_
N
0
dc
i
dz
expz=sdz 4:3
On the basis of eq. (4.3), we can appreciate that the measured signal is,
strictly, not a linear descriptor of the concentration of adsorbate but for many
cases it will very nearly be so. Figure 4.3 describes the apparent change in signal
that would be registered for a constant mass of adsorbate in a number of
dierent modes of adsorption on the surface of an optical biosensor. On the
basis of eq. (4.3), the relative signal (normalized to the type 1 case) for each
adsorption mode would be type 1, 1; type 2, 1.01; type 3, 0.56; and type 4, 0.23.
A variety of ingenious means (of which SPR [20] is perhaps the best known)
have been developed to maximize the changes in the reected beam of light so
as to make it a viable means for experimental measurement. Although the
physics of each dierent experimental detection method dier somewhat, all
optical biosensor technology share basic characteristics as outlined in eqs.
(4.1)(4.3). More is said about the physics and instrumentation particular to
SPR in Chapters 2 and 3. For the remainder of the current chapter we will
discuss kinetic models that are relevant to dierent aspects of the process of
adsorption as studied by optical biosensor technology. Although we will not
specically transform the concentration units of the adsorption prole into an
equivalent signal as dened by eq. (4.3), we ask the reader to keep in mind the
likely consequences of the transformation for some of the more complicated
adsorption geometries.
4.2 Dening Factors of the Adsorption Event
For adsorption to occur two events must happen in sequence: rst the solute
must be transported to the surface, and second the solute must successfully
interact with the surface to form the adsorbate. We will discuss each of these
events in greater detail in subsequent sections; however, a basic understanding
of the distinction between the two regimes can be gained by casting the
adsorption event in terms of a simple transport/reaction process [21] [eq. (4.4)].
C
i

k
a
k
b
fC
i
g
k
0
1
k
2
C
i
ads
4:4
(A)
(B)
1
2
3
4
1
0.5
0
0 50 100
Z (nm)
150 200 250 300
e
x
p
(
-
z
/
)
Figure 4.3 (A) Four dierent modes of solute adsorption with the adsorptive surface
at which the light is totally internally reected shown on the vertical left
axis. Type 1 adsorption: the solute adsorbs directly on the surface plane
and maintains its shape. Type 2: the solute adsorbs and undergoes a post-
adsorption isomerization to maximize its interaction with the surface.
Type 3: the solute adsorbs to matrix sites on a supporting polymer
molecule in an evenly dispersed manner. Type 4: the solute adsorbs on a
supporting polymer molecule preferentially on the bulk solution side. On
the basis of eq. (4.3), the relative signal (normalized to the type 1 case) for
each adsorption mode would be type 1, 1; type 2, 1.01; type 3, 0.56; and
type 4, 0.23. (B) The decrease in evanescent eld strength with distance
normal to the surface for a decay constant s200 nm (approximate case
for light used in TIR experiment of wavelength 600 nm).
88 Chapter 4
In this formulation, C
i
, {C
i
} and (C
i
)
ads
represent the concentration of solute
of type i, the concentration of solute of type i spatially close to the surface
(termed intimate solute) and the concentration of adsorbed solute of type i (the
adsorbate), respectively. Equation (4.4) describes the process whereby solute
approaches and leaves the surface at a rate dened by phenomenological
transport coecients k
a
and k
b
. Once in physical proximity to the surface,
the intimate solute may react to form adsorbate according to an association
rate function k
0
1
and a reverse rate constant k
2
, where both k
0
1
and k
2
have
units of s
1
. In the simplest instance the association rate function may be
further decomposed [eq. (4.5)] into an intrinsic second-order rate constant k
1
,
the initial total concentration of matrix sites,
6
(C
x
)
tot
and some unitless function
describing the surface site occupation, f(f), where fn(C
i
)
ads
/(C
x
)
tot
(n is the
average number of sites covered by the adsorbed solute). More will be said
about the surface function, but for now we may treat f(f) as describing the
fraction of total matrix sites available for participation in further adsorption of
solute.
k
0
1
k
1
C
x
tot
f f 4:5
In the simple scheme set forth in eq. (4.4), the rate of formation of the intimate
solute is given by eq. (4.6).
dfC
i
g
dt
k
a
C
i
k
b
fC
i
g k
0
1
fC
i
g k
2
C
i
ads
4:6
The so-called transport-limited case represents the situation in which the rate
of physical encounter between solute and surface dictates the rate of formation
of the adsorbate species. This situation occurs for non-equilibrium cases when
[k
1
0
{C
i
} k
2
(C
i
)
ads
] c [k
a
C
i
k
b
{C
i
}], making the rate of formation of the
intimate solute eectively zero (d{C
i
}/dt -0). It results in the eective associ-
ation rate being given by eq. (4.7a). In the limit of zero adsorbate concentra-
tion, (C
i
)
ads
E0, the eective forward rate parameter f
i
, can be approximated
by a ratio of reaction and transfer rate constants with the further approxima-
tion for f
i
allowable if the transport coecient is much smaller than the reaction
rate (k
b
{k
0
1
) [eq. (4.7b)]. In the alternative limit of zero solute concentration,
C
i
E0, the eective dissociation rate parameter b
i
is given by eq. (4.7b), again
with the additional approximation valid if k
b
{ k
0
1
(here K
R
0
k
0
1
/k
2
). For a
given degree of surface occupation an eective adsorption partition constant
6
To speak of a total concentration of adsorption sites one must assign a reaction volume to the
surface and treat the adsorbed layer as a three-dimensional phase. An alternative treatment is to
consider the surface layer as a two-dimensional plane with adsorbate and binding site concentra-
tions given in terms of mol m
2
. This results in the usual set of units. This results in an unusual set
of units for the resulting equilibrium (m
3
mol
1
) and association rate constants (m
3
mol
1
s
1
). We
prefer the former treatment as it is conceptually similar to solution chemical kinetics. Despite the
dierences, the two treatments are formally equivalent.
(unit-less) can then be determined by the ratio of eective forward and reverse
rate parameters [(K
i
)
e
f
i
/b
i
] (Eq. 4.7b).
dC
i
ads
dt
k
0
1
k
a
C
i
k
2
C
i
ads
k
b
k
0
1
_ _
k
2
C
i
ads
4:7a
f
i
E
k
0
1
k
a
k
b
k
0
1
Ek
a
; b
i
E
k
b
k
2
k
b
k
0
1
E
k
b
K
0
R
;K
i
eff

k
0
1
k
a
k
2
k
b
4:7b
In the antithetical, reaction-limited kinetic regime [k
0
1
{C
i
} k
2
(C
i
)
ads
] {
[k
a
C
i
k
b
{C
i
}], the concentration of intimate solute is little perturbed by the
reaction component of eq. (4.6), allowing a quasi-equilibrium approximation
such that {C
i
} E(k
a
/k
b
)C
i
K
T
C
i
(where K
T
represents an equilibrium partition
constant for formation of the intimate solute) leaving the reaction rate to be
written as eq. (4.8a). In this case, the eective association rate parameter, f
i
, in
the limit of zero adsorbate [(C
i
)
ads
E0] and the eective dissociation rate
constant, b
i
, in the limit of zero solute (C
i
E0) and the eective adsorption
partition constant, (K
i
)
e
f
i
/b
i
, can be described by eq. (4.8b).
dC
i
ads
dt
k
0
1
K
T
C
i
k
2
C
i
ads
4:8a
f
i
Ek
0
1
K
T
; b
i
Ek
2
; K
i
eff

k
0
1
k
a
k
2
k
b
4:8b
Although we have presented the two extreme antithetical kinetics regimes the
system under study may exhibit mixed kinetic behavior, in which case no
simplifying relationship can be deduced and the numerical solution of two
coupled dierential equations [eqs. (4.6) and (4.9)] must be performed with the
apparent phenomenological rate constants varying between the two limits set
out in eqns. (4.7b) and (4.8b) depending on the relative concentrations of
solute, sites for adsorption and adsorbate.
dC
i
ads
dt
k
0
1
C
i
f g k
2
C
i
ads
k
a
4f
i
4k
0
1
K
T
and
k
b
K
0
R
4b
i
4k
2
4:9
4.2.1 Mass Transfer
In optical biosensors, the solute is introduced to the adsorptive surface in
either a ow-through or a cuvette-type system. In the ow-through biosensor
(Figure 4.4), the adsorptive surface constitutes the base of a microuidic
channel through which solution ows at some average velocity v*. As the
solute is taken up by adsorption it is replaced anew by solute owing in from
further down the channel. In the cuvette-type biosensor (Figure 4.5), the
90 Chapter 4
adsorptive surface constitutes the base of a microcuvette into which solute is
added. To eect ecient mass transfer, the solution is subjected to shear forces
generated by aspiration cycles from a free wall jet device (see Chapter 3) or by a
stirring rod suspended near the top of the solution.
The most general approach for describing the mass transfer process rst
involves the spatial discretization of the solution volume comprising the bio-
sensor device, followed by numerical solution of a continuity equation (for the
cases of non-turbulent ow [22] describing the diusion, convection and
Reflected Light
z
x
(A)
(B)
Figure 4.4 (A) Diagram depicting the general characteristics of a ow-through type
biosensor. Solute is introduced as a plug into a microuidic channel at
controlled ow rate. The adsorptive surface at which light is totally
internally reected constitutes one side of the microuidic channel. As
solute is adsorbed to the surface, it is replaced by the solute owing
through behind allowing the liquid phase concentration for certain exper-
imental regimes to be considered a constant. (B) Idealized representation
of the velocity eld associated with ow in an optical biosensor based on
the parabolic ow assumption [eq. (4.12)].
Reflected Light
(A)
(B)
x
z
Figure 4.5 (A) Diagram depicting the general characteristics of a cuvette-type bio-
sensor. Solute introduced by injection into an open cuvette is subjected to
shear forces by an oscillating stirrer. The adsorptive surface at which light
is totally internally reected is positioned at the base of the cuvette. As the
solute is adsorbed to the surface it is depleted from solution meaning that
except for cases of insignicant depletion the solution phase concentration
of solute must be calculated by dierence between the total and adsorbed
concentrations. (B) An idealized representation of the velocity eld asso-
ciated with stirring in a cuvette type optical biosensor based on the
stagnant point ow assumption [eq. (4.13)].
92 Chapter 4
adsorption of the solute component) [eq. (4.10)]. Here D
i
and v
i
denote the
solutes diusion coecient and linear velocity.
7,8,9
mass transport reaction
dC
i
dt
_ _
r D
i
rC
i
rC
i
v
i
k
0
1
C
i
k
2
C
i
ads
4:10a
dC
i
ads
dt
k
0
1
C
i
k
2
C
i
ads
4:10b
For ecient mass transport, we require that, for points in close spatial prox-
imity to the adsorptive surface, the absolute values of the mass transport term
in eq. (4.10a) be either greater than the forward reaction term, for association
experiments, or greater than, the reverse reaction term, for experiments exam-
ining dissociation of adsorbate.
10
From inspection of eq. (4.10a), we note that
we may inuence this ratio by changing the transport parameters D
i
and v
i
or
alternatively modifying the reaction parameters C
i
, (C
x
)
tot
and f(f) (the last two
parameters housed in the association rate function k
0
1
). For a particular system
of interest, one can usually vary C
i
and (C
x
)
tot
by careful experimental design
leaving the transport parameters D
i
and v
i
as the important variables to be
considered. With regard to the diusion of solute, we note that D
i
for a
spherical solute of radius r
i
existing in an aqueous solvent of viscosity Z at
temperature T can be written [23] as eq. (4.11):
D
i

kT
6pZr
i
4:11a
r
i

3 v
i
M
i
4pN
A
3
_
4:11b
where k is Boltzmanns constant, N
A
is Avogadros number, v
i
is the partial
specic volume of the solute and M
i
is the solute molecular weight. In the
context of achieving ecient mass transfer in biosensor adsorption experi-
ments, we are interested how D
i
depends on the environmental variables of
solution temperature and composition [2427] (Table 4.2). With respect to
temperature, we note that the viscosity of water (and dilute aqueous solutions)
between temperatures of 0 and 501C can be described very well by an empirical
exponential dependence (where Z
01C(LIQUID)
1.8 10
3
kg s
1
m
1
), allowing
an estimation
11
of the solute diusion coecient that shows a signicant
increase with temperature over this range.
7
Eq. (4.10) must be appropriately modied for spatial boundary conditions, i.e. to account for the
motion /reaction of solute at or near physical boundaries.
8
The operator r when applied to a function f, rf, implies the sum of spatial derivatives, e.g. @f/
@x +@f/@y +@f/@z.
9
The reaction component is presented as a simple mechanistically concerted process for simplicity.
For sequential based reaction processes this must be modied.
10
That is, (dC/dt)
reaction
/(dC/dt)
transport
-0.
11
Valid for solutes that will not signicantly change their average dimensions upon experiencing a
change in temperature, e.g. as might a protein upon temperature-induced protein unfolding.
With regard to the dependence of the solute diusion coecient on solution
composition, the most pertinent changes for consideration are those due to the
addition of viscogenic agents often required for biochemical stability, most
commonly glycerol or sucrose [25], or those brought about by the use of a
tethered polymer gel layer [26,27]. Table 4.2 describes the estimated diusion
coecient for a spherical solute as a function of added glycerol or sucrose as
viscogenic agents (designated c
v
) at 201C [25]. A common feature of optical
biosensor experimentation is the use of a chemically inert tethered polymer
support such as carboxymethyldextran to act as a point of covalent attachment
for biological ligands [17,18]. Over a small distance scale the tethered polymer
support has the potential to act as a porous chemical phase acting to retard
diusion of the solute in a size dependent manner. A simple estimate
12
of the
potential for reduction in diusion coecient of a spherical solutes diusion
coecient upon entering a porous gel has been provided by Ogston et al.
[26,27]. They treated the polymer phase as a randomly arranged collection of
very long cylindrical rods having a radius Dr and a fractional volume occupa-
tion j. Using this approach, one may estimate the dependence of the diusion
coecient of a spherical solute on the size of solute and the concentration of
rod like polymer
13
constituting the gel phase D
i
(r
i
, j) as shown in Table 4.2.
Although reasonable estimates of the diusion coecient
14
can be gained
from experiment or theory, estimation of the velocity prole for a particular
conguration requires detailed device dimensions as input parameters in sim-
ulations based on the NavierStokes equations [28,29]. However, some general
conclusions can be made about the velocity proles operating in ow-through
Table 4.2 Dependence of solute diusion coecient on solution parameters
capable of variation in optical biosensor experimentation.
Particular Dependence Relevant equation describing diusion coecient
Temperature (for Water) [24] D
i

kT
6p 0:0018 exp0:0236T 273:1 r
i
Weight Concentration of
viscogenic agent
i
[25] (at
201C)
D
i

kT
6p 0:00103 Ac
v
Bc
v
r
i
Concentration and dimensions
of the supporting gel phase
[26,27]
D
i
D
i
j0
exp
r
i
Dr
_ _
j
p
_ _
i
For glycerol, A2.18e-6, B9.0e-9, for sucrose, A1.11e-6, B1.96e-8.
12
With some experimental support.
13
Yarmush et al. [32] have additionally raised the possibility that the diusion coecient of solute
in the gel will be sensitive to changes in the concentration of adsorbate. This will have the eect of
decreasing the rate of mass transfer to lower regions of the gel in an occupancy-dependent manner.
14
Diusion is assumed isotropic.
94 Chapter 4
biosensor designs and in cuvette-type biosensor designs by assuming idealized
velocity proles that are related to simpler cases.
15
For the ow-type biosensor,
one approach has been to treat the problem as ow in an open-ended pipe
and hence describe the velocity prole [3034] as the simple parabolic type
(Figure 4.4b) where the linear velocity in the x direction, v
x
, is expressed as a
function of position of the height of the channel, z, which has a maximum
height of z
max
[eq. (4.12)].
v
x
z 4v
x
z
max
=2
z
z
max
z
2
z
2
max
_ _
4:12
As can be noted from Figure 4.4b, the linear x component of the velocity of
solute i approaches zero at the channel walls and is maximal in the central
region. Simple approximate models for predicting the velocity prole are
harder to come by for the cuvette-type biosensor because stirring is carried
out dierently in dierent machine types. However, some information can be
gained from a limiting case analysis in terms of the stagnation point ow
assumption in a stirred cuvette [35,36]. In this treatment, a reduced subsection
of the ow prole is regarded as emanating from the position of the stirrer bar
and is expected to disperse symmetrically as it approaches the surface (leaving a
so-called stagnant point on the adsorptive surface in the area immediately
below the point of stirring and a stagnant ow region next to the surface)
[Figure 4.5b and eq. (4.13)].
v
x
0:164B
3
2
Z
r
_ _
1
2
xz 4:13a
v
z
0:164B
3
2
Z
r
_ _
1
2
z
2
4:13b
In these equations, B is a multi-term ow constant, dependent on device
geometry for which values can be found in Ref. [35] and references therein.
With regard to coordinate position, the point of origin is the point on the
surface immediately below the stirring device. As with the parabolic ow
assumption in the ow-through device, the predicted velocity for the cuvette
device approaches zero at the surface.
The full numerical solution of the continuity equation ([eq. (4.10)] can be a
daunting task. However, nearly all problems involving mass transfer to a
surface may, with varying degrees of success, be decomposed into simpler
compartment models [34,37,38] (Figure 4.6) reminiscent of the descriptive
model outlined in eq. (4.4). The basis of the two-compartment simplication
15
Here we consider convective ow up to the plane of the adsorptive surface. For situations where
the adsorption sites are attached to a polymer gel layer we assume that the polymer layer will
disturb ow and consider the surface of the tethered polymer layer as the relevant surface for
dening the velocity prole. Some authors also have considered the penetration of ow into any
such gel layer [39].
lies in the observation (noted above for the stagnant point ow and the
parabolic ow cases) that the velocity prole of a uid owing over a
stationary surface approaches zero at the limit of contact with the surface
(i.e. z 0). This zero ow region or stagnant layer may be considered as
extending out a small distance, d, from the surface. A neighboring similar-
sized region beyond this distance d, termed the bulk, is considered to move
Figure 4.6 (A) Two-compartment model for describing mass transport from bulk
solution to a surface. The region of zero ow, v 0, is assumed to extend
out a stagnant layer distance d from the surface, beyond this the bulk
solution moves at the average velocity of v
BULK
. Transport from the bulk
to the surface phase across zero ow region occurs predominantly via
diusion. (B) Ratio of steady-state stagnant ow region solute concen-
tration to bulk solute concentration for dierent values of d and dierent
extents of initial adsorption rate and concentration of matrix adsorption
sites, represented as a combined parameter k
1
(C
x
)
tot
. The value of d has
been estimated on the basis of the 5% approximation discussed in the text.
The value of (C
x
)
tot
is calculated on the basis of the volume of the stagnant
ow region, i.e. (C
x
)
tot
decreases with increasing d. (C) Typical dierential
rate plots for a system displaying Langmuir adsorption kinetics, which is
highly inuenced by mass transfer eects [dotted line: simulated using eqs.
(4.9) and (4.14a)]. Dierential rate plot for adsorption prole with the
identical reaction kinetics but without mass transfer eects (solid line).
The mass transfer limited data only approaches that of the Langmuir case
at near maximal saturation of the adsorption matrix sites.
96 Chapter 4
with a common averaged velocity, v
BULK
. For relatively fast ow rates,
transfer of solute from one region to another in the bulk will occur primarily
by convection, allowing for the assumption of a single homogeneous com-
partment having a solute concentration designated C
BULK
that is determined
solely by the ow rate. Due to the assumed zero ow in the stagnant layer
region close to the surface, transfer of solute from the bulk solution to the
adsorptive surface over the distance d will occur predominantly by diusion.
For such a simplied system the rate of solute transfer into the stagnant layer
and the subsequent degree
16
of mass transfer limitation for an adsorption
reaction (for a given receptor density, concentration of solute and intrinsic
reaction kinetics) can be estimated via eq. (4.14).
dfC
i
g
dt
E
D
i
d
2
C
BULK
fC
i
g k
0
1
fC
i
g 4:14a
fC
i
g
C
BULK
E
D
i
=d
2
D
i
=d
2
_ _
k
0
1
4:14b
Values of d for dierent ow congurations and magnitudes of the stirring/
ow rate have been calculated [28]; typical values
17
for current biosensor
devices are of the order of micrometers [28,36]. In general, the parameter d
representing the thickness of the stagnant layer will become smaller upon
Figure 4.6 Continued
16
Upper limit calculated by assuming irreversible reaction, i.e. k
2
0 s
1
and assuming steady-state
d{C
i
}/dt 0. The assumption of k
2
0 is equivalent to the case of initial adsorption rate and hence
mass transfer limitations will be at their greatest during initial adsorption rates.
17
For solutes with diusion coecients of the magnitude of moderately sized globular proteins
(DE1 10
10
m
2
s
1
).
increasing the rate of ow.
18
Using eq. (4.14b), we have estimated the eect for
the initial adsorption rate for various receptor densities and intrinsic kinetics of
the adsorbing system in Figure 4.6a and the results are shown in Figure 4.6b.
As can be noted, mass transfer limitations will be greatest for an adsorbing
system (solute plus matrix site) having intrinsically rapid adsorption kinetics
under experimental conditions where the adsorptive surface comprises a highly
concentrated array of matrix sites and the uid ow/stirring rate and the
solutes diusion coecient are both of small magnitude. Such considerations
have led to the general recommendation that optical biosensor experiments be
conducted under conditions of low matrix concentration and high ow/stirring
rates [4245]. As illustrated in Figure 4.6c, mass transfer-limited adsorption
data can be readily identied by the appearance of a downward curvature in
plots of the adsorption rate versus the concentration of adsorbate [46]. In
experimental studies of adsorption, the best approach is the elimination of any
mass transfer eects by good experimental design. When this is technically not
possible, a number of approximate strategies based on the two-state model
described in eqs. (4.6) and (4.9) have been developed [33,37,38]. More will be
said about the practical aspects of this extension to the standard analysis in
Chapters 9 and 10. Now, however, having addressed the basics of the mass
transfer process in optical biosensor experimentation, we turn our attention to
the reaction component of the adsorption event.
4.2.2 Adsorption Mechanisms
In this section we formalize the discussion of adsorption mechanisms by decon-
structing the subject into smaller component pieces.
19
We then use these building
blocks to describe a wide range of adsorption behavior. By doing this we hope
both to increase awareness of the pool of candidate models available for initial
consideration and to educate the reader as to the richness in variety of adsorption
behavior that is not conned to just the standard 1:1 binding model or variants
thereof. In eq. (4.4) we have essentially pictured the adsorption reaction as the
partition of a single solute for which the association rate function k
0
1
is sensitive
to the concentration of adsorbate in a manner determined by the true
18
One estimate of the magnitude of d is on the basis of a 5% approximation, i.e. the value of d at
which the rate of solute transport into the stagnant region is less than 5% of the rate at which
solute would enter via diusion, i.e. d E0.05D
i
/v(d). More detailed approaches to estimate d or a
d-like parameter can be found in refs. [35,36] for cuvette systems and [4041] for ow systems.
19
Standard practice in the selection of a mechanistic model to describe experimental kinetic and
equilibrium adsorption data involves selection of a number of likely candidate models followed by
non-linear regression to nd the model that best ts the data in a statistically relevant manner.
There are a number of dierent strategies for performing such non-linear regression analysis of the
data. Depending on the mathematical sophistication of the experimenter, this may be done using
user-written software or user-dened models in a non-linear regression software package. Alter-
natively, a number of semi-black box routines are provided by the instrument manufacturers or
interested third parties. A good starting point for information on the subject of modeling data by
non-linear regression analysis can be found in ref. [47]. Discussion of the relative benets of
dierent tting strategies is outside the scope of this chapter.
98 Chapter 4
second-order adsorption association rate constant, k
1
, the starting (hypo-
thetical
20
) total concentration of matrix sites, (C
x
)
tot
, and the surface function,
f(f), describing the concentration of available sites for a given extent of surface
occupation. By considering the process in this fashion we may neatly delineate
our discussion of adsorption mechanisms into a number of separate sections. In
the rst section we will concern ourselves with general models of idealized
partition occurring at a constant concentration of matrix sites [i.e. where f(f)
equals 1 and therefore k
0
1
equals k
1
(C
x
)
tot
]. In the second section we will focus on
how competition might aect the partition process of the single solute. In the
third section we will describe how the surface function changes with dierent
extents of surface occupation for dierent modes of adsorption.
4.2.2.1 Idealized Partition Processes
In chemistry, partition refers to the distribution of a solute between two
immiscible liquid phases at innitely dilute concentrations of the partitioning
solute. In adsorption experiments, a similar situation is achieved when the
solute is dilute and the adsorbate concentration is exceedingly low, thereby
allowing us to equate f(f)(C
x
)
tot
with (C
x
)
tot
. By examining the kinetics
associated with this low-concentration region we may successfully divorce
our analysis of adsorption from complications associated with any changes
associated with the surface function. A single molecular adsorption event may
occur as a mechanistically concerted adsorption event, observed to be occurring
in a single elementary step [eq. (4.15)].
21
fC
1
g
k
0
1
k
2
C
1
ads
4:15
Alternatively, a single molecular adsorption event may occur as a mechanis-
tically stepwise process, whereby dierent parts of the adsorption process occur
in a sequential manner with each event itself being dened by a unique
characteristic time scale. Although potentially an indenite number of inter-
mediate forms of adsorbate may exist, e.g. (C
1a
)
ads
, (C
1b
)
ads
, . . ., we express the
general concept in a limited two-state model [eq. (4.16)].
fC
1
g
k
0
1
k
2
C
1a
ads

F
k
4
C
1b
ads
4:16
The rate of interconversion between (C
1a
)
ads
and (C
1b
)
ads
is respectively
governed by forward and reverse rate constants F and k
4
(units of s
1
). If the
forward rate of conversion is independent of the local concentration of matrix
sites, F takes on a constant value of a simple rst-order rate constant. If,
20
By hypothetical we mean that the total concentration of adsorbate at maximal saturation
may dier from the total starting concentration of matrix sites if an adsorbate molecule covers
multiple potential binding sites.
21
From here on we will make specic reference to the solute type, i.e. in this case we refer to the
intimate concentration of solute of type 1 given by {C
1
}.
however, the rate of interconversion is dependent on the local concentration of
matrix sites, F represents a rate function composed, in the simplest case, of a
second order rate constant k
3
(units of l mol
1
s
1
), the total concentration of
matrix sites, (C
x
)
tot
(and a stepwise specic surface function, f
3
(f
1a
, f
1b
) that
describes the fractional availability of nearby matrix sites and is dependent on
the concentration of both types of adsorbed species.) The rate equation
[describing the formation of adsorbate for a mechanistically concerted reaction
denoted by eq. (4.15)] was given in eqs. (4.7)(4.9) and is shown in Figure 4.7a.
0
0
500 1000
Time (s)
(
C
i
)
a
d
s
(
M
)
(
C
i
)
a
d
s
(
M
)
0.2
0
0.2
(A)
(B)
0 500 1000
Time (s)
(C
1,1
)
ads
(C
1a,1
)
ads
(C
1b,1
)
ads
Figure 4.7 Adsorbate concentration vs. time in dierent partition processes. (A) Mech-
anistically concerted homogeneous partition (single class of solute and matrix
site resulting in formation of single class of adsorbate (C
1,1
)
ads
. (B) Mecha-
nistically stepwise homogeneous isomerization (single class of solute and
matrix site, adsorbate formation proceeds by way of a two-step reaction
resulting in formation of two classes of adsorbate (C
1a,1
)
ads
and (C
1b,1
)
ads
.
100 Chapter 4
For a stepwise reaction, the rate of formation of the dierent classes of
adsorbate
22
is given by eq. (4.17) (Figure 4.7b).
dC
1a
ads
dt
k
0
1
fC
1
g k
2
C
1a
ads
FC
1a
ads
k
4
C
1b
ads
4:17a
dC
1b
ads
dt
FC
1a
ads
k
4
C
1b
ads
4:17b
Assuming that the system is reversible, the most typical diagnostic for
identifying adsorbate isomerization
23
is from analysis of a dissociation phase
adsorption experiment [48]. Depending on the relative extents of occupation of
the intermediate states, a bimodal signature of the rate plot of d(C
1
)
ads
/dt vs.
(C
1
)
ads
will be apparent. Another diagnostic sign of adsorbate isomerization
will be an apparent incompatibility between the dissociation rate calculated
from analysis of the association phase and that calculated from the dissociation
phase [48]. Examples of such isomerizing systems have been encountered in
experimental studies of multivalent solutes; examples include antibody binding
studies [49] and studies of multimeric proteins [50]. A detailed treatise of the
theory can be found in ref. [51].
4.2.2.2 Eect of Competing Reactions
Depending on the nature of the experimental system (i.e. the solute, the
adsorptive surface and the solvent conditions), there may exist a degree of
heterogeneity in the eective strength of the interaction. This apparent heter-
ogeneity may result from heterogeneity of the solute or heterogeneity of the
matrix adsorption sites. Whatever the cause, analysis of heterogeneous binding
data will result in an apparent distribution of rate constants whose inuence
will be felt according to the relative concentrations of the dierent types of
solute and matrix adsorption sites. For the case where heterogeneity exists in
the solute alone, such that a number p of dierent intimate solute types
[{C
1
}, . . ., {C
p
}] exist, the observed rate of adsorption is given by the summation
shown in eq. (4.18a). For the case where heterogeneity is caused by a certain
number, q, of dierent types of matrix sites [(C
x
)
tot(1)
, . . ., (C
x
)
tot(q)
], the
corresponding summation describing the rate of formation is that shown by
eq. (4.18b). For the case where both the solute and the matrix site show
heterogeneity, the double summation in eq. (4.18c) is required for an adequate
description of the rate of adsorption.
24
jp
j1
dC
j;1
ads
dt
k
0
1j;1
C
j
_ _
k
2j;1
C
j;1
ads
_ _
4:18a
22
Again here limited to two for convenience.
23
Another potential case of adsorbate isomerization is due to reaction of the adsorbate with other
adsorbate molecules. This situation will be dealt with in the section on adsorbate clustering.
24
The heterogeneity reaction can be extended to include conformational change.
kq
k1
dC
1;k
ads
dt
k
0
11;k
C
1
f g k
21;k
C
1;k
ads
_ _
4:18b
jp
j1
kq
k1
dC
j;k
ads
dt
k
0
1j;k
C
j
_ _
k
2j;k
C
j;k
ads
_ _
4:18c
In the above formulation, the subscripted rate function, k
0
1(j, k)
and rate
constant k
2(j, k)
refer to the operative rate parameters for interaction between
solute of type j and matrix site of type k to produce an adsorbate (C
j, k
)
ads
.
Svitel et al. [52] have examined techniques for identifying and quantifying such
heterogeneity in biosensor data for interactions obeying simple 1:1 binding
models. Phenomenological descriptions of heterogeneity such as those de-
scribed by eqs. (4.18ac) also encompass competition reactions (Figure 4.8).
One well-known example of a biological competition reaction is that discovered
by Vroman et al. [53], in which the heterogeneous proteins of blood plasma
were found to adsorb on glass with dierent characteristic time scales and
binding anities.
4.2.2.3 Surface Functions for Dierent Modes of Adsorption
For a surface denuded of adsorbate, all potential matrix sites are available for
adsorption, making the surface function equal to 1. However, upon adsorption
of solute the fraction of available matrix sites will change in a manner dictated
by the way in which adsorbate molecules aect the likelihood of subsequent
adsorption events. The mathematical form of this dependence is known as the
surface function [5459] and its exact nature will dene the kinetic and equi-
librium adsorption isotherms [60], i.e. the dependence of the rate and equilib-
rium extent of adsorption on the concentration of intimate solute for a given
adsorbate concentration (and set solution composition and temperature). Here
we make use of simple geometric approximations to describe the physical
nature of the solute and adsorptive surface and treat attractive and repulsive
behavior through simple interaction potentials such as hard particle approxi-
mations or square well potentials (mesoscopic models). These mesoscopic
models are used here to describe the surface function for various adsorptive
behaviors. Such approaches are especially suited to the aims of this chapter,
in which kinetic models are used to comment on the phenomenon of adsorption
from the viewpoint of a measurement device that provides a single observ-
able experimental parameter related to the total concentration of all species
adsorbed at the interfacial layer. With this proviso, we begin our descrip-
tion of surface functions by rst more carefully dening the meaning of the
quantity f.
For the simple case of a spherical adsorbing solute, the surface function is
expressed in terms of the quantity f
i
, which expresses the ratio of the concen-
tration of adsorbate to the total concentration of matrix sites multiplied by n
i
, a
102 Chapter 4
term representing the average number of matrix sites physically covered by
adsorbate of type i [eq. (4.19)].
25
f
i

n
i
C
i
ads
C
x
tot
E
r
i
ads
pr
2
i
A
tot
4:19
At very low densities of matrix sites, such that adsorption at one site does not
inuence adsorption at another, n
i
1. At high matrix densities relative to the
size of the solute (see Figure 4.1), the matrix will approach the continuum limit
0
0
500 1000
Time (s)
(
C
i
)
a
d
s
(
M
)
0.2
(A)
(
C
i
)
a
d
s
(
M
)
0.2
(B)
0 500 1000
Time (s)
(C
1,2
)
ads
(C
1,1
)
ads
(C
2,1
)
ads
(C
1,1
)
ads
0
Figure 4.8 Adsorbate concentration versus time for dierent competition processes.
(A) Mechanistically concerted heterogeneous partition [one class of solute
and two classes of matrix site resulting in formation of two classes of
adsorbate (C
1,1
)
ads
and (C
1,2
)
ads
]. (B) Mechanistically concerted hetero-
geneous partition [two classes of solute and one class of matrix site
resulting in formation of two classes of adsorbate (C
1,1
)
ads
and (C
2,1
)
ads
].
25
The surface function will be a function of the relative orientation for an asymmetric adsorbate but
we do not treat that complication here.
(approximation shown in brackets). In this case, one may estimate the value
of n on the basis of geometric arguments alone. For planar adsorptive
surfaces characterized by an area density r
x
* of matrix sites (molecules m
2
)
or three-dimensional adsorptive surfaces characterized by a volume density
of matrix sites r
x
(molecules m
3
), we may approximate n
i
from knowledge
of the physical area and volume covered by the adsorbate yielding eqs. (4.20a)
and (4.20b), respectively.
n
i
!r
x
pr
2
i
at high densities for 2D surfaces 4:20a
n
i
!r
x
4=3pr
3
i
at high densities for 3D surfaces 4:20b
In discussing surface functions, we recognize that adsorption reactions can
be categorized into two general types [60], those capable of being saturated (i.e.
a xed number of adsorption sites) and those incapable of being saturated
(non-xed number of adsorption sites, usually corresponding to multilayer
formation) (Figure 4.9). In optical biosensor experimentation involving bio-
logical macromolecules, both types of adsorption are encountered [61,62] so we
discuss the general characteristics of each in turn.
Surface Phases Capable of Being Saturated. In the case of adsorptive surfaces
that are capable of being saturated, the simplest form that the surface function
can take is that suggested by Langmuir [63], whereby adsorption to any
particular matrix site is independent of all others and each binding event
decreases the total number of matrix sites by one (such that n
i
1). For a single
class of solute adsorbing to a single class of matrix site, we have [eq. (4.21)].
f
11;1
f
1;1
1 f
1;1
4:21
Provided that the necessary conditions concerning independence and capacity
for saturation described above are met, the Langmuir model may be extended to
describe the case of a heterogeneous solute or heterogeneous matrix site. For the
most general case of both heterogeneous solute and matrix [Eq. (4.18c)], a
separate surface function for each of the k types of matrix sites, f
1(j,k)
, can be
written as eq. (4.22), again within the remit of the Langmuir requirements.
f
1j;k
f
1;k
; . . . ; f
p;k
_ _
1
p
j1
f
j;k
4:22
If the matrix sites are densely packed relative to the matrix site spacing, the
adsorption of a solute molecule will inuence the potential for adsorption of
other solute molecules in the vicinity. This inuence may be repulsive, in which
case the surface function will decrease more sharply with increasing adsorbate
concentration than a corresponding hypothetical Langmuir isotherm
26
[60].
26
By corresponding hypothetical Langmuir isotherm we mean one dened by a surface function in
which the total concentration of matrix sites (C
x
)
tot
is not the true concentration of matrix sites but
that corresponding to the maximal adsorbate concentration (C
i
)
ads
.
104 Chapter 4
(
C
i
)
a
d
s
C
i
(A)
(B)
(C)
Figure 4.9 (A) Illustration of two general cases of adsorption behavior. Lower trace:
saturation due to a xed number of sites and monolayer formation
(saturation implies that no further adsorption will take place despite
increasing the concentration of solute). Upper trace: the type of adsorp-
tion that cannot be saturated due to multilayer adsorption. The dotted line
represents monolayer coverage. (B) Diagram showing apparent monolayer
coverage for saturated adsorption. (C) Diagram showing apparent multi-
layer formation for adsorption not capable of being saturated.
Alternatively, if the inuence is attractive, the extent of decrease in the surface
function versus the Langmuir case will lessen. At very high extents of attraction
the surface function may even mimic that associated with the Langmuir case [64].
The simplest manner of incorporating symmetrical repulsive interactions into
the surface function is to approximate the repulsive interaction by an equiv-
alent interaction of hard circles for planar (two-dimensional) surface phases
and hard spheres for three-dimensional surface phases [54,57] (Figure 4.10).
Such approaches usually take one of two general forms, depending on the
extent of the reversibility of the adsorption event or the degree of mobility of
the surface phase. For highly mobile surfaces and/or rapidly dissociating
systems, equilibrium uid models based on scaled particle theory (SPT) expan-
sions [65,66] are particularly useful for describing the surface function. For the
case of adsorption of a single class of solute to a 2D continuum of a single class
of matrix sites we have eq. (4.23).
f
11;1
f
1
exp ln 1 f
1

2f
1
1 f
1
f
1
1 f
1

2
_ _ _ _
4:23
The SPT models can easily accommodate a range of dierent size adsorbates
existing on the surface by making the surface function for a given elementary
step and a certain class j of solute, a function of the dierent extents of
fractional coverage of all the dierent adsorbate types on a single class of
surface matrix sites f
1,1
, f
2,1
, . . . , f
n,1
, e.g. f
1(j,1)
(f
1,1
, f
2,1
, . . . , f
n,1
). For a
range of dierent sized adsorbate molecules characterized by radii r
i
and
surface densities r
i
* (for a two-dimensional adsorptive surface), the surface
function describing the probability of nding a free matrix site is approximated
by eq. (4.24), for convenience expressed in radii of adsorbate on a near
continuum of matrix sites on the surface.
f
1;j;1
r
1
; . . . ; r
n
exp ln 1 p
i
R
2
i
_ _
_ _
2p
i
R
i
_ _
1 p
i
R
2
i
_ _
R
j
_ _
i
1 p
i
R
2
i
p
2
i
R
i
_ _
2
1 p
i
R
2
i

2
_ _
R
2
j
__
4:24
For irreversible adsorption reactions (k
2
0 s
1
), which occur on an ads-
orptive surface with matrix sites xed in position (termed immobile), the form
of the surface function will be more appropriately described by a modeling
approach based on the phenomenon of random sequential adsorption (RSA)
[67,68]. At high levels of surface coverage, the surface distribution of adsorbate
will be, on average, less eciently placed and this more selsh packing will
result in a reduced maximal extent of adsorption, f
max
. For monolayer
adsorption of one class of solute to a planar continuum surface, the maximal
extent of surface coverage for a reversible (or alternatively highly mobile
surface) adsorption reaction is f
max
0.906, and for an irreversible adsorption
106 Chapter 4
Figure 4.10 Adsorption behavior showing a saturation limit: repulsive adsorbate
adsorbate interactions. (A) Surface functions for adsorption of a single
class of monomeric spherical solute to a single class of adsorptive surface
sites versus fractional surface coverage for (1) Langmuir case [eq. (4.21)],
(2) 2D equilibrium uid calculation [eq. (4.23)], (3) random sequential
adsorption calculation [eq. (4.25)] and (4) 3D equilibrium uid calcu-
lation [eq. (4.26)]. (B) Illustration of the rapid decrease in the surface
functions for the particle model based cases [lines 24 in (A)]. Sub-
optimal placement leads to a rapid decrease in the number of available
sites. Here black circles represent adsorbate and red halos denote area
excluded for subsequent solute adsorption.
reaction occurring on an immobile adsorptive surface the corresponding
value of f
max
is 0.546 [54]. The surface function from such irreversible
random sequential adsorption of a single class of solute to a planar surface
has been calculated by Monte Carlo-based computer simulations [55]. A
compact polynomial description of the results of such simulations is given by
eq. (4.25).
f
1;1;1
f
1

1 f
1
=f
max
f g
3
1 0:812 f
1
=f
max
f g 0:2336 f
1
=f
max
f g
2
0:0845 f
1
=f
max
f g
3
4:25
At high extents of surface coverage, somewhere between the limits of
irreversible adsorption to an immobile surface and reversible adsorption to a
highly mobile phase, one may encounter kinetics reminiscent of a glassy state to
ordered state phase transition [69]. In this regime the rate at which the surface
function changes from the random sequential model to the equilibrium SPT
model will be determined by the kinetics of reorganization of the surface phase
dictated by the rate constants k
1
and k
2
and the adsorptive surface phase matrix
site diusion constants [70].
For a three-dimensional surface phase such as that presented by matrix sites
existing on a carboxymethyldextran gel layer derivatized with specic matrix
sites, the calculation/estimation of fractional site coverage and maximal extent
of occupation is more dicult.
27
In the volume transcribed by the gel layer a
signicant volume fraction is already occupied by the gel itself [31]. Addition-
ally, the fractional availability of matrix sites will in part be determined by mass
transfer into the gel layer. As solute will enter the surface phase from the top,
these surface sites will be preferentially occupied, possibly preventing access to
matrix sites below [32]. A possible starting point for the calculation of approxi-
mate surface functions for a highly derivatized gel matrix has been suggested
[71], which involves treating the gel layer as a 3D phase with evenly dispersed
matrix sites. Using such an approach one may calculate
28
f
1
(f
1
) for a
single class of adsorbing solutes using the 3D version of the SPT equation
[66] [eq. (4.26)].
f
11;1
f
1
exp ln 1 f
1

7f
1
1 f
1

_ _
7:5f
2
1
1 f
1

2
_ _
3f
3
1
1 f
1

3
_ _ _ _ _ _
4:26
For the alternative case of attractive surface interactions resulting in prefer-
ential cluster formation, the surface function will decrease less sharply than for
the purely repulsive surface interaction case with increasing adsorbate
27
For freely accessible insertion of spheres into an open volume, f
max
varies between 0.636 and
0.7405 for the equilibrium random and jamming close packing limit.
28
Using eq. (4.20b) for the estimation of n
1
and neglecting the presence of the gel layer as a rst
approximation.
108 Chapter 4
concentration and may even approach the behavior dened by the surface
function for the Langmuir isotherm [64]. For a highly mobile/reversible
surface, clustering may occur via a post-adsorption interaction in a stepwise
reaction event (Figure 4.11a) and may be viewed as a type of isomerization
reaction as derived in eq. (4.17). Alternatively, clustering of adsorbate may
occur as part of the elementary process of adsorption (Figure 4.11b) [72].
With regard to the rst mode of cluster formation (monomer adsorption
followed by cluster growth), we present a cluster isomerization/growth
model based on monomeric adsorbate addition and monomeric adsorbate
loss in eq. (4.27) on a single class of matrix sites. In this model, an adsorbate
cluster containing i monomers is specied by (C
1
)
ads[i]
and the intrinsic
association and dissociation rate constants for clusters of size i on the
surface as k
3[i]
and k
4[i]
, respectively. By varying the rate constants for
formation of a cluster of i monomers, k
3[i]
and dissociation of an adsorbate
from a cluster of i monomers into monomeric adsorbate and a cluster of
size (i 1) monomers, k
4[i]
, between large and low values one can describe
the widest possible range of behavior corresponding to a high mobility and
low mobility surface phase, for the case where the attractive potential exerted
Figure 4.11 Adsorption behavior showing a saturation limit: attractive adsorbate
adsorbate interactions resulting in cluster formation happening as a
result of (A) post-adsorption association due to surface mobility and (B)
association occurring in concert with the adsorption event.
by the cluster does not increase the rate of solute adsorption but only eects
the rate of adsorbate desorption.
dC
1
ads1
dt
k
0
1
fC
1
g k
2
C
1
ads1
Z
i2
k
3i
C
1
ads1
C
1
adsi
k
4i
C
1
adsi
_ _
4:27a
d C
1

adsi
dt
k
3i
C
1

ads1
C
1

adsi1
k
4i
C
1

adsi
k
3i1
C
1

ads1
C
1

adsi
k
4i1
C
1

adsi1
for 2 ioz
4:27b
d C
1

adsz
dt
k
3z
C
1

ads1
C
1

adsz1
k
4z
C
1

adsz
for i z 4:27c
For cluster formation occurring in 2D, an initial approximation of the
system can be made by modeling each cluster of i monomers as a circular
aggregate of radius r
i
r
1
i, thus allowing the use of surface functions described
for the 2D continuum surface phase [eq. (4.24)].
29
Figure 4.12a describes the change in surface function for the situation when,
on the time-scale of adsorption, clustering is rapid. In the alternative case,
where cluster growth occurs contemporaneously with the adsorption event, we
model cluster formation by dening an attractive potential that contributes a
stabilizing energy, DE
c
, to monomer adsorbed in an area surrounding an
already adsorbed monomer or cluster. For simplicity, the stabilizing potential
can be considered as a square well that projects a small distance d from the
cluster perimeter and the distance of closest approach.
30
When comparing
monomer adsorption to two equal areas of surface, one existing within the
attractive well and the other outside it, the overall anity of the monomer for
the adsorptive surface is modied by a unitless factor K
C
[eq. (4.28)].
K
C
exp
DE
C
RT
_ _
4:28
Equation (4.28) represents an equilibrium stabilization factor. However, for
a kinetic model one needs to parse out the contributions into the individual
forward and reverse rate constants. In principle, the stabilizing eect could be
29
Note that in reality the cluster will be a chain of aggregated adsorbate having a shape that will
display fractal-like characteristics. This fractal-like quality will have an eect on the adsorption
kinetics. This topic is dealt with in Chapter 5.
30
For reasons of stability of the solution of equations, we actually consider that the zone of
attractive potential associated with adsorbate species lies between the distance of closest approach
(r
1
+r
i
) and a smaller distance (r
1
+r
i
d). Such an approximation will not signicantly change
the form of the eect upon the surface function. Additionally, this approximation will allow for the
estimation of an approximate surface function for this clustering mode up to fairly high degrees of
surface occupation (however, it will be less reliable as f approaches its maximum value). The
distance d should be chosen so that it is less than r
1
.
110 Chapter 4
Figure 4.12 (A) Surface function for the post-adsorption cluster formation case as a
function of the extent of cluster formation (line 1, k
3
/k
4
1; line 2, k
3
/
k
4
100; line 3, k
3
/k
4
1000). Inset describes the corresponding size
distributions of adsorbate on surface for the three dierent values of k
3
/
k
4
note that for the lowest value of k
3
/k
4
the adsorbate exists exclusively
as monomer. (B) Surface functions for the case of cluster formation
occurring in concert with adsorption. Line 1 describes f
1(1,1)
and line 2
describes f
1(1,2)
. The interaction distance was set at d 0.3r
1
. The addi-
tional line represents the calculation of the surface function for the
reduced radius [see eq. (4.30a) and footnote 31].
housed either entirely in k
1
or entirely in k
2
. For purposes of discussion the
eect is divided equally by modifying the idealized adsorption partition rate
constant of monomer to cluster by the factor O(K
C
) and the adsorption
dissociation rate constant of monomer from a cluster by the factor 1/O(K
C
).
In this case, the intrinsic rate constants describing monomer adsorption to a
region of surface within the attractive cluster potential are described by
eq. (4.29). For an already occupied surface the attractive well around each
adsorbate cluster eectively constitutes a dierent class of matrix sites to that
existing outside the zone of attractive potential, similar to the heterogeneous
case in eq. (4.18b), for which there are two classes of matrix site.
k
11;2
k
11;1
K
C
_
and k
21;2

k
21;1
K
C
p 4:29
For such a case we may estimate the surface function for the rst class of
sites (matrix alone) using eq. (4.24). The surface function for the second class
of sites can be calculated as the dierence between surface functions calculated
on the basis of circular species of true radius and species of modied radius r
i

d [see footnote 31 eq. (4.30a)]. The surface function specically associated
with each cluster composed of i monomers, f
1(1,2){i}
, can be parsed out by
multiplying eq. (4.30a) by the surface area of the potential region around the
i-sized cluster and dividing by the total area associated with all regions of
attractive potential [Eq. (4.30b)].
f
1;1;2
Ef
1;1;1
r
1
d ; . . . ; r
n
d f
1;1;1
r
1
; . . . ; r
n
4:30a
f
1;1;2i
f
1;1;2
C
1

adsi
r
i
r
1
d
2
z
k1
C
1

adsk
r
k
r
1
d
2
4:30b
By modeling each growing cluster as a circle that grows and shortens by either
addition or loss of a monomer unit, the rate of formation of each cluster of k
monomers can be expressed using eqs. (4.31ac). Figure 4.12b describes the
change in surface functions for this mode of adsorption with varying adsorbate
levels.
d C
1

ads1
dt
k
0
11;11
C
1
f g k
21;11
C
1

ads1
k
0
11;22
C
1
f g k
21;22
C
2

ads2
4:31a
d C
1

adsk
dt
k
0
11;2k
C
1
f g k
21;2k
C
1

adsk
k
0
11;2k1
C
1
f g k
21;2k1
C
k

adsk1
4:31b
d C
1

adsz
dt
k
0
11;2z
C
1
f g k
21;2z
C
1

adsz
4:31c
112 Chapter 4
Surface Phases Capable of Supporting Multilayer Growth. The treatment of
multilayer growth is a complex problem which has received a number of
reviews [73,74] and is important in many areas as diverse as semiconductor
preparation [75], gas wetting phenomena [76] and bio-nanotechnology [77].
Here, the aim is to provide a basic introduction to the subject that goes
beyond the usual cursory mention of the BET isotherm [78] by using simple
models to outline some limiting case behaviors of the major types of multi-
layer adsorption growth (Figure 4.12). As it vastly decreases the complexity
while still providing much chemical insight, we will restrict our kinetic
description of multilayer growth that proceeds eectively irreversibly
(i.e. k
2
-0 s
1
).
If the fractional coverage and pertinent rate parameters applicable to the rst
and subsequent layers are additionally appended by the subscripts {L1},
{L2}, . . . , then a full kinetic description of the system can be made by solving
one of the candidate surface functions for the primary adsorption layer
{L1}
and for each consecutive surface layer above layer one, {L2}, {L3}, . . . , etc.
To aid our discussion of multilayer formation we will employ the formalism
just described which pictures cluster formation as a form of heterogeneous
adsorption. By varying the value of k
1(1,1)
and K
C
[eq. (4.29)] for each adsorp-
tion layer, one can eectively describe many dierent modes of multilayer
formation.
One of the simplest multilayer growth modes is that of the Frankvan
der Merwe type, shown in Figure 4.13a [79]. This type of multilayer growth
involves sequential deposition of one layer to top of the previous layer. This
multilayer formation results from the fact that adsorption occurring in concert
with cluster formation is highly favored over adsorption in the absence of
cluster formation e.g. for an arbitrary layer i, k
1(1,2){Li}
ck
1(1,1){Li}
. In this
limit, the overall vertical rate of growth of the multilayer will be much slower
than its rate of lateral growth and adsorption of the new layer will generally
occur after deposition in the underlying layer has achieved a signicant cov-
erage. In analogy with condensation or crystallization phenomena, one may
liken the rst adsorption event to each new layer as a nucleation event which is
followed by a growth event (layer growth proceeding to coverage) [73]. As the
layer becomes signicantly covered, the chance of another nucleation event
becomes greater and the process may begin again. If the horizontal and vertical
adsorption rates are approximately equal [k
1(1,2){Li}
Ek
1(1,1){Li}
] a dierent type
of multilayer adsorption behavior, known as VolmerWeber growth [80] or
island growth, is observed (Figure 4.13b). In this case, adsorption mounds are
formed separate from each other. Similar reasoning can be used to describe the
phenomena of columnar multilayer growth shown in Figure 4.13c [81]. In this
situation, the vertical adsorption rate far outweighs the lateral adsorption rate
[k
1(1,2){Li}
{k
1(1,1){Li}
], leading to the growth of columns that may or may not be
vertical, depending on the initial orientation of the rst adsorbing molecules
and/or the degree of surface roughness.
Although a complete description of the adsorption kinetics for multilayer
growth is beyond the scope of this chapter, a fair approximation of the basic
(A)
(B)
(C)
Figure 4.13 Illustration of dierences in adsorption growth behavior. (A) Near
sequential formation of one layer after another due to preference for
lateral (edge to edge) growth classed as Frankvan der Merwe-type
growth. (B) Island formation adsorption behavior (also known as
VolmerWeber adsorption) due to similar preferences for lateral and
vertical growth modes. (C) Columnar growth behavior resulting from
strong preference for vertical growth modes only.
114 Chapter 4
behavior of all three preceding types can be arrived at on the back of the
following simplications:
1. The rst adsorption layer is allowed to form lateral interactions (cluster
formation as described in the previous sections).
2. Deposition of solute to the primary adsorptive surface to form the rst
surface layer is considered chemically distinct from adsorption to any
other surface layer, i.e. (k
1
)
{L1}
a(k
1
)
{LY}
, where Y denotes an adsorbate
layer greater than 1, i.e. Y41.
3. Deposition on top of the rst adsorbed surface layer is {L2} chemically
equivalent to deposition on any other higher layer ({L3},{L4} . . . ), i.e.
(k
1
)
{LY}
(k
1
)
{LW}
, where Y41 and W41.
4. The total surface area available for adsorption to a particular multilayer,
A
tot{LY}
above the rst layer (i.e. {LY} for Y41) at any stage of the
experiment is equal to f
{L(Y1)}
A
tot{L1}
. This statement is equivalent to
saying that there can be no unsupported growth. The remaining free
surface area not covered by adsorbate on the rst layer is equal to
(1f
{L1}
)A
tot{L1}
.
5. Surface functions for each layer of growth are calculated on the basis of an
assumed continuous surface area.
31
On the back of the preceding postulates the set of rate equations describing
the adsorption rate to the rst layer can be written using either Eq. set 4.27 or
4.31 and then used again to express the rate of adsorption to each particular
layer above the rst, using Eq. set 4.31.
4.3 Summary and Conclusions
The study of adsorption and interfacial phenomena is indeed a very important
subject in biology. Table 4.1 describes some of the essential roles that adsorp-
tion phenomena play in the fundamental processes of life. In this chapter, we
have examined how adsorption phenomena can be studied using optical
biosensor technology. After discussing pertinent features of the optical biosen-
sor measurement technique, we examined some of the physical chemistry
behind the process of adsorption. We formalized our discussion of adsorption
by breaking it down into its component pieces. We covered the process of mass
transfer to the surface and looked at how under some circumstances this could
be rate limiting. Under conditions in which mass transfer eects are negligible,
we described how the form of the adsorption progress curve would be deter-
mined by dierences in the adsorption mechanism. We further deconstructed
our analysis of adsorption mechanism into two separate discussions. The rst
31
Obviously this assumption will be weak when the preceding layer is not clustered and in this case
so-called edge eects will play some role. However, as the degree of cluster formation becomes
greater the assumption will become stronger.
concerned itself with dierent modes of idealized partition at zero adsorbate
concentration. Here, we reviewed several fundamental types of adsorption/
partition including homogeneous concerted, homogeneous stepwise and hetero-
geneous concerted and stepwise modes. In the second discussion, we examined
how each new adsorbate addition would aect the likelihood of the next
adsorbate addition. We then introduced and reviewed the dierent forms that
the surface function
32
may take for dierent types of adsorption events. By
extension, we examined how both attractive and repulsive interactions between
adsorbate molecules or between adsorbate and intimate solute molecules would
aect such surface functions. We also examined the eect on the surface
function of multilayer growth and introduced some of the basic modes that
such multilayer growth might take.
Many reviews tend to focus on adsorption mechanisms in the absence of
mass transport considerations or alternately put their focus on mass transfer
limitations while treating adsorption phenomena with overly simplistic 1:1
binding models. We feel that this review has lled a gap by providing an
introduction to the general features associated with the adsorption of macro-
molecules to surfaces by focusing on both areas. Throughout this chapter we
have tried to enter discussions from the viewpoint of a biochemist investigating
biologically related adsorption phenomena involving macromolecules. In this
vein, we have not concentrated on some of the typical topics more preferred by
chemists such as discussions of the energetic dierences between physisorption
and chemisorption.
33
Equally, we presented the discussion of mass transfer
eects in non-transformed quantities as opposed to the approaches preferred
by the (generally) more mathematically oriented engineering community.
However, with these caveats out in front we have unashamedly tried to engage
the reader with some of the complexity (and wonders) associated with the
biophysical approach to the study of adsorption. Interfacial events form such a
part of our everyday living that it is not overly dramatic to nish this review
with a line from a short poem by Vroman [82]:
All we can create and cry is interface
4.4 Questions
1. a. With respect to the phenomenon of adsorption dene the following
terms:
solute
adsorbate
binding/matrix site.
32
A probability function describing the likelihood of solute nding an available site on the surface
for adsorption.
33
Distinctions which have less functional meaning when examining the non-covalent adsorption of
very large molecules.
116 Chapter 4
b. With respect to the natures of both the solute and matrix site, discuss
what is meant by the terms:
distinct array of binding sites
continuum of binding sites.
2. a. Optical biosensors measure the amount of adsorbate directly, as
opposed to chromatographic procedures, which measure the amount
of adsorbate by calculating the dierence between the total and free
solute concentrations during and after adsorption. Give three advan-
tages of such a direct measurement technique for quantifying the
adsorption process.
b. If the eciency of detecting adsorbed solute using an optical biosensor
decays exponentially with distance normal to the surface, comment on
what type of adsorption reactions and adsorption geometries would be
the most straightforward to categorize.
3. a. Starting with the basic transport/reaction scheme outlined in eqn. (4.4),
derive the limiting case kinetic behavior for (i) transport-limited and
(ii) reaction-limited adsorption.
b. Comment on how mass transport might be accounted for in a com-
pletely general fashion regardless of tube/cell geometries.
c. What approximate forms of the general approach given in your answer
to (b) are useful for describing mass transport in a ow-through and
cuvette-type biosensor?
4. a. Adsorption reaction mechanisms can be described completely generally
as partition events in which the partition rate is a function of the extent
and type of surface occupation. Discuss the above statement in terms
of the components that constitute the association rate function k
0
1
[eq. (4.5)].
b. Write down kinetic mechanisms for the following types of adsorption
behavior where unless specied the adsorption is of a simple concerted
type:
adsorption of a homogeneous solute to a homogeneous array of
binding sites
multi-step adsorption pathway of a homogeneous solute to a homo-
geneous array of binding sites
adsorption of heterogeneous solute to a homogeneous array of
binding sites
adsorption of homogeneous solute to a heterogeneous array of
binding sites.
5. Adsorption reactions can be categorized into two general types, those
capable of being saturated (i.e. a xed number of adsorption sites) and
those incapable of being saturated (a non-xed number of adsorption
sites, usually corresponding to multilayer formation).
a. What surface functions are applicable to the following types of satu-
rable adsorption reactions?:
Langmuir adsorption
irreversible adsorption of spherical solute to a continuum array
of binding sites
reversible adsorption of spherical solute to a continuum array of
binding sites
reversible adsorption of spherical solute to a continuum array of
binding sites with adsorbate clustering leading to monolayer formation.
b. Describe in general terms the dierences between the three types of
multilayer adsorption growth with respect to the adsorbate preference
for forming lateral or longitudinal contacts with the already adsorbed
solute.
6. Mass transport-limited kinetics are benecial for concentration determi-
nation of the analyte in a sample. Why?
4.5 Symbols
v
i
partial specic volume of solute
(C
1
)
ads[i]
concentration of adsorbed monomer clusters on surface of size i
monomers
(C
i
)
ads
concentration of adsorbed solute of type i
(C
x
)
tot
initial total concentration of matrix sites
(K
i
)
e
eective adsorption partition constant [(K
i
)
e
f
i
/b
i
]
{C
i
} concentration of solute of type i spatially close to the surface
(termed intimate solute)
B multi-term ow parameter, dependent upon device geometry
b
i
eective dissociation rate parameter
C
BULK
solute concentration in the bulk solution
c
i
weight concentration of component i
C
i
concentration of solute of type i
c
v
weight concentration of viscogenic agents
D
i
solutes diusion coecient
DE
c
stabilizing energy (in cluster formation)
F forward rate constant (units of s
1
)
DS change in signal
f fractional site/area coverage
f
max
maximal adsorbate site coverage
f(f) unitless function describing the surface site occupation, where
Fn(Ci)ads/(Cx)tot (here n is the average number of sites
covered by the adsorbed solute)
f
3
(f
1a
,f
1b
) stepwise specic surface function, describes the fractional avai-
lability of nearby matrix sites
f
i
eective forward rate parameter
118 Chapter 4
Z aqueous solvent of viscosity
i component i
k Boltzmanns constant
k
1
intrinsic second-order rate constant
k
0
1
association rate function (unit: s
1
)
k
2
dissociation rate const (unit: s
1
)
k
3
second-order rate constant (units of l mol
1
s
1
)
k
a
phenomenological transport coecient
k
b
phenomenological transport coecient
k
3[i]
intrinsic association rate constant for clusters of size i on the
surface
k
4[i]
intrinsic dissociation rate constant for clusters of size i on the
surface
K
0
R
K
0
R
k
0
1
/k
2
K
T
partition constant for formation of the intimate
solute
M
i
solute i molecular weight
n refractive index
N
A
Avogadros number
n
i
average number of matrix sites occupied by adsorbate i
S signal (optical)
T temperature
v* average velocity of liquid in ow cell
v
BULK
velocity of the bulk of the liquid
v
i
linear velocity of solute
v
x
linear velocity in the x direction
z axis normal to the sensor surface
z
max
max height of ow channel
r
x
volume density of matrix sites (molecules m
3
)
r
x
* area density of matrix sites (molecules m
2
)
s decay constant (unit: distance)
Dr radius of cylindrical polymer rods
j fractional volume occupation of polymer
d small distance, the zero ow region extends out a
surface
4.6 Acknowledgements
From my time in the U.K. I would like to thank Professor Christopher M.
Dobson for providing me with space to work in his laboratory, his keen interest
in my research and his continued friendship. During this period I would like to
acknowledge the nancial support of the Human Frontiers Science Program
(HFSP) which nanced my stay at the Cambridge University Chemical
Laboratories (20032007) via the award of a HFSP Long Term Fellowship.
From my time in Japan I would like to thank Professor Haruki Nakamura and
Assoc. Professor Fumio Arisaka for the amazing support which they have
provided me and for which I am deeply indebted. I would like to acknowledge
nancial assistance from a RIKEN grant for Research and Development of
Next-Generation Integrated Life Simulation Software. Finally I am deeply
appreciative of the valuable assistance provided by Dr. N. Hirota.
References
1. A.E. Johnson, Trac, 2005, 6, 1078.
2. I. Mellman and G. Warren, Cell, 2000, 100, 99.
3. M.B. Neiditch, M.J. Federle, A.J. Pompeani, R.C. Kelly, D.L. Swemm,
P.D. Jerey, B.L. Bassler and F.M. Hughson, Cell, 2006, 126, 1095.
4. K.A. Janes, J.G. Albeck, S. Gaudet, P.K. Sorger, D.A. Lauenburger and
M.B. Yae, Science, 2005, 310, 1646.
5. W. Cho and R.V. Stahelin, Annu. Rev. Biophys. Biomol. Struct., 2005, 34,
119.
6. K. Kristiansen, Pharmacol. Ther., 2004, 103, 21.
7. R. Benton, Cell. Mol. Life. Sci., 2006, 63, 1579.
8. J.R. Pugh and I.M. Raman, Biophys. J., 2005, 88, 1740.
9. J. McGhee and P. von Hippel, J. Mol. Biol., 1974, 86, 469.
10. A.A. Aderem and D.M. Underhill, Annu. Rev. Immunol., 1999, 17, 593.
11. S. Abraham, S. Brahim, K. Ishihara and A. Guiseppi-Elie, Biomaterials,
2005, 26, 4767.
12. R.J. Green, M.C. Davies, C.J. Roberts and S.J. Tendler, J. Biomed. Mater.
Res., 1998, 42, 165.
13. D.J. Winzor and W.E. Sawyer, Quantitative Characterization of Ligand
Binding, Wiley-Liss, New York, 1995.
14. L. Gorton (Ed.), Biosensors and Modern Specic Biospecic Analytical
Techniques, in Comprehensive Analytical Chemistry, Series Ed. D. Barcelo,
Elsevier, Amsterdam, 2005, p. xliv.
15. J. Fraden, Handbook of Modern Sensors: Physics, Designs and Applications,
Springer, New York, 2004.
16. M. Terpstra, Biosens. Bioelectron., 1993, 8, ii.
17. R. Cush, J.M. Cronin and J.M. Stewart, Biosens. Bioelectron., 1993, 8, 347.
18. U. Jonnson, L. Fagerstram and B. Ivarsson, Biotechniques, 1999, 11, 620.
19. F. de Fornel, Evanescent Waves: From Newtonian Optics to Atomic Optics,
Springer-Verlag, Berlin, 2001, p. 1.
20. H. Raether, in Physics of Thin Films, G. Haas, M.H. Francombe and
R.W. Homann (Eds.), Vol. 9, Academic Press, New York, 1977, p. 145.
21. D. Shoup and A. Szabo, Biophys J., 1982, 40, 33.
22. L. Edelstein-Keshet, Mathematical Models in Biology, SIAM, Philadelphia,
PA, 2005.
23. H.C. Berg, Random Walks in Biology, Princeton University Press, New
Jersey, 1993, Chap. 3.
24. R.C. Weast (Ed.), CRC Handbook of Chemistry and Physics, 62nd edn.,
CRC Press, Boca Raton FL, 1981, Table F42.
120 Chapter 4
25. S. Uribe and J.G. Sampedro, Biol. Proc. Online, 2003, 5, 103.
26. A.G. Ogston, B.N. Preston and D.J. Wells, Proc. R. Soc. London A, 1973,
333, 297.
27. A. Pluen, P.A. Netti, R.K. Jain and D.K. Berk, Biophys. J., 1999, 77, 542.
28. V.G. Levich, Physicochemical Hydrodynamics, Prentice Hall, Englewood
Clis, NJ, 1962.
29. R.B. Bird, W.E. Stewart and E.N. Lightfoot, Transport Phenomena, Wiley,
New York, 1960.
30. R.W. Glaser, Anal. Biochem., 1993, 213, 152.
31. P. Schuck, Biophys. J., 1996, 70, 1230.
32. M.L. Yarmush, D.B. Patankar and D.M. Yarmush, Mol. Immunol., 1996,
33, 1203.
33. D.G. Myszka, X. He, M. Dembo, T.A. Morton and B. Goldstein, Biophys. J.,
1998, 75, 83.
34. C. Wofsy and B. Goldstein, Biophys. J., 2002, 82, 1743.
35. M. Elimelech, X. Jia, J. Gregory and R. Williams, Particle Deposition
and Aggregation: Measurement, Modelling and Simulation, Butterworth
Heinemann, London, 1998.
36. D.A. Edwards, SIAM J., 2000, 105, 1.
37. P. Schuck and A.P. Minton, Anal. Biochem., 1996, 240, 262.
38. R. Karlsson, H. Roos, L. Fagerstram and B. Persson, Methods Comp.
Methods Enzymol., 1994, 6, 99.
39. J. Witz, Anal. Biochem. 1999, 270, 201.
40. B.K. Lok, Y.L. Cheng and C.R. Robertson, J. Colloid Interface Sci., 1983,
91, 104.
41. S. Sjolander and C. Urbaniczky, Anal. Chem., 1991, 63, 2338.
42. D. Hall, Anal. Biochem., 2001, 288, 109.
44. D.G. Myszka, T.A. Morton, M.L. Doyle and I.M. Chaiken, Biophys.
Chem., 1997, 64, 127.
45. R. Karlsson and A. Falt, J. Immunol. Methods, 1997, 200, 121.
46. C. Calonder and P.R. Van Tassel, Langmuir, 2001, 17, 4392.
47. W.H. Press, S.A. Teukolsky, W.T. Vetterling and B.P. Flannery, Numerical
Recipes in C, Cambridge University Press, Cambridge, 2002.
48. T.M. Morton, D.G. Myszka and I.M. Chaiken, Anal. Biochem., 1995, 227,
176.
49. K.M. Mu ller, K.M. Arndt and A. Plu ckthun, Anal. Biochem., 1998, 261,
149.
50. N.L. Kalinin, L.D. Ward and D.J. Winzor, Anal. Biochem., 1995, 228, 238.
51. P.R. Van Tassel, L. Guemouri, J.J. Ramsden, G. Tarjus, P. Viot and
J. Talbot, J. Colloid Interface Sci., 1998, 207, 317.
52. J. Svitel, A. Balbo, R.A. Mariuzza, N.R. Gonzales and P. Schuck, Biophys. J.,
2003, 84, 4062.
53. L. Vroman, A.L. Adams and M. Klings, Fed. Proc., 1971, 30, 1494.
54. P. Schaaf, J. Voegel and B. Senger, J. Phys. Chem. B, 2000, 104, 2204.
55. P. Schaaf and J. Talbot, Phys. Rev. Lett., 1989, 62, 175.
56. I. Lundstrom, Prog. Colloid Polym. Sci., 1985, 70, 76.
57. R.C. Chatelier and A.P. Minton, Biophys. J., 1996, 71, 2367.
58. P. Wojciechowski and J.L. Brash, J. Colloid Interface Sci., 1989, 140, 239.
59. S. Stankowski, Biochim. Biophys. Acta, 1983, 735, 352.
60. A.W. Adamson, in Physical Chemistry of Surfaces, Wiley, New York,
1982, Chapter 16.
61. T.J. Halthur, A. Bjorklund and U.M. Eloson, Langmuir, 2006, 22, 2227.
62. D.R. Hall, N.N. Gorgani, J.G. Altin and D.J. Winzor, Anal. Biochem.,
1997, 253, 145.
63. I. Langmuir, J. Am. Chem. Soc., 1918, 40, 1361.
64. J.J. Ramsden, G.I. Bachmanova and A.I. Archakov, Phys. Rev. E, 1994,
50, 5072.
65. H. Reiss, H.L. Frisch and J.L. Lebowitz, J. Chem. Phys., 1959, 31, 369.
66. J.L. Lebowitz, E. Heland and E. Praestgaard, J. Chem. Phys., 1965, 43,
774.
67. M.C. Bartelt and V. Privman, Int. J. Mod. Phys., 1991, 5, 2883.
68. J.W. Evans, Rev. Mod. Phys., 1983, 65, 1281.
69. P.L. Krapivsky and E. Ben-Naim, J. Chem. Phys., 1995, 100, 6778.
70. X. Jin, G. Tarjus and J. Talbot, J. Phys. A, 1994, 27, L195.
71. D. Hall, Optical Biosensor Based Studies of Protein, Adsorption: Theory
and Measurement, PhD Thesis, University of Queensland Press, Brisbane,
2000.
72. A.P. Minton, Biophys. J., 1999, 76, 176.
73. P.G. Vekilov and J.I.D. Alexander, Chem. Rev., 2000, 100, 2061.
74. D.D. Vvdensky, A. Zangwill, C.N. Luse and M.R. Wilby, Phys. Rev. E,
1993, 48, 852.
75. F.A. Ponce and D.P. Bour, Nature, 1997, 386, 351.
76. U.G. Volkmann and K. Knorr, Phys. Rev. Lett., 1991, 66, 473.
77. J. Texter and M. Tirrell, AIChE J., 2001, 47, 1706.
78. S. Brunauer, P.H. Emmett and E. Teller, J. Am. Chem. Soc., 1938, 60, 309.
79. F.C. Frank and J.H. van der Merwe, Proc. R. Soc. London, Ser. A, 1949,
198, 205.
80. M. Volmer and A. Weber, Z. Phys. Chem., 1926, 119, 277.
81. G.S. Bales and A. Zangwill, J. Vac. Sci. Technol., 1991, 9, 145.
82. As quoted by J.D. Andrade, in Surface and Interfacial Aspects of
Biomedical Polymers, Plenum Press, New York, 1985, Preface to Vol. 1.
122 Chapter 4
CHAPTER 5
Kinetic and Thermodynamic
Analysis of LigandReceptor
Interactions: SPR Applications
in Drug Development
NICO J. DE MOL AND MARCEL J.E. FISCHER
Department of Medicinal Chemistry and Chemical Biology, Utrecht Institute
for Pharmaceutical Sciences, Faculty of Science, Utrecht University,
P.O. Box 80082, 3508 TB Utrecht, The Netherlands
5.1 Introduction
Increasing evidence can be found that describing receptor ligand interactions in
terms of a lock-and-key model is no longer adequate. Receptors can be
regarded as part of a molecular machinery, in which ligand binding forms a
trigger to activate or deactivate the machinery. According to this view, it is no
longer sucient to know how the key ts into the lock, but we should also nd
out the mechanism with which the key opens and closes the lock. In other
words, in drug design we would be interested not only in the anity of the
ligand for the receptor, but also in the changes of a biological receptor molecule
when it forms a complex with a ligand. Such changes may involve conforma-
tional adaptation, changes in solvation (i.e. ordering of water molecules) and
changes in molecular exibility.
Kinetics is a rather underdeveloped aspect of ligandreceptor interactions. It
is readily conceivable that in some cases, such as in dynamic regulation of signal
transduction processes, kinetic control prevails rather than anity control.
Rapid onset of formation and an optimum lifetime of the complex can be ne
tuned by appropriate association and dissociation kinetics. Explicit references
to the biological signicance of binding kinetics are rather scarce; some
examples are given by Schreiber [1]. Other examples include the serial triggering
123
of T-cell receptors [2] and the activation of the epidermal growth factor
receptor ErbB-1 [3].
Elucidation of the molecular architecture, using especially X-ray and NMR
techniques has been of crucial importance for understanding how a ligand
protein or proteinprotein interaction functions in the molecular machinery.
However, for a more complete understanding of the dynamic processes
underlying receptor activation, kinetic and thermodynamic studies of ligand
receptor interactions are needed. It is increasingly acknowledged that, to fully
appreciate relevant molecular properties of potential drug candidates in a drug
design process, there is a need for thermodynamic and kinetic studies [48].
Traditionally, vant Ho analysis has been used for thermodynamic studies.
More recently, the use of sensitive calorimetric techniques in drug research is
emerging [9,10]. Stopped-ow has been the method of choice to study kinetics
of molecular interactions. With SPR one now can derive kinetic and thermo-
dynamic parameters from a single set of experiments. SPR allows to follow the
mass change on the sensor surface in real time, yielding anity and kinetic
data. Thermodynamic and kinetic parameters can be derived from a series of
experiments in a temperature range. The combination of kinetic and thermo-
dynamic information from well-designed SPR experiments is unique and oers
an added value compared with separate techniques for kinetic and thermo-
dynamic information: it allows even a full transition state analysis of the
binding process from a single data set.
In this chapter, we describe examples of thermodynamic and kinetic analysis
of biomolecular interactions using SPR-based approaches that we have devel-
oped in recent years. These examples apply mainly to peptides, interacting
monovalently or bivalently with important signal transduction proteins, contain-
ing Src Homology 2 (SH2) and SH3 domains. These signal transduction proteins
are attractive targets for drug design. The underlying investigations are aimed at
validation of the SPR-based approach, at gaining insight into the mechanism of
the binding process and nally at using this insight in ligand design.
To be able to exploit SPR fully, a few initial problems had to be solved. Part
of these problems originated from the fact that in our investigations cuvette-
based SPR instruments were used. As discussed in Chapters 3 and 4, in ow-
based SPR instruments (e.g. Biacore), a continuous ow of the sample
enhances diusion of analyte towards the sensor surface. In cuvette-based
instruments, the hydrodynamic properties are controlled by constant agitation
of the bulk solution in contact with the sensor surface. The cuvette-based
design oers the following advantages: (1) open architecture allowing manual
interventions during a run and (2) long association times without extensive
consumption of often precious biological material. A disadvantage might be
that during the binding process the concentration of unbound analyte in the
cuvette is not constant. In this chapter, a correction for this eect is described.
Another complication associated with the cuvette design is that in the dissocia-
tion phase the analyte released is not removed from the bulk solution. This
problem has been solved by adding competing ligand to prevent rebinding of
released analyte during the dissociation phase.
124 Chapter 5
When the association rates are high compared with the diusion in the bulk
solution, mass transport limitation (MTL) occurs.
1
Association and dissociation
are aected by MTL to the same extent. We describe a simple method to estimate
the extent of MTL. As MTL can be easily included in a simple kinetic model, the
experimental association curves can be analyzed. Another problem, not related
to instrumental design, is that in principle the anity of the analyte for the
immobilized ligand at the sensor surface, as obtained in a direct SPR assay, is not
necessarily identical with that in solution. A method is described to obtain
thermodynamic binding constants for the interaction in solution, using competi-
tion experiments with a concentration range of the ligand of interest. Using this
approach, several ligands can be studied using the same sensor surface.
We should emphasize that in this chapter the focus is not so much on theory,
but rather on application. We would like to give the reader practical tools to
obtain reliable kinetic and thermodynamic parameters on the binding pro-
cesses. In order to achieve this, we need to provide the corresponding
conceptual and theoretical background. Following the outlined approach,
reliable kinetic and thermodynamic parameters can be obtained, which can
greatly increase our knowledge of binding processes. Later in this chapter we
show examples of how kinetic and thermodynamic analysis of interactions
using SPR can support chemical biology studies in general and rational
structure-based drug design in particular.
5.2 Anity and Kinetics of a Transport-limited
Bimolecular
2
Interaction at the Sensor Surface
In a standard SPR assay, one of the interacting partners (the ligand) is
immobilized on the sensor surface. The other component (the analyte) is added
in the solution, in our case in a cuvette. In our experiments, the ligand is usually
a peptide provided with a linker, to increase the distance between the binding
epitope and the matrix on the sensor surface, avoiding steric hindrance between
the bound analyte and the sensor matrix (see Figure 5.1).
The linker is also provided with a free NH
2
terminus, for covalent coupling
to the sensor surface using EDC/NHS chemistry.
3
This system guarantees a
homogeneous surface by uniform coupling of the ligand through the NH
2
group. The analyte is a protein with generally a much higher molecular weight
than the ligand. This increases the sensitivity of the assay, as the change in
SPR angle is proportional to the amount of bound mass. Hydrogel SPR
sensor chips are used, containing carboxymethylated dextran chains on a
50 nm gold surface (Figure 5.1), either from Biacore (Uppsala, Sweden) or
Xantec (Mu nster, Germany). Negatively charged ligands, e.g. peptides
1
For a more detailed treatment of mass transport limitation and diusion, see Chapter 4, Section
4.2.1.
2
A bimolecular interaction is a biomolecular interaction of only one analyte (A) which binds with
one ligand (B) to form complex (AB).
3
For further details, see Chapter 7.
125 Kinetic and Thermodynamic Analysis of LigandReceptor Interactions
containing phosphotyrosines (pY), are more dicult to immobilize, due to lack
of preconcentration at the sensor matrix, caused by electrostatic repulsion
between the negatively charged peptide and the negatively charged carboxyl
groups on the dextran chains. In such cases 1 mol l
1
NaCl is added to the
coupling buer to diminish electrostatic repulsion [11]. Our SPR instruments
(IBIS and Autolab ESPRIT) have two cuvette cells: a sample cell and a
reference cell. The two cells are treated in an identical way, the only dierence
being that only the sample cell contains immobilized ligands. The net SPR
signal (the signal in the reference cell subtracted from the signal in the sample
cell) is used for further analysis. Subtraction of the reference signal allows
correction for bulk eects due to addition of the analyte, for transient
temperature eects and for non-specic binding which occurs incidentally.
In a series of experiments at dierent analyte concentrations, the anity of
the analyte for the immobilized ligand can be assayed in several ways. The
method preferred by us is non-linear tting of the SPR signal at equilibrium
with a Langmuir binding isotherm. Alternative methods are based on the
kinetics of the interaction. These methods for determining the anity of the
analyte will be described in the following sections.
5.2.1 Anity Constants Derived from Equilibrium SPR Signals
For a simple bimolecular interaction with molecules A and B forming the
complex AB, the equilibrium association constant K
A
and dissociation constant
K
D
are given by eqs. (5.1a) and (5.1b):
K
A

AB
AB
; with K
A
in l mol
1
5:1a
Figure 5.1 Schematic view showing (a) the SPR sensor matrix existing of dextran
chains with carboxymethyl groups on a gold surface (50 nm) before
coupling, (b) immobilization of the ligand after coupling and (c) binding
of analyte to the surface.
126 Chapter 5
K
D

AB
AB
; with K
D

1
K
A
in mol l
1
5:1b
where the brackets [A], etc., indicate concentration of the molecules. In a well-
designed anity experiment, several analyte concentrations are used, which
should be in a range around the K
D
value. In SPR experiments, [AB] and [B] are
not approached as concentrations in solution, but as amounts at the surface
expressed as SPR signal. The amount of complex AB is proportional to the
shift in SPR angle [expressed in millidegrees (m1) or so-called response units
(RU)]. A conversion factor can be calculated for SPR response to concentra-
tions in the volume of the 100 nm dextran layer at the sensor surface (see, e.g.,
Box 5.1). The shift in SPR angle is recorded as function of time in a
sensorgram. In Figure 5.2, an example of sensorgrams, based on the net SPR
signal (R
sample cell
R
reference cell
), at dierent analyte concentrations is shown.
Using the kinetic evaluation software supplied with SPR instruments, the
shift in SPR angle at equilibrium (R
eq
) is readily determined (see Section
5.2.2.1). Frequently, the data are represented in a Scatchard plot (R
eq
/[analyte]
vs. R
eq
) as a straight line. However, large errors can occur in Scatchard plots,
especially at low concentrations, with small amounts of binding [13], therefore
we prefer non-linear regression using the Langmuir binding isotherm [eq. (5.2)],
in which [A] is the free analyte concentration and B
max
is the maximum binding
capacity in m1, when all binding sites on the sensor surface are occupied.
R
eq

A
A K
D
_ _
B
max
5:2
Examples of plots with ts according to the Langmuir binding isotherm are
shown in Figure 5.3.
Figure 5.2 Sensorgrams (net signal) of binding of v-Src SH2 protein to immobilized
EPQpYEEIPIYL-peptide. Start of dissociation is indicated by the arrow.
v-Src SH2 concentrations form top to bottom: 500, 333, 208, 125, 83.3 and
62.5 nmol l
1
. For further details, see ref. [12].
5.2.1.1 Correction for Depletion of Free Analyte Concentration in
the Cuvette
In a cuvette, the free analyte concentration decreases due to binding to the
sensor. This section describes how depletion of analyte can be quantied and
corrected for. The change in SPR angle (in m1) due to a binding process is
directly related to the amount
4
of bound material per mm
2
. Under equilibrium
conditions the amount of bound analyte is proportional to R
eq
(in m1) and the
surface of the sensor S (in mm
2
) in contact with the bulk solution. To relate
the amount of bound analyte to a decrease in the free analyte concentration, the
molecular weight (MW) of the analyte and the volume of the bulk solution
(V
bulk,
in liters) must be known. The depletion correction can be calculated
using eq. (5.3).
A
free
A
0
R
eq
S 10
9
122 MWV
bulk
5:3
Here [A]
0
is the initially added analyte concentration in the bulk and [A]
free
is
the corrected analyte concentration, both in nmol l
1
.
Two examples of depletion correction are presented in Figure 5.3 in (A) for
v-Src SH2 with molecular weight 12 300 Da and in (B) Syk kinase tandem SH2
Figure 5.3 SPR signal at equilibrium as function of analyte concentration. The lines
are the ts with the Langmuir binding isotherm [eq. (5.2)]. Data without
depletion correction, open circles; with depletion correction (see Section
5.2.1.1), closed circles. (A) Binding of v-Src SH2 domain (conditions as in
Figure 5.2). (B) Binding of Syk kinase tandem SH2 domain. For further
details on this interaction, see ref. [14].
4
For the IBIS and Autolab ESPRIT instruments used by us, an SPR signal of 122 m1 corresponds
to 1 ng mm
2
at 25 1C.
128 Chapter 5
(Syk tSH2) with MW is 29 800 Da. With eq. (5.3), using values for S (6 mm
2
)
and V
bulk
(35 ml, the applied volume), the depletion correction is calculated as
0.114 nmol l
1
/m1 for v-Src SH2 protein and 0.0472 nmol l
1
/m1 for Syk tSH2
protein. It is obvious from Figure 5.3 that for Syk tSH2 the depletion
correction has a larger eect: without correction K
D
is 8.7 nmol l
1
and with
correction K
D
is 5.9 nmol l
1
. Although the correction factor for v-Src protein
is larger due to the lower molecular weight, the depletion correction has only a
limited eect in this case: K
D
goes from 294 nmol l
1
without correction to
252 nmol l
1
with correction. For Syk tSH2 the eect of correction on K
D
is
much larger. Owing to the high anity, the Syk tSH2 concentration used in the
assay is very low (Figure 5.3). Depletion caused by binding to the sensor surface
has a large eect at low concentrations. Another factor with direct eect on
depletion correction is the binding capacity B
max
of the sensor surface, as it is
directly proportional to R
eq
[see eq. (5.2)]. To minimize depletion and the need
for correction, a low binding capacity is advised. In general, a value of B
max
of
100 m1 (or 500 RU in Biacore systems) is more than sucient for reliable
assays.
5
In equilibrium anity assays using, e.g., the Langmuir binding isotherm, the
depletion correction can be readily calculated by entering the proper numbers
in eq. (5.3) and by using a spreadsheet, the correction can be automatically
calculated for every data point. Problems may arise when the sensorgrams are
used for kinetic analysis. If the depletion correction is large, the free analyte
concentration will substantially diminish during the association phase and
second order kinetics might apply [15]. In our experience, as long as the
depletion correction is below 10% of the total analyte concentration, rst-
order kinetics can be safely used [11].
From eq. (5.3) it can be concluded that if B
max
is below 100 m1, for medium
strong interactions (K
D
E100 nmol l
1
) and analyte molecular weights higher
than 10 kDa, depletion corrections will be smaller than 10%. In kinetic analysis
of high-anity interactions as in the case of Syk tSH2 (see Section 5.4.1), one
should be aware of the occurrence of second-order kinetics. In these systems,
reliable kinetic analysis is possible based on rst-order association kinetics, on
surfaces with low B
max
(B50 m1) and only at higher concentrations, such that
depletion correction remains below B10%.
5.2.2 Anity Constants and Rate Constants Derived from
Kinetic Analysis
In the previous section we focused on equilibrium anity assays based on R
eq
.
Alternatives are oered by kinetic analysis based on the shape of the sensor-
grams, which can be useful when the association rate is slow. Especially in ow-
based SPR instruments lengthy association times to reach equilibrium may
cause problems due to large analyte consumption.
5
Or even lower, depending on the sensitivity of the instrument.
5.2.2.1 k
obs
Kinetic Analysis
Assuming a simple bimolecular interaction with analyte A interacting with
immobilized ligand B, forming the complex AB at the sensor surface, ideally the
SPR signal vs. time (R
t
) is given by eq. (5.4) [16].
R
t

k
on
AB
max
k
on
A k
off
1 e
k
on
Ak
off
t
_ _
5:4
Here k
on
and k
o
are the association and dissociation rate constants for
formation and dissociation of the complex AB, respectively. Note that R
t
is
proportional to the amount of complex AB. A new parameter k
obs
6
is dened as
k
obs
k
on
[A] +k
o
. Using software that is generally supplied with SPR instru-
ments, a t of the curve of R
t
vs. time yields k
obs
. When k
obs
is plotted vs. [A],
k
on
can be obtained from the slope of the curve and k
o
from the intercept. The
anity can be calculated as K
A
k
on
/k
o
or K
D
k
o
/k
on
. In Figure 5.4,
examples of k
obs
analysis are shown, using the data sets of Figure 5.3. The
parameters of the k
obs
analysis for v-Src SH2 are included in Table 5.1.
Although the plots are linear, as expected from theory, the results deviate
from the equilibrium analysis. Now for v-Src SH2 a K
D
value of 11.9 nmol l
1
is
found, which is almost two orders higher than obtained from the Langmuir
binding isotherm. For Syk tSH2 no anity could be determined using this
analysis because the intercept is slightly below zero. The reason for these
deviations is that these interactions are severely aected by mass transport
limitation (MTL), as appears in Sections 5.2.2.2 and 5.3. Under such condi-
tions, eq. (5.4), which forms the base for this analysis, no longer holds. Schuck
and Minton [17] showed with theoretical data how MTL inuences the k
obs
vs.
[A] plot. Further examples of how MTL aects the outcome of k
obs
analysis can
be found in the literature [17,18].
As shown above, a straight k
obs
plot by no means indicates that reliable
kinetic data can be derived. Unfortunately, a number of examples of erroneous
interpretations of kinetic data can be found in the literature, especially using
k
obs
-analysis or closely related methods. To avoid this pitfall, one should
be absolutely sure that MTL is not involved. A number of simple self-
consistency tests, e.g. comparing data from equilibrium and kinetic analyses,
should be performed before interpreting such kinetic analysis [20]. A simple
experiment to test whether MTL is involved is to measure the eect of addition
of binding ligand during analyte dissociation to prevent rebinding (see Section
5.3.2.2). The k
obs
analysis presented in the previous section has severe short-
comings as the outcome is very sensitive to MTL. In the following, an
alternative model is oered which also includes MTL. This approach is based
on analysis of sensorgrams and curve tting according to predened binding
models.
6
In earlier publications regarding kinetic evaluations [16], this parameter is denoted k
s
.
130 Chapter 5
5.2.2.2 Global Kinetic Analysis with a Simple Bimolecular
Binding Model
The real-time information on the mass changes resulting from the interaction
can be used to study various binding models, also including MTL. In this
Table 5.1 Kinetic and anity parameters for v-Src SH2 protein binding to
immobilized EPQpYEEIPIYL-peptide (experimental data shown
in Figure 5.2), as derived with dierent approaches (see text).
Method parameter Binding isotherm k
obs
analysis
CLAMP global analysis
Model 1 Model 2
K
D
(nmol l
1
) 252 13 11.9
a
308
a
250
a
B
max
(m1) 323 8 363 3 323
k
on
(l mol
1
s
1
) 9.2 (0.3) 10
4
3.35 10
4
7.99 10
6
k
o
(s
1
) 1.1 (0.8) 10
3
0.01 2
b
L
m
(ms
1
)
c
6.3 10
6
Res Ssq
d
4.19 2.00
a
Calculated from k
o
/k
on
.
b
Experimental value from dissociation in the presence of peptide to prevent rebinding.
c
Calculated from k
tr
with conversion factor (see Box 5.1).
d
Residual sum of squares [19], indicates quality of the t.
Figure 5.4 k
obs
plots for binding of v-Src SH2 (closed circles) and Syk tSH2
(open circles) to immobilized ligands. Analysis of datasets presented in
Figure 5.3.
section, the emphasis is on the rate constants of a simple, transport-limited
bimolecular reaction, more complicated binding models are presented in
Section 5.4. In this chapter, the kinetic analysis is treated using basic chemical
kinetics concepts applied to experimental SPR data. In Chapter 4, kinetics is
treated with concepts based on physical solute absorption to surfaces.
7
In general, dierential rate equations for species binding to the sensor surface
can be derived from a binding model. Experimental sensorgrams can be tted to
a model and one can analyze how far the experimental data agree with the model.
Furthermore, parameters such as rate constants and maximum binding capacity
can be derived from the ts. In a global analysis such t procedures are applied to
several curves obtained at dierent analyte concentrations simultaneously, using
the same t parameters. To explain the procedure we use a simple bimolecular
binding model with and without mass transport step (see Scheme 5.1).
For model 1, the following dierential rate equations can be derived:
Association :
dAB
dt
k
on
AB k
off
AB 5:5a
Dissociation :
dAB
dt
k
off
AB 5:5b
The analytical solution for the association phase is eq. (5.4) and for dissociation
it is eq. (5.6), where [AB] is directly proportional to R
t
, the net SPR signal at
time t, and R
eq
the net SPR signal at dissociation time zero. Equation (5.6)
describes rst-order exponential decay from which k
o
can be obtained.
Dissociation : R
t
R
eq
e
k
off
t
5:6
Model 2 is somewhat more complicated: again in the association phase the time
dependence of [AB] is given by the dierential rate equation [eq. (5.5a)], but
now also the time dependence of diusion of analyte from the bulk ([A]
0
) to the
sensor surface ([A]) has to be taken into account. The accompanying rate
equations are given in eqs. (5.7a) and (5.7b).
dA
0
dt
k
tr
A
0
k
tr
A 5:7a
1: Bimolecular model 2: Bimolecular model + transport step
AB B A
on
off
k
k
+
AB B A
A A
on
off
tr
tr
k
k
k
k
+
0
Scheme 5.1 Binding models for a simple bimolecular reaction (1) and a bimolecular
reaction including a transport step of analyte from the bulk to the sensor
surface (2).
7
Remark: this is also the reason why terminology, symbols, etc., dier in Chapters 4 and 5.
132 Chapter 5
dA
dt
k
tr
A
0
k
tr
A k
on
AB k
off
AB 5:7b
Several programs can be used for solving such dierential equations by
numerical integration. We used the program CLAMP developed for tting
experimental sensorgrams [19].
8
Consistency of the ts is greatly improved by
tting several curves for dierent analyte concentrations simultaneously with the
same kinetic parameters in a so-called global analysis. Examples of global kinetic
analysis with CLAMP are shown in Figure 5.5, with the data set of Figure 5.2.
To emphasize the kinetic phase, only a relatively short association time interval
was allowed. The quality of the ts compared with the experimental data is
indicated by the residual sum of squares (res Ssq) parameter [19]. For models 1
and 2 this is rather similar (Table 5.1). However, the initial linear phase observed
for the higher concentrations is not very well tted with model 1, and the t
returns a k
o
value of 0.01 s
1
. This linear phase is indicative for MTL [21].
From experiments in the presence of competing peptide to prevent rebinding of
protein during the dissociation phase (see Section 5.3.2.2), a much higher
experimental value of 2 s
1
for k
o
is obtained. Therefore, we conclude that
this model does not yield a satisfactory description of the kinetic parameters.
The experimental values of K
D
and B
max
are known from the binding
isotherm (Figure 5.3), k
o
is known from dissociation experiments and k
on
can be calculated from k
o
/K
D
. Therefore, all parameters in model 1 are known,
allowing simulation of the sensorgrams based on model 1 (Figure 5.5B). This
simulation demonstrates that in practice the association phase proceeds much
slower than expected for the high on-rate of 8 10
6
l mol
1
s
1
. This
Figure 5.5 Global analysis using the program CLAMP of the association phase of
binding of v-Src SH2 protein to a pY-containing peptide (data set as in
Figure 5.2). Black lines, experimental curves, red lines, tted curves. (A)
Fit with bimolecular model 1, all t parameters are left free. (B) Simula-
tion of model 1 with xed parameters for B
max
(320 m1), k
on
(8 10
6
l mol
1
s
1
) and k
o
(2 s
1
). (C) Fit with transport model 2, with
k
o
xed on experimental value. See text for further details.
8
Currently, the features of CLAMP are included in a more extended biosensor data analysis tool
named Scrubber2 from the results of David Myszka (see http://www.cores.utah.edu/interaction/
software.html).
underscores MTL: due to the high on-rate, diusion of analyte to the sensor
surface is much slower and becomes rate limiting. In model 2, a step for
transport of analyte from the bulk to the sensor surface is included. This model,
using the xed experimental k
o
value of 2 s
1
, gives an excellent t to the
experimental data. The diusion of analyte to and from the sensor surface is
assumed to be equal and is characterized by the rate constant k
tr
.
From eq. (5.7b) it follows that the units of k
tr
obtained from the CLAMP t
are m1 s
1
l mol
1
, as [B] and [AB] are in m1 and the tted curves are SPR signal
(m1) vs. time (s). The value of k
tr
in m1 s
1
l mol
1
can be converted to the mass
transport coecient [21] (L
m
) in ms
1
(see Box 5.1). For v-Src SH2 protein
(MW12.3 kDa) this conversion factor is 6.6610
13
; applying this conversion
yields L
m
in ms
1
. In Table 5.1, the anity and kinetic parameters derived from
global analysis of the dataset of Figure 5.2, using models 1 and 2, are included.
The results illustrate once more that in this severely transport-limited system
the outcome of k
obs
analysis is not reliable. The anity from MTL model 2
agrees perfectly with that from equilibrium analysis using the binding isotherm
(Table 5.1). This is not surprising, as in a considerable part of the tted curves
the signal is at equilibrium (see Figure 5.5C). In model 2, the use of an
experimental value for k
o
is crucial for the outcome. In general, it helps to
use in the ts xed experimental values for, e.g., k
o
and B
max
.
Box 5.1 Conversion of k
tr
(in m1 s
1
l mol
1
) into the mass
transfer coecient L
m
(in ms
1
)
To calculate L
m
two conversions have to be applied: (1) from m1 to mol m
2
and (2)
from l mol
1
to m
3
mol
1
.
1. The SPR signal in m1 corresponds to a xed amount of material/surface unit.
For the IBIS and Autolab ESPRIT instruments used in these studies, 122 m1
corresponds to 1 ng mm
2
or 10
3
g m
2
. Taking into account the molecular
weight (MW) of the analyte, 1 m1 corresponds to
10
3
122 MW
mol m
2
.
2. 1 l mol
1
is 1 dm
3
mol
1
; this corresponds to 10
3
m
3
mol
1
. Combining 1 and
2, the conversion factor from m1 s
1
l mol
1
to ms
1
is
10
6
122 MW
. For v-Src SH2
protein with an MW of 12.3 kDa, the conversion factor is 6.66 10
13
. From
the t with model 2, k
tr
is found to be 9.49 10
6
m1 s
1
l mol
1
. Applying the
conversion factor, this corresponds to L
m
6.3 10
6
ms
1
.
In Biacore instruments, the SPR signal is expressed in response units (RU); 1 RU
corresponds to 1 pg mm
2
. As 122 m1 corresponds to 1 ng mm
2
(see above),
1000 RU corresponds to 122 m1, and 1 m1 is 8.2 RU.
In this chapter, calculations are based on m1. By using the conversion factor of 8.2,
these calculations can be adapted for RU-values.
It is surprising that the experimental data can be described by such relatively
simple models. For example, usually not all binding sites are equal: within the
134 Chapter 5
dextran layer of the sensor the binding sites more close to the gold surface have a
higher intrinsic contribution to the SPR signal, due to the exponentially decay-
ing evanescent eld (Chapter 2). Furthermore, especially on high binding
capacity surfaces approaching saturation of binding, heterogeneity of binding
sites is expected. The global kinetic analysis presented here is attractive because
it yields thermodynamic and kinetic parameters. However, one should be careful
in the interpretation of kinetic parameters as in the t procedure these can be
mutually correlated [22] and in more complicated binding models the separate
steps may not be completely kinetically resolved, as described in Section 5.4.
5.3 Detecting Mass Transport Limitation: A Practical
Approach
Kinetic and anity analysis with simple models can lead to large errors when
MTL is unaccounted for, as shown in Section 5.2. Therefore, it is necessary to
detect MTL. In this section, practical methods are described to nd transport
limitation.
9
Furthermore, we describe here two approaches to obtain true
o-rates
10
from severely MTL-aected dissociation phases.
5.3.1 Eect of Viscosity Change on the Association Phase
The essence of MTL is that the on-rate is high and diusion of analyte from the
bulk phase to the biosensor (and partly in the biosensor dextran layer [23])
becomes rate limiting. Viscosity changes of the bulk solution will aect
diusion of the analyte and this should be visible in the sensorgrams of an
MTL-controlled interaction. We performed experiments with increasing
amounts of glycerol to increase the viscosity. In Figure 5.6, the eect of
glycerol on the association of the GST fusion protein of the Lck SH2 domain
to immobilized pY-peptide is shown.
As expected, increasing the viscosity slows down the association. No eect of
glycerol in the applied amounts on the anity was observed (equilibrium signal
not aected). Kinetic analysis of data sets obtained with a range of glycerol
concentrations, using model 2 (Scheme 5.1), yields a series of k
tr
values and L
m
transport coecients. For ow cells, the ux to the sensor surface due to mass
transfer (i.e. L
m
) was derived to be proportional to D
2/3
[24]. The diusion
coecient D is reciprocally related to the viscosity Z, according to the Stokes
Einstein equation, and therefore L
m
should be proportional to Z
2/3
. From
Figure 5.7, it appears that a plot of L
m
vs. Z
2/3
yields a linear relation as
predicted for ow systems. For a cuvette instrument, the hydrodynamics might
be dierent, as the bulk solution is subject to continuous agitation. Actually,
with this data range it is not possible to discern the ow model from other
models, as a plot of L
m
vs. Z also has an excellent linear correlation.
9
For a more basic treatment of mass transport phenomena, see Chapter 4.
10
From theoretical considerations by Schuck and Minton [17], it follows that MTL aects
association and dissociation to the same extent.
Figure 5.6 Eect of glycerol on the association phase of 30 nmol l
1
Lck SH2 GST
fusion protein to immobilized Ahx-EPQpYEEIPIYL-peptide. Solid line,
no glycerol; dashed line, 5% glycerol; dot-dashed line, 7.5% glycerol.
Reprinted from ref. [11], Copyright (2000), with permission from Elsevier.
Figure 5.7 Relation between mass transport coecient (L
m
) from bulk solution to the
sensor surface and viscosity (Z) as predicted for the hydrodynamics in a
ow cell. Determined in a cuvette based instrument for binding of Lck
SH2 GST fusion protein to immobilized EPQpYEEIPIYL-peptide in the
presence of 0 to 10% glycerol.
136 Chapter 5
Attempts to correlate L
m
values with molecular weight have not been successful.
This is probably caused by dierences among individual sensors surfaces and the
fact that L
m
depends not only on diusion in the bulk solution, but also on
diusion within the sensor surface dextran layer as proposed by Schuck [23].
5.3.2 Transport Limitation in the Dissociation Phase
The high on-rate compared with diusion also aects the apparent dissociation
kinetics. If diusion is slow and the on-rate is high, a considerable amount of
dissociated analyte will rebind before there is an opportunity to diuse away
from the sensor surface into the bulk. This implies that if the association phase
is transport limited, the dissociation is also transport limited. In cuvette
instruments used in a static mode (i.e. released analyte is not removed from
the cell), rebinding will always occur due to the equilibrium between free
released analyte in the bulk and bound analyte on the sensor surface, even
under conditions that transport limitation does not apply!
We present two independent methods to assay true dissociation rates, which
gave comparable k
o
values. The rst method takes rebinding into account and
the second uses added competing ligands during dissociation to prevent rebinding.
5.3.2.1 Rebinding Model for Transport-limited Dissociation
If transport limitation applies, the dissociation phase for a simple bimolecular
interaction on a homogenous surface will no longer be described by rst-order
decay kinetics according to eq. (5.6). A high on-rate compared with diusion
away from the biosensor compartment and the availability of free binding sites
on the surface will increase rebinding of released analyte. Schuck and Minton
[17] have developed a two-compartment model as an approximate description
for the dissociation phase under ow conditions. This model is characterized by
the dierential eq. (5.8).
dR
t
dt

k
off
R
t
1
k
on
k
tr
B
max
R
t
5:8
In this model k
tr
(in m1 s
1
l mol
1
) has the same meaning as in model 2
(Scheme 5.1) and represents transport between the bulk and the sensor surface.
If k
tr
ck
on
no transport limitation will occur and eq. (5.8) then changes to
eq. (5.5b). (B
max
R
t
) represents the amount of free binding sites and will be
proportional to the amount of rebinding.
An example of application of this model using the program REBIND
11
is
shown in Figure 5.8. Continuous wash steps were performed to remove released
analyte. Initially, dissociation proceeds fast (Figure 5.8A), as at the start of the
dissociation practically all binding sites are occupied and rebinding is negli-
gible. Soon more binding sites become available, slowing dissociation due to
11
Kindly provided by Dr. Peter Schuck.
rebinding. Continuous renewal of the bulk solution increases the apparent
dissociation rate. The dissociation phase with the wash steps is excellently
matched by the model in eq. (5.8) (Figure 5.8B) with B
max
xed at the
experimental value derived from the binding isotherm [eq. (5.2)].
Using the sum of squared residuals (SSR) analysis of REBIND, a large
interval of 0.06 ok
o
o0.95 s
1
falls within 5% of the best SSR value (see
Figure 5.9, lower curve). In this interval, k
o
appears to be strongly correlated
with k
on
/k
tr
. This especially occurs if (k
on
/k
o
)(B
max
R
t
) c 1 [see eq. (5.8)],
which is the case under conditions of considerable transport limitation (k
on
and
B
max
R
t
are large).
If a surface is used with a lower B
max
, eects of transport limitation can be
somewhat diminished; however, in a severe transport-limited system, B
max
should be unrealistically low to prevent transport limitation completely (see
Section 5.3.3). Introduction of experimental values for k
on
and k
tr
, as derived
from global analysis of the association phase with, e.g., model 2 greatly
improves the results. As can be seen in Figure 5.9, the SSR analysis indicates
a discrete value for k
o
if k
on
/k
tr
is kept at the value from the association phase.
The obtained k
o
-value of 0.38 s
1
is close to the value found for the same
interaction (0.6 s
1
), in the presence of large amounts of competing peptide to
avoid rebinding (see Section 5.3.2.2.). The found value is also in accordance
with the rapid dissociation found for other SH2 domains [25].
For several reasons, this result is impressive. First, the slow decay in the
dissociation phase in Figure 5.8 in no way suggests such a high k
o
(see also
Figure 5.10). The results conrm that ignoring transport limitation can yield
Figure 5.8 Sensorgrams of binding of Lck SH2 GST fusion protein to immobilized
EPQpYEEIPIYL peptide. (A) Solid line, dissociation without renewal of
bulk solution; dashed line, dissociation with continuous wash steps to
remove released protein from the cuvette. (B) Upper lines: dotted line,
experimental data for dissociation with continuous wash steps; continuous
line, t of the data with the program Rebind based on dierential eq. (5.8).
Below: residuals of the t. Reprinted from ref. [11], Copyright (2000), with
permission from Elsevier.
138 Chapter 5
several orders too low k
o
values, as can be seen in Table 5.1. Second, the fact
that the transport parameter, k
tr
, derived independently from the association
phase, gives consistent results in the dissociation phase is a solid experimental
conrmation that both the association phase and the dissociation phase are
aected to the same extent by transport limitation, as concluded from
theoretical considerations [17].
In practice, it appears that dissociation curves that can be tted well with eq.
(5.8) can also be well tted by a double exponential dissociation equation for
two independent binding sites with each their own dissociation rate [17]. In
many cases such a t will result in an artifact and the obtained rates are not
meaningful. Before concluding from a double exponential t that two dierent
binding sites are involved, a simple consistency check should be performed, e.g.
by adding competing peptide to diminish/prevent rebinding (see Section 5.3.2.2).
5.3.2.2 Competing Ligand to Prevent Rebinding During
Dissociation
As indicated in the previous section, under transport-limited conditions the
dissociation phase is considerably inuenced by rebinding of released analyte.
In principle, this rebinding can be prevented by adding an excess of competing
ligand with high anity for the analyte. In Figure 5.10, examples are shown of
dissociation in the presence of increasing amounts of competing ligand for a
monovalent Lck protein and bivalent binding Syk protein.
Figure 5.9 Sum of squared residuals (SSR) as a function of k
o
. Broken line, analysis
with k
on
/k
tr
derived from global kinetic analysis of the association phase
based on model 2. Reprinted from ref. [11], Copyright (2000), with
The eect of the ligands is dramatic and illustrates that for a transport-limited
interaction, the o-rate can be several orders larger than expected from the
dissociation phase without competing ligand. Although the anity of these
proteins for the ligands is rather high, relatively high concentrations are needed
to prevent rebinding completely. In control experiments with high concentrations
(4200 mmol l
1
) of non-binding peptides, the dissociation rate is at maximum,
only a bulk eect, a higher baseline is seen, as also occurs in Figure 5.10B, for
4.4 10
4
mol l
1
ITAM peptide. At high concentration of binding peptide the
dissociation phase approaches a monophasic exponential decay (see Figure 5.11)
and the curve can be tted with eq. (5.6) to derive the o-rate.
In both cases the dissociation rate is very high. For the Lck protein,
rebinding (Figure 5.11A) still seems not to be completely suppressed in the
presence of even 10
4
mol l
1
peptide. Especially at low R values with more free
sites on the surface available for rebinding [see model eq. (5.8)], deviation from
rst-order decay kinetics is observed; however, the rst ca. 80% of the decay
can be approximated by the exponential function. The half-lifetime is about 1 s
and k
o
is 0.6 0.1 s
1
, close to the value obtained from the rebinding model
with xed experimental value for k
on
/k
tr
(0.38 s
1
; see Section 5.3.2.1).
The dissociation of the Syk protein (Figure 5.11B) is also speeded up in the
presence of competing peptide and shows monophasic exponential decay, with
k
o
0.13 0.02 s
1
, close to that of comparable mono- and bivalent interac-
tions involving SH2 domains [25,26]. The high dissociation rate in combination
with the high anity (K
D
5 nmol l
1
) is intriguing. As a rule, high-anity
monovalent interactions have slow dissociation rates, hence the complex has
long half-lifetimes, e.g. for the avidinbiotin complex it is over 1 week [27]. The
binding of a double phosphorylated ITAM-peptide with the tandem SH2
Figure 5.10 Eect of dierent concentrations of competing peptide ligand on the
dissociation rate. (A) 200 nmol l
1
Lck SH2 GST fusion protein with
EPQpYEEIPIYL peptide and (B) 5 nmol l
1
Syk tandem SH2 domain
with g-ITAM peptide. Reprinted from ref. [11], Copyright (2000), with
permission from Elsevier (A) and from ref. [14], with permission from
Wiley-VCH (B).
140 Chapter 5
domain is bivalent, existing of two weak monovalent interactions (see Section
5.4.1 for more structural details). For such multivalent interaction, the dis-
sociation rate approaches that of a single (much weaker) monovalent interac-
tion contributing to the bivalent binding, when competing (monovalent) ligand
is present [28]. Therefore, in the presence of proper ligands, multivalent
interactions oer the opportunity to combine high anity with high o-rates
and short lifetime of the complex. In signal transduction processes, the
combination of high anity and short half lifetimes might be decisive for
specicity and transiency of proteinprotein interactions. It is remarkable that
multivalent interactions, e.g. involving tandem-SH2 and tandem-SH3 domains,
are abundant in signal transduction processes.
5.3.2.3 Experimental Procedure to Assay High O-rates
O-rates approaching 1 s
1
, as obtained in the previous section, are at the limit
of what can be accurately measured by SPR. For really fast decay kinetics we
think that the open cuvette structure is an advance as it allows direct
accessibility for manual handling. Our protocol developed for assaying rapid
dissociation kinetics in cuvette-based instruments starts with setting the instru-
mental sampling rate high (5 data points s
1
). The instrument is operated in
one-channel mode. A 25 ml volume of the analyte protein, preferably with a
concentration that saturates 490% of the binding sites,
12
is pipetted manually
into the sample cell. After reaching equilibrium of binding, the measurement is
Figure 5.11 Dissociation of bound proteins in the presence of competing ligands to
prevent rebinding. (A) Lck SH2 GST fusion protein in the presence of
100 mmol l
1
peptide. (B) Syk tandem SH2 protein with 220 mmol l
1
of
bivalent binding ITAM-peptide. The monophasic-exponential or rst-
order decay ts [eq. (5.6)] are indicated by the dotted lines. Reprinted
from ref. [11], Copyright (2000), with permission from Elsevier (A).
12
This analyte concentration is approximately 10 K
D
.
started and very quickly 10 ml of a high-concentration competing peptide is
pipetted manually. The required concentration of peptide to prevent rebinding
has to be determined experimentally (Figure 5.10). The data points of the
resulting sensorgram can be exported, processed in a spreadsheet and tted to
an exponential function. The data points within 1 s after peptide addition are
discarded, as these are often aected by distortions. The experiment is repeated
at least in triplicate with ample manual washing steps in between.
5.3.3 Quantitative Considerations on Mass Transport Limitation
As explained previously, MTL occurs if the reaction (binding) ux is much
higher than the transport ux of analyte from the bulk solution to the sensor.
These uxes are described by the transport coecient L
m
and the Onsager
coecient L
r
for the reaction ux [21]. A quantitative measure for MTL is
expressed in eq. (5.9).
MTL
L
r
L
m
L
r
5:9
If the analyte transport is totally rate limiting in the binding kinetics (L
r
cL
m
),
MTL will approach 1. L
m
is directly related to k
tr
as dened in model 2
(Scheme 5.1) for the association phase and eq. (5.8) for the dissociation phase.
L
m
in m s
1
is obtained by conversion of k
tr
as indicated in Box 5.1 in Section
5.2.2.2. The Onsager coecient of reaction ux (L
r
) in ms
1
is obtained from
eq. (5.10) [21].
L
r
k
on
B 5:10
Here k
on
is converted to m
3
mol
1
s
1
units.
13
At the start of the interaction [B]
equals B
max
, the maximum binding capacity of the sensor surface, i.e. the
concentration of free analyte binding sites on the sensor surface. B
max
can be
obtained from the Langmuir binding isotherm [eq. (5.2)] or from global kinetic
analysis in m1. Using the same approach as explained in Box 5.1 B
max
is
converted to [B] in mol m
2
with eq. (5.11).
B
B
max
122

10
3
MW
mol m
2
5:11
The extent of MTL allows one to consider whether MTL can be avoided by
adaptation of the experimental conditions, e.g. by lowering the binding
capacity on the sensor surface or increasing the diusion rate by increasing
ow or agitation of the bulk solution in the cell.
The global kinetic analysis with model 2 yields L
r
1.72 10
3
ms
1
and L
m
6.3 10
6
ms
1
for the interaction with v-Src SH2 as analyte
14
13
10
3
times k
on
in l mol
1
s
1
.
14
MW12.3 kDa.
142 Chapter 5
(see Figure 5.5). Applying eq. (5.9), under these conditions, MTL is practically
1 for this rather high binding capacity surface (B
max
320 m1). To reduce MTL
to 0.5 (L
r
L
m
), the binding capacity B
max
should be reduced to about 1 m1,
which is not a feasible assay condition. Increasing ow and agitation will not be
sucient, either. Assuming that a 5-fold increase in L
m
can be obtained, the
interaction will still be completely transport limited. In practice, the increase in
L
m
will be rather limited due to hydrodynamics (stagnant layer) and the
dimensions of the (ow) cells and because diusion within the dextran matrix
on the sensor is not sensitive for the ow conditions [23].
In general, for an analyte protein of approximately 40 kDa molecular weight,
L
m
is around 3 10
6
ms
1
. This number can vary somewhat depending on type
of sensor chip. Applying eq. (5.9), this implies that for a surface with
B
max
100 m1 (corresponding to 2 10
8
mol m
2
) that L
m
4L
r
if
k
on
o2 10
2
m
3
mol
1
s
1
, corresponding to k
on
o2 10
5
l mol
1
s
1
. This
number agrees very well with predictions from various MTL models [17,29].
The size of the analyte, of course, inuences the diusion rate and the k
on
value,
but as a rule of thumb, transport limitation will aect binding kinetics to a
dextran-based sensor surface, if k
on
is larger than 10
5
l mol
1
s
1
.
In summary, lowering the binding capacity and increasing the ow rate can
prevent MTL only in the case of slightly transport-inuenced kinetics. In
practice, we assume that the L
m
/L
r
ratio can be improved at most by about a
factor 5 on changing the experimental conditions. As a consequence, in
moderately and severely transport-limited cases an eect of MTL on the
kinetics cannot be avoided.
5.3.3.1 Flow or Cuvette?
One can ask whether a ow or cuvette instrument is to be preferred when it
comes to transport limitation. Although dierences in hydrodynamic behavior
may exist between a ow and a cuvette instrument and detailed hydrodynamic
models have been derived for ow cells [23,24], in practice, no signicant
dierences in transport uxes between ow and cuvette systems have been
observed. This is illustrated by the agreement of the L
m
value of
9.8 10
6
ms
1
for IL-2 (MW 14 kDa) obtained in a ow-instrument (Biacore)
using a model similar to model 2 [30] and the value of L
m
6.5 10
6
ms
1
for the v-Src SH2 domain (MW 12.3 kDa) obtained in a cuvette instrument
(Table 5.1). We conclude that in practice no signicant dierences exist in the
extent of MTL between cuvette and ow instruments.
5.4 Global Kinetic Analysis of Complex Binding
Models
After describing simple bimolecular interactions, we shift to more complex
binding mechanisms with conformational changes, dimerization, multi-
component interactions, multivalent binding, etc. In addition to structural
information as derived from NMR and X-ray analysis, kinetic information and
insight into the mechanism is valuable for understanding the binding process
and is of special interest for rational drug design. We describe examples of
applications of global kinetic analysis with more complex models to illustrate
this point.
5.4.1 Global Kinetic Analysis Including Mass Transport and a
Conformational Change
For better understanding of our rst example, the bivalent binding of Syk
tandem SH2 domain (Syk tSH2) with a surface loaded with ITAM-peptide, we
start with the description of the structural aspects. The interaction of an ITAM-
derived ligand with Syk tSH2 involves bivalent binding of two phosphotyrosine
containing sequences on the ITAM-peptide with the two SH2 domains of the
Syk protein. This interaction plays an important role in, amongst others, signal
transduction of the IgE receptor (FceRI) and the B-cell antigen receptor [31].
An X-ray structure of Syk tSH2 with an ITAM peptide is shown in Figure 5.12.
The linker part in the ITAM, between the two phosphotyrosine-containing
sequences, hardly interacts with the protein [32]. The two SH2 domains in the
Figure 5.12 X-ray structure of Syk tandem SH2 domain (ribbon) with doubly
tyrosine phosphorylated ITAM (sticks). Both phosphotyrosines are
indicated in red. PDB entry 1A81 [32]. Reprinted with adaptation from
ref. [14], with permission from Wiley-VCH.
144 Chapter 5
Syk protein (labeled N-SH2 and C-SH2) are connected by a exible coiled coil
linker, giving some exibility in the inter-SH2 distance.
The association phase of this interaction was subjected to a global kinetic
analysis (Figure 5.13). The association phase was assayed at dierent tempera-
tures as part of a complete thermodynamic analysis (see Section 5.6 for
thermodynamic analysis based on SPR).
At 11 1C, deviation is observed using model 2 (Scheme 5.1) in global analysis
in the association phase as when the signal approaches equilibrium. It is
conceivable that a bivalent interaction occurs in two discrete steps [29], as
indicated in Box 5.2. After initial monovalent binding, the second step involves
a conformational (intramolecular) change, leading to a high-anity bivalently
bound complex AB* (model 3, Scheme 5.2). This interaction is certainly
transport limited, as we see a strong eect of added ligand on rebinding in
the dissociation phase (Figure 5.10B). Also indicative of transport limitation is
the initially linear association phase, especially at higher analyte concentrations
[21]. Therefore, a transport step is included in model 3.
Figure 5.13 Association phase sensorgrams for the interaction of Syk tandem
SH2 domain with an immobilized ITAM peptide. Global kinetic
analysis of the experimental data using CLAMP, according to the
indicated models in red. Reprinted from ref. [14], with permission from
Wiley-VCH.
Model 3: transport
conformation model
Model 4: transport
dimer model
Model 5: 1PP-bivalency
model
*
0
AB AB
AB B A
A A
k
k
k
k
k
k
+
B A A AB
AB B A
A A
k
k
k
k
2
0
+
+ 2 AB B AB
AB B A
k
k
k
k
+
+
Scheme 5.2 Complex binding models used in global kinetic analysis of association
phases.
Box 5.2 Relation of K
b
, K
conf
and K
obs
in the conformation
change model
Conformation change model 3 consists of two binding steps (see also Scheme 5.2): an
initial binding event occurs, characterized by the equilibrium association constant K
b
.
Second, a structural change in the bound state occurs, characterized by equilibrium
constant K
conf
, leading to a higher anity complex. For a bivalent interaction as in
the case of Syk t SH2 binding to doubly phosphorylated ITAM peptides, this last step
includes a structural arrangement of the complex with a second intramolecular
binding event.
K
b
and K
conf
are dened by eqs. (5.12) and (5.13):
K
b

AB1
AB
5:12
K
conf

k
conf
k
conf
AB2
AB1
5:13
The observed equilibrium association constant K
obs
is dened by eq. (5.14):
K
obs

AB1 AB2
AB
5:14
Note that [AB1] +[AB2] corresponds to the total amount of bound analyte, which is
proportional to the change in SPR signal R. Substitution of eqs. (5.12) and (5.13) in
eq. (5.14) yields
K
obs
K
b
1 K
conf
5:15
A plot of R
eq
vs. [A] will also for this case obey a binding isotherm t as demonstrated
in Figure 5.3, and from the t B
max
and K
obs
are obtained.
Applying model 3 gives an excellent t (Figure 5.13B); in the ts, B
max
and k
o
were kept at experimental values.
15
The t yields the parameters k
tr
, k
on
, k
conf
and k
conf
; the values especially of k
on
, k
conf
and k
conf
may not be reliable as
they are strongly correlated (see Table 5.2). According to model 3, K
obs
contains
contributions of the initial binding step characterized by association constant K
b
15
B
max
is derived from the Langmuir binding isotherm [eq. (5.2)], which also holds for model 3 (see
Box 5.2); k
o
is derived from the experiment shown in Figure 5.11B.
146 Chapter 5
and of the second intramolecular binding step or conformation change K
conf
(see Box 5.2). Notwithstanding uncertainty in the rate constants, a Monte Carlo
run
16
with CLAMP gives consistent values for K
b
(k
on
/k
o
) and K
conf
(k
conf
/
k
conf
) of 1.9 10
7
l mol
1
and 18.5, respectively. This value of K
b
is signicantly
higher than the 7.7 10
5
l mol
1
found for the monovalent binding of mono-
phosphorylated ITAM peptide [33]. It is likely that the two binding steps in
model 3 are not completely resolved. First, no obvious biphasic association
phase exists in Figure 5.13; second, by increasing the temperature above
30 1C, we obtain an excellent t using the bimolecular transport model 2
(Figure 5.13C), indicating that the two binding steps can no longer be discerned.
Calculation of K
obs
from the t parameters using eq. (5.15) (Box 5.2) yields a
value of 3.6 10
8
l mol
1
, in excellent agreement with K
obs
obtained from the
binding isotherm at 11 1C (3.4 10
8
l mol
1
). Although the values obtained for
K
b
and K
conf
may not be physically meaningful, they can be used to calculate
solid K
obs
values with eq. (5.15).
5.4.2 Unusual Kinetics: Intermolecular Bivalent Binding to the
Sensor Surface
A second example from our work where global kinetic analysis plays a central
role in elucidating the binding mechanism to a SPR sensor surface is Grb2
protein binding to an immobilized bivalent polyproline (PP) peptide (2PP). 2PP
contains two PP binding epitopes derived from the SOS protein
17
separated by a
short linker moiety. The Grb2 protein exists of two SH3 domains and one SH2
domain connected by two exible linkers [34]. The two SH3 domains can each
bind to one of the PP epitopes of 2PP in a bivalent mode [34,35]. In Figure 5.14,
a model of the bivalent complex of Grb2 protein with 2PP is shown.
It was expected that the kinetics of this bivalent interaction could be
described by a model similar to that used for the Syk tSH2-ITAM interaction
(Section 5.4.1); instead, a dierent binding model emerged. The binding of
Grb2 to immobilized 2PP SPR sensor surfaces with dierent binding capacities
is shown in Figure 5.15. The form of the curve describing the association phase
Table 5.2 Parameters and their correlation from calculations used in Figure
5.13B.
Parameter Value Correlation 1 Correlation 2 Correlation 3
1 k
tr
(m1 s
1
l mol
1
) 5.4 10
7

2 k
on
(l mol
1
s
1
) 1.19 10
6
0.74
3 k
conf
(s
1
) 2.81 0.72 1.000
4 k
conf
(s
1
) 0.16 0.72 1.000 1.000
16
In a Monte Carlo run the t is repeated for a dened number of cycles with variation of the
start parameters within a dened range.
17
The SOS-Grb2 interaction plays an important role in the signal transduction cascade of numerous
receptors, controlling among others cell proliferation and dierentiation, platelet aggregation and
T-cell activation [36].
appears to be sensitive to the binding capacity: at high binding capacities, the
association phase looks conventional with a steady increase until equilibrium is
reached. Lowering the binding capacity yields biphasic association, with a very
fast initial increase, taking only a few seconds, followed by a slower increase
(Figure 5.15B and C). The SPR signals at equilibrium apparently comply with
the Langmuir binding isotherm [eq. (5.2)], as shown in Figure 5.16 for medium
binding capacity. At high and medium capacity surfaces the binding isotherm
Figure 5.14 Model based on X-ray structure of Grb2 (PDB entry: 1GRI, ribbons)
which a double polyproline peptide docked on the SH3 domains (sticks).
Figure 5.15 Sensorgrams of Grb2 protein (range 3302000 nmol l
1
) binding to
immobilized 2PP-peptide with various binding capacities (Table 5.3).
(A) and (B) net signal (reference cell subtracted from sample cell); (C)
data from the sample cell only due to non-specic binding in the
reference cell; lowest curve is the baseline not containing Grb2 protein.
148 Chapter 5
gave comparable K
D
values (Table 5.3). At low binding capacities equilibrium
is not reached within 20 min (Figure 5.15C).
The data from binding of Grb2 to the 2PP surfaces with various binding
capacities have been subjected to global kinetic analysis, exploring several
models. As shown in Figure 5.17A, the data from very high binding capacity
(Figure 5.15A) could be readily tted with a conformation change model
(model 3, Scheme 5.2). Interestingly, the transport step could be omitted from
the model, giving identical results. This suggests that in this case the interaction
is not transport limited, in agreement with the calculated value of
k
on
5.8 10
4
l mol
1
s
1
, which is below the indicated value for transport
limitation (Section 5.3.3).
At medium high and low capacity surfaces, model 3 shows systematic devia-
tions from the experimental data as shown in Figure 5.17B: the slope of the slow
phase in the ts changes much less with the concentration than experimentally
observed. X-ray structures suggest that Grb2 dimers can be formed [37,38] and
therefore the data were tted with dimer model 4 (Scheme 5.2), as can be seen in
Figure 5.16 Equilibrium analysis with binding isotherm of Grb2 protein to a medium
high capacity surface of 2PP peptide (data from Figure 5.15B).
Table 5.3 Data from binding isotherm of Grb2 binding to 2PP surfaces.
Binding capacity B
max
(m1) K
D
(nmol l
1
)
High 1520 55 535 40
Medium 218 3 460 30
Low 73 3 ND
a
a
ND Not Determined (equilibrium not reached).
Figure 5.17C. The residual sum of squares parameter from the t with model 4
was 1.71, compared with 2.50 for model 3. Also for the low capacity surface
tting with model 4 gave good results (Figure 5.17D). According to model 4
binding of the rst Grb2 molecule (A in the model) to immobilized 2PP (B in the
model) facilitates binding of a second Grb2 molecule. However, we doubt that on
the surface such physical Grb2 dimer will be formed, because we cannot
demonstrate the formation of dimers in solution upon addition of 2PP-peptide
to Grb2, either by chemical cross-linking or dynamic light scattering, in line with
published results [37].
An alternative to bivalent intramolecular binding is intermolecular bivalent
binding (see Scheme 5.3b). In the exible dextran matrix of the sensor chip, the
distance between 2PP epitopes could easily adapt to facilitate intermolecular
Figure 5.17 Global kinetic analysis of Grb2 binding to 2PP surfaces with CLAMP.
(A) High binding capacity surface, tted with conformation change
model 3. (B) Medium capacity surface, tted with conformation change
model 3. (C) Medium capacity surface, tted with dimer model 4. (D)
Low binding capacity surface, tted with dimer model 4. Details of these
models are given in Scheme 5.2.
150 Chapter 5
binding. If intermolecular bivalent binding occurs, this should be observed for a
monovalent 1PP surface (Scheme 5.3c).
In Figure 5.18, the association kinetics of 1PP surfaces with low and high
binding capacity are shown. The curves can be readily approximated with a
bivalent binding model: rst monovalent binding of Grb2 protein to immobilized
1PP, followed by bivalent binding to a second free 1PP ligand (model 5, Scheme
5.2). As expected, the rate of the slow bivalent binding step depends on the
binding capacity: with a high capacity it will be easier to nd a second 1PP ligand
Scheme 5.3 Schematic representation of various binding modes of Grb2 protein to
1PP and 2PP surfaces.
and the rate will be higher. Interestingly, the anity of the initial fast binding step
as derived from the kinetic analysis is approximately 12mmol l
1
, which is similar
to the anity of monovalent binding of 1PP to Grb2 in solution [35].
How can the intermolecular bivalent binding mode to the 2PP surface be
reconciled with the observed association kinetics which can be well described by
the dimer model 4 (Figure 5.17C and D)?
We propose a three-step mechanism as outlined in Scheme 5.4: the rst step is
monovalent (or bivalent) binding of Grb2 to one 2PP ligand; after this, bivalent
binding with another 2PP ligand occurs (model 5). This intramolecular binding
prepares a perfect second docking site for another Grb2 molecule as the inter-PP
distance is already optimal for bivalent binding. This model also explains why
the slow phase is much more delayed in case of low binding capacity (Figure
5.17): it is more dicult to nd a second partner for divalent binding. The
proposed binding model for the 2PP surface cannot be dened in all detail in the
CLAMP program, e.g. no bivalent ligands can be dened and no discrimination
between single and double occupation of two 2PP ligands can be made.
Actually, model 4 is a simplied approximation of the possible binding modes.
It is possible that in the dimer complex (step3, Scheme 5.4) dimeric interactions,
as found in the X-ray structure of Grb2 are involved, as the Grb2 molecules are
forced to be together. For the monovalent 1PP surface, such dimer formation
cannot take place and the kinetics can be adequately described with model 5.
5.4.3 Global Kinetic Analysis: Concluding Remarks
At rst sight, in both examples in this section we have a protein that binds
bivalently to an immobilized ligand. However, the outcome is surprisingly
dierent. This is illustrative for the use of models in global kinetic analysis: one
Figure 5.18 Association phase sensorgrams of Grb2 protein binding to a low (A) and
high (B) capacity monovalent 1PP surface. Global kinetic analysis with
CLAMP according to bivalent binding model 5 (see Scheme 5.2).
152 Chapter 5
should be very careful in the interpretation of the applied binding model. In
complex situations, good ts of the experimental curves can be derived from
more than one model. This raises the question of how far the physical reality is
reected by these models. A model will only provide an approximation: not all
reactions can be included in detail, as described above. Unless the kinetic steps
are well resolved, the calculated kinetic parameters may not be reliable, as they
will also be strongly correlated. Hence it is useful to introduce all possible
experimental values in the calculations.
Other complications may arise from the fact that not all ligands are equally
accessible: the sensor with immobilized ligand comprises a three-dimensional
Scheme 5.4 Sequential steps in the proposed model for binding of Grb2 to the 2PP
surface, ultimately leading to the formation of dimers.
volume. Especially for high binding capacity sensors, partition into this
volume may be hampered near saturation of binding due to crowding [23].
With multivalent binding analytes the hydrogel of the dextran on the sensor
may become more cross-linked, leading to a more compact structure during
the binding process. As the SPR signal decreases exponentially with distance
from the gold surface (Chapter 2), this might lead to an accounted signal
increase vs. the model. In spite of all these considerations, sometimes the
quality of the ts with complicated binding models can be stunning (Figures
5.17 and 5.18).
As a rule, a proposed binding model should be simple and supported by
experimental evidence; additionally, it should include all possible xed experi-
mental parameters. Global kinetic analysis is a unique tool, providing insight into
the binding mechanism, the kinetics of an interaction and the role of protein
dynamics. It can inspire new ideas for molecular design and drug development, for
example, the length and rigidity of the linker between the two phosphotyrosine-
containing binding epitopes in ITAM-mimetic constructs binding to Syk tSH2 [14].
Ample examples exist using the simple bimolecular models 1 or 2; applica-
tions of more complicated models are rather scarce. Such examples are the
binding of IL-2 to the heterodimeric IL-2 receptor [30], binding to a hetero-
geneous surface with two dierent ligands [39] and the kinetic analysis of
amyloid bril elongation [40]. Deviation from the simple 1:1 model is already
indicative of a more complex binding mechanism.
5.5 Anity in Solution Versus Anity at the Surface
In SPR measurements, interactions take place at the sensor surface, which is not
always representative of interactions in solution. This is certainly true for divalent
analytes, such as antibodies and GST fusion proteins that form dimers and show
an avidity eect when binding to a surface [41]. The amount of analyte binding to
the sensor surface in the presence of a competing ligand in solution is inuenced
by the anity of the analyte for this ligand. If the anity is high, a relatively large
amount of analyte will be in complex with the ligand in solution and only a small
amount of analyte will be available for binding to the surface, resulting in a lower
shift in SPR angle. Using this model, Morelock et al. developed a method to
obtain thermodynamic binding constants in solution [42]. Based on this, we
derived a tting model for data from competition experiments with constant
analyte and varying ligand concentrations in solution (Box 5.3) [43].
An example of competition experiments is shown in Figure 5.19. Experi-
ments were performed at various pH values to determine the shift in pK
a
of the
phosphotyrosine upon binding [43]. The equilibrium dissociation constant at
the chip (K
C
) was determined at each pH and these values were used in the ts.
The experimental data was tted with eq. (5.19) (Box 5.3), using experimental
values for [A]
tot
and K
C
, while the independent variable in the t is the ligand
concentration [B]
tot
.
154 Chapter 5
Box 5.3 Thermodynamic binding constants for binding in
solution
In an SPR competition experiment with ligands for the analyte present both on the
sensor surface and in solution, the two binding equilibria are as follows:
1. Interaction between analyte A and immobilized ligand B on the sensor chip
(Bc), yielding complex ABc on the sensor. The dissociation constant (K
C
) is
K
C

ABc
ABc
5:16
[Bc] and [ABc] are in millidegrees; when all Bc sites are occupied [ABc] B
max
.
2. Interaction in solution between ligand B in solution with A to form complex
AB. The dissociation constant (K
S
) is
K
S

AB
AB
5:17
Note that K
S
is a thermodynamic binding constant.
In analogy with eq. (5.2), the amount of binding onto the surface can be described by
a binding isotherm:
R
eq

Z
K
C
Z
_ _
B
max
5:18
where [Z] is the total concentration of analyte [A]
tot
minus the amount of analyte in
the complex AB ([Z] [A]
tot
[AB]), and eq. (5.18) changes to
R
eq

A
tot
AB
K
C
A
tot
AB
_ _
_ _
B
max
5:19
The amount of complex AB in solution is a function of the anity in solution (K
S
)
and eq. (5.17) can be rewritten:
K
S

A
tot
AB
_ _
B
tot
AB
_ _
AB
5:20
From this equation, it appears that [AB] is a quadratic function of the type
ax
2
+bx+c 0, for which the solution is given by the square root equation
AB
K
S
A
tot
B
tot
_ _
K
S
A
tot
B
tot
_ _
2
4A
tot
B
tot
_
2
5:21
Substituting eq. (5.21) for [AB] into eq. (5.19) yields an equation that ts data from
competition experiments. Fitting with [A]
tot
kept at the experimental value and [B]
tot
as independent variable provides K
S
and B
max
. Note that eq. (5.19) contains K
C
. The
value of K
C
is obtained from a separate experiment. Best t is expected when
[A]
tot
Z K
C
. B
max
is the maximum binding capacity upon complete saturation, and
not the binding capacity in the absence of competing ligand.
In order to verify the reliability of our approach for obtaining anity data in
solution and to see if anity at the sensor surface is signicantly dierent from
that in solution, K
C
and K
S
values are compared in Table 5.4. The data illustrate
that the anity of dimer proteins (the GST fusion protein and not-heated
Grb2-SH2; see below) at the surface is larger than in solution. This can be
explained by the avidity eect, occurring when the dimer binds bivalently to two
ligands at the surface. The case of the Grb2-SH2 protein without GST part is
interesting. It has been reported that this protein occurs as a dimer.
18
Probably
this is an artifact due to the expression as a GST fusion protein, which is known
to form dimers through the GST part [44]. From size-exclusion chromatography
we estimate that our GST-cleaved Grb2-SH2 protein contains B60% dimer.
The dimer is metastable and upon heating to 50 1C the monomer is irreversibly
formed [38]. Before heating, the anity to the sensor surface is higher, due to the
large amount of dimer. K
S
is higher before heating, suggesting that the anity of
the ligand for the dimer in solution is lower than for the monomer.
For the pYVNV-peptide binding to monomer Grb2-SH2, consistent values
for K
S
are obtained (230260 nmol l
1
), notwithstanding large dierences in K
C
(7.9 for the GST fusion protein to 790 for full-length Grb2) used in the
Figure 5.19 Data from SPR competition experiments to determine the binding
constant K
S
in solution. The analyte is Lck-SH2 GST fusion protein
(50 nmol l
1
), the immobilized ligand and the ligand in solution are
identical [a phosphotyrosine 11-mer peptide derived from the hamster
middle-T-antigen (hmT)]. Experiments were at dierent pH: from left
to right pH 9, 6.8 and 5. The lines are the ts with the substituted
eq. (5.19) (see Box 5.3). Reprinted from ref. [43], Copyright (2002), with
18
Dimer formation by domain swapping of an a-helix [38].
156 Chapter 5
calculations to obtain K
S
. Moreover, the solution anities agree with literature
values obtained using other techniques. This strengthens our condence in the
competition approach.
As a rule, the anity of monomer proteins in solution is the same as at the
sensor surface, even for the bivalent binding Syk tSH2! An exception seems to
be the binding of the full-length Grb2 protein to the SH2 domain.
19
In the case of a bivalent binding analyte with two (identical) binding sites such
as an antibody, the expression for [AB] will be dierent from eq. (5.21). Now we
have to take into account that not all occupied antibody remains in solution, as
monovalently occupied antibodies are able to bind to the sensor surface. When a
certain fraction of all binding sites are occupied, a statistically determined
distribution exists over double-bound, single-bound and unbound antibodies in
solution. Unbound antibodies, and also a single-bound antibody with a ligand
from solution, can bind to the sensor surface. We have adapted the expression
for [AB] (Box 5.3) to the statistical distribution [45] and it appears that this
correction has only a modest eect on the resulting K
S
value (o10%).
In summary, the approach derived in Box 5.3 cannot be used for every
binding model. Especially when the Langmuir binding isotherm is not suitable
for tting R
eq
as a function of analyte concentration, this approach will not be
valid. However, for more complicated binding models obeying the Langmuir
binding isotherm, such as the two-step model proposed for Syk tSH2 (Box 5.2),
reliable K
S
values can be obtained. In this case K
S
will be an apparent binding
constant, containing the various contributions to K
obs
(see Box 5.2). The
competition experiments as described in this section are very attractive in drug
research: the anity of a range of potential drug candidates can be assayed at
the same surface! In general the standard error in K
S
is larger than in K
C
.
Processes in solution may not always be representative for processes at
sensor surfaces or in biological systems. We are convinced that in some cases
interaction at a surface might be a better model than interaction in solution,
Table 5.4 Comparison of anity at the sensor surface (K
C
) and in solution
(K
S
), calculated from competition experiments using the substituted
eq. (5.19) (see Box 5.3).
Protein+peptide K
C
(nmol l
1
) K
S
(nmol l
1
)
Lck-SH2 GST fusion protein
a
+hmT-peptide 6 60
Grb2-SH2 GST fusion protein +pYVNV-peptide 7.9 260
Grb2-SH2 not heated
b
+pYVNV-peptide 134 1800
Grb2-SH2 heated +pYVNV-peptide 220 255
Full length Grb2 protein +pYVNV-peptide 790 230
Syk tSH2 +g-ITAM-peptide 5.4 5.7
v-Src SH2 +hmT-peptide 220 234
a
If explicitly indicated the dimer forming GST moiety is present.
b
Heating to 50 1C converts the Grb2-SH2 dimer irreversibly to the monomer (see text).
19
A possible explanation for this dierence is discussed in ref. [12].
especially with multivalent interactions. For example, the Sos-protein
20
contains multiple (six) polyproline sequences to recognize Grb2 SH3 domains
[36]. Several Grb2 molecules might bind bivalently to these sequences in one
Sos molecule in dierent combinations. For this a surface loaded with polypro-
line ligands might be a better model than 1:1 interactions in solution.
5.6 Thermodynamic vant Ho Analysis Using SPR
Data
As described in the Introduction, it is no longer opportune to describe ligand
receptor interactions in terms of a rigid lock-and-key concept. Binding of a
receptor by a ligand can inuence the dynamics, induce allosteric changes of the
receptor or, very importantly, have an eect on bound water molecules. All this
can be vital for the biological eects in a biomolecular interaction. In this
section we will concentrate on SPR-based assay of thermodynamic parameters,
to reveal the biomolecular recognition process, to help understand it and to
exploit it for improved rational drug design (see Box 5.4) [5,6,8].
5.6.1 vant Ho Thermodynamic Analysis
vant Ho thermodynamic analysis requires the measurement of the anity at
a range of temperatures. As the SPR signal is extremely sensitive to temperature
changes, complications may arise when measuring at temperatures deviating
from room temperature, as some time may be needed before complete thermal
equilibration is reached. An example is shown in Figure 5.20: in the reference
cell a temperature eect is observed in addition to a bulk eect. It takes about
100 s to reach thermal equilibrium. The temperature eect is not visible in the
net signal (Figure 5.20). The best approach is to use R
eq
for the anity assay as
by the time R
eq
is reached the system will be in thermal equilibrium. It is
important, especially when kinetics are assayed, that the sample is at the correct
temperature and that the injection system is well thermostated. The design of
newer generations of SPR instruments, e.g. Autolab ESPRIT, is optimized for
anity assays in a temperature range 1045 1C.
The simplest case of binding is a bimolecular interaction, such as that
between v-Src SH2 and hmT-peptide as described above. The vant Ho
thermodynamic analysis of this interaction is shown in Figure 5.21. The data
can be readily tted with eq. (5.26) and the resulting thermodynamic para-
meters are included in Table 5.5. The data show that the anity at the sensor
surface (K
C
) matches that in solution (K
S
), using the method described in
Section 5.5. The convex form of the curve indicates a negative value for the heat
capacity (DC
p
) and this interaction appears to be enthalpy driven. A concise
description of how to interpret thermodynamic parameters in terms of mole-
cular events during the binding process is given in Box 5.5. A more detailed
interpretation can be found in ref. [12].
20
This is a crucial interaction in the activation of the Ras signaling pathway described in ref. [36].
158 Chapter 5
Box 5.4 Thermodynamics of binding
For the simple bimolecular interaction between A and B yielding the complex AB, the
change in Gibbs free energy (DG) is related to the standard Gibbs energy change
under standard conditions (1 mol l
1
of A and 1 mol l
1
of B, at 25 1C):
DG DG
RT ln
AB
AB
_ _
RT lnK
a
RT lnK
d
5:22
where R is the gas constant and T is the absolute temperature. DG1 consists of a heat
component released or taken up during the binding process (enthalpy, DH1), and an
entropy (DS1) component related to the change in the degree of order of the system
due to binding:
DG
DH
TDS
RT lnK
a
5:23
For proteinprotein interactions and other biomolecular interactions DH1 and DS1
change with temperature. The temperature dependence of DH1 and DS1 can be
described in terms of the heat capacity (DC
p
) as given in eqs. (5.24) and (5.25):
DH
DH
DC
p
T T
5:24
DS
DS
DC
p
lnT=T
5:25
DC
p
is assumed constant within the applied temperature range. T1 is the reference
temperature, usually 25 1C, and DH1(T1) and DS1(T1) are the values of DH1 and DS1
at this temperature.
From eqs. (5.23)(5.25) follows the integrated vant Ho equation, eq. (5.26),
which describes the temperature dependence of the anity constant K
A
:
lnK
A

DH
RT

DS
R

DC
p
R
T T
T
_ _
ln
T
T
_ _ _ _
5:26
This expression can be used to t K
A
values derived at various temperatures versus
1/T, and yields the thermodynamic parameters DH1, DS1 and DC
p
. A concise
description of the interpretation of these parameters in terms of molecular events
related to the binding process is given in Box 5.5.
Box 5.5 How to interpret thermodynamic binding parameters?
It is important to realize that in a thermodynamic analysis two situations are
compared: the situation after the process (e.g. binding) is completed, vs. the situation
before the process. This is why we look at the dierence in the thermodynamic
parameters (indicated as D) including the whole of the process, e.g. also the solvent.
The thermodynamic analysis of a single interaction usually tells us whether the
binding is entropy or enthalpy driven, but it is not possible to interpret molecular
processes more in detail, e.g. whether replacement of water molecules is involved or
whether protein dynamics decreases. The power of thermodynamic analysis for drug
design lies in the combination of 3D structural information and the study of
structurally related compounds. This can give detailed insight on how specic
structural features contribute to binding energetics. The most important thermo-
dynamic parameters in molecular structural events are:
DH binding enthalpy represents the heat eects involved in the interaction. It can be
directly experimentally determined with calorimetric measurements. The heat eects
are caused by the formation and disruption of non-covalent bonds (hydrogen and
ionic bonds and van der Waals interactions) and can involve bonds between the
reactants, but also bonds of solvent reorganization and conformational rearrange-
ments of the reactants during the binding process. A large part of DH is due to bulk
hydration. In drug design, more water molecules at the interaction interface may
extend the complementarity of the surfaces and H-bond networks [9]. This is
favorable for enthalpy, but disadvantageous due to a loss in entropy, and contributes
to the phenomenon of entropyenthalpy compensation (see text).
DS binding entropy can in general be interpreted in terms of degree of order and
disorder of the system. This might comprise designed restriction of conformational
freedom and rotation of chemical bonds involved in binding. Also, hydration can be
a major factor for entropy, e.g. in hydrophobic binding: the burial of water-accessible
surfaces and resulting release of water molecules can contribute to binding due to
increases in entropy.
DC
p
heat capacity is almost entirely ascribed to solvent eects and is considered of
high information content. DC
p
can be determined directly in calorimetric experiments
over a temperature range. DC
p
can be interpreted in terms of solvent-accessible
hydrophobic and polar surface areas, buried in the binding process [5]. A decrease in
accessible hydrophobic surface upon binding has a negative eect on DC
p
; that of
polar surface a positive eect. Experimental DC
p
values have been related to dierent
degrees of success to 3D structural information on these surface changes upon
binding. Problems arise due to solvation eects and release or ordering of water
molecules during binding. Ordering of water molecules in the binding interface has a
large negative contribution to DC
p
[4].
Figure 5.20 Temperature eect on SPR sensorgrams. Left: signals of sample cell
(solid line) and reference cell (broken line). Right: net signal (reference
cell minus sample cell). Reprinted from ref. [46], Copyright (2003), with
160 Chapter 5
5.6.2 Comparison of SPR Thermodynamics with Calorimetry
The heat eects of an interaction can be directly measured using calorimetry.
Especially the introduction of isothermal titration microcalorimetry (ITC)
instruments with improved sensitivity has greatly advanced the use of calori-
metry in biomolecular interactions [9]. A debate is ongoing on the equivalence
of enthalpy values from vant Ho analysis (DH1
vH
), compared to those from
calorimetry (DH1
cal
): discrepancies between DH1
vH
, from several techniques for
anity assay and DH1
cal
have frequently been observed [4750]. In calorimetric
assays the total of the heat eects is assayed, e.g. heat of dilution, of mixing and
heat eects due to changes in buer protonation and solvent equilibria linked
with the binding process [47], which go beyond the intrinsic enthalpy contribu-
tion of the simple equilibrium A+B"AB. On the other hand, linked
equilibria like that of buers and solvent will also inuence the anity and
Figure 5.21 vant Ho plot for binding of hmT-peptide with v-Src SH2 domain. (J)
Anity at the sensor surface, K
C
; (K) in solution, K
S
. The lines are
calculated using eq. (5.26). Reprinted with permission from ref. [12],
Copyright (2005) American Chemical Society.
Table 5.5 Thermodynamic parameters derived from vant Ho thermo-
dynamic analysis shown in Figure 5.21.
Parameter v-Src SH2 with 11-mer hmT-peptide
DH1 (kcal mol
1
)
a
9.4 0.6
TDS1 (kcal mol
1
)
a
0.3 0.6
DC
p
(cal mol
1
K
1
) 920 160
DG1 (kcal mol
1
)
a
9.1
K
D
(nmol l
1
) 220 30
a
At reference temperature 25 1C.
DH1
vH
. The situation may even become more complicated when DC
p
is not
constant with temperature which can occur in multi-step binding processes.
Horn et al. demonstrated that, when experimental setup and analysis are
performed correctly, there is no statistically signicant dierence between
DH1 values [51]. This holds even for complicated binding models including a
conformational equilibrium as shown in Box 5.2.
It should be remarked that vant Ho analysis is peculiar in its error
estimation. Recently, Tellinghuisen demonstrated that the usual way of error
estimation in vant Ho analysis is actually not correct and that the errors in
DG1, DH1, DS1 and DC
p
are a function of temperature, leading to relatively
large errors in DC
p
[52]. The number of thermodynamic studies using SPR is
rather limited [12,14,50,5355]. In our experience, the thermodynamic para-
meters from SPR vant Ho analysis often compare fairly well with those from
ITC [12,14]. As an example, in Figure 5.22 we show the match between our
SPR data and ITC data from various studies in an entropyenthalpy compen-
sation (EEC) plot for a wide range of ligands for the Lck and v-Src SH2
domains [46].
EEC is a universal phenomenon in biomolecular interactions in water and is
generally a problem for the medicinal chemist as a gain in enthalpy, e.g. by
adding hydrogen bonds to strengthen the binding, will be counteracted by a
loss in entropy [56].
Figure 5.22 Entropyenthalpy plot for binding of various ligands to the Src- or
Lck-SH2 domains. Open symbols: data derived from various ITC
studies. Closed circles: data derived from SPR competition experiments,
with peptide and peptoid ligands. Reprinted from ref. [46], Copyright
(2003), with permission from Elsevier.
162 Chapter 5
Combining SPR and calorimetry to explore fully the thermodynamics and
kinetics of interacting systems might provide an optimal approach. To explore
the information content of especially DC
p
values fully, ITC is generally a better
choice than SPR thermodynamic analysis. A disadvantage of ITC is that it
requires much more material (at least 100 times as much as is needed for SPR).
The strong point of the SPR technique is the anity data and kinetics derived
from the same data set.
5.6.3 Transition State Analysis Using Eyring Plots
SPR analysis has the unique feature that kinetic and anity information can be
obtained from one experiment. This implies that thermodynamic experiments
can be performed by analyzing the temperature dependence of k
on
and k
o
.
Eyrings transition state theory provides the fundamental conceptual frame-
work for understanding rates of chemical processes [57]. The transition state
(AB
#
) is the high energy state along the pathway of reactants to product (for a
binding process, the unbound species to the complex).
AB
k
1
k
1
AB
#
!
k
2
AB 5:27
Based on eq. (5.27), the thermodynamic equilibrium constant for formation
of the transition state is dened as K
#
[AB
#
]/[A][B]. Applying statistical
mechanics, we obtain the Eyring equation state that holds for a rate constant k:
k
k
B
T
h
K
#
5:28
where k
B
is Boltzmanns constant (1.381 10
23
J K
1
) and h is Planks
constant (6.626 10
34
J s). K
#
is related to DH
#
and DS
#
(the activation
enthalpy and entropy, respectively) in the same way as K
A
is related to DH1
and DS1 [eqs. (5.22), (5.23) and (5.26)]. This implies that for a linear Eyring plot
[ln(kh/k
B
T) vs. 1/T] the data can be tted with eq. (5.29).
In
kh
k
B
T
_ _

DH
#
R
1
T

DS
#
R
5:29
A non-linear Eyring plot can be tted with eq. (5.26); such ts yield the
activation parameters DH
#
, DC
#
p
and DS
#
.
Transition state analysis using Eyring plots derived from SPR data have been
published elsewhere [12,14,53,54]; an example is given in Figure 5.23. The k
o
values at various temperatures were determined in the presence of high
concentrations of competing ligand to prevent rebinding (Section 5.3.2.2).
The k
on
values were derived from k
on
k
o
/K
D
. It appears that the Eyring
plot for k
o
is linear, indicating that DC
#
p
is zero between the complex and the
transition state (vice versa). The plot for k
on
shows a convex curvature,
indicating that DC
#
p
for formation of the transition state from the reactants is
not zero. DC
#
p
has the same absolute value as that derived for DC
p
from the
vant Ho analysis of K
A
for the same interaction (Figure 5.21), because going
from the complex to the transition state (k
o
) DC
#
p
is zero. The Eyring plot for
k
on
(Figure 5.23a) is interesting as it displays non-Arrhenius kinetics above
20 1C, i.e. at higher temperature k
on
decreases. Non-Arrhenius kinetics have
been frequently found for protein folding. A general explanation for this
phenomenon is that at higher temperatures a wide region of conformational
space is visited and the probability of a exible ligand or part of a protein
having the proper conformation for binding or folding, decreases [58]. Such a
model makes sense for the binding of a pYEEI-peptide to an SH2 domain, as
the high-anity binding can be regarded as a two-pronged plug into a two-
holed socket in need of suitable positioning of the pY and I residue for binding
[59]. If, for instance, binding starts with the pY moiety in its binding pocket, at
higher temperatures it will be more dicult to have the I residue in the correct
position to allow high-anity binding.
A transition state analysis can give additional information, as is also
illustrated by the comparison of binding phosphotyrosine ligands to the v-Src
SH2 domain vs. Grb2 protein. The anity and DG1 values are comparable, even
the activation energies DG
#
are nearly identical (Figure 5.24). However, transi-
tion state analysis reveals large dierences in DH
#
and DS
#
. Both DH
#
values are
negative, indicating that upon formation of the transition state, heat is released.
The high barrier of activation energy is caused by unfavorable activation
entropy contribution. For binding to v-Src SH2 this contribution is about
4 kcal mol
1
more unfavorable. This means that formation of the transition
state of the Src SH2 domain from the reactants involves a higher degree of
ordering than that of Grb2 SH2. Upon binding, the dynamics of Grb2 and v-Src
SH2 domains is decreased to the same extent, leaving the large dierence in DS
#
unexplained [12]. The dierence in thermodynamic behavior can probably be
Figure 5.23 Eyring plots for the interaction of v-Src SH2 domain with hmT
11-mer phosphopeptide. (a) k
on
, tted with eq. (5.26); (b) k
o
, tted
with equation (5.29). Reprinted with permission from ref. [12],
Copyright (2005) American Chemical Society.
164 Chapter 5
attributed to the role of water molecules, which form a hydrogen bonding
network at the binding interface between ligand and v-Src SH2 protein [7] upon
formation of the transition state, which has an entropy price. On the other hand,
the water molecules in the network make a favorable enthalpy contribution to
the transition state, explaining the favorable DH
#
. For binding to Grb2 SH2,
such a role of water molecules is not inferred, only direct contact exists between
the ligand and the protein. The above described transition state analysis has
been criticized, because it is assumed that every activation leads to complex
formation [60]. Alternatively, Arrhenius analysis is advocated [60].
In the interpretation of the data, 3D information from X-ray and NMR
analysis is essential. However, the 3D structures alone cannot provide informa-
tion on energetic contributions determining the binding process. Especially in
cases where dynamics of ligand and receptor or solvent eects are involved,
results of computational chemistry can be expected to be disappointing. The
contribution of entropy to the free binding energy can be very large and may
inuence the anity by several orders of magnitude. Thermodynamic and
kinetic analysis can help to quantify the extent of these contributions and to
generate ideas to exploit them in molecular design.
5.7 SPR Applications in Pharma Research: Concluding
Remarks and Future Perspectives
In this chapter, we have emphasized the role of kinetics and thermodynamics in
biomolecular interactions. Notwithstanding the impressive contributions of
structural biology and computational chemistry, our understanding and the
ability to predict anities of receptorligand interactions remain poor. It is
Figure 5.24 Energy transitions at 25 1C as a function of the binding coordinate for
phosphorylated peptide binding to v-Src SH2 domain (solid lines) and to
Grb2 SH2 domain (dotted lines). Reprinted with permission from
ref. [12], Copyright (2005) American Chemical Society.
increasingly acknowledged that for a more accurate notion, thermodynamic
and kinetic studies of biomolecular interactions are needed. Modern calori-
metric and SPR techniques are the tools to perform such studies and deserve a
place in the toolbox of rational design used by the medicinal chemist and
chemical biologist.
The high information content of SPR data with kinetic and anity informa-
tion is unique and allows full thermodynamic and kinetic characterization of an
interaction, including transition state analysis, as shown in Section 5.6.3. There
are also limitations to the use of vant Ho analysis: even with anity data with
relatively small standard errors, DC
p
has a large error. DC
p
is important as it
can give insight into the nature of the surface area buried upon binding and on
the role of water molecules in the binding process (Box 5.5). Using ITC, in
general DC
p
can be more accurately determined by measuring DH at a wide
range of temperatures.
The best of both worlds is to use SPR and ITC for thermodynamic analysis,
with special emphasis on careful experimental design and on the limitations of
each method. A concern associated with SPR data is that anity for a ligand
immobilized on a sensor surface might not be identical with that for the ligand
in solution. In this respect it is relevant to introduce linkers between the binding
epitope and the dextran matrix of the sensor chip, as described in this chapter.
In many cases no dierence appears, especially for monovalent binding.
Using the approach for assays of K
S
with competition experiments as outlined
in Box 5.3, we nd comparable anities at the surface and in solution (see
Table 5.4 and Figure 5.21). On the other hand, binding to a surface might
be a better model for a biological interaction involving multivalency than
(monovalent) interactions in solution.
Apart from drugreceptor studies, SPR is also useful for other aspects of drug
research. Studies on binding to serum proteins are relevant for distribution
properties of drugs [61]. Many important drug targets are membrane-bound
proteins, e.g. G-protein coupled receptors (GPCRs). Technology to follow passive
and active absorption to membrane interfaces using SPR is under development,
as is drug binding to metabolizing enzymes [62]. SPR biosensor systems with
supported monolayers and tethered bilayer membranes are under development,
but not standard technology yet [63]. In an approach called ligand shing,
crude tissue extracts and cell homogenates are screened for potential ligands or
targets using SPR [62]. In such approaches, identication of bound species is
crucial, with mass spectrometry (MS) as the ideal platform. In particular, matrix-
assisted laser desorption/ionization time-of-ight MS (MALDI-TOFMS) and
electrospray ionization MS (ESI-MS) are powerful tools for protein identication.
It is therefore not surprising that it has been attempted to integrate SPR and MS
for proteome analysis, as reviewed in ref. [64]. SPR serves two main purposes in
proteome analysis: (1) to conrm and possibly quantify specic binding and (2) to
act as a micro purication support for further analysis. MALDI analysis directly
on a chip surface is possible [65]. Problems may arise due to the small amount of
captured protein on the chip, the many handling steps of the procedure and the
166 Chapter 5
acidity of the matrix material. In another approach, analytes are eluted and
collected in a recovery system. In principle then the whole range of MS
techniques, including analysis of digested samples, is available for identication.
In drug research, high-throughput screening (HTS) plays an essential role in
screening large libraries of compounds. Until recently, the use of SPR technol-
ogy was hampered by the limited number of surfaces on one sensor chip. In
view of the high commercial potential, in the slipstream of the development of
array technologies, SPR imaging applications are emerging using microarrays
on chips. Examples include Biacores Flexchip microarray device [66] and IBIS
Technologies IBIS-iSPR instrument (see Chapter 6) and other SPR imaging
techniques [67,68] (Genoptics and K-MAC, respectively). This eld is devel-
oping rapidly and is extremely promising. We expect an increasing impact of
SPR technology on drug research. This will be enhanced by further develop-
ments of SPR technology for pharma applications, such as high-throughput
screening, further integration of SPR and MS and mimicking membrane
environments and protein ensembles on SPR surfaces.
5.8 Questions
1. What causes mass transport limitation (MTL) in the kinetics of SPR
experiments?
2. How can MTL be diminished by experimental design?
3. How can one perform kinetic analysis under MTL conditions?
4. How can depletion of the analyte in a cuvette-based system be calculated?
5. What is the strength of SPR as a tool in drug development research?
6. Explain why a high loading of the ligand aects the determination of the
anity constant. Describe at least two ways to solve this and determine
from the sensorgram given below the anity constant of an antibody
antigen reaction in nmol l
1
.
5.9 Symbols
[A] concentration of analyte A at the sensor surface (mol l
1
)
[A]
0
initially added analyte concentration in the bulk (mol l
1
)
[A]
free
analyte concentration corrected for depletion (mol l
1
)
[A]
tot
total analyte concentration used in a competition experiment (mol l
1
)
[B] concentration of free analyte binding sites on the sensor surface
(mol m
2
or m1)
A molecule A (usually analyte)
AB
#
transition state on formation of the AB complex
B molecule B (usually ligand)
B
max
maximum binding capacity (in m1)
D diusion coecient (m
2
s
1
)
DC
p
heat capacity (J mol
1
K
1
)
DC
p
#
activation heat capacity (J mol
1
K
1
)
DG Gibbs free energy (of binding) under non-standard conditions
(J mol
1
)
DG
#
Gibbs activation free energy for formation of the transition state
(J mol
1
)
DG
0
Gibbs free energy (of binding) under standard conditions (J mol
1
)
DH
#
activation enthalpy (J mol
1
)
DH
0
change of enthalpy (upon binding) under standard conditions
(J mol
1
)
DS
#
activation entropy (J mol
1
K
1
)
DS
0
change of entropy (upon binding) under standard conditions
(J mol
1
K
1
)
h Plancks constant (6.6262 10
34
J s)
K
#
equilibrium association constant for formation of the transition state
(l mol
1
)
K
A
equilibrium association constant (l mol
1
)
k
b
Boltzmans constant (1.381 10
23
J K
1
)
K
C
equilibrium dissociation constant for anity at the chip (mol l
1
)
k
conf
rate constant for conformation change AB -AB* (s
1
)
K
conf
equilibrium constant for conformation change (k
conf
/k
-conf
)
k
-conf
rate constant for conformation change AB* -AB (s
1
)
K
D
equilibrium dissociation constant (mol l
1
)
k
obs
k
on
[A] +k
o
(s
1
)
k
o
dissociation rate constant (s
1
)
k
on
association rate constants for formation of the complex AB
(l mol
1
s
1
)
K
S
equilibrium dissociation constant in solution (mol l
1
)
k
tr
transport rate for diusion from bulk to sensor surface (in
m1 s
1
l mol
1
)
L
m
transport coecient from bulk solution to sensor surface (ms
1
)
L
r
Onsager coecient for reaction ux (ms
1
)
MW molecular weight of analyte (Da)
168 Chapter 5
R gas constant (8.3143 J K
1
mol
1
)
R
eq
net increase in SPR angle at binding equilibrium (reference cell
subtracted from sample cell (m1)
R
t
net increase in SPR angle at time t (reference cell subtracted from
sample cell (m1)
S surface of the sensor in the cells (mm
2
)
T absolute temperature (K)
T
0
reference temperature (298 K)
V
bulk
volume of the bulk solution in the cells (l)
Z viscosity (cP)
The examples presented in this chapter originate from work in the Department
of Medicinal Chemistry and Chemical Biology, Utrecht Institute of Pharma-
ceutical Sciences, Utrecht University, The Netherlands. We wish to thank the
Head of the Department, Professor Rob M.J. Liskamp, for making it possible
to study exciting new compounds in our SPR assays. Many members of
the Department have contributed: we especially thank Dr. Frank J. Dekker,
Dr. Rob Ruijtenbeek and Dr. Ir. John A.W. Kruijtzer. We are grateful for the
cooperation with Dr Isabelle Broutin (Universite Rene Descartes, Paris) and
Professor Albert Heck (Biomolecular Mass Spectrometry, Utrecht University).
We thank The Netherlands Organization for Scientic Research (NWO) for
nancial support.
References
1. G. Schreiber, Curr. Opin. Struct. Biol., 2002, 12, 41.
2. S. Valitutti and A. Lanzavecchia, Immunol. Today, 1997, 18, 299.
3. A.E. Lenferink, E.J. van Zoelen, M.J. van Vugt, S. Grothe, W.
van Rotterdam, M.L. van De Poll and M.D. OConnor-McCourt, J. Biol.
Chem., 2000, 275, 26748.
4. G.A. Holdgate, A. Tunniclie, W.H. Ward, S.A. Weston, G. Rosenbrock,
P.T. Barth, I.W. Taylor, R.A. Pauptit and D. Timms, Biochemistry, 1997,
36, 9663.
5. K.P. Murphy and E. Freire, Pharm. Biotechnol., 1995, 7, 219.
6. H.J. Bo hm and G. Klebe, Angew. Chem. Int. Ed. Engl., 1996, 35, 2588.
7. D.A. Henriques and J.E. Ladbury, Arch. Biochem. Biophys., 2001, 390,
158.
8. J.E. Ladbury, Chem. Biol., 1996, 3, 973.
9. R. Perozzo, G. Folkers and L. Scapozza, J. Recept. Signal Transduct. Res.,
2004, 24, 1.
10. G.A. Holdgate and W.H. Ward, Drug Discov. Today, 2005, 10, 1543.
11. N.J. de Mol, E. Plomp, M.J. Fischer and R. Ruijtenbeek, Anal. Biochem.,
2000, 279, 61.
12. N.J. de Mol, F.J. Dekker, I. Broutin, M.J. Fischer and R.M. Liskamp,
J. Med. Chem., 2005, 48, 753.
13. K. Zierler, Trends Biochem. Sci., 1989, 14, 314.
14. N.J. de Mol, M.I. Catalina, F.J. Dekker, M.J. Fischer, A.J. Heck and
R.M. Liskamp, Chembiochem, 2005, 6, 2261.
15. P.R. Edwards, C.H. Maule, R.J. Leatherbarrow and D.J. Winzor, Anal.
Biochem., 1998, 263, 1.
16. D.J. OShannessy, M. Brigham-Burke, K.K. Soneson, P. Hensley and
I. Brooks, Anal. Biochem., 1993, 212, 457.
18. D.R. Hall and D.J. Winzor, Anal. Biochem., 1997, 244, 152.
19. T.A. Morton and D.G. Myszka, Methods Enzymol., 1998, 295, 268.
20. P. Schuck and A.P. Minton, Trends Biochem. Sci., 1996, 21, 458.
21. L.L. Christensen, Anal. Biochem., 1997, 249, 153.
Chem., 1997, 64, 127.
23. P. Schuck, Biophys. J., 1996, 70, 1230.
24. S. Sjolander and C. Urbaniczky, Anal. Chem., 1991, 63, 2338.
25. S. Felder, M. Zhou, P. Hu, J. Urena, A. Ullrich, M. Chaudhuri, M. White,
S.E. Shoelson and J. Schlessinger, Mol. Cell. Biol., 1993, 13, 1449.
26. J.R. Faeder, W.S. Hlavacek, I. Reischl, M.L. Blinov, H. Metzger,
A. Redondo, C. Wofsy and B. Goldstein, J. Immunol., 2003, 170, 3769.
27. N.M. Green, Biochem. J., 1963, 89, 585.
28. J. Rao, J. Lahiri, R.M. Weis and G.M. Whitesides, J. Am. Chem. Soc.,
2000, 122, 2698.
29. D.G. Myszka, X. He, M. Dembo, T.A. Morton and B. Goldstein, Biophys.
J., 1998, 75, 583.
30. D.G. Myszka, P.R. Arulanantham, T. Sana, Z. Wu, T.A. Morton and
T.L. Ciardelli, Protein Sci., 1996, 5, 2468.
31. C.S. Navara, Curr. Pharm. Des., 2004, 10, 1739.
32. K. Futterer, J. Wong, R.A. Grucza, A.C. Chan and G. Waksman, J. Mol.
Biol., 1998, 281, 523.
33. R.A. Grucza, J.M. Bradshaw, V. Mitaxov and G. Waksman, Biochemistry,
2000, 39, 10072.
34. S. Yuzawa, M. Yokochi, H. Hatanaka, K. Ogura, M. Kataoka,
K. Miura, V. Mandiyan, J. Schlessinger and F. Inagaki, J. Mol. Biol.,
2001, 306, 527.
35. D. Cussac, M. Vidal, C. Leprince, W.Q. Liu, F. Cornille, G. Tiraboschi,
B.P. Roques and C. Garbay, FASEB J., 1999, 13, 31.
36. B.E. Hall, S.S. Yang and D. Bar-Sagi, Front. Biosci., 2002, 7, d288.
37. S. Maignan, J.P. Guilloteau, N. Fromage, B. Arnoux, J. Becquart and
A. Ducruix, Science, 1995, 268, 291.
38. N. Schiering, E. Casale, P. Caccia, P. Giordano and C. Battistini,
Biochemistry, 2000, 39, 13376.
39. M.B. Khalifa, L. Choulier, H. Lortat-Jacob, D. Altschuh and T. Vernet,
Anal. Biochem., 2001, 293, 194.
170 Chapter 5
40. M.J. Cannon, A.D. Williams, R. Wetzel and D.G. Myszka, Anal.
Biochem., 2004, 328, 67.
41. J.E. Ladbury, M.A. Lemmon, M. Zhou, J. Green, M.C. Boteld and
J. Schlessinger, Proc. Natl. Acad. Sci. USA, 1995, 92, 3199.
42. M.M. Morelock, R.H. Ingraham, R. Betageri and S. Jakes, J. Med. Chem.,
1995, 38, 1309.
43. N.J. de Mol, M.B. Gillies and M.J. Fischer, Bioorg. Med. Chem., 2002, 10,
1477.
44. P. Nioche, W.Q. Liu, I. Broutin, F. Charbonnier, M.T. Latreille,
M. Vidal, B. Roques, C. Garbay and A. Ducruix, J. Mol. Biol., 2002,
315, 1167.
45. M.J. Fischer, C. Kuipers, R.P. Hofkes, L.J. Hofmeyer, E.E. Moret and
N.J. de Mol, Biochim. Biophys. Acta, 2001, 1568, 205.
46. F.J. Dekker, N.J. de Mol, P. Bultinck, J. Kemmink, H.W. Hilbers and
R.M. Liskamp, Bioorg. Med. Chem., 2003, 11, 941.
47. H. Naghibi, A. Tamura and J.M. Sturtevant, Proc. Natl. Acad. Sci. USA,
1995, 92, 5597.
48. P.D. Ross and M.V. Rekharsky, Biophys. J., 1996, 71, 2144.
49. J. Thomson, Y. Liu, J.M. Sturtevant and F.A. Quiocho, Biophys. Chem.,
1998, 70, 101.
50. G. Zeder-Lutz, E. Zuber, J. Witz and M.H. Van Regenmortel, Anal.
Biochem., 1997, 246, 123.
51. J.R. Horn, D. Russell, E.A. Lewis and K.P. Murphy, Biochemistry, 2001,
40, 1774.
52. J. Tellinghuisen, Biophys. Chem., 2006, 120, 114.
53. A. Baerga-Ortiz, S. Bergqvist, J.G. Mandell and E.A. Komives, Protein
Sci., 2004, 13, 166.
54. Y.S. Day, C.L. Baird, R.L. Rich and D.G. Myszka, Protein Sci., 2002, 11,
1017.
55. J. Deinum, L. Gustavsson, E. Gyzander, M. Kullman-Magnusson,
A. Edstrom and R. Karlsson, Anal. Biochem., 2002, 300, 152.
56. J.D. Dunitz, Chem. Biol., 1995, 2, 709.
57. G.A. Peterson, Theor. Chem. Acc., 2000, 103, 190.
58. D. Truhlar and A. Kohen, Proc. Natl. Acad. Sci. USA, 2001, 98, 848.
59. J.M. Bradshaw, R.A. Grucza, J.E. Ladbury and G. Waksman, Biochem-
istry, 1998, 37, 9083.
60. D.J. Winzor and C.M. Jackson, J. Mol. Recognit., 2006, 19, 389.
61. A. Frostell-Karlsson, A. Remaeus, H. Roos, K. Andersson, P. Borg,
M. Hamalainen and R. Karlsson, J. Med. Chem., 2000, 43, 1986.
62. M.A. Cooper, Nat. Rev. Drug Discov., 2002, 1, 515.
63. M.A. Cooper, J. Mol. Recognit., 2004, 17, 286.
64. C. Williams and T.A. Addona, Trends Biotechnol., 2000, 18, 45.
65. R.W. Nelson, J.R. Krone and O. Jansson, Anal. Chem., 1997, 69,
4363.
66. D. Wassaf, G. Kuang, K. Kopacz, Q.L. Wu, Q. Nguyen, M. Toews,
J. Cosic, J. Jacques, S. Wiltshire, J. Lambert, C.C. Pazmany, S. Hogan,
R.C. Ladner, A.E. Nixon and D.J. Sexton, Anal. Biochem., 2006
351, 241.
67. J.B. Fiche, A. Buhot, R. Calemczuk and T. Livache, Biophys. J., 2007, 92,
935.
68. S.O. Jung, H.S. Ro, B.H. Kho, Y.B. Shin, M.G. Kim and B.H. Chung,
Proteomics, 2005, 5, 4427.
172 Chapter 5
CHAPTER 6
Surface Chemistry in SPR
Technology
ERK T. GEDIG
XanTec Bioanalytics GmbH, Mendelstrasse 7, D-48149 Mu nster, Germany
6.1 Introduction
Whereas the previous chapters focused mainly on the physical basics of surface
plasmon resonance, instrumentation and assay design, we will now turn to the
heart of SPR and related biosensors: the sensor chip and its surface chemistry.
It is here where the biomolecular interaction takes place and although dimen-
sionally extremely small (the coating thickness is measured in nanometers), the
sensor chip surface has a tremendous inuence on the performance of a
biosensor and on the quality of the data retrieved.
The nanocomposite coating on a biosensor chip usually consists of the
following elements, as depicted schematically in Figure 6.1:
The substrate is a mechanic carrier (e.g. glass) covered with a thin metal
layer (e.g. gold; beige layer in Figure 6.1), which enables surface plasmons
to be excited. The substrate material and geometry should be compatible
with the instrument optics.
A passivation or adhesion linking layer (grayblue in Figure 6.1) that links
the gold surface with the immobilization matrix.
The immobilization matrix, which is a critical element as it is in direct
contact with the ligand and the sample and thus inuences the specicity
and other key characteristics of the biosensor (green brushes in Figure 6.1).
The immobilized ligand (blue Y-shapes in Figure 6.1), usually a biomole-
cule, is linked to the immobilization matrix and should interact selectively
with the analyte (white squares).
This chapter focuses on the treatment of the substrate, the adhesion linking
layer and the immobilization matrix, including chemistries to couple the ligands
173
to the matrix. The Introduction is followed by a description of the adhesion
linking layer and the immobilization matrix. A thorough treatment is given of
the immobilization protocols of the ligand to the immobilization matrix, one of
which is shown in Figure 6.2. Further, structural features of chip surfaces for
different applications are covered and, nally, an overview is provided that
should be helpful in selecting the optimal surface for a given experiment.
6.1.1 General Aspects of Surfaces for Biomolecular Interaction
Analysis
Recognition processes between biomolecules are the key to a thorough under-
standing of almost all processes in living organisms. To characterize biomole-
cular recognition, direct optical biosensors are excellent tools as they allow for
fast and quantitative analysis without the need for labels. Although direct
detection is elegant, label-free detection bears the inherent disadvantage that
not only the desired specic components of a biomolecular interaction con-
tribute to the sensor signal, but also non-specic binding (NSB) of other matrix
components. This is a major difference in comparison with methods employing
labels which detect the labeled component only and are insensitive to other,
Figure 6.1 Nanoarchitecture of a typical SPR sensor chip: the gold-coated (beige)
transparent substrate is covered by an adhesion linking layer (grayblue)
to which the immobilization matrix is grafted (here a brush-structured
hydrogel in green). Ligands (blue Y-shapes) are covalently coupled to the
hydrogel chains to bind analyte molecules specifically (white squares).
174 Chapter 6
unlabeled species. Although NSB can be referenced out by the use of a second
channel, the data quality is much better when the ratio of specic vs. non-
specic interactions is high.
The second important distinction from traditional methods such as blotting
or other solid-phase techniques is the multiple reuse of the chips. Due to
significant variations between derivatized chip surfaces and also for cost and
time reasons, chip surfaces are reused as long as possible but at least for a series
of continuous measurements. Complete regeneration requires that after each
measurement cycle the chip surface is brought back to its original state.
Proteins and many other biomolecules tend to adsorb irreversibly on untreated
metal, glass or plastic surfaces, often losing 90% or more of their activity. The
result would be a chip surface with a low density of active ligand, which can
only be partially regenerated and is practically unusable after a few cycles, as
shown in Figure 6.3.
The time-resolved nature of direct optical techniques leads to additional
requirements of the chip surface. In contrast to common solid-phase assays
yielding one data point per assay, a biosensor delivers the full time course of the
interaction. As outlined in Chapters 4 and 5, this high information content can be
used to determine either the analyte concentration or usually with a differently
structured surface the association and dissociation kinetics of the interaction.
To understand the underlying processes that occur on the surface of a sensor
chip, it is helpful to take a closer look at the basic forces which determine the
interaction of biomolecules at the molecular level, listed in Table 6.1. Two of the
interactions, the electrostatic (ionic) and the hydrophobic interaction, account
for approximately 85% of the overall energy and are therefore the most
relevant. If the chip surface and the interacting species are charged, the extent
and sign of the electrostatic interaction can be manipulated within a relatively
Time
Phases: baseline - EDC/NHS - baseline - ligand - baseline - ethanolamine - baseline
Sensorgram of ligand immobilization
SPR-dip
shift
R
ligand
Figure 6.2 Sensorgram of capturing ligand immobilization to a dextran hydrogel sensor
surface by EDC/NHS chemistry (see Protocol 6.3), a commonly used
immobilization method. The double-headed arrow reects the resulting
amount of immobilized ligand (R
ligand
) in the deactivated hydrogel.
175 Surface Chemistry in SPR Technology
broad range through modications of the ionic strength and pH of the buer.
An increasing salt concentration screens charged groups and usually has a
practically pH-independent repulsion effect on hydrophilic and charged immo-
bilization matrices because ion pairs are formed which can neutralize the
charged domains. At low ionic strengths,
1
the pH of the buer becomes more
important, as the electrostatic interactions dominating in this regime are
governed by the ligands overall charge, which is positive at pH values below
B
A
n
D
t
Figure 6.3 Unstable immobilization and partial denaturation of ligand (red) and
analyte (blue) on an incompatible surface (A) results in incomplete
regeneration, yielding in an increasing baseline and decreasing binding
capacity from cycle to cycle (B).
Table 6.1 Molecular forces contributing to biomolecular interaction processes
[1,2].
Force Energy (kJ mol
1
) Distance dependence
Hydrophobic interaction Up to 15 Not applicable
Electrostatic interaction Up to 12.5 r
2
Hydrogen bonding Up to 4 r
6
Van der Waals forces Up to 0.4 r
10
1
I.e. below 0.1 M.
176 Chapter 6
the ligands isoelectric point and negative in the more alkaline pH range.
Depending on the charge of the sensor chip surface, strong attractive or repulsive
forces between a dissolved species and the surface can be the consequence.
The hydrophobic interaction cannot be easily controlled, as it is not an
attractive force but the exclusion of hydrophobic domains of low surface
energy in highly energetic solvents such as water. Hydrophobic interactions
are induced by a changed degree of organization of water molecules and are
therefore an entropy effect. The entropy of the total system (water and
interacting molecules) will increase when a hydrophobic interaction takes place.
Generally, substances which strengthen the inner structure of water, such as
most salts, increase the hydrophobic contribution (the salting-out effect is
caused by a change in entropy); chaotropic substances such as guanidine or
ethanol lower it. The inuence of the pH is moderate, although a pH close to
the pI of the interacting species usually minimizes the extent of electrostatic
forces and thus increases the relative inuence of hydrophobic interactions. As
a consequence, precipitation of protein molecules often occurs at the pI caused
by the decreased electrostatic repulsion between molecules. As hydrophobic
interactions can lead to partial unfolding and as a consequence to significant
activity losses of immobilized proteins, the fraction of hydrophobic domains on
the sensor chip surface should with the exception of surfaces for immobili-
zation of membrane proteins be kept as low as possible.
Keeping the above in mind, it is obvious that especially the surface charge
and surface energy of biosensor chips have to be carefully controlled in order to
achieve high immobilization yields, to minimize non-specic interaction of the
matrix, to retain the biological activity of the immobilized ligand and to achieve
high signal-to-noise ratios. Other means are the charge distribution and den-
sity, surface structure and functionality, which are key factors to adapting the
sensor chip surface to particular applications.
6.1.2 Selection of the Optimal Surface
Surface plasmons at the interface between a metal and a dielectric material have
a combined electromagnetic wave and surface charge character, as shown in
Chapter 2. This combined character results in the electric eld component
perpendicular to the surface being enhanced near the surface and decaying
exponentially with distance (Figure 6.4). The eld in this perpendicular direc-
tion is said to be evanescent, reecting the bound, non-radiative nature of
surface plasmons. In the sample buer above the metal, the decay length of the
eld, d
d
, is of the order of half of the excitation wavelength of light and is
usually dened as the distance over which the intensity of the evanescent eld
drops to 1/e, i.e. to about 37%. In most commercial instruments the wavelength
of light is between 600 and 800 nm and d
d
is in the range 300400 nm.
One important consequence of the exponential decay of the evanescent eld
intensity is that a typical SPR biosensor is practically blind at distances beyond
600 nm from its surface. The high signal-to-noise ratios achievable with
evanescent eld sensors are partially due to this insensitivity towards changes in
the bulk phase. In addition, identical interactions give rise to different signal
shifts, depending on how close to the sensor surface the interaction occurs. A
receptorligand binding (but also non-specic interaction processes) observed
within the rst 10 nm from the metal surface result in an almost three times
higher response than the same process at a distance of 300 nm.
The transport of solutes to the chip surface by convection and diffusion has a
profound effect on the signal, in addition to solute binding to the immobilized
ligand. The product of these three processes results in a sensorgram which is
consequently inuenced by factors contributing to any of those processes. In
linear ow cells used in many commercial instruments, convection can be
controlled simply by adapting the ow rate. At a certain ow rate, a so-called
stagnant layer is formed on the sensor, as described in Chapter 4. High ow
rates induce small stagnant layers; however, a stagnant layer of less than 2 mm is
difcult to achieve. Cuvette geometries, such as those employing the free wall-
jet principle, allow variation of more parameters, as for example the distance of
the injector tip from the sensor surface.
2
In all of these cases, the analyte is more
or less efciently transported to a distance of a few micrometers from the sensor
chip surface, but still relatively far away from the evanescent eld and any
three-dimensional surface structures. Here, the unstirred diffusion layer begins,
through which transport is solely effected by diffusion,
3
as depicted schemati-
cally in Figure 6.5.
Figure 6.4 The intensity of the evanescent eld decays exponentially with increasing
distance from the metal layer. The SPR angle shift is proportional to the
change of effective refractive index close to the surface.
2
As described in Chapter 3 Section 3.3.2.
3
Mass transport limitation effects are clearly described in Chapter 5, Section 5.3.
178 Chapter 6
The diffusion rate through this layer depends largely on the diffusion constant
of the analyte molecules, which varies from 3 10
6
to 6 10
6
cm
2
s
1
for low
molecular weight substances to below 10
7
cm
2
s
1
for macromolecules with
molecular weights of several hundred kDa. This corresponds to an average
diffusion time through the unstirred layer from less than 1 s for small molecules
up to several seconds for high molecular weight compounds. The effective
concentration of larger molecules in the evanescent eld but still not yet bound
to the ligands at the sensor surface will then be typically lower than the
concentration in the bulk solution. Even at the maximal ow rates, mass
transport is often limiting with typical protein molecular weights above
10 kDa. Due to a higher diffusion constant, the relative binding rate of small
molecules is usually not affected by diffusion phenomena even at high ligand
densities.
It is usually not possible to enhance the diffusion rate significantly above a
certain upper limit because the maximal ow rate is limited by the volume of
the sample, i.e. by the volume of the injection loop. In cases where diffusion
limitation should be strictly avoided, as for example in kinetic analysis, the only
way to circumvent diffusion-related problems is to decrease the density of the
immobilized ligand and/or increase the analyte concentration. On the other
hand, if mass transport limited binding is desired, as for concentration deter-
mination, then apart from decreasing the ow and thus the diffusion rate,
higher immobilization densities of the ligand can be chosen. In both cases low
and high immobilization densities a correctly selected surface nanoarchitec-
ture can be extremely useful for controlling the amount of immobilized ligand.
Figure 6.5 Dimensions of the evanescent eld and the unstirred diffusion layer.
Analyte molecules are transported rst mostly by convection and then
solely by diffusion.
The nal step in the biomolecular interaction process is the (specic) capture
of the analyte by the immobilized ligand, which is governed by the kinetics of
the interaction and the properties of the nanoenvironment around the ligand
analyte pair.
With kinetic analysis, it is essential that the binding sites are neither sterically
hindered nor their afnity affected by the immobilization process, as both can
lead to heterogeneously distributed afnities [3,4]. As shown in Figure 6.6, rate
Figure 6.6 Distribution analysis (this method is briey discussed in Chapter 12) of the
complex experimental surface binding kinetics of myoglobin binding to
monoclonal antibody immobilized in the carboxymethyldextran matrix of
a Biacore CM5 sensor chip. The two-dimensional rate and afnity con-
stant distribution indicates heterogeneous binding sites with different
afnities. Reproduced with kind permission from Ref. [3].
180 Chapter 6
and equilibrium constants can be determined based on a novel distribution
analysis method. A detailed characterization of the distribution of binding
properties provides a useful tool for the optimization of surface immobilization
towards the efcient functionalization of biosensor surfaces with uniform high-
afnity binding sites and for studying immobilization processes and surface
properties.
Sufcient spacing between ligands is helpful for decreasing steric problems,
but the requirement of surface homogeneity at the nanoscale remains. There-
fore, directed immobilization through capture molecules can become necessary
to maintain the homogeneity of a binding site population which might other-
wise be compromised by the random distribution of the covalent bonds within
the immobilization matrix. Finally, the surface charge can also have an effect
on the interaction kinetics and on the extent of non-specic interactions, so this
feature should also be controlled.
Assays with the mere purpose of analyte quantication are generally less
critical, as here a certain degree of diffusion limitation is allowable if not
desired. In most cases, high immobilization densities are advantageous in order
to maximize the specic signal and to deplete the unstirred layer of analyte
during the initial phase of the interaction phase. Under these conditions,
analyte binding is heavily diffusion controlled and the then linear slope
of the binding curve is directly proportional to the analyte concentration in the
sample, a prerequisite for a precise quantitative assay.
6.2 Adhesion Linking Layers for Gold, Glass and
Plastics
As outlined in the previous section, it is necessary to protect the sensitive
biomolecular ligands from the usually incompatible chip substrate material.
Also, suitable functional groups for ligand immobilization have to be intro-
duced. This is mostly achieved by coating the substrate with a bioinert layer
that contains functionalizable groups, typically carboxylates. As one of the key
characteristics of these layers their hydrophilicity makes them water soluble,
they would be washed away without the use of an adhesion linking layer, which
promotes their adhesion to the substrate material. Therefore (as shown in
Figure 6.1), a typical biochip coating contains at least three functional com-
ponents: an adhesion linking layer, a bioinert matrix and ligands coupled to the
matrix. In this section, the adhesion linking layer is described.
Ideally, the adhesion linking layer provides a stable link between substrate
material and immobilization matrix and also shields the substrate from the
sample buer with a dense and homogeneous lm. With respect to the expo-
nentially decaying strength of the evanescent eld, thicknesses above 10 nm would
significantly decrease the sensitivity of the sensor and can lead to baseline drift
caused by swelling effects. On the other hand, thicknesses below 1 nm usually
result in unstable and inhomogeneous coatings, so preferably the thickness of the
adhesion linking layer is between 2 and 5nm. Finally, the refractive index
of the adhesion linking layer should be lower than that of the substrate
material. Depending on the surface-exposed material of the biochip substrate
typically gold or glass different routes are chosen to address the above
requirements.
6.2.1 Adhesion Linking Layers for Metal Surfaces
The cationic surfaces of many transition metals are soft electron pair acceptors
and exhibit a strong afnity towards soft electron pair donors such as thiols,
disulfides and thioethers. Due to the negative imaginary part of their electro-
magnetic wavefunction in combination with their chemical inertness, gold,
silver or platinum are suited for use in SPR and as electrochemical sensors.
Alkyl derivatives of above-mentioned functional groups with a chain length of
410 carbon atoms assemble spontaneously on such substrates and form self-
assembled monolayers (SAMs) with high packing densities [5]. Shorter thiols also
assemble, but the SAMs are not well dened and relatively unstable [6]. Mono-
functional mercaptoalkyls yield hydrophobic surfaces having contact angles
higher than 1001. Bifunctional derivatives form monolayers with dened chem-
istry which are useful intermediates for covalent coupling of ligands or for further
derivatization [7]. Typical examples for such compounds are 16-hydroxy-
hexadecane-1-thiol and the corresponding carboxylated compound 15-carboxy-
pentadecane-1-thiol. The adsorption of these long-chain thiols usually takes place
in 15 mM ethanolic solutions in 824h. Although the formation of a monolayer
is almost complete after a few minutes, the initially formed monolayer is not well
ordered and contains many gauche defects within the chains. Over time, the layer
becomes more ordered and well packed. In addition to thiols, dithiols and
thioethers are also suitable, as all of these groups exhibit sufciently high
adsorption energy on surfaces of Group 712 metals, which is typically in the
range 4050 kJ mol
1
B50% of the strength of a CC bond. As discussed below
in more detail, the stabilizing characteristics of the resulting surface, i.e. its
inertness against adsorption of proteins and other sample components, depends
critically on the functionality of the surface exposed to the sample. However, the
long-term stability of SAMs is limited as they show desorption after a few weeks
of exposure to buer or serum [8].
A less common approach for the derivatization of noble metal surfaces
is the adsorption of positively charged or mercapto-derivatized polymers to
the negatively charged metal surface. Due to electrostatic attraction and
the cooperative effect of several adsorption sites, stable monolayers can be
formed.
6.2.2 Adhesion Linking Layers for Inorganic Dielectrics
The classical modication route for glass, ceramics and other oxidic surfaces is
treatment with silanes which are able to form stable silyl ether links with
182 Chapter 6
exposed hydroxyl groups [9]. A frequently used and versatile silane is GPTMS
(glycidylpropyltrimethoxysilane), which is reactive towards amino, sulfhydryl
and hydroxyl groups but can alternatively be hydrolyzed to yield a vicinal
diol or can be further oxidized to an aldehyde or carboxylate. Another example
is APTES (aminopropyltriethoxysilane), which is useful for coupling activated
carboxyls, aldehydes or glycidyl moieties. Under optimal conditions the silanes
assemble on the surface of a thoroughly cleaned [10] substrate in a uniform
monolayer [11]. If properly prepared, such an arrangement allows the attach-
ment of the bioinert matrix or directly of the biomolecular ligand in a similarly
uniform fashion.
6.2.3 Adhesion Linking Layers for Plastics
Although at present gold and glass dominate as chip surface materials used for
direct optical biosensors, it is foreseeable that future low-cost devices will
increasingly rely on injection molded consumables made from, e.g., poly(methyl
methacrylate) (PMMA), polystyrene, polycarbonate or cycloolen copolymers
(COC). The surface of these materials is a complex, heterogeneous mix of
amorphous and crystalline regions consisting of mostly hydrophobic polymer
chains which often slowly migrate and rearrange over time. Modication of this
kind of substrate usually begins with an oxidative pretreatment either via a wet
etch step [12] or more reproducibly oxygen plasma treatment [13]. The
immobilization matrix can then be coupled either directly or via subsequently
adsorbed stabilizing polymer layers [14].
Regardless of the substrate material, adhesion linking layers can be further
applied via plasma deposition, allowing fast and simultaneous processing of
large batch volumes, and can yield homogeneous coatings with different
chemical functionalities at relatively low cost per unit. Typical thick lm
preparation methods such as dip or spin coating are less suited as the necessary
coating thickness of a few nanometers is difcult to control reproducibly with
these techniques.
6.3 Bioinert Matrices
While the aforementioned adhesion mediators are optimized in terms of
maximal interaction with the substrate material, it is this characteristic that
induces a significant level of non-specic binding from the sample. Therefore,
addition of a bioinert topcoat, or immobilization matrix, becomes necessary.
6.3.1 Non-specic Adsorption of Biomolecules
Non-specic adsorption of heterogeneous macromolecules such as proteins or
larger aggregates on surfaces is a central aspect in the design of biocompatible
materials. However, the complex multistep interaction process depends on the
complex composition and concentration of potential adhering components in the
bulk and at the surface, which is still not fully understood. For example, the
initial adsorption of plasma proteins is dominated by the small protein (albumin),
present at higher concentrations in the bulk, to be replaced later by larger
proteins such as immunoglobulin G (IgG) and brinogen. This sequential
adsorption is called the Vroman sequence [15], which explains the competitive
adsorption of plasma proteins for a limited number of surface sites. Reorgani-
zation and stapling of biomolecules occur, which depend on several factors
and surface parameters. Dynamically, adsorbing species undergo three phases
[16]: transport to the surface by convection and diffusion (see Figure 6.5),
reversible (labile) attachment on rst contact, spreading and conformational
rearrangement.
Whereas the transport rate is an effective means of controlling the extent of
protein adsorption, the initial phase cannot be inuenced by the surface
functionality and structure. It is possible, however, to tune the degree of
protein attachment and subsequent processes by optimizing the nanoarchitec-
ture and molecular-scale functional design of the surface. In this context, it
should be stressed that the interaction of surface and proteins with the solvent
water plays an major role as the surrounding water molecules compete with
potential attachment sites.
The high internal energy of water (cohesion) favors the exclusion of low-energy
sites, i.e. hydrophobic domains which consequently tend to agglomerate the
already mentioned hydrophobic interaction which is an entropy phenomenon
caused by the organization grade of water molecules. This driving force leads to
conformational changes after adsorption of proteins to hydrophobic moieties.
Proteins attached to poorly hydrated or attracting surfaces regardless of
whether non-specifically attached or immobilized undergo high deformation
into a pancake-like structure [17]. This deformation process is typically irrevers-
ible and deactivates the protein. If the adsorbed proteins retained some lateral
mobility, clustering can occur [18] as a secondary process, which highlights the
difculty of models based on simple Langmuir isotherms. An essential prereq-
uisite of a bioinert matrix is therefore not just hydrophilicity but also a high
degree of hydration. In general, the more hydrophilic the surface, the greater is
the surfacewater interaction and the higher degree of water molecule organiza-
tion and hence less chaotic behavior. Hydrophobic molecules will merely irre-
versibly adsorb on the hydrophilic surface, e.g. by unfolding hydrophobic sites
that favor the increase in entropy of the total system upon binding. This primary
interaction is the likely basis for two distinct types of response seen towards
hydrophobic vs. hydrophilic surfaces [19]. Moreover, at a hydrophobic surface,
mutual water molecule hydrogen bonding is disrupted and significant surface
dewetting occurs, extending into bulk water.
In addition to hydrophilicity, the surface charge is an important factor. It is
obvious that proteins will be attracted by oppositely charged structures,
although at medium and high ionic buer strengths and physiological pH
ionic groups are usually surrounded by counter-ions, so if the charge density of
the surface is not too high, this effect is often less significant than expected. In
this context, one should take into account that the hydration sphere of ionized
184 Chapter 6
groups binds water molecules and therefore contributes to the bioinertness of
the surface. If not accompanied by irreversible secondary processes, such as
hydrophobic adsorption or covalent coupling, the electrostatic interaction is
typically fully reversible at high ionic strengths. If the underlying mechanism of
a protein adsorption process is unknown, increasing the ionic strength is a
useful measure to discriminate between ionic and hydrophobic interactions as
the latter become stronger and are irreversible under these conditions.
In addition to its reversible character, electrosorption is a relatively well-
dened process and can be controlled by pH-induced charge shifts of adsorb-
ates with a dened pI. Furthermore, it is self-terminated when electrostatic
neutrality is reached. However, it should be noted that even proteins bear-
ing the same charge as a strongly charged surface can show significant
adsorption overriding the net Coulombic force in low ionic strength solu-
tion [20]. Oppositely charged domains on the protein interacting with the
homogeneously charged surface are likely to be a factor, driving out small
counter-ions, an entropy-driven phenomenon referred to as counter-ion
evaporation [21].
In accordance with the above, a molecular-level structureproperty relation-
ship survey of differently functionalized surfaces [22] revealed that most
surfaces that resist the adsorption of proteins incorporate groups that exhibit
several characteristics, including hydrophilicity and neutral overall charge.
Another factor is the presence of hydrogen bond donors and the absence
of hydrogen bond acceptors, indicating the role of hydrogen bonding as a
third force contributing significantly to biomolecular interactions. However,
carbohydrate-based surfaces, many of which are highly protein resistant, do
not follow this rule, indicating that not all aspects of surfaceprotein interac-
tions are covered by this model. Again, water may be the key to this observa-
tion, as uncharged carbohydrates are claimed to orient up to three layers of
water [23].
6.3.2 Bioinert Hydrogels
In addition to choosing suitable functionalities at the molecular scale, the
bioinertness of a surface can be controlled by structural means. Polymers
grafted on the surface of a so-called hydrogel are effective in preventing the
adsorption of proteins. As the amount of protein adsorbed is the result of the
interplay between the bare surfaceprotein interactions, the competition
between the proteins and the polymer chains striving to be in the vicinity of
the surface, the polymerprotein interactions and the conformational statistics
of both protein and polymer molecules, it is obvious that variations in the
coatings nanoarchitecture result in different adsorption characteristics of the
coated surface. Surface stabilization by hydrogels is due to the decreasing
degrees of freedom of surface-bound polymer chains when a protein molecule
approaches the coated surface (Figure 6.7). As this leads to an energetically
unfavorable entropic loss, the system is more stable if dissolved macromolecular
species do not interfere with the polymer layer.
Theoretical studies using the mean-eld theory
4
[24] demonstrated that branched
and loop polymers are more effective in preventing protein adsorption than linear
exible chains [25], as they are more rigid and can provide a denser concentration
of desired functional groups (Figure 6.8). Dendritic polyglycerols, for example,
have shown good protein resistance and proven superior to dextran coatings used
to reduce non-specic binding [26]. However, other features important for evanes-
cent eld biochips, such as immobilization capacity, homogeneity at the nanoscale
and diffusion characteristics, should also be considered and these usually favor the
use of better dened linear structures as the immobilization matrix.
An additional feature of hydrogels for SPR-based biosensors is that the
hydrogel can be tuned to t the evanescent eld. Ligands can not only be
coupled directly to the surface but can also be immobilized in the evanescent
volume, making a higher loading of ligands feasible. However, if the hydrogel
becomes too thick and dense, ligands and analyte molecules may not penetrate
into the evanescent eld. The quality of the bioinert hydrogel layer in combi-
nation with the type of ligand which should be coupled determines to a great
extent the quality of the biomolecular interaction results. A list of common
protein-compatible polymers is given in Table 6.2; an in-depth evaluation of the
properties of three-dimensional hydrogels can be found in Section 6.4.2.
Figure 6.7 Entropic stabilization of surfaces: polymer segments in the vicinity of an
approaching macromolecule have fewer degrees of freedom than those
which can move freely in all directions. Compression of equal charges
along the polymer chains adds to the repulsive effect, preventing
non-specic binding.
4
A theory that describes the behavior of surface-grafted polymer chains. At moderate surface
concentrations (semi-dilute regime), the statistical mechanical properties of the chains can be
evaluated by using an analogy with the classical mechanics of a particle moving in a mean potential
eld. The analytical solutions of this theory provide a convenient way to explore the conforma-
tions and the interactions of such surfaces.
186 Chapter 6
In conclusion, bioinert matrices can practically quantitatively eliminate non-
specic binding even in complex samples such as serum or fermentation broths
and signal-to-background ratios can increase by one to two orders of magnitude.
6.4 Choosing the Optimal Nanoarchitecture
As outlined above and in Section 6.5, the molecular-scale functionalities of
sensor chip surfaces mediate the immobilization of ligands and are further
effective means of minimizing the non-specic background. However, their
Figure 6.8 Selected protein-compatible functionalities.
Table 6.2 Examples of hydrophilic, protein-compatible polymers.
Class Polymer
Polysaccharides, natural Dextran
Alginic acid
Hyaluronic acid
Heparin
Chitosan
Pectin
Polysaccharides, modied Carboxymethyldextran
Carboxymethylcellulose
Polyalcohols Poly(vinyl alcohol)
Poly(hydroxyethyl methacrylate)
Polyalcohol and polyether Polyglycerol (dendrimer)
Polyethers Poly(ethylene glycol)
Poly(propylene glycol)
Polycarboxylates Poly(acrylic acid)
Polyamines Poly(L-lysine)
proper selection alone is usually not sufcient to achieve reliable quality of the
sensorgram. Structural characteristics in the submicrometer range must also be
considered. In other words, the quality of the nanoarchitecture (e.g. hydrogel)
determines efcient coupling for accurate rate and equilibrium constant meas-
urements. A homogeneous distribution of binding sites and homogeneity of
their nanoenvironment are vital when it comes to kinetic analysis. Further,
depending on the application, different immobilization capacities are required
and sometimes extra features, such as a lter functionality, may be advanta-
geous. All these issues are addressed by the design of the nanostructured
coating or bioinert matrix.
Coatings for bioanalytical and biomedical devices are usually divided into
two major groups: two-dimensional (2D) planar coatings and three-dimen-
sional (3D) hydrogels. With SPR biosensor chips, both types are employed. In
Table 6.3, a practical overview is given of the application of 2D and 3D
nanostructures on sensor surfaces.
Table 6.3 Overview of nanostructures for different sensor applications.
Application Suggested structure
Proteinprotein, assay Hydrogel, 100500 nm, lowmedium
density. Linear polycarboxylate or
carboxylated polysaccharide
Proteinprotein, kinetics Well-stabilized carboxylated 2D surface
Hydrogel o50 nm, low density. Linear
polycarboxylate
ProteinDNA or polysaccharide Hydrogel, 100500 nm, lowmedium
density. Linear polycarboxylate
Proteinpeptide or small molecule Hydrogel, 4500 nm, high density. Linear
polycarboxylate or carboxylated
polysaccharide
Proteincell, virus or particle Thin hydrogel, o20 nm, high density.
Linear polycarboxylate or carboxylated
polysaccharide. Well-stabilized
carboxylated 2D surface
DNADNA, small molecule or peptide Hydrogel, 4300 nm, high density. Linear
polycarboxylate
DNAprotein Hydrogel, 100500 nm, lowmedium
density. Carboxylated and streptavidin-
modied polysaccharide
Cell, virus, particle, lipid bilayersany
species
Thin hydrogel, o20 nm, high density.
Linear polycarboxylate or carboxylated
polysaccharide, partially alkyl
derivatized. 2D mercaptoalkyl SAM
Assays in serum or culture medium Hydrogel, 100500 nm, lowmedium
density. Polysaccharide with reduced
carboxylation level
Ligand shing from crude samples,
assays in whole blood
Hydrogel, 300500 nm, lowmedium
density with lter layer
2D SPR microarrays Hydrogel, 4100 nm. NHS preactivated
linear polycarboxylate
188 Chapter 6
6.4.1 Two-dimensional Surfaces
Two-dimensional surfaces are well suited for the detection of large, particulate
analytes such as viruses or even whole cells, as a hydrogel would be inaccessible
for such bulky species, i.e. would keep them outside the evanescent eld. Due to
the same reasons and the higher eld intensity close to the chip surface, 2D
structures are also a good choice for the immobilization of high molecular
weight ligands, cell fragments or lipid mono- and bilayers.
2D surfaces (Figure 6.9) are often chosen for applications where a low
immobilization capacity is important, such as the kinetic analysis of medium to
high molecular weight compounds. As the density of immobilized ligand is
limited to a maximum of one monolayer (12 ng mm
2
for a typical protein),
depletion of analyte from the diffusion layer during the interaction phase can
hardly occur. Also, the diffusion of the analyte to and from the ligand is not
hindered by an extended immobilization matrix, both contributing to kinetic-
ally controlled binding. This free accessibility also minimizes rebinding of
dissociated analyte to free binding sites during the dissociation phase, which
can happen in extended hydrogels and results in the measured dissociation rate
constant (apparent k
o
) being slower than the true k
o
.
In kinetic studies, homogeneity of 2D coatings is essential, as a narrow
distribution of the immobilized ligands activity and accessibility is a prerequisite
for a correct t of the resulting sensorgrams with the appropriate kinetic model.
Figure 6.9 Schematic illustration of 2D surface architecture. Ligand molecules are
immobilized on a few nanometers thick, planar immobilization matrix and
are readily accessible to analyte molecules.
Here, problems can occur, as at the nanoscale level the untreated surface of a
sensor chip is not a well-dened, homogeneous, two-dimensional plane. In fact,
most materials, regardless of whether glasses, noble metals or plastics, show an
irregular nanotopology with a roughness well above 20 nm, which corresponds
to 410 times the diameter of the immobilized ligand.
The structure of gold surfaces used in SPR sensors depends on the deposition
method and, due to the higher deposition energy, is denser and more homo-
geneous if the gold is sputtered (Figure 6.10) compared with vapor-deposited
lms, which consist of loosely adhering gold clusters. More problematic are
molecular-scale inhomogeneities, although they are sometimes useful as they
increase the surface available for immobilization. These steps, gaps and tips
represent discontinuities which can lead to pinhole defects of coatings, thus
inducing local non-specic interactions or deactivation of sensitive immobilized
ligands (Figure 6.11). Non-specic binding sites are of particular relevance with
high molecular weight molecules and aggregates, as due to cooperative effects
even a low fraction of such attachment points can be sufcient to induce
irreversible adsorption of larger species.
For these reasons, the frequently employed SAMs of long-chain
mercaptoalkyls can be problematic and it is advantageous to coat them with
a second, smoothing ideally thin polymeric layer. The latter further
increases the stability of the coating and can provide a spacer functionality
which is useful for increasing the accessibility of the immobilized ligand.
An alternative and popular approach is the use of short polyether usually
PEG chains terminally linked to the SAM layer, which has a similar effect.
Figure 6.10 STM image of a sputtered gold surface. Crystalline 111 domains create
terrace-like nanostructures with roughness of around 30 nm. Scale bar
for STM in picoamps.
190 Chapter 6
6.4.2 Three-dimensional Hydrogels
Whereas with a sufciently sensitive detector 2D surfaces form a good basis for
many biomolecular interaction studies, numerous applications remain which
require a ligand density far above one monolayer, as either the analyte is
relatively small (o10 kDa) or the optics are not sufciently sensitive. For this
purpose, surface-grafted 3D hydrogels have been developed with thicknesses
from below 10 up to more than 1000 nm (see Figure 6.1). These structures
provide an increased density of attachment sites and what is more important
use a higher volume fraction of the evanescent eld for the specic interaction
than 2D coatings. A welcome side-effect of this approach is that the absolute
inuence of non-specic factors such as bulk shifts decreases proportionally
with increasing occupation of the evanescent eld volume by the ligandanalyte
pair. Further advantages of hydrogels are the improved stabilization against
non-specic interactions, their spacer functionality that keeps the immobilized
ligand away from the surface, thus improving its accessibility, and the protec-
tion of sensitive ligands against denaturation, especially when the surface is
being dried. Hydrogels can also help to solubilize hydrophobic ligands which
would otherwise tend to form insoluble aggregates on the chip surface.
In principle, surface-grafted hydrogels can be made from any water-soluble
polymer, but in practice polycarboxylates, polyethers and polyols are preferred
as they show significantly lower background than, for example, polyamines. As
in the early 1990s the rst commercial sensor chip coatings were derived from
well-characterized, dextran-based solid phases for afnity chromatography, the
most popular matrix material today is carboxymethylated dextran. Other
carboxylated polysaccharides such as alginate, pectin, carboxymethylcellulose
or hyaluronic acid can also be used, but are less common. Because these
polymers are of natural origin, their structure is often irregular, i.e. they can be
more or less branched or form hyperstructures, such as helices or suprabers,
thus contributing to heterogeneous binding site populations [3].
Especially in the hydrated state, carbohydrates are relatively bulky mole-
cules, occupying a considerable fraction of the evanescent eld volume and
rendering it unavailable for ligand immobilization. The bulky structure can also
Figure 6.11 Schematic illustration of possible defects of SAMs adsorbed on a
molecular scale nanorough gold surface. Such defects can cause local
non-specic interactions.
hinder free diffusion of the analyte. For these and other reasons, efforts have
been made to replace polysaccharides by better dened synthetic polycarboxy-
lates with a smaller molecular footprint, as illustrated in Figure 6.12. In
addition to these steric issues, pure polycarboxylate coatings have the advan-
tage that they cannot form ester crosslinks upon EDC/NHS activation, a
common side-reaction with carboxymethylated carbohydrates. After activa-
tion, the NHS polyesters are moderately hydrophobic, which is a prerequisite
for spotting ligand microarrays used for parallel detection with 2D SPR optics:
on a hydrophilic surface the spots run and merge. On the other hand, hydro-
phobicity can make a polycarboxylate hydrogel insoluble and collapse, i.e.
precipitate on the chip surface, if too activated, so optimization of the activation
level is important.
Regardless of whether a natural polysaccharide or synthetic polymer, two
main parameters can be varied to control immobilization capacity and diusion
characteristics of a 3D surface: the thickness and the density of a hydrogel. The
thickness depends on the molecular weight and structure of the polymer. The
thickness of surface-grafted polymer monolayers ranges from a few nm up to
2 mm with immobilization capacities between 1 and 4100 ng mm
2
. Typically,
due to the limited penetration depth of the evanescent eld, thicknesses between
20 and 200 nm are employed; thicker hydrogels can be useful to achieve ultra-
high immobilization densities for the detection of low molecular weight anal-
ytes. Hydrogels with a thickness above 1 mm are also useful to keep particulate
contaminations or air bubbles outside the evanescent eld, resulting in a very
robust surface. In such structures, heavy diffusion limitation is often observed
with a mean diffusion time of several seconds across the hydrogel.
Immobilization capacity can also be controlled by the hydrogel density. In
addition, variation of the chain density can increase the selectivity as it can
Figure 6.12 Structural differences between polysaccharide- (left) and synthetic poly-
carboxylate (right)-based surface-grafted hydrogels. See text for details.
192 Chapter 6
result in a ltering effect excluding macromolecules above a certain molecular
weight. For mere detection or analyte quantication purposes it is advisable to
choose relatively high densities as this not only maximizes the signal but also
minimizes non-specic effects. However, care must be taken that the hydrogel
density is not too high, as analytes with a molecular weight in excess of some
tens of kDa can agglomerate in the upper layer of dense hydrogel matrices, clog
the pores and prevent other analyte molecules from diffusing into the lower
parts of the sensing layer. If the analyte carries hydrophobic domains, agglom-
eration and in the worst case even (partial) collapse of the hydrogel can also
occur. Irregularly shaped binding curves are a typical indication for this
phenomenon. In addition, steric hindrance can result in the aforementioned
heterogeneous accessibility and afnities of the immobilized binding sites.
It should be stressed that the above is correct for medium and high molecular
weight analytes only. For low molecular weight compounds (below 2 kDa),
hydrogels with a maximum immobilization capacity should be chosen, as these
molecules are too small to be affected by high ligand densities and also diffuse
fast enough to avoid mass transport limitations. Signal maximization and thus
a maximum binding site density are the rst priority here, so the hydrogel of
choice for these applications should have a high thickness and medium to high
density, allowing a high loading of ligands on the hydrogel.
Composite structures, i.e. hydrogels with additional lter layers, can be
advantageous for samples containing particulate matter, such as blood or crude
fermentation broths. An inert, relatively thick hydrogel on top of the ligand-
derivatized layer excludes particles and cells which might be present in the
sample, as shown in Figure 6.13. These surfaces can also be used for assaying
because such a layer acts as a diffusion barrier, so that most of the interaction
takes place in a diffusion-controlled manner. The result is an extended linear
slope of the binding curve which is easy to t and facilitates the calculation of
analyte concentrations.
Charge should also be considered when selecting chip coatings for a partic-
ular application. Generally, the surface should be uncharged to minimize ionic
non-specic interactions. Although most biomolecules are negatively charged
and consequently repelled by the usually anionic chip surfaces, numerous
applications exist where at least partially positively charged species are present
which show a significant adsorption to polycarboxylate surfaces. Biomedical
samples (e.g. serum) are typical examples of such analyte matrices and espe-
cially here a low background is crucial as very often the analyte concentrations
are low. As a certain amount of COOH groups is usually required for electro-
static preconcentration of the ligand during the immobilization process and its
covalent attachment, the carboxyl density has to be optimized in these cases.
With dextran-based hydrogels it could be shown that a reduction of the
carboxymethylation level from one COOH group per anhydroglucose unit to
one group per 48 units effectively reduces ionic non-specic interactions with a
still sufcient density of attachment sites.
Situations may occur, however, where the charge density or the strength of
the carboxyl functions is not sufcient. An example is the immobilization of
compounds with a pI below 4. Within the pH range of the usually applied
preconcentration buers, these molecules are still negatively charged, rendering
their electrostatic accumulation in the hydrogel matrix difcult. Below pH 4, an
increasing percentage of the hydrogel-bound carboxylate groups is protonated,
i.e. uncharged, further enhancing this difculty. Introduction of strongly acidic
groups such as sulfonate groups into the matrix is helpful as they stay unpro-
tonated at sufciently low pH and can attract even relatively acidic species. A
way to introduce sulfonate groups is activation with sulfo-NHS instead of the
non-sulfonated compound.
6.5 Coupling Procedures for Ligand Immobilization
Coupling the ligand to the matrix on the sensor chip surface is a critical step
when pursuing biomolecular interaction studies. The immobilization strategy
should be chosen such that the immobilization level is sufcient and can be
controlled. Also, the activity and steric accessibility of the ligand has to be
preserved. In some cases, e.g. when the regeneration of the ligandanalyte pair
is impossible or results in unacceptable activity losses of the ligand, repeated
removal and re-immobilization of the ligand are advantageous, as this renders
fresh ligand for each analyte binding cycle. Generally, immobilization methods
can be divided into adsorptive, covalent, ionic and capture molecule-mediated
coupling.
Figure 6.13 Schematic view of a hydrogel with an additional lter layer. Only the
short polymer chains within the evanescent eld are carboxyl function-
alized and can immobilize the ligand. The larger polymers carry no
ligands. Their function is to prevent cells, cell debris and particulate
contamination from entering the sensitive volume close to the chip
surface.
194 Chapter 6
6.5.1 Adsorptive Immobilization
Adsorptive methods, typically the immobilization to low-energy surfaces such
as plastics via hydrophobic interactions, are the simplest and most popular
approach in solid-phase assays as this process occurs spontaneously upon
contact of a protein-containing buer with a hydrophobic surface [27,28]. Less
common is adsorption on noble metals, such as gold or silver, through strong
thiolmetal bonds, although the method is easy to perform and except for
thorough cleaning requires no prior surface preparation. Due to unavoidable
structural changes, usually a large part of the adsorbed ligand is denatured [29].
Typically, less than 10% of the immobilized binding sites remain active and
sterically accessible after adsorptive immobilization. Another disadvantage is the
poor stabilization of the resulting surface against non-specic interactions and
the impossibility of complete regeneration after analyte binding. Therefore, such
surfaces usually have to be blocked with suitable blocker proteins
5
and cannot be
regenerated. As stable adsorption requires cooperative forces of several surface-
afne functional groups, the adsorbate must have a sufcient number of such
residues and a molecular weight of at least 10 kDa. This method is generally not
suitable for non-derivatized nucleotides, peptides or small molecules.
6.5.2 Preconcentration Methods Prior to Covalent
Immobilization
Covalent methods are relatively straightforward, can give high coupling yields
and form stable covalent bonds between the ligand and a suitable biocompatible
sensor chip coating. Therefore, they are usually the rst choice when evaluating
different ways to immobilize biomolecules on a chip surface. A disadvantage is
the randomly oriented coupling which occurs equally at active and not active
sites of the ligand and hence can affect the afnity of at least a fraction of the
immobilizate. In extreme cases especially with small ligands, for example
peptides complete deactivation can even occur. Non-uniform coupling of the
ligand may lead to distribution of rate and equilibrium constants, which can be
determined with the distribution analysis method described in Chapter 12.
Regardless which of the covalent coupling methods described further below
is chosen, care must be taken that the ligand molecules approach close enough
to the reactive groups as otherwise no reaction can take place. This basic rule of
interaction before reaction is an important prerequisite for good coupling
yields, but it can be difcult to realize on well-stabilized surfaces which are
designed to suppress any interaction. This obvious dilemma can be addressed in
two ways, as follows.
6.5.2.1 Electrostatic Preconcentration
If the surface is usually negatively charged, electrostatic preconcentration
can take place by adjusting the pH of the coupling buer so that the net charge
5
Typical examples are BSA or casein; other ready-made cocktails are commercially available.
of the ligand is opposite to the surface charge. If the ionic strength of the
coupling buer is sufciently low, typically below 20 mM, and the pI of the
ligand is close to or optimally above the pI of the surface, the electrostatic
attraction between surface and ligand will override the hydrophilic or steric
stabilization. Under these conditions, the ligand accumulates at the surface
until electrostatic neutrality is reached the so-called preconcentration. Inter-
estingly, due to the aforementioned counter-ion evaporation effect, sometimes
electrosorption also occurs, when the overall charge of the ligand is neutral or
even the same as the surface.
Protocol 6.1 Electrostatic preconcentration
Preparation of the coupling buers. The pHof the buer should be at least 0.5opI
Protein
to ensure a positive net charge of the protein, which is required for electrostatic
interaction with the negatively charged surface. Also, a very low ionic strength is
essential. Prepare 5 mM buers from formic and acetic maleic acid carefully titrated
with 0.1 M NaOH and avoid overtitration as this will increase the salt concentration.
Buer ranges:
pH 3.04.0: sodium formate
pH 4.05.5: sodium acetate
pH 5.56.0: sodium maleate
These buers are subject to rapid microbial growth and after sterile ltration
should therefore be aliquoted, stored frozen until use and always used fresh. Do not
add preservatives such as sodium azide, as azide interferes with later activation steps.
Procedure to check the electrostatic preconcentration:
1. Mount a carboxylated sensor chip in the instrument. Make sure that the liquid
handling system is free from any protein contamination, since even minor
amounts of desorbed proteins will concentrate on the charged sensor surface.
2. Optional: to remove hydrophobically bound proteins from tubes, etc., clean the
ow system with an easily desorbable detergent solution followed by doubly
distilled water. If SDS is used, wash with a 10 min pulse 50 mM glycine HCl pH
8.5 afterwards.
3. Elute electrostatically adsorbed contaminants from the surface for 10 min with 2
M NaCl, 10 mM NaOH.
4. Check the baseline with coupling buer. After 1015 min, almost no drift should
be observed.
5. Inject protein solution to verify the preconcentration conditions. If no interaction
occurs, the ionic strength or pH should be lowered. Increasing the protein
concentration is optional but usually not helpful. Repeat the last three steps
until a sufcient preconcentration of the protein is observed on the sensor
surface. Note: some ligands come with significant salt contamination from
previous purication steps or as preservative (for example, ammonia salts, Tris,
sodium azide). These additives are often not stated on the product data sheet and
can interfere severely with the preconcentration process and also quench the
active groups. This results in a significantly reduced immobilization yield and
therefore the ligand should generally be microdialyzed into the coupling buer.
196 Chapter 6
6. Elute the protein from the surface for 25 min with elution buer.
7. Wash with water until the baseline is stable.
Procedure for coupling ligand using the electrostatic preconcentration effect:
8. Prepare the buer for covalent activation of choice and activate the surface. See
paragraphs below for details.
9. Wash as quickly as possible with coupling buer.
10. Inject the protein solution for 550 min. Since some of the carboxylate functions
have been converted into non-charged groups by the matrix activation proce-
dure, the preconcentration will usually be somewhat slower compared with the
unactivated surface. The immobilization yield can be increased by repeated
reinjection of the protein solution.
11. When working at low pH, inject water for 10 min to complete the coupling
reaction, which proceeds slowly under these conditions.
12. Quench remaining active groups as described below.
13. Optional: remove physisorbed proteins with regeneration buer.
14. Wash with coupling buer. The amount of covalently bound protein can now be
determined by comparison with step 4.
15. Switch to suitable running buer and start interaction experiments.
Note: At least two analysis cycles (see next chapter) are needed to equilibrate and
stabilize the coating for reliable and accurate measurements of the rate and afnity
constants.
The preconcentration effect is independent of the ligand concentration, it
works even at concentrations as low as a few mg ligand ml
1
until the coupling
buer is practically quantitatively depleted of ligand. As the electrostatic forces
work in a cooperative manner, the efciency of the preconcentration is pro-
portional to the molecular weight of the ligand. Achievable ligand densities
depend mainly on the surface nanoarchitecture, i.e. 2D or 3D coatings, and on
the thickness and density of the latter. With hydrogels as thick as 1000 nm, high
protein densities of well above 100 ng mm
2
(and 410
6
RU or 4101 SPR angle
shift) can be reached.
Apart from changing the ligand concentration and contact time, the ligand
density is controlled by adjusting the ionic strength of the coupling buer or
varying the activation level. The latter is the preferred variable as lowering the
density of activated groups also prevents multisite coupling, which can lead to
unwanted crosslinking and ligand deactivation. In this context it should be
noted that the preconcentration after activation is typically lower than that of an
underivatized surface. This is caused by partial conversion of charged groups
into neutral and insoluble moieties, which results in a decrease of the overall
charge density and also in a lower hydrogel hydration, which in turn can lead to
shrinking, and thus reduction of the available volume for ligand immobilization.
6.5.2.2 Dry Immobilization
This less frequently employed method is useful when it is not possible to
preconcentrate a ligand electrostatically on to the sensor chip surface. This is
the case with ligands having a pI below 4, such as DNA, with relatively small
and acidic compounds such as some peptides or if the surface is not charged. In
addition to bearing functionalities which are able to provide covalent coupling,
the ligand to be immobilized must be robust enough to survive the drying which
is necessary to bring it close to the reactive groups on the sensor surface.
Suitable sensor surfaces have to be preactivated oline (see Protocol 6.2).
Alternatively, NHS preactivated chips are commercially available.
Care should be taken that the surface functional groups present on the surface
cannot react with each other, as such self-quenching drastically reduces the
coupling yield. Commonly used carboxymethyldextran hydrogel surfaces, for
example, are not suitable for dry immobilization, as the NHS esters form esters
with abundant hydroxyl functionalities of the polysaccharide matrix, leading to
deactivation and unwanted crosslinking, a side-reaction that occurs although
to a lesser extent also during the standard wet activation process. The
advantages of dry immobilization are high coupling yields and the possibility
of using less reactive coupling chemistries, as for example epoxide activation,
which require harsher reaction conditions, such as elevated temperatures.
Protocol 6.2 Dry immobilization (o-line)
1. Place chips in a suitable receptacle and incubate the surface for 10 min. with
2 MNaCl, 10 mMNaOH.
2. Wash thoroughly with doubly distilled (dd) water.
3. Prepare the buer for covalent activation and activate the surface. See para-
graphs below for details.
4. Wash thoroughly with dd water and remove liquid traces with a quick centrifuge
spin or a sharp jet of clean compressed air/nitrogen. For reproducible drying the
direction of the jet of air/nitrogen is important, e.g. 451 to the surface. It is
essential that no droplets dry on the surface as even dd water leaves contami-
nation behind which might interfere with later processing steps.
5. Place chips in a dry and clean receptacle that is small enough to be placed in a
vacuum desiccator.
6. Spot 1030 ml of a solution of at least 0.3 mg ligand ml
1
low ionic strength
coupling buer manually on to the chip surface so that the sensing area is entirely
covered. Commercial spotters, preferably contactless spotters, may be used to
produce a microarray to be measured with an SPR imaging instrument.* Dry as
fast as possible in a vacuum desiccator but without desiccative to avoid potential
hydrolysis.
Note: some ligands come with significant salt contamination from previous
purication steps or as preservative (for example ammonia salts, Tris, sodium
azide). These additives are often not stated on the product data sheet and can
quench the active groups. As this results in a significantly reduced immobilization
yield, the ligand should generally be microdialysed into coupling buer before
spotting.
7. Incubate for 16 h depending on the activation employed. With less reactive
chemistry and robust ligands, additional heating to elevated temperatures up to
90 1C might be necessary.
198 Chapter 6
8. If the chips are not used immediately, put them in a freezer at this stage of the
process.
9. Dissolve unbound ligand over 1 h with running buer, wash with dd water and
dry as described above.
10. The chips are now ready for interaction analysis.
*Commercial spotters usually spot 0.11 nl and the spot diameter may vary
between 50 and 300 mm, which depends on the hydrophilicity of the surface. Be
aware that the ligand concentration in nanograms per square mm should be recal-
culated from the spot size, concentration of ligand in the spotting liquid and the
coupling yield which can vary between a few and 50%.
6.5.3 Covalent Activation Chemistries
Some common activation chemistries to covalently immobilize the ligand to the
chip surface are listed in the following sections. Most are based on carboxylic acid
residues as this functional group is present on many common immobilization
matrices.
6.5.3.1 Amine Coupling via Reactive Esters
Due to its exibility, relative ease of use, high coupling yields and robustness,
amine coupling via reactive esters is the most frequently employed immobili-
zation method. The reaction conditions for coupling proteins, peptides and
small molecules to carboxylated chip surfaces are well characterized and
extensive optimization studies have been performed [30]. Typically, the sur-
face-bound carboxyl groups are rst activated by a carbodiimide and converted
into active ester intermediates, which are then aminolyzed by lysines or the
N-terminal NH
2
group of the ligand.
For more than 50 years, carbodiimides have been used to mediate the forma-
tion of amide bonds between a carboxylate and a primary or secondary amine
[31,32]. Over time a number of different reagents has been developed, all forming
a reactive O-acylisourea intermediate with the COOH function, which is reactive
towards nucleophiles, such as primary and secondary amines, hydrazides, pri-
mary and some secondary alcohols, thiols and an important side-reaction
water. Whereas many carbodiimides, for example dicyclohexylcarbodiimide
(DCC), are water-insoluble and of limited use for bioconjugation, others are
readily soluble in water and therefore the reagents of choice for water-based
activation methods [33]. 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC)
is most frequently employed for conjugating biomolecules. Not only the reagent
itself but also the isourea reaction byproducts are water soluble, facilitating
activation and chip processing. EDC is usually supplied as the hydrochloride,
which is both hygroscopic and labile in the presence of water. It should be stored
in a dry container at 20 1C and allowed to warm to room temperature before
opening in order to prevent water condensation, which will cause decomposition
over time. Ideally, the EDC container should be opened under dry gas, such as
N
2
. For the same reason, EDC-containing activation buers should always be
prepared fresh and not from stock solutions. Note that inactive EDC is among
the most frequent reasons for failed immobilizations. EDC activation proceeds
effectively from pH 4.5 to 6.5 in 2-(N-morpholino)ethanesulfonic acid (MES) or
phosphate buer. Amine- or carboxylate-containing buers should be avoided as
they would react with the carbodiimide.
As the O-acylisourea intermediate hydrolyses within seconds, it would not
survive the time from activation until injection of the ligand. Therefore, it has
to be converted into a more stable active ester. This is achieved by adding the
water-soluble compound N-hydroxysuccinimide or its sulfonated form
N-hydroxysulfosuccinimide to the activation mix. In contrast to O-acylisourea,
the thus formed (sulfo-)NHS ester intermediate has a half-life from minutes at
pH 89 up to several hours at pH 45.
EDC/(sulfo)NHS coupling (Figure 6.14) is highly efcient but one should be
aware that side-reactions can occur, for example esterication with abundant
hydroxyl groups of polysaccharide hydrogel coatings. The resulting esters,
however, are subject to hydrolysis, as are also temporarily formed thioesters or
acylated histidine side-chain nitrogens [34]. The high efciency of the reaction
can result in multisite coupling and crosslinking if larger ligands with many
lysines are immobilized. Finally, the transesterication tendency of NHS esters
Figure 6.14 Reaction scheme of COOH activation with EDC/NHS and ligand
coupling.
200 Chapter 6
should be noted as it can lead to unwanted cross-activation of ligand carboxyl
groups or carboxylate buer components.
The aminolysis of (sulfo-)NHS esters occurs smoothly at slightly alkaline pH,
i.e. in carbonate or borate buers (pH 810) where most amines are deproton-
ated. This is therefore the reaction condition of choice for immobilization of
small ligands and many peptides. For the immobilization of larger species, such
as most proteins, it is advisable to create conditions where only a small fraction
of the lysine residues is reactive in order to prevent multisite coupling and
preserve the proteins activity. A low pH between 4 and 6 is therefore recom-
mended in these cases and also required in order to achieve the above-described
electrosorption effect. As an additional advantage, coupling the amine terminus
of proteins is favored over random immobilization through lysine residues at
lower pH values. The reason for this effect is the lower pI of N-terminal amines
compared with lysine NH
2
groups. Such optimization of the pH can result in a
significantly increased activity, especially of sensitive proteins.
In some cases, especially with labile ligands or when preparing high-capacity
surfaces for low molecular weight compound analysis, slight crosslinking of the
immobilized ligand by injecting a short pulse of low concentration EDC/NHS
over the surface-bound protein can result in additional stability of the immo-
bilized ligand. However, the applicability of such an additional crosslinking
step should be tested empirically.
Protocol 6.3 The EDCNHS coupling procedure
1. Prepare a solution of 0.5 M N-hydroxysuccinimide (NHS) in 0.5 M 2-(N-morpho-
lino)ethanesulfonic acid (MES) buer, pH 5.5. Optionally, 50 mM sulfo-NHS in
50 mM MES, pH 5.3 may be used.
2. Prime a COOH-modied sensor chip according to the procedure to check the
electrostatic preconcentration as described above. Elute electrostatically adsorbed
contaminants from the surface for 10 min with 2 M NaCl, 10 mM NaOH.
3. Check the baseline with coupling buer. After 1015 min, almost no drift should
be observed.
4. Make a 5100 mM solution of solid N-ethyl-N
0
-(dimethylaminopropyl)carbodii-
mide (EDC) in the NHS/MES buer from step 1. The optimal EDC concentra-
tion depends on the coating and the ligand. A low EDC concentration (around
10 mM) is recommended for linear polycarboxylate coatings and for the immo-
bilization of larger ligands with abundant lysine residues. Planar coatings and
CMD hydrogels can be activated with higher EDC concentrations, especially if
the ligand is small.
5. Inject this activation mix over the sensor chip for 120 min. Variation of the
incubation time allows control of the activation level. To achieve a lower activa-
tion level, which is recommended, for example, for polycarboxylate hydrogels, an
activation time o10 min should be chosen.
6. For the standard wet immobilization protocol, wash briey with water or slightly
acidic coupling buer and continue with injection of ligand. For dry immobili-
zation, wash with 5 mM acetic acid and dry according to Protocol 6.2.
6.5.3.2 Amine Coupling Through Reductive Amination
Carbonyl groups such as aldehydes, ketones and glyoxals can react with amines
to form labile Schi base intermediates, which can be reduced to yield stable
secondary amines (Figure 6.15). The Schi base formation is efcient at both
low and high pH and sodium borohydride and sodium cyanoborohydride are
the most commonly used reduction reagents. The difference is that
cyanoborohydride is about ve times milder [35] and in contrast to sodium
borohydride does not reduce free aldehyde groups, i.e. leads to higher
coupling yields by shifting the reaction equilibrium towards the product side.
Reductive amination of Schi bases is a highly selective reaction that
proceeds smoothly under mild conditions and unlike coupling through amide
Figure 6.15 Oxidation of a polysaccharide hydrogel and immobilization of an amine-
containing ligand through reductive amination.
202 Chapter 6
bonds preserves the positive charge of the reacted amine. It is therefore an
interesting alternative for the immobilization of sensitive ligands to polysaccha-
ride surfaces which, after periodate oxidation, display a sufcient density of
aldehyde groups in the form of cyclic hemiacetal structures [36]. Useful chip
substrates for this method are either coated with non-derivatized dextran or
carboxymethyldextran with a low degree of carboxymethylation. The latter has
the advantage that it allows electrosorptive preconcentration of protein, i.e.
working with lower ligand concentrations in the immobilization buer. In
addition to coupling via active esters and reductive amination, several other
amine reactive chemistries exist but are not further described here as their use
with evanescent eld-based biosensors is limited.
Protocol 6.4 Amine coupling through reductive amination
1. Prime the chip surface as described above (see points 2 and 3 in the EDC/NHS
coupling Protocol 6.3).
2. Inject 10 mM sodium periodate in water over 30 min. To control the density of
aldehyde groups, the concentration and contact time may be varied.
3. If working without applying the electrosorption effect, inject the ligand at a
concentration of at least 1 mg ml
1
in in phosphate, borate or carbonate buer
(pH 710) plus 10 mM sodium cyanoborohydride (Caution: extremely toxic!). pH
89 is optimal and other buers may be used provided that they do not contain
competing amines (such as Tris). If protein ligands are electrostatically precon-
centrated, a lower pH and lower ionic strength are required. As the reaction
proceeds slowly under these conditions, long contact times of several hours should
be chosen.
4. After completion of the coupling reaction, inject 0.5 M Tris or ethanolamine HCl
(pH 8.0) plus 10 mM sodium cyanoborohydride for 1 h. Alternatively, the excess
aldehydes may be reduced by injection of 50 mM sodium borohydride in 0.1 M
sodium carbonate (pH 9).
5. Wash with running buer of choice until the baseline is stable and start the
interaction analysis cycle.
6.5.3.3 Thiol Coupling
An alternative to coupling of proteins through their amine functionalities is the
use of thiol groups after reduction of pendant disulde bridges (Figure 6.16).
Choosing this immobilization strategy makes sense if amines are not present or a
sufcient number of disulde groups are available which are located far enough
Figure 6.16 Reduction of disulde-containing compounds.
from the proteins active site. In such cases a higher activity of the immobilizate
can be achieved compared with the more randomly oriented amine coupling
[37]. Immobilization of the reduced proteins can occur either by disulde
exchange using activated disulfides, by nucleophilic substitution of halo-
acetyls or by alkylation of maleimides. The rst method formation of disulf-
ides is reversible under reducing conditions, a relevant parameter for full
regeneration of chip surfaces [38]. The last two chemistries yield stable
bonds which cannot be cleaved again. Whereas the site-directed coupling can
result in a higher activity of the immobilizate, the additionally introduced
disulde, acetyl or maleimide residues can lead to slightly increased non-specic
interactions.
Prior to thiol coupling, the protein of interest has to be reduced. This is
usually done using thiol-containing compounds such as dithiothreitol (DTT),
2-mercaptoethanol, 2-mercaptoacetic acid or 2-mercaptoethylamine. For a
more convenient purication of the reaction mixture after the reduction step,
immobilized reducing agents may be used. The use of complexing additives is
generally recommended to prevent reoxidation that is catalyzed by traces of
heavy metal ions. If using alternative protocols it should be ensured that no
reducing reagents remain in the reduced protein solution.
Protocol 6.5 Protein disulde reduction procedure
1. Add 0.5 ml of thiolated agarose to a column and equilibrate with 0.1 M phosphate
buer pH 8.0.
2. Activate the column by adding 1 ml of 10 mM DTT in 0.1 M phosphate buer pH
8.0 containing 1 mM EDTA.
3. Wash the column with 20 column volumes of 0.1 M phosphate buer pH 8.0.
4. Add the protein to be reduced in 0.1 M phosphate buer pH 8.0 containing 1 mM
EDTA. Recover fractions and collect all samples. Read the absorbance at 280 nm
to determine which fractions contain the eluted protein. Pool these fractions,
aliquot and freeze at 20 1C until required.
During the thiol-disulde exchange, the free thiol of the reduced ligand reacts
with an activated disulde on the chip surface and forms a new mixed disulde
with release of a leaving group usually pyridyl disulde (Figure 6.17). The
latter is easily transformed into pyridine-2-thione, a non-reactive compound
not capable of participating in further mixed disulde formation. Thus only the
surface bound end of the mixed pyridyl disulde has potential for becoming
attached to the sulfhydryl-containing ligand.
This disulde exchange reaction occurs over a broad pH range and in a
variety of buers, including acidic, low ionic strength regimes used for electro-
static ligand preconcentration. As already mentioned, the resulting disulde
bond is cleaved under reducing conditions and is also affected by thiol buer
additives.
204 Chapter 6
Protocol 6.6 Thiol-disulde exchange procedure
1. Prime the disulde-derivatized chip (see points 2 and 3 in the EDC/NHS coupling
Protocol 3 as described above).
2. Inject 100 mM DTT or other reducing agents in 0.1 M phosphate buer (pH 8.0)
for 20 min.
3. Wash for 5 min with running buer.
4. Inject 10 mM pyridyl disulde in 0.1 M phosphate buer +20% ethanol for 20
min. (Note: dissolve pyridyl disulde in 100% ethanol before diluting to 10 mM
with phosphate buerethanol mixture).
5. Wash for 15 min with running buer.
6. Switch to water as running buer and wash the system for 10 min until the
baseline is stable.
7. Inject 10100 mg reduced protein ml
1
coupling buer plus 1 mM EDTA for 1020
min. Significant preconcentration should occur in this step.
8. Quench excess active disulfides with 1 mM mercaptoethanol in 0.1 M acetate
buer (pH 4.2), 1 M NaCl for 30 min (prepare fresh).
Thiol Coupling via Maleimides. For irreversible immobilization via thiol
groups, maleimide or haloacetyl coupling is the method of choice
(Figure 6.18). The alkene group of the maleimides undergoes an alkylation
reaction with the sulfhydryls and forms a stable thioether bond. The speci-
city of this coupling for sulfhydryl groups is high at neutral pH [39],
whereas at higher pH some cross-reactivity with amino groups may
occur [40]. At pH 7 the reaction of maleimides with sulfhydryls is about
Figure 6.17 Pyridyl disulde-mediated disulde exchange.
1000 times faster than its reaction with amines [41]. As a further side-
reaction especially at higher pH the maleimide group may undergo
hydrolysis to an open maleamic acid form, before but also after the reaction
with a sulfhydryl. In addition to the maleimide method described here,
sulfhydryl-containing ligands can also be immobilized through haloacetyl or
epoxy activation (see Section 6.5.3.5).
Protocol 6.7 Maleimide coupling of sulfhydryl groups
1. Activate a carboxylated surface with EDC/NHS as described in Protocol 6.3. Use
of a 10-fold diluted activation mixture should be considered as such a reduced
activation level is sufcient to obtain a good immobilization yield with most
proteins.
2. Convert the active NHS esters into amino groups by injection of 1 M ethylene-
diamine hydrochloride (pH 6.0) for 10 min.
3. Quench remaining reactive groups with 1 M ethanolamine hydrochloride (pH
8.5) for 15 min.
4. Wash for 10 min with 0.1 M HCl.
5. Inject 2050 mM heterobifunctional reagent, N-g-maleimidobutyryloxysuccini-
mide, in HBS for 10 min.
6. Wash for 5 min with dd water.
7. Inject the reduced protein (Protocol 6.5) in suitable coupling buer (see Protocol
6.1) for 10 min.
8. Wash for 5 min with HBS.
9. Hydrolyze excess unreacted maleimido groups by a 10 min exposure to 0.1 M
NaOH. Other, milder solutions may be used for inactivation but require a longer
contact time.
Alternatively, maleimide preactivated chips, which are commercially available,
may be used. In this case, start with step 6.
Note: as with most other protocols, at least two analysis cycles (see the next
chapter) are needed to equilibrate and stabilize the coating for reliable and accurate
measurements of the rate and afnity constants.
Figure 6.18 Maleimide coupling of sulfhydryl groups.
206 Chapter 6
6.5.3.4 Immobilization of Aldehydes Through Hydrazide
Groups
Aldehyde groups can be valuable reactive sites when immobilizing carbohydrates
or glycoproteins [42]. In the case of carbohydrates, often the anomeric aldehyde,
i.e. the product of the mutarotation, is used (Figure 6.19); they can also be created
by mild oxidation of sugar residues by sodium periodate. Alternatively, amino
groups can be converted with glutaraldehyde. As carbohydrate groups
are typically located at some distance from the ligands binding site, immobiliza-
tion through them might be considered a useful method to retain the activity
of immobilized glycoproteins. However, as immobilization using the EDC/
NHS method can also be performed under very mild conditions, the differ-
ence between the two methods is often smaller then expected; a comparative study
of antigen binding capacities of immobilized monoclonal antibodies showed
almost no difference between EDC/NHS- and hydrazide-mediated coupling [43].
Hydrazide-derivatized sensor chips can be generated in situ by EDC/NHS
activation of carboxylated surfaces followed by reaction with hydrazine or the
bifunctional and less toxic adipic acid hydrazide solution. The activation level
and thus the density of immobilized ligand can be controlled through concen-
tration of the EDC/NHS mixture and by the reaction time. To avoid cross-
linking of 3D surface structures, the concentration of the hydrazine and
dihydrazide should be chosen high enough. The coupling reaction itself pro-
ceeds best at slightly acidic pH and is therefore compatible with the conditions
chosen for electrostatic preconcentration. Small amounts of sulfate catalyze the
reaction. The hydrazone bond is relatively stable at neutral and alkaline pH but
somewhat labile in acidic buers. To avoid ligand leakage, it should be
stabilized by reduction with cyanoborohydride.
Due to their typically neutral to acidic nature, carbohydrates cannot
be immobilized using the electrostatic preconcentration effect (Protocol 6.1).
The method to follow is either injection at high concentrations (several tens
of mg ml
1
) or oline immobilization using the dry method according to
Protocol 6.2.
Figure 6.19 Hydrazide coupling of a carbohydrate through its anomeric aldehyde end
group.
Protocol 6.8 Glycoprotein oxidation
1. Dissolve the protein at a concentration of 310 mg ml
1
in 10 mm sodium phos-
phate, 0.1 M NaCl (pH 6.2).
2. From a freshly prepared sodium periodate stock solution add a sufcient volume
to reach a nal concentration of 510 mM.
3. Carry out reaction for 30 min in the dark.
4. Immediately purify the oxidized protein into slightly acidic immobilization buer
(see above) using a suitable column or spin column. Ensure that the puried
protein contains no periodate contamination.
5. Pool protein fractions, eventually dilute with immobilization buer to a con-
centration of 30300 mg ml
1
and inject immediately (point 5 in Protocol 6.9),
as due to intermolecular Schi base formation self-polymerization, deactiva-
tion and precipitation may occur over time. If solutions have to be stored, freeze
at 20 1C.
Protocol 6.9 Hydrazide activation of a carboxylated surface
and coupling of aldehyde-derivatized ligands
1. Prime the surface according to Protocol 3 as described above.
2. Activate for 3 min with 0.2 M EDC, 0.05 M NHS. For lower ligand densities
dilute this activation mixture.
3. React for 15 min with 0.1 M adipic acid hydrazide hydrochloride (pH 8.0).
4. Quench remaining NHS esters over 30 min with 1 M ethanolamine hydrochloride
(pH 8.5).
5. Inject oxidized ligand according to Protocol 6.8 and react for at least 20 min.
6. Stabilize hydrazone bonds by reduction with 50 mM sodium cyanoborohydride
(Caution: toxic! prepare in a fume hood) in 0.1 M acetate buer (pH 4.0) for
20 min.
7. Wash with running buer of choice until the baseline is stable and start interac-
tion analysis.
6.5.3.5 Coupling Through Epoxy Groups
Epoxy-mediated immobilization is a robust method with a long tradition,
especially for immobilization of carbohydrate ligands in afnity chromato-
graphy [44]. With optical biosensors this technique is relatively seldom
employed, as fully carboxymethylated dextran hydrogels are difcult to acti-
vate and the previously described alternative methods work under milder
reaction conditions. However, for coupling of carbohydrates, but also for
selective reaction with amine and sulfhydryl nucleophiles, epoxy coupling can
208 Chapter 6
be an interesting alternative (Figure 6.20). Note that crosslinking can occur as a
side-reaction.
Protocol 6.10 describes epoxy sensor chip activation and dry immobilization
of carbohydrates. It can be applied if the ligand is robust enough to survive
drying and heating at strongly alkaline pH.
Figure 6.20 Epoxy activation of a dextran matrix and coupling of a carbohydrate.
Protocol 6.10 Epoxy activation of chip surfaces and coupling of
carbohydrates
1. Prime dextran or partially carboxymethylated dextran-coated chip according to
Protocol 6.3 as described above. Fully carboxymethylated dextran is difcult to
epoxy activate and should not be used.
2. Activate for 15 min with 2% epichlorohydrin (Caution: toxic!) in 0.1 M potassium
hydroxide. For a lower activation level and a lower ligand density, the pH and
reaction time should be decreased.
3. Wash chip thoroughly with water and dry according to Protocol 6.2 as described
above.
4. Prepare 100 ml of a solution of 10100 mg ligand ml
1
10 mM NaOH and cover the
sensing area of the chip with a few ml of this solution so that it is totally covered by
the liquid.
5. Dry thoroughly.
6. Heat for 30 min at 7090 1C. Lowering the reaction temperature is another
possiblity to decrease the coupling yield.
7. Wash for 2 h with running buer.
8. Wash with water, dry as described above (Protocol 6.2) and install the chip to
start an analysis cycle.
In addition to using the above-described methods, ligands can be covalently
immobilized via radical substitution, photocoupling or other crosslinking
chemistries. Generally, most methods employed in bioconjugate synthesis or
for the preparation of supports for afnity chromatography are also suitable
for the derivatization of sensor chip coatings.
6.5.4 Electrostatic Methods
Electrostatic/ionic immobilization is a technique frequently used for the attach-
ment of oligonucleotides to polyamine-coated microarray substrates. As it is
non-selective and incurs considerable electrostatic forces towards the immobi-
lized ligand, the drawbacks of this method are similar to those of adsorptive
techniques (see Sections 6.3.1 and 6.5.1). It is therefore suitable for robust ligands
only. A specic feature of this technique is its potential reversibility by changing
the ionic strength and/or pH or through the addition of chelating additives.
However, this results in labile immobilization complexes, significant leakage of
ligand and, as a consequence, an unstable baseline. Therefore, this method is still
seldom employed for biosensors. An exemption is the use of immobilized
nitrilotriacetic acid (NTA) or iminodiacetic acid derivatives or similar chelating
groups, which allow reversible immobilization of His
6
-tagged molecules through
complex formation with transition metal ions, preferably Ni
21
[4547].
6
Such
complexes are relatively stable, although interestingly more sensitive to changes
in conditions than NTA afnity columns. For total regeneration of the chip
surface they can be cleaved by chelating agents such as EDTA.
6
Protocol 6.12.
210 Chapter 6
Protocol 6.11 Electrostatic adsorption of DNA on (poly)amine
coated sensorchips
1. Prime aminomodied chip surface for 10 min with 0.1 M HCl.
2. Wash for 2 min with degassed 5 mM sodium acetate (pH 5.0).
3. Inject 1100 mM DNA in degassed 5 mM sodium acetate (pH 5.0) for 5 min. A
baseline increase between 3000 and 6000 mRIU (about 20004000 RU or 200
300 mdeg) should be observed.
4. Equilibrate for at least 15 min with running buer as used for the interaction
experiment. Depending on the ionic strength and pH, a certain fraction of the
electrosorbed DNA is desorbed. 2 M NaCl plus 10 mM NaOH results in quan-
titative desorption.
5. For a more stable immobilization, activate the surface with epichlorohydrin as
described in Protocol 6.10 and use amino- or sulfhydryl-derivatized DNA.
Protocol 6.12 Ni
21
-mediated immobilization of His
6
-tagged
ligands [47]
1. Mount the sensor chip derivatized with NTA or another suitable chelator.
2. Prepare the immobilization buer: 10 mM HEPES, 0.15 M NaCl, 0.005% Tween
20 (pH 7.5). Higher NaCl concentrations and a higher pH lower the immobili-
zation yield; lowering the pH to 6.9 increases it.
3. Optional: microdialyze a 50200 nM solution of ligand into immobilization
buer.
4. Condition the surface for 5 min with 0.5 M Na EDTA (pH 8.5).
5. Wash for 2 min with immobilization buer.
6. Inject 0.3 M NiCl
2
or NiSO
4
for 2 min. Depending on the density of chelating
groups, a 2060 RU baseline increase should be observed.
7. Wash for 2 min with immobilization buer.
8. Inject 50200 nM ligand in immobilization buer for 25 min. Through pH and
ionic strength of the immobilization buer the amount of immobilized ligand can be
controlled via the ligand concentration and contact time and should be relatively low
as a certain number of unoccupied Ni complexes are required for stabilization, i.e.
continuous rebinding of the weakly bound (K
D
o10
6
) His
6
-tagged immobilizate.
The binding is improved at very low ow rates and at high pH.
9. Equilibrate with sample buer until the baseline is stable. HEPES was found to
give the best results but PBS and Tris also work. Low concentrations (maximum
50 mM) of EDTA in the sample buer stabilize the assay as it scavenges
contaminating metal ions.
10. Inject analyte in sample buer and analyze interaction.
11. Instead of regenerating the ligandanalyte interaction, the undissociated ligand
analyte pair can be quantitatively removed with a 3 min injection of 0.5 M Na
EDTA (pH 8.5). If the surface is of good quality, about 100 such regenerations
can be carried out without significantly affecting the surface capacity. This also
means that the same chip can be used for many different proteins.
6.5.5 Directed Immobilization
Oriented coupling with retention of the ligands activity and structure can be
achieved via indirect immobilization through site-specic capture molecules on
the chip surface. These can be proteins such as antibodies or protein A [48], but
also smaller, more robust moieties such as biotin [37,49]. While such indirect
methods seem to be less problematic and at least theoretically should render
a higher percentage of ligands still active after the immobilization process, one
should be aware that bulky (protein) linkers occupy a significant fraction of the
evanescent eld volume, which is then not available for ligand molecules and
can result in smaller signals and especially in the case of high molecular
weight analytes and hydrogel matrices might cause steric hindrance. Fur-
thermore, larger capture molecules can alter the afnity prole of the immo-
bilized ligand or induce non-specic interactions. An interesting feature of
regenerable afnity-based ligand surfaces such as protein A is that the ligand
can be removed together with the analyte during regeneration. This is advan-
tageous if it is difcult to break the ligandanalyte interaction.
When using biotinylated molecules, one should take into account that
biotinylation is usually carried out with NHS-activated biotin derivatives.
However, covalent coupling of unbiotinylated ligand using the EDC/NHS
method as the coupling chemistry is identical but usually gives higher immo-
bilization yields.
Protocol 6.13 Immobilization of biotinylated ligands on
streptavidin-modied surfaces
1. Prepare a 120 mg ml
1
solution of ligand in compatible buer. The choice of
buer is not critical as the biotinstreptavidin interaction is strong (K
D
c10
15
)
and takes place under a wide range of conditions. More important are minor
contaminations of free biotin often present with biotinylated biomolecules. As
biotin is usually much smaller then the biotinylated ligand, it diffuses faster to the
yet unoccupied binding sites and can drastically reduce the amount of bound
ligand. As a general precaution one should therefore thoroughly microdialyze (or
otherwise purify) the ligand into the corresponding buer.
2. Equilibrate a streptavidin-modied sensor chip with immobilization buer.
3. Inject ligand solution for 130 min. The immobilization level can be controlled
through the ligand concentration and contact time. Although the binding itself
proceeds rapidly, larger ligands generally require more time to reach the biotin
binding pockets than do smaller ones.
4. Run 23 test analysis cycles. Depending on the strength of the regeneration agent,
it might happen that a more or less significant fraction of the immobilized ligand
is desorbed again. This is due to a few low-afnity binding sites but is also caused
by dissociation of the tetrameric streptavidin into monomeric subunits. In this
case re-inject ligand solution over the chip and repeat analysis cycles until a stable
baseline is reached.
212 Chapter 6
Protocol 6.14 Immobilization of antibodies on protein A-
modied surfaces
1. Equilibrate a protein A-modied sensor chip with sample buer for 5 min until
the baseline is stable. Alternatively, protein A can be immobilized on a carboxy-
lated chip surface following Protocol 6.3.
2. Inject 550 mg of IgG in physiological sample buer for 3 min.
3. Wash for 2 min with sample buer.
4. Inject the analyte sample and record the sensorgram.
5. Optional: inject sample buer and monitor dissociation.
6. Strip the analyte together with ligand IgG with two 1 min pulses of 0.1 M HCl.
The protein A surface is now ready for immobilization of fresh IgG. Protein A is
robust and survives several hundred regeneration cycles.
6.5.6 Immobilization of Membrane Proteins
Although direct adsorptive immobilization of protein ligands is not optimal
for specic biomolecular interaction analysis (see Section 6.5.1), hydrophobic-
ally driven adsorption can be a valuable method for indirect immobilization
of membrane-bound receptor molecules. Usually, these proteins denature
when removed from the lipid bilayer membrane and if not stabilized with
detergents require a more or less intact membrane environment to display their
normal functionality. This can be achieved by either rst integrating t
he transmembrane species into vesicles and liposomes, which are then fused
on a hydrophobic surface or on an amphiphilic hydrogel [50,51], or by
rst immobilizing the membrane protein on a partially alkyl-derivatized
hydrogel sensor chip followed by on-chip reconstitution with a lipiddetergent
mixture [52]. The approach using a hydrogel layer has the advantages that
the lipid bilayer is supported by a hydrated structure, better resembling the
natural environment, and that the preparation process is more robust, i.e. not so
easily affected by impurities as is the case with strongly hydrophobic surfaces.
Protocol 6.15 Preparation of mixed micelles [50]
1. Prepare a lipid lm by pipetting a small volume of 10 mM lipid in chloroform
solution into chloroform-washed round-bottomed glass tubes and drying this
solution under a nitrogen stream (fume hood). Slight heating in a 30 1C water-
bath accelerates the drying process.
2. Evacuate for at least 2 h in a vacuum desiccator to remove traces of solvent.
3. Add 25 mM of the detergent octylglucoside (OG) in 9 mM HEPES (pH 6.4) and
135 mM NaCl to yield a 3.3 mM lipid solution. If other detergents are used, the
volume of the detergent solution should be adjusted such that the ([detergent]
cmc
detergent
)/[lipid] ratio is between 2 and 3.5. Note that the optimal ratio depends
on the lipid and the detergent and must be identied empirically from case to case.
Protocol 6.16 On surface reconstitution of lipid bilayers without
and with immobilized receptor protein [50]
Priming the surface:
1. Mount a partially alkyl-derivatized carboxylated hydrogel coated sensor chip.
2. Wash the surface for 5 min with 20 mM Chaps.
Optional immobilization of receptor protein:
3. Activate the surface for 7 min with 0.2M N-ethyl-N-dimethylaminopropylcarbodii-
mide and 50 mM N-hydroxysuccinimide.
4. Inject 10100 mg of receptor protein in suitable coupling buer (see Protocol 6.1)
plus 20 mM octyl glucoside for 1020 min.
5. Quench remaining active groups for 15 min with 1 M ethanolamine pH 8.5 plus
20 mM octylglucoside.
Formation of lipid bilayer and reconstitution of receptor protein:
6. Inject the mixed micelles (prepared as described in Protocol 6.15) for 2 min and
slow ow.
7. Wash with running buer and start interaction experiments.
Note: The surface can be completely regenerated by two 2 min injections of 20 mM
Chaps or 50 mM octyl glucoside.
Alternatively, liposomes or cell fragments can be modied with suitable tags
such as biotin, oligonucleotides or His
6
, which are then specifically bound
by the corresponding capture molecules immobilized on a hydrophilic surface.
The ligand (receptor protein) density can be controlled either via the lipid/
transmembrane protein ratio during the vesicle preparation or when using the
on-chip reconstitution protocol by varying the NHS activation level (Protocol
6.12, step 3).
In summary, there is no ideal, general-purpose immobilization strategy and
the optimal choice depends on the experimental approach, and the nature of
the ligand, analyte and sample buer. Generally, it is desirable that the
immobilization chemistry does not interfere with the activity and binding
characteristics of the ligand by linking to the sensor chip surface. In most
cases, covalent methods give good results. As a rule of thumb, it is advisable to
begin experiments with standard amino coupling chemistry, combined with
low- to medium-density hydrogels (see Chapter 6.4.2), and, if unsuccessful, to
try alternative, site-directed immobilization strategies.
As can be seen in Figure 6.21, the standard amino coupling (Protocol 6.3),
due to its robustness and good immobilization yields is the most popular
method, followed by the direct immobilization of biotinylated ligands. How-
ever, one should keep in mind that the popularity of different immobilization
methods is significantly predetermined by the commercial availability of suit-
able sensor chip surfaces. Table 6.4 might be helpful in identifying
214 Chapter 6
Figure 6.21 Relative use of different immobilization techniques in 2003 [53].
Table 6.4 Immobilization strategies for different ligands with protocol
numbers in parentheses (the recommended methods are listed rst).
Ligand Immobilization method Remarks
Proteins, general EDC/NHS (6.3) Thiol
exchange, maleimide
(6.6, 6.7) Reductive
amination (6.4) Ni
21
/
His
6
(6.12) Biotin/
streptavidin (6.13)
Direct adsorption
Dry coupling may be
necessary if pI o 4, as
preconcentration
becomes difcult
Membrane proteins On-surface reconstitution
(6.16) Incorporation
into labeled micelles
Requires labeled lipids
Antibodies EDC/NHS (6.3) Protein A
(6.14) Thiol exchange,
maleimide (6.6, 6.7)
Ni
21
/His
6
(6.12)
Hydrazide (6.9) Biotin/
streptavidin (6.13)
Obey different afnities
After oxidation of
glycosyl residues Critical
for steric reasons
Antibody fragments,
abodies
EDC/NHS (6.3) Thiol
exchange, maleimide
(6.6, 6.7)
Glycoproteins EDC/NHS (6.3) Hydrazide
(6.9) Thiol exchange,
maleimide (6.6, 6.7)
Peptides EDC/NHS (6.3) (6.2)
Reductive amination
(6.4) EDC/NHS-
activated peptides to
Dry coupling may be
necessary as
preconcentration is
immobilization methods for the most common ligands, although each method
is not suitable in each specic case.
6.6 Conclusions and Outlook
In this chapter, an introduction to surface chemistry for SPR and similar
biosensors has been presented. A few key factors, namely the composition and
nanostructure of the immobilization matrix, determine the performance and
characteristics of the sensor chip and can be used to optimize the system for a
particular application. As a direct consequence of the increasing variety of ligand
analyte interactions which can be analyzed by SPR, a choice of immobilization
methods has been developed, the most important of which have been described.
It can be concluded that the data quality delivered by optical biosensors is
only as good as the surface of the biochip, hence careful selection of the most
optimal surface for the corresponding experiment is necessary. Because SPR is
a very sensitive yet non-selective method, artifacts caused by defective or
Table 6.4 (continued )
Ligand Immobilization method Remarks
amines Biotin/
streptavidin (6.13) Ni
21
/
His
6
(6.12) Epoxide
coupling (6.10)
difcult. Dilute sample
to avoid crosslinking
Small molecules EDC/NHS (6.3) (6.2)
Reductive amination
(6.4) Biotin/streptavidin
(6.13) Hydrazide (6.9)
Reverse EDC/NHS to
amino surface Epoxide
coupling (6.10)
Chemistry depends on
functional groups of
ligand. Dry coupling
may be necessary as no
preconcentration effect
DNA,
oligonucleotides
EDC/NHS, dry (6.3) (6.2)
Epoxy coupling (6.10)
Biotin/streptavidin
(6.13)
EDC/NHS and epoxy
coupling require amino-
modied oligos but give
higher immobilization
yields
DNA, native PCR
products
Electrosorption (6.11)
EDC/NHS, dry (6.3)
(6.2)
Combine with epoxy
activation
Carbohydrates Hydrazide (6.9) Epoxy
coupling (6.10)
Cells, cell fragments Specic capture antibodies
Electrosorption to 2D
amine surfaces Lectins
Streptavidin (6.13)
If biotinylation possible
Viruses, fragments Specic capture antibodies
Electrosorption to
amine surfaces
216 Chapter 6
suboptimal surfaces become immediately visible, stressing the requirement for
rigorous optimization.
Despite the enormous progress in the elds of bionanotechnology and
surface science during recent years, surface design, especially for analytical
and biomedical devices, remains a demanding challenge. The interplay of
multifunctional macromolecules, ions and other dissolved species with water
molecules at interfaces results in a complex scenario which is difcult to
describe with theoretical models. Therefore, several phenomena and processes
at the molecular scale remain not fully understood. It is predictable, however,
that the last white spots on the map of surface science will probably disappear
within the next 10 years, contributing to sensor chips with increased signal-
to-noise ratio and a more homogeneous distribution of binding sites. In this
context, it is likely that todays frequently employed polysaccharide matrices of
microbial origin will be gradually replaced by better dened synthetic polymers
with optimized functionality and structure.
In addition to optimization of the surface homogeneity and minimizing non-
specic background, a clear demand exists for universal and easy to perform
immobilization techniques which are standardized and can be used for any
ligand irrespective of its chemistry, size or structure. On the way to such
universal solutions we will probably see the development of individual, appli-
cation-related chip surfaces with corresponding kits allowing reliable and even
fully automated immobilization of the most frequently employed ligand classes.
Surfaces for SPR microarrays to be used with 2D detectors are another area of
further development as the sheer number of proteins in the proteome creates an
increasing demand for parallel characterization in multiplex analysis.
6.7 Questions
1. Try to scale the dimensions roughly by drawing perpendicular to the surface
(1) the gold layer thickness of an SPR device, (2) the antibody coating, (3)
the evanescent eld and (4) the stagnant layer in mass transport-controlled
kinetics if we use a ow cell height of (5) two white blood cells.
2. Why is it important to check the quality of the sensor surface in an SPR
image or SPR dip shape check? What is the meaning of a shallow and wide
SPR curve?
3. The most popular immobilization chemistry is the EDC/NHS protein
coupling to a carboxymethylated dextran chip. Predict the sensorgram of
the coupling of a protein (molecular weight 10 kDa) with high isoelectric
point including ethanolamine deactivation.
4. Draw a typical sensorgram of three analysis cycles if the ligand is decoupled
after each regeneration step (e.g. 50% loss of ligand after a regeneration step).
5. Contact of the sensor surface with air may result in serious effects on the
performance of the sensor chip. Which coatings show the most dramatic
effects and should be prevented from any contact with air? And which
coating is robust?
6. A common observation with afnity biosensors is the dependence of the
electrostatic adsorption (preconcentration) and thus the immobilization
capacity from the ow rate of the liquid handling system. Would you expect
a higher or lower capacity with increasing ow rate? Explain the phenomenon.
7. a. Small ligands or oligonucleotides are frequently immobilized via biotin-
streptavidin interaction. Why?
b. An alternative can be the direct immobilization through reactive groups.
What should be obeyed in this case?
c. Discuss the advantages and disadvantages of indirect vs direct immobi-
lization.
References
1. B. Alberts, D. Bray, J. Lewis, M. Ra, K. Roberts and J.D. Watson,
Molkularbiologie der Zelle, 2nd edn., VCH, Weinheim, 1990.
2. J. Israelachvili, Intermolecular and Surface Forces, 2nd edn., Academic
Press, San Diego, 1992.
3. P. Schuck, L.F. Boyd and P.S. Andersen, Measuring protein inter-
actions by optical biosensors. In Current Protocols in Protein Science, John
Wiley & Sons, New York, 1999, 20.2.1-20.2.21. (Annual Report 1 Z01 OD
010485).
4. J. Svitel, A. Balbo, R.A. Mariuzza, N.R. Gonzales and P. Schuck, Biophys.
J., 2003, 84, 40624077.
5. R.G. Nuzzo and D.L. Allara, J. Am. Chem. Soc., 1983, 105, 44814483.
6. C. Pale-Grosdemange, et al., J. Am. Chem. Soc., 1991, 113, 1220.
7. C.D. Bain, J. Evall and G.M. Whitesides, J. Am. Chem. Soc., 1989, 111,
71557164.
8. N.T. Flynn, T.N.T. Tran, M.J. Cima and R. Langer, Langmuir, 2003, 19,
1090910915.
9. P. Van der Voort and E.F. Vansant, J. Liq. Chromatogr. Relat. Technol.,
1996, 19, 27232752.
10. J.J. Cras, C.A. Rowe-Taitt, D.A. Nivens and F.S. Ligler, Biosens. Bioelectron.,
1999, 14, 683688.
11. P. Silberzan, L. Leger, D. Aussere and J. J. Benattar, Langmuir, 1991, 7,
16471651.
12. S.R. Holmes-Farley, R.H. Reamey, T.J. McCarthy, J. Deutch and G.M.
Whitesides, Langmuir, 1985, 1, 725740.
13. U. Lappan, H.-M. Buchhammer and K. Lunkwitz, Polymer, 1999, 40,
40874091.
14. A. Delcorte, P. Bertrand, E. Wischerho and A. Laschewsky, Langmuir,
1996, 13, 51255136.
15. L. Vroman, J.S. Mattson and C.A. Smith, Science, 1974, 184, 585586.
16. M.A.C. Stuart, Surfact. Sci. Ser., 1998, 75, 125.
218 Chapter 6
17. P. Billsten, M. Wahlgren, T. Arnebrandt, J. McGuire and H. Elwing,
J. Colloid Interface Sci., 1995, 175, 7782.
18. S. Ravichandran and J. Talbot, Biophys. J., 2000, 78, 110120.
19. P. Vadgama, Annu. Rep. Prog. Chem., Sect. C, 2005, 101, 1452.
20. A. Wittemann, B. Haupt and M. Ballau, Phys. Chem. Chem. Phys., 2003,
5, 16711676.
21. C. Czeslik, R. Jansen, M. Ballau, A. Wittemann, C.A. Royer, E. Gratton
and T. Hazlett, Phys. Rev. E, 2004, 69, 021401.
22. E. Ostuni, R.G. Chapman, R.E. Holmlin, S. Takayama and
G.M. Whitesides, Langmuir, 2001, 17, 56055620.
23. P.M. Claesson, in Biopolymers at Interfaces, ed. M. Malmsten, Marcel
Dekker, New York, 1998, Vol. 75, pp. 281320.
24. S.I. Milner, T.A. Witten and M.E. Cates, Europhys. Lett., 1988, 5,
413418.
25. I. Szleifer, Physica A, 1996, 244, 370388.
26. C. Siegers, M. Biesalski and R. Haag, Chem. Eur. J., 2004, 10, 28312838.
27. B.W. Morrisey and C.C. Han, J. Colloid Interface Sci., 1978, 65, 423431.
28. C.J. Van Oss and J.M. Singer, J. Reticoendothelial Soc., 1966, 3, 2940.
29. J.E. Butler, L. Ni, R. Nessler, K.S. Joshi, M. Suter, B. Rosenberg,
J. Chang, W.R. Brown and L.A. Cantarero, J. Immunol. Methods, 1992,
150, 7790.
30. B. Johnsson, S. Lo fas and G. Lindquist, Anal. Biochem., 1991, 198, 268277.
31. D.G. Hoare and D.E. Koshland, J. Am. Chem. Soc., 1966, 88, 20572058.
32. J.C. Sheehan and J.J. Hlavka, J. Org. Chem., 1956, 21, 439441.
33. J.C. Sheehan, J. Preston and P.A. Cruickshank, J. Am. Chem. Soc., 1965,
87, 24922493.
34. P. Cuatrecasaes and I. Parikh, Biochemistry, 1972, 11, 22912299.
35. L. Peng, G.J. Calton and J.W. Burnett, Appl. Biochem. Biotechnol., 1986,
14, 9199.
36. S.P. Massia and J. Stark, J. Biomed. Mater. Res., 2001, 56, 390399.
37. B. Renberg, I. Shiroyama, T. Engfeldt, P. Nygren and A.E. Karlstro m,
Anal. Biochem., 2005, 341, 334343.
38. D.J. Oshanessy, M. Brigham-Burke and K. Peck, Anal. Biochem., 1992,
205, 132136.
39. D.G. Smyth, O.O. Blumenfeld and W. Konigsberg, Biochem. J., 1964, 91,
589595.
40. C.F. Brewer and J.P. Riehm, Anal. Biochem., 1966, 18, 248255.
41. G. Gorin, P.A. Martin and G. Doughty, Arch. Biochem. Biophys., 1966,
115, 593597.
42. D.J. Oshannessy and R.H. Quarles, J. Appl. Biochem., 1985, 7, 347355.
43. B. Johnsson, S. Lo fas, G. Lindquist, A. Edstro m, R.-M. Mu ller Hillgren
and A. Hansson, J. Mol. Recognit., 1995, 8, 125131.
44. L. Sundberg and J. Porath, J. Chromatogr., 1974, 90, 8798.
45. D.J. OShannessy, K.C. ODonnell, J. Martin and M. Brigham-Burke,
Anal. Biochem., 1995, 229, 119124.
46. P.D. Gershon and S. Khilko, J. Immunol. Methods, 1995, 183, 6576.
47. L. Nieba, S.E. Nieba-Axmann, A. Persson, M. Ha ma la inen, F. Edebratt,
A. Hansson, J. Lidholm, K. Magnusson, A.F. Karlsson and A. Plu ckthun,
Anal. Biochem., 1996, 252, 217228.
48. J. Quinn, P. Patel, B. Fitzpatrick, B. Manning, P. Dillon, S. Daly,
R. OKennedy, M. Alcocer, H. Lee, M. Morgan and K. Lang, Biosens.
Bioelectron., 1999, 14, 587595.
49. P. Nilsson, B. Persson, M. Uhle n and P.A. Nygren, Anal. Biochem., 1995,
224, 400408.
50. P.A. Ohlsson, T. Tja rnhage, E. Herbai, S. Lo fas and G. Puu, Bioelectro-
chem. Bioenerg., 1996, 38, 137148.
51. M.A. Cooper, A. Hansson, S. Lo fas and D.H. Williams, Anal. Biochem.,
2000, 277, 196205.
52. O.P. Karlsson and S. Lo fas, Anal. Biochem., 2002, 300, 132138.
53. R.L. Rich and G. Myszka, J. Mol. Recognit., 2005, 18, 139.
220 Chapter 6
CHAPTER 7
Measurement of the Analysis
Cycle: Scanning SPR
Microarray Imaging of
Autoimmune Diseases
RICHARD B.M. SCHASFOORT,
a
ANGELIQUE M.C.
LOKATE,
b
J. BIANCA BEUSINK,
a
GER J.M. PRUIJN
b
AND GERARD H.M. ENGBERS
c
a
b
Department of Biomolecular Chemistry, Nijmegen Centre for Molecular
Life Sciences, Institute for Molecules and Materials, Radboud University
Nijmegen, P.O. Box 9101, NL-6500, HB Nijmegen, The Netherlands;
c
IBIS
Technologies B.V. P.O. Box 1242, 7550 BE Hengelo, The Netherlands,
Internet: www.ibis-spr.nl
7.1 Introduction
Surface plasmon resonance (SPR) [1] is a direct sensing analysis technique with
a unique combination of features. Since it is a label-free technique it does not
involve changes of the structure of the (bio)molecules that may be involved in
the biomolecular interaction and the technique yields information on the
concentration (how many analyte molecules are present in a complex sample),
the afnity (how strong an interaction is) and the kinetics (how fast an
interaction is) of biomolecules and their interactions [2].
Almost every SPR-experiment applies the same basic procedure in which the
rst step, the so-called immobilization (see also Chapter 6), involves covalent
binding of a ligand to the surface of the sensor followed by the so-called
analysis cycle [3]. In order to detect the presence or a specic interaction of a
221
certain analyte in a sample, prior to analysis, a capturing entity (the ligand)
should be immobilized on the sensor surface [4]. Although the ligand can be any
type of (bio)molecule, more than 80% of the capturing agents used in SPR
analysis are related to proteins (e.g. antibody, antigen, receptor, peptide).
In the previous chapter, various chemical protocols are given for the attach-
ment of ligands to the sensor surface. The present chapter covers the analysis
cycle for measuring biomolecular interactions with SPR detection. After a
discussion of the analysis cycle, it focuses on the use of SPR in immunoassays,
including the latest developments of new SPR instrumentation: scanning SPR
microarray imaging for monitoring autoimmune diseases. SPR imaging in
combination with microarrays for genomics and proteomics applications is
rapidly evolving. To demonstrate the features and benets of multianalyte
microarray-based biomolecular interactions, the analysis cycle of autoimmune
antibody interactions in serum monitored with SPR imaging is discussed.
7.2 The Analysis Cycle
After immobilization of a preferably pure ligand, the analysis cycle starts with
the introduction of a buer solution in the measurement cell, in order to
generate a (stable) baseline. Subsequently, a sample, containing the ligands
binding partner (the analyte), is introduced into the measurement cell and the
analyte binds to the ligand, leading to the selective accumulation of the analyte
on the sensor surface. This causes an increase in the refractive index near the
sensor surface. This change in refractive index can be measured in real time.
The selectively accumulated mass on the sensor chip surface (which is generally
expressed in pg mm
2
) correlates linearly with the change in the refractive
index near the sensor surface measured by the SPR instrument [5]. A rule of
thumb is that for an instrument that uses light with a wavelength of 670 nm,
1 ng mm
2
protein gives a signal of about 100 millidegrees (m1) angle shift. In
Biacore instruments, 10 resonance units (RU) correspond to approximately
1 m1 SPR angle shift. When light of another wavelength is used then the angle
shift should be calibrated.
The response in SPR experiments of which a typical example is shown in
Figure 7.1 directly reects the change in the refractive index of the liquid phase
near the surface of the sensor chip. The observed response is the sum of the
response caused by association of analyte with the ligand and that caused by
the difference in refractive index between the solution and the baseline buer.
The latter is called the bulk effect [6] and is caused by the presence of
dissolved material including buer components, biomolecules and salt. The
bulk effect needs to be subtracted from the observed response in order to
measure the true binding of the analyte molecules. If a reference channel is
available in the SPR instrument, then the bulk effect can be determined
accurately. This reference channel may also correct for non-specic adsorption
of sample components to the surface of the sensor chip. Using baseline buer as
the solvent of the analyte will minimize the bulk effect. In experiments for
222 Chapter 7
qualitative determination of whether the analyte will bind or not, careful
adjustment of the buer is not critical. However, for quantitative kinetic
determinations in which a series of different analyte concentrations are used,
it is important that buer exchange is performed reliably, for instance by
dialysis or gel ltration. Tuning the refractive index of the background buer to
mimic the sample and using the baseline buer for dilution of the sample is a
way to minimize the bulk effect.
After the association process, buer solution is introduced into the system,
which initiates the dissociation phase. During the dissociation phase, the
specifically and non-specifically bound compounds may detach from the
surface, resulting in a decrease in response. Because in the dissociation
phase the background buer is usually the same as the buer used during the
baseline measurement, the exact amount of associated material can be
Time
Phases: baseline, association, dissociation, regeneration, baseline
Injection
Bulk shift
Figure 7.1 A typical sensorgram obtained from an instrument with a reference
channel showing the phases of an analysis cycle. Two channels/spots
are measured; one with ligand and the other without ligand as a reference
channel. The refractive index bulk shifts can be clearly observed from the
reference channel. Double arrow represents the specic response. (1)
Baseline phase: initially, baseline buer is in contact with the sensor
surface to establish the baseline. For sensor calibration purposes, the
injection of a calibration liquid (e.g. 1% glycerol spiked in baseline buer)
can be incorporated in this phase (not shown). (2) Association phase:
sample containing the target component is injected, the capturing
elements on the sensor surface bind the target component resulting in
complex formation. (3) Dissociation phase: On injection of baseline
buer, target components (and also non-specifically bound molecules)
dissociate from the surface. Note: in this sensorgram, dissociation as an
exponential decay is very low. (4) Regeneration phase: the regeneration
solution (e.g. low-pH buer) is injected to remove the remaining bound
analyte. (5) After this phase, the cycle is completed and a new experiment
can start by establishing the baseline again. If remaining accumulated
mass is present, the baseline level will increase.
223 Measurement of the Analysis Cycle
calculated. It should be mentioned that in order to prevent rebinding effects in
the dissociation phase, it is sometimes preferred to add ligand to the dissoci-
ation buer (see Chapter 5, Section 5.3.2.2). This may induce a bulk shift which
should be subtracted using the response obtained in the reference channel.
Subsequently, a regeneration solution that, for instance, induces a strong pH
shift (including a strong bulk refractive index shift to low SPR angles) is
injected and the ligandanalyte complex dissociates (preferably) completely.
Injection of the baseline buer closes the analysis cycle and brings the system
into its starting situation and the completed sensorgram is now available for
analysis. Exactly timed liquid handling procedures are required to determine
accurate kinetic rate constants of biomolecular interactions.
In experiments where the sensorgram is compensated for bulk shift effects
(for instance, when multichannel instruments are used in which one channel
serves as reference), the association and dissociation phase of the sensorgram
allow the calculation of the reaction rate and the thermodynamic equilibrium
constants. If this is available, it is performed using instrument-specic software.
Alternatively, it can be performed with general kinetic evaluation software.
Furthermore, from the sensorgram the effectiveness of the regeneration pro-
cedure can be determined. Incomplete regeneration will result in residual
analyte and/or non-specifically bound components on the sensor surface, which
will increase the baseline, whereas too harsh sensor regeneration will result in a
decreased binding capacity of the sensor in subsequent analysis cycles
1
.
7.3 Buer Solutions for Measuring the Analysis Cycle
7.3.1 Baseline Buer
The running, background or baseline buer should create optimal conditions
for the binding of the ligand to the analyte. For biomolecular interactions the
baseline buer is usually a physiological buer with sufcient salt and (near-)
neutral pH. Phosphate-buered saline (PBS) is a standard baseline buer.
Alternatively, 10 mM HEPESKOH (pH 7.4), 0.15 M NaCl is often used. The
addition of a surfactant, e.g. Tween 20 (0.030.05%) to the buer is favored in
order to minimize non-specic binding. The surfactant not only enhances the
ratio between specic and non-specic binding but also helps to prevent air
bubble adsorption on the surface. However, surfactants should not be added if
hydrophobic surfaces are used for non-covalent attachment of membrane-
bound components. Moreover, a surfactant should not be used during the
immobilization cycle (see Chapter 6). Sometimes 3 mM EDTA is added to trap
remaining bivalent cations (e.g. Mg
21
or Ca
21
), which may interfere with the
carboxylate groups in the hydrogel layer. Blocking components may help
reducing non-specic binding, e.g. 3% BSA or HSA if patient serum is used.
1
See Figure 6.3 in Chapter 6 for a sensorgram depicting unstable immobilization and partial
denaturation of the ligand.
224 Chapter 7
In the experiments on samples from rheumatoid arthritis (RA) patients
described in Section 7.6, normal sheep serum was applied.
All buer solutions should be prepared with Milli-Q water and preferably
ltered through a 0.22 mm lter and degassed before use.
7.3.2 Regeneration Solution
For repeated use of the same sensor chip, the surface should be regenerated by
removal of analyte and any other non-covalently bound material. However, the
ligand should be kept intact and should not be inactivated or denatured during
the regeneration phase. Commonly used solutions for regeneration include low-
pH buers, e.g. 10 mM glycineHCl (pH 1.5). Optimal sensor regeneration
includes a pH shock and regeneration is preferably performed as two consec-
utive steps of, for instance, 30 s each, rather than one step of 60 s. If a negatively
charged gel-type sensor is used, not only the specic bond between ligand and
analyte will break during low pH regeneration, but also the hydrogel will
collapse and the analyte will be more or less squeezed out of the hydrogel. If the
ligand or analyte cannot withstand low pH, sometimes a high alkaline pH is
used, e.g. 10 mM NaOH (pH411). Alternative regeneration solutions have a
high salt concentration and the salts used include chaotropic agents that are
chemicals, such as urea and guanidine hydrochloride, that disrupt hydrogen
bonding in aqueous solutions. It should be noted that concentrated solutions of
these agents may denature proteins, because they also interfere with hydro-
phobic interactions.
7.4 SPR-based Immunoassays
In general, an immunoassay is a laboratory technique based on the binding
between an antigen and its homologous antibody in order to identify and
quantify the specic antigen or antibody in a sample. In classical immunoas-
says, the detection of the concentration of an analyte relies on signals generated
by various labels (uorescent dyes, enzymes or radioisotopes) attached to
antigens or antibodies [7]. Labeling may disrupt the binding sites involved in
the interaction. Moreover, labeling induces heterogeneity of the biomolecular
interaction because in most cases labeling of a specic molecule (e.g. an
antibody) is not homogeneously distributed. In addition, the label itself might
interact with the capturing ligands, leading to false positives. A way to
circumvent some of the problems is to use a labeled secondary binding
molecule, but this extra step requires an additional, high-afnity binding
compound and it will also increase the required analysis time. SPR-based
biosensors measure proteinprotein binding directly as a shift in surface-bound
masses. Since SPR is a label-free analysis technique, it is preferred for studying
biomolecular interactions.
The observed binding rate of analyte to the ligand at the sensor surface may
be limited either by the rate of interaction (kinetics) or the rate of supply of
analyte to the surface (mass transport). Large amounts of ligand on the surface
lead to higher interaction rates and low sample ow rates lead to a slower
supply of analyte to the surface. Both of these factors contribute to mass
transport-limited binding. These considerations are important for kinetic
evaluation of association rate constants k
a
and dissociation rate constants k
d
.
If mass transport limits the observed binding rates, then the apparent associ-
ation rate constant from the simple model A+B#AB will be lower than the
true value. A high ligand loading leads to rebinding effects during the disso-
ciation phase and lower apparent dissociation rate constants will be measured
for AB#A+B. In order to prevent rebinding, the ligand which is immobi-
lized to the surface may be added to the dissociation buer too. Rebinding has
been demonstrated at high ligand loadings in hydrogels. It affects the calcula-
tion of the afnity constant dramatically and deviations of as high as two
orders of magnitude have been reported [8].
Concentration measurements require a high loading of ligand on the surface
and constant hydrodynamic conditions (ow rate or mixing speed). If the
analyte dissociates from the surface it should not rebind at free sites while
diffusing out of the hydrogel. Kinetic rate experiments require a low surface
loading of the ligand, ideally the lowest ligand loading that still gives a
measurable response after analyte binding. Mass transport limitation and other
kinetic evaluation considerations are treated in detail in Chapters 4 and 5.
Figure 7.2 shows different assay formats that are appropriate for SPR including
direct, competitive, inhibition and sandwich assays.
7.4.1 Direct Assay
The direct assay is described by A+B#AB. In this type of assay, antibodies
directed against the antigen are immobilized on the sensor surface (ligand).
Sample solution containing the antigen (analyte) is then incubated with the
sensitized sensor surface. The signal increase resulting from antigen binding
correlates with the amount of antigen in the sample. Direct assays can also be
designed with the antigen coupled to the surface and detection of the binding of
the specic antibody. This is the case in the example shown in Section 7.6 of an
immunoassay for the monitoring of autoimmune autoantibodies in the serum
of rheumatoid arthritis patients.
7.4.2 Competition Assay
This type of assay is typically applied for the detection of low molecular weight
antigens that do not generate sufcient signal, while accumulating at the sensor
surface. In this assay format specic antibodies are immobilized on the sensor
surface and sample solution that contains the antigen is mixed with an antigen
conjugate. Because of its high molecular weight, the conjugate enhances the
SPR angle shift. The antigenconjugated antigen mixture is incubated with the
sensor surface. The difference in signal between a reference sample containing
226 Chapter 7
only conjugated antigen and the sample solution indicates the amount of
antigen in the sample. In this assay, high antigen concentrations in the sample
will result in low signals (less conjugated antigen can be bound).
Competition assays are often used for the detection of toxic compounds (see
also examples in Chapter 11). The maximum signal is generated when no free
(toxic) analyte is present. When the signal is too low, possibly the analyte is
present in the sample or the ligands are denatured or poisoned by the sample
and no longer active while the analyte is not present. Both outcomes are
harmful and will require action. If the competition assay shows a response of
the unconjugated antigen as in the direct assay, then severe calibration proce-
dures have to be performed. Preferably in SPR competition immunoassays the
conjugated antigen is a large refractive index label (e.g. a latex bead or gold
nanoparticle) loaded with the antigen.
7.4.3 Inhibition Assay
In this assay format, the target antigen is immobilized on the sensor surface.
Sample solution containing the antigen is mixed with specic antibodies in
Figure 7.2 Immunoassay formats commonly used in SPR measurements. (a) Direct
assay: the ligand (antibody) is immobilized on the sensor surface interac-
tion with the analyte (here antigen) yields a detectable refractive index
shift. (b) Competition assay for measuring small molecules where direct
capturing of the antigen yields insufcient refractive index shift, while the
conjugated antigen is large enough for a measurable refractive index shift
(see also Figure 11.4). (c) Inhibition assay where the analyte is the same
molecule as the immobilized ligand. Antibody is added to the sample in
excess. The analyte forms conjugates with the antibody, inhibiting the
binding to the sensor surface (see also Figure 11.2). (d) Sandwich assay
with secondary antibody.
excess and incubated with the sensor surface. Antibodies will bind both to
antigen in solution and to antigen that is bound to the sensor surface. The
difference in signal between a blank sample that does not contain the antigen
and the sample solution indicates the amount of antigen in the sample. In this
assay, high antigen concentrations in the sample result in low signals (less
antibodies remain to bind to the surface). Because antibodies have high
molecular weight, their binding is directly detected.
7.4.4 Sandwich Assay
In sandwich assays, antibody molecules against the analyte are immobilized on
the sensor surface, capturing the analyte molecules when sample solution is
incubated with the sensor surface. In the next step, a secondary antibody binds
specifically with either the antigen or the antigen-bound antibody. The antigen
is captured by a sandwich of two antibodies. Only very high afnity capture
antibodies should be used, in order to avoid a mixture of afnities of each
component in the sandwich. Several steps can be build in, but this complicates
the analysis.
Often, a highly specic goat/rabbit/sheep anti-mouse IgG is immobilized as
the rst capturing agent which traps a monoclonal (mouse) antibody for the
antigen. Streptavidinbiotin linkers are often used because of the high afnity
constant, which will not interfere with the rate and equilibrium constants of the
analyteligand pair. The increase in signal (as a result of antigen binding) is
proportional to the amount of antigen in the sample. Washing the surface with
buer is followed by the injection of a secondary antibody. The high molecular
weight of the secondary antibody is usually sufcient for monitoring the
binding process. For further signal enhancement, antibody conjugates with
colloidal gold or latex particles as refractive index label can be used [9].
7.5 Detection of Multiplex Analysis Cycles Using
Scanning SPR Imaging
SPR imaging can be considered as a new trend (see Chapter 12), although the
principle was published decades ago [10,11]. Currently computers are fast
enough to process digital images allowing kinetic measurements of biomolec-
ular interactions in real time. This allows the monitoring of multiple analysis
cycles on a microarray by an SPR imaging instrument. In our studies, the
instrument for biomolecular interaction sensing by imaging SPR (IBIS iSPR,
IBIS Technologies, Hengelo, The Netherlands) is used for this purpose and is
represented schematically in Figures 7.3 and 7.4.
Fixed-angle instruments in the Kretschmann conguration (Chapter 3) are
based on the relationship between a small change in the intensity of the reected
light and the mass of bound analyte, i.e. a xed incident light angle is employed
and mass changes are estimated from the intensity of the reected light.
However, the applicable dynamic range and linear relationship of this
228 Chapter 7
experimental setup are limited [12] and the optimal incident angles of each spot
of a microarray differ considerably when ligand or analyte panels with different
molecular weights are monitored, as explained in Chapter 3. Therefore, choosing
one xed angle to monitor many biomolecular interactions in a microarray will
Figure 7.3 Schematic representation of the optical conguration of the IBIS iSPR
instrument. An 840 nm light beam is passed through a p-polarizer and a
focusing lens before reaching the hemispherical prism and gold sensor.
The incident light is reected by the angle operated mirror with a max-
imum scanning angle of 81. The light beam permits imaging of a 50 mm
2
sensor surface. The light reects from the sensor and passes through
another lens before the CCD camera transfers the uncompressed images
to the software.
Figure 7.4 Schematic representation of the liquid handling system of the IBIS iSPR in
ow cell conguration. Samples are injected from the thermostated sample
rack, using the needle of the liquid handler robot arm connected to the
ow cell.
result in only qualitative data. Kinetic rate constants, however, require accurate
quantitative data including subtraction of bulk effects for all the spots of the
microarray simultaneously.
7.5.1 Dynamic Range of Scanning SPR Imaging
The continuous SPR dip angle scanning mode of the IBIS iSPR provides a
large dynamic range, thereby facilitating accurate, simultaneous measurements
of, for instance, the presence and concentration of large proteins and small
peptides in the same sample.
In Figure 7.5, the difference in refractive index dynamic range of angle
scanning and xed-angle measurements is shown. The actual SPR angle
depends on the refractive index in the evanescent eld of the liquidgold
interface and microarrays spotted with different ligands show an initial distri-
bution of shifts resulting in baselines at different angle positions. A problem
exists when the biomolecular interaction is followed at xed angle caused by the
non-linear shape of the SPR curve as the baselines from spots of different
ligands and immobilized biomolecular masses show different levels of reectiv-
ity. In SPR imaging instruments, the SPR angles of many regions of interest
(ROIs) should be monitored simultaneously. Comparison/subtracting baselines
of different reectivities is essentially incorrect. In fact, the slope of the reec-
tivity vs. angle curve is not constant and every spot requires its own calibration.
The only reliable parameter that directly reects the concomitant mass
change on an SPR sensor is the SPR angle [13]. Thus, only when the SPR
angle is monitored for all spots on a microarray independently can the
Linearity scanning angle and fixed angle SPR
0
500
1000
1500
2000
2500
0 0.005 0.01 0.02 0.025 0.03 0.035
refractive index
0
100
150
200
50
A
n
g
l
e

s
h
i
f
t

i
n

m
R
%
)
0.015
scanning angle (left Y-axis)
fixed angle (right Y-axis)
Figure 7.5 Relationship of the angle shift of the SPR dip (m1) (left axis) and the
change in refractive index (Dn) of glycerol dilution series in PBS in the
scanning angle mode. On the right axis the reectivity change (DR%) is
shown for this range of refractive indices at xed angle settings (the value
of the xed angle is in the inection point of the left-hand ank of the SPR
curve; see Chapter 3, Figure 3.2). Starting angle is at absolute reectivity
of about 40%.
230 Chapter 7
magnitude and afnity of biomolecular interactions be reliably compared
among all spots of the microarray. In the IBIS instrument, the SPR angle of
all spots of a microarray can be determined continuously. Thanks to its novel
angle-scanning principle with tting algorithms to calculate the resonance angle
of the SPR dip for each ROI, this instrument quanties the protein mass
distribution on the microarray directly.
The unique, innovative character of the IBIS imaging SPR instrument lies in
the accuracy of SPR angle determination of multiple ROIs in combination with
real-time imaging of the complete sensor surface. Inspection of the sensor
surface using the microscopic SPR image shows the quality of spots (in terms of
coating homogeneity), which is an important feature for generating reliable
data (see Figures 7.7, 7.8 and 7.9). SPR angles of microarrays of more than
500 ROIs can be determined simultaneously, with the maximum number of
ROIs being limited by the computing power and the lateral resolution of the
SPR technique.
In Figure 7.6, an image of a microarray is shown, which is spotted, in situ, in
the IBIS iSPR instrument. Under scanning conditions, only homogeneous
spots with equally distributed pixel intensities give reliable SPR results. A
special needle in the liquid handling system automatically spots about 50 nl on
a dened location creating the microarray. Drying of the spots can be pre-
vented if the humidity is close to 100%. Evaporation can be delayed by
applying non-interfering compounds in the spotting buer, e.g. glycerol.
However, this may impair the reproducibility of the results. In combination
with angle scanning and inspection of the SPR dip in terms of depth and width,
an indication of the quality of the sensing surface can be obtained. If, e.g., bare
gold surfaces are used for ligand immobilization, often cauliower images can
be observed caused by irreproducible drying effects which can result in unre-
liable sensorgrams. Preferably hydrophilic coatings such as hydrogels are
Figure 7.6 SPR image of 5 6 spots generated in situ with the liquid handling system
of the IBIS iSPR instrument. Right image: ROIs (up to 500) can be
distributed automatically or manually by the operator of the instrument.
applied for optimal results. The best contrast can be obtained when homoge-
neous intensities in one spot are at the left ank of the SPR curve. If the light is
not homogeneously distributed over the entire surface, it has an effect on the
shape (width) of the reectivity vs. angle plot or SPR dip, but will not affect the
minimum resonance condition.
7.5.2 Liquid Handling Procedures
For the IBIS iSPR instrument, different types of ow cells and cuvettes are
available and alternatively the instrument can be combined with lab-on-a-
chip devices for controlling the sample ow over a microarray. Flow cells
proved to be the best with regard to performance, accuracy and precision,
providing the lowest limit of detection and best response stability. Cuvette-
based systems offer the user maximum exibility. An advantage of a cuvette-
based systems for studies of kinetic rate constants is that higher mass transport
can be obtained towards the interaction site [14]. A disadvantage is that the
mixing parameters such as injection, needle height and mixing speed need to be
optimized and controlled in order to prevent shear stress at the sensor surface.
For example, particles in the sample can hit the sensor surface, leading to
damage of the microarray.
7.5.3 Determination of the Limit of Detection Using Multiplex
Analysis Cycles
In order to obtain an indication of the limit of detection (LOD) a microarray
containing various densities of two different ligands was spotted on a sensor
chip. Regeneration of the sensor with 10 mM glycineHCl (pH 1.5) provided
the ability to test several antibody concentrations on one sensor chip.
Figure 7.7 shows the sensorgram [angle shift (m1) as a function of time (s)] of
eight analysis cycles of the interaction of the biotinylated peptides and -proteins
with an anti-biotin antibody. The different curves represent the interaction with
the various amounts of biotinylated peptides spotted using Top Spot [15,16] as
indicated in the gure caption. The curves varying over time (x-axis) represent
the different concentrations of the analyte, anti-biotin IgG. From the analysis
cycles of every spot, in principle overlay plots of the interactions with the
different analyte concentrations can be generated in order to calculate the on-
and o-rate constants and afnity equilibrium constants. The afnity constants
can be compared with a one-shot or single injection analysis, a term recently
introduced by Bio-Rads Proteon system. When a serially diluted ligand is
immobilized on the surface, the LOD can be expressed as the smallest change in
refractive index of the bulk solution or refractive index change in the evanes-
cent eld caused by accumulated mass.
As indicated before, if compounds with an increased refractive index replace
the background electrolyte buer solution, a linear relationship exists between
the surface coverage (in %) and the molecular weight of the adsorbate.
232 Chapter 7
Therefore, a low molecular weight compound (e.g. a small peptide) will
contribute less to the accumulated mass per surface area than a high molecular
weight compound (e.g. an immunoglobulin). The LOD can be used to char-
acterize an instrument; however, in Chapters 4 and 5 it has been shown that the
LOD also depends strongly on the afnity constant of biomolecular
Response curves of immobilized peptide in femtomole
0
50
100
150
200
250
300
0.1 1 10 100 1000
[IgG] (pM)
A
n
g
l
e

s
h
i
f
t

(
m
d
e
g
)
417
208
104
52
26
0
Figure 7.7 Sensorgram of eight analysis cycles (raw data) of a biotinantibody
interaction microarray (shown in Figure 7.8) at different ligand and
analyte concentrations for the determination of the on and o rate
constants. Injection of a single analyte was done, which is similar to the
one-shot analysis approach as described in Chapter 12. An injection of the
highest analyte mouse anti-biotin mouse IgG (150 kDa) concentration
(133 pM) resulted in a gradual decrease in responses of the spots loaded
with different ligand biotinylated peptide (2.4 kDa) concentrations ob-
tained by spotting using Top Spot [14,15] of 417, 208, 104, 52 and 26 fmol
and two controls, non-biotinylated peptide and the background. After
this rst analysis cycle, a lower mouse (53 pM) anti-biotin mouse IgG
(150 kDa) analyte concentration was injected. Only non-specic dissoci-
ation can be observed at higher concentrations of the analyte. In total
eight analysis cycles were carried out over the range of 133, 53, 13, 5.3,
1.3, 0.53, 0.13 and 0 pM anti-biotin mouse IgG. Each curve shows the
transient 2 m1 response as a result of the change in ow direction. The few
spikes are caused by air bubbles disturbing the SPR signal. The inset
shows the doseresponse curves of the spots as obtained from this
sensorgram after a xed interaction time just after the association phase.
In Figure 7.8, the SPR view of this microarray is shown.
interactions [14]. In our test system using the biotinstreptavidin pair, the
specifically bound analyte will not dissociate significantly because of the high
afnity constant of this couple. At high concentrations of the analyte only a low
non-specic dissociation effect occurs, as shown in Figure 7.7.
The molecular mass of the analyte (in this case the anti-biotin IgG, B150 kDa)
gives an indication of how many molecules are needed to obtain a significant
response compared with the signal-to-noise ratio. The amount of analyte bound
to the immobilized ligand can be derived from the measured angle shift;
1 m1 10.8 pg mm
2
, resulting in 1.62 amol IgG in the case of an ROI of
150 150 mm and 65zmol or 4.1 10
4
IgG molecules bound to the minimal
ROI of 3030mm. Figure 7.7 shows the importance of both a balanced
combination of the ligand and analyte concentrations for concentration meas-
urements (maximum angle shift) and kinetic parameter determination (low
rebinding, low mass transport limitation). Figure 7.8 shows the corresponding
uorescence image after removing the sensor chip from the system. More on
combining uorescence detection and SPR excitation can be found in Chapter 9.
Secondary antibodies can be used not only to introduce uorescent labels to
check the system but also for linking to refractive index labels such as latex
particles and colloidal gold to improve sensitivity (see Section 7.4). The second-
ary antibody linked to the mass label will induce a measurable shift, thereby
improving the LOD if the mass label ts into the evanescent eld and in the
hydrogel. The use of secondary antibodies can even increase the specicity of a
reaction; however, an additional interaction step has to be included which
overrules the favorable label-free one-step measurement. The application of SPR
imaging for multi-analyte detection in one step is shown in the following section.
Figure 7.8 Imaging SPR vs. uorescence microscopy. Microarray of the spotted
ligands using Top Spot [15,16] after interaction with 133 pM Alexa Fluor
488-labeled mouse anti-biotin IgG. The top two rows contain various
peptide concentrations spotted in duplicate: from left to right, 417 fmol of
a non-biotinylated peptide as a negative control followed by 26, 52, 104,
208 and 417 fmol of the biotinylated peptide. The bottom two rows
contain various IgG concentrations spotted in duplicate: 6.67 fmol of a
non-biotinylated IgG as a negative control followed by 0.42, 0.83, 1.67,
3.33 and 6.67 fmol of biotinylated IgG. In the SPR imaging view (a) the
non-biotinylated ligand (left column; peptide and IgG) can be clearly
seen, whereas in the uorescence image (b) these spots are not observed.
234 Chapter 7
7.6 Monitoring of Autoantibodies in Serum of
Rheumatoid Arthritis Patients
The potency of SPR imaging is demonstrated by monitoring the analysis
cycles of serum autoantibodies of rheumatoid arthritis (RA) patients that
bind specifically to citrullinated peptides in a single step. Autoimmune diseases
are characterized by the presence of high-afnity autoantibodies directed
against self-proteins, such as rheumatoid factor for RA [17], Sm for systemic
lupus erythemathosus (SLE) [18] and Ro/SS-A and La/SS-B for Sjo grens
syndrome [18]. Although at least some autoantibodies are known to be
involved in cell and tissue damage, their mechanistic role in the patho-
genesis of the disease is generally not known [19]. Nevertheless, the specicity
of autoantibody responses highlights their potential as important tools for
improved diagnosis, disease classication and prognosis. Miniaturized
multiplex assays can deliver a ngerprint of a patients autoantibody repertoire
requiring only a limited amount of patient material. During the last decade,
various research groups have made important contributions to the application
of multiplex assays for autoantibody detection. In 2002, Robinson et al.
employed protein and peptide ligand arrays, representing candidate auto-
antigens, to survey autoantibody binding [20]. Arrays of in situ synthesized
peptides can also be generated with photolithography to perform antibody
characterization [21]. Another approach is to apply virtual arrays in a
homogeneous assay system with addressable beads [22]. However, in
these systems at least one of the interactants has to be labeled, which may
disrupt the binding sites involved in the interaction, leading to false negatives.
In addition, the label itself might interact with the immobilized proteins,
leading to false positives [23]. A way to circumvent these problems is to use a
labeled secondary binding molecule, which might result in improved assay
sensitivity.
In this section, scanning SPR microarray imaging is used to measure the
presence of anti-citrullinated protein antibodies (ACPA) in the sera of RA
patients. Recently, it was shown that the so-called citrulline amino acid (which
can be generated post-translationally from arginine) is a critically important
moiety of the antigenic determinants targeted by RA-specic autoantibodies
[24]. Cyclic citrullinated peptides (CCPs) are widely used as antigenic targets in
ELISA-based diagnostic tests for RA. The sensitivity and specicity are 71 and
99%, respectively [25].
7.6.1 Experimental Conditions for Serum Measurements
7.6.1.1 Serum Samples
The sera were obtained from the Department of Rheumatology, University
Hospital Nijmegen. Sera were collected from patients visiting the outpatient
clinic who had been diagnosed as having RA according to the revised criteria of
the American College of Rheumatology. To assess specicity further, we
analyzed a group of serum samples from healthy individuals and groups of sera
from patients with osteoarthritis and SLE, obtained from various clinics
and hospitals. Sera were stored at 80 1C until used. Anti-CCP2 ELISA
was performed by IMMUNOSCAN RA (Euro-Diagnostica, Arnhem, The
Netherlands), in accordance with the manufacturers instructions with the
recommended 25 Uml
1
cut-o.
7.6.1.2 SPR Microarray Interaction Studies
SPR detects changes in refractive index in the hydrogel (200 nm) which is linked
to the gold surface. Due to the small molecular weight of the synthetic peptides
used (B1500 Da), the contrast of the immobilized array to the background is
not high. To visualize the array for determining the ROIs, human IgG was also
spotted.
Incubation, washing and regeneration were performed in an automated way
using liquid handling procedures (LHPs) in the instrument for biomolecular
interaction sensing (IBIS iSPR, IBIS Technologies Hengelo, The Netherlands).
A serum sample plug of 400 ml (diluted 1:50 in PBS, 0.03% Tween 20) was
guided backwards and forwards over the array in a ow cell at a rate of 1 ml s
1
.
The serum sample plug was surrounded by two air plugs to prevent the
diffusion of serum components into the buer. Between all steps the ow cell
was rinsed with PBS, 0.03% Tween 20. The array was regenerated by injection
of 400 ml of 10 mM glycineHCl (pH 1.5) twice for 30 s. Two incubation
regeneration cycles were completed before applying the sera in order to block
non-specic binding sites and create an optimal reactive sensor surface. Anal-
ysis of the data were done using the software supplied.
7.6.2 Results and Discussion of Monitoring Analysis Cycles for
Autoantibody Screening
In our proof-of-principle experiments, a 24-spot microarray containing human
IgG, and also two different linear citrullinated peptides I and II and the
corresponding control peptides (containing arginine instead of citrulline), were
spotted (1 nl per spot) on the sensor surface. Because none of the peptides
contained lysine residues, coupling is expected to occur exclusively via the
N-terminal primary amino groups, thereby ensuring oriented, end-on immo-
bilization of the peptides. One set of peptides was synthesized with a C-terminal
biotin tag. These peptides were used to investigate possible differences in
immobilization efciency by assessing the SPR angle shift that resulted from
incubation of the array with an anti-biotin antibody. The corresponding SPR
angle shifts were 211 10 m1 and 212 12 m1 (m1 milli degree) for the
citrullinated and arginine-containing peptide I, respectively. From these data,
it can be concluded that there was no difference in immobilization efciency
between the two peptides.
After placing the spotted sensor chip on top of the hemisphere prism in the
ow cell-based instrument, self-dened liquid handling procedures (LHPs) were
236 Chapter 7
used to increase the inter-experiment reproducibility. Serum incubation, wash-
ing and regeneration were done in an automated manner. In the instrument an
angle scan including the capture of up to 80 images is carried out in 8 s. When a
spot meets the optimal SPR conditions, the reected light at a certain ROI
within this spot will reach a minimum value resulting in a dark spot. Images of
the reected light at three different scan angles after incubation with serum for
1 h are shown in Figure 7.9. The SPR dip of each ROI can be visualized in a
reectivity versus angle plot as shown. The 0 m1 setting is an arbitrary value,
which can be set before each individual experiment by manual adjustment over
a range of 101 to obtain the best angle range. At an incident angle of 700 m1
the background area near the array is in resonance (left image), at 200 m1 the
spots coated with the arginine control peptides are in resonance (middle image)
and at 100 m1 the spots coated with citrullinated peptide I are in resonance
(right image). By curve tting of the reected light intensities as a function of
scan angles for each spot of the microarray, the exact value of the SPR dip
angle in m1 can be calculated and plotted in a sensorgram (Figure 7.10). As
described above, the y-axis of the sensorgram does not represent an arbitrary
reectivity parameter, but contains the exact values for the SPR angle at
maximum resonance. These exact SPR angles can be normalized for easy
comparison of all the different sensorgrams of all the spots and are linearly
correlated with the refractive index, corresponding to the mass of protein
bound to the sensor surface.
Figure 7.10 shows the sensorgram of an analysis cycle obtained during
incubation of the microarray with a RA serum. ROIs in a background region
nearby the array and ROIs within the spots of immobilized human IgG and
of two arginine-containing peptide I and II controls showed relatively small
angle shifts. Binding to the two corresponding citrullinated peptides I and II,
however, resulted in angle shifts of 400 and 250 m1, respectively. The intra-array
variation was very low, as illustrated by the sensorgrams of the quadruplicate
interactions shown in Figure 7.10. The noise, i.e. the baseline difference in
resonance angle between the highest and the lowest values, measured over 100
time points (1000 s) in one individual ROI, was 1.35m1, making angle shifts of
3 m1 significant.
In Figure 7.11, the repeatability of the serum interactions with the microar-
ray by an automated liquid handling procedure is shown. At the start of the
sensorgram, two injections of a 1% glycerol solution were made for sensor
calibration purposes. The sensor chips could be used for up to 50 analysis cycles
by treatment of the sensor surface with two repetitive incubations with 10 mM
glycineHCl (pH 1.5) for 30 s after each serum incubation step. Furthermore,
alternating injections of normal sheep serum and three RA patient sera were
performed. After completing these six serum incubations, the total protocol
was repeated with the same injection sequence to determine the stability of the
array. Sequential measurements of patient sera on a single spot showed
variations of less than 5% in binding to the citrullinated peptide, even when
six interactionregeneration cycles were performed between the two sequential
measurements. At the end the sensor surface was again calibrated with two
Figure 7.9 SPR curves obtained from the microarray by scanning the angles in a range
of 3.41. The three reectivity images are from the same array, but obtained at
different angle positions as shown by the arrows. After the procedure of
angle scanning in steps of 50m1 a reectivity image of the surface is obtained
and gray values of the region of interests are stored in the database. By
viewing the image at different angle positions after, e.g., a spotting procedure,
the quality of the image can be checked for optimal resonance conditions. If,
e.g., a spot is not homogeneous then the depth and width of the SPR curve
becomes higher and broader. Here three images of the microarrays are
shown where in the rst image the background is in resonance. The middle
image shows the spots of peptide II in resonance after interaction with serum
antibodies of an RA patient and the image on the right shows the spots at the
angle position where spots with peptide I are in resonance after interaction
with serum compounds. For details of the spots and specic interaction with
autoantibodies, see Section 7.7. Gray values of ROIs are averaged and
determined for each angle position. The SPR curves from each ROI are tted
in order to calculate the exact minimum.
238 Chapter 7
F
i
g
u
r
e
7
.
1
0
S
e
n
s
o
r
g
r
a
m
o
f
2
4
m
u
l
t
i
p
l
e
a
n
a
l
y
s
i
s
c
y
c
l
e
s
o
f
R
A
p
a
t
i
e
n
t
a
n
t
i
b
o
d
i
e
s
w
i
t
h
p
e
p
t
i
d
e
s
i
m
m
o
b
i
l
i
z
e
d
o
n
t
h
e
S
P
R
s
e
n
s
o
r
c
h
i
p
.
A
f
t
e
r
t
h
e
a
s
s
o
c
i
a
t
i
o
n
p
h
a
s
e
(
A
)
,
t
h
e
a
r
r
a
y
w
a
s
w
a
s
h
e
d
w
i
t
h
P
B
S
,
0
.
0
3
%
T
w
e
e
n
2
0
(
B
)
a
n
d
r
e
g
e
n
e
r
a
t
e
d
w
i
t
h
t
w
o
s
e
q
u
e
n
t
i
a
l
s
t
e
p
s
o
f
1
0
m
M
g
l
y
c
i
n
e
H
C
l
(
p
H
1
.
5
)
f
o
r
3
0
s
(
C
)
.
B
e
t
w
e
e
n
t
h
e
s
e
q
u
e
n
t
i
a
l
g
l
y
c
i
n
e
H
C
l
s
t
e
p
s
t
h
e
a
r
r
a
y
w
a
s
r
i
n
s
e
d
w
i
t
h
P
B
S
,
0
.
0
3
%
T
w
e
e
n
2
0
(
D
)
.
B
l
u
e
(
u
p
p
e
r
)
a
n
d
d
a
r
k
g
r
e
e
n
l
i
n
e
s
(
l
o
w
e
s
t
)
a
r
e
t
h
e
q
u
a
d
r
u
p
l
i
c
a
t
e
s
o
f
c
i
t
r
u
l
l
i
n
a
t
e
d
p
e
p
t
i
d
e
I
a
n
d
t
h
e
c
o
r
r
e
s
p
o
n
d
i
n
g
a
r
g
i
n
i
n
e
c
o
n
t
r
o
l
,
r
e
s
p
e
c
t
i
v
e
l
y
;
l
i
g
h
t
g
r
e
e
n
(
s
e
c
o
n
d
u
p
p
e
r
)
a
n
d
p
u
r
p
l
e
l
i
n
e
s
a
r
e
t
h
e
q
u
a
d
r
u
p
l
i
c
a
t
e
s
o
f
c
i
t
r
u
l
l
i
n
a
t
e
d
p
e
p
t
i
d
e
I
I
a
n
d
t
h
e
c
o
r
r
e
s
p
o
n
d
i
n
g
a
r
g
i
n
i
n
e
c
o
n
t
r
o
l
,
r
e
s
p
e
c
t
i
v
e
l
y
.
R
e
d
l
i
n
e
s
a
r
e
f
r
o
m
t
h
e
h
u
m
a
n
I
g
G
(
n
8
)
a
n
d
t
h
e
b
r
o
w
n
l
i
n
e
s
a
r
e
f
r
o
m
r
e
g
i
o
n
s
c
l
o
s
e
t
o
t
h
e
a
r
r
a
y
(
b
l
a
n
k
c
o
n
t
r
o
l
s
)
(
n
8
)
.
T
h
e
e
n
l
a
r
g
e
d
s
e
c
t
i
o
n
s
h
o
w
s
t
h
e
f
o
u
r
i
n
d
i
v
i
d
u
a
l
c
u
r
v
e
s
f
r
o
m
t
h
e
q
u
a
d
r
u
p
l
i
c
a
t
e
s
o
f
c
i
t
r
u
l
l
i
n
a
t
e
d
p
e
p
t
i
d
e
s
I
a
n
d
I
I
.
injections of a 1% glycerol solution. One can observe that the baseline increases
by 20 m1 after 12 injections.
In Figure 7.12, a typical sensorgram (raw data) is shown of seven analysis
cycles of different RA sera of three RA patients and control sera without anti-
CCP antibodies and normal control serum. In addition, the responses of an
SLE patient serum and an osteoarthritis patient serum are shown. We tested 50
RA sera and 29 control sera (9 SLE patient sera; 10 osteoarthritis patient sera;
10 normal healthy control sera). The interaction of sera with the microarray
was quantied by calculating the ratio between the angle shifts observed for the
citrullinated peptide I and the corresponding arginine control peptide (hereafter
designated C/R ratio) upon binding of serum antibodies. The mean C/R ratio
obtained with scanning SPR microarray imaging for the RA sera which were
tested positive in a CCP2-ELISA (RA CCP+) is 8.7. The C/R ratio for the
RA-sera that were tested negative in the CCP2-ELISA (RA CCP), normal
Figure 7.11 Sensorgram (raw data) is shown showing the repeatability of serum
interactions with a sensor array using an automated liquid handling
procedure. At the beginning and at the end of the sensorgram a
calibration cycle (Z) is performed for sensor calibration purposes by
duplicate injections of a 1% glycerol solution (delta 160 m1). Then 12
analysis cycles from A to L were performed with subsequent alternating
injections of a negative sheep serum (samples A, C, E, G, I and K) and
positive RA patient samples (B, D, F, H, J and L). The positive sample
was in three concentrations in serial dilution [analysis cycle (B, H); 1/100
and (D, J) 1/200 diluted and (F, L); 1/400 times diluted]. One can observe
that the baseline increased by 20 m1 after 12 injections.
240 Chapter 7
controls, SLE patients and OA patients where between 0.9 and 1.2. A detailed
study of these 79 patient sera with ELISA and SPR test results were published
recently [27].
The currently most widely applied target for the detection of anti-citrullinated
protein antibodies, cyclic citrullinated peptides, are recognized by only about
70% of RA patients [23]. Due to the heterogeneity of the anti-citrullinated
protein response in RA, the use of additional citrullinated peptides may allow
the detection of such antibodies in patients that are not reactive with the
peptides used in the CCP2 ELISA. The use of microarrays monitored by SPR
imaging will facilitate the simultaneous detection of the various anti-citrullinated
protein antibodies.
7.7 Features and Benets of Monitoring Analysis
Cycles with SPR Imaging
Compared with xed-angle based SPR imaging systems, the scanning mode of
the IBIS system measures the relevant angle of the SPR dip, permitting valid
comparison of ROIs and correction for bulk effects. Due to the integrated
liquid handling system in combination with the XYZ robot, it is possible to
run the experiments automatically.
Figure 7.12 SPR sensorgram (raw data) of seven analysis cycles of different RA sera
and control sera with CCP and its negative control. Between different
sera the microarray was regenerated with 10 mM glycineHCl (pH 1.5).
(A) RA patient serum without anti-CCP antibodies; (B) SLE patient
serum; (C) RA patient 1 serum; (D) normal sheep serum control; (E) RA
patient 2 serum; (F) osteoarthritis patient serum; (G) RA patient 3
serum.
Although in the past many autoantigens have been identied and characterized,
to date most of the assays that have been developed to detect autoantibodies are
ELISA based, thus allowing only separate analyses for each type of autoanti-
bodies. The need for multiplex analysis systems has been emphasized by recent
observations that the specicity of the detection of anti-CCP antibodies in
RA patient sera can be increased by monitoring reactivities with both the
citrulline- and arginine-containing peptides in parallel [26]. In all of the multi-
analyte autoantibody studies, a secondary antibody conjugate was necessary to
visualize bound antibodies. Here, we show that SPR imaging of protein/peptide
microarrays provides a method that allows one-step multi-analyte detection of
autoantibodies in patient sera, which does not require additional reagents
to visualize antibody binding. Additional advantages of this system are that
the SPR dip angle scanning principle allows accurate and simultaneous monitor-
ing of biomolecular interactions between molecules of varying masses and that
the ligand-containing sensor chips can be efciently regenerated and re-used.
Although the sensitivity of detecting RA-specic autoantibodies by the scan-
ning SPR microarray imaging system under the conditions applied in this study
is slightly lower than that of the ELISA systems, we expect that further
optimization will lead to similar sensitivities. Moreover, for the low-titered
sera a signal amplication step, e.g. by using a (gold-labeled) secondary antibody,
may raise the sensitivity to levels that allow positive signals for all anti-CCP
autoantibody-containing sera.
The initial results obtained with 50 rheumatoid arthritis patient sera (and 29
control sera) proves that SPR imaging is a promising, multi-analyte label-free
technique for the development of diagnostic tests. Although kinetic evaluation
of the analysis cycle including the determination of the thermodynamic
equilibria is beyond the scope of this chapter (see Chapters 4 and 5), the
application of measuring analysis cycles of microarrays has been demonstrated
for RA patient autoantibody detection. In order to quantify the response of RA
patient sera, the ratio of the binding to the citrulline- and arginine-containing
peptides is used. An accurate calculation of this ratio is obtained only by
comparison of the SPR angle shifts (and not reectivity) measured in the
scanning SPR microarray imaging system.
7.8 Conclusion
In this chapter, the analysis cycle of a biomolecular interaction has been
discussed, and exemplied by experiments using an SPR imaging instrument.
The preferred assay format is a direct assay described by A+B#AB,
while other formats can be used for specic cases generally where the molecular
weight of the analyte is too low for direct detection. The challenge of the
small intrinsic refractive index label of the analyte can be circumvented by
using alternative assay formats: competition, inhibition or sandwich. Real-time
monitoring of the binding of biomolecules using SPR imaging allows the
analysis of association and dissociation rate constants for determination of
242 Chapter 7
the afnity constants of many biomolecular interactions. The imaging
feature of the scanning SPR microarray technology is of importance to check
the quality of the sensor surface and to carefully identify the regions of
interest in order to obtain high reproducibility. In addition, the accurate
measurement of the SPR dip angle for monitoring the presence of autoanti-
bodies in sera of autoimmune patients allows the comparison of each curve
(e.g. citrulline vs. arginine), including subtraction of the common mode
effect (i.e. bulk shift jumps). Scanning SPR microarray imaging will be of
great use in any eld that requires multi-analyte detection in an automated
analysis cycle.
7.9 Questions
1. After spotting a microarray of peptide ligands on a sensor surface for SPR
imaging, it is wise to also immobilize a high molecular weight ligand.
Why?
2. If a spot or selected ROI is not homogeneous, it will affect the shape of the
SPR curve. How? Draw possible SPR dip curves.
3. Ligands can be spotted on the sensor surface in different concentrations.
However, the degree of immobilization caused by diffusion of the ligand
to the surface is mass transport controlled. How is spotting of ligands to
the surface in exact serial dilution possible?
4. Inspection of the surface at several angles is a strong feature of scanning
angle imaging instruments. Describe the benets.
5. In SPRI instruments, ligands can be diluted over the sensor surface,
reference spots can be dened, positive and negative controls can be
spotted and the ratios can be determined. Baseline correction, subtraction
and ratio determinations can be applied in the kinetic evaluation process
as shown in this chapter. Explain why scanning angle SPR imaging
instruments show in principle higher reproducibility than xed-angle
SPR imaging instruments. Explain why reectivity data (y-axis in a
sensorgram) plotted as a function of time introduces artifacts. How can
these artifacts be corrected?
7.10 Acknowledgement
The authors thank The Dutch Technology Foundation STW for funding
projects TMM 6209, Proteomics on a Chip, and TMM 6635, Screenchip.
References
1. B. Liedberg, C. Nylander and I. Lundstro m, Sens. Actuators, 1983, 4, 299.
2. J.M. McDonnell, Curr. Opin. Chem. Biol., 2001, 5, 572.
3. R. Karlsson and A. Fa lt, J. Immunol. Methods, 1997, 200, 121.
4. B. Johnsson, S. Lo fa s and G. Lindquist, Anal. Biochem., 1991, 198, 168.
5. D.G. Myszka, J. Mol. Recognit., 1999, 12, 279.
6. K. Nagata, Significance of the real-time analysis of biological interactions,
in Real-Time Analysis of Biomolecular Interactions, K. Nagata, H. Handa,
(Eds.), Springer, 2000, pp. 39.
7. D. Wild (Ed.), The Immunoassay Book, Stockton Press, New York,
1994.
8. N. Murai and M. Yoshida, Molecular Chaperones in Part 4 Chapter 1, in
Real-Time Analysis of Biomolecular Interactions, K. Nagata, H. Handa
(Eds.), Springer, 2000, pp. 8795.
9. G.A.J. Besselink, R.P.H. Kooyman, P.J.H.J. van Os, G.H.M. Engbers and
R.B.M. Schasfoort, Anal. Biochem., 2004, 333, 165167.
11. C.E.H. Berger, T.A.M. Beumer, R.P.H. Kooyman and J. Greve, Anal.
Chem., 1998, 70, 703706.
1999, 71, 3928.
13. E. Stenberg, B. Persson, H. Roos and C. Urbaniczky, J. Colloid Interface
Sci., 1991, 143, 513526.
14. N.J. de Mol, E. Plomp, M.J. Fischer and R. Ruijtenbeek, Anal. Biochem.,
2000, 279, 61.
15. O. Gutmann, R. Niekrawietz, R. Kuehlewein, C.P. Steinert, S. Reinbold,
B. de Heij, M. Daub and R. Zengerle, Analyst, 2004, 129, 835840.
16. B. de Heij, M. Daub, O. Gutmann, R. Niekrawietz, H. Sandmaier and
R. Zengerle, Anal. Bioanal. Chem., 2004, 387, 119122.
17. R. A. Mageed, in Manual of Biological Markers of Disease, W. J.
van Venrooij, R. N. Maini (Eds.), Kluwer, Dordrecht, 1994, Chapter
B1.1, pp. 127.
18. J.G. Routsias, A.G. Tzioufas and H.M. Moutsopoulos, Clin. Chim. Acta,
2004, 340, 1.
19. F.A. van Gaalen, J. van Aken, T.W.J. Huizinga, G.M.Th. Schreuder, F.C.
Breedveld, E. Zanelli, W.J. van Venrooij, C.L. Verweij, R.E.M. Toes and
R.R.P. De Vries, Arthritis & Rheum., 2004, 50, 21132121.
20. W.H. Robinson, C. DiGennaro, W. Hueber, B.B. Haab, M. Kamachi, E.J.
Dean, S. Fournel, D. Fong, M.C. Genovese, H.E. Neuman de Vegvar,
K. Skriner, D.L. Hirschberg, R.I. Morris, S. Muller, G.J. Pruijn, W.J. van
Venrooij, J.S. Smolen, P.O. Brown, L. Steinman and P.J. Utz, Nat. Med.,
2002, 8, 295.
21. S.P. Fodor, J.L. Read, M.C. Pirrung, L. Stryer, A.T. Lu and D. Solas,
Science, 1991, 251, 767773.
22. R.J. Fulton, R.L. McDade, P.L. Smith, L.J. Kienker and J.R. Kettman
Jr., Clin. Chem., 1997, 43, 17491756.
23. M.A. Cooper, Anal. Bioanal. Chem., 2003, 377, 834.
24. G.A. Schellekens, B.A. de Jong, F.H. Van den Hoogen, L.B. Van de Putte
and W.J. van Venrooij, J. Clin. Invest., 1998, 101, 273281.
244 Chapter 7
25. W.J. van Venrooij, A.J. Zendman and G.J.M. Pruijn, Autoimmun. Rev.,
2007, 6, 3741.
26. A. Vannini, K. Cheung, M. Fusconi, J. Stammen-Vogelzangs, J.P. Drenth,
A.C. DallAglio, F.B. Bianchi, L.E. Bakker-Jonges, W.J. van Venrooij,
G.J.M. Pruijn and A.J.W. Zendman, Ann. Rheum. Dis., 2007, 66, 511516.
27. A.M.C. Lokate, J.B. Beusink, G.A.J. Besselink, G.J.M. Pruijn, R.B.M.
Schasfoort, J. Am. Chem. Soc., 2007; DOI: 10.1021/ja075103X.
CHAPTER 8
Advanced Methods for SPR
Imaging Biosensing
ALASTAIR W. WARK, HYE JIN LEE AND
ROBERT M. CORN
Department of Chemistry, University of CaliforniaIrvine, Irvine,
CA 92697, USA
8.1 Introduction
Microarray biosensors have become an invaluable biotechnological tool for the
rapid, multiplexed detection of surface bioanity interactions. For example,
nucleic acid microarrays are currently being applied to the areas of genomics
[1,2], genetic testing [3], gene expression [4,5], and single nucleotide polymorphism
(SNP) genotyping [6,7]. Many researchers are also interested in developing
protein microarrays for application in the areas of proteomics [810] and drug
discovery [11,12]. In addition, the detection and proling of multiple protein
biomarkers in biological uids (e.g. blood, serum, urine) by antibody microarrays
is a potentially powerful method for the diagnosis of diseases and the monitoring
of subsequent therapeutic treatments [13,14].
An attractive alternative to traditional uorescence-based microarray detection
methods is the surface-sensitive optical technique of surface plasmon resonance
imaging (SPRI). SPRI, also denoted SPR microscopy [1517], was originally
applied to the study of surface morphology in phospholipid monolayers and other
thin surface lms [15]. Since those initial eorts, SPRI has evolved into a primary
method for the measurement of bioanity adsorption onto biopolymer micro-
arrays by detecting changes in the local refraction index upon binding [1831].
This is a signicant advantage in the analysis of biological samples where labeling
multiple biomarkers with uorophores or nanoparticles is often not possible.
Over the last decade, a number of improvements have been made in SPRI in
terms of instrumentation, surface microarray fabrication and microuidic formats
for the multiplexed measurement of surface adsorption kinetics with microarrays.
A brief review of some of these advances is given in Section 8.2.
246
While multiplexed microarray analysis of bioanity interactions is a valuable
research tool, even more specicity and sensitivity can be obtained by coupling
the bioanity process to an enzymatic transformation. This coupling is often
employed in solution-phase biotechnological processes; for example, coupling of
DNA hybridization with polymerase chain reaction (PCR) leads to the process
of PCR amplication in genomic DNA samples [32,33]. However, the direct
incorporation of solution enzymatic methods such as PCR into a parallel
microarray format is dicult, because any intermediate solution species will
diuse on to neighboring array elements. Instead of using solution enzyme
chemistry, it is better to utilize a surface enzyme reaction on biomolecules
attached to the microarray surface. The use of surface enzyme chemistry ensures
that the parallel nature of the multiplexed microarray assay remains intact. In
Section 8.3, we describe some of the basic equations that we have developed for
the use of surface enzyme kinetics in SPRI, and a new enzymatically amplied
SPRI methodology to detect DNA down to femtomolar concentrations.
Finally, in Section 8.4 we describe how the incorporation of DNA-coated gold
nanoparticles into the enzymatically amplied SPRI methodologies can further
increase the sensitivity of these array bioanity measurements. DNA-coated gold
nanoparticles were originally employed with SPRI by He et al. to detect DNA in a
sandwich assay format at a concentration of 10pM [30]. Using a nanoparticle-
enhanced surface ligation strategy, we demonstrate SNP genotyping with SPRI at
concentrations as low as 1 pM, and with a combination of poly(A) polymerase
and DNA-coated gold nanoparticles, we can detect microRNA in biological
samples down to a concentration of 10fM [34].
8.2 Advances in SPRI Instrumentation and Surface
Chemistry
Since the rst demonstration in 1997 of the use of SPRI and microarrays for
monitoring DNADNA interactions [35,36], our group has continued to
develop this technology focusing in particular on (i) optimizing the SPRI
instrumental set-up and (ii) designing new chemical strategies for the surface
attachment of dierent biomolecular probes in an array format. A schematic
diagram of an SPRI apparatus is shown in the middle inset in Figure 8.1, where
the output from a collimated white light source is passed through a polarizer
and then directed onto the prism/chip assembly at a xed optimal angle. The
reected p-polarized light is then collected via a narrow-band interference lter
centered at 830 nm onto a CCD camera. We have found that this white light
NIR lter combination has several advantages compared with the use of a
visible monochromatic laser beam such as improved sensitivity and the removal
of interference fringes from the SPR image that are often problematic when
using coherent laser excitation [37]. A number of commercial SPR imaging
instruments are now available from GWC Technologies, Biacore (HTS Bio-
systems), IBIS Technologies and GenOptics; for more on instrumental aspects,
see Chapter 3.
247 Advanced Methods for SPR Imaging Biosensing
In conjunction with SPRI instrumentation, the use of well-characterized and
robust surface chemistries to tether biological molecules to gold surfaces in an
array format is extremely important for successful SPRI biosensing measure-
ments. Consequently, we have developed a variety of surface immobilization
strategies for covalently attaching dierent biomolecular probes to chemically
modied gold surfaces [3841]. These attachment chemistries can be used along
with either UV photopatterning [39] or microuidic techniques [42] to create
arrays of multiple, independently addressable elements on a single surface.
Figure 8.1 shows a set of representative SPRI images obtained from various
studies that our group has performed [18,34,4352]. A detailed summary of the
biopolymer microarrays (e.g. DNA, RNA, peptide, protein, carbohydrate)
prepared for dierent biomolecular targets is listed in Table 8.1.
Figure 8.1 A simplied SPRI set-up (middle inset) and representative SPRI dierence
images obtained for the detection of bioanity interactions with biopol-
ymer microarrays. From top center and clockwise, protein biomarker
binding to an antibody microarray, antibody binding to a peptide micro-
array, protein interactions with a His-tagged protein microarray, lectin
binding to a carbohydrate line array, response regulator protein binding to
a dsDNA microarray, 16mer ssDNA hybridization adsorption on an
ssRNA microarray and 16mer ssDNA hybridization adsorption on an
ssDNA microarray. The middle inset shows a simplied SPRI set-up.
Briey, the output from a collimated white light source (L) is passed
through a polarizer (P) with the resulting p-polarized light directed on to a
high index prism/sample assembly at an optimal incident angle. The
reected light from the sample assembly is then collected via a narrow
band pass lter (F) on to a CCD camera (C).
248 Chapter 8
A third area in which SPRI biosensing has improved is in examining the
thermodynamic and kinetic parameters of surface bioanity interactions. A
simple one-step surface bioanity adsorption process involving the specic
binding of a target biomolecule (T) to a surface-attached probe (P) or ligand
can be described by the reaction
T P
k
a
k
d
TP 8:1
Table 8.1 Summary of dierent surface bioanity measurements performed
using SPRI and biopolymer microarrays.
Microarray probes
(on surface)
Target biomolecules
(in solution) K
ads
(M
1
) Ref.
DNA DNA
ssDNA ssDNA (1618 bases) 2.0 10
7
45,53
DNA RNA
ssDNA ssRNA (18 bases) 1.8 10
7
53
ssDNA 16S ribosomal RNA N 86
Messenger RNA 42
DNA Proteins
Biotinylated DNA Streptavidin N 35
ssDNA ssDNA-binding protein N 87
dsDNA Mismatch-binding proteins N 87
dsDNA Phosphorylated OmpR 1.6 10
8
50
dsDNA Phosphorylated VanR 6.6 10
6
50
Peptides Antibodies and proteins
Flag-peptide Anti-Flag 1.5 10
8
51
S-peptide S-protein 1.7 10
7
52
Carbohydrates Lectins
Mannose Concanavalin A 2.2 10
7
88
Galactose Jacalin 5.6 10
6
88
Proteins Antibodies and DNA
His-tagged ubiquitin Anti-ubiquitin N 89
His-tagged Flag Anti-Flag N 89
His-tagged RFP Anti-RFP N 89
His-tagged TATA box-
binding protein
dsDNA N 89
Antibodies Protein biomarkers
Anti-b
2
-microglobulin b
2
-Microglobulin 1.4 10
8
43
Anti-cystatin C Cystatin C 1.0 10
8
43
RNA Proteins
RNA aptamer factor IXa 1.6 10
7
71
RFP, red uorescent protein; ssDNA, single-stranded DNA; dsDNA, double-stranded DNA; N,
not measured with SPRI.
where TP is the surface-bound targetprobe complex and the rates of T adsorp-
tion and desorption are dened by k
a
and k
d
, respectively (see Figure 8.2a). An
example of such a reaction would be DNA hybridization adsorption on a DNA
microarray. Figure 8.2b shows an SPRI dierence image of a four-component
ssDNA microarray following exposure to two complementary 16mer DNA target
sequences. A positive increase in percentage reectivity (D%R) is observed at only
the perfectly matched array elements due to the formation of duplexes via
hybridization adsorption. Fresnel calculations and experimental evidence show
that if the SPRI response remains below 10%, then it is directly proportional to
the relative surface coverage (y) of complementary DNA [53], where y G
TP
/G
tot
and G represents the molecular surface density. At equilibrium, the fractional
surface coverage reaches a steady state and this equilibrium surface coverage (y
eq
)
is given by the Langmuir adsorption isotherm:
y
eq

K
ads
T
1 K
ads
T
8:2
Figure 8.2 (a) Schematic showing target ssDNA (T) hybridization adsorption on
ssDNA microarray elements (P). (b) SPRI dierence image obtained for
sequence-specic 16mer target ssDNA hybridization adsorption on
ssDNA microarray elements. (c) A representative plot of relative surface
coverage (y) as a function of target DNA concentration. The solid line
is a Langmuir isotherm t to the data, from which a value of
K
ads
2 10
7
M
1
was determined.
250 Chapter 8
where the Langmuir adsorption coecient (K
ads
) is dened as K
ads
k
a
/k
d
.
Figure 8.2c shows a typical Langmuir isotherm plot of y versus concentration for
16mer DNA from which a value of K
ads
210
7
M
1
was obtained [45,53]. At
low surface coverages, the Langmuir isotherm depends linearly on DNA con-
centration and eq. (8.2) becomes y
eq
K
ads
[T]. In DNA hybridization adsorp-
tion, the lowest DNA concentration observed with SPRI was 1 nM, which
corresponds to a y
eq
of 0.02.
In addition to equilibrium measurements, further information on surface
bioanity interactions can be obtained by measuring the kinetic parameters k
a
and k
d
. For the case of adsorption from a solution of concentration [T] on an
unoccupied surface, the time-dependent fractional surface coverage, y(t), is
given by
yt y
eq
1 e
k
a
Tk
d
t
8:3
where y
eq
is the equilibrium value for y at a particular bulk concentration, as
given using eq. (8.2). For the case when y 1 at t 0, the desorption rate can be
described by
yt y
eq
e
k
d
t
8:4
Equations (8.3) and (8.4) have been used frequently to analyze the adsorp-
tion of biomolecules on surfaces, especially with SPR [54] and SPRI [52].
Compared with single-channel SPR measurements, real-time SPRI has a great
advantage in being able to determine simultaneously the rates of target
adsorption/desorption at multiple dierent probe elements on a single chip.
Figure 8.3a shows the design of a poly(dimethylsiloxane) (PDMS) micro-
channel ow cell that facilitates well-controlled and reproducible sample
delivery to each array element in addition to signicantly reducing the sample
volume (B10 ml). SPR imaging kinetic experiments are performed using a
continuous ow-through microchannel to prevent mass transport limitations,
whereas equilibrium measurements are obtained under stopped-ow conditions
using a larger volume ow cell (B100 ml). Additionally, a specially designed
water-jacketed ow cell, which controls the system temperature to within
0.1 1C, is used to perform temperature-dependent studies. We have applied
real-time SPRI to determine accurately the k
a
and k
d
of S protein binding to an
array composed of various S peptide analogues [52]. For this study, custom-
written software that rapidly acquires and organizes large data sets during real-
time acquisition was used (see Figure 8.3b). These real-time SPRI methods have
also been employed to characterize quantitatively our surface enzyme methods
described in Section 8.3.
Finally, we have also continued to pursue the development of improved
SPRI biosensing methods; one example is the design of novel multiple-layered
chip structures that support the generation of long-range surface plasmons
(LRSPs) [45,5558]. Essential to the LRSPR chip design shown in Figure 8.4 is
the use of Cytop as an inert optically transparent material whose refractive
index is very close to that of water, with the depth and position of the LRSPR
Figure 8.3 (a) SPRI raw image of a peptide array created from parallel PDMS
microuidic channels to deliver heterobifunctional cross-linker and probe
molecules before being replaced by a second serpentine PDMS channel to
create a continuous ow cell for use in kinetics measurements. The dotted
lines indicate regions of interest (ROIs) on the array where the change in
percent reectivity is measured as a function of time. (b) Simultaneous
real-time SPRI measurements obtained for each ROI [peptide elements
and poly(ethylene glycol) background] when the array was rst exposed to
a 150 nM solution of S protein and then rinsed with phosphate buer
only. The SPRI signal increases then decreases in response to protein
adsorption/desorption. The adjacent poly(ethylene glycol) regions are
used to correct the SPRI signal for background eects. Reprinted with
permission from reference 52.
252 Chapter 8
Figure 8.4 (a) SPR reectivity curves of p-polarized light as a function of incident
angle for a long-range SF10 glass/Cytop (1180 nm)/Au (32 nm)/water
conguration (J) and a conventional SF10 glass/Au (45 nm)/water SPRI
conguration (B) at excitation wavelength 814 nm. The solid lines show
the results of a theoretical t to the data using a multiple phase Fresnel
calculation. Refractive indices (n) used in the calculation: n(SF10) 1.711,
n(Cytop) 1.336, n(Cr) 3.186 +3.47i, n(Au) 0.185 +5.11i and
n(H
2
O) 1.328. (b) Calculated relative optical eld intensities for the
LRSPR and conventional SPR assemblies whose schematics are shown as
insets. The conventional SPR eld intensity is multiplied by 100 for
visibility. Reprinted with permission from reference 45.
mode strongly dependent on the thickness of both the Cytop and gold lms.
Compared with conventional surface plasmons, LRSPs possess longer surface
propagation lengths, higher electric eld strengths and narrower resonance
curves. This is demonstrated in Figure 8.4a, which shows in situ scanning angle
reectivity measurements obtained for both a conventional SPRI chip design
[SF10/gold (45 nm)/water] and that of a LRSPR chip [SF10 prism/Cytop
(1180 nm)/gold (32 nm)/water] [45].
To demonstrate the advantages of LRSPR, DNA microarrays were prepared
on both conventional and long-range chips. Repeated measurements of DNA
hybridization adsorption showed an B20% increase for the LRSPR D%R
signal compared with regular SPRI chips. Perhaps of more interest is the much
higher surface electric eld strengths associated with the generation of LRSPs.
Figure 8.4b displays the results of n-phase Fresnel calculations [59,60] which
demonstrate that the LRSP electric eld intensity at the gold sensing surface is
higher by as much as 10
3
compared with that associated with conventional
surface plasmons. These enhanced elds should lead to further research
advances in surface plasmon uorescence spectroscopy (SPFS) studies [61].
8.3 Surface Enzymatic Transformations for Enhanced
SPRI Biosensing
To enhance further the biosensing capabilities of SPRI and overcome some of
the diculties associated with detecting very low target concentrations
(o1 nM), we have developed a number of advanced methodologies which
couple surface enzyme reactions and bioanity interactions on biopolymer
microarrays [18,34,46,48,49]. In this section, we rst outline a simple theoretical
framework that quantitatively describes the catalysis reaction of an enzyme in
bulk solution with a surface immobilized substrate. The application of this
model to analyze real-time SPRI and SPFS measurements of RNase H surface
activity is also discussed. Next, we provide two examples of surface enzymatic
processes on nucleic acid microarrays that are coupled with surface bioanity
interactions to enhance the sensitivity and selectivity of the biosensor. These
processes are (i) the enzymatically amplied detection of genomic DNA using
RNase H and RNA microarrays and (ii) the application of surface ligation
chemistry for the fabrication of RNA microarrays from DNA microarray
templates for the study of proteinRNA aptamer interactions.
8.3.1 Measuring Surface Enzyme Kinetics
The quantitative characterization of the surface enzyme reactions utilized in
advanced SPRI biosensing methods is extremely important, as an enzyme can
react orders of magnitude slower on a surface as compared with in solution.
Here, we describe a simple kinetic model that includes enzyme adsorption,
desorption and the surface enzymatic reaction and apply it to the analysis of real-
time SPRI and SPFS measurements of the surface hydrolysis of RNADNA
254 Chapter 8
heteroduplexes
1
by RNase H. This surface reaction (see Figure 8.5 inset) forms
the basis of our enzymatic amplication methodology described in Section 8.3.2.
For the case of a 1:1 binding of an enzyme molecule (E) to a surface-immobilized
substrate (S), we have identied three processes that control the overall reaction
rate:
E
xN
!
k
m
E
x0
8:5
S E
x0

k
a
k
d
ES 8:6
ES !
k
cat
S
E
x0
8:7
Figure 8.5 A series of real-time SPRI data curves obtained for the RNase H hydrol-
ysis of surface RNADNA (R1-D1) array elements at dierent concen-
trations of RNase H solutions [0.5 nM (K), 1.0 nM (), 2.0 nM (m) and
4.0 nM (E)]. The experimental curves were globally tted using eqs. (8.8)
(8.10) to obtain the parameter values k
a
3.4 (0.2) 10
6
M
1
s
1
,
k
d
0.10 (0.05) s
1
, k
cat
1.0 (0.1) s
1
and b180 (20). The solid
lines are the tted theoretical kinetic curves. The inset is a schematic
illustration of the RNase H hydrolysis of surface RNADNA hetero-
duplexes. Reprinted with permission from reference 64.
1
A heteroduplex is a double-stranded molecule of nucleic acid composed of two single comple-
mentary strands derived from dierent sources.
where E
(xN)
and E
(x0)
are the bulk and surface enzyme concentrations,
respectively, ES is the surface enzymesubstrate complex, S* is the surface-
bound product and k
cat
is the surface reaction rate for the enzyme complex.
When microuidics are used for solution delivery, the enzyme diusion can be
described by a steady-state mass transport coecient (k
m
) that can also be
written as D/d, where D is the diusion coecient for the enzyme and d is the
steady-state diusion layer thickness [62,63].
The kinetic equations for this reaction scheme can be expressed in terms of
the relative fractional surface coverages of each of the three surface species
(denoted y
x
G
x
/G
tot
, where x S, ES or S*):
y
S
y
ES
y
S
1 8:8
dy
ES
dt

k
a
E1 y
ES
y
S
k
d
k
cat
y
ES
1 b1 y
ES
y
S
8:9
dy
S
dt
k
cat
y
ES
8:10
In eq. (8.9), [E] is the bulk enzyme concentration and b is the dimensionless
diusion parameter [62,63], dened by
b
k
a
G
tot
k
m
k
a
G
tot
d
D
8:11
These equations were derived in a series of recent papers [18,64,65] and can
be solved using simple Euler integration methods with the initial conditions
y
S
1 and y
ES
y
S*
0 at time t 0 to yield three time-dependent surface
coverages y
ES
(t), y
S*
(t) and y
S
(t) that can be separately proled over the course
of the surface reaction.
The kinetic model described above has been applied to the quantitative
analysis of time-resolved SPRI and SPFS measurements of the catalytic
behavior of RNase H on surface-immobilized RNADNA heteroduplexes.
Figure 8.5 shows the D%R loss observed in real-time SPRI measurements due
to the selective removal of RNA from the surface at enzyme concentrations
ranging from 0.5 to 4 nM. The analysis of the SPRI kinetic curves (shown as
solid lines in Figure 8.5) using eqs. (8.8)(8.10) yielded best-t values of
3.4 10
6
M
1
s
1
, 0.1 s
1
, 1.0 s
1
and 180 for k
a
, k
d
, k
cat
and b, respectively.
Also, the analysis showed y
ES
(t) remained very small (B10
3
) throughout the
reaction with k
cat
c k
a
[E]. This means that once adsorbed, RNase H reacts
very quickly and is immediately released from the surface. The observed
k
cat
value of 1.0 s
1
is signicantly faster than k
cat
measured in a similar study
[65] for the surface hydrolysis of double-stranded DNA by Exonuclease
III (k
cat
0.01 s
1
). In this surface reaction, the time-dependent SPRI signal
initially increased before an eventual overall decrease due to the loss of
the complementary strand in the surface DNA duplex. This behavior is
256 Chapter 8
qualitatively very dierent from that of the RNase H reaction in Figure 8.5,
where at no point was a net increase in SPRI signal observed.
Since enzyme adsorption and surface duplex hydrolysis both contribute to
the measured SPRI response, some assumptions on the relative contributions of
both reaction steps are required when using our kinetic model. Therefore, to
characterize the surface enzyme reaction completely it is important to obtain an
independent set of in situ surface kinetic measurements. In the case of RNase H,
this was achieved using uorescently labeled RNA and applying the technique
of SPFS [64]. The loss in SPFS signal due to enzymatic hydrolysis of the surface
attached RNA is a direct measure of y
S
(t) and was analyzed in a similar manner
to the SPRI data for various enzyme concentrations. The k
a
, k
d
and k
cat
best-t
values obtained from the SPFS measurements were almost identical with the
values obtained from the SPRI data described above, indicating that uores-
cence labeling of RNA does not signicantly aect k
cat
.
8.3.2 RNase HAmplied Detection of DNA
The rst major breakthrough in the application of surface enzyme processes to
greatly improve the sensitivity of SPRI bioanity measurements was the use
of the RNase H surface process for the amplied detection of DNA [48,49].
Figure 8.6a shows how this enzymatic amplication methodology works. An
RNA probe microarray is exposed to a solution containing both target DNA
and RNase H. When a target DNA molecule binds to a complementary RNA
probe array element, RNase H recognizes the formed RNADNA heterodu-
plex and selectively hydrolyzes the RNA strand (Step 1). The target DNA is
released back into solution and is free to bind to another surface RNA probe
and RNase H will again hydrolyze the RNA in the surface heteroduplex (Step 2).
This repeated target bindingenzymatic hydrolysistarget release will continue
until all the RNA probe molecules are removed from the surface (Step 3).
Amplication is achieved because only a very small number of target DNA
molecules are required to produce a large change in SPRI signal.
The sensitivity of RNase H amplied SPRI was found to be sucient for the
detection of DNA at concentrations as low as 1 fM. Consequently, we have
been able to directly detect target sequences in a human genomic DNA sample
without PCR amplication. Figure 8.6b shows an SPRI dierence image of a
three-component RNA microarray following 4 hours exposure to a genomic
DNA sample. At the R1 and R2 elements, which contain sequences designed to
bind specically to the TSPY gene on the Y chromosome, a decrease in SPRI
signal (0.7%) was observed
2
, whereas no change in SPRI signal was observed
at either the R3 elements or the array background. The concentration of the
TSPY gene sequences in commercially available male genomic DNA was
estimated to be 7 fM [48]. In addition to human genomic DNA, this enzyma-
tically amplied SPRI method can be applied to the ultrasensitive detection and
identication of DNA and RNA from viruses and bacteria.
2
Note that the minimum change in SPRI reectivity that can be measured is around 0.08%.
Figure 8.6 (a) Schematic outlining enzymatically amplied SPRI methodology using
an RNA microarray for the detection of target DNA. (b) An SPRI
dierence image obtained for the detection of male genomic DNA in the
presence of RNase H. A schematic of the three-component RNA micro-
array is also shown where R1 and R2 are specically designed to bind to
the TSPY gene on the Y chromosome and R3 is a negative control
sequence. Reprinted with permission from reference 48.
258 Chapter 8
8.3.3 Fabrication of RNA Microarrays with RNA-DNA Surface
Ligation Chemistry
A second example of how surface enzyme processes can be applied to enhance
the biosensing capabilities of SPRI is the use of ligation chemistry to create
RNA microarrays from DNA microarrays. The ability to fabricate stable and
active single-stranded RNA (ssRNA) microarrays is essential for a successful
RNase H amplied SPRI measurement and will also promote the use of RNA
microarrays for the study of RNAprotein, RNARNA and other bioanity
interactions. Despite the many potential benets of RNA microarrays, at
present only a handful of reports are available on the fabrication of RNA
microarrays in the literature [48,49,6670]. The use of modied RNA sequences
such as biotinylated RNA [66,69,70] or thiol-modied RNA [48,49] can be time
consuming and costly with also the increased possibility of RNA degradation
during both the modication and surface attachment procedures. To address
these problems, we have recently developed two novel RNA fabrication strat-
egies utilizing two dierent enzymes: (i) T4 DNA ligase and (ii) T4 RNA ligase
[47,71]. Both of these enzymes catalyze the ligation of unmodied RNA onto a
DNA microarray, but only T4 DNA ligase requires the use of a complementary
DNA template sequence [47].
Figure 8.7 outlines a simplied schematic for the creation of an ssRNA
microarray using T4 DNA ligase. A single-stranded (ssDNA) microarray is
rst prepared by chemically attaching 3
0
-thiolated, 5
0
-phosphorylated ssDNA.
These anchor DNA array elements (D
A
) are then exposed to a solution
containing T4 DNA ligase and both probe RNA (R
P
) and template ssDNA
(D
T
), which will hybridize to the D
A
surface. Following the formation of a
phosphodiester bond between the 5
0
-phosphate of D
A
and the 3
0
-hydroxyl
group of R
P
the array surface is thoroughly rinsed with 8 M urea in order to
denature and remove the DNA template and any T4 DNA ligase, resulting in
the creation of biologically active ssRNA array elements. The SPRI dierence
image [b a] in Figure 8.7 shows a D%R increase of 2.2 0.3% following the
ligation of 24mer RNA [47]. We also found that the original ssDNA microar-
ray could be regenerated with RNase H hydrolysis allowing the surface ligation
process to be repeated to create a new RNA microarray. This is because RNase
H specically cleaves the phosphodiester bonds in the RNA component of the
DNA-RNA heteroduplex (formed by complementary DNA, D
C
, hybridiza-
tionadsorption on R
P
) to produce 5
0
-phosphate and 3
0
-OH termini. This
ligation-hydrolysis cycle could be repeated up to three times using the same
ssDNA microarray without any degradation in SPRI signal. In addition, when
the ligated microarray was used for the detection of 1 pM DNA via RNase H
amplication, the initial rate of change in D%R observed was over 20 times
faster than that observed with microarrays prepared using thiol-modied
RNA. This is attributed to an increase in RNase H activity due to the anchor
DNA sequence increasing the distance of R
P
from the gold surface.
A second example of the application of surface ligation chemistry for the
fabrication of ssRNA microarrays is the use of the enzyme T4 RNA ligase,
which does not require a complementary DNA template sequence [71]. To
create a multi-component RNA microarray using this approach, solutions
containing both enzyme and RNA probe were delivered onto individual array
DNA elements via spotting. Following completion of the ligation reaction, the
microarray was rinsed with 8 M urea to remove any enzyme and non-ligated
ssRNA from the surface. As outlined in Figure 8.8a, the prepared bioactive
ssRNA array elements can then be directly applied for the study of RNA
protein bioanity interactions. Figure 8.8b shows an SPRI dierence image
obtained for the adsorption of 20 nM human factor IXa (fIXa) on a three-
component RNA aptamer microarray. The array pattern is shown in Figure
8.8c, where R
F
has a strong binding anity for fIXa with both R
C
and R
R
aptamer variants acting as negative controls. In addition to the screening of
dierent aptamer sequences, these ligated RNA microarrays can also be
applied to the ultrasensitive detection of protein biomarkers [71].
8.4 Nanoparticle-amplied SPRI Biosensing
A second method for increasing the sensitivity of SPRI bioanity measure-
ments by several orders of magnitude is the use of functionalized gold
Figure 8.7 Simplied schematic of the fabrication of a renewable ssRNA microarray
via RNADNA ligation chemistry followed by RNase H hydrolysis. The
ssRNA microarray was created by the selective ligation of RNA (R
P
) to
D
A
elements of a two-component DNA microarray in the presence of a
DNA template (D
T
). Hybridization of complementary DNA (D
C
) on to
the ligated ssRNA followed by the selective hydrolysis of R
P
using RNase
H regenerates the original 5
0
-phosphorylated ssDNA surface. The right
inset shows a representative in situ SPRI dierence image [b a] obtained
by subtracting images acquired before and after RNA ligation. Reprinted
with permission from reference 47.
260 Chapter 8
nanoparticles, an approach which is also completely compatible with the use of
surface enzyme chemistry. Nanoparticle-amplied SPRI detection of DNA was
originally demonstrated by He et al. who used DNA-modied nanoparticles to
achieve a detection limit of 10 pM in a sandwich assay format [30]. Although
the dependence of the SPR response as a function of both nanoparticle size and
the nanoparticleplanar lm separation distance has been investigated by
several groups [7274], nanoparticles are not yet routinely applied for enhanced
SPRI bioanity sensing. In this section, two novel methodologies combining
surface enzyme reactions and nanoparticle-enhanced SPRI measurements are
Figure 8.8 (a) Schematic showing the creation of an RNA aptamer microarray via
RNADNA surface ligation chemistry using T4 RNA ligase. DNA and
RNA bases are shown as open circles and lled circles, respectively. Step
(i): microarray elements composed of 5
0
-phosphorylated ssDNA are
individually exposed to solutions containing ssRNA aptamer molecules
and T4 RNA ligase. This results in the formation of a phosphodiester
bond between the 3
0
-hydroxyl of the ssRNA and the 5
0
-phosphate of the
ssDNA. Step (ii): after ligation, the surface is rinsed with 8 M urea to
remove the enzyme and non-ligated ssRNA. The ssRNA aptamer micro-
array is then used for the study of proteinaptamer binding events.
(b) SPRI dierence image obtained for the specic adsorption of 20 nM
fIXa on the R
F
array elements of a three-component microarray. (c) A
pattern of a three-component aptamer array where R
F
has a high binding
anity towards fIXa and R
T
and R
R
are negative control aptamer
sequences.
described for (i) the analysis of single nucleotide polymorphisms (SNPs) in
genomic DNA and (ii) the femtomolar detection of microRNAs (miRNAs
3
) in
a total RNA sample. In both studies, DNA-modied gold nanoparticles
(B13 nm diameter) were utilized with a maximum absorbance at 525 nm.
The surface plasmon excitation wavelength in our SPRI measurements is
830 nm. Therefore, the amplication in SPRI signal is primarily due to changes
in refractive index and not absorptive coupling between the gold nanoparticles
and thin gold lm. A theoretical description of nanoparticle SPR can be found
in Chapter 2.
8.4.1 Single Nucleotide Polymorphism Genotyping
The development of new methods for the rapid, multiplexed detection and
identication of single nucleotide polymorphisms (SNPs) in human genomic
DNA samples is attracting major interest as these methods can be used to
accelerate the discovery of specic disease-related mutations and also for large-
scale human genetic variation studies. In principle, SPRI measurement with
DNA microarrays is an ideal candidate technique for SNP genotyping. Various
researchers have employed both conventional SPRI and single-channel SPR
formats to detect single base mismatches in DNA by hybridization adsorption
[24,7780]. However, the SPRI detection limit of 1 nM described in Section 8.2
is insucient for SNP genotyping of unamplied genomic DNA samples where
the detection of single base pairs in a DNA sequence at a concentration of 1 pM
or better is required.
In order to demonstrate that SPRI can be used for multiplexed SNP geno-
typing, we developed a novel approach that involves the sequence-specic ligation
of target DNA to an ssDNA microarray followed by the nanoparticle-amplied
SPRI detection of the surface ligated product. As outlined in Figure 8.9, a
solution containing the 36mer target (T), a 5
0
-phosphorylated ssDNA ligation
probe (L) and the enzyme Taq DNA ligase (E) is introduced to a two-component
ssDNA microarray containing the probes P
G
and P
A
[44]. The two array probe
DNA sequences dier only at their 3
0
-terminal nucleotide (G for P
G
and A for
P
A
). The target DNA molecules hybridize simultaneously to the ligation probes
and the array probes, resulting in the formation of two dierent surface com-
plexes, LTP
G
and LTP
A
. Single base pair selectivity is achieved because Taq
DNA ligase will only catalyze the formation of a phosphodiester bond between
the juxtaposed P and L probes when they are both perfectly complementary to the
hybridized target. Here, the ligation probe will be specically ligated to P
G
but not
to P
A
. Following an 8 M urea wash to remove any target, non-ligated probe and
enzyme from the microarray, SPRI detection of the ligated LP
G
array elements is
achieved through the hybridization adsorption of gold nanoparticles modied
with ssDNA 16mers (L
C
) which are complementary to the ligation probe
sequence.
3
MicroRNAs are a new class of small, non-coding RNA molecules (1923mers) that can regulate
gene expression in both plants and animals [75,76].
262 Chapter 8
Prior to performing experiments using real genomic DNA samples, the
sensitivity and single base specicity of this advanced SPRI methodology was
established using synthetic target oligonucleotides. A four-component array
was designed consisting of three 20mer probe sequences (P
A
, P
C
and P
G
) that
are identical except for the last nucleotide at their 3
0
-termini (A, C and G,
respectively) and also a poly T sequence (P
N
), which serves as a negative
control. Specic ligation of the 16mer L probe to only one of the array
components could be detected for 36mer T concentrations as low as 1 pM [44].
Having established the detection limit of this SNP genotyping method, we
then applied the same microarray design to screen the PCR products of human
genomic DNA samples for a possible point mutation in the BRCA1 gene that is
associated with breast cancer. In this case, probe P
G
is the perfect complement
to the wild-type allele, probe P
C
is the perfect complement to the mutant allele
in NA13710, while probe P
A
is the perfect complement to the mutant allele in
NA14637. Both mutant samples were obtained from the Coriell Institute [44].
The SNP genotyping results for the wild-type and NA14637 samples are shown
in Figures 8.10a and b, respectively. The SPRI dierence image in Figure 8.10a
shows an increase in reectivity only at the P
G
array elements, indicating that
the genotype of the wild-type DNA sample is a CC homozygote. However, in
Figure 8.10(b), a D%R increase was observed at both the P
G
and P
A
array
elements, identifying the NA14637 sample as a CT heterozygote. The
NA13710 sample (data not shown) was identied as a CG heterozygote.
These experiments successfully demonstrate for the rst time that SPRI can be
Figure 8.9 Schematic of SNP genotyping method based on the combined use of
surface ligation chemistry and nanoparticle enhanced SPRI. Reprinted
with permission from reference 44.
applied for multiplexed SNP genotyping achieving both high specicity and
sensitivity. The detection limit of 1 pM is comparable to that typically reported
for uorescence imaging measurements of DNA microarrays [81,82] and we
expect that this detection limit will further improve with the incorporation of
enzymatic amplication methods.
8.4.2 MicroRNA Detection
A nal example of an advanced SPRI biosensing methodology is the use of both
a surface poly(A) polymerase reaction and nanoparticle amplication for the
ultrasensitive microarray detection of microRNAs (miRNAs) down to a con-
centration of 10fM. In the previous example, improved sensitivity was obtained
from nanoparticle adsorption only. However, in this case, both the enzyme
reaction and DNA-coated nanoparticles contribute to the amplication of the
Figure 8.10 SNP genotyping of 1 nM PCR amplicons of two genomic DNA samples
(a) wild-type and (b) NA14637. Both SPRI dierence images measure the
hybridization adsorption of L
C
-modied nanoparticles on a four-
component DNA microarray (P
A
, P
C
, P
G
and P
N
) after the surface
ligation and denaturation steps are completed. The SPRI dierence
image (left) and corresponding line proles (middle) are shown next to
the array pattern (right). An increase in SPRI signal was observed only at
the P
G
array elements for the (a) wild-type sample identifying it as being a
CC homozygote. For the (b) NA14637 mutant sample, an SPRI signal
increase was observed at both the P
G
and P
A
(boxed) elements identifying
it as a CT heterozygote. Reprinted with permission from reference 44.
264 Chapter 8
SPRI response. As indicated in the left inset in Figure 8.11, poly(A) polymerase
selectively catalyzes the multiple addition of adenosine residues to the 3
0
-OH end
of ssRNA molecules resulting in the formation of long poly(A) tails. To char-
acterize the surface polymerase reaction, 5
0
-thiol-modied, 3
0
-OH ssRNA was
covalently attached to the gold surface via a surface thiolmaleimide reaction
[34,49] to create an ssRNA monolayer with free 3
0
-hydroxyl groups that are
accessible to the enzyme. Figure 8.11 shows an SPRI dierence image along with
the corresponding line prole following the completion of the polyadenylation
reaction on a microarray where the fractional surface coverage of ssRNA ranged
from a full monolayer to 10
4
. The diluted array elements were prepared by
mixing a 1 mM solution of thiol-modied ssRNA with a 1 mM solution of a non-
interacting thiol-modied ssDNA sequence of similar length in a ratio that varied
from 1:10 to 1:10
4
. An hour was sucient for completion of the polyadenylation
reaction to gather the data in Figure 8.11, clearly showing that SPRI reectivity
changes could be detected at ssRNA surface coverages of around 10
3
.
A further gain in SPRI sensitivity can be achieved by the hybridization
adsorption of T
30
DNA-coated nanoparticles on the surface polyadenylated
ssRNA. Figure 8.12 compares the total change in SPRI reectivity measured
after both polyadenylation and nanoparticle adsorption with the SPRI signal
Figure 8.11 SPRI dierence image (top middle) and corresponding line prole
(bottom) obtained after a surface polyadenylation reaction (left inset)
on a microarray composed of array elements with relative ssRNA
surface coverages ranging from 10
1
to 10
4
. The array elements were
created by diluting a 1 mM solution of thiol-modied ssRNA with 1 mM
thiol-modied ssDNA solution prior to surface immobilization and
surface polyadenylation reaction at the 3
0
-end of the surface-attached
ssRNA substrate.
obtained after the polyadenylation step only [34]. When a 10 nM nanoparticle
solution was employed, the SPRI signal responded linearly at ssRNA fractional
surface coverages ranging from 10
6
to 10
4
. At higher coverages, the SPRI
responsivity decreases with signal saturation observed at a DR of B25%, as
expected from theory [53]. The remarkably low value of 10
6
corresponds to an
ssRNA surface density of 5 10
6
molecules cm
2
or just B10,000 RNA
molecules on a 500 mm square array element! These measurements can be used
to estimate what concentration of target miRNA in solution would be required
to achieve a similar surface coverage following hybridization adsorption onto a
complementary nucleic acid microarray. At low miRNA concentrations (C),
the fractional surface coverage (y) can be described using the equation
y K
ads
C. Given a Langmuir adsorption coecient (K
ads
) of 10
8
M
1
, a y of
10
6
corresponds to an miRNA concentration of 10 fM.
Figure 8.13 shows how this polyadenylationnanoparticle amplication
scheme can be applied to the ultrasensitive detection of miRNAs using SPRI
[34]. The recent surge in interest in miRNAs has led to an increased need for
new methods that can perform multiplexed, quantitative analyses at very low
target concentrations [8385]. The rst step (i) in our miRNA detection
Figure 8.12 SPR reectivity changes as a function of ssRNA surface coverage after
surface polyadenylation reaction (K) and after nanoparticle (NP)
adsorption on a surface where the polyadenylation reaction has already
occurred (). The inset shows the scheme of the hybridizationadsorp-
tion of T
30
-coated gold nanoparticles (Au-NP) on the poly(A) tail
formed at the 3
0
-end of the surface-attached ssRNA molecules. The
dashed line indicates the minimum detectable SPRI response signal.
Reprinted with permission from reference 34.
266 Chapter 8
methodology involves the sequence-specic hybridization adsorption of the
miRNA target on a single-stranded locked nucleic acid (LNA) microarray.
LNAs are nucleic acid analogues containing one or more nucleotides modied
with an extra bridge connecting the 2
0
-O and 4
0
-C atoms of the ribose moiety;
the RNALNA binding strength (B10
8
M
1
) is at least 10 times greater than
that associated with RNADNA heteroduplex formation [83]. The presence of
the surface-bound miRNA is then detected with SPRI following polyadenylat-
ion (step ii) and T
30
-coated nanoparticle adsorption (step iii). Initial measure-
ments were performed using synthetic analogues of the target miRNA
molecules with the SPRI response was found to increase linearly over a
concentration range of 10500 fM. Both the detection limit and linear response
range could be adjusted by varying the nanoparticle concentration. At miRNA
concentrations of 100 pM and above, polyadenylation was sucient to detect
the presence of miRNA without nanoparticle amplication.
As a nal demonstration, the surface polyadenylation-nanoparticle amplica-
tion methodology was applied to the multiplexed detection of three dierent
miRNAs present in a total RNA sample extracted from mouse liver tissue. A
four-component microarray was constructed containing three LNA probes
designed to bind to the known miRNA sequences [34,84], miR-16, miR-122b
and miR-23b, with a DNA probe used as a negative control. A 250 ng RNA
Figure 8.13 Detection of microRNAs using a combination of surface polyadenylation
chemistry and nanoparticle-amplied SPRI. Step (i): hybridization adsorp-
tion of miRNA onto a complementary LNA array element. Step (ii):
poly(A) tail addition at the 3
0
-end of surface-bound miRNAs using poly(A)
polymerase. Step (iii): hybridizationadsorption of T
30
-coated Au nano-
particles to poly(A) tails. Reprinted with permission from reference 34.
sample in a volume of 500 ml was circulated over the microarray surface repeat-
edly for 4 h followed by surface amplication. Analysis of the resulting SPRI
dierence image (Figure 8.14a) and corresponding line prole (Figure 8.14c)
clearly shows the miR-122b sequence as the most abundant with a measured DR
of 9.7% [34]. Having already calibrated the linear SPRI response range using
synthetic analogues of the target sequences we were able to estimate miRNA
concentrations of 20fM, 50 fM and 2 pM for miR-16, miR-23b and miR-122b,
respectively. Further verication of the concentration of the least abundant
miRNA (miR-16) was obtained by repeating the measurement with the addition
of 100 fM synthetic miR-16. As shown in Figures 8.14b and c, a 5-fold increase in
SPRI signal was observed only at the miR-16 probe array elements. These
experiments demonstrate unequivocally that polyadenylationnanoparticle am-
plied SPRI measurements can be used to prole miRNA sequences quantita-
tively in biological samples. The current detection limit of 5 amol (10fM in 500 ml)
Figure 8.14 Quantitative analysis of miRNAs from 250 ng of mouse liver total RNA
using polyadenylationnanoparticle amplied SPRI measurements. (a)
SPRI dierence image obtained by subtracting images acquired before
and after the nanoparticle amplication step. (b) An SPRI dierence
image obtained from a separate chip using the same total RNA concen-
tration as the top image plus the addition of 100 fM synthetic miR-16.
(c) Comparison of line proles taken from both SPRI dierence images
with the solid and dashed lines corresponding to top and bottom
images respectively. (d) Schematic of the four-component LNA probe
microarray and the line prole location. Reprinted with permission from
reference 34.
268 Chapter 8
opens up the possibility of complete miRNA proling with SPRI using very small
amounts of sample (o250 ng).
8.5 Summary and Outlook
This chapter highlights some of the recent advances that we have made in
improving the microarray-based biosensing capabilities of SPRI. The continual
development of new surface attachment chemistries and array fabrication
methods on gold lm surfaces has resulted in a large variety of biomolecular
interactions that have been studied in a multiplexed format using SPRI.
Furthermore, the selectivity, sensitivity and applicability of SPRI bioanity
measurements can be greatly enhanced by incorporating surface enzyme reac-
tions into the detection scheme. The ability to detect unlabeled DNA sequences
at concentrations as low as 1 fM using RNase H amplied SPRI is a remark-
able improvement compared with the previous nanomolar detection limit of
SPRI based on hybridization absorption only. Also, we found the combined
use of surface enzyme reactions and nanoparticle-enhanced SPRI to be partic-
ularly powerful. This allowed us to demonstrate for the rst time that SPRI can
be applied for SNP genotyping; the surface Taq DNA ligase reaction enhanced
the specicity of the SNP detection while DNA-coated nanoparticles amplied
the SPRI response. Even greater improvements in sensitivity were achieved
through the combined use of a surface poly(A) polymerase reaction and DNA-
coated nanoparticles, which we used to simultaneously detect multiple
miRNAs present at concentrations ranging from 20 fM to 2 pM in a total
RNA sample.
The level of sensitivity obtained using these combined enzymenanoparticle
amplied SPRI methods is equal to or better than that of uorescence imaging
measurements of DNA microarrays, where a detection limit of 1 pM is typically
reported [79,80]. Another strength of SPRI versus uorescence is the ability to
monitor directly multi-anity interactions involving the adsorption of two or
more species on a single surface probe, which is important for the development
of more complex multi-step assays. Furthermore, the nucleic acid enzymes
described in this chapter are only some of the many possible enzymes (e.g.
proteases, kinases) that can be utilized on biomolecules attached to microarray
surfaces. In the case of more complex surface enzymatic processes, the com-
bined use of SPFS and SPRI measurements will support the development of
more complicated surface kinetic models and also help optimize the surface
enzyme reaction through changes in surface density and vertical spacing of the
immobilized substrate. The surface enzyme reactions at gold surfaces described
in this chapter can also be coupled with uorescence or electroactive labeled
molecules; this opens up the possibility of creating new enzymatic methods that
combine SPRI with other surface-sensitive detection techniques. We expect
that, as more advanced surface amplication methods are developed for
ultrasensitive microarray bioanity measurements, the technique of SPRI will
continue to grow in importance as an invaluable research tool that can be
applied to many areas of modern biotechnology research.
8.6 Questions
1. A signicant improvement in SPRI instrumentation is obtained by using a
non-coherent light source instead of coherent laser excitation. (A) Suggest
some reasons why and (B) outline other features of the optical set-up
required for SPRI measurements.
2. Most surface bioanity measurements utilize the specic adsorption of
target biomolecules (T) from solution onto a surface that has been
chemically modied with probe biomolecules (P). If the target and probe
interact in a simple 1:1 ratio, then in the absence of bulk transport the
surface reaction can be represented in the form
T P
k
a
k
d
TP
where TP is the surface bound targetprobe complex. Both P and TP are
surface species, and in the Langmuir approximation their surface concentra-
tions G
P
and G
TP
are linked to the total concentration of surface sites G
tot
by
G
tot
G
P
G
TP
A. If we dene y as the fraction of occupied surface sites, y G
TP
/G
tot
,
write down the dierential equation for the time evolution of y (i.e. dy/
dt ) using k
a
, k
d
and the bulk concentration [T].
B. Find the steady-state equilibrium surface coverage y
eq
, which is
obtained in the steady-state approximation (dy/dt 0). This equation
denes the Langmuir adsorption coecient K
ads
k
a
/k
d
. What is y
eq
at
a bulk concentration equal to 1/K
ads
?
3. One of the main factors that determine the SPRI detection limit for DNA
hybridization adsorption is the lowest fractional surface coverage of target
DNA on the gold chip surface that can actually be measured. At low
target concentrations the solution to question 2B can be approximated as
y
eq
K
ads
[T]. In the text, the DNA detection limit reported for long-range
SPRI measurements was 1 nM, with a value of 1 pM described for
nanoparticle-enhanced SPRI detection. In the case of RNase H amplied
SPRI, a detection limit of 1 fM was reported. For the detection limits of
1 nM, 1 pM and 1 fM, calculate y
eq
, G
TP
and estimate the number of target
DNA molecules contributing to a change in signal within an array element
that has an area of 500 500 mm. Assume K
ads
2 10
7
M
1
and a
surface density of DNA probe molecules of 5 10
12
molecules cm
2
with
all probes available for target binding.
4. In this chapter, several examples where surface enzyme reactions form
the basis of advanced SPRI biosensing measurements are demon-
strated. Equations (8.8)(8.11) describe the reaction of an enzyme in
solution with a surface-immobilized substrate that couples Langmuir
adsorption/desorption kinetics with the rate of the surface enzyme
270 Chapter 8
reaction. The time-dependent SPRI signal D%R(t) responds to both en-
zyme adsorption and the loss of DNA or RNA substrate from the gold
surface (see Figure 8.5, inset) and can be represented by D%R(t) y
ES
y
S*
.
A. Using simple Euler integration methods (which can be easily applied in
Excel, IGOR Pro or any mathematics software) with the initial con-
ditions y
s
1 and y
ES
y
S*
0 at time t 0, try to simulate the surface
kinetic reaction described in eqs. (8.8)(8.11). Generate two sets of
plots showing y
ES
, y
S*
, y
S
and D%R(t) versus time using the following
parameter values:
Parameter Set 1 where k
cat
4k
a
[E]:
k
a
1 10
6
M
1
s
1
k
d
0.025 s
1
k
cat
2.5 s
1
[E] 2.5 10
7
M
b0.
Parameter Set 2 where k
cat
ok
a
[E]:
k
a
1 10
6
M
1
s
1
k
d
0.025 s
1
k
cat
0.025 s
1
[E] 2.5 10
7
M
b0.
B. To demonstrate the impact of mass transport eects on the kinetics of
the surface enzyme reaction, replot Parameter Set 2 using b 50.
5. In both this chapter and Chapter 2, gold nanoparticles are used for the
sensitive detection of biomolecular interactions. What is the fundamental
dierence in applying these particles in SPR instruments?
6. In Sections 8.3.2 and 8.4.2, two very dierent enzymatic amplication
methods are applied. Describe and compare both these methods.
7. A detection limit of 5 amol (10 fM in 500 ml) of miRNA with SPRI is
shown for very small amounts of sample (o250 ng). Which two tricks
were applied to achieve this limit of detection?
This research was funded by the National Institutes of Health (2RO1
GM059622-04) and the National Science Foundation (CHE-0551935). The
authors would like to thank Dr. S. Fang for obtaining the SPRI dierence
image shown in Figure 8.11.
References
1. D.J. Lockhart and E.A. Winzeler, Nature, 2000, 405, 827.
2. P.O. Brown and D. Botstein, Nat. Genet., 1999, 21, 33.
3. R.B. Stoughton, Annu. Rev. Biochem., 2005, 74, 53.
4. J.M. Thomson, J. Parker, C.M. Perou and S.M. Hammond, Nat. Methods,
2004, 1, 47.
5. R.L. Stears, T. Martinsky and M. Schena, Nat. Med., 2003, 9, 140.
6. Y.P. Bao, M. Martin Huber, T. Wei, S.S. Marla, J.J. Storho and U.R.
Mu ller, Nucleic Acids Res., 2005, 33, e15.
7. D. Gerion, F. Chen, B. Kannan, A. Fu, W.J. Parak, D.J. Chen,
A. Majumdar and A.P. Alivisatos, Anal. Chem., 2003, 75, 4766.
8. P.F. Predki, Curr. Opin. Chem. Biol., 2004, 8, 8.
9. H. Zhu and M. Snyder, Curr. Opin. Chem. Biol., 2003, 7, 55.
10. G. Macbeath and S.L. Schreiber, Science, 2000, 289, 1760.
11. S.F. Kingsmore, Nat. Rev. Drug Discov., 2006, 5, 310.
12. P. Bertone and M. Snyder, FEBS J., 2005, 272, 5400.
13. M. Srivastava, O. Eidelman, C. Jozwik, C. Paweletz, W. Huang, P.L.
Zeitlin and H.B. Pollard, Mol. Genet. Metab., 2006, 87, 303.
14. W. Gao, R. Kuick, R.P. Orchekowski, D.E. Misek, J. Qiu, A.K. Greenberg,
W.N. Rom, D.E. Brenner, G.S.S. Omenn, B.B. Haab and S.M. Hanash,
BMC Cancer, 2005, 5, 110.
15. W. Hickel, D. Kamp and W. Knoll, Nature, 1989, 339, 186.
17. E. Yeatman and E.A. Ash, Electron. Lett., 1987, 23, 1091.
18. H.J. Lee, A.W. Wark and R.M. Corn, Langmuir, 2006, 22, 5241.
19. K.S. Phillips, T. Wilkop, J.-J. Wu, R.O. Al-Kaysi and Q. Cheng, J. Am.
Chem. Soc., 2006, 128, 9590.
20. L.K. Wolf, D.E. Fullenkamp and R.M. Georgiadis, J. Am. Chem. Soc.,
2005, 127, 17453.
21. T. Wilkop, Z. Wang and Q. Cheng, Langmuir, 2004, 20, 11141.
22. M. Kyo, K. Usui-Aoki and H. Koga, Anal. Chem., 2005, 77, 7115.
23. V. Kanda, P. Kitov, D.R. Bundle and M.T. McDermott, Anal. Chem.,
2005, 77, 7497.
24. A. Okumura, Y. Sato, M. Kyo and H. Kawaguchi, Anal. Biochem., 2005,
339, 328.
25. M. Kyo, T. Yamamoto, H. Motohashi, T. Kamiya, T. Kuroita, T. Tanaka,
J.D. Engel, B. Kawakami and M. Yamamoto, Genes Cells, 2004
9, 153.
26. J.S. Shumaker-Parry and C.T. Campbell, Anal. Chem., 2004, 76, 907.
27. J.S. Shumaker-Parry, M.H. Zareie, R. Aebersold and C.T. Campbell, Anal.
Chem., 2004, 76, 918.
28. V. Kanda, J.K. Kariuki, D.J. Harrison and M.T. McDermott, Anal.
Chem., 2004, 76, 7257.
29. E.A. Smith and R.M. Corn, Appl. Spectrosc., 2003, 57, 320A.
30. L. He, M.D. Musick, S.R. Nicewarner, F.G. Salinas, S.J. Benkovic, M.J.
Natan and C.D. Keating, J. Am. Chem. Soc., 2000, 122, 9071.
31. L.A. Lyon, W.D. Holliway and M.J. Natan, Rev. Sci. Instrum., 1999, 70,
2076.
32. S.F. Gonzalez, M.J. Krug, M.E. Nielsen, Y. Santos and D.R. Call, J. Clin.
Microbiol., 2004, 42, 1414.
272 Chapter 8
33. S. Sengupta, K. Onodera, A. Lai and U. Melcher, J. Clin. Microbiol., 2003,
41, 4542.
34. S. Fang, H.J. Lee, A.W. Wark and R.M. Corn, J. Am. Chem. Soc., 2006,
128, 14044.
35. C.E. Jordan, A.G. Frutos, A.J. Thiel and R.M. Corn, Anal. Chem., 1997,
69, 4939.
36. A.J. Thiel, A.G. Frutos, C.E. Jordan, R.M. Corn and L.M. Smith, Anal.
Chem., 1997, 69, 4948.
1999, 71, 3928.
38. B. Frey and R.M. Corn, Anal. Chem., 1996, 68, 3187.
39. J.M. Brockman, A.G. Frutos and R.M. Corn, J. Am. Chem. Soc., 1999,
121, 8044.
40. A.G. Frutos, J.M. Brockman and R.M. Corn, Langmuir, 2000, 16, 2192.
41. E. Smith, M.J. Wanat, Y. Cheng, S.V.P. Barreira, A.G. Frutos and R.M.
Corn, Langmuir, 2001, 17, 2502.
42. H.J. Lee, T.T. Goodrich and R.M. Corn, Anal. Chem., 2001, 73, 5525.
43. H.J. Lee, D. Nedelkov and R.M. Corn, Anal. Chem., 2006, 78, 6504.
44. Y. Li, A.W. Wark, H.J. Lee and R.M. Corn, Anal. Chem., 2006, 78, 3158.
45. A.W. Wark, H.J. Lee and R.M. Corn, Anal. Chem., 2005, 77, 3904.
46. H.J. Lee, Y. Li, A.W. Wark and R.M. Corn, Anal. Chem., 2005, 77, 5096.
47. H.J. Lee, A.W. Wark, Y. Li and R.M. Corn, Anal. Chem., 2005, 77, 7832.
48. T.T. Goodrich, H.J. Lee and R.M. Corn, J. Am. Chem. Soc., 2004, 126,
4086.
49. T.T. Goodrich, H.J. Lee and R.M. Corn, Anal. Chem., 2004, 76, 6173.
50. E.A. Smith, M.G. Erickson, A.T. Ulijasz, B. Weisblum and R.M. Corn,
Langmuir, 2003, 19, 1486.
51. G.J. Wegner, H.J. Lee and R.M. Corn, Anal. Chem., 2002, 74, 5161.
52. G.J. Wegner, A.W. Wark, H.J. Lee, E. Codner, T. Saeki, S. Fang and R.M.
Corn, Anal. Chem., 2004, 76, 5677.
53. B.P. Nelson, T.E. Grimsrud, M.R. Liles, R.M. Goodman and R.M. Corn,
Anal. Chem., 2001, 73, 1.
54. D.J. OShannessy, M. Brigham-Burke, K.K. Soneson, P. Hensley and
I. Brooks, Anal. Biochem., 1993, 212, 457.
55. G.G. Neminger, P. Tobiska, J. Homola and S.S. Yee, Sens. Actuators B,
2001, 74, 145.
56. F. Yang, G.W. Bradberry and J.R. Sambles, Phys. Rev. Lett., 1991, 66,
2030.
57. J.C. Quail, J.G. Rako and H.J. Simon, Opt. Lett., 1983, 8, 377.
58. D. Sarid, Phys. Rev. Lett., 1981, 47, 1927.
59. W.N. Hansen, J. Opt. Soc. Am., 1968, 58, 380.
60. http://www.corninfo.ps.uci.edu/calculations.html.
61. A. Kasry and W. Knoll, Appl. Phys. Lett., 2006, 89, 101106.
62. C. Bourdillon, C. Demaille, J. Moiroux and J. Saveant, J. Am. Chem. Soc.,
1999, 121, 2401.
64. S. Fang, H.J. Lee, A.W. Wark, H.M. Kim and R.M. Corn, Anal. Chem.,
2005, 77, 6528.
65. H.J. Lee, A.W. Wark, T.T. Goodrich, S. Fang and R.M. Corn, Langmuir,
2005, 21, 4050.
66. E.J. Cho, J.R. Collet, A.E. Szafranska and A.D. Ellington, Anal. Chim.
Acta, 2006, 564, 82.
67. J.R. Collett, E.J. Cho, J.F. Lee, M. Levy, A.J. Hood, C. Wan and A.D.
Ellington, Anal. Biochem., 2005, 338, 113.
68. T. Mori, A. Oguro, T. Ohtsu and Y. Nakamura, Nucleic Acids Res., 2004,
32, 6120.
69. T.G. McCauley, N. Hamaguchi and M. Stanton, Anal. Biochem., 2003,
319, 244.
70. M.B. Murphy, S.T. Fuller, P.M. Richardson and S.A. Doyle, Nucleic Acids
Res., 2003, 31, e110.
71. Y. Li, H.J. Lee and R.M. Corn, Nucleic Acids Res., 2006, 34, 6416.
72. L. He, E.A. Smith, M.J. Natan and C.D. Keating, J. Phys. Chem. B, 2004,
108, 10973.
73. L.A. Lyon, D.J. Pena and M.J. Natan, J. Phys. Chem. B, 1999, 103, 5826.
74. L.A. Lyon, M.D. Musick and M.J. Natan, Anal. Chem., 1998, 70, 5177.
75. R.W. Carthew, Curr. Opin. Genet. Dev., 2006, 16, 203.
76. D. Bartel, Cell, 2004, 116, 281.
77. T. Endo, K. Kerman, N. Nagatani, Y. Takamura and E. Tamiya, Anal.
Chem., 2005, 77, 6976.
78. J. Liu, S. Tian, L. Tiefenauer, P.E. Nielsen and W. Knoll, Anal. Chem.,
2005, 77, 2756.
79. A.W. Peterson, L.K. Wolf and R.M. Georgiadis, J. Am. Chem. Soc., 2002,
124, 14601.
80. K. Nakatani, S. Sando and I. Saito, Nat. Biotechnol., 2001, 19, 51.
81. H.-P. Lehr, M. Reimann, A. Brandenburg, G. Sulz and H. Klapproth,
Anal. Chem., 2003, 75, 2414.
82. T. Livache, E. Maillart, N. Lassalle, P. Mailley, B. Corso, P. Guedon, A.
Roget and Y. Levy, J. Pharm. Biomed. Anal., 2003, 32, 687.
83. M. Castoldi, S. Schmidt, V. Benes, M. Noerholm, A.E. Kulozik, M.W.
Hentze and M.U. Muckenthaler, RNA, 2006, 12, 913.
84. T. Babak, W. Zhang, Q. Morris, B.J. Blencowe and T.R. Hughes, RNA,
2004, 10, 1813.
85. P.T. Nelson, D.A. Baldwin, L.M. Scearce, J.C. Oberholtzer, J.W. Tobias
and Z. Mourelatos, Nat. Methods, 2004, 1, 155.
86. B.P. Nelson, M.R. Liles, K. Frederick, R.M. Goodman and R.M. Corn,
Environ. Microbiol., 2002, 4, 735.
87. J.M. Brockman, B.P. Nelson and R.M. Corn, Annu. Rev. Phys. Chem.,
2000, 51, 41.
88. E.A. Smith, W.D. Thomas, L.L. Kiessling and R.M. Corn, J. Am. Chem.
Soc., 2003, 125, 6140.
89. G.J. Wegner, H.J. Lee, G. Marriott and R.M. Corn, Anal. Chem., 2003, 75,
4740.
274 Chapter 8
CHAPTER 9
Surface Plasmon Fluorescence
Techniques for Bioafnity
Studies
WOLFGANG KNOLL, AMAL KASRY, JING LIU,
THOMAS NEUMANN,
1
LIFANG NIU, HYEYOUNG
PARK,
2
HARALD PAULSEN,
3
RUDOLF ROBELEK,
DANFENG YAO AND FANG YU
4
Max Planck Institute for Polymer Research, Ackermannweg 10, D-55128
Mainz, Germany
9.1 Introduction
Among the various sensing principles proposed for bioafnity studies, optical
evanescent wave techniques have gained the lead in popularity. Next to
evanescent ellipsometry [1] and the various optical waveguide platforms [2,3],
surface plasmon resonance (SPR) spectroscopy [46], in particular, has found
widespread applications and has demonstrated its potential for the sensitive
detection of bioanalytes in numerous examples [79]. Since its introduction as a
method for bioafnity studies in 1983 by Liedberg et al. [10] and since the
presentation of the rst commercial instrument by Biacore in 1990 [11], the
number of papers published has grown exponentially to currently more than
1500 contributions each year (Figure 9.1). This success story is largely based on
the fact that SPR represents a label-free detection principle the mere presence
of the bound analyte slightly changes the optical architecture at the sensor
surface which is probed by the surface plasmon mode propagating along this
1
Present address: Grafnity, Heidelberg, Germany.
2
Present address: Samsung, Seoul, Korea.
3
Johannes Gutenberg University of Mainz, FB Biology, Gresemundweg 2, D-55099 Mainz,
Germany.
4
Present address: Stanford University, Stanford, CA, USA.
275
metal/dielectric interface. Moreover, these SPR-based detection principles offer
very attractive sensitivities for in situ and real-time monitoring of bio-analytes.
Another reason for the rapid growth of SPR biosensing applications is
directly linked to the fact that the required functionalization of the sensor
surface can be achieved fairly easily using a variety of surface modication
protocols ([1214], and including Chapter 6). One particularly simple strategy is
based on the self-assembly of a monolayer of thiol derivatives on Au surfaces
[15]. Figure 9.2 shows the schematics of such a multilayer architecture devel-
oped for the functionalization of SPR Au chips with a two-dimensional binding
matrix for protein- and DNA-sensing processes. First, a biotinylated thiol
derivative is co-assembled together with OH-terminated diluent molecules (in a
molar ratio of 1:9) in a binary mixture, allowing for the specic binding of a
monolayer of streptavidin, a tetrameric protein with an extremely high afnity
(K
A
10
15
M
1
) for biotin. Two of the four binding sites face the aqueous
phase and thus can be used to couple other biotinylated molecules or small
objects to the sensor surface generating the designed architecture with the
specicity for the analyte of interest.
Even though this two-dimensional binding matrix fulls many criteria
required for biosensor coatings, e.g. the efcient exposure of binding sites
which are specic for the particular analyte while simultaneously minimizing
non-specic adsorption [16], it is wasting sensitivity by monitoring binding
200
400
600
800
1000
1200
1400
1600
0
20
40
60
80
100
120
140
1990 1995 2000 2005
Year
N
u
m
b
e
r

o
f

p
u
b
l
i
c
a
t
i
o
n
s
/
y
e
a
r
Surface plasmons
Surface plasmon fluorescence
Figure 9.1 Number of publications per year dealing with surface plasmon resonance
spectroscopy (red curve) and with surface plasmon uorescence spectro-
scopy (blue curve).
276 Chapter 9
events only within a thin two-dimensional slice of the analyte solution that is
probed by the evanescent surface plasmon wave reaching some 150 nm out into
the buer [6]. Hence it was a natural next step to use a quasi-three-dimensional
binding matrix based on a polymer brush that is partially coupled to the
substrate and reaches out into the analyte solution some 100 nm in the case of
Biacores dextran brush, matching the extent of the evanescent surface plasmon
eld. This way, binding events away from the interface, but still within the
evanescent tail of the surface plasmon mode probing the brush, also contribute
to the sensor signal (although with an exponentially decreasing weighting
factor). Roughly 35 times more protein compared with the 2D matrix bound
per unit area of the sensor surface is monitored, thus enhancing the sensitivity
of the technique considerably.
However, if very small analyte molecules with a low molecular mass are to be
detected or if only very low coverage on the sensor surface (even for large
proteins) can be obtained, the resulting changes of the effective refractive index
probed by the surface plasmon are too minute to be detected. An example is
given in Figure 9.3; the angular scans taken in the Kretschmann conguration
LaSFN9
Au
SAM
Biotin
Streptavidin
Biotinylated
probe
Target DNA
Figure 9.2 Biofunctional interfacial architecture based on the self-assembled mono-
layer (SAM) of a binary mixture of thiol derivatives (i.e. 10% of a
biotinylated thiol in 90% of OH-terminated thiols) designed for the stable
attachment of a monolayer of streptavidin as a generic binding matrix. In
this example, this layer is employed for the coupling of biotinylated DNA
oligonucleotide catcher probes (blue strands) for surface hybridization
studies with target strands from solution (red strands).
277 Surface Plasmon Fluorescence Techniques for Bioafnity Studies
of an SPR surface hybridization experiment with sensor architecture as shown
in Figure 9.2 before and after binding of the target analyte are virtually
superimposed (a). Consequently, also the kinetic mode (b) shows no change
in the reected intensity monitored as a function of time after the injection of
the target solution. Hence neither of these modes of operation allows for label-
free detection of this hybridization event.
We should point out, however, that this complete lack of sensitivity is a direct
consequence of the specic design of this experiment: in order to reduce the
Coulombic cross-talk between neighboring binding sites, especially for studies
in low ionic strength buers, the individual catcher strands were separated from
each other by coupling them via a biotin linker to the streptavidin monolayer,
thus generating a very dilute binding matrix and, hence, a low analyte coverage
even at 100% hybridization efciency. It has been shown that other catcher
binding matrices based, e.g., on thiolated probe strands do allow label-free
detection of target hybridization, but at the expense of a dense binding matrix
resulting in significant cross-talk between neighboring hybridization sites [17].
9.2 Surface Plasmon Fluorescence Spectroscopy
(SPFS)
An obvious way to enhance sensitivity and to push the limit of detection (LOD)
to lower surface coverage is the use of uorescent chromophores covalently
attached to the analyte molecules. In this approach, the resonantly excited
46 48 50 52 54 56 58 60 62
0.0
0.2
0.4
0.6
0.8
1.0
R F / cps
reflectivity:
before
after
angle /deg
hybridization
fluorescence:
before
after
0 200 400 600 800 1000 1200 1400 1600
0
1
R
time t /sec
reflectivity
0.0
5.0x10
5
1.0x10
6
1.5x10
6
2.0x10
6
F / cps
fluorescence
(a) (b)
rinsing
injection of target solution
Figure 9.3 Hybridization experiments on small amounts of chromophore-labeled
target strands to the probe matrix in the Kretschmann conguration. (a)
Recorded reectivities, R, and uorescence intensities, F, as a function of
the angle of incidence before (full squares and dotted line) and after the
binding of the target strands to the surface immobilized catcher probes
(open circles and solid line). (b) Kinetic experiment with injection of the
target solution and a rinsing step, as indicated by the arrows. While
reectivity remains unchanged, the uorescence increase demonstrates
successful binding.
278 Chapter 9
surface plasmon waves excite the uorophores and their emitted photons can
be monitored by a simple detection unit attached to a conventional SPR setup
[1821], shown in Figure 9.4 for the Kretschmann conguration, but can also
be realized with grating couplers.
5
Just as in surface-enhanced Raman spectroscopy [22], here too one leverages on
the strong optical eld enhancements that can be achieved at resonant excitation of
surface plasmons. This is shown in Figure 9.5, comparing attenuated total internal
reection (ATR) results with mere total internal reection (TIR) data. The coupling
prism was Ag metal coated on the lower half of the base area only and then spin-
coated with a thin polymer layer that was doped with a low concentration of
uorescent chromophores [23]. In this way, a mere vertical shift of the prism relative
to the (horizontal) plane of incidence in the reection setup of the Kretschmann
conguration allows for recording of the reected intensities as a function of
incidence angle for either TIR or ATR mode of operation. Simultaneously, the
emitted uorescence can be monitored for both cases. While the intensity moni-
tored for TIR shows only a at background signal (originating mostly from the
intrinsic uorescence excited by the laser beam passing through the high-index
prism), the intensity recorded in the case of the excitation of a surface plasmon
mode exhibits the strong angle dependence expected for this resonance behavior.
The peak intensity shows an enhancement of more than a factor of 100 for this Ag
substrate. Even with an Au coating, which is typically used for chemical stability
laser 632.8 nm
polarizers
photodiode
flow cell
filter
2
goniometer
lock-in
amplifier
photon-
counter
motor-
steering
PC

chopper
laser-shutter
r
lens
shutter
controller
prism
attenuator
PMT
Figure 9.4 Schematic of a Kretschmann surface plasmon spectrometer with added
uorescence detection unit consisting of a collection lens, an attenuator
(optional), a set of lters for the separation of scattered light and a
photomultiplier tube (PMT) or a (color) CCD camera for the microscopic
mode of operation (cf. Figure 9.22).
5
As shown in Figure 9.19a.
reasons when using aqueous buer solutions in biosensor studies, the enhancement
can account for more than an order of magnitude.
The slight angular displacement between the minimum angle in the reec-
tivity scan and the uorescence intensity peak position is a consequence of the
resonant excitation of the surface plasmon waves interfering with the directly
reected laser beam: the re-radiated and out-coupled plasmon light destruc-
tively interferes with the plane wave of the laser reected at the prism/metal
interface provided their relative phase shift corresponds to p or 1801. For the
surface plasmon, being a damped oscillator, this phase shift relative to the
reected laser is reached just slightly above the angle for maximum excitation
strength at resonance. Strictly, the uorescence gives the correct angular
position of the maximum eld intensity upon excitation of the surface plasmon.
The intensity prole normal to the metal/dielectric interface at the correct
resonance angle decays exponentially in both directions into the metal as well as
into the dielectric phase. This is shown in Figure 9.6a calculated for a multilayer
conguration consisting of a high-index glass prism (LaSFN9, n1.85 at
l633nm)/50nm Au/water). The intensity distribution normal to the interface
20 25 30 35 40
0.0
0.2
0.4
0.6
0.8
1.0

c
r
e
f
l
e
c
t
i
v
i
t
y

R
angle /deg
SPS Ag
TIR Glass
0
2x10
5
4x10
5
6x10
5
8x10
5
1x10
6
f
l
u
o
r
e
s
c
e
n
c
e
/
c
p
s
SPFS Ag
TIRF Glass
Figure 9.5 Comparison of the reectivities in the (attenuated) total reection mode of
excitation of a surface plasmon wave in an angular scan at a Ag/(thin)
polymer lm/air interface (lled circles) with just total internal reection at
a glass/polymer/air interface (open circles). Given are also the uorescence
intensities recorded as a function of time for the two modes of operation
[dashed line, total internal reection uorescence (TIRF); solid line, sur-
face plasmon uorescence spectroscopy (SPFS)]. The polymer lm was
doped with Cy5, a commonly used uorophore that can be excited at
l 633 nm.
280 Chapter 9
suggests trying to bring the analyte molecules as close to the metal surface as
possible in order to place their chromophores into the highest possible optical eld.
One has to take into account other effects on chromophores in the excited
state near a metal substrate, which can lead to, e.g., quenching of uorescence.
Qualitatively, the various distance regimes are summarized in Figure 9.7. As
shown in Figure 9.7a, in the immediate proximity of the metal, the resonant
excitation energy (Fo rster) transfer from the excited state of the donor chromo-
phore to the acceptor states of the metal substrate leads to strong quenching of
the uorescence emission, largely compromising the eld enhancement
obtainable in this conguration. At intermediate distances, shown in Figure
9.7b, an efcient back-coupling of the excitation energy from the vibrationally
relaxed excited state of the chromophore to the metal substrate leads to the
excitation of a red-shifted plasmon mode that can re-radiate via the prism at its
respective resonance angle. This back-coupling effect can be used to enhance
further the probability of uorescence emission; however, the recording of this
uorescence light through he prismis inconvenient from a practical point of view.
By far the easiest mode of operation is to monitor the uorescence emitted
directly from chromophores sufciently separated from the substrate surface
illustrated in Figure 9.7c. The dye molecules are still within the substantially
Prism Au layer Water
Dextran brush
O
p
t
i
c
a
l

i
n
t
e
n
s
i
t
y

I
s
/
I
o
0
0
0
1
2
100
200
4
8
12
16
(a)
(b)
Q
u
e
n
c
h
i
n
g

p
r
o
f
i
l
e
AF-Ab
z/nm
Figure 9.6 (a) Intensity prole of a surface plasmon mode at resonance (solid line)
and the relative uorescence intensity emitted by a chromophore as a
function of its separation from the metal surface (dashed line); (b) sche-
matic representation of a polymer (dextran) brush used to assemble a
quasi three-dimensional binding matrix functionalized by binding sites
(circles) for the chromophore (Alexa)-labeled antibody (AF-Ab).
enhanced optical eld of the surface plasmon mode, but they are not quenched
at all. This combination of eld enhancement and uorescence detection forms
the basis for the largely enhanced sensitivity applied for a wide range of
bioafnity studies shown on selected examples in this chapter.
9.3 Interface Kinetics Based on the Langmuir
Adsorption Model
Evanescent wave biosensors offer an easy way to measure the kinetics of the
reversible binding of a biomolecule from solution to a binding site (typically
another biomolecule) immobilized on the sensor surface. Although theoretical
aspects are treated in depth in Chapters 4 and 5, a brief analysis of kinetics is
described here, which is relevant for Section 9.4. The conventional treatment
starts with a simple 1:1 interaction model [24], equivalent to the Langmuir
adsorption model [25], which is the simplest physically plausible isotherm based
Dipole-dipole coupling Surface plasmon
back-coupling
Free emission
prism
metal
dye
dielectric
(a)
(b)
(c)
Figure 9.7 Schematic representation of the various coupling schemes for chromo-
phores near a metal surface. (a) If the dye is within the range of the Fo rster
resonance energy transfer regime (B57 nm), the (dipolar) coupling
between the donor chromophore and the acceptor (states of the) metal
substrate results in efcient quenching of uorescence; (b) at intermediate
distances, efcient back-coupling leads to excitation of a red-shifted
surface plasmon mode that can re-radiate via the prism and only at
sufcient distances free emission of the chromophore uorescence is
observed (c).
282 Chapter 9
on three assumptions:
the adsorption cannot proceed beyond monolayer coverage;
all sites are equivalent and the surface is uniform;
the ability of a molecule to adsorb to a given site is independent of the
degree of occupation of neighboring sites.
If these conditions are met, the dynamic equilibrium is given by
A B
k
on
k
off
AB 9:1
assuming that A is the molecule binding from solution (analyte) and B is the
species immobilized on the sensor surface (ligand). The forward and reverse
reaction rates are described by the adsorption (association) rate constant k
on
and the desorption (dissociation) rate constant k
o
, respectively. The associa-
tion process results in the formation of the complex [AB] and is described by
d AB
dt
k
on
A B 9:2
and the dissociation rate of the complex [AB] is given by
d AB
dt
k
off
AB 9:3
Once a dynamic equilibrium is established, the rates of both processes are
equal, i.e.
k
on
A B k
off
AB 9:4
Hence the equilibrium constants can be expressed by the rate constants
according to
K
a

AB
A B

k
on
k
off
9:5
and
K
d

A B
AB

k
off
k
on
9:6
where K
a
and K
d
are the afnity constant and the dissociation constant,
respectively. This formalism is mathematically identical with that of the treat-
ment of the interaction in the homogeneous phase. However, at the solidliquid
interface the transport (diffusion and convection) of A from the bulk solution
to the interface must be taken into account (cf. Chapter 5).
9.3.1 Mass Transport-controlled Kinetics
If the surface concentration of B (ligand) is very large and the mass transport
rate k
m
is small compared with the association rate constant k
on
, i.e.
k
m
{k
on
[B], the interaction is controlled by the mass transport rate. Then
the complex formation rate is solely dependent on the bulk concentration of
analyte A and the binding signal increases linearly with time:
d AB
dt
k
m
A
bulk
9:7
This can be used for the concentration analysis of analyte A, since the slope
of the initial stage of the binding curve is proportional to the analyte concen-
tration. Theoretically, the linear range of the dose-response curve has no
limitation at the lower concentration side. If the reaction rate is fully mass-
transport limited, the sensor surface acts like an innite sink and [A
surface
] 0.
In this case, k
m
for all practical situations can be described by [26]
k
m
0:98D=h
2
3
v=bx
1
3
9:8
where D is the diffusion coefcient, h and b are the height and the width of the
ow cell, respectively, v is the volumetric ow rate and x is the distance from the
ow cell entrance.
9.3.2 Interaction-controlled Kinetics
If the mass transport rate is much larger than the association rate constant or if
the surface concentration of the immobilized species is low, i.e. k
m
ck
on
[B],
then [A
surface
] [A
bulk
] and the binding rate can be expressed as
d AB
dt
k
on
A B k
off
AB 9:9
The surface concentration of the free binding site, [B], is the difference between
the concentration of the complex at saturation, [AB
max
], and the current
complex concentration, [AB]:
B AB
max
AB 9:10
Combining eqs. (9.9) and (9.10) and considering that the response R scales
linearly with the complex concentration [AB], one obtains
dR
dt
k
on
c
0
R
max
R k
off
R 9:11
284 Chapter 9
where c
0
is the concentration of the analyte and R
max
is the saturation signal at
sufciently high analyte concentration. The solution of eq. (9.11) yields
R
k
on
c
0
R
max
k
on
c
0
k
off
1 e
k
on
c
0
k
off
t
R
0
1 e
k
on
c
0
k
off
t
9:12
This is shown schematically in Figure 9.8a as the association phase.
For the dissociation phase, c
0
0, hence
dR
dt
k
off
R 9:13
and the solution becomes (cf. Figure 9.8a, dissociation phase)
R R
0
e
k
off
t
9:14
R
e
s
p
o
n
s
e
R
e
s
p
o
n
s
e
0
1
0.5
Concentration
Time
Association
Dissociation
K
d
(a)
(b)
Figure 9.8 Schematics of a typical interaction analysis involving a kinetic association
and dissociation process (a) and an equilibrium analysis based on a 1:1
Langmuir model.
6
(b) Both methods allow for the determination of the
intrinsic afnity constant of the interaction partners.
6
Note the logarithmic scale of concentration.
Equations (9.12) and (9.13) can be used to give k
on
and k
o
from a single set
of association/dissociation experiments using non-linear curve tting.
9.3.3 Equilibrium Analysis
Once a dynamic equilibrium has been reached, the net effect of the association
and dissociation process is zero, i.e.
dR
dt
k
on
c
0
R
max
R
eq

k
off
R
eq
0 9:15
where R
eq
is the equilibrium response at a given analyte concentration c
0
.
Therefore, the equilibrium signal reects the afnity constant K
a
and dissoci-
ation constant K
d
of the interaction couple.
This can be converted to a format which resembles the 1:1 Langmuir
isotherm:
R
eq

c
0
K
A
R
max
c
0
K
A
1
c
0
R
max
c
0
K
d
9:16
The isotherm is an S-shaped curve if using logarithmic axis for the concentra-
tion, as shown in Figures 9.8b and 9.14. On the application of an analyte
concentration of c
0
K
d
, R
eq
is half of the saturation response R
max
.
9.4 Applications of the Kinetic Model
Although this pseudo-rst-order kinetic model has been used very successfully
in qualitative studies (such as demonstration of interactions between biomol-
ecules), the determination of the kinetic rate constants of binding is often
complicated by the fact that most binding curves deviate from the single
exponential time course expected for a simple pseudo-rst-order reaction.
Apart from the experimental causes (e.g. the sample depletion, noise, drift,
impurity), major concerns for the deviation are focused on mass transport/
rebinding effects, multivalent interactions/avidity effects, heterogeneity in the
immobilized ligands/matrix effects and complex binding mechanisms.
On improving the experimental design (e.g. by using high ow rates and low
surface capacities) and applying advanced analysis algorithms (e.g. the Global
Analysis [27], tting association and dissociation phase data for a series of
concentrations simultaneously), the contribution of most of these effects can be
minimized or even corrected [28].
9.4.1 Surface Hybridization Reactions of Oligonucleotides
The rst set of examples that we discuss concern hybridization studies between
surface-attached catcher probe oligonucleotides and chromophore-labeled
286 Chapter 9
targets from solution. From among the feasible experimental schemes [29], a
few typical ones are summarized in Figure 9.9. The most direct approach is
given in (a): the sensor surface is functionalized by a single-stranded DNA
oligonucleotide catcher probe with a base sequence specic for a target strand
from solution which carries the uorophore. Upon binding, the chromophores
are placed in the resonantly enhanced evanescent eld of the surface plasmon
mode and emit strong uorescence light. The number of emitted photons is
directly related to the number of bound target strands. This scheme will be
discussed in detail below.
An alternative strategy is presented in Figure 9.9b with the corresponding
experimental results given in Figure 9.10. In this case, the probe strand at the
sensor surface carries the chromophore which emits uorescence light at a level
that reects the compromise between the evanescent character of the excitation
eld and the quenching prole for energy transfer, as discussed above. Upon
the binding of the target analyte, the resulting double strand stiens and thus
stretches, thereby pushing the chromophore at the end of the probe strand
further away from the (quenching) sensor metal surface. The result is a net
increase in the uorescence intensity because the (slight) decrease in the optical
excitation is by far overcompensated by the increase in emission intensity owing
to the reduced quenching upon the growth in chromophoremetal distance.
This principle for the recording of a hybridization reaction combines the best of
Figure 9.9 Various detection schemes for hybridization reactions between surface-
attached catcher probes and oligonucleotide targets from solution: (a) the
targets carry the chromophores that emit uorescence photons; (b) the
catchers are labeled with a uorophore which upon hybridization are
placed further away from the substrate surface; (c) the probes are labeled
with donor and the targets with acceptor dyes, leading upon hybridization
to efcient energy transfer.
both worlds: the sensitivity of uorescence spectroscopy with the attraction of a
label-free analyte molecule.
The last concept that is outlined in the schematic cartoon in Figure 9.9c is
based on the popular resonant energy transfer reaction between (the excited
state of) a donor chromophore, attached, e.g., to the probe strand and (the
ground state of) an acceptor dye coupled to the analyte target molecule [30].
Upon hybridization, the two chromophores come sufciently close to each
other to allow for efcient energy transfer, resulting in a variety of spectral
changes of the observed emission that range from donor emission quenching to
sensitized acceptor uorescence emission to the appearance of novel spectral
bands. The strong distance dependence of the energy transfer being efcient
only within the Fo rster radius of typically 57 nm allows for a number of
detailed investigations, e.g. at the single molecule level, of the structural and
dynamical aspects of DNA hybridization reactions in solution and at surfaces.
According to the theoretical treatment of surface hybridization reactions
within the Langmuir model, one has a number of experimental options for the
determination of the relevant parameters, i.e. k
on
and k
o
and from there the
afnity constant K
a
. Starting from an unoccupied, bare probe matrix, one can
follow the association phase of the surface hybrid formation after injection of a
target solution at a concentration that should be, at least, in the range of
the half-saturation value c
1/2
given by the afnity value for this hybrid
(c
1/2
K
d
K
a
1
). Once the saturation coverage (for this given bulk
0 1000
0
2000 3000 4000 5000
20000
40000
60000
80000
100000
120000
140000
time/s
f
l
u
o
r
e
s
c
e
n
c
e
/
c
p
s
+ labeled probe
rinsing
+ target (MM0, unlabeled)
+ 75 %
Figure 9.10 Binding and hybridization experiment between chromophore-labeled
probe strands at the sensor surface and unlabeled targets in solution.
On binding, stretching of the double strand relative to the single probe
strand leads to a net increase in uorescence intensity.
288 Chapter 9
concentration) is reached, one can record the dissociation process by contin-
uously rinsing the ow cell with pure buer solution, thus triggering the
dissociation of the hybrids and the free target strands being washed out.
If one measures these association and dissociation phases systematically at
different concentrations of the injected target solution but for a very limited
time only (e.g. for 10 min), and with regeneration steps between such that each
time association starts from a bare probe, one can derive reliable values for k
on
and k
o
because they will be averaged over the concentration range investi-
gated. This global analysis allows one, in particular, to check whether the
dissociation rate constant, k
o
, is concentration (and coverage) independent as
predicted by the Langmuir model, and it shows whether the association
process, indeed, speeds up linearly with increasing bulk concentration c
0
.
Finally, by monitoring equilibrium coverage as a function of c
0
, e.g. by
recording angular uorescence scans after a plateau has been reached in the
association phase, one can perform titration experiments which yield the
afnity constant directly (cf. Figure 9.8b). By comparing the various param-
eters determined by the different techniques, one can check for internal con-
sistency of the prediction of the Langmuir model and, can test its applicability
to surface hybridization reactions.
The application of kinetic SPFS measurements for the quantication of
hybridization reactions and, in particular, for the discrimination of single
nucleotide polymorphisms (SNPs) between the catcher probe P2 and dierent
target strands is illustrated in Figure 9.11ac. The architecture of the sensor
coating was identical in all three experiments, the only difference being the
sequence of the various 15mer oligonucleotide target strands used [20].
The fully complementary strand T2 (MM0) (cf. Figure 9.11d) binds very
fast at the target concentration employed (c
0
1 mM), reaches then a stable
level of the emitted uorescence intensity and comes o the surface only very
gradually with a barely detectable loss of uorescence intensity upon rinsing
(Figure 9.11a).
The kinetic behavior changes dramatically if a target solution is injected with
a strand sequence that differs by only a single nucleotide within the recognition
sequence of 15 bases (T1, cf. Figure 9.11d): after the injection of the target
solution a slower uorescence increase reects the reduced association rate
constant although the nal intensity measured after about 15 min reaches
almost the same level as in the MM0 case (Figure 9.11b). The most pronounced
difference is seen during the dissociation phase induced by ushing the ow cell
with buer: it takes 23 h, but then the surface-bound hybrids are completely
dissociated and the targets rinsed out of the ow cell. Continuous ow of the
buer solution through the cell prevents rebinding of target compounds to
the sensor surface. As indicated by the full black curve, the whole process
is described by a single exponential decay as predicted by the Langmuir model
[cf. eq. (9.14)].
A further significant change in the kinetic response is seen if a target strand
representing a mismatch 2 (MM2) situation is owing through the cell: the
uorescence intensity barely increases. The full black curves are ts to the
1.5 x 10
6
1.0 x 10
6
5.0 x 10
5
0.0
1.5 x10
6
1.0 x10
6
0.0
1.5 x 10
6
1.0 x 10
6
5.0 x 10
5
5.0 x 10
5
0.0
1000 0 2000 3000 4000 5000 6000 7000
0 15000 5000 10000
0 1000 2000 3000 4000 5000
Time / sec
Time / sec
Time / sec
I
f
l
/
c
p
s
I
f
l
/
c
p
s
I
f
l
/
c
p
s
P2/T2: MM 0
P2/ T1: MM 1
P2/T3: MM 2
(a)
(b)
(c)
(d) Probes:
DNA-P2: 5-biotin-T15-TGT ACA TCA CAA CTA-3
PNA-P2: biotin- TGT ACA TCA CAA CTA-NH
2
Targets:
T1: 3-ACA TGC AGT GTT GAT - cy5 - 5
T2: 3-ACATGT AGT GTT GAT - cy5 - 5
T3:
3-ACA TGC ACT GTT GAT - cy5 - 5
6
O
O
O
N
H
O
NH
2
Figure 9.11 Kinetic experiments with association and dissociation between surface-
attached probe strands P2 and target strands from solution with different
complementarity, i.e. MM0, T2 (a); MM1, T1 (b); MM2, T3 (c). Target
concentration in each case: c
0
1 mM. (d) The base sequences for the
probes and targets, respectively, used in Figures 9.119.13 and 9.19.
290 Chapter 9
experimental data based on the Langmuir model with rate constants summa-
rized in Table 9.1.
A global analysis with the P2/T2 system cannot yield reliable k
on
values
because the dissociation is so slow, as shown in Figure 9.11a, that no mean-
ingful t to the nearly at intensity curve (measured typically for a few minutes
only) would be obtained. However, upon the introduction of a single mismatch
in the base sequence of the target (T1, in Figure 9.11d) the duplex is consid-
erably destabilized, thus the dissociation is enhanced and the loss of uores-
cence intensity can be measured within the 10 min rinsing phase of the global
analysis. This is shown in Figure 9.12a for a probe matrix assembled from
biotinylated peptide nucleic acids (PNA-P2), the neutral mimic of the corre-
sponding DNA-T2 (cf. Figure 9.11d). The recognition sequence was identical
with the DNA P2; however, the spacer to the biotin group was a stretch of six
ethylene oxide-containing units [30] instead of the 15 thymines in the case of the
DNA probe strands.
The rate constants k
a
(k
on
c
0
+k
o
) [cf. eq. (9.12)] derived for the PNA-P2T1
hybridization is plotted in Figure 9.12b. From the slope of the linear depend-
ence on the bulk concentration c
0
the association rate constant is obtained as
k
on
3.1 10
3
M
1
s
1
, slightly lower than that for the DNA-P2T1 case (cf.
Table 9.1).
Upon ushing the cell with buer and monitoring the dissociation phase, one
can determine the k
o
rate constant(s) as a function of coverage according to
eq. (9.14). The corresponding measurements are also shown in Figure 9.12a.
Note that after starting the buer solution an instant decrease in the uores-
cence intensity can be observed, indicating the starting point of the dissociation
phase. This drop originates from removal of free uorophores (targets) in the
solution near the surface ushed out by the buer. From the decrease in the
uorescence intensity during the rinsing process the dissociation rate constant
was determined as k
o
2.5 10
4
s
1
, which is the average of all values
obtained by tting each dissociation part of the measurement (during rinsing)
using eq. (9.14). Thus, the afnity constant, K
a
(K
a
k
on
/k
o
), is found to be
K
a
1.2 10
7
M
1
for the PNA-P2T1 MM1 hybrid, which is of the same
order as for the DNADNA hybrid.
For a titration experiment, a series of angular scans were taken after target
solutions of 1, 5, 10, 20, 50 and 100 nM had been injected and equilibrium for
each new bulk concentration was reached. The corresponding series of angular
scans is shown in Figure 9.13a. Several features are noteworthy. (1) No
Table 9.1 Rate constants k
on
and k
o
and the afnity constant K
a
k
on
/k
o
,
derived from the kinetic experiments shown in Figure 9.11.
Parameter T
2
(MM0) T
1
(MM1) T
3
(MM2)
k
on
(M
1
s
1
) 3.7 10
4
8.9 10
3
10
k
o
(s
1
) 7 10
6
3.7 10
4
7.7 10
4
K
A
(M
1
) 5.3 10
9
2.4 10
7
1.3 10
4
significant shift of the surface plasmon minimum angle was observed, the
various reectivity curves are virtually superimposed, indicating negligible
increase in the optical thickness upon forming the PNADNA duplex. (2)
The bulk solution uorescence excited by light transmitted through the 50 nm
Au substrate at 451 (below the critical angle y
c
47.31) was measured at each
concentration. As demonstrated in Figure 9.13b, this intensity is a linear
function of the target concentration (from 1 to 100 nM), due to the direct
excitation of the uorophores in the bulk solution. (3) In Figure 13c the
open squares are the data from angular scans maximum intensity (after rinsing
0 400 800 1200 1600
Time/sec
50
100
200
500
10
20
0 100 200 300 400 500 600
0.0012
0.0018
0.0024
0.0030
0.0036
k
a

/
s
e
c
-
1
Concentration c
0
/nM
c
0
/ nM
F
l
u
o
r
e
s
c
e
n
c
e

/

1
0
5

c
p
s
(a)
(b)
1.5
1.0
0.5
0
Figure 9.12 (a) Global analysis of the association and dissociation phase of DNA
target T1 hybridization to PNA-T2 probe matrix in a solution containing
10 mM phosphate buer. (b) k
a
k
on
c
0
+k
o
obtained from tting the
data (open squares) of (a) as a function of target concentration c
0
.
292 Chapter 9
for 3 min) and the solid curve is a simulated Langmuir t using eq. (9.16) with
an afnity constant of K
a
1.7 10
8
M
1
, more than an order of magnitude
lower than the afnity constant for the DNADNA hybrid (cf. Table 9.1).
The observed behavior of binding of the targets from solution to the surface-
immobilized probe strands can be essentially understood from the Langmuir
model. The isotherms describing the three different match/mismatch situations
in Figure 9.11 are displayed in Figure 9.14 by the three different S-shaped
binding curves (solid, dashed and dotted black curves). They are separated
0.0
5.0x10
4
1.0x10
5
1.5x10
5
F
l
u
o
r
e
s
c
e
n
c
e

@
4
5
/
c
p
s
Concentration c
0
/nM
0 20 40 60 80 100
0 20 40 60 80 100
0.0
8.0x10
4
1.6x10
5
2.4x10
5
F
l
u
o
r
e
s
c
e
n
c
e

@

5
6
.
6
/
c
p
s
Concentration c
0
/nM
100 n
50 n
20 n
10 n
5 n
1 n
backgroun
c
(2
(1
45 50 55 60 65
0.0
0.2
0.4
0.6
0.8
1.0
Angle/deg
R
e
f
l
e
c
t
i
v
i
t
y

R
0.0
0.8
1.6
100 nM
50 nM
20 nM
10 nM
5 nM
1 nM
background
c
0
(1
Kinetic Exp. Transmission Exp.
F
l
u
o
r
e
s
c
e
n
c
e
/
1
0
5

c
p
s
2.4
(a)
(b)
(c)
Figure 9.13 (a) Angular scans taken after saturation was reached for T2 target
solutions (1, 5, 10, 20, 50 and 100 nM), hybridizing to a PNA-P2 matrix,
together with the background uorescence intensity. Open squares are
reectivities. Angular uorescence curves before (open symbols) and
after rinsing (full symbols) the ow cell with buer following the
equilibrium binding of targets from solution. (b) Fluorescence intensity
measured at y 451, (below critical angle), giving information of the
bulk concentration c
0
. (c) Plots (open squares) of the maximum uores-
cence intensity (y 56.61) taken from (a) versus target concentration c
0
.
Solid line: t by a Langmuir isotherm with K
a
1.7 10
8
M
1
.
from each other on the (bulk) concentration axes by the 23 orders of magni-
tude difference in their respective afnity constants, K
a
, resulting in a shift of
the curves according to the respective half-saturation constants K
d
(cf. the red
dashed line at F50% and the corresponding red dotted lines to the abscissa).
Based on these plots one expects, for the kinetic experiments performed at a
bulk concentration of c
0
1 mM (cf. the blue dashed line in Figure 9.14), that
the intensities for both the MM0 and the MM1 target strands reach almost the
same level near saturation because this concentration is way above the respec-
tive half-saturation concentrations for both targets. On the other hand, the
MM2 strands at this concentration are barely hybridizing to the interfacial
matrix, because for this low-afnity pair the bulk concentration is not high
enough to induce any significant binding.
This experiment also proves that unspecic adsorption does not play any role
in the observation of these binding curves: most of the molecular details of the
three analytes are identical, e.g. the length of the strand, the charge (number
and density) and almost all bases. Hence any non-specic interactions are
virtually the same for all targets. This points to one important design principle
for optimized SNP detection: by quantifying the afnities for targets of dier-
ent mismatches to the surface probe, one can identify the bulk concentration
best suited to differentiate between the two sequences in a (static end-point)
titration experiment. In order to discriminate MM1 from MM0 in the above
example, a bulk concentration of c
0
3 nM would be ideal because this
1E-13 1E-12 1E-11 1E-10 1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-3 0.01
0
20
40
60
80
100
Probe P2/Target:
T3 (mismatch 2)
T1 (mismatch 1)
T2 (mismatch 0)
S
u
r
f
a
c
e

c
o
v
e
r
a
g
e

/
%
Concentration c
O
/M
Figure 9.14 Langmuir isotherms of three targets, MM0, MM1 and MM2, with
significantly different c
1/2
(K
d
) values (cf. the dashed red line at
F50%). The concentration of the kinetic experiments displayed in
Figure 9.11 is given by the blue line at c
0
1 mM.
294 Chapter 9
concentration is right in between the c
1/2
values of the two strands. The con-
centration of 1 mM used in the kinetic measurements shown in Figure 9.11
obviously was too high. At this concentration both targets hybridize to the
extent that almost a complete monolayer is formed, and hence would give
about the same intensity when measured with a uorescence scanner after
binding. Of course, in a sensor platform such as the one described here, i.e. with
the full in situ and real-time recording of the whole association and dissociation
process by surface plasmon uorescence spectroscopy, one could easily dis-
criminate between different mismatches by the significantly different rate
constants. Already after a few minutes or even seconds of recording the early
phase of the association process a clear discrimination of the MM0 from the
MM1 targets is feasible.
When dealing with extremely dilute solutions, one needs to take into account
limitations given by the diffusion of the analyte from the bulk of the ow cell
across the unstirred layer to the sensor/liquid interface. This mass transfer-
limited regime is sketched schematically in Figure 9.15. With the closed loop of
the ow cell one can easily switch, e.g. from buer to the injection of a solution
with a particular (low) analyte concentration. As was outlined above, the
analyte molecules, A
0
, then have to diffuse across the unstirred layer and bind
to the functional groups, B, at the sensor surface. This leads to a linear increase
in the uorescence intensity as a function of time after injection of the analyte
solution [cf. eq. (9.7)] with a slope that is a linear function of the bulk
concentration of A
0
.
Peristaltic pump
Sample
Flow cell
and fluidic system
Figure 9.15 The sensor ow cell with closed loop of analyte solution being pumped
from the sample container through the measuring cell passing along the
sensor surface (left). For ultra-low analyte concentrations this mass
transfer-limited regime leads to a linear increase in the sensor (uores-
cence) signal proportional to the bulk analyte concentration c
0
.
An example of mass transfer-limited binding curves is given in Figure 9.16b.
The analyte in this case was an amplicon of 125 bases shown schematically in
Figure 9.16a, prepared by heating the double-stranded PCR products to 96 1C,
followed by injecting the solution into a low ionic strength buer at 0 1C
[31,32]. This procedure was shown to separate the duplexes and to stabilize the
single strands via Coulombic repulsion, thus preventing rapid re-hybridization
in the bulk. Since the catcher strands at the sensor surface were PNAs, the
surface hybridization of the PCR amplicons to the catcher sequences at the
detector was not inuenced by the low ionic strength. The bulk concentrations
used in these experiments (1100 pM) were much lower than the half-saturation
value (about 3 nM [30]) for this hybridization reaction. Moreover, only the very
0.1 1 10 100 1000
baseline stability limit
100 fM
S
l
o
p
e
/
c
p
s

m
i
n
-
1
Concentration/pM
0 50 100 150 200 250
10 pM
1 pM 2 pM
F
l
u
o
r
e
s
c
e
n
c
e
/
1
0
4

c
p
s
Time/min
10
5
10
4
10
3
10
2
10
1
100 pM
8
4
0
- - - - -
(b)
(c)
(a)
Figure 9.16 (a) Schematics of the supramolecular architecture with catcher oligonuc-
leotides (blue) assembled at the sensor surface for the study of hybrid-
ization reactions with PCR amplicons (black/red); (b) time-dependent
recordings of the diffusion and binding of amplicons from low-
concentration solutions to the surface-attached PNA probes; (c) calibration
curve obtained by plotting the slopes of the linear increase in the
uorescence intensity in (b) as a function of the respective bulk analyte
concentration. The limit of detection [as dened by the intersection of the
calibration curve (red line) with the baseline stability limit (blue dashed
line)] is at c
0
100 fM.
296 Chapter 9
early stage of the diffusion/adsorption process was monitored. The linear
increase in the uorescence intensity was recorded for a maximum of 30 min
in the case of the lowest concentration to only a few seconds for the highest
concentration (c
0
100 pM). In each case this was followed by a rinsing step
and regeneration induced by 10 mM NaOH to break all hybrids.
By plotting the slope of the binding curves as a function of the corresponding
bulk concentration, one nds a linear relation as shown in Figure 9.16c, which
further proves the validity of the treatment of the data by the mass transfer-
limited regime. The intersection of this calibration plot with the baseline
stability limit, which indicates the background drift of the uorescence intensity
when measured in the absence of any target molecules (see also Figure 17b),
yields a limit of detection of c
LOD
100 fM, far below that of label-free
detection with SPR.
E
v
a
n
e
s
c
e
n
t

f
i
e
l
d
Prism
Gold
SAM
Dextran
Antigen
Dye-labeled
antibody
Gold
Detection limit
LOD500 aM
-10 0 10 20 30 40 50 60
5x10
3
6x10
3
7x10
3
8x10
3
9x10
3
1x10
4
1x10
4
1x10
4
1x10
4
1x10
4
2x10
4
F
l
u
o
r
e
s
c
e
n
c
e
/
c
p
s
Time /minutes
(a)
(b)
(c)
10
-2
10
0
10
2
10
4
10
6
10
0
10
3
10
6
Baseline deviation level
B
i
n
d
i
n
g

s
i
g
n
a
l
/
c
p
s

m
i
n
-
1
Concentration / fM
(d)
0 200 400 600
1
2
3
333 fM
buffer
3.3 pM
33 pM
67 fM
333 fM
F
l
u
o
r
e
s
c
e
n
c
e
/
1
0
4
c
p
s
Time / minutes
F
l
u
o
r
e
s
c
e
n
c
e
Figure 9.17 Protein binding studies in the mass transfer-limited regime. (a) Interfacial
binding matrix based on a dextran brush. (b) Control of the baseline
stability: solid lines give drift stability, the average value of which (plus
three times the standard deviation) denes the limit of detection (LOD).
(c) Time-dependent uorescence increase upon binding of chromophore-
labeled antibodies from solutions of different (low) concentrations to the
dextran matrix-immobilized antigens. (d) Calibration curve obtained by
plotting the slopes of the linear uorescence increase taken from (c) as a
function of the bulk concentration. Note the linear dependence
(slope 1), covering a concentration range of six orders of magnitude.
9.4.2 Protein Binding Studies
All examples discussed so far were based on the recording of uorescence from
chromophores covalently bound to the analyte molecule of interest. This does
not constitute a major limitation for the detection of PCR amplicons because
the use of uorophore-labeled primers is well established. However, there might
be other situations where the attachment of a uorescent label is not possible or
bears the risk of changing the characteristics of the interaction with afnity
partners in a significant way, e.g. when dealing with proteins. In this case one is
interested in developing detection schemes that offer the sensitivity provided by
uorescence spectroscopy, but do not require the analyte to carry the uor-
ophore directly. Competitive replacement assays, obviously, fulll that require-
ment: the analyte of interest replaces a chromophore-labeled ligand that was
preadsorbed on the binding sites at the sensor surface.
Much of the optical principles for uorescence-based detection of hybridi-
zation reactions between a surface-attached probe oligonucleotide and a target
strand from solution applies to the study of protein binding reactions at
surfaces in very much the same way [33]. However, generally, proteins are
significantly more delicate in their behavior and, in particular, are more
sensitive to the proper control of their interfacial interaction potentials. A
direct consequence is the need for a significantly more complex interfacial
architecture for the functionalization of sensor surfaces which allows for the
attachment of ligands or other binding partners in the proper orientation,
exibility, without the loss of specicity or even the risk of partial denaturing
and with the control of any unwanted non-specic binding (NSB).
A number of different strategies for the attachment of binding partners for
the recognition and binding reaction with proteins from solution have been
reported. For the case of bioafnity studies using surface plasmon uorescence
spectroscopy, the dextran brush introduced by Biacore turned out to be ideally
suited for this purpose [34]. As can also be seen in Figure 9.6, the extent of the
brush with its roughly 100 nm thickness matches very well the extent of the
evanescent eld of the surface plasmon. The functionalization of the brush via
its COOH groups along each dextran chain allows for the covalent attachment
of antigens. If the antigens are recognized by the corresponding chromophore-
labeled antibody, binding with high afnity from solution, strong uorescence
emission occurs when excited by a resonant surface plasmon eld. This is
shown schematically in Figure 9.17a. This matrix also minimizes intensity
losses due to quenching, because most of the binding events occur sufciently
far away from the metal surface. Figure 9.17b shows a series of stability tests,
i.e. recordings of the baseline uorescence as a function of time without any
analyte injected. The average value of the slope of these curves (plus three times
the standard deviation) gives the baseline stability limit needed for the deter-
mination of the limit of detection. The mass transfer-limited linear increase in
the uorescence intensity seen after the injection of antibody solutions can be
sensitively monitored down to extremely low concentrations, as shown in
Figure 9.17c. The plot of the slope of the intensity increase as a function of
298 Chapter 9
the bulk protein concentration again gives a straight line (Figure 9.17d),
conrming the validity of the diffusion model. The intersection of the calibra-
tion line with the baseline stability limit again denes the LOD of this detection
method and is found to be as low as c
LOD
500 aM (5 10
16
M), far below
LODs ever observed with SPR [21].
The highest concentration used in the above measurements was 1 nM, barely
reaching the half-saturation concentration of this antibodyantigen pair with
its afnity in the range of K
A
2 10
8
M
1
. However, when working with a
6.7 nM solution one can monitor the uorescence signal at various stages of the
antibody adsorption and correlate it with the resulting angular shift of the SPR
signal at this coverage. In this way, one can calibrate the observed uorescence
intensity in terms of the number of proteins binding to the matrix. For the LOD
of 500 aM one nds that the monitored uorescence increase corresponds to the
diffusion and binding of about 10 proteins per mm
2
per minute, approaching
the range of single molecule detection.
All uorescence data presented so far were based on the recording of quasi-
monochromatic uorescence intensity by photon counting with the emitted
light passing through a narrow bandpass lter for stray light suppression.
However, an alternative means of light detection, this time with the full spectral
information of the emitted uorescence, is given by the use of a spectrometer
that disperses the light into its spectral components. In this way, not only the
existence of an analyte can be monitored through its uorescence emission, but
also additional processes at the sensor surface which lead to spectral changes of
the emission can be recorded.
The principle of this approach is illustrated in Figure 9.18 for a protein layer
coupled to the sensor matrix by the well-established His-tag strategy. A generic
streptavidin monolayer was introduced to couple a layer of biotinylated
nitrilotriacetic acid molecules which after activation by exposure to 500 mM
NiCl
2
solution can be used to attach a monolayer of a recombinant protein/
chromophore aggregate, i.e. the light-harvesting complex LHCII known from
chloroplasts, that was modied by a stretch of six histidines at the C-terminus.
Figure 9.18a illustrates the nal architecture, and Figure 9.18b shows the SPR
protocol of the assembly process: after binding of the Ni
21
ions, the injection of
the protein (dark arrows) leads to the formation of an oriented LHCII
monolayer of a density which is controlled by the streptavidin matrix. After
rinsing the multilayer with pure buer (white arrows) followed by a 0.35 M
solution of EDTA, a high-afnity chelator for divalent ions, all Ni
21
ions are
released and the aggregates can be specifically disassembled between the NTA
on the surface and the His-tag at the protein, which is then rinsed out. This
process is fully reversible, as demonstrated by the virtually identical SPR
responses of the repeated assembly and disassembly steps.
After binding the unmodied LHCII bands found for the protein in solution
can be identied also for the surface-immobilized protein layer (Figure 9.18c,
red curve) in the uorescence spectrum of wavelength-resolved SPFS. Injecting
the unmodied protein mixed with protein labeled with an acceptor dye
(DY-730 from Dyomics, Jena, Germany), the blue curve in Figure 9.18c
indicates the occurrence of an efcient intermolecular energy transfer between
the proteins in the monolayer. This is documented by the appearance of the
red-shifted emission shoulder that originates from the energy transfer from the
unmodied donor protein (with its 12 chlorophyll and three carotenoid mole-
cules in the complex) and the lower-energy acceptor chromophore chemically
attached to some of the proteins. The degree of energy transfer depends
strongly on the molar ratio of the two proteins in the immobilized monolayer
(not shown).
9.5 Novel Approaches to SPFS
In most cases, the Kretschmann conguration will be the coupling scheme of
choice not only for SPR but also for SPFS because the evanescent character of
Wavelength/nm
Time/min
R
e
f
l
e
c
t
i
v
i
t
y
(a)
(c)
(b)
I
n
t
e
n
s
i
t
y

/

1
0
4

c
p
s
6
80 40 0
4
2
0
0.20
0.30
600 700 800 900
120
0.35M EDTA
500M NiCl
2
Figure 9.18 (a) Interfacial architecture assembled from a binary thiol SAM, a layer of
streptavidin, biotinylated nitrilotriacetic acid (NTA), Ni
21
ions and a
layer of His-tagged light harvesting complex (LHCII) proteins, some with
attached acceptor chromophore. (b) Assembly protocol with injection of
LHCII after binding of Ni
21
ions (dark arrows), rinsing step (light
arrows) and injection of EDTA for the removal of Ni
21
. (c) Wave-
length-resolved uorescence spectra of a monolayer of LHCII (red curve)
and of a similar monolayer in which a fraction of the protein complexes
was chemically modied by an acceptor dye (DY-730 from Dyomics)
emitting at l 758 nm (blue curve).
300 Chapter 9
the surface wave being excited from the back of the prism allows for easy
attachment of the ow cell from the front side and the recording of the resulting
uorescence emission o the base plane of the prism. However, the alternative
scheme for surface plasmon excitation by laser light based on a surface grating
structure offers a number of advantages for biosensing in general and for
surface plasmon uorescence spectroscopy in particular. For example, the need
for a relatively high index prism for the momentum matching condition when
working in aqueous buer solutions can barely be met by cheap plastic prisms
and requires expensive specialty glasses. Even then, relatively high angles of
incidence are needed. In the following, novel instrumental solutions are given
for the SPFS challenge.
9.5.1 Grating Coupling for SPFS
The use of a grating coupler offers an important element of exibility in that the
choice of the grating periodicity, L, and hence the modulus of the grating
vector, G2p/L, allows for the tuning of the angle of resonant excitation, y,
for a given laser wavelength to any desired value from normal incidence to
essentially y 901, according to the momentum matching condition for grating
coupling:
k
psp
k
photon
siny mG 9:17
The grating coupler is illustrated in Figure 9.19. Since the laser is incident from
the front side, as shown in Figure 9.19a, the Au metal coating can be of any
thickness provided that it remains opaque and no light is reected from the
back side (an advantage in the context of quality control for the mass produc-
tion of such sensor chips). Moreover, the substrate can be made of any
material, e.g. a plastic chip that can be surface structured very easily and at
low cost by hot embossing with a master grating (which offers the additional
advantage that all gratings are virtually identical in terms of their grating
constant and amplitude, Fourier components, roughness, etc.). As in this case
the laser beam passes through the ow cell before exciting the surface plasmon
wave at the sensor (grating) surface, at higher analyte concentrations, this may
lead to an additional uorescence contribution, i.e. from chromophore-labeled
analyte molecules owing through the cell, resulting in a so-called bulk-jump
upon injection of the analyte solution. However, for very dilute concentrations
this is of no concern.
Figure 9.19b gives an example of the angular dependence of the reectivity
and of the corresponding uorescence for both, p- and s-polarized laser
excitation. At higher angles the reectivity curves show the appearance of
the 1st-order diffraction intensity, resulting in a slight decrease of the spec-
ularly reected intensity (not seen, of course, in the uorescence intensity which
only reects the interfacial intensity of the surface plasmon mode). Note also
the pronounced angular shift between the reectivity minimum angle and the
angle for maximum uorescence intensity.
A typical concentration titration experiment between surface bound
DNA-P2 probes and T1 targets (MM1) in solution is given in Figure 9.19c.
The interfacial architecture on the grating surface was the binary SAM/
streptavidin-based matrix (shown in Figure 2) used also for the hybridization
experiments in the Kretschmann conguration. The open squares are the
experimental uorescence data points and the red curves are calculations based
on the Langmuir model with k
on
and k
o
parameters summarized in Table 9.2.
The k
on
values were obtained by tting the association phases recorded after
injecting bulk solutions of different concentrations, while the nal dissociation
experiment (at c
0
0 nM) with its exponential decay of the uorescence inten-
sity could be tted by a single k
o
rate constant (cf. Table 9.2). Using
the various k
on
values and the k
o
value of the dissociation experiment for
the determination of the K
A
values, one nds, in agreement with the Langmuir
0 50 100 150 200 250
Concentration/nM
4
3
2
1
0
2 4 6 8 10 12 14 16 18 20 22
p-light SPS
s-light SPS
p-light fluorescence
s-light fluorescence
R
e
f
l
e
c
i
v
i
t
y

R
Incident angle/deg
F
l
u
o
r
e
s
c
e
n
c
e
/
1
0
6

c
p
s
1.0
0.5
0
2
3
4
0
1
PMT
O-ring
Cover glass Au coating
substrate
flow cell

0 100 200 300 400 500 600 700
0
1x10
5
2x10
5
3x10
5
4x10
5
5x10
5 Rinse
10 nM
(without t-20)
10 nM
20 nM
50 nM
100 nM
250 nM
F
l
u
o
r
e
s
c
e
n
c
e
/
c
p
s
Time/minutes
(a)
(b)
(c)
(d)
out
in
Figure 9.19 (a) Schematics of the excitation of surface plasmons by a grating on the
back of a ow cell, with uorescence detection through the cell. (b)
Angular reectivity (open black symbols) and uorescence scans (closed
blue symbols) in grating coupling. The Au gating was functionalized with
a probe layer to which a chromophore-labeled target strand was hybrid-
ized. (c) Titration of surface attached P2 probe matrix with DNA target
T1 at different concentrations. The association and dissociation rate
constants were derived from the ts to the data points (red curves) and
are given in Table 9.2. (d) Langmuir isotherm constructed from the
uorescence intensities from (c) after the equilibrium coverage at each
new concentration was reached, with data t using K
A
7 10
7
M
1
(red
curve).
302 Chapter 9
model, that the afnity constant for the MM0 hybrid does not depend on the
bulk concentration. The values obtained vary within a factor of two, well
within the range of accuracy achievable in these experiments.
As shown in Figure 9.19d, by plotting the uorescence intensities recorded at
the end of each association phase after injection of a new bulk solution,
representing the newly established equilibrium surface coverage, one obtains
again a Langmuir adsorption isotherm similar to that constructed from angular
uorescence scans presented for a Kretschmann experiment in Figure 9.13. The
red curve is a t with an afnity constant of K
A
6.7 10
7
M
1
. Given the fact
that such an equilibrium titration experiment is a fundamentally different
approach for the determination of afnity constants, the agreement with the
kinetically determined K
A
values adds strong support to the application of the
Langmuir model for the analysis of hybridization reactions between surface-
attached oligonucleotide probes and target strands in solution. Deviations are
only seen for very long (highly charged) PCR amplicons in low ionic strength
buers or for high target densities on the sensor surface. These deviations are
well explained by Coulomb interactions between probes and targets and
between neighboring hybridization sites as we are dealing with highly charged
interfaces and relatively dense oligo-electrolyte brushes.
9.5.2 Long-range Surface Plasmons for SPFS
Before the introduction of another very promising mode of operation in SPFS,
i.e. the use of long-range surface plasmons (LRSP) as the excitation light
source, we briey review a few basics of normal surface plasmons. These
non-radiative, surface-bound electromagnetic modes propagate along a (noble)
metal/dielectric interface with an optical eld that peaks at the interface and
decays exponentially both into the metal and into the dielectric. The classical
Kretschmann conguration with a prism as the coupling element allows for the
required energy and momentum matching between the exciting (laser) photons
and the surface plasmon modes, provided that the refractive index of the prism
is sufciently high compared with the dielectric medium in contact with the
metal surface. For (bio-)sensor applications with the transducer operating
typically in water (n 1.33 at l 633 nm) this requires that the thin metal
Table 9.2 Rate constants, k
on
and k
o
, and the afnity constant, K
A
k
on
/k
o
,
determined from the kinetic titration experiment presented in
Figure 9.19.
Concentration (nM) k
on
(M
1
s
1
) k
o
(s
1
) K
A
(M
1
)
10, no Tween 3.4 10
4
14 10
7
20 4.7 10
4
19 10
7
50 2.3 10
4
9.2 10
7
100 2.4 10
4
9.6 10
7
250 2.4 10
4
9.6 10
7
0 2.5 10
4
layer that guides the surface wave is in optical contact with a high index glass,
e.g. LaSFN9 (Schott) prism base.
This coupling geometry is shown schematically in Figure 9.20a. A typical
angular reectivity scan simulated for this conguration at a laser wavelength
of 633 nm and for a 50 nm Ag layer in contact with water is given in
Figure 9.20b. One important feature of the resonantly excited bound wave
seen as the narrow dip in the reectivity spectrum is the extent of its evanescent
eld (at resonance) into the dielectric medium. This is given in Figure 9.20c for
the architecture in Figure 9.20a, but simulated for 40 nm of Au in order to
match the experiment described below. One nds the typical penetration depth
of PSP modes in the region of L
z
178 nm (red dashed lines in Figure 9.20c).
metal
n =1.33
40 45 50 55 60 65 70
0.0
0.2
0.4
0.6
0.8
1.0
R
e
f
l
e
c
t
i
v
i
t
y
-500 0 500 1000 1500 2000
0.0
0.5
1.0
1.5
2.0
2.5
F
i
e
l
d

I
n
t
e
n
s
i
t
y

(
H
y
)
40 45 50 55 60 65 70
0.0
0.2
0.4
0.6
0.8
1.0
R
e
f
l
e
c
t
i
v
i
t
y
(e)
(f)
(c) (b)
(a)
Incident angle / deg
Incident angle / deg z / nm
600 nm
1.33
metal
(d )
n =1.33
500 0 500 1000 1500 2000
0.0
0.5
1.0
1.5
2.0
2.5
F
i
e
l
d

I
n
t
e
n
s
i
t
y

(
H
y
)
z / nm
50 nm Ag 40 nm Au
50 nm Ag 40 nm Au
Figure 9.20 (a) Schematics of the Kretschmann conguration for surface plasmon
excitation. (b) Angular reectivity scan simulated for a high refractive
index prism (n 1.85 at l 633 nm, 50 nm Ag with e 17 +0.7i, in
contact water of n 1.33). (c) Optical eld distribution, normal to
the interface for the architecture given in (a) but calculated for a 40 nm
Au layer (e 12.3 +1.29i) (dashed blue lines). For easier comparison
the eld is normalized to its value at the metal/water interface. (d)
Kretschmann conguration for the excitation of LRSP: a low refractive
index cladding layer is deposited on the high refractive index prism
(n 1.33), followed by the metal coating (Ag with d 50 nm), operating
in water. (e) Reectivity scan simulation for the architecture given in (d).
Note the two minima at lower and at higher angles than in normal
SPR in (b). (f) Optical eld distribution for a layer architecture as in (d),
but calculated for a thin Au layer (d 40 nm and e 12.3 +1.29i, with
a cladding layer n 1.29) to simulate the experimental conditions. Note
the substantial asymmetry in the eld distribution with the significantly
enhanced decay length into the analyte solution {in order to show this
difference from normal SPR better [cf. (c)], we scaled both elds to the
value at the metal/water interface}.
304 Chapter 9
The evanescent character of this surface plasmon mode guarantees the surface
sensitivity and selectivity of the technique which does not suffer from bulk
contributions to the detected (uorescence) signal, e.g. by scattering from larger
objects such as cells.
If a very thin metal lm is sandwiched between two dielectric media of
(nearly) identical refractive indices, n
d
, plasmon modes excited at the two
opposite interfaces will interact with each other provided that the metal
layer is sufciently thin (do40 nm), hence the optical elds within the
metal start to overlap, establishing a transverse standing wave. This interaction
lifts the dispersion degeneracy of the two identical evanescent waves and
two new, coupled modes appear, a symmetrical and an antisymmetric wave
(referring to their transverse electric eld distribution) [35]. The latter has
attracted considerable interest, because its electric eld across the metal
lm, responsible for the energy dissipation by the lossy metal, is largely reduced
and the propagation length of the mode is considerably increased. Hence
this mode is also called long-range surface plasmon (LRSP) as opposed
to the short-range surface plasmon (SRSP) mode which is subject to enhanced
dissipation.
A suitable way to excite LRSPR is the prism-coupled Kretschmann cong-
uration given in Figure 9.20d [36,37]. In this case, a high refractive index glass
prism is rst coated with a low refractive index cladding layer (e.g. PTFE,
n E1.33 and a thickness of d E600 nm), followed by the deposition of a 50 nm
thin Ag layer. The simulated angular reectivity spectrum of such a sandwich
sample in water (n 1.33 at l 633 nm) is given in Figure 9.20e. The LRSP
and SRSP modes can be seen as the sharp dip in the reectivity spectrum at
small angles and the much shallower and broader feature at higher angles,
respectively, in roughly symmetrical angular positions relative to normal SPR
(cf. Figure 9.20b and e). The magnetic eld prole, H
z
, of the LRSP normal to
the planes of the sandwich layers, scaled to its value at the metal/analyte
solution (water) interface, is given in Figure 9.20f. For this simulation we chose
40 nm gold as the metal layer and a refractive index for the PTFE cladding of
n 1.29 in order to match the actual experimental conditions (see below). The
asymmetric refractive index prole of the cladding layer as substrate and water
as superstrate with their slightly different refractive index and also with their
different thicknesses (500 nm vs. innite, respectively) results in a rather asym-
metric eld distribution on both sides of the metal lm.
Reduced damping is responsible for the extended propagation of LRSPs
and also for the narrow angular resonance seen in the reectivity spectrum in
Figure 9.20e. This has triggered interest in using these modes also in optical
biosensors for detecting thin coatings covering the metal lm, e.g. by an
adsorbed protein or DNA analyte layer [8]. The sharp dip in the resonance
with its angular position shifting up the binding of these biopolymers to slightly
higher values should result in a change in reectivity (xed-angle mode) which
should be substantially larger for an LRSP mode than for a classical SPR wave.
However, the extension of the LRSP optical eld reaching out much further
into the buer solution reduces the effect of a thin adsorbed coating on the
dispersion of the plasmon mode, hence the shift of the mode, Dy, is consider-
ably lower than in SPR.
When using LRSP in conjunction with surface-plasmon eld-enhanced
uorescence detection, the analyte molecules that carry a chromophore label
and bind from solution to a correspondingly functionalized sensor surface will
be excited by the evanescent modes of the surface plasmons with the optical
intensity of the LRSP eld being significantly enhanced at the interface and
extending much further out into the solution compared with the situation in
normal SPR (compare the red dashed lines in Figure 9.20c and f ).
The enhancement for LRSP-excited uorescence becomes particularly obvi-
ous if one places the chromophore layer with the analyte further away from the
Au surface [38]. The weaker decay of the LRSP eld out into the analyte
solution should be seen in the difference of uorescence excitation by the two
modes of operation if the chromophores are placed at a certain distance away
from the interface. To this end, two samples were prepared, one for normal
SPR and another one suited for LRSP excitation. In both cases the test analyte,
i.e. the chromophore-labeled antibodies, were adsorbed on a 500 nm thick
PTFE layer on top of the Au substrate in order to probe the differences in the
decay lengths of the two plasmon elds. As expected, the measured uorescence
excited by the normal SPR almost completely decayed to a barely detectable
value (Figure 9.21, full triangles), whereas the uorescence intensity of
the identical protein layer probed by the LRSP eld shows a high value
(Figure 9.21, full circles): the peak intensity ratio amounts to a factor of 33.
9.5.3 Fluorescence Imaging and Color Multiplexing
As an example of the use of surface plasmon uorescence microscopy for
hybridization studies in the format of an mn sensor array, we present data
obtained from experiments with quantum dots (QDs) as uorescent probes.
QDs are small, inorganic, semiconducting nanocrystals that possess unique
optical properties, the most significant being (i) their broad absorption spectra
and (ii) the composition- and/or size-engineered emission properties, making it
possible to excite different QDs simultaneously with a single wavelength light
source, but monitoring their luminescence light emitted at different wave-
lengths in a color-multiplexed recording mode [39].
The example for the parallel read-out of bioafnity reactions that we
briey describe concerns a simple 4 1 array composed of four Au stripes
(cf. Figure 9.23). With this array being oriented relative to the exciting laser beam
at an angle near the SPR reectivity minimum, corresponding to the highest
surface plasmon eld enhancement and thus the highest uorescence intensities,
the various target solutions were injected into the ow cell. Upon hybridization
of the targets to the oligonucleotide catcher probe on the Au electrodes, the
chromophore tags were excited by the evanescent tail of the propagating surface
plasmon waves. The uorescence photons emitted from the electrode array were
imaged by a color CCD camera as shown in Figure 9.22.
306 Chapter 9
laser
polarizers
PC
beam
expander
goniometer
objective lens
CCD - camera
Figure 9.22 Experimental setup in the Kretschmann conguration for surface
plasmon eld-enhanced uorescence microscopy.
46 48 50 52 54 56 58
0.0
0.2
0.4
0.6
0.8
1.0
Incident Angle/deg
R
e
f
l
e
c
t
i
v
i
t
y
/
R
0.0
4.0x10
4
8.0x10
4
1.2x10
5
1.6x10
5
F
l
u
o
r
e
s
c
e
n
c
e
/
c
p
s
Reflectivity
Fluorescence
LR SP
SPR
Figure 9.21 Comparison of the angular reectivity scans recorded with normal SPR
(open circles) and LRSP (open triangles), together with the simultane-
ously measured angular uorescence intensity curves (full circles for
SPFS and full triangles for the LRSP uorescence). The setup consisted
of a prism, 500 nm thick cladding layer (in LRSP excitation), Au layer
(40 nm in both setups, yellow slice) and 500 nm PTFE coating on top of
the Au (both setups) and the chromophore-labeled protein layer (red
slice) adsorbed from solution.
The functionality of the four stripes in Figure 9.23 was chosen as follows: the
top one was covered by a poly(ethylene glycol) (PEG)thiol passivation SAM,
working as a negative control. The other three Au stripes were, from the
bottom, functionalized with P1, P2 and a mixture of P1 and P2, respectively
(cf. the nucleotide sequences given in Figure 9.11d). The target sequence T1
conjugated with quantum dots emitting at l 655 nm (T1
0
-QD
655
with a red
emission color) and the targets T2, coupled to quantum dots with an emission
wavelength of l565 nm (T2
0
-QD
565
emitting in the green) could both
be excited with a green HeNe laser at l 543 nm as the light source, using
Cr/Ag/Au (2/30/7 nm) as a multi-layer metal lm thermally evaporated via a
mask on to the glass slide for the preparation of the 4 1 array [40].
The DNA hybridizing experiment presented here was initiated by injecting
rst a c
0
200 nM solution of the red-emitting T1
0
-QD
655
target sample in PBS.
Next, the T2
0
-QD
565
solution was applied to the system and allowed to
hybridize to the surface for some time. Now the green uorescence of the
QD
565
could be observed from the electrode that was exclusively P2 function-
alized. The electrode containing a mixture of P1 and P2, however, changed its
color from red to yellow due to the red/green/blue (RGB) color addition of the
green uorescence originating from T2
0
-QD
565
with the red uorescence
from the targets T1
0
-QD
655
, as has been observed in a similar way for the
case of inkjet prepared sensor spots [39] (cf. the uorescence spectrum given in
Figure 9.23). Only the electrode protected with the PEGthiol shows no uo-
rescence signal, conrming the excellent passivation properties of a PEG SAM.
PEO
P1 + P2
P2
P1
500 550 600 650 700 750
0
2000
4000
6000
8000
10000
12000
P
L

I
n
t
e
n
s
i
t
y
/
c
p
s

Wavelength/nm
T1 - QD
655
: 3-CGT GGA CTG AGG ACA-QD
655
-5
T2 - QD
565
: 3- ACA TGT AGT GTT GAT -QD
565
-5
Figure 9.23 SPFM images showing the hybridization following sequential
introduction of T1
0
and T2
0
.
308 Chapter 9
9.6 Conclusions
This chapter has demonstrated that the combination of surface plasmon eld
enhancement mechanisms with uorescence detection schemes allows for a
number of bioafnity studies that go far beyond the classical SPR concept. A
critical point that is sometimes raised concerns the fact that in this way the
attraction of SPR as being a label-free detection scheme is sacriced. This
statement needs a few comments. First, whenever the mere presence of the
analyte induces a change in the optical interfacial architecture at the transducer
surface which is sufcient to be monitored by the (slight change in the
dispersion relation of the) surface plasmon wave, there is, of course, no need
to attach a chromophore, thus risking an unwanted change in the interaction
between the afnity partners of interest. However, even in the case of label-free
detection by SPR, one of the partners has to be surface immobilized. This is a
constraint at least as severe as the attachment of a chromophore. Second,
applying uorescence detection schemes does not necessarily mean that the
analyte has to be dye-modied. The stretching of a chromophore-labeled probe
strand on the SPFS transducer surface upon hybridization of an unlabeled
target strand illustrated in Figure 9.10 is an example of this. Although not
discussed in this chapter explicitly, the SPFS equivalent of the widespread
ELISA assay with a catcher antibody immobilized on the sensor surface,
followed by the binding of the unlabeled analyte from solution and the nal
recording of the decoration (binding) of a second, chromophore-conjugated
antibody in this sandwich assay, is another prominent example of this concept.
Third, the major advantage of using uorescence detection is the significant
extension of the range of analyte concentrations that can be monitored. For
example, we demonstrate that binding of proteins can be quantied by SPFS
from solutions at sub-femtomolar concentrations, which is several orders of
magnitude better than can be seen by SPR. So what does one sacrice? One
only gains considerably in the LOD of an analyte. Finally, the detection of
uorescence can be seen as a second information channel that allows for the
simultaneous but independent observation of two analytes, e.g. a protein by
SPR via its optical mass and the other one, e.g. a (relatively) small oligonuc-
leotide via its uorescence signal. On top of this, color multiplexing in uores-
cence detection allows for multiple simultaneous recordings of analyte binding
to the sensor surface, a feature that does not exist in SPR.
9.7 Questions
1. If you look at Figure 9.3a, the minimum of the reectivity and the
maximum of the uorescence intensity measured in the angular scans do
not occur at the same angle of incidence. Why?
2. Why is the uorescence of a chromophore near a metal surface reduced
(quenched). What is a typical dyesurface distance for which this
happens?
3. Assume that the Langmuir adsorption behaviour is correct; what
type of measurements would give you the afnity constant for a binding
reaction between a surface-attached receptor and a ligand from
solution?
4. Imagine that you study the hybridization reactions of two different target
strands (T1 and T2, differing in their sequence by one nucleotide)
from solution to surface-attached catcher probe DNA strands, with
K
A
(T1)10
9
M
1
and K
A
(T2)10
7
M
1
. Which concentration of the
targets would you chose to maximize mismatch discrimination?
5. What is the analogy between long-range surface plasmons and coupled
pendulums?
Some of the results presented in this chapter were collected during the PhD
projects of Thorsten Liebermann, who was the rst in our group to document
surface plasmon eld-enhanced uorescence detection of hybridization reac-
tions. We are grateful to Eva-Kathrin Sinner and Danfeng Yao for their help
with PCR, to Peter E. Nielsen, Andrea Germini, Stefano Sforza, Roberto
Corradini and Rosangela Marchelli for their support with PNAs and to Jakub
Dostalek for helpful discussions concerning long-range surface plasmons.
Financial support came from various projects funded by the EU (QLK1-
2000-31658, DNA-Track) and by the Deutsche Forschungsgemeinschaft
(KN 224/13-1, 2).
References
1. T. Moses and Y.R. Shen, Phys. Rev. Lett., 1991, 67, 20032037.
2. J.D. Swalen, J. Phys. Chem., 1979, 83, 14381445.
3. E. Aust, S. Ito, M. Sawodny and W. Knoll, Trends Polym. Sci., 1994, 2,
313323.
4. E. Burstein, W.P. Chen, Y.J. Chen and A. Hartstein, J. Vac. Sci Technol.,
1972, 11, 10041019.
5. H. Raether, Surface Plasmons, Springer Tracts in Modern Physics, Springer,
Berlin, 1988.
6. W. Knoll, Annu. Rev. Phys. Chem., 1998, 49, 569638.
7. W. Knoll, M. Zizlsperger, T. Liebermann, S. Arnold, A. Badia, M. Liley,
D. Piscevic, F.-J. Schmitt and J. Spinke, Colloids Surf. A, 2000, 161,
115137.
8. A.W. Wark, H.J. Lee and R.M. Corn, Anal. Chem., 2005, 77, 3904.
9. J. Homola, Anal. Bioanal. Chem., 2003, 377, 528.
10. B. Liedberg, C. Nylander and I. Lundstro m, Sens. Actuators, 1983, 4,
299304.
310 Chapter 9
11. S. Lo fa s and B. Johnsson, J. Chem. Soc., Chem. Commun., 1990,
15261528.
12. C.E. Jordan, A.G. Frutos, A.J. Thiel and R.M. Corn, Anal. Chem., 1997,
69, 49394947.
13. J. Spinke, M. Liley, H.J. Guder, L. Angermaier and W. Knoll, Langmuir,
1993, 9, 18211825.
14. N. Huang, J. Vo ro s, S.M. De Pani, M. Textor and N.D. Spencer,
Langmuir, 2002, 18, 220.
15. L. Ha ussling, H. Ringsdorf, F.J. Schmitt and W. Knoll, Langmuir, 1991, 7,
18371840.
16. F. Yu, D. Yao and W. Knoll, Anal. Chem., 2003, 75, 26102617.
17. K.A. Peterlinz, R.M. Georgiadis, T.M. Herne and M. Tarlov, J. Am.
Chem. Soc., 1997, 119, 34013402.
18. J.P. Attridge, P.B. Daniels, J.K. Deakon, G.A. Robinson and G.P.
Davidson, Biosens. Bioelectron., 1991, 6, 201214.
19. T. Liebermann and W. Knoll, Colloids Surf. A, 2000, 171, 115130.
20. T. Liebermann, W. Knoll, P. Sluka and R. Herrmann, Colloids Surf. A,
2000, 169, 337350.
21. F. Yu, B. Persson, S. Lofas and W. Knoll, J. Am. Chem. Soc., 2004, 126,
89028903.
22. S. Ushioda and Y. Sasaki, Phys. Rev. B, 1983, 27, 14011404.
23. W. Knoll, H. Park, E. Sinner, D. Yao and F. Yu, Surf. Sci., 2004, 570,
3042.
24. P. Schuck, Rev. Biophys. Biomol. Struct., 1997, 26, 541566.
25. P.W. Atkins and J. de Paula, Physical Chemistry, 7th edn., Oxford
University Press, 2006.
Chem., 1997, 6, 127137.
27. T.A. Morton, D.G. Myszka and I.M. Chaiken, Anal. Biochem., 1995, 227,
176185.
28. C.R. MacKenzie, T. Hirama, S.J. Deng, D.R. Bundle, S.A. Narang and
N.M. Young, J. Biol. Chem., 1996, 271, 15271533.
29. T. Neumann, M.L. Johansson, D. Kambhampati and W. Knoll, Adv.
Funct. Mater., 2002, 12, 575586.
30. H. Park, A. Germini, S. Sforza, R. Corradini, R. Marchelli and W. Knoll,
Biointerphases, 2006, 1, 113122.
31. D. Yao, J. Kim, P. Nielsen, E. Sinner and W. Knoll, Biophys. J., 2005, 88,
27452751.
32. D. Yao, F. Yu, J. Kim, J. Scholz, P. Nielsen, E.K. Sinner and W. Knoll,
Nucleic Acids Res., 2004, 32, 177192.
33. E.K. Sinner, K. Kobayashi, T. Lehmann, T. Neumann, B. Prein, J. Ru he,
F. Yu and W. Knoll, in Biopolymers at Interfaces, M. Malmsten, Ed., 2nd
edn., Marcel Dekker, New York, 2003, pp. 583607.
34. F. Yu, B. Persson, S. Lofas and W. Knoll, Anal. Chem., 2004, 76,
67656770.
35. D. Sarid, Phys. Rev. Lett., 1981, 47, 19271930.
36. D. Sarid, R.T. Deck, A.E. Craig, R.K. Hickernell, R.S. Jameson and
J. Fasano, Appl. Opt., 1982, 21, 39933995.
37. G.G. Nenninger, P. Tobiesoeka, J. Homola and S. Yee, Sens. Actuators B,
2001, 74, 145151.
38. A. Kasry and W. Knoll, Appl. Phys. Lett., 2006, 89, 101106.
39. R. Robelek, L. Niu, E. Schmid and W. Knoll, Anal. Chem., 2004, 15,
61606165.
40. L. Niu and W. Knoll, Anal. Chem., 2007, 79, 26952707.
312 Chapter 9
CHAPTER 10
SPR Imaging for Clinical
Diagnostics
ELAIN FU,
a
TIMOTHY CHINOWSKY,
b
KJELL NELSON
a
AND PAUL YAGER
a
a
Department of Bioengineering, University of Washington, Seattle,
WA 98195, USA;
b
Verathon Corporation, Bothell, WA 98021, USA
10.1 Introduction
Surface plasmon resonance (SPR)-based detection for clinical applications
has attracted significant interest in recent years. Studies have demonstrated
the detection of clinically relevant analytes such as prostate-specic antigen [1],
C-reactive protein [2] and human ferritin [3] at physiologically relevant con-
centrations using both commercial and custom SPR-based systems. The com-
mercial market for SPR imagers is competitive and includes instruments from
multiple companies including Biacore, IBIS, GWC Technologies, Biosensing
Instruments, K-Mac and Lumera (for more companies and websites, see
Chapter 3). However, these instruments have generally been designed to target
the research market. There is currently no commercial SPR imaging-based
instrument on the market that qualies as a true point-of-care (POC) clinical
diagnostic system, i.e. a portable SPR imaging-based detection platform for the
rapid, simultaneous measurement of multiple analytes in human samples that
could be used in a clinical setting by minimally trained personnel.
SPR imaging is well suited for use in a system for detecting clinical analytes
for several reasons:
1. Detection is rapid. Because detection is based on changes in refractive
index (RI) near the surface, specic binding events are quantied as
they occur, without time-consuming rinses of unbound reagents; this
reduces time-to-result compared with other methods such as uorescence
detection.
313
2. Detection does not depend on labeling of the target substances. Labeling
of a substance may affect its binding kinetics and afnity, in addition to
increasing the complexity of the assay and the cost of reagents.
3. Immunoassays based on SPR imaging are capable of direct detection of
many clinically relevant analytes. In cases in which the sensitivity or RI
resolution (i.e. limit of detection) is not adequate, signal amplication
methods can be used to improve detection sensitivity or RI resolution.
4. SPR imaging only requires simple optics that can be miniaturized to a size
that is suitable for POC diagnostics in clinical applications.
5. SPR imaging can detect multiple types of binding events simultaneously
using spatially addressed capture arrays. The number of independent
measurements possible is limited by the lateral resolution of the instru-
ment and the density of the components in the capture array.
6. The non-specic binding to the SPR surface by interferents present in
complex samples can be mitigated in imaging systems by the ability to
have multiple reference surfaces that may be used to correct for back-
ground signals.
7. Recent advances in microuidics allow preconditioning of small-volume
samples upstream of the SPR surface to minimize the impact of non-
specic surface fouling. In addition, preconditioning of a complex sample
may significantly reduce the range of expected sample RI, thus simplifying
sample analysis.
In this chapter, we discuss the main requirements of a POC SPR imaging-
based clinical diagnostic system: (1) mechanical and optical simplicity for
miniaturization and low cost, (2) adequate imager performance for the detec-
tion of the analytes of interest, (3) robust operation despite temperature and
bulk RI variations and (4) formatting of bioassays that allow for multiplexed,
rapid, quantitative measurement of analytes from a complex sample.
10.2 Achieving Miniaturization and Low Cost
SPR-based detection is inherently simple, with the potential for dramatic
miniaturization and low cost. To date there have been multiple reports of
miniature non-imaging SPR detection systems, many focusing on the use of
ber optics [4,5]. Alternatively, Kawazumi et al. [6] have described a very
compact (16 9 6 cm
3
) instrument for the multi-channel detection of small
analytes using a laser diode line source to create two-dimensional angle scans of
the surface over a limited angular range of 51. The smallest reported device is
the Texas Instruments (TI) multi-channel Spreeta sensor (B$50 per sensor),
which uses a design based on angular interrogation without moving parts [7].
The Spreeta chip has been used as the basis for highly integrated multi-channel
sensor systems [812].
There has been substantially less progress in the development of miniature
SPR imaging systems. Imagers described in the literature are generally research
314 Chapter 10
instruments that are mechanically complex (e.g. include bulky rotation arms
[13]) or specialized, costly instrumentation [14,15]. In this section, we describe
application-driven SPR imager design principles that stress mechanical and
optical simplicity.
10.2.1 Tuning in SPR Imager Design
SPR imagers require adjustability in incident wavelength or angle in order to
achieve maximum sensitivity and dynamic range for the target range of sample
RI. Several tuning methods that minimize mechanical complexity and thus
increase compactness and ruggedness are described below.
A simple and compact imager design that uses wavelength tuning has been
reported by Fu and co-workers [16,17] The center imaging wavelength was
varied over 65 nm by tilting a low-cost interference lter with respect to the
collimated source beam. This wavelength range is more than adequate to detect
the small changes in RI produced in the binding of sub-monolayers of
biomolecules. An additional advantage of this design is the ability to extend
the range of operating wavelengths by replacement of the interference lter [13].
Compact SPR imager designs that operate by angle tuning have also been
reported [18,19]. Angle interrogation systems typically use bulky rotation
stages to rotate the source and detector about the prism. One option to avoid
this is to rely on the intrinsic eld of view of the optics. Imaging optics accept
rays emitted from an object point at a range of angles and focus them to a single
image point. Sufciently optimized optics will have a eld of view adequate to
intercept a range of angles large enough to permit adjustment-free operation
over a useful range of incident angles. Figure 10.1a shows a prototype portable
SPR imager developed for the rapid detection of small molecules in saliva that
is based on this design consideration [18]. In this device, the angle of incidence
is adjusted by mechanically translating a single LED across the optical axis of
its collimator. Mechanical motion in this device may be eliminated entirely
if, rather than translating the source, multiple light sources (e.g. an array of
LEDs), positioned at various locations perpendicular to the optical axis, are
switched on in turn [19].
10.2.2 Compact Design of Additional SPR Imager Optical
Elements
In addition to wavelength or angle tuning, all SPR imagers require polarization
control and sufcient optical path length to permit image formation. Strategies
for meeting these two requirements in a compact package are discussed below.
Many designs use the simple geometry shown in the top drawing of
Figure 10.1b and require mounting of the light source and detector at the
corresponding angle to the prism surface (e.g. for an equilateral prism, 601 from
the normal to the sensing surface). For compact packaging it may be advan-
tageous to fold the optical path. In the middle drawing of Figure 10.1b, two
315 SPR Imaging for Clinical Diagnostics
(a)
Microfluidic
card
Card mated
with optics and
external fluidics
Tablet PC
for instrument
control
(b)

(c)

Collimating lenses
LCD polarizer
Imaging lenses
Beam steering prisms
Detector
Light
source
Figure 10.1 (a) Portable SPR imager developed for the detection of small molecules in
saliva. Folding of the optical path is a key factor in the compact size.
Reproduced with permission from reference 18. (b) Schematic of the
optical path for various prism congurations. (c) Schematic of the
compact optical module that occupies approximately 25 10 5 cm
3
.
316 Chapter 10
additional equilateral prisms are added to direct light downwards, perpendic-
ular to the sensing surface (the angle of incidence for the reecting surfaces is
above the critical angle, so no aluminization is needed). This conguration
allows mounting of optical components along a plane parallel to the sensing
surface. In the bottom drawing of Figure 10.1b, two further 451 prisms are
added, creating a conguration in which the optical axes are parallel to the
sensing surface.
Polarization control, i.e. switching between transverse magnetic (TM)
polarization (to excite surface plasmons) and transverse electric (TE) polariza-
tion (to normalize the SPR images for non-uniformity of measured intensity), is
often implemented using a rotation stage, adding bulk and cost to the instru-
ment. An alternative is the use of a liquid crystal-based device that incorporates
a polarizer and a vertically-aligned nematic liquid crystal phase shifter. Appli-
cation of an appropriate AC voltage rotates the polarization 901 from TM to
TE. The necessary voltage source can be compactly built into the light source
controller.
Many of the design elements described in this section have been implemented
in a portable SPR imager (see Figure 10.1a). The mechanism of tuning is angle
interrogation via small range translation of a single near-infrared LED source
and the use of stationary wide-eld imaging optics. The polarization of the light
source is electronically switchable from TM to TE through the use of a liquid
crystal switchable polarizer. Folding of the optical path is a key factor in
producing the compact size of the instrument. The result is an optical module,
shown schematically in Figure 10.1c, that occupies approximately 25 10 5
cm
3
and resides within the imager shown in Figure 10.1a.
10.3 Optimizing Imager Performance
For any target application, the performance requirements for an SPR imager
can be broken down into two categories: RI resolution and lateral resolution
over the eld of view. Optimization of these performance parameters in SPR
imagers is discussed below.
10.3.1 Refractive Index Resolution
An SPR imager designed for clinical diagnostics requires adequate RI resolu-
tion for the detection of the lowest expected levels of the target analyte(s).
In SPR imaging, the signal-to-noise ratio (SNR), the ratio of the SPR signal
(the change in reectivity) to noise (the uncertainty in the reectivity measure-
ment), can be expressed as
SNR
I
p

dR
dn

Dn
R
p 10:1
where I is the illumination intensity, R is the reectivity, Dn is the change in RI
and the sensitivity,
dR
dn
, is the derivative of reectivity with respect to RI [19]. The
reectivity will change only slightly during measurement of a typical binding
event, so for measurement of the RI resolution, Dn
LOD
, eq. (10.1) can be
expressed as
Dn
LOD
/
1
I
p

dR
dn
10:2
Optimization of the RI resolution translates into maximizing the quantity
I
p

dR
dn
. The main parameters available to the user for maximizing
dR
dn
are the
interrogation wavelength and angle. Greater
dR
dn
occur at longer wavelengths
[20,21]. Choice of angle is slightly more involved because the incident angle
which maximizes
dR
dn
varies depending on the sample RI, but for a chosen
illumination wavelength such an angle can be found. Increasing I should be
accompanied by changes in the detector that enable an increased number of
photoelectrons to be accumulated without saturating the detector.
One way to increase I is to increase the range of imaging wavelengths and
incident angles (Dl and Dy) for a given measurement. However, adjustment of
either Dl or Dy will increase I but reduce
dR
dn
. Determination of the optimal Dl
and Dy is done by balancing these effects. The contour plot in Figure 10.2a
shows the dependence of 1
I
p
dR=dn on Dl and Dy, indicating that the
positive effects of the increase in I may outweigh the negative effects of a
decrease in
dR
dn
on the product
I
p

dR
dn
[19]. The largest
I
p

dR
dn
is predicted for
Dl on the order of 30 nm and Dy on the order of 31. Nelson et al. reported a
threefold increase in the SNR of their system (leading to improved RI resolu-
tion) through the use of a source lter with a 10 times wider bandpass (10 nm
compared with a 1 nm wide bandpass lter) [13].
Image averaging and normalization can also be used to improve the RI
resolution. An example of this is shown in Figure 10.2b [18]. The Raw curve
represents the root-mean-square RI noise values over eight sample regions for
each of four RI solutions. The Normalized curve represents root mean square
RI noise values that were calculated by rst normalizing the brightness of each
sample region to the corresponding region of the intensity reference sector of
the image and then averaging in the same manner. Because the reference sector is
not in resonance, any changes in its brightness are expected to be due to light
source variations, camera uctuations and other systematic effects. Image aver-
aging and normalization successfully reduce measured noise to a level that is
inversely proportional to the square root of the number of pixels averaged
[17,18,22]. In contrast, the noise level of the un-normalized data decreases at a
slower rate for high levels of averaging, indicating the presence of noise compo-
nents that are correlated across pixels and thus cannot be reduced with averaging.
Uniform SPR response is required across the instrument eld of view in
order to make quantitative comparisons between multiple sites. A comparison
of the normalized reectivity of multiple regions across the image provides a
useful measurement of the response uniformity. An example of this is shown in
Figure 10.2c [18]. During sensing operations, the incident angle is typically
adjusted such that the reected intensity is approximately one-third up the left
318 Chapter 10
(a)
-1 0 1
0.5
1
1.5
2
log10(dn/dR / I)
-
4
-
-
-
3
-
3
-
-
2
S
p
e
c
t
r
a
l

w
i
d
t
h
,

l
o
g
1
0

n
m
-
4
-
3
.
5
-
3
.
5
-
3
-
3
-
2
.
5
-
2
(b) 10
-4
10
-5
10
-6
10
0
10
1
10
2
10
3
10
4
Number of pixels averaged, N
R
I

m
e
a
s
u
r
e
m
e
n
t

n
o
i
s
e
,

10
-7
Normalized
Raw
Dark
-0.5 0.5
Angular width, log10 degrees
(c)
0.8
0.6
0.4
R
e
f
l
e
c
t
i
v
i
t
y
62 63 64 65
Incident angle, degrees
Figure 10.2 (a) Predicted dependence of 1
I
p
dR=dn on light source properties.
The calculation assumes a source at 670 nm and a three-layer system
composed of BK-7 glass, gold and water. The largest 1
I
p
dR=dn is
predicted for Dl on the order of 30 nm and Dy on the order of 31.
Reproduced with permission from reference 19. (b) Noise as a function of
pixels averaged with and without normalization for a portable SPR
imager (described in Section 10.2). Error bars indicate the standard
deviation across eight sample regions and four solutions of different RI.
Shown for comparison are the equivalent RI noise values measured with
no illumination and a line showing the expected inverse square root
relationship between the number of pixels averaged and RI noise.
Reproduced with permission from reference 18. (c) Normalized reec-
tivity (reected intensity of a gold surface in water divided by the
reected intensity measured using a bare glass surface) of 16 regions
across the image for a portable SPR imager (described in Section 10.2).
Reproduced with permission from reference 18. (d) Improved focus over
the eld of view by tilting the image plane according to the Scheimpflug
condition. (e) Image of approximately 100 mm features taken with a
portable SPR imager (described in Section 10.2). The test pattern is
composed of PDMS and water. Reproduced with permission from
reference 18.
side of the curve (vertical line in Figure 10.2c). Binding to the sensor surface will
cause a local increase in RI, a shift of the SPR curve to higher angles and an
increase in reected intensity. Changes in intensity will be approximately linear
with respect to RI, provided that the operating point remains on the portion of
the SPR curve which is close to linear [23].
10.3.2 Lateral Resolution Over the Field of View
The SPR imager should have adequate lateral resolution to distinguish the surface
structures of interest for the target application. In the case of an instrument
designed to make measurements on multiple analytes simultaneously, the lateral
resolution will determine the minimum area of the independent capture regions
and also place an upper limit on the number of possible parallel measurements.
Specications for the target application will inuence the desired quality of focus
over the eld of view of the instrument and the choice of operating wavelength.
One factor that affects the lateral resolution of an SPR imager is the quality
of focus over the instrument eld of view. A notable feature of SPR imager
optics is that the object (i.e. the SPR sensing surface) is tilted relative to the
optical axis of the imaging optics. If standard imaging optics are used, much of
the object is either closer or further away than is required for best focus. SPR
imagers in the literature often accept this limited depth of eld. A potentially
better performance alternative is to use a tilted image plane as dictated by the
Scheimpflug condition [24], i.e. if the object and image are tilted such that the
(d)
Normal incidence on detector
74

incidence on detector
The
Scheimpflug
condition
(e)
Figure 10.2 (Continued)
320 Chapter 10
object plane, the image plane and the lens plane meet in a single line, the entire
image will be in focus. The use of a tilted image plane becomes most important
when a large depth of eld is needed and the collimation of the input light has
been relaxed in order to increase light throughput. For an SPR imager, the
object tilt is often fairly large and meeting the Scheimpflug condition may
present some difculties (e.g. typical lens mounts for commercially packaged
imagers will block light that has a sufciently oblique angle of incidence and
some image detectors have properties that make them unsuitable for use at
oblique angles of incidence). Figure 10.2d demonstrates the quality of focus
that can be gained by satisfying the Scheimpflug condition.
As stated above, the change in the SPR reectivity per unit of RI is greater at
longer wavelengths. However, the use of longer wavelengths also produces
surface plasmons with longer propagation lengths (e.g. approximately 25 mm
for a gold layer at 830 nm [13] compared with 0.1 mm for a gold layer at 488 nm
[25]). This surface plasmon propagation length presents an inherent limit to the
lateral resolution of the imager, but only in the dimension parallel to surface
plasmon propagation. de Bruijn et al. [26] have demonstrated the utility of
rotating an object such that the critical dimension is perpendicular to the
direction of surface plasmon propagation. The lateral resolution for the critical
dimension is then determined only by the optics of the system. A balance
between the lateral resolution and the RI resolution should be established given
the requirements for the target application. For example, a portable SPR
imager operating at 880 nm on gold (described in Section 10.2) can easily
resolve 100 mm features (Figure 10.2e) with an RI noise level of approximately
7 10
6
RI units (RIU) [18]. Given a 10 7.4 mm eld of view, the instrument
has the potential to monitor simultaneously over 1000 independent samples.
Additionally, averaging and normalization (described in Section 10.3) over a
greater region can decrease the RI noise level at the trade-o of fewer total
independent samples. An RI noise level of 2 10
6
RIU would limit the
number of independent samples to approximately 400.
10.4 Robust Operation
A requirement of any clinical diagnostic system is robust operation. The system
should perform to specications on complex human samples (such as blood,
urine and saliva) under varying conditions. For SPR-based instrumentation,
temperature uctuations and variations in the sample bulk RI are of particular
concern. In this section, we focus on these two critical issues.
10.4.1 Effects of Temperature Fluctuations
For an SPR-based system, temperature effects may occur in a variety of ways.
First, surface plasmons, which sense the effective RI of the sample in a volume
adjacent to the metal/sample interface, are affected by changes in the temper-
ature-dependent RI of the sample. The solvent component of a typical sample
will experience the largest RI temperature dependence. For example, a dier-
ence in temperature of approximately 0.11C in a sample of water corresponds
to a 10
5
RIU change, near to or greater than the (limit of detection) LOD of
many reported SPR-based instruments [20,27]. Second, temperature variations
in the metal layer may also significantly affect the generation of surface
plasmons. At higher temperatures, surface plasmons may be significantly
damped due to increased phononelectron scattering in the metal [28]. The
result is a broadening of the SPR curve and a decrease in
dR
dn
with increasing
temperature [28]. Third, variations in temperature may also affect the operation
of critical components of the system such as the light source output, detector
response and sensor geometry. Finally, changes in ambient temperature may
affect the rate and/or efciency of the binding event(s) of interest. For example,
Zeder-Lutz et al. [29] found a 10-fold change in the ratio of the association rate
to disassociation rate for an antibodyantigen binding event over the temper-
ature range 5301C.
Measurable effects of temperature variation for a given sample on SPR-
based systems are exhibited as changes in the shape (e.g. width) of the SPR
curve and the location of the minimum. For imaging systems, these changes in
the SPR curve appear as a change in the measured reectance at xed wave-
length and angle. The result is a temperature-induced signal that is indistin-
guishable from the signal due to a targeted binding event. Thus, minimizing
temperature-induced SPR effects (or compensating for them) is critical to
achieving robust instrument operation.
10.4.2 Strategies to Alleviate the Effects of Temperature
Fluctuations
There are two general strategies to alleviate the effects of temperature variation
on SPR-based instrumentation: temperature stabilization and compensation.
The most basic form of temperature stabilization is the judicious choice of
materials in the fabrication of imager components. For example, use of con-
ducting materials in imager elements that contact the sample cells and upstream
sample tubing can greatly aid in sample temperature stabilization [8,9].
Active heating or cooling of the system to achieve a predetermined optimal
setpoint temperature is another temperature stabilization method. Thermo-
electric devices, based on the PeltierSeebeck effect,
1
are often used in
SPR-based instrumentation since the devices are compact, lightweight and
inexpensive. They allow for both heating and cooling and are relatively easy to
implement (e.g. they require no coolant). Optimal performance of these devices
requires good thermal coupling between the heating/cooling elements and the
sample and proper insulation of the instrument housing [8], so care should be
taken in the choice of materials used in instrument fabrication. Temperature
control to within 0.011C or 10
6
RIU in a laboratory setting has been reported
with a Spreeta-based system using a thermoelectric device [11].
1
The conversion of electric voltage to heat ux and vice versa.
322 Chapter 10
Temperature compensation can be achieved through the use of a reference
channel to correct the sample signal for changes in temperature. In this method,
two SPR signals are acquired simultaneously. One measurement is performed
on an active channel that contains a sensor surface functionalized for the target
application. A second measurement is performed on a reference channel that is
resistant to the binding events of interest, but is otherwise similar to the active
channel. The compensated signal is calculated by subtracting the reference
channel signal from the active channel signal. Using this method, Kawazumi
et al. reduced the effects of long-term temperature drift in their system to nearly
the same level as that measured for short times [6]. However, the use of
temperature compensation over a wide-range of temperatures has limits. An
investigation of the efcacy of reference channel compensation over the
temperature range 4401C for a Spreeta-based system showed significant
degradation in reference channel compensation performance at the tempera-
ture extremes (calibration was performed at intermediate temperatures) [8].
In the case of a POC instrument that requires operation over a wide range of
ambient temperatures or has a small mass, a combination of these methods
may be needed. Naimushin et al. [8] reported the use of both active temperature
control and reference channel compensation in their portable SPR instrument.
They used a thermoelectric module to control the temperature of their Spreeta-
based system over widely varying ambient temperatures, from 4 to 371C, to
within 0.71C or a noise level of approximately 7 10
5
RIU. Additionally, use
of a reference channel reduced this 0.71C change in temperature to below the
noise limit of their instrument, 10
7
RIU.
10.4.3 Bulk RI Compensation
In addition to temperature change, a change in composition of the sample is
another significant contributor to sample bulk RI variation. In the case of a
clinical diagnostic system, natural variation in the bulk RI of complex human
samples may be of particular concern. Changes in sample bulk RI may also
occur intentionally in the course of an assay, such as during the binding and
rinse stages.
Reference channel compensation can effectively correct for variations in bulk
RI [3033] due to inherent differences in sample composition and also varia-
tions in bulk RI due to temperature. Several reports using Biacore instruments
with active temperature control demonstrated compensation levels of 9899%
for bulk RI changes on the order of 10
3
RIU [3436]. Reference channel
compensation has the potential to correct for additional interfering inuences
such as non-specic binding [8] (see the discussion below) and drift due to
instability in the functionalization of surface layers. Accurate compensation for
these effects relies on the assumptions that (1) the reference channel can be
treated to have identical non-specic binding and drift characteristics as the
sample channel and (2) the reference channel is resistant to the binding event of
interest. Implementation of a system that even approximately satises these
assumptions may be challenging.
A second effective bulk RI compensation method that is much less dependent
on surface effects is internal reection refractometry (IRR). In this method, a
region of the substrate without the metal layer is used to perform measurements
of reectivity near the critical angle to determine the bulk RI. These IRR
measurements can be set up to be acquired simultaneously with the SPR
measurements. The compensated signal is then calculated by subtracting the
IRR signal from the SPR signal. The accurate measurement of the bulk RI of
solutions of varying temperature and composition has been demonstrated using
this method [37]. Chinowsky and co-workers implemented this method in a
Spreeta-based instrument [38,39] and achieved compensation of bulk RI
variations of 5 10
3
RIU to within (24) 10
5
RIU. Effective compensation
using this method requires calibration over the RI range of interest [39].
As mentioned above, the issue of variability in sample RI is of particular
concern for an SPR-based clinical diagnostic system that processes raw human
samples. For example, the RI of blood varies by 1.4 10
5
RIU per 10 mg dl
1
of glucose [40] which, can range over 100 mg dl
1
in uncontrolled diabetics.
Urine samples can also vary greatly in RI, with a reported range of 6.4 10
3
RIU [41]. Based on the clinical application, bulk RI compensation after sample
preconditioning (described below) and adjustment of imaging conditions to
those optimal for a given sample may be needed.
10.5 SPR Imaging Assays
SPR imaging assays can enable multiplexed, rapid, quantitative analysis of
complex human samples for a multitude of clinical applications. The target
application will determine the required sensitivity, LOD and speed of the assay,
and also provide constraints on the sample volume, the stringency of sample
preconditioning and the range of analyte concentrations. The optimal imaging
assay for an application will contain design features that balance performance
requirements with complexity and cost of implementation. In this section, we
discuss key design considerations, sample handling and the complexity and cost
of assay implementation.
10.5.1 Microuidic Immunoassay Design for Small Molecule
Analytes
A straightforward example of a microuidic SPR imaging assay [18,42] (see
also Chapter 7 for SPR immunoassays) is shown in Figure 10.3. In this assay,
sample uid containing an unknown concentration of analyte is mixed with a
known amount of buer containing antibody and then passed to the detection
region. The binding of free or singly-bound antibody (antibody molecules with
only one of their two binding sites occupied) to the surface-immobilized analyte
within the detection region results in a change in the RI that is quantied by
SPR imaging. The design of a multiplexed assay, the choice of assay quanti-
cation method and the choice of assay parameters such as ow rate, reagent
324 Chapter 10
Figure 10.3 Parallel immunoassays for the anti-epileptic drug phenytoin conducted
on a portable SPR imager (described in Section 10.2). (a) Difference
image shows the outcome after 3 min of anti-phenytoin binding to the
BSAphenytoin-modied surface. Solutions contained 75 nM anti-
phenytoin IgG in phosphate buer premixed with 50, 25 or 0 nM
phenytoin (left to right streams, respectively). Flow is from bottom to
top. The dark region at the bottom of the difference image is a non-
fouling poly(ethylene glycol) (PEG)-terminated thiol surface. The chan-
nel dimensions (60 mm deep, 3.8 mm wide) and ow rates (750 nl s
1
total,
250 nl s
1
per ow stream) used in this assay ensured laminar ow with
negligible diffusion between the streams. The scale bars represent 1 mm
in each dimension. Squares demarcate the pixels used to calculate the
average intensity values plotted in (b). Contrast has been adjusted for
display. (b) The average intensities demonstrate that both the rate of
binding and the total coverage of the antibody anti-correlate with the
amount of competitor in the sample. Solid lines are the average intensi-
ties in the PEG regions and show transient refractive index changes
during the initial (wash in) phase of the experiment. After 1.5 min, the
transients have stabilized and the remaining intensity changes can be
solely attributed to antibody binding.
concentration and sample volume are discussed below in the context of this
example.
Simultaneous analysis of multiple samples may be accomplished using
controlled parallel ows. In the multiplexed SPR imaging assay shown in
Figure 10.3, three samples of phenytoin (a small-molecule therapeutic drug) in
buer (premixed with antibody) are measured simultaneously. Laminar ow,
with negligible diffusion between the streams (i.e. low Reynolds number and
high Peclet number), ensures independent measurements of each sample. This
single-channel design provides for the self-balancing of ow rates, eliminating
the need for precise control of uid resistances in different channels, as ow rate
variations will affect the rate of antibody binding. The SPR signals demonstrate
that both the rate of binding and the total coverage of antibody anti-correlate
with the concentration of phenytoin. The detection of 25 nM phenytoin for an
application in which the low end of the therapeutic concentration in saliva is
200 nM [43] provides a comfortable LOD margin for potentially necessary
manipulations (including dilution) of the sample. Note the rapid assay time of
less than 3 min. Further multiplexing can be achieved by patterning multiple
capture zones within a given uid stream for the simultaneous detection of
multiple analytes [44].
On-board positive and negative controls may be integrated similarly. For
example, parallel ow streams alongside the sample stream in a single channel
may be used for a run-time two-point calibration. Proper functioning of the
calibrators requires that the antibody concentrations and degree of non-specic
binding or interference be identical in all streams an important consideration
when using solutions with known composition as a reference for complex
samples.
The selection of appropriate assay parameters is largely based on the method
used for quantifying the assay; typically either an analysis of the end-point
signal or binding rate. Both quantication methods require detailed knowledge
of the concentration gradients at the leading and trailing edges of the sample
due to Taylor dispersion [45,46]. Measurement of the end-point signal is less
demanding on image processing capabilities but requires precisely timed deliv-
ery of the sample volume with respect to the end-point measurement. Binding
rate analysis may reduce the time to result but is only valid at relatively low
fractional surface coverage, complicating comparisons between high- and low-
rate assay outcomes. This method of quantication also requires well-behaved
sample introduction to the binding region, with regard to leading edge Taylor
dispersion, such that an appropriate, well-dened time period can be used to
evaluate the binding rate. Comparisons between streams will naturally require
that dispersion effects be similar.
The choice of ow rate, reagent concentration and sample volume is critical
for optimal assay operation. Selection of ow rate is based on a balance
between time to result and reagent consumption, and also the performance
characteristics of the method used to drive the uids (e.g. pressure-driven ow).
Higher ow rates reduce the time to result by increasing the ux of molecules of
interest (e.g. antibodies) with access to the sensing surface. Conversely, lower
326 Chapter 10
ow rates increase the time to accumulate adequate signal, but reduce the total
volume of sample and other reagents consumed [4749]. Low ow rates also
increase the time available for diffusion transverse to convective ow, increas-
ing potential cross-talk between parallel streams in the format illustrated in
Figure 10.3. The choice of reagent concentration, e.g. the concentration of
antibody mixed into the sample, will mainly affect the dynamic range of the
assay and should be selected appropriately for the target range of analyte
concentrations. Achieving adequate sensitivity for low analyte concentrations
will require similarly low antibody concentrations. However, this will increase
the time needed to obtain sufcient surface coverage for adequate signal and
affect the optimal choice of other assay parameters.
10.5.2 Assay Compatibility with Complex Samples
Complex human samples contain substances that interfere with the detection of
surface binding events in SPR-based assays. This interference may occur either
through non-specic adsorption of substances in the sample to the detection
surface (increasing the background signal) or through interactions between
substances in the bulk sample with the analyte such that transport or binding of
the analyte to the detection surface is impeded. These effects will create either
an inated or deated signal and are problematic for the implementation of a
quantitative SPR-based assay. Two strategies for dealing with these effects will
be discussed in the next section.
The non-specic adsorption signal from complex human samples in SPR-
based assays can be significant. As discussed previously, reference channel
compensation may be used to subtract out the signal due to non-specic
adsorption if an appropriate reference channel can be created. For example,
Navratilova and Skladal described a competition assay in which the sample
channel contained urine samples with human serum albumin (HSA) and
anti-HSA, while the reference channel contained urine samples with HSA only,
i.e. without anti-HSA [50]. Implementation of a reference channel is not as
straightforward in all assay designs. For SPR measurements of analyte binding
directly to the detection surface (relevant for large analytes), the analogous
reference channel would require selective removal of the analyte of interest
from the sample an impractical requirement. An alternative is to prepare a
reference region in the assay channel with a surface chemistry that is similar
to that in the sample channel but does not bind the analyte of interest (as
described above). The practical utility of this method depends heavily on the
specic surface chemistries available and the ease with which they can be
incorporated into the device.
For some samples, compensation for non-specic adsorption alone is not
adequate and processing of the sample to remove interfering substances before
arrival at the detection region is needed. Yager et al. [51] described a two-stage
preconditioning process for human saliva samples. The rst stage consisted of a
mechanical lter and the second stage consisted of an H-lter [52,53]. They
reported removal of 73% of the glycoprotein content (predominantly mucins)
Figure 10.4 (a) Example of an integrated diagnostic card created using a combination
of polymeric laminate, PDMS and gold-coated glass. Reproduced with
permission from reference 51. (b) Schematic of the on-card and o-card
processes. Not all of the challenges of integration have been accom-
plished in the current design. Most notably, the valves and pumps have
yet to be integrated onto the card. Reproduced with permission from
reference 51.
328 Chapter 10
and retention of 92% of the small molecule analyte using the mechanical lter.
The combination of both stages removed 98% of the glycoprotein content
while retaining 27% of the small molecule analyte. These preconditioning
processes permitted quantitative operation of an immunoassay for the detec-
tion of a small molecule analyte in saliva. Preconditioning of a complex sample
in this manner has the added benet of narrowing the target range of sample
RI, thus reducing the burden on tuning of the optical system to the optimal
imaging conditions for a given sample. This is especially useful for samples as
variable across populations and over time as saliva.
10.5.3 Assay Implementation
An attractive approach to POC diagnostics for high performance with low per-
test cost is to combine a permanent SPR imager (and uid workstation) with a
disposable card incorporating all sample-contacting elements. This approach
also eliminates the need for cleaning the instrument between samples to prevent
carry-over. In the example shown in Figure 10.4, many of the steps necessary
for analysis of a saliva sample (subsequent to mechanical ltration) have been
combined on to a disposable diagnostic card [51]. In this card design, the
sample and waste are retained, avoiding sample contact with the electrical and
optical components of the permanent instrument.
The selection of materials for use in a disposable should balance application
requirements against the inherent cost of the material plus the processing costs
[51]. Fortunately a 50 nm layer of gold over 1 cm
2
does not contribute
significantly to the cost of the disposable, provided that an inexpensive fabri-
cation method can be found. Lamination of laser-cut polymer sheets is
extremely cost-effective for rapid prototyping [54] but cannot be used to form
sub-micrometer features. Another material, polydimethylsiloxane (PDMS), is
excellent for the production of sub-micrometer features, but is permeable to
many small and hydrophobic molecules and is not readily or economically
formed in high-throughput production. The example card of Figure 10.4 was
created using a combination of PDMS to form a herringbone mixer [55] with
small features, gold-coated glass for the SPR imaging surface and polymeric
laminate for the bulk of the disposable. Practical devices used in POC settings
will likely require a variety of materials and manufacturing methods.
10.6 Conclusion
SPR imaging is a detection technique that is well suited for use in a clinical
detection platform. In this chapter, we have discussed the key requirements of
an SPR imaging-based POC diagnostic system and described progress towards
implementation of these requirements into an integrated system. Assuming that
some of the remaining challenges to miniaturization of instruments can be
overcome, we expect to see fully integrated POC SPR imaging-based diagnostic
systems introduced into the marketplace in the next 5 years.
10.7 Questions
1. What are the main advantages of SPR imaging as a detection method for
POC diagnostics?
2. What are the main challenges for SPR-based detection with complex
samples?
3. How can temperature uctuations affect SPR-based measurements?
4. Describe temperature compensation using a reference channel.
5. Describe the competing effects of an increase in Dl or Dy on RI
resolution.
References
1. L. Huang, G. Reekmans, D. Saerens, J. Freidt, F. Frederix, L. Francis,
S. Muyldermans, A. Campitelli and C. Van Hoof, Biosens. Bioelectron.,
2005, 21, 483.
2. W. Hu, H. Hsu, A. Chiou, K. Tseng, H. Lin, G. Chang and S. Chen,
Biosens. Bioelectron., 2006, 21, 1631.
3. S. Chou, W. Hsu, J. Hwang and C. Chen, Biosens. Bioelectron., 2004, 19,
999.
4. R. Slavik, J. Homola and E. Brynda, Biosens. Bioelectron., 2002, 17, 591.
5. G. Garabedian, C. Gonzalez, J. Richards, A. Knoesen, R. Spenser,
S. Collins and R. L. Smith, Sens. Actuators A, 1994, 43, 202.
6. H. Kawazumi, K. Vengatajalabathy Gobi, K. Ogino, H. Maeda and
N. Miura, Sens. Actuators B, 2005, 108, 791.
7. J. Melendez, R. Carr, D. Bartholomew, K. Kukanskis, J. Elkind, S. Yee,
C. Furlong and R. Woodbury, Sens. Actuators B, 1996, 35, 212.
8. A. Naimushin, S. Soelberg, D. Bartholomew, J. Elkind and C. Furlong,
Sens. Actuators B, 2003, 96, 253.
9. G. Neuert, S. Kufer, M. Benoit and H. Gaub, Rev. Sci. Instrum., 2005, 76,
054303.
10. A. Sesay and D. Cullen, Environ. Monit. Assess., 2001, 70, 83.
11. S. Soelberg, T. Chinowsky, G. Geiss, C. Spinelli, R. Stevens, S. Near,
P. Kauman, S. Yee and C. Furlong, Environ. Technol., 2005, 32, 669.
12. M. Suzuki, F. Ozawa, W. Sugimoto and S. Aso, Anal. Bioanal. Chem.,
2002, 372, 301.
13. B. Nelson, A. Frutos, J. Brockman and R. Corn, Anal. Chem., 1999, 71,
3928.
14. S. Wang, S. Boussaad, S. Wong and N. Tao, Anal. Chem., 2000, 72, 4003.
15. S. Boussaad, J. Pean and N. Tao, Anal. Chem., 2000, 72, 222.
16. E. Fu, J. Foley and P. Yager, Rev. Sci. Instrum., 2003, 74, 3182.
17. E. Fu, T. Chinowsky, J. Foley, J. Weinstein and P. Yager, Rev. Sci.
Instrum., 2004, 75, 2300.
18. T. Chinowsky, K. Johnston, T. Edwards, K. Nelson, E. Fu and P. Yager,
Biosens. Bioelectron., 2007, 22, 22082215.
330 Chapter 10
19. T. Chinowsky, A. Mactutis, E. Fu and P. Yager, Proc. SPIE, 2004, 5261,
173182.
20. J. Homola, S. Yee and G. Gauglitz, Sens. Actuators B, 1999, 54, 3.
21. K. Johansen, H. Arwin, I. Lundstrom and B. Liedberg, Rev. Sci. Instrum.,
2000, 71, 3530.
22. T. Chinowsky, J. Quinn, D. Batholomew, R. Kaiser and J. Elkind, Sens.
Actuators B, 2003, 91, 266.
23. J. Shumaker-Parry and C. Campbell, Anal. Chem., 2004, 76, 907.
24. W. Smith, Modern Optical Engineering, McGraw-Hill, New York, 2000.
25. C. Berger, R. Kooyman and J. Greve, Rev. Sci. Instrum., 1994, 65, 2829.
26. H. Debruijn, R. Kooyman and J. Greve, Appl. Opt., 1993, 32, 2426.
27. J. Homola, H. Vaisocherova, J. Dostalek and M. Piliarik, Methods, 2005,
37, 26.
28. H. Chiang, Y. Wang, P. Leung and W. Tse, Opt. Commun., 2001, 188, 283.
29. G. Zeder-Lutz, E. Zuber, J. Witz and M. Van Regenmortel, Anal. Biochem.,
1997, 246, 123.
30. R. Karlsson and R. Stahlberg, Anal. Biochem., 1995, 228, 274.
31. G. Nenninger, J. Clendenning, C. Furlong and S. Yee, Sens. Actuators B,
1998, 51, 38.
32. W. Peng, S. Banerji, Y. Kim and K. Booksh, Opt. Lett., 2005, 30, 2988.
33. H. Roos, K. Magnusson, R. Karlsson, in Proceedings of the 8th Inter-
national Conference on Solid-State Sensors and Actuators and Eurosensors
IX, Stockholm, Sweden, volume 2, p. 517520. General Chairman: Ingemar
Lundstran Foundation for Sensor and Actuator Technology Stockholm,
Sweden, 1995.
34. R. Ober and E. Ward, Anal. Biochem., 1999, 271, 70.
35. A. Frostell-Karlsson, A. Remaeus, H. Roos, K. Andersson, P. Borg,
M. Hamalainen and R. Karlsson, J. Med. Chem., 2000, 43, 1986.
36. R. Karlsson, M. Kullman-Magnusson, M. Hamalainen, A. Remaeus,
K. Andersson, P. Borg, E. Gyzander and J. Deinum, Anal. Biochem.,
2000, 278, 1.
37. J. Grassi and R. Georgiadis, Anal. Chem., 1999, 71, 4392.
38. T. Chinowsky, A. Strong, D. Bartholomew, S. Jorgensen-Soelberg,
T. Notides, C. Furlong and S. Yee, Proc. SPIE, 1999, 3857, 104113.
39. T. Chinowsky and S. Yee, Proc. SPIE, 2002, 4578, 442453.
40. K. Zirk and H. Poetzschke, Med. Eng. Phys., 2004, 26, 473.
41. A. Wolf and V. Pillay, Am. J. Med., 1969, 46, 837.
42. K. Nelson, N. Geisler, K. Tandon and P. Yager, presented at Micro Total
Analysis Systems, Tokyo, 2006.
43. H. Liu and M. R. Delgado, Clin. Pharmacokinet., 1999, 36, 453.
44. K. Nelson, J. Foley, A. Mashadi-Hossein, P. Yager, presented at Micro
Total Analysis Systems, Boston, MA, 2005.
45. A. Bancaud, G. Wagner, K. D. Dorfman and J. Viovy, Anal. Chem., 2005,
77, 833.
46. J. Ruzicka and E.H. Hansen, Flow Injection Analysis, Wiley, New York,
1988.
47. A. Lionello, J. Josserand, H. Jensen and H.H. Girault, Lab Chip, 2005, 5,
1096.
48. A. Lionello, J. Josserand, H. Jensen and H.H. Girault, Lab Chip, 2005, 5,
254.
49. M. Zimmermann, E. Delamarche, M. Wolf and P. Hunziker, Biomed.
Microdevices, 2005, 7, 99.
50. I. Navratilova and P. Skladal, Supramol. Chem., 2003, 15, 109.
51. P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M. Tam and B. Weigl,
Nature, 2006, 442, 412.
52. J. Brody and P. Yager, Sens. Actuators A, 1997, 58, 13.
53. J. Brody, P. Yager, R. Goldstein and R. Austin, Biophys. J., 1996, 71, 3430.
54. P. Yager, D. Bell, J.P. Brody, D. Qin, C. Cabrera, A. Kamholz, B.H.
Weigl, presented at Micro Total Analysis Systems, Ban, Canada, 1998.
55. A. Stroock, S. Dertinger, A. Ajdari, I. Mezic, H. Stone and G. Whitesides,
Science, 2002, 295, 647.
332 Chapter 10
CHAPTER 11
The Benets and Scope of
Surface Plasmon Resonance-
based Biosensors in Food
Analysis
ALAN McWHIRTER AND LENNART WAHLSTRO
M
General Electric Healthcare, Biacore AB, Rapsgatan 7, SE-754 50 Uppsala,
Sweden
11.1 Introduction
Although surface plasmon resonance (SPR)-based biosensors are most
commonly associated with protein interaction analysis in the life sciences and
drug discovery, the technology is also well established in the food industry. SPR
biosensors are used commercially to screen for drug residues and to quantify
vitamins in food matrices. The main advantages of SPR-based detection over
alternative analytical techniques such as microbiological assays (MBA) include
ease of use, simpler and faster sample preparation and reduced assay time from
days to minutes in some cases. SPR biosensors offer the clearest advantages in
speed over alternative techniques that rely on biological readouts such as
inhibition of microbial growth for detecting antibiotics.
In addition to a review of the practical aspects of SPR-based food analysis,
quality control and safety aspects will be discussed in this chapter. These issues
are often driven by regulations set by national authorities in addition to
widespread consumer demand that the information on product packaging is
a true indicator of the content. Several case studies will be covered such as the
detection of growth-promoting hormone residues in meat, multi-analyte
screening of veterinary drug residues in body uids, antibiotic screening in
honey and the detection of adulterants in dairy products. Together, these
application areas illustrate the tremendous scope of SPR-based biosensors in
333
food safety and quality control, elds with ever-increasing performance
demands on throughput and data quality.
11.1.1 Why Analyze Food?
The packaging of most food products provides information about its contents,
including additives such as preservatives or vitamins. In addition, a label may
assure the consumer that the product has been tested and shown to be free of
harmful or undesired substances, that the raw ingredients are from approved
suppliers or that the crops or animals from which the product is derived were
reared according to ethically accepted standards and that the production
process was performed in compliance with safety and hygiene standards.
Many of these criteria are strictly regulated and demand rigorous adherence
to internationally agreed standards. Failure to comply with these standards can
have serious economic and even legal implications for manufacturers and it is
clearly in their interest to be as condent as possible that their products are
compliant or, if not, to have the opportunity to rectify the problem as early in
the production process as possible.
Whatever the aim of the analysis, such as conrming the presence of
benecial vitamin supplements in baby food or detecting potentially harmful
residues of growth-promoting hormones in meat, the technology used must
provide that information rapidly, reliably and economically. Same-day results
allow closer control of product quality and minimize the risk of costly product
recalls. This chapter seeks to show how label-free analysis of food products and
raw materials using the Biacore Q SPR-based biosensor satises these criteria.
11.1.2 Food Analysis Steps and SPR Assay Formats
Minimal sample preparation is one of the most striking advantages of
SPR-based assays (Figure 11.1). For example, using the Biacore Qex Kit
Sulfonamides simply requires homogenization and centrifugation of meat
samples or dilution of honey or milk samples, whereas lengthy and potentially
hazardous organic solvent or solid-phase extraction is required by alternative
technologies such as ELISA, TLC or HPLC. The savings in cost and time for
sample extraction, together with direct analysis, are a highly attractive option
for busy routine screening laboratories.
Usually, only minimal sample preparation such as homogenization, extraction,
centrifugation and ltration is required and accurate measurements may be
made even in complex matrices. The risk of matrix effects interference due
to components other than the analyte itself in the sample or sample extract
giving false positives is minimized, since the contact time of a few minutes
between sample and specic binding protein within the continuous ow system
of the instrument favors high-afnity interactions over low-afnity matrix
interactions. The longer contact times between proteins and sample in a typical
334 Chapter 11
Figure 11.1 1. Prepare the sample often a simple dilution or homogenization/
centrifugation step. 2. Prepare the assay using a Qex Kit. A growing
range of Qex Kits is available for the analysis of many commercially
important food additives and contaminants, including vitamins and
veterinary drug residues. 3. Perform an assay on Biacore Q. Wizards
guide the user through all steps from assay setup to analysis. 4. Analyze
the results by interpolation results on a calibration curve. Biacore Q
software automatically evaluates results and generates a report.
335 SPR-based Biosensors in Food Analysis
ELISA (approximately 1 h) significantly increase matrix effects, causing loss of
sensitivity and an increased risk of false positives.
Typically, an active derivative of the analyte of interest is immobilized on the
gold sensor surface of a chip. A xed concentration of specic binding protein
(often an antibody) is then mixed with the sample and injected over the sensor
surface. The SPR response is inversely proportional to the concentration of the
sample (Figure 11.2), which may be estimated by interpolation of the response
on a calibration curve. The sensor surface needs to be regenerated after each
sample injection, ready for the next sample to be tested (Figure 11.3).
Although the sensitivity, specicity and working range of assays are deter-
mined to a large extent by the afnity of the interacting partners, high
immobilization levels are also desirable in concentration assays. This ensures
that binding is highly mass transport limited and less dependent on afnity.
Although SPR is an optical detection technique, measurements may be made
on colored, opaque or even turbid solutions. When a sample or calibrant
solution containing an interacting partner is injected over the prepared sensor
surface, the response is recorded in real time as a sensorgram and is directly
related to the concentration of the interacting partner in solution. In the
analysis of food samples, the binding partner of interest may be a relatively
small molecule such as a vitamin or an antibiotic. Since it is difcult to interpret
accurately the small changes in SPR response elicited by interactions involving
Figure 11.2 Inhibition assays are used for measuring the concentration of low
molecular weight target molecules in solution. Note that the ligand
which is immobilized on the sensor surface is similar to the target
molecule (analyte). An excess of the detecting molecule is added to the
samples prior to injection. The remaining amount of free detecting
molecule binds to the sensor and is measured. Inhibition assays are
based on solution competition, in which the high molecular weight
detecting molecules that remain in solution are free to bind to analyte or
analyte analogue on the sensor surface. The SPR response is inversely
related to the concentration of the target molecule in solution. More
assay formats are described in Chapter 7 (see Figure 7.2).
336 Chapter 11
small molecules, the assays performed on Biacore Q using Qex Kits are based
on inhibition or surface competition to alleviate the problem.
Interacting partners in solution with a mass below 5 kDa can be difcult to
measure directly with high precision in SPR-based assays because they elicit
only small changes in mass at the sensor surface. Inhibition assays are useful in
such cases, in which a puried preparation of the interacting partner in the
sample is immobilized on the sensor surface (Figure 11.2). A high molecular
weight detecting molecule that can bind to the interacting partner under
analysis is then added in excess to the sample. At equilibrium, a proportion
of detecting molecules remains in solution and is free to bind to the sensor
surface. In inhibition assays, the binding response is inversely related to the
analyte concentration.
Difculties may be encountered in designing an inhibition assay if, for
example, the interacting partner in the sample cannot easily be immobilized
on the sensor surface. An alternative format in such cases is surface competition
assay, in which the analyte is covalently conjugated to a larger molecule and
then added to the free analyte solution prior to injection. This high mole-
cular weight complex then competes with free analyte for binding sites on
the interacting partner immobilized on the sensor surface (Figure 11.4).
The competing complex should be significantly larger than the analyte so that
the SPR response is attributable almost entirely to the high molecular weight
analogue.
The shelf-life of enzyme conjugates used in ELISA-based methods is typi-
cally 612 months and the production of consistent batches is difcult and time
consuming. In contrast, prepared sensor surfaces are highly stable and the
immobilization procedure is reproducible. The shelf-life of the surface, in fact,
Figure 11.3 The entire course of the interaction may be followed in a real-time trace
(sensorgram), allowing the user to visualize association, nal binding
level and dissociation. Analysis is fully automated and the sensor surface
may be regenerated after each sample injection, ready for the next sample
to be tested.
is limited primarily by the chemical stability of the immobilized binding partner
itself.
Regardless of assay format, regeneration of the sensor surface with a
regeneration solution is normally required between sample cycles to remove
interacting partners bound to the surface. The ideal regeneration solution will
sufciently remove all traces of bound material, but at the same time it must be
sufciently mild not to perturb irreversibly the biological activity of the
immobilized partner.
11.2 Biacore Q and Qex Kits the Workhorse
of Food Analysis
Biacore Q (Figure 11.5) is a fully automated SPR-based instrument from
Biacore AB (Uppsala, Sweden), designed for routine analysis. Samples and
reagents are injected automatically to ensure reproducibility and reliability of
results by elimination of the variability inherent in manual procedures. Biacore
Q Control Software is wizard-based for immobilization and analysis and is
optimized to allow exibility in assay design in addition to routine testing.
Results from individual tests are clearly and automatically presented within a
few minutes of the start of analysis. Methods under revision or routinely used
may be protected by a password while changes are registered to comply with
GLP/GMP for routine measurements.
To make the analysis process as simple and consistent as possible, an
extensive range of Qex Kits have been created for use with the Biacore Q
system. An overview of the currently available kits, designed for quantifying or
Figure 11.4 Surface competition assays are indirect assays, in which a high molecular
weight complex competes with free analyte for binding sites on the
ligand. The response is inversely related to analyte concentration (see
also Figure 7.2).
338 Chapter 11
screening specic, commercially important food additives and contaminants, is
given in Table 11.1. Analyte levels in the low picomolar range can typically be
measured, but the dynamic range must be empirically determined and is
dependent on the experimental conditions and the molecular weights of the
interaction partners.
Although Biacore Q includes a surface preparation unit that allows easy
immobilization of small molecules on the sensor surface, several Qex Kits are
delivered with pre-immobilized sensor chips, further reducing manual prepa-
ration. It is easy to alternate between different Qex Kits on the same system.
As indicated in Table 11.1, Qex Kits, containing stability-tested reagents,
are available for the quantication of several vitamins and for screening
veterinary drug residues such as growth promoters and antibiotics in animal
products. In addition to the Qex Kit range, the procedures leading to AOAC
certication of Qex Kit Folic Acid (a B group vitamin especially recom-
mended during pregnancy) are described here.
11.2.1 Qex Kits for Screening Veterinary Drug Residues
Sulfonamides are a family of broad-spectrum synthetic bacteriostatic antibiotics,
which can be used against most Gram-positive and many Gram-negative
organisms and protozoa. The resistance of animal pathogens to sulfonamides
is widespread as a result of more than 50 years of therapeutic use, but despite
this, they are still widely used in combination with other medication. Strict
withdrawal periods must be observed before slaughter to prevent the entry of
residues into the human food chain. The assays used to ensure compliance with
EU and US regulations must be sufciently sensitive to detect the maximum
permitted residue level of 100 ppb in edible tissue. Immunoassays are frequently
Figure 11.5 Biacore Q, specially adapted to the demands of routine screening and
detection in the food industry. The system can be used for different
assays while wizard-based software controls all processes from immobi-
lization to data interpretation.
T
a
b
l
e
1
1
.
1
Q
e
x
K
i
t
s
f
o
r
t
h
e
d
e
t
e
c
t
i
o
n
o
f
d
r
u
g
r
e
s
i
d
u
e
s
a
n
d
v
i
t
a
m
i
n
s
:
s
p
e
c
i
c
a
t
i
o
n
s
.
K
i
t
(
t
a
r
g
e
t
c
o
m
p
o
n
e
n
t
s
)
I
n
t
e
n
d
e
d
m
a
t
r
i
x
L
i
m
i
t
o
f
d
e
t
e
c
t
i
o
n
T
h
r
o
u
g
h
p
u
t
S
u
l
f
o
n
a
m
i
d
e
s
(
2
0
s
u
l
f
o
n
a
m
i
d
e
s
)
P
o
r
c
i
n
e
m
u
s
c
l
e
a
n
d
k
i
d
n
e
y
,
h
o
n
e
y
a
n
d
a
v
i
a
n
m
u
s
c
l
e
,
s
e
r
u
m
a
n
d
p
l
a
s
m
a
2
.
7
p
p
b
(
e
q
u
i
v
a
l
e
n
t
t
o
1
6
.
9
p
p
b
i
n
t
i
s
s
u
e
)
4
0
s
a
m
p
l
e
s
i
n
5
h
S
u
l
f
a
d
i
a
z
i
n
e
B
i
l
e
a
n
d
t
i
s
s
u
e
s
o
f
s
l
a
u
g
h
t
e
r
e
d
p
i
g
s
0
.
6
n
g
m
l
1
4
0
s
a
m
p
l
e
s
i
n
8
h
S
u
l
f
a
m
e
t
h
a
z
i
n
e
B
i
l
e
a
n
d
t
i
s
s
u
e
s
o
f
s
l
a
u
g
h
t
e
r
e
d
p
i
g
s
0
.
3
n
g
m
l
1
4
0
s
a
m
p
l
e
s
i
n
8
h
S
t
r
e
p
t
o
m
y
c
i
n
B
o
v
i
n
e
m
i
l
k
,
p
o
r
c
i
n
e
m
u
s
c
l
e
a
n
d
k
i
d
n
e
y
,
h
o
n
e
y
2
8
m
g
l
1
(
m
i
l
k
)
,
5
0
m
g
k
g
1
(
k
i
d
n
e
y
)
,
6
9
m
g
k
g
1
(
m
u
s
c
l
e
)
,
1
5
m
g
l
1
(
h
o
n
e
y
)
4
0
s
a
m
p
l
e
s
i
n
8
h
(
m
i
l
k
)
,
2
0
s
a
m
p
l
e
s
i
n
8
h
(
m
u
s
c
l
e
/
k
i
d
n
e
y
)
,
4
0
s
a
m
p
l
e
s
i
n
8
.
5
h
(
h
o
n
e
y
)
C
h
l
o
r
a
m
p
h
e
n
i
c
o
l
C
h
i
c
k
e
n
m
u
s
c
l
e
,
h
o
n
e
y
,
b
o
v
i
n
e
m
i
l
k
,
s
h
e
l
l
s
h
a
n
d
o
t
h
e
r
m
a
t
r
i
c
e
s
0
.
0
2
p
p
b
(
c
h
i
c
k
e
n
)
,
0
.
0
7
p
p
b
(
h
o
n
e
y
)
,
0
.
0
2
5
p
p
b
(
m
i
l
k
)
,
0
.
0
7
3
p
p
b
(
s
h
e
l
l
s
h
)
8
0
s
a
m
p
l
e
s
i
n
2
4
h
(
c
h
i
c
k
e
n
)
,
2
0
s
a
m
p
l
e
s
i
n
1
0
h
(
h
o
n
e
y
/
s
h
e
l
l
s
h
)
,
4
0
s
a
m
p
l
e
s
i
n
1
0
h
(
m
i
l
k
)
b
-
A
g
o
n
i
s
t
s
B
o
v
i
n
e
,
p
o
r
c
i
n
e
a
n
d
s
h
e
e
p
l
i
v
e
r
0
.
0
2
n
g
g
1
(
m
a
b
u
t
e
r
o
l
)
1
.
4
6
n
g
g
1
(
p
i
r
b
u
t
e
r
o
l
)
1
6
s
a
m
p
l
e
s
i
n
3
6
h
C
l
e
n
b
u
t
e
r
o
l
B
o
v
i
n
e
u
r
i
n
e
0
.
2
7
n
g
m
l
1
4
0
s
a
m
p
l
e
s
i
n
9
h
R
a
c
t
o
p
a
m
i
n
e
U
r
i
n
e
a
n
d
l
i
v
e
r
0
.
1
1
n
g
m
l
1
(
u
r
i
n
e
)
,
0
.
1
7
n
g
g
1
(
l
i
v
e
r
)
2
0
s
a
m
p
l
e
s
i
n
1
5
h
(
u
r
i
n
e
)
,
2
0
s
a
m
p
l
e
s
i
n
1
7
h
(
l
i
v
e
r
)
T
y
l
o
s
i
n
(
A
a
n
d
B
)
H
o
n
e
y
5
.
7
m
g
k
g
1
4
0
s
a
m
p
l
e
s
i
n
1
2
h
B
i
o
t
i
n
C
e
r
e
a
l
s
,
m
i
l
k
p
o
w
d
e
r
,
m
i
l
k
-
o
r
s
o
y
a
-
b
a
s
e
d
i
n
f
a
n
t
f
o
r
m
u
l
a
,
f
o
r
t
i
e
d
b
e
v
e
r
a
g
e
s
,
v
i
t
a
m
i
n
p
r
e
m
i
x
e
s
a
n
d
d
i
e
t
a
r
y
s
u
p
p
l
e
m
e
n
t
s
1
n
g
m
l
1
2
0
s
a
m
p
l
e
s
i
n
6
h
F
o
l
i
c
a
c
i
d
V
a
r
i
o
u
s
h
e
a
l
t
h
a
n
d
n
u
t
r
i
t
i
o
n
a
l
f
o
o
d
s
t
u
s
,
c
e
r
e
a
l
p
r
o
d
u
c
t
s
1
n
g
m
l
1
2
0
s
a
m
p
l
e
s
i
n
6
h
V
i
t
a
m
i
n
B
2
C
e
r
e
a
l
s
,
m
i
l
k
p
o
w
d
e
r
,
m
i
l
k
-
o
r
s
o
y
a
-
b
a
s
e
d
i
n
f
a
n
t
f
o
r
m
u
l
a
,
f
o
r
t
i
e
d
b
e
v
e
r
a
g
e
s
,
v
i
t
a
m
i
n
p
r
e
m
i
x
e
s
a
n
d
d
i
e
t
a
r
y
s
u
p
p
l
e
m
e
n
t
s
1
7
.
1
n
g
m
l
1
2
0
s
a
m
p
l
e
s
i
n
6
h
V
i
t
a
m
i
n
B
1
2
C
e
r
e
a
l
s
,
m
i
l
k
p
o
w
d
e
r
,
m
i
l
k
-
o
r
s
o
y
a
-
b
a
s
e
d
i
n
f
a
n
t
f
o
r
m
u
l
a
,
f
o
r
t
i
e
d
b
e
v
e
r
a
g
e
s
,
v
i
t
a
m
i
n
p
r
e
m
i
x
e
s
a
n
d
d
i
e
t
a
r
y
s
u
p
p
l
e
m
e
n
t
s
0
.
0
6
n
g
m
l
1
2
0
s
a
m
p
l
e
s
i
n
1
2
h
P
a
n
t
o
t
h
e
n
i
c
a
c
i
d
(
v
i
t
a
m
i
n
B
5
)
C
e
r
e
a
l
s
,
m
i
l
k
p
o
w
d
e
r
,
m
i
l
k
-
o
r
s
o
y
a
-
b
a
s
e
d
i
n
f
a
n
t
f
o
r
m
u
l
a
,
f
o
r
t
i
e
d
b
e
v
e
r
a
g
e
s
,
v
i
t
a
m
i
n
p
r
e
m
i
x
e
s
a
n
d
d
i
e
t
a
r
y
s
u
p
p
l
e
m
e
n
t
s
4
.
4
n
g
m
l
1
2
0
s
a
m
p
l
e
s
i
n
1
2
h
340 Chapter 11
used as screening tests but they tend to detect only one or two compounds from
the sulfonamide family and often give a positive result for the acetylated
metabolites. In contrast, assays based on Biacore Q used together with Qex
Kit Sulfonamides are capable of detecting at least 20 compounds within the
sulfonamide family and have the added advantage of not recognizing the
acetylated metabolites.
The sulfonamide-based antimicrobial compounds sulfadiazine and sulfa-
methazine are commonly incorporated into porcine feedstus for therapeutic
and prophylactic reasons. As tissue extraction is a time-consuming process, a
common strategy for the detection of residues of these antibiotics in slaugh-
tered animals is the random screening of body uids (typically bile, urine or
serum) by ELISA as an indicator of their presence in edible tissue. An
indication of violative tissue concentrations of sulfamethazine or sulfadiazine
from the bile screen then requires conrmation by HPLC analysis of the
suspected tissue. Qex Kit Sulfamethazine and Qex Kit Sulfadiazine, how-
ever, can be used to quantify these specic sulfonamide residues in both bile
and tissues of slaughtered pigs.
Biacore Q in combination with Qex Kit Sulfamethazine and Qex Kit
Sulfadiazine has been tested on-site at the kill line in a slaughterhouse to
determine how the screens performed in the cold and humid conditions of a
modern pig processing factory [1]. No adverse effects on any mechanical or
electrical components were encountered either during the testing period or after
the instrument was returned to the laboratory.
Bile samples were rst screened as a predictor of sulfonamides in the
corresponding animal tissue and positive samples were conrmed using HPLC.
Samples were compared with values from a calibration series containing 5%
guaranteed sulfonamide-free pig bile. It is important to construct calibration
curves in the presence of bile, as this medium appears to be responsible for an
appreciable matrix effect. False positive prediction rates were calculated and
compared with a standard enzyme immunoassay (EIA). For sulfamethazine,
false-positive rates were 0.14% for the Biacore Q assay and 1.54% for EIA and
for sulfadiazine the rates were 0.34% and 1.44%, respectively. No false-
negative results were obtained using the Biacore method. EIA analysis, in
contrast, delivered false-negative rates of 0.14 and 0.24% for sulfamethazine
and sulfadiazine, respectively. Despite the complexity of bile as a matrix, direct
analysis is possible using the Biacore Q assay without clean-up or extraction.
The results are available within 34 min, permitting analysis in real time in the
slaughterhouse and, importantly, before the meat can enter the human food
chain.
Qex Kit Sulfamethazine and Qex Kit Sulfadiazine can be adapted to
muscle extracts using a simple and low-cost mechanical extraction procedure
involving homogenization and centrifugation. It is therefore possible to use
these assays for both on-site screening and conrmation, thereby avoiding the
need for HPLC. It has been demonstrated, however, that a small number of
false positives were found using Biacore Q in a screening method [2]. Analysis
of samples containing N
4
-acetyl metabolites always gives higher results than
HPLC, due to the cross-reactivity of the antibodies, whereas HPLC only
detects parent sulfonamides. This highlights the importance of using real
incurred samples in evaluating assays as opposed to the simpler and more
widely used technique of spiking blank samples. The validation data reported
in this work were gathered from pigs that had been fed with sulfonamide-
supplemented feed and then withdrawn for a specied time before slaughter
and analysis.
Streptomycin, in combination with penicillin, is the most commonly used
antibiotic for the treatment of mastitis. It is also used for the treatment of
bacterial honeybee diseases, such as European foulbrood, and to control re
blight, a devastating bacterial disease affecting fruit trees during blossom.
Public health concerns persist, however, based on unlicensed use of the drugs
and on lack of compliance with the withdrawal periods. Residues of antibiotics
present a potential hazard to the consumer in terms of toxicity, allergic reaction
and the development of bacterial resistance. Regulatory authorities have
established permitted residue limits for streptomycin and dihydrostreptomycin
in milk, porcine kidney and muscle. None have been set for honey. The
permitted levels are based on extensive toxicological, pharmacological and
microbiological data from clinical trials and specify a concentration of a drug
residue that is considered safe (200 mg l
1
in milk, 1 mg kg
1
in kidney and
0.5 mg kg
1
in muscle).
Biacore Q and Qex Kit Streptomycin may be used to determine the
concentration of residues of streptomycin and dihydrostreptomycin in a wide
range of foodstus including whole milk, honey, porcine kidney and muscle.
Chloramphenicol is a broad-spectrum antibiotic with excellent pharmacoki-
netic properties. In humans, however, its use is often associated with the
adverse development of aplastic anemia, a rare but serious blood disorder.
For this reason, it is been banned from use in food-producing animals,
including honeybees, in the EU, USA, Canada and many other countries.
Chloramphenicol, however, is still widely available in developing countries
and is in common use in animal production. Affected products include poultry,
shrimps and honey. As a prohibited substance, zero tolerance applies. Methods
of detection, therefore, require very low detection limits and with the intro-
duction of increased testing, high sample throughput is also an important
factor. Conventional methods of detection include MBA, immunoassays,
chromatography and mass spectrometry. Some of these methods lack the
necessary sensitivity whereas others require long sample preparation. A
provisional limit of detection of 0.3 ppb has been proposed as a level that
any applied method should be able to achieve, but many food companies have
introduced their own limits, forcing the development of ultra-sensitive assays
able to detect o0.1 ppb.
Qex Kit Chloramphenicol can be used to screen for residues of chlor-
amphenicol and chloramphenicol glucuronide in poultry muscle, honey, whole
milk [3] and shellsh.
b-Agonists are used in human and veterinary medicine as bronchodilators
and also as agents to induce relaxation of the uterus. At high doses
342 Chapter 11
(approximately 10 times the therapeutic dose), they are effective growth pro-
moters in farm animals, increasing protein deposition and decreasing the fat
mass. Although the economic benets of this practice can be substantial, toxic
residues of the drugs can remain in the meat. Several cases of acute food
poisoning from the ingestion of contaminated meat containing the b-agonist
clenbuterol have been reported. Consequently, b-agonists are banned from use
in livestock production in many countries. Biacore Q and Qex Kit b-Agonists
can be used to screen for a wide range of b-agonist residues in liver samples
from cattle, sheep and pigs. The assay has the advantage of using minimal
organic solvent.
In the EU, although not in the USA, all b-agonists, including ractopamine,
are prohibited from use in livestock production. For monitoring purposes,
however, the US Food and Drug Administration (FDA) has established specic
tolerance levels for residues of ractopamine hydrochloride in edible swine
tissues of 0.05 ppm in muscle and 0.15 ppm in liver. In the EU, there is zero
tolerance to the presence of ractopamine in food of animal origin. Immunoas-
says have been developed as screening tests for ractopamine. HPLC/mass
spectrometry is used for conrmatory analysis, analyses which require signifi-
cant sample preparation. In comparison, Qex Kit Ractopamine exploits the
specicity of a high-afnity ractopamine-binding protein, providing a rapid
screening method where results are obtained in hours rather than days.
Some countries permit the use of certain antibiotics to treat diseased bee
colonies, raising the possibility that antibiotic residues will contaminate the
honey. Tylosin is a macrolide antibiotic that inhibits peptide growth in suscep-
tible microorganisms. Its use in apiculture is permitted in the USA, but banned
in the EU. Tylosin A is the predominant form but, at the low pH associated with
honey, tylosin A is converted to tylosin B (desmycosin), which is also biologically
active. The detection of both compounds in honey is essential to avoid the risk of
underestimation. Traditional methods of analyzing tylosin include screening
methods such as ELISA and conrmatory methods such as LC/MS/MS.
11.2.2 Qex Kits for Quantifying Vitamin Content
Fortied foods with vitamins are subject to regulations from government
bodies and consumer organizations to justify the labeling claims. Quality
control procedures for the conrmation of vitamin levels are often rate limiting
in the throughput of an entire processing facility. Assays that combine rapid
analysis time with accurate results are therefore in demand. To reduce the
overall assay time and increase throughput, in addition to simplifying assay
preparation and reducing errors inherent in manual processes, Qex Kits
containing pre-immobilized sensor surfaces are available for quantifying the
vitamin content of biotin, folic acid, vitamin B
2
, vitamin B
12
and pantothenic
acid. Many of these kits are certied as Performance Tested Methods by the
AOAC. Consistency between the quality-controlled chip surfaces ensures that
data can be reliably compared, even in multi-center studies.
Biotin plays an important role in the metabolism of both humans and
animals. As such, it is commonly added as a supplement to various health
and nutritional foodstus such as cereals. Traditional methods of analyzing
biotin include MBA, which require lengthy sample preparation and can take up
to 2 days to obtain results. In SPR assays, food samples are prepared using
simple extraction procedures, minimizing the pretreatment time. This indirect
SPR inhibition assay exploits the high afnity of a biotin antibody to measure
the biotin content in samples. Independent AOAC International (Association
of Analytical Communities)-approved studies have demonstrated that analysis
with Biacore Q and Qex Kit Biotin correlates well with the ratied MBAs.
Collaborative studies performed on a broad range of sample matrices have
shown that the Biacore method also correlates well with other standard MBAs
across a range of concentration values [4].
Although MBA has long been the method of choice for the analysis of folic
acid, a widespread food supplement and an especially important dietary
requirement during pregnancy, it is now possible to quantify folic acid content
in hours rather than days, using SPR. Samples are simply mixed with a xed
amount of folic acid antibodies and injected into the system.
Vitamin B
2
(riboavin) is routinely added to foods as a nutrient during
processing and as a colorant in some cases. Traditional methods for the
quantication of vitamin B
2
are MBA, HPLC and uorimetry. Qex Kit
Vitamin B
2
PI assay (supplied with pre-immobilized chips) is designed as an
indirect inhibition assay. Independent studies at the Nestle Research Centre,
Lausanne, Switzerland, have shown an excellent correlation between Qex Kit
Vitamin B
2
PI and a validated HPLC method across a range of matrices,
detecting both fortied and non-fortied levels [5].
Vitamin B
12
is routinely added as a supplement to a variety of processed foods
such as infant formulas, breakfast cereals and nutritional products. Reliable
quantication can be complicated by the low concentrations (nanogram quan-
tities) normally present in typical samples and the presence of interfering
compounds in sample matrices. Traditional methods of analyzing vitamin B
12
include MBA, which results in a typical sample turnover of 23 days.
Pantothenic acid (vitamin B
5
) has multiple functions and is essential for
normal growth and development in humans. It is routinely added to many
types of fortied foods, including dairy and non-dairy infant formulas, break-
fast cereals and pet food. Any analytical method must therefore be sufciently
robust to perform well in a wide diversity of matrices. The most widely
used method to determine pantothenic acid levels is an MBA employing
Lactobacillus plantarum. Qex Kit Pantothenic Acid is an inhibition assay
designed for use with Biacore Q.
11.2.3 AOAC Certication of Qex Kits
AOAC International is the foremost independent body providing internation-
ally recognized standards of testing, with over 120 years of experience in method
344 Chapter 11
testing and validation and the expertise of more than 2700 members worldwide.
Certication of an analysis requires documented evidence of the quality of
several assay parameters and provides assurance that an independent third party
has tested the assay and found that the product fullls all performance claims.
Qex Kit Folic Acid, for example, received AOAC approval after rigorous
performance trials carried out in independent laboratories [4]. A summary of the
procedures leading to certication and the results is presented below.
In the certication procedures, various assay parameters were determined:
specicity, accuracy, repeatability, reproducibility, precision, recovery, limit of
detection (LOD) and limit of quantication (LOQ). To ascertain the specicity
of the assay, cross-reactivity of the monoclonal antibody to folic acid provided
in the kits was tested with a target analyte, 5-methyltetrahydrofolic acid
(5MTHF) and a related vitamer, tetrahydrofolic acid (THF). The cross-
reactivities, calculated from inhibition curves, were 100% for 5MTHF and
4% for THF. To determine the accuracy of the assay, four accredited labora-
tories measured folic acid concentration in ve types of sample (milk powder,
milk-based infant formula, soya-based infant formula, cereal and premix) using
a microbiological technique and Qex Kit Folic Acid. The results from the two
assays did not differ significantly.
Repeatability and reproducibility were tested by 10 operators in four labora-
tories in a total of 22 samples of cereals, milk powder, premix, milk-based infant
formula and soya-based infant formula. Repeatability (the variation in the
results within a laboratory, obtained by the same operator) was between 0.6 and
3.6% and reproducibility (the variation in results from different laboratories and
operators) was between 5.0 and 9.9%. The intra- and inter-kit precisions of folic
acid concentrations were measured in soya-based infant formula. The results
showed not only that variation between kit batches was insignificant but also
that variation was low even when calibration solutions were prepared on
different occasions.
Three operators analyzed milk powder, milk-based infant formula, soya-based
infant formula and cereal samples on different occasions and recorded recovery
eciencies of between 87 and 95%. The LOD for folic acid was below 1 ng ml
1
and the LOQ was between 2.0 and 70ng ml
1
. Analysis of a sample spiked with
folic acid at 2.0 ng ml
1
gave a concentration within 4% of the expected
concentration (CV 2.7%, n 4) and analysis of a 70ngml
1
sample gave a
concentration within 4.5% of the expected concentration (CV 3.3%, n 4).
11.3 Examples of Applications for SPR-based
Biosensors in Food Analysis
11.3.1 Quantifying Antibiotics in Honey
Although antibiotics or their residues are generally not permitted in honey,
some may be used in the control of bee diseases, provided that they are
employed at the right time, using the right method and correct dosage.
Early in 2002, all honey imports from China to the EU were temporarily
stopped after the discovery in some samples of unacceptably high levels
of the antibiotic chloramphenicol. Chloramphenicol has been banned from
use in apiculture in the EU, North America and many other countries.
A Minimum Required Performance Limit of 0.3 mg kg
1
has been established
for chloramphenicol assays but many food companies have introduced their
own limits, forcing the development of ultrasensitive assays able to detect levels
below 0.1 mg kg
1
.
Conventional methods of chloramphenicol detection include MBA, immuno-
assays, chromatography and mass spectrometry. Sensitivity is not the only
important factor in determining the suitability of a method; in addition, sample
preparation must be quick and easy with a rapid sample turnover to cope with
the ever-pressing testing schedule. The Qex Kit portfolio (Table 11.1) includes
tests for the presence of several antibiotics used in the honey industry. Although
optimized for honey, the assays may also be applicable to other bee products
such as royal jelly in addition to other food matrices such as muscle and kidney.
Streptomycin residues are not permitted in European apiculture, but the
antibiotic is permitted in the USA for the treatment of American foulbrood, a
serious bacterial disease of honeybees. Withdrawal of its use must take place at
least 4 weeks before the main honey ow. HPLC with post-column derivatization
and uorescence detection is one possible analytical method for screening but
requires elaborate sample preparation. Enzyme immunoassays are also available
although these tests have a tendency to generate large amounts of false-positive
results due to cross-reactivity. LC/MS/MS with a sample preparation that
involves an SPE step and concentration before being applied to LC is used for
conrmation. Qex Kit Streptomycin is a rapid and simple screening alternative
for streptomycin, requiring little more than pH adjustment of the sample.
Sulfathiazole is the only sulfonamide permitted for the treatment of
European foulbrood in the USA, but no sulfonamides are permitted in the
EU. Again, LC with different detection modes may be used for conrmation.
Sample preparation for LC/MS/MS and LC/uorescence-based detection can
be tedious, involving hydrolysis (to counter the effects of the formation of
glucose adducts, giving erroneously low results), pH adjustment, liquidliquid
extraction, evaporation, SPE and concentration before injection. Further, the
majority of biological assays for sulfonamides in honey tend to have very
limited cross-reactivity within the sulfonamide family, do not reach the
required sensitivity for honey analysis or tend to cross-react with N-acetyl
metabolites. Qex Kit Sulfonamides is a rapid method for the detection of more
than 20 sulfonamides in honey. There is good sample recovery (approximately
95%) from both real and spiked samples and analysis is rapid, with a sample
injection time of 90 s.
11.3.2 Screening for Veterinary Drug Residues
In animal production, the outbreak of disease or sub-optimal growth of
livestock can lead to considerable economic losses. Veterinary drugs, therefore,
346 Chapter 11
are often administered under prescription to food-producing animals for the
prevention, control and treatment of disease. Product licenses for all veterinary
drugs state a withdrawal time that must be observed by producers before
treated animals can be sent for slaughter. In recent years, a number of highly
publicized instances drew attention to violations of the stated withdrawal
periods, unlicensed use or deliberate abuse of the drugs, all leading to a fall
in consumer condence.
Although the use of hormone preparations such as the b-agonist clenbuterol
to enhance animal growth is forbidden in the EU, isolated cases of malpractice
have come to the attention of regulatory authorities. Johansson and Hellena s
together with the Swedish National Food Administration (NFA) have devel-
oped rapid SPR-based assays for detecting clenbuterol in urine and hair
samples [6,7]. As a complement to controls in slaughterhouses, controls are
also made during animal husbandry so that violations can be detected as early
as possible. Although convenient to sample, trace quantities of clenbuterol in
urine can only be detected up to 1 week after exposure to the drug. On the other
hand, residues can be detected in hair samples several months after exposure.
The test for clenbuterol and related compounds is based on an inhibition assay
in which clenbuterol is rst immobilized on a sensor surface over which specic
antibodies are injected after incubation with test material.
In order to cope with the demand for high throughput testing, industry is
increasingly calling for tests that can screen for families of veterinary drugs
rather than testing for each individual member of these groups. Qex Kits for
screening families of veterinary drug residues in foodstus have been developed
to address this demand for generic testing. Qex Kit Sulfonamides, for exam-
ple, contains the reagents necessary to detect at least 19 members of the
sulfonamide family of drugs in porcine muscle, including sulfamethazine,
sulfadiazine, sulfathiazole and sulfaquinoxaline, all in one analysis cycle. A
major advantage of this assay is that there is no cross-reactivity with inactive
N-acetyl metabolites, thus reducing the probability of false-positive results.
Samples giving values above the maximum residue level (MRL) can then be
further analyzed by conrmatory methods where the specic sulfonamide
present can be identied and quantied.
11.3.3 Milk Testing
Penicillins and cephalosporins, members of the b-lactam family of antibiotics,
are the most frequently used group of antibiotics for the treatment of bacterial
infections, e.g. mastitis (inammation of the udder) in dairy cows. Conse-
quently, they are also the most frequently occurring type of drug residues in
milk. The most commonly used methods for the detection of b-lactams in milk
are based on inhibition of microbial growth. These tests are inexpensive, but
they are very time consuming, requiring days to complete an analysis. In recent
decades there has been an increase in the number of rapid receptor- and
antibody-based tests, for example, the enzyme-based Penzym test from UCB
Bioproducts and the Charm II test from Charm Sciences, based on the use of
whole bacterial cells in combination with labeled tracer. Simpler receptor-
binding assays have also been developed, such as the SNAP test from IDEXX
Laboratories, the b-STAR test from UCB Bioproducts and the Charm Safe
Level test from Charm Sciences. Rapid immunoassays are also available, such
as the Parallux test from IDEXX Laboratories. These tests offer high sensitivity
and specicity, but they are difcult to automate and are therefore limited in
their capacity for increased sample throughput.
An assay for the detection of b-lactam residues in milk has been developed in
Biacore Q. Most Biacore assays for the detection of drug residues in various
foods use antibodies as the detection molecule but in work by Gustavsson et al.,
a broad spectrum b-lactam receptor protein was used instead [8]. The advan-
tage of using such a receptor protein instead of antibodies is that a generic assay
can be designed, i.e. the whole group of b-lactams can be detected. This type of
assay makes it possible to detect the active form of the b-lactam ring structure,
an important consideration, as the legislated residue limits only cover the active
form. Three assays were developed, all based on the same receptor protein, a
carboxypeptidase. The b-lactam ring structure is easily hydrolyzed, resulting in
inactivation of the substance. This, together with the very strong interaction
between b-lactams and the receptor protein, made it necessary to use an assay
different in design from most previous assays in which antibodies are mixed
with the sample and injected over the sensor surface containing immobilized
antibiotics. The formats of the b-lactam assays based on the enzymatic activity
were similar, but differed in some important aspects. The b-lactam receptor
protein carboxypeptidase hydrolyzes a tripeptide into a dipeptide, but this
reaction is inhibited in the presence of b-lactams. Milk sample is mixed with the
receptor protein and the tripeptide and the enzymatic reaction is allowed to
proceed. Antibodies directed against the dipeptide (two-peptide assay) or
tripeptide (three-peptide assay) are then added and the sample is injected over
the sensor surface to which the respective di- or tripeptide is immobilized. The
excess unbound antibody is free to bind to the surface.
SPR results of milk samples from different producers were in good agree-
ment with those from a number of commercially available screening tests
commonly used for milk analysis and also with an HPLC method. The SPR
assays based on the enzymatic activity of the b-lactam receptor are highly
precise and are sufciently sensitive to detect several b-lactams at or near their
respective MRLs set in European legislation. These assays are thus viable
alternatives for automated screening of b-lactam antibiotics in milk.
The compositional standards for most milk products require that they
contain no other proteins than milk proteins, unless declared. The low price
of some non-milk proteins makes them attractive as potential adulterants in
dairy products. Soya protein is probably the most commonly used non-milk
protein in milk substitutes such as simulated yogurts, coffee whiteners and
frozen desserts and it is likely to be a potential adulterant.
A multiplex biosensor immunoassay was developed by Haasnoot et al. [9] for
the simultaneous detection of soya, pea and soluble wheat proteins (SWP) in
348 Chapter 11
milk powder by using polyclonal antibodies raised against these plant proteins
as the immobilized binding partner [9]. The specicity of the assay was
conrmed by analyzing proteins extracted from other food products. Both
anti-soya and anti-pea antibodies partially cross-reacted with extracts from
sources other than those to which the polyclonal antibodies were originally
raised, showing that the assay could be used for tracing a broad range of
non-milk proteins.
Supplementing bovine milk with cheaper milk from other species is nancially
attractive. Within the EU, however, it is mandatory for producers to state the
type of milk used for manufacturing dairy products as this may be important
information for allergic individuals. Methods used to detect adulteration must
be rapid and have an LOD of o1%.
In an assay developed for bovine milk proteins, two monoclonal antibodies
(Mabs) to bovine b-casein were used as immobilized binding partners [10]. After
conrming the specicity of the assay by testing bovine milk protein solutions of
k-casein, b-casein, g-casein, a-casein and whey proteins such as a-lactalbumin
and b-lactoglobulin, high responses were obtained for bovine milk whereas
responses for ewes and goats milk were much lower. From the milk species
tested, therefore, the two Mabs were sufciently specic for cows milk to enable
adulterated milk to be detected. An inhibition assay was also developed in which
k-casein is immobilized on the sensor surface and the binding response is
measured after incubation of milk samples with Mabs to bovine k-casein.
Different combinations were tested, aiming for a fast assay with 50% inhibition
at around 0.5% cows milk in the milk of ewes and goats.
Both assay formats performed well, giving o0.1% LODs and run times of
around 5 min. Although the direct assay is characterized by its simplicity (single
reagent format), the use of small amounts of antibodies and a broad calibration
curve (0.110% cows milk), the inhibition assay has other advantages due to
the possible application of non-puried Mabs, the higher responses, the higher
sensitivity at relevant low percentages of cows milk and superior robustness
(4800 cycles per chip).
Bioactive minor proteins such as IgG, folic acid binding protein and
lactoferrin are becoming valued for their benecial effects as additives to foods
such as infant formula. Direct biosensor-based immunoassays have been
developed by Indyk et al., in which approximately 60 samples can be analyzed
within 1218 h, with each cycle completed in 815 min [11]. Briey, sensor
surfaces are prepared with anti-bovine IgG, folic acid derivative or anti-bovine
lactoferrin and analytes are prepared for analysis by simple dilution. The
specicity of immobilized ligands is evaluated by measuring cross-reactivity
against individual casein and whey proteins at levels expected in milk.
In comparison with alternative techniques (LC, ELISA, RID) for the analy-
sis of milk products, the mean determined protein levels are consistent
with ranges reported in the literature from assays employing alternative tech-
niques. The minimal requirement for sample manipulation reduces the risk of
recovery losses commonly associated with procedures such as ltration and
centrifugation.
11.3.4 Detecting Antibodies to Salmonella in Meat
Salmonella is a major cause of food-borne bacterial infections with a significant
proportion of these cases associated with the consumption of pork. There are
several stages during which the meat can become contaminated, e.g. infections
may arise at the farm, during transport and in the slaughterhouse. It is
important, therefore, to be able to monitor the whole production process.
Serological assays are currently the most commonly used methods and,
although time consuming, they provide useful information indicating the
prevalence of Salmonella on individual farms. Evidence indicates that
these ubiquitous monitoring steps can help reduce the frequency of human
Salmonella infections.
A more rapid SPR-based assay has been developed to detect antibodies
against Salmonella in pig sera in a busy routine setting [12]. The assay is based
on the immobilization of antigenic bacteria-derived lipopolysaccharides (LPS)
on a sensor surface (the Salmonella serotypes are dened according to the
sugars present on the LPS). The assay was compared with a commercially
available ELISA-based Salmonella assay.
The results of the assays were very similar. Large differences in the preva-
lence of Salmonella antibodies at different farms were noted, with some farms
having positive sera in almost all deliveries while others had only negative sera
in all deliveries. These results indicate that provided enough sera are taken, it is
feasible to screen pig herds for Salmonella antibodies using an SPR-based
biosensor assay. Where a quick result is needed, for instance when a pig herd is
suspected of infection with a notiable disease such as classical swine fever or
foot and mouth disease, SPR-based testing can provide an answer within a few
minutes and control measures can be implemented promptly.
11.3.5 Genetically Modied Organisms
In the wake of an investigation in 2003 into the content of food products, the
Swedish National Food Administration (NFA) found gene-modied organisms
(GMOs) in several foods that were not declared as such. GMOs were found in
14% of food products containing soya or corn. Although labeling demands are
generally being observed, trace amounts of GMOs are nding their way into
nal products. In addition, both the source and means of introduction of
GMOs into the manufacturing process remain to be conclusively identied.
Gene-modied crops are known to be able to spread in nature and, in addition,
raw ingredients must pass through many environments from cultivation to the
outlets, including growers, mills, lorries, silos and deliverers, all of which
present a potential point of entry for GMOs into the manufacturing process
of an ostensibly GMO-free food product. Sensitive assays have been developed
by Feriotto and colleagues to address this issue [13,14] based on DNA
hybridization.
In a protocol described by Feriotto et al., hybridization was monitored by
immobilizing PCR products generated using DNA isolated from normal or
350 Chapter 11
transgenic soyabeans on a sensor surface. Oligonucleotides or PCR-generated
probes from suspected sources were then injected over the prepared surface.
PCR-generated probes are far more efcient than oligonucleotides in detecting
GMOs, allowing the detection of trace quantities. Their results indicate that
the need for automated detection systems for GMO screening in food makes
SPR-based biosensors strong candidates for routine procedures.
11.4 Conclusions
Although the screening and quantication assays using Biacore Q and Qex
Kits or custom-built assays in this chapter cover a wide range of additives to
and contaminants of food products, they are all characterized by a set of
common advantages. SPR assays are highly automated, they are easily con-
trolled and evaluated via software, no uorescent or radioactive labels are
required and the samples seldom require clean-up or solvent extraction. In
addition, the availability of kits providing pre-immobilized sensor surfaces
further reduces the manual input required by the user and provides a stand-
ardized sensor surface. Lastly, the possibility to easily regenerate and reuse
sensor surfaces with no significant loss in assay performance is a major
attraction in the quest for an assay that delivers consistent data over several
hundred runs, permitting valid comparative studies between geographically
disparate locations or at different times.
Together with the inherent robustness of the technology and the proven
resistance of Biacore Q to stressful industrial environments, these features add
to the efciency and productivity of screening and quantication. For example,
the ability to identify the presence of veterinary drug residues in animal
carcasses on the slaughterhouse oor means that steps can be taken to address
the problem immediately, before the meat has left the factory. Import controls
are a further area in which a short screen-to-result time can allow regulators to
act on imported food containing banned substances before it has become
widely distributed throughout the recipient country.
There remains scope for the development and application of an instru-
ment that can handle more samples simultaneously in order to control food
safety hazards within the production chain. To this end, a multi-channel,
high-throughput biosensor prototype instrument was recently developed at
Biacore [15] and is designed to be used in conjunction with a commercially
available automatic sample pipetting station in both laboratory and abattoir
environments. A number of differences exist between this instrument and its
predecessors. The most important of these changes is the prototype auto-
sampler design which allows eight samples to be collected and analyzed
simultaneously. The application wizard permits automatic result evaluation.
The analysis time for one full 96-well microtitre plate for the sulfon-
amide assay was approximately 50 min as opposed to 9 h on a conventional
instrument.
Although SPR-based biosensors remain to make significant inroads into
many areas of food safety testing and analysis, all comparative tests to date
demonstrate that in addition to delivering data of a quality at least as high as its
main competitors, it is safe and easy to use, with clear advantages of speed and
throughput.
11.5 Questions
1. In SPR instruments the ligand is always coupled to the sensor surface.
Explain why the terminology ligand and analyte in inhibition tests is
sometimes confusing.
2. Why should we apply other assay formats for low molecular weight
protein detection?
3. Explain the four main SPR assays and give reasons for choosing these
assays in typical applications.
4. Try to address the complexity of the determination of kinetic rate con-
stants in sandwich assays.
5. Why are SPR assays as shown in the Biacore kits attractive for the
analysis of compounds in the food industry?
References
1. G.A. Baxter, M.C. OConnor, S.A. Haughey, S.R.H. Crooks and
C.T. Elliott, The Analyst, 1999, 124, 13151318.
2. P. Bjurling, M. Caselunghe, C. Jonson, B. Persson, G.A. Baxter,
M. OConnor and C.T. Elliott, The Analyst, 2000, 125, 17711774.
3. V. Gaudin and P. Maris, Food Agric. Immunol., 2001, 13, 7786.
4. L. Wahlstro m and G.A. Baxter, Biacore J., 2005, 5, 811.
5. Biacore Application Note 51, Product No. BR-9003-55 (available from
www.biacore.com).
6. M.A. Johansson and K.-E. Hellena s, Food Agric. Immunol., 2003, 15,
197205.
7. M.A. Johansson and K.-E. Hellena s, Int. J. Food Sci. Technol., 2004, 39,
891898.
8. E. Gustavsson, P. Bjurling, J. Degelaen and A. Sternesjoe, Food Agric.
Immunol., 2002, 14, 121131.
9. W. Haasnoot, K. Olieman, G. Cazemier and R. Verheijen, J. Agric. Food
Chem., 2001, 49, 52015206.
10. W. Haasnoot, N.G.E. Smits, A.E.M. Kemmers-Voncken and M.G.E.G.
Bremer, J. Dairy Res., 2004, 71, 322329.
11. H.E. Indyk, E.L. Filonzi and L.W. Gapper, J. AOAC Int., 2006, 89,
898902.
12. R. Achterberg, J. Maneschijn-Bonsing, R. Bloemraad, M. Swanenburg
and K. Maassen, Biacore J., 2005, 5, 1618.
352 Chapter 11
13. G. Feriotto, M. Borgatti, C. Mischiati, N. Bianchi and R. Gambari,
J. Agric. Food Chem., 2002, 50, 955962.
14. G. Feriotto, S. Gardenghi, N. Bianchi and R. Gambari, J. Agric. Food
Chem., 2003, 51, 46404646.
15. C. Situ, S.R.H. Crooks, G.A. Baxter, J. Ferguson and C.T. Elliott, Anal.
Chim. Acta., 2002, 473, 143149.
CHAPTER 12
Future Trends in SPR
Technology
RICHARD B.M. SCHASFOORT
a
AND PETER SCHUCK
b
a
b
Protein
Biophysics Resource, DBEPS, National Institute of Biomedical Imaging and
Bioengineering, National Institutes of Health, Bethesda, MD 20892, USA
12.1 Introduction
We anticipate that surface plasmon resonance (SPR) technology will remain
the gold standard in direct biomolecular interaction sensing during the next
decade. Although in the past only one company (Biacore AB) mainly domi-
nated the market (B90%) of high-quality SPR systems including optics, liquid
handling and sensor chips, new players and new trends can be identied. As
highlighted in Chapter 3, with 25 SPR-related companies the market is now
more open than ever before and competition between companies takes place on
several aspects of the SPR system. The customers for instruments will prot
further from this competition, oering more exibility, innovation and cost-
eectiveness. The eld of clinical analysis including proteomics in all its facets is
expected to undergo a revolutionary change with the introduction of new multi-
analyte diagnostic SPR systems. Mainly the combination of SPR imaging and
dedicated microuidic lab-on-a-chip may drive the technology to another
level of commercialization where point of care (POC) devices for specic
applications may be the ultimate objective. For applications in the eld of
kinetic and thermodynamic characterization of molecular binding parameters,
higher computational power readily available with desktop PCs brings more
sophisticated approaches within reach.
In this Chapter, anticipated trends regarding the SPR systems are high-
lighted. Necessarily, this may sometimes have a speculative character and we
354
recognize that, unavoidably, such an attempt may be regarded as biased to
some extent by the interests and unique perspectives of the authors. However,
we hope that this Chapter will, at the minimum, provide a description of some
exciting prospects for the coming decade in the main areas of SPR development.
It is structured along the three essential units that make up SPR instruments:
(1) the detection instrumentation, (2) the uidics and (3) sensor surfaces/chips.
These units are complementary to each other and all critical for the quality and
performance of the total SPR system and they are complemented by the data
analysis. In the following, we will discuss our perception of the trends in each of
these elds.
12.2 Trends in SPR Instrumentation
Although the improvements and functionality of the Biacore line of instru-
ments during the almost two decades of SPR technology are impressive, new
trends in segmented parts in the market are appearing for more instruments.
The proteomics area demands screening of a large number of analytes in
complex samples, exceeding by far the currently readily available number of
420 and multi-analyte parallel diagnostics of kinetic parameters are desired.
The FLEXChip instrument of Biacore is the answer for screening many
biomolecular interactions simultaneously. For the next 5 years it is predicted
that hundreds to thousands of simultaneous biomolecular interactions will
need to be measured for screening the quality of binders. SPR technology
should follow the protein microarray technology and kinetic parameters should
be determined in a highly parallel manner. Instruments with SPR imaging will
meet these requirements. Another new development is the combination of SPR
with a complementary technology, so-called hyphenation SPR. Some trends
are given in this Chapter and we expect that the combination with mass
spectrometry (SPRMS) will increase in importance. The implementation of
lab-on-a-chip technology with SPR is the second important trend for a huge
number of new applications.
Right from the beginning of the introduction of SPR technology, one goal
was the use of SPR for point of care diagnostic devices, but several attempts
were unsuccessful. Why is this development of point of care SPR devices so
problematic? Is it the sensitivity which should compete with other immuno-
chemical tests (e.g. dipsticks) using labels? SPR has a resolution in the order of
10
5
refractive index change. If new instruments are developed based on the
plasmonic eect in nanoparticles as described in Chapter 2, a new generation of
instruments with unmatched high sensitivity is foreseen. However, the limiting
factor will not be the sensitivity of refractive index changes, but the ratio
between specic and non-specic binding. This determines the limit of detection
for concentration measurements and therefore label-free SPR cannot compete
with the best labeling technology. In the latter, any non-specically interacting
unlabeled species will not contribute to the signal. The nanoparticle SPR
approach is unequivocally highly relevant; however, it will take many years
355 Future Trends in SPR Technology
to implement other important aspects for the study of biomolecular interac-
tions, such as uidics, nanoparticle surface functionalization and the investiga-
tion of this technology for the determination of reliable kinetic parameters.
The rst Biacore instrument in 1990 contained a so-called optogel for use as
the medium between the prism in the optical unit of the instrument and the
sensor chip, which turned out to be a crucial factor in the early development of
the instruments. The optogel, which is considered part of the success of Biacore,
ensures optical contact, simplifying exchange of the sensor chip. Some com-
mercial SPR players in the market still apply refractive index matching oil,
although alternatives have been published. For example, Masadome et al. [1]
developed a refractive index matching polymer lm for a portable SPR system.
It is expected that successful SPR instruments will have a solution for the
optical matching problem and will no longer need RI matching oil. Several
manufacturers of commercial SPR instruments already use alternative optical
arrangements with (expensive) disposable prisms. An alternative optical ar-
rangement using polymeric diractive optical coupling elements (DOCEs) was
published by Thirstrup et al. [2], which is an attractive solution to prevent the
optical matching problem.
12.2.1 SPR Imaging
SPR imaging has the ability to obtain a microscopic view of the sensor surface
and dene certain regions of interest (ROIs) for measurement of many
biomolecular interactions at the same time. Reference spots, positive and
negative controls to determine the non-specic binding and/or cross-over
interactions, as well as triplicates or higher replicates of identical interactions
for checking the variability of the sensor surface are helpful to obtain reliable,
accurate and valid data. Common mode eects caused by temperature changes,
bulk refractive index shift or ow direction shifts can be compensated using
these reference spots. However, as explained in Chapter 3, simple reectivity
measurements can only give qualitative and not quantitative data which are
necessary for in-depth studies of the kinetics of the binding process on spotted
microarrays with dierent ligands. For every spot the shift of the SPR dip
should be followed to allow subtraction and referencing. The data should not
include an instrumental artifact but should show the real kinetic data of the
parallel interaction process. Nevertheless, for expert users it is still possible to
obtain relevant data from reectivity instruments in carefully conducted
experiments. In Chapter 3, Section 3.4.6, the companies that have SPR imaging
instruments (including FLEXChip) are briey described.
12.2.2 Hyphenation SPR Technology
New developments are combinations of SPR with a complementary technol-
ogy, also referred to as hyphenation SPR. Often, these technologies are o-
line and should be used serially, but both technologies share the SPR sensor
356 Chapter 12
chip, which can be installed in the SPR instrument as well as inspected with the
other hyphenation technique. An exception to this strategy is the combina-
tion with SPR excited uorescence, which is treated in-depth in Chapter 9.
12.2.2.1 SPRMS
Krone et al. [3] rst combined SPR and mass spectrometry (MS), which created
a unique approach for protein investigations. This technique was adapted and
modied in several laboratories [47] and some of the commercial instruments
are now also equipped with basic analyte recovery capability for MS. The basic
idea is to follow up the characterization of interactions between proteins and
surface-immobilized ligands by SPR with the determination of the identity of
the bound proteins or peptides using MS. This has applications in protein
interaction discovery in proteomics, and also the characterization of protein
modications critical for the interaction. Considering the SPR sensor surface as
a miniaturized chromatographic matrix, SPRMS is reminiscent also of the
traditional anity chromatographic purication preceding MS in protein
discovery, but obviously with added real-time quantication of the capturing
and elution process, in addition to exhibiting dierent elution behavior due to
the much smaller scale matrix. Several fundamentally dierent interface
approaches have been developed by dierent groups, for example (1) direct
use of the SPR chip surface for MS by mounting the preloaded SPR chip on a
MALDI platform [3] or analyzing the surface with a SELDI protein chip reader
[7]; (2) on-chip digestion followed by microuidic elution, recovery in a
reversed-phase microcapillary column and ESI-MS/MS sequencing of peptides
from the digest [5]; and (3) analyte dissociation, microuidic elution and
collection followed by external digestion and MALDI-TOF sequencing [4,9].
Recent progress in SPRMS includes improved methods and operations,
increased limits of detection, multi-protein analysis and protein-complex
delineation. With the subsequent design of SPR protein arrays, SPRMS enters
into the eld of high-throughput protein interaction discovery and miniaturized
diagnostics.
Below we briey highlight an example of SPR detection of an enrichment
and elution process of specically bound analyte as an application of SPRMS
by Visser et al. (Hecks group) at the University of Utrecht (The Netherlands)
[8]. In this work, an SPR-based chemical proteomics approach in combination
with nano-liquid chromatography/electrospray tandem mass spectrometry
(nanoLC/ESI-MS/MS) was used to obtain both semi-quantitative and quali-
tative data on enriched proteins obtained from SPR sensor surfaces. The aim of
the research was the characterization and identication of autoimmune anti-
bodyantigen complexes present in RA patient sera. The autoimmune anti-
bodies can be enriched using a citrullinated peptide (i.e. a specic peptide where
the arginine is replaced by citrulline) as a ligand which is immobilized on a gel-
type G-COOH sp IBIS sensor chip. The work is connected to the research
described in Chapter 7, which should be consulted for details and references.
The analysis cycle can be followed in real time as shown in Figure 7.10. A three-
step sequential elution procedure (see Figure 12.1) was conducted to increase
the selectivity and to permit discrimination between non-specic and specic
binding.
It is necessary to dissociate non-specically and less-specically bound
material with low anity from the sensor chip in the rst elution step. Then
selectively bound proteins should be competitively dissociated in a second
elution step, which could in this case be achieved using a competing ligand
peptide (Pept.pos) to that immobilized ligand. This diers from other
approaches where reversible chemical denaturation is applied for the analyte
release. This present approach works for analyte molecules where the capture is
kinetically aided by rebinding arising from mass transport limitation. If the
protein dissociates from the surface in the presence of the competitor, it will
bind to free ligand in solution and rebinding to the surface will not occur. (Note
that for SPR instruments the ligand is considered to be immobilized to the
surface. The specic elution with ligand can be confused with an inhibition
assay format as described in Figures 7.2 and 11.2.) In the third elution step, the
2. Peptide: pos.(citr)
C
O
O
H
C
O
O
H
C
O
O
H
C
O
O
H
C
O
O
H
C
O
O
H
C
O
O
H
C
O
O
H
1. EDC/NHS
3. EtAm
TB34 serum
Sequential
elution:
1. Pept.neg.
2. Pept.pos.
3. Gly
Pept.neg
1. Trypsin,O/N, RT
2. LC-ESI-MS/MS
Pept. pos
Gly
Figure 12.1 Schematic set-up of the autoimmune antibodyprotein complex detection
and enrichment procedure in three steps from a specic RA serum (TB34)
using a citrullinated peptide as the ligand for binding of specic auto-
immune antibodies (Pept.pos a specic peptide, where the arginine is
replaced by citrulline). The rst elution was with negative peptides
(Pept.neg where the arginine was not replaced by citrulline), followed
by a second elution with the ligand in solution (Pept.pos) to replace the
bound fraction with free competing ligands in solution and nally a full
regeneration step (Gly).
358 Chapter 12
sensor was regenerated with a low-pH glycineHCl (gly) buer, removing the
remaining bound material from the sensor disc. Each of the fractions was
collected and digested. The protein content was characterized using a nano-LC/
ESI-LTQ set-up.
Preliminary results showed a three-fold increase in dissociation of the second
elution with the competing ligand with respect to the rst elution. The
remaining bound material can be eluted completely in the regeneration step
with the low-pH (gly) buer. As with other methods for SPRMS interfacing,
to obtain a fraction containing sucient protein which was specically bound
to the citrullinated peptide and which has been successfully eluted, is not easy.
The eluted specic protein should be as clean as possible without interference
from highly abundant non-specic proteins or non-isolated competing free
ligands for successful detection with the LC/ESI-MS/MS. The o-line combi-
nation of SPR and nano-LC/ESI-MS/MS needs further to be optimized and
may lead to enrichment and identication of autoimmune antibodyantigen
complexes using immobilized target peptides combined with the sequential
elution approach. New results will be published by Carol-Visser and colleagues
soon.
Some of these diculties regarding high selective loading and recovery may
be addressed with improved microuidic liquid handling. As illustrated by
Gilligan et al. [9], the manipulation of small distinct liquid volumes in the
microuidics with reversible and oscillatory ow patterns [10] can provide
signicantly increased amounts of recovered material and controlled washing
conditions. In this approach, small liquid plugs containing the loading sample,
washing buers and elution solution are separately transferred to the sensor
surface such that multiple cycles of loading, washing and recovery can take
place, enriching the eluent concentration in the same plug of recovery buer [9].
Further, by virtue of the oscillatory ow pattern once the specic liquid
volumes are covering the ow cells, the contact times can be extended (while
maintaining high mass transfer to the capturing molecules at the surface) until
the slow binding kinetics arising from the low antigen concentration has
achieved a plateau. The latter sample handling technique has been applied,
for example, to minimize sample consumption and optimize detection eciency
in an SPR based assay for the detection of anti-idiotypic antibodies in patient
sera [11]. Another eective approach for improved recovery from the SPR
surface is the use of a larger surface [12].
12.2.2.2 Other Hyphenated SPR Techniques
A way to enhance sensitivity and to push the limit of detection (LOD) to lower
surface coverages is the use of uorescent chromophores covalently attached to
the analyte molecules. In this approach of surface plasmon uorescence
spectroscopy (SPFS), the resonantly excited surface plasmon waves excite the
uorophores [13] and their emitted photons can be monitored by a simple
detection unit attached to a conventional SPR set-up. In Chapter 9 the features
and benets of such an SPR excited uorescence instrument are described. The
preferred mode of operation is to monitor the uorescence emitted directly
from chromophores suciently separated from the substrate surface. The dye
molecules are still within the substantially enhanced optical eld of the surface
plasmon mode, but they are not quenched. This combination of eld enhance-
ment and uorescence detection forms the basis for the largely enhanced
sensitivity applied for a wide range of bioanity studies (for selected examples,
see Chapter 9).
As shown in the previous section, SPR excitation in the SPFS instrument can
be combined with instrumentation assembled at the liquid side. SPR optics in
the Kretschmann conguration require only the assembly of optical compo-
nents at the prism side and therefore approaching the instrument at the wet side
with ow cells in microuidic cartridges, cuvettes or even lab-on-a-chip devices
is in principle feasible. The implementation of lab-on-a-chip devices is discussed
in Section 12.3.3.
The lateral resolution of SPR, which is equal to the propagation length of the
plasmon wave (e.g. for a wavelength of 680 nm it is of the order of 10 mm), is
sometimes unacceptably large for the imaging of small features, such as ligand
clusters, aggregates or brils with submicrometer dimensions. If a higher lateral
resolution is needed, then other combinations of instruments or hyphenation
SPR technology are required. A combination of atomic force microscopy
(AFM) and SPR [14] seems in principle a very interesting approach. Combining
the dynamic topographical data from AFM with the kinetic data from SPR
may be applied in many elds of materials research, particularly in the design of
biomaterials, where dynamic surface changes, such as protein aggregate
adsorption/desorption processes, play a role. Another approach is the exploita-
tion of SPR as a complementary surface analytical tool to scanning probe
microscopy (SPM). The potential for a combined SPMSPR approach for the
analysis of biomaterial surfaces was demonstrated in 1994 [15]. It is expected,
although speculative, that commercial SPR imaging combinations with AFM,
STM [16], SPM, etc., will enter the market in the next 5 years.
12.2.3 Nanoparticle SPR
SPR phenomena are not restricted to planar multilayers as discussed in this
book: for metal particles, usually gold, with dimensions much smaller than the
wavelength of the interacting light, surface plasmon eects can be much more
prominent (see Chapter 2, Section 2.5.2). Nanostructured surfaces, such as
nanoholes, can also be applied to exploit surface plasmon/plasmonic eects for
sensing biomolecular interactions [17]. The use of metal nanoparticles as
surface plasmon-assisted eld ampliers is described in Section 8.4. However,
these particles can also be exploited as intrinsic refractive index sensors,
analogous to the more familiar planar SPR experiments (for a review, see
[18]). The physical basis of this application is the light extinction (absorption
and scattering) which is heavily dependent on the nanoparticles dielectric
constant, size and geometry and also on the dielectric constant of the
360 Chapter 12
surrounding medium. Van Duynes group [19] developed a silver nanoparticle-
based LSPR nanosensor which yields ultrasensitive biodetection with extremely
simple, small, light, robust and low-cost instrumentation. They used this LSPR
spectroscopy to detect less than 1 pM up to micromolar concentrations of
biological molecules. However, the silver nanoparticles applied are intrinsically
less inert as sensing elements than gold nanoparticles. Gold nanoparticles of
dened dimensions can be coated on a substrate in order to enable the easy
exchange of liquids similar to at SPR instruments. Hong and Kao [20]
developed such a gold nanoparticle-coated lm to achieve highly spatially
resolved biosensing that is based on localized SPR. It was reported that unlike
the planar gold lm employed for conventional SPR sensing, the gold nano-
particle lm relies exclusively on shifting the peak extinction wavelength for the
detection of biomolecular interactions and that it does not depend critically on
the angle of incidence.
Magnetic particles, which can be recruited to the sensor surface by magnetic
elds, are dicult to use in Kretschmann-operated SPR sensors, where the
magnetic eld would need to be raised from the prism using bulky coils.
However, for the phenomena used in giant magnetic resistors (GMRs) [21], an
electrical current is owing to attract magnetized particles to a sensing metal
line. A simple electrical conguration can be made using gold lines both as an
electric current actuator to attract the particles and as an SPR sensing device.
Essentially, it should be possible to apply SPR imaging in a hyphenation
approach of magnetism and SPR. However, to our knowledge, such a
combination has not yet been realized.
12.3 Trends in Fluidics
For the binding to the SPR surface to have maximum sensitivity and the signal
to reect the intrinsic kinetic and thermodynamic properties of the molecules
under study, microuidics appears to be the most powerful approach. Attrac-
tive features are excellent baseline stability, low sample consumption and the
potential for relatively high mass transfer when high ow rates and thin
channels are used. As already described in Section 3.3.1 and Figure 3.9, ow
channels are formed, in principle, by pressing a grooved surface against the
sensor chip. Biacore introduced a microuidic cartridge for sample delivery in
1990, with incorporated pneumatic values that allow closing of specic
channels, thereby providing control over the ow paths. This allowed the
sample or buer liquid volumes to be directed to dierent surface spots. In this
microuidic cartridge, also integrated was an (HPLC-like) injection loop that
allowed a xed sample volume to be chased by running buer and owed over
the surface.
Although this technology has been shown to be fairly powerful and far
superior with regard to stability in comparison with simple cuvette-based
systems, some improvements were subsequently addressed in dierent designs.
For example, the constraints of the injection loop with nite xed sample
volumes allowing only limited sample contact times and ow rates were
overcome with the oscillatory ow technique [10]. This permits the use of
sample plugs of smaller volume, yet allowing simultaneously very long contact
times while maintaining high ow rates and mass transfer. This can be very
important when working with limited sample volumes, for better characterizing
the thermodynamic binding parameters by permitting the binding progress to
reach the steady state [10] and be utilized to improve the sensitivity for analyte
detection at low concentrations [11] and in interfacing SPR with MS as outlined
above [9].
To improve the usage of each sensor ow channel, Biacore has more recently
introduced the option of hydrodynamically addressing dierent spots within
each channel, thus multiplexing the use of each sensor ow channel as shown in
the Section 3.5, Figures 3.35 and 3.36. A dierent microuidic design was
implemented in the commercial Prote-On system by Bio-Rad, where a criss-
cross pattern of microuidic channels allows one to measure the binding of a
series of several samples to several dierent sensor spots in parallel a design
particularly suited to SPR imaging (see Figure 3.29). The latter multi-analyte
detection permits experimental designs that do not require surface regenera-
tion. The trend for multiple ligands to be attached to the sensor surface will
continue in the future.
The body of literature on lab-on-a-chip is expanding at a fast pace, but the
eld is still far from mature. The combination of lab-on-a-chip devices with
SPR sensing is even in its infancy. Advanced techniques for microarray
spotting are reviewed in more detail in Section 12.3.1, and prospects for point
of care use of SPR devices are discussed in Section 12.3.2.
Limitations inherent in laminar ow-based microuidics arising from diu-
sion and viscosity can be overcome using dierent ow principles for sample
delivery. This includes the use of the electroosmotic ow principle to pump
liquids in microchannels, which will be treated in Section 12.3.3. A combination
of ow with a separation system is the microuidic free-ow electrophoresis
device for proteomics, which is reviewed in Section 12.3.4. Another highly
promising development is digital microuidics using electrowetting principles,
as reviewed in Section 12.3.5.
12.3.1 Microarray Spotting on Gold
With the introduction of protein microarrays in the late 1990s, there was
tremendous excitement about the potential of protein arrays to improve further
our understanding of protein expression, function and structure on a scale
approaching the proteome. However, there was also from the beginning
hesitancy by many scientists to adopt a technology that is often still perceived
as unstable and irreproducible. Protein microarrays were developed largely by
extending technologies used for gene chips [22]. Most protein arrays as
currently developed rely on detection technologies that apply uorescent tags.
Denitely, uorescence detection methods are successful for gene chips, but
362 Chapter 12
much less convenient with protein chips due to the heterogeneity of proteins,
diculties in the synthesis of conjugates and the potential for non-specic
binding. Also, signal-producing reactions in solution catalyzed by commonly
used enzyme-linked antibodies are dicult to implement in an immunoassay
microarray format, where the product will diuse away from the surface,
diminishing the potential for the discrimination of spots. However, in Section
8.3 an enzymatic amplication method is described for SPR sensing, which
applies a localized non-soluble precipitate, which is detected by SPR.
Protein microarrays can be generally divided into two main categories:
capture arrays and interaction arrays. Capture arrays have immobilized
molecules such as antibodies or chemically treated surfaces, which bind
generally with high anity to a specic, known ligand. Interaction arrays have
ligands which are used to identify functions or determine directly interactions
with other, frequently unknown, analytes. The detection is generally accom-
plished with a non-interfering labeled conjugate in a two-step sandwich. Is SPR
microarray imaging the answer for measuring biomolecular interactions in a
reliable way without the need for labeling? In Chapter 3 many instruments
including FLEXChip, Lumera, K-MAC, Bio-Rad and IBIS are described
which can implement microarray technologies.
A main challenge is how to spot the ligands to the SPR sensor surface.
Initially, pioneers in the eld used laboratory-made equipment as designed and
published by Brown and co-workers (see http://cmgm.stanford.edu/pbrown).
Recently, several commercial array systems have been developed for printing
target DNA on microscope slides which can be adapted to create protein
microarrays. In short, every arrayer consists of an XYZ robot of which the
printhead travels between a microtiter plate, containing the ligand and the
sensor surface, which is a standard 3 1 inch glass microscope slide. Approxi-
mately 16 spots per mm
2
can be printed when spots have a pitch width of
250 mm. The dierences between the available systems are mainly based on the
structure of the printhead: there are several dierent ways of picking up a small
amount of ligand solution and printing small amounts of this solution in an
orderly and systematic way on the sensor surface. Examples of arrayer
methodologies include piezo technology (inkjet dispensers), quills (split-pen)
and pen and ring systems (see the webpage cited above). As a consequence,
the systems dier considerably in reliability, accuracy, capacity and the
required (starting) volume of sample.
One of the main diculties that distinguish DNA and protein arraying is that
DNA spots do withstand drying of the surface while many protein ligands are
prone to various conformation changes and denaturation after evaporation of
the liquid, which frequently aects biological binding properties or may
eliminate binding altogether. In the next section we present briey a strategy
for producing high-density protein microarrays using a DNA coding technique
in a conned microuidics set-up including a self-assembly process of proteins
onto individual addressable microstructures. Spots of 10 mm are feasible, which
correspond to the propagation length of the plasmon wave, allowing the
creation of a microarray with 2500 individually addressable protein spots per
mm
2
. This is the physical maximum for Kretschmann-congured SPR instru-
ments.
12.3.1.1 DNA Coding Technology
The so-called DNA coding technology is proposed in Figure 12.2. First a substrate
is coated with a homogeneous coating (e.g. streptavidin). To dene the array, a
PDMS device is fabricated with channels and reservoirs (see Figure 12.3).
The reservoirs in the PDMS device are lled with multiple solutions of
biotinylated single-stranded DNA, each having a specic sequence. When these
solutions ow through the channels the single-stranded DNA is coupled to the
streptavidinated substrate surface. After immobilization of the rst lane, the
streptavidin coating is inactivated with an excess of biotin. Next, a second,
equal PDMS device is placed in a perpendicular position and the reservoirs are
lled with multiple solutions of dierent mixtures of complementary DNA
conjugates. Each of the DNA conjugates hybridizes in a self-assembly process
to its specic lane. At each individual crossing of lanes the self-assembled
Figure 12.2 Schematic presentation of the construction of a microarray using DNA
coding technology. First, lanes of biotin-labeled reverse oligonucleotides
are immobilized on the streptavidin surface. The spot is dened by
hybridizing a mixture of complementary DNAprotein conjugate in
the perpendicular direction. During the self-assembly process the forward
complementary oligonucleotides are immobilized and an ultra-high-
density protein microarray can be created. The DNA coding technology
can be extended to fabricate, for example, a 50 50 microarray.
364 Chapter 12
molecules will dene a dierent array square spot. For a high-density protein
array, a mixture should be made of a DNA sequence coupled to a protein
(proteinssDNA or DNA conjugate). To address each spot, a mixture of
dierent proteinDNA complementary codons should be prepared, which in
principle can be produced by molecular synthesis and can be automated using
liquid handler robots.
The physical dimension of the array chip is dened by the channel width of
the PDMS device and by the number of spots in the array. Although the lateral
resolution of SPR imaging is limited by the propagation length of the surface
plasmon wave (e.g. B10 mm) in the clean room of the MESA+ Institute,
University of Twente, even 1 mm wide lines can be processed. The novel method
allows the fabrication of a high-density multi-ligand array, without exposing
the sample to air, thus avoiding denaturation of proteins. It is important that
non-specic binding should be suppressed to a high degree to have optimal
benet from the self-assembly process.
Kanda et al. [23] used PDMS microuidic devices as a surface to pattern the
gold and adsorb antigens from the sample this device had the potential to
produce arrays of 64 or more spots. In the biochip laboratories of the
University of Twente, Beusink and colleagues developed PDMS devices for
spotting ligands for immobilization in small, spatially separated lanes on a
sensor chip. In Figure 12.3 a six line-spotter is shown for the immobilization of
six ligands (unpublished results).
A critical advantage of the proposed coding system is that it can create ultra-
high-density microarrays without using droplet-based spotting procedures.
Line spotter microuidic devices allow us to fabricate conned microarrays
Figure 12.3 Left: PDMS line-spotter (50 mm lane) on top of an iSPR sensor chip.
Right: iSPR image of the six lines produced by the line spotter and the
ROIs in red. The rst DNA coded lanes can be created and the PDMS
device can be removed. In principle, drying of the DNA-lanes is allowed
to occur. A second similar device can be placed perpendicular to the rst
creating a 6 6 array of dierent protein ligands using the DNA coding
technique.
in a two-step process, as shown here, for an immunochemical biomolecular
interaction. Although the technology is still in its infancy, applications
are foreseen in the areas of disease monitoring, proteomics and genomics
studies.
12.3.2 Prospects for SPR-based Point of Care Devices
Clinicians are beginning to use point of care (POC) testing of compact (in
terms of size and weight) and exible clinical chemistry testing devices suitable
for use close to the patient. These analytical devices are designed to move
diagnostic testing out of central laboratories into sites closer to the patient. In
the future, households might also be equipped with small, user-friendly devices
to monitor the daily health status based on measurements of small samples [24].
Miniaturization of devices will oer advantages when rapid and selective
monitoring is required, e.g. of cardiac markers for diagnosing acute myocardial
infarction [25] or whole blood chemistry relevant to intensive care medicine
[26,27].
Why are POC SPR devices not yet in the market? We can identify the
following ve reasons:
1. SPR-based concentration measurement is never as sensitive as labeled
techniques because of the intrinsic and inherent drawback of SPR: the
detection of non-specic binding interferes with the specic binding signal
intended to be measured. In contrast, an immunoassay that detects only
the label will not be susceptible to signal interference from non-specic
binding of unlabeled proteins.
2. For POC analysis there is still not yet the absolute necessity to detect
kinetic rateequilibrium constants of biomolecular interactions, which are
the unique features of SPR.
3. Current SPR instruments are still bulky and expensive and are not in a
state for production for high-volume markets. The SPREETA chip of
Texas Instruments was the ultimate conguration for POC devices but
this application has not developed as successfully as expected.
4. The advantages of direct detection and speed of SPR appeared to be not
(yet) the crucial factor making it suitable for POC devices and giving it a
decisive advantage, e.g. over dipstick tests, which seem to be fast enough
for current applications.
5. The cost aspect of expensive labeling is not (yet) the remaining argument
to replace labeled tests for SPRPOC devices.
If POC tests can be designed where the kinetics of the biomolecular interaction
determine the outcome of the test, then SPRPOC devices are attractive for
the market and can be developed with a great intrinsic advantage. However,
at the present time, in the absence of this type of POC test, it is still question
able whether SPR can compete with current label technology in POC
devices.
366 Chapter 12
12.3.3 Implementation of Lab-on-a-Chip Devices for SPR
Systems
In 1990, Biacore introduced a fully automated pneumatic valve-operated
microuidic cartridge for biospecic interaction analysis, which was at that
time the most high-tech and advanced uidic system developed in a commercial
instrument. However since the introduction in 1990 by Manz et al. [28] of
miniaturized total analysis systems (m-TAS), an enormous research eort
[29,30] has taken place in the area of miniaturized devices or labs-on-a-chip
with thousands of papers in 2006. It may once have seemed an utopian dream
to create highly parallel and automated microfabricated devices for SPR
systems. However, considering the new lab-on-a-chip trends for SPR systems
in this section, we hope to convince the reader that the rapidly unfolding reality
of lab-on-a-chip technologies for new SPR instruments will pave the way to
achieving highly parallel and automated microanalyses of biological processes.
Table 12.1 contains a compilation of some potential building blocks for lab-
on-a-chip devices that are useful in combination with SPR detection. In this
chapter, only a few of these building blocks will be described in detail. The most
important trends regarding the implementation of potential lab-on-a-chip
approaches into SPR systems are discussed.
12.3.3.1 Pumping Liquids Using Electroosmotic Flow
in Microuidic Devices with Gold Layers
When an electric eld is applied in the longitudinal direction of a charged
surface (substrate) or capillary, the cations close to the wall move towards the
cathode (Figure 12.4). Wrapped in the layer of cations, the bulk solution is
transported in the direction of the cathode. The only plane of friction is
between the stationary layer at the capillary wall and the layer of cations
which is in motion. As illustrated in Figure 12.4, unlike in pressure-driven
systems with parabolic ow prole (Poiseuille ow), the velocity of the bulk is
constant resulting in a at ow prole (plug ow). The stagnant diusive layer
Table 12.1 Potential lab-on-a-chip building blocks for SPR devices.
Lab-on-a-chip
building block Description
Channel Formed by dry/wet etching, molding or soft lithography (e.g.
PDMS)
Pump Enables uid transport which can be driven electro- or
hydrodynamically
Mixer Splitting or coiling of a laminar ow
Separator Based on capillary or free ow electrophoresis (CE, FFE)
Collector Using hydrodynamic addressing or address ow
Spotting Line, criss-cross or dead end spotting
Detector Functionalized gold surface; patterned gold sensor patch
of e.g. 2 mm (see Chapter 6, Figure 6.5) is reduced to the thickness of the double
layer (B10 nm), which is dependent on the ionic strength of the buer solution.
This may result in improved mass transport of the analyte to the immobilized
ligand. The velocity of the electroosmotic ow (EOF) is given by the following
equation [31]:
v
EOF
e
0
e=4pZ zE 12:1
where e
0
and e denote the dielectric constants of vacuum and the buer,
respectively, and z denotes the zeta potential, the potential at the rst moving
layer at the capillary wall. In order to modify the magnitude and direction of
the EOF, either the lateral electric eld E or the zeta potential should be
modied. The principle of modifying or even reversing the EOF inside a
capillary has been presented before, including the control of ows in integrated
devices [32,33]. As the EOF is generated at the channel wall, the viscosity at the
wall is one of the determining factors for the ow velocity. If a conducting gold
layer is deposited on the glass surface, the electric eld will be inuenced and
changes in the EOF arise. Moreover, this electric eld may have an eect on the
surface plasmons, which, in turn, may aect detection of biomolecular inter-
actions using the SPR phenomenon.
The gold patches for SPR sensing are not directly connected to a power
supply but are oating. A lateral eld is over the gold layer and the
equipotential of the gold is considered to be the average between the potentials
in the electric eld in the liquid at both sides of the gold electrode. Under
conditions where the voltage dierence is low, o1 V, no reactions occur, no
electric current will ow and the double layer of the metal/electrolyte interface
EOF
E
Figure 12.4 The electroosmotic ow prole is plug ow, while hydrodynamic pump-
ing gives a laminar ow prole. The electric eld (E) will drag cations in
the double layer of the wall (B10 nm) to the cathode (negative electrode).
If a conducting gold layer is deposited on the glass surface, the electrical
eld will be short-circuited and aected by the gold through reduction
and oxidation processes at the gold surface. The gold layer can be
considered as a bipolar electrode.
368 Chapter 12
behaves as an insulator. In contrast, when the voltage dierence over the
electrodes becomes too large, an electric current will ow through the metal and
at both ends electrochemical processes or even electrolysis of the liquid will
occur, resulting in bubble generation.
If, on the other hand, the gold electrode is connected in an electronic circuit,
then a changed potential of the electrode will modify the charge distribution
near the metal surface and hence alter the ionic distribution in the double layer.
In principle, these ionic changes will be measured with SPR, because changes
in ionic concentration in the electrical double layer will alter the refractive index
in the evanescent eld. This so-called electrochemical SPR or E-SPR phenom-
enon depends on the ionic composition of the electrolyte and on the electron
concentration in the metal. As described in the literature [34], to a rst
approximation, the refractive index of the double layer can be considered to
vary with the change in the charge of the double layer. Lioubimov et al. [35]
described a combination of oscillating electric potential and SPR measurement.
Tadjeddine [36] discussed a multilayer model of the electrochemical interface in
combination with SPR phenomena.
An example of these eects can be seen in Figure 12.5: when an external
voltage is applied to the channel, one side of the bipolar gold strip turns darker,
while the other side turns lighter. There is a gradient of reected light along the
gold chip. When the voltage polarity is switched, the dark and light ends of the
strip exchange accordingly. A delayed switch-on eect was also observed, in
agreement with experiments observed by Lioubimov et al. [35]: the gradient in
the reected light did not appear instantaneously, indicating a true electro-
chemical process changing local refractive indices, as opposed to a physical
SPR eect based on electron charge concentration dierences in the gold,
which may induce a change in free electron plasma oscillation.
12.3.4 Lab-on-a-Chip Implementation Using Free Flow
Electrophoresis and SPR Imaging
for Proteomics-on-a-Chip
A so-called proteomics-on-a-chip device should permit the separation, detection
and identication of new biomarkers in a micro-fabricated device.
A common problem with miniaturized separation systems based on, e.g.,
capillary electrophoresis is the low sample loading capacity. This problem,
together with the often low concentration of relevant bioactive compounds,
puts severe demands on the detection system. If a separated sample is owing in
peaks over the SPR sensor area, the contact time of the sample with the surface
is limited, making it impossible to detect proteins of low abundance. Therefore,
a peak owing over the sensor area should be trapped (stopped) in order to
allow diusion of analyte to the sensor area. Lammertink et al. [37] described a
peak recirculation approach to allow coupling of CE with SPR with higher
contact times with the sensor area. However, the absolute number of low-
abundant biomolecules in volumes of less than 1 nl will put a severe constraint
on the SPR biosensing system. Therefore, we shifted to another combination of
separation technology and SPR detection.
Recently, we have developed a microuidic free ow electrophoresis (FFE)
device for coupling the device with SPR imaging as a separation and detection
system for biomarker discovery.
FFE is a continuous separation method, providing continuous bands and
thus virtually unlimited amounts of separated components. In FFE an electric
eld is applied perpendicular to the ow direction and charged molecules are
Figure 12.5 Three screenshots (A, B, C) (unpublished results obtained by M. van der
Ploeg) of an SPR imaging experiment in a microuidic channel acquired
at the University of Twente laboratories. The electric eld drops from left
to right. Two containers as shown in Figure 12.4 are outside this image
and contain the Pt electrodes At A, no external voltage was applied, at B
an external voltage of less than +2 V over the gold electrode was applied
(note that the gray values at the two ends of the gold strip dier), at C an
external voltage of o2 V was applied. The reectivities at the two ends
of the gold strip exchanged as a result of the electric eld switch. A
10 mM HEPES buer was used in the channel.
370 Chapter 12
deected from the carrier ow direction, in a way that is controlled by the
electrophoretic mobility, the ow velocity and the electrical eld strength.
Figure 12.6 depicts the layout of the FFE section of the device [38]. Micro-
uidic FFE (m-FFE) was introduced by Raymond et al. [39]. Although the peak
capacity in FFE is limited, a continuous supply of separated components is
benecial for the successful integration with anity detection. In our group, we
patented a new approach to an FFE device in combination with detection [40].
In this system, the separated proteins at the outlet of the FFE chip enter a so-
called SPR anity area where biomolecular interactions are studied at a
microarray. The binding can be followed with imaging surface plasmon
resonance detection (iSPR). Although the concepts shown here are still in their
infancy, a rst operational device with a separation and detection section is
expected to be available soon.
An improved free-ow isoelectric focusing chip has been fabricated by
Kohlheyer et al. [41] and tested with a set of uorescent isoelectric focusing
markers. For the rst time, high-resolution results could be obtained in such a
microuidic device. As illustrated in Figure 12.7, the chip contains ve inlets.
These inlets are used to infuse the separation chamber with dierent ampholyte
solutions to build up a pH gradient inside the separation chamber, perpendi-
cular to the ow.
Figure 12.8 illustrates how the pH gradient is developing with the increasing
residence time of the ampholytes and the sample in presence of the electrical
eld. The pH gradient starts with a step gradient and nally results in a linear
pH gradient, with all components focused at their isoelectric points. Due to the
use of a step pH gradient, the focusing times and, more importantly, Joule
to CCD camera
flow direction
monochromatic and
p-polarized light
total internal reflection
at the glass/gold
interface
SPR angle
circular prism
Figure 12.6 The FFESPR combination.
Figure 12.7 IEF Chip layout.
Figure 12.8 pH gradient development in a FFIEF device.
372 Chapter 12
heating are reduced. During experiments eight dierent isoelectric focusing
markers were used to visualize the separation eciency and the linearity of the
pH gradient (see Figures 12.8 and 12.9).
A new FFE chip has been fabricated, which contains a gold surface for SPR
measurements (Figure 12.10). This chip was placed inside an IBIS SPR
instrument and rst preliminary results could be obtained as shown in Figure
12.10 (left). The photograph shows the gold region inside the chip with a
centered sample stream. The surrounding water is in resonance, while the
centered sample (2-propanol) is out of resonance.
It is denitely a trend that the integration of lab-on-a-chip devices and SPR
(imaging) instruments will be developed further for new applications in the life
sciences. Broadening the options of parallelization and assay implementation,
including sample treatment on-a-chip as shown here by the FFE principle,
would certainly contribute to an increase of the range of applications. A
limitation of miniaturization is imposed by the lateral resolution of SPR
imaging, which is B10 mm and can only be improved by using other plasmonic
strategies, for instance using nanoparticle sensing (see Section 12.3 and the
discussion above). However, current spotting technologies are still an order of
magnitude (250 mm) away from these physical limitations.
12.3.5 Digital Microuidics
A new trend has been observed in the development of a microuidic chip, in
which single cells or excreted compounds from these single cells can be
diagnosed for (early) diseases in a exible and versatile way by combining
two existing platforms. Although it can be categorized among hyphenation
Figure 12.9 Experimental results with uorescent IEF markers.
techniques, we consider this combination as a new trend in the combination of
microuidics and SPR imaging. So-called digital microuidics using electro-
wetting (EW) principles can actuate cell-containing droplets and SPR imaging
detects the binding anity of the cell or excretion products from this single cell
to a variety of specic ligands. This is the topic of current research being carried
out in the groups of van den Berg and Mugele at the University of Twente.
EW is arguably the most versatile tool for the manipulation of individual
droplets in digital microuidic systems (DMS) [42]. In certain respects, EW-
based DMS is similar to other lab-on-a-chip concepts: there are several
macroscopic reservoirs on the chip, for the analyte uid, for reagents and
perhaps also products. However, in contrast to conventional lab-on-a-chip
methods which use continuous ow, in DMS all substances are handled in
discrete amounts of individual droplets, which can be detached from on-chip
reservoirs and moved along certain paths to dedicated locations, where
(bio)chemical reactions can be initiated or optical/electrical measurements
can be performed.
The basic experimental setup of EW is shown in Figure 12.11. An aqueous
droplet rests on a solid substrate with an electrode on top. Typically the
substrate is a glass slide and the electrode a transparent indium tin oxide (ITO)
layer, covered by an additional insulating layer of either SiO
2
/Si
3
N
4
or a
Teon-like polymer with thickness of the order of 1 mm.
Under this condition, the contact angle y can be reduced by several tens of
degrees by applying a voltage U across the electrodes. This follows from the
electrowetting equation [42]:
cosy cosy
Y
e
0
e
d
2ds
lv
U
2
cosy
Y
Z 12:2
Figure 12.10 Left: FFE+SPR chip (20 20 mm). Right: an image showing the SPR
gold region inside the FFE chip.
374 Chapter 12
where s
lv
is the liquidvapor interfacial tension, d the thickness of the insulator
layer, e
d
its dielectric constant and Z the dimensionless electrowetting number.
This principle allows actuation of drops via a dierent wetting by two
adjacent electrodes on the same side of the droplet [43]. The basic actuation
scheme is shown in Figure 12.11b. If only one electrode is activated, the contact
angle reduction takes place only on one side and hence the drop experiences a
net force. This is generally sucient to let droplets make a unit step, within a
2D pattern of electrodes laid out by the chip designer.
To avoid the necessity of immersing a wire, the drop is typically sandwiched
between two parallel plates (Figure 12.11c). The top plate then provides the
second electrode, which may or may not be patterned also. Apart from serving
as a second electrode, the top surface also reduces water evaporation drama-
tically, which can otherwise be a substantial problem. Droplets can also be
provided with an oil environment. This will eliminate evaporation altogether,
while the oil will also form a lm between the droplets and the substrate, which
can reduce actuation problems related to adsorption and pinning.
The use of EW to actuate droplets has been used successfully many times
already for simple liquids, typically salt solutions with concentrations ranging
from zero (deionized water) to saturation [44], with the most active groups
being based in the USA (Kims group at UCLA and Fairs group at Duke
University).
12.3.5.1 Cell Diagnosis and Monoclonal Antibody Screening
Using SPR Imaging and Digital Microuidics
In principle, digital microuidics allows for various kinds of diagnostic tests on
cell-laden droplets: optical, electrical, chemical and even mechanical testing. This
diagnosis can take place on a exible and custom-designed detector array on
chip. Here, cells can be exposed to a variety of ligands (e.g. antibody, receptor or
binding epitope), with each type of ligand being grafted to an isolated patch of
gold, patterned on the lower glass surface of the microuidic chip.
Screening hybridoma cell lines for the production of specic monoclonal
antibodies is an example of the application of the EWSPR imaging combina-
tion. In biotechnology, the development of monoclonal antibodies is a com-
plex, time consuming and thus expensive procedure involving the generation,
Figure 12.11 Sketch of electrowetting setups. (a) Basic setup with homogeneous
substrate. (b) Substrate with two independent electrodes. (c) System
with parallel plate geometry with the top electrode replacing the
immersed wire. Taken from Mugele and Baret [42].
maintenance and screening of thousands of hybridoma clones. The condent
early identication of hybridomas that produce the best candidate antibodies is
a critical step in successful, cost-ecient development. During the exposure of
the cell to a specic sensor patch, the cell will secrete monoclonal antibodies
into the droplet. While the cell itself will not be bound to the surface, the
secreted antibodies will diuse within the droplet and be measured in real time
with respect to their binding to specic ligands at the surface. The cell can then
be further processed for cultivation of the hybridoma clone, also using the
digital microuidics principle for actuation.
12.4 Trends in Sensor Surfaces
Although dimensionally extremely small, the quality of the sensor chip surface
coating has a tremendous inuence on the performance of an SPR biosensor.
SPR imaging of an area of 1 cm
2
of various high-quality sensor chips shows
defects in more than 90% of the chips, leading to potential aected sensor-
grams and artifacts. For instance, irreproducible drying eects, caused, for
example, by adsorbed air bubbles, often show cauliower images of the surface.
Further, dust particles that are always present in the air of non-sterile
environments can be irreversibly adsorbed on the surface. The imaging feature
reveals the quality of the sensor surface and inhomogeneous coatings can be
visualized. The operator of the imaging instrument is able to reject suspicious
sensor areas. The homogeneity of the nanoarchitecture of the sensor surface
can be checked with a reectivity image where the SPR angle is set in the
inection point of the left-hand ank of the SPR curve. For instance, in the
IBIS-iSPR system of IBIS Technologies either an SPR image can be measured
with improved contrast or the SPR image can be transposed further to an
articial color image. In order to avoid contamination, it is extremely im-
portant to expose the uncovered sensor chip as briey as possible to the open
atmosphere. Biacore introduced in 1990 a cassette for the sensor chip that only
is opened automatically inside the Biacore instrument.
Hydrogels with a thickness above 1 mm are useful to keep particulate
contaminants or air bubbles outside the evanescent eld, resulting in a very
robust surface. However, heavy diusion limitation is observed in such
structures with a mean diusion time of several seconds for small molecules
across the hydrogel.
12.4.1 Smart Polymer Brushes
In the past, an enormous variety of sensor coatings were used for SPR
detection. It is clearly demonstrated that the gold surface needs to be shielded
from the inuence of complex samples. The modication of surfaces with thin
polymer lms is still used to tailor surface properties such as the adsorption
behavior of ligands, wettability and biocompatibility. For example, a poly-
styrene microtiter plate for ELISA measurements was mimicked in an SPR
376 Chapter 12
set-up by spin coating or spraying of a polystyrene polymer dissolved in toluene
solution [45]. A table-top spin coater for this purpose can be obtained from, for
example, Eco Chemie (Utrecht, The Netherlands) to deposit polymers on the
sensor surface for SPR measurements. However, hydrophobic surfaces are
prone to poor wettability, ligand desorption due to physical attachment and the
adsorption of air bubbles detrimental to signal reproducibility. A hydrophilic
coating is much more reliable in SPR biosensors. The hydrogel-based carbox-
ymethylated dextran layer is the most popular matrix for SPR biosensors due
to its high coupling yields and reliability with regard to ligand immobilization.
The reaction conditions to couple proteins, peptides and small molecules to
carboxymethylated dextran surfaces are well characterized and extensive
optimization studies have been performed [46].
Figure 6.12 a smart composite hydrogel layer is shown, which consists of
an unreactive, relatively long polymer chain in low density on top of a thin
ligand layer. Such a long polymer composite layer will exclude particles and
cells which might be present in the sample. Particulate samples such as whole
blood or crude fermentation broths are thus ltered from the surface and do
not have access to the evanescent eld of the sensor surface. The synthesis of
hydrogel layers to surfaces can be performed by grafting polymers with reactive
end-groups on to surfaces, resulting in so-called polymer brushes. The advan-
tage of polymer brushes over other surface modication methods (e.g. self-
assembled monolayers) is their mechanical and chemical robustness, coupled
with a high degree of synthetic exibility towards the introduction of a variety
of functional groups. There is also increasing interest in the use of functional or
diblock copolymer brushes for smart or responsive surfaces, which can change
a physical property (hydrophilicity, biocompatibility, swelling) upon an ex-
ternal trigger, such as heat (in the case of materials with a lower critical solution
temperature), pH or salt concentration [47].
Two methods of grafting polymers to surfaces are applied to create polymer
brushes: either (1) via chemical bond formation between reactive groups on the
surface and reactive end-groups or (2) by physisorption of block copolymers
with sticky segments. This grafting to approach is experimentally simple,
but has some limitations. It is very dicult to achieve high grafting densities
because of steric crowding of reactive surface sites by already adsorbed
polymers. Relying on non-covalent adsorption of polymers to surfaces makes
the adsorption a reversible process and such brushes desorb from the sensor
surface resulting in an apparent o-rate, which has nothing to do with the
dissociation of the analyte from the ligand.
Surface-initiated polymerizations [48] (also called grafting from) from
initiators bound to surfaces are a powerful alternative to control the function-
ality, density and thickness of polymer brushes with almost molecular preci-
sion. First, the gold sensor surface should be modied with an initiator-bearing
self-assembled monolayer (for example, thiols on gold). The sensor surface is
then exposed to solutions containing catalyst and monomer (plus solvent if
necessary). Ideally, the polymerization is only surface initiated and no poly-
merization in solution takes place. In order to achieve maximum control over
brush density, polydispersity and composition, a controlled polymerization is
highly desirable. Over the last few years, this eld has evolved rapidly and many
polymerization strategies have been used to grow polymer brushes on gold [49].
The surface-initiated route to polymer brushes initiated by Fukudas group [50]
has expanded tremendously over the last 9 years. It opens up new possibilities
for creating smart or responsive surfaces and we have only seen the beginning of
research in this direction. Polymer brushes have interesting physical properties
that are primarily related to the fact that the polymers are covalently tethered to
the surface while the other end of the chain is freely moving in solution. No
doubt this will lead to new applications and to the improvement of SPR devices
regarding the controlled swelling and shrinking of polymer brushes for the
purpose of enhancing the sensitivity of specic biomolecular interactions and
for improving the limit of detection of low molecular weight analytes, which
cannot be detected reliably without the use of these smart polymer brushes.
12.4.2 Photoactivation of Surfaces for Immobilization
A trend is observed in combining lab-on-a-chip unit operations and SPR
imaging as treated in Section 12.3.2. After a chip-based separation process
integrated in the SPR instrument, ligands should be immobilized on the surface
of the sensor chip. For example, the FFESPR combination requires a method
to trap separated molecules on the surface. It is desired that an interfering
method triggers the activation of the surface, in order to allow covalent
coupling of the separated ligands on a desired spatially resolved area. Light
of a specic wavelength and intensity leads to activation of specic groups in
the hydrogel, permitting covalent coupling of ligands. Benzophenone-contain-
ing substrates with low non-specic binding properties are a rst choice [51].
The use of benzophenone as photoactivator has important advantages: benzo-
phenones are more stable than other photoactivators such as arylazides [52]
and can be manipulated in ambient light; activation wavelengths are around
350 nm, a range not damaging towards most proteins; and benzophenone can
be activated by light repeatedly without chemical degradation.
Two general approaches were pursued in our laboratories at the University
of Twente for the preparation of photoactivatable substrates (unpublished
results):
1. Chemical coupling of 4-benzoylbenzoic NHS and benzophenone-4-iso-
thiocyanate on to an aminated hydrogel.
2. Chemical coupling of neutravidin on to a carboxylated hydrogel followed
by binding of biotindPEG
3
benzophenone.
A relatively simple method for immobilizing benzophenone moieties consists of
the chemical coupling of 4-benzoylbenzoic NHS or benzophenone-4-isothio-
cyanate on to amino group-containing substrates. First substrates of choice are
coatings of branched polyethylenimine or hydrogels containing simple primary
amino groups. For example, the incubation of bare gold with a diluted (0.01%)
378 Chapter 12
aqueous solution of branched polyethylenimine (10 kDa) leads to good cover-
age of the sensor surface. The advantage is that the eciency of immobilization
by photoactivation can be followed in real time.
Another approach is to prepare, illuminate and evaluate gold sensor
modications in an IBIS imaging SPR instrument during the SPR measure-
ment, using an LED as a light source for photoactivation. Then the process of
immobilization and the eect of the photoactivation can be followed in real
time. Neutravidin has been covalently coupled on to carboxylated hydrogel
(XanTec HC 200m) and binding of biotindPEG
3
benzophenone to this
matrix was shown to be successful (unpublished results).
Figure 12.12 Schematic overview of the concentration gradient immunoassay meth-
od. (A) Device geometry (not drawn to scale). Three uid streams
converge within a single channel: one containing the analyte of interest
(sample), one containing an antibody against the analyte and (option-
ally) a reference or control stream. Following convergence, the uids
ow down a channel for some distance to permit diusive mass
transport among the uid streams. This portion of the channel surface
is functionalized with PEG to resist protein adsorption. After inter-
diusion establishes a gradient of antibodyanalyte complexes trans-
verse to ow, the uids then encounter the sensing surface that is
functionalized with surface-bound analyte. (B) Schematic view of
competition between solution-phase analyte and surface-immobilized
analogue for antibody binding sites. Diusion of solution-phase ana-
lytes into the antibody stream establishes a gradient of occupied
antibodies. Only antibodies with at least one open binding site may
bind to the surface, leading to an SPR signal change. The view is
through the channel from the outlet. (C) Cartoon of SPR dierence
image obtained after antibody accumulation to the sensing surface. The
assay shift is the dierence between the width of the center ow
stream and the region of the surface on which antibodies accumulate.
Axes indicate the view direction: +x is through the channel in the
direction of uid ow. Reproduced from Analytical Chemistry with
permission of the American Chemical Society.
12.4.3 Gradient Chemistries
Gradient chemistry strategies are attractive in combination with SPR imaging.
In SPR imaging instruments, the optimal degree of coupling can be spatially
resolved.
An example of the gradient chemistry strategy was recently published by
Yagers group (see also Chapter 10). Concentration gradient immunoassay
(CGIA) [53] is capable of the direct measurement of low molecular weight
analytes in less than 10 min with simultaneous controls on a single uid sample
(Figure 12.12).
Question: how can we coat an SPR sensor chip device in a steady gradient? In
the previous example, a diusion regime between two dierent ow regions
causes a gradient at the intersection of the two ows. Gradients can also be
created by timely contact of a ligand with the sensor chip. A microuidic device
is necessary to create the gradients in the chip which is represented schemati-
cally in Figure 12.13.
In the sensing lane area of Figure 12.13 ligands are immobilized on sensor
patches. In the non-specic binding section the analyte will bind non-specically
to the surface, while the ligand was not in contact with this section, hence ligands
are not present on the surface. In the common mode rejection (clean) section, the
analyte did not contact the surface and the area can be used for common mode
rejection compensation for bulk refractive index shifts of buer or regeneration
liquids or for temperature correction. It is a challenge to determine the eec-
tiveness of this new strategy for a certain application, however; the implementa-
tion of a microuidic device and an SPR imaging instrument is needed. Not only
dened areas of sensing lane, non-specic binding section or common-mode
rejection section, but also a gradient of contact times of ligands and analytes can
easily be created in a microuidic device.
Another approach is described in Section 12.3.2. A pH gradient can be
obtained which may be used either for gradient coupling or for gradient
denaturation of bound ligands. The technique may also be used for nding
Figure 12.13 In a microuidic device a sample can be injected slowly at point B and
reversed after a certain exposure to the sensor spots. The diusion rate
of the compounds and the contact time determine the spatial resolved
accumulation or gradient of ligands and analytes. Because the supply of
ligand and analyte in the sample is only from one side (B) to the other
compartment (C) there is a gradual contact time dierence of the
sample with the sensing areas on the left (red) and right side (yellow
to blank). Here the ligand coating solution (green antibody) is reversed
at the fourth spot. The analyte (blue stars) is reversed at the seventh
spot. The nal three spots as indicated here in this linear ow channel
will never be exposed to the ligand or analyte containing solution.
380 Chapter 12
the optimal conditions for reversible regeneration or elution properties, which
should be as mild but as complete as possible. The combination of smart
chemistries, microuidic chips and SPR imaging will gain great potential for
the search for the best interface behavior of biomolecular interactions of
ligands with their specic analyte.
12.5 Trends in Measuring Reliable Kinetic Parameters
12.5.1 Introduction
The aphorism God made the solid state, he left the surface to the devil,
attributed to Wolfgang Pauli [54] or Enrico Fermi [55], seems to apply equally to
the measurement of protein interactions in solutions versus at surfaces. Surface
binding measurements, as facilitated by SPR and illustrated in this book, provide
unique opportunities, among them small sample volumes [56], high anity
measurement with signal-to-noise ratio independent of K
D
and micropurication
for interaction discovery [57] and multi-protein binding studies [58]. However,
this comes at the price of immobilizing one binding partner to the surface. For
proteins, this in itself bears the possibility that its conformational ensemble may
be skewed or even signicantly altered, even when using chemically and/or
structurally uniform attachment strategies. Further, while in free solution all
molecules experience the same environment, the microenvironment at the surface
may be strongly variable depending on the location on the surface (e.g. from
surface roughness) or depending on the location within an inhomogeneous
matrix (e.g. from the obligate density distribution perpendicular to the surface
of grafted polymers [59,60] or from ligand gradients perpendicular or parallel to
the surface created during immobilization [61]). In solution the conformational
ensemble usually exchanges rapidly enough for the binding thermodynamics to
be well described by single values but, as a result of the immobilization and
localization to the surface, this is frequently no longer the case for the surface
sites. As a consequence, it should be expected that the surface binding energetics
can experience a dispersion and assume a continuous distribution of parameters.
In this chapter new insights will be evaluated regarding this distribution analysis
of rate and equilibrium constants.
Methods have been introduced that allow the characterization of this
functional distribution of binding parameters of the ensemble of surface sites
from experimental surface binding data [6264], and, in particular, we have
recently shown that it can be applied to the global analysis of SPR kinetic traces
[65,66]. The goal is two-fold. First, a more detailed characterization of the
distribution of binding properties should provide a useful tool in the optimiza-
tion of surface immobilization, to study immobilization processes and surface
properties in more detail, towards the ecient functionalization of biosensor
and protein chip surfaces with uniform high-anity binding sites. Second, if
one is able to identify a peak in the distribution reecting high-anity sites,
distinct from a range of other low-anity and non-specic sites impaired
in their function, this should allow one to characterize the interaction
properties more reliably as they may reect solution conditions. To some
extent, that may be possible with conventional tools by an ad hoc assumption of
two classes of discrete sites, but this articial a priori constraint to two classes of
sites may be too crude and introduce bias in the results. It is preferable to have a
more rened, assumption-free method from which the determination of
number of classes of sites may be made, if appropriate, as a nal inter-
pretative step considering the full continuous distribution. This model invokes
better insight in the biomolecular binding process than simple discrete binding
models.
That the experimental data carry enough information for the characteriza-
tion of the functional distribution of binding parameters is ensured by the
excellent signal-to-noise ratio and reproducibility of SPR binding traces.
Further, it is a common observation that the errors encountered from tting
with simple discrete binding models are not randomly distributed, but fre-
quently show highly reproducible systematic residuals [67]. This indicates that
discrete binding models may not be suciently detailed to account for the
observed processes. In some cases this may well be the result of more
complicated binding schemes (see examples in Chapters 4 and 5 and models
reviewed in [58]). However, in cases where there is no independent supporting
evidence for higher-order binding modes, a natural extension of a simple
bimolecular reaction scheme is to account for an ensemble of surface sites
with continuous distribution of binding parameters. This can frequently
provide excellent ts of data commensurate with the high signal-to-noise ratio
and reproducibility of SPR binding traces.
As described in Chapters 4 and 5, another potential diculty in the
quantitative interpretation of surface binding kinetics is mass transport limita-
tion. As a penalty from the measurement taking place at the surface, for the
surface binding kinetics to reect only the chemical reactions of interest
the soluble binding partner has to diuse suciently rapidly from bulk to the
surface, such that the probability of binding to a surface site remains a
relatively rare event and no concentration gradients of the soluble binding
partner occur. The latter will depend on the diusivity of the soluble analyte,
the time-scale of the binding reaction and the density of immobilized sites [65].
As introduced in previous chapters, if this condition is not fullled, the binding
process will be governed by mass transport limitation, where the experimental
data will only indirectly reect the molecular binding properties, and instead be
governed by macromolecular transport properties, such as diusion coecients
and transient non-specic interaction with the surface, as well as the perme-
ability of the immobilization matrix (if one is used) [66,67]. Recently, the
analysis of the distribution of surface sites was extended to include compart-
ment-like rst-order corrections for mass transport-inuenced surface binding.
For the rst time, this permits the diculties of surface heterogeneity and mass
transport, both of which frequently occur in experimental SPR data, to be
simultaneously addressed. This extends the range where the distribution
analysis can be applied and also permits a more detailed study of the origin
and behavior of mass transport limitation.
382 Chapter 12
12.5.2 The Model for Distribution Analysis of Rate
and Equilibrium Constants
With the terminology of Chapter 5, the traces for dierent analyte concentrations
[A] starting to interact at t
0
for a duration t
c
with a single class of surface sites of
maximum signal B
max
, as described in eqs. (5.4) and (5.6), can be combined as
R
t
k
off
; K
D
B
max
k
off
; K
D
k
on
A
k
on
A k
off
1 e
k
on
Ak
off
tt
0
t
0
ot t
0
t
c
1 e
k
on
Ak
off
t
c
t
0

e
k
off
tt
c
t4t
0
t
c
(
12:3
The expression R
t
(k
o
, K
D
) highlights the signal contribution arising from the
sites with equilibrium dissociation constant K
D
and dissociation rate constant
k
o
(or with equilibrium association constant K
A
1/K
D
and association rate
constant k
on
k
o
/K
D
, respectively). The binding kinetics expressed in eq.
(12.3) combines the association and dissociation phase and reects a family
of curves as shown, for example, in Figure 4.2 in Chapter 4. Depending on the
rate and equilibrium constants, the shape of the curves follows dierent
patterns and the following analysis is concerned with the inverse problem of
seeking the combination of such patterns that ts experimental data best.
For the description of a continuous distribution of surface sites B
max
(k
o
,
K
D
) with a range of chemical o-rate constants k
o
and equilibrium dissocia-
tion constants K
D
, we dene the innitesimal quantity B
max
(k
o
*,
K
D
*)dk
o
dK
D
as the population of surface sites (in signal units) with an o-
rate constant between k
o
* and k
o
*+dk
o
and the equilibrium constant
between K
D
* and K
D
*+dK
D
. (With the usual transformation k
on
k
o
/K
D
,
the distribution can be transformed to an equivalent distribution of on-rate and
equilibrium constants; see below and Figure 12.14.) The total measured signal,
R
tot,t
, can then be expressed as a Fredholm integral equation:
R
tot;t

Z
K
D;max
K
D;min
Z
k
off;max
k
off;min
R
t
k
off
; K
D
dk
off
dK
D
12:4
The computational problem consists in a global least-squares t of the integral
eq. (12.4) to experimental binding traces R
(exp)
t
([A]) acquired at dierent analyte
concentrations. This can be accomplished by discretization of the two-dimen-
sional space of binding parameters with a mesh of (k
o,i
, K
D,i
) values:
R
exp
t
A
X
N
i1
R
t;i
k
off;i
; K
D;i
12:5
and numerical optimization of B
max,i
(k
o,i
, K
D,i
), which is the discrete repre-
sentation of the distribution B
max
(k
o
, K
D
) [65].
In case of mass transport-limited binding, the distribution analysis model can
be extended [66]. Briey, it is approximated on a discretized mesh of binding
parameters B
max,i
(k
o,i
, K
D,i
). However, in this case the binding traces for a
single class of sites do not follow eq. (12.3). As outlined in Chapter 5, mass
transport-limited binding can be approximately described by a two-compart-
ment model that introduces, as a rst-order approximation of spatial inhomo-
geneity, the distinction of analyte at the surface [A] and in the bulk [A]
0
.
Equation (5.7) can be generalized to the case of many sites with binding
properties (k
o,i
, K
D,i
), which results in the following dierential equation for
the signal arising from each site:
dR
t;i
dt
k
on;i
A B
max;i
R
t;i

k
off;i
R
t;i
for all i
dA
dt
k
tr
A
0
A

X
N
j1
dR
t;j
dt
12:6
with the abbreviations R
t,i
and B
max,i
denoting R
t
(k
o,i
, K
D,i
) and B
max,i
(k
o,i
, K
D,i
), respectively, and k
on,i
k
o,i
/K
D
. After a fast initial transition
period after application of [A]
0
, the surface concentration [A] assumes an
Figure 12.14 Binding of soluble microglobulin to a monoclonal IgG immobilized on
a carboxymethylated dextran surface (F1 chip) [66]. (A) Association
and dissociation traces at an analyte concentration of 1 nM. (B)
Contour plot of the anity and rate constant distribution calculated
from the 1 nM binding data of (A), on a grid of K
D
and k
o
value as
indicated by the small circles. In this presentation, lines of constant k
on
values are diagonals and logk
on
values can be read as logk
o
logK
D
.
(C) Association and dissociation traces at a concentration of 100 nM.
(D) Surface site distribution calculated from the 100 nM trace alone.
384 Chapter 12
approximately constant value (steady-state conditions), which leads to
dR
t;i
dt
k
on;i
A
0
R
max;i
R
t;i

k
off;i
R
t;i
k
on;i
R
max;i
R
t;i

k
tr
X
N
j1
dR
t;j
dt
12:7
This is a system of rate equations coupled through the third term, which is the
one describing the mass transport inuence. The coupling reects the physical
possibility that an analyte dissociating from one class of sites may rebind to
sites from another class. Again, optimization of the B
max,i
(k
o,i
, K
D,i
) values in a
least-squares t of experimental data [eq. (12.5)] provides a discretized repre-
sentation of the continuous distribution B
max
(k
o
, K
D
).
In order to make this distribution analysis computationally feasible, it is
combined with maximum entropy [68] or Tikhonov regularization. This is a
technique for stabilizing ill-conditioned or underdetermined data inversion
problems and well known in many areas of physical data analysis [69].
Importantly, it acts to suppress detail in the distribution that is not statistically
warranted by the information content of the data. This will lead to the most
parsimonious, or broadest, distribution consistent with the data. The software
(termed EVILFIT in our laboratory), is implemented on the MATLAB plat-
form and is freely available from the authors. A simplied graphical user
interface is anticipated for the future.
12.5.3 Examples of the Distribution Analysis Method
When inspecting the results from the distribution analysis method, it is
important to be aware of the eects of regularization providing the simplest
distribution of all that may be consistent with the data, following Occams
razor. For example, by virtue of the regularization, it is possible to solve
underdetermined problems, such as to obtain full two-dimensional distribu-
tions of anity constants and kinetic rate constants from association and
dissociation curves at a single concentration.
This is illustrated in Figure 12.14, which in panels B and D depict the
distributions in units of logk
o
and logK
D
[in this presentation, lines of constant
k
on
are diagonals (since logk
on
logk
o
logK
D
) and the volume under the
peaks gives the total binding capacity of sites within the given parameter range].
Panel A shows the association and dissociation traces of 1 nM antigen binding
to an immobilized antibody and panel B shows the distribution obtained from
this single trace. As can be expected on the basis of the limited information
carried in these data, the distribution has only very broad features. It is
conservative in a sense that it suggests order of magnitude of K
D
and k
o
values, rather than attempting to provide single values. This aspect is high-
lighted by comparison with the distribution obtained at a single analyte
concentration of 100 nM (Figure 12.14C and D, respectively). Since higher
concentrations provide more information, which can be discerned here from
the biphasic shape of the association trace, the distribution has more detailed
features: it displays a narrower main peak for the high-anity sites, alongside
some minor peaks for sites with lower anity.
This example demonstrates how the level of detail in the distributions is
automatically adjusted according to the information carried in the experimen-
tal data, such as to provide the most conservative interpretation. Importantly,
the width of the distribution should not be interpreted in a sense that it
necessarily reects the exact distribution of surface sites, but that it reects
the most detailed statement that can be made about the surface sites given the
data. For example, the single 1 nM trace may well be modeled by a single
discrete 1:1 binding model, providing apparently unambiguous unique values
of binding and rate constants. However, whether or not this number would
reect the true binding parameters cannot be decided on the basis of the single
curve. Indeed, as the additional data at the higher concentration show even
from visual inspection of the biphasic shape, there is heterogeneity of the
surface sites and the discrete single-site model to the 1 nM data would have
been misleading. Similarly, there may still be ner structure in the true
distribution of surface sites than displayed in Figure 12.14D, but the given
nite signal-to-noise ratio does not permit closer characterization.
Although the analysis of single traces can have important applications, for
example, for the study of binding sites prior to the exposure to chemical
regeneration, obviously the goal of the analysis and the mode in which the
distribution analysis is usually applied is the global analysis of all traces at all
concentrations or even global analysis of dierent ow rates. The relationship
between information content of the data and resolution of the resulting
distribution is the same for global analyses. For this example, such a global
analysis including curves at more analyte concentrations has been carried out
[66], showing similar features to Figure 12.14D.
The question of whether the observed width of the distribution arises from
true microheterogeneity of the surface sites or from the nite signal-to-noise
ratio in the data is addressed in the next example (Figure 12.15) [58]. Although
a single major peak is obtained in the distribution (Figure 12.15C), some degree
of heterogeneity may be discerned. The polydispersity of the binding sites is
supported by a comparison of the quality of t of the distribution model
(residuals in red in Figure 12.15B) with the best-t single-site model (residuals
in blue in Figure 12.16B and dotted line in Figure 12.15A).
The opportunity to study the behavior of binding sites in their dependence on
the surface employed for immobilization is highlighted in Figure 12.16. This
shows the same antibodyantigen interaction as in Figure 12.14, but immobi-
lized on a long-chain carboxymethyldextran surface (Biacore CM5). On this
surface, the binding is mass transport limited and the global analysis of binding
traces at 0.1, 1, 10 and 100 nM with the model eq. (12.7) results in a distribution
that exhibits a tail of low-anity sites that amounts to 425% of all surface
sites. Such a broad population of low-anity sites was not observed in the
analogously conducted experiment using the short-chain carboxymethyldex-
tran surface [66]. Interestingly, the possible role of non-specic binding causing
ow rate-independent contributions to mass transport limitation is supported
386 Chapter 12
by theoretical expectation of transient matrix interactions slowing the eective
diusion time of analyte through the matrix [66,69].
Finally, the surface site distribution can be applied to the study of immobi-
lized biomolecules that exhibit naturally multiple classes of sites. This is
highlighted, for example, in the study by Vorup-Jensen et al. [70] on acidic
residues of brinogen exposed during tissue decay and their role as a pattern
Figure 12.15 Surface site distributions obtained from an antigen interacting with the
antibody immobilized on a long-chain carboxymethyldextran surface
(Biacore CM5) [58]. (A) Experimental association and dissociation
traces (black) at three dierent concentrations. The blue dotted line is
the best-t model based on a single discrete 1:1 interaction. (B)
Residuals of the t with a single discrete 1:1 interaction model (blue)
and with the continuous distribution model (red). (C) K
D
k
o
distribu-
tion from the data in (A). For details, see [58].
recognition motif in the recognition by integrin a
X
b
2
(Figure 12.17). Vorup-
Jensen et al. concluded that the increased anity for the recognition of
damaged, i.e. partially unfolded or proteolyzed, brinogen plays an impor-
tant role in tissue repair as a danger signal enabling enhanced recognition by
leukocyte cell-surface receptors [70].
12.5.4 Conclusions and Perspectives of the Distribution
Analysis Model
The binding model for a continuous distribution of sites is a natural extension
of the 1:1 discrete binding site model that takes into account the heterogeneity
of binding site properties commonly arising from heterogeneity in the surface
microenvironment (surface roughness, charge and density distribution from
polymer matrices) and from surface immobilization (steric, spatial and/or
chemical heterogeneity of the attachment). At the same time, the model can
account for low to moderate degrees of mass transport limited binding. This
addresses the two most commonly observed complications in the use of SPR to
study interactions of biological macromolecules [67].
Interestingly, heterogeneity in the spatial dimension parallel to the surface
(as opposed to the functional heterogeneity) was addressed in Chapter 7, where
it was shown how SPR imaging enables one to obtain a microscopic inspection
Figure 12.16 Binding of soluble microglobulin to a monoclonal IgG immobilized to a
long-chain carboxymethylated dextran surface (CM5 chip) [66]. Shown
is a plot of the anity and rate constant distribution calculated from the
global analysis of traces obtained at analyte concentrations of 0.1, 1, 10
and 100 nM.
388 Chapter 12
of the spatial heterogeneity of the surface before and after a biomolecular
interaction process. It has often been observed that an initial homogeneous
spot shows heterogeneity after the binding process, implying that a distribution
of a kinetic process takes place over the surface caused by clustering of
biomolecules or by an initial heterogeneous distribution of binding sites. In
this way, the functional and spatial heterogeneity may be linked. In the case of
Figure 12.17 Interaction between brinogen and integrin a
X
I domain. The binding
of soluble integrin a
X
I domain to native, proteolyzed or guanidine-
treated immobilized brinogen was probed by applying 10 a
X
I con-
centrations from 0.28 to 10.6 mM (A). For comparison, sensorgrams in
(B) show the injections of the I domains at the highest applied
concentration of 10.6 mM over the surfaces with native, proteolyzed
or guanidine-treated brinogen (the end of injection phase is indicated
with arrows). Two-dimensional o-rate constant and anity distribu-
tions show a relatively homogeneous ensemble for native (C), but more
heterogeneous ensembles for denatured (D) and plasmin-treated (E)
brinogen. Reproduced with permission from reference 70.
such an observation, it might be advantageous to apply the distribution
analysis method to describe and verify the biomolecular interaction process.
With the goal of the SPR experiment to characterize the binding anity and
rate constants, the model is parsimonious in the assumptions and the regular-
ization ensures conservative data interpretation, automatically adjusting to the
information content of the experimental traces. In addition to a range of
binding constants, signal-average single numbers for the anity constant and
for the chemical rate constants may be determined by integration of the peaks.
The model naturally displays multiple classes of sites, without the need for
postulation of their number a priori, if their detection is statistically warranted
by the data. Even if some of the sites are not of interest, such as low-anity or
non-specic sites considered surface-related artifacts, identifying them and
accounting for their signal contributions is particularly important in order not
to bias the characterization of the high-anity sites of interest.
In recent years, we have routinely applied this approach, for example, to
the study of antibodyantigen interactions [71,72] and have observed that the
distribution model in most cases provides a t of the data stringently within
the noise of data acquisition (such as in the examples with systems drawn from
the Biacore startup kit shown in the present chapter and in Chapter 6, where
simple discrete 1:1 models would fail and the extension to more complex
binding schemes seems inappropriate since biologically not supported). This
supports the view that the surface site heterogeneity model captures a true
feature of analyte binding to surface-immobilized sites, which SPR binding
data commonly have a sucient signal-to-noise ratio to display readily.
A second area of application for the distribution model is the optimization of
sensor surfaces and surface immobilization. This is a very important and active
area of research (see, e.g., Chapter 6). Although the best sensor surface to
permit uniformly active protein attachment and to exhibit low non-specic
binding will certainly depend on the nature of the proteins and analytes
involved, the model for distribution of kinetic and anity parameters currently
provides the most detailed tool for the functional characterization of the
ensemble of surface sites. As has been shown [66], the combined mass
transport/surface heterogeneity model can also provide more detailed insights
into the physical nature of the transport process and help to understand how
dierent sensor surface properties may impact the relationship between the
measured surface binding kinetics and the intrinsic chemical kinetics of the
interacting macromolecules.
Similarly, SPR imaging of (protein) microarrays with multiple spots of the
same ligand immobilized at dierent densities is a promising approach, where
the global analysis will become important. In this case, also, the combination of
multi-spot global analysis with the distribution method may lead to a more
detailed description of the ligand functional binding properties. We expect that
with the introduction of reliable SPR imaging instruments with hundreds of
spots, kinetic o-rate screening of samples [73] will be extended to screening of
the anity constant of many analytes to ligands (multi-kinetics) in only a single
injection.
390 Chapter 12
12.6 Final Comments
In this chapter, trends in SPR technology regarding instrumentation, uidics,
sensor surfaces and kinetic analysis have been described. It is speculative to lay
down the likely degree of impact in the next few years regarding these trends in
SPR technology. However, considering that biomolecular recognition and the
quantitation of molecular interactions have become central in the study of
structure and function of proteins and biological pathways, in molecular
medicine and in biotechnology and pharmaceutical development, and consid-
ering SPR as being one of the most mature, widespread and direct detection
principle, it seems certain to continue to undergo rapid development and
expansion of applications. We hope that this chapter and indeed the entire
book will have persuaded the reader to share this enthusiasm for SPR and
surface-related research and technologies. If the surface is the domain of the
devil, as suggested by the well known aphorism cited above, it seems tting that
it should provide us with many highly exciting new research avenues and
opportunities to shed light on biological processes.
12.7 Questions
1. The implementation of a lab-on-a-chip device with integrated gold surface
in an SPR instrument is not easy. What will happen with the channel
surface and the gold surface if we transport analyte directly from diluted
serum to the gold surface using electroosmotic ow (EOF) as pumping
mechanism? Describe the eects.
2. What happens when an analyte in serum with low isoelectric point is not
pumped by pressure from a sample loop but transported using the EOF
principle?
3. The main analytical problem of coupling SPR and MS is that the
detection limits of both techniques should be matched. What are the
discrepancies and explain why SPR has in principle a lower detection limit
than MALDI-MS. Does this also apply to small molecules?
4. The application of PDMS devices is benecial to improve mass transport
to the surface. Calculate the increase in the ow rate if we replace a
standard ow cell of 1 cm1 cm200 mm (LWD) with a PDMS
device with channels of 5 mm200 mm25 mm (LWD). What is the
factor of sample reduction if we apply an equal ow rate in these ow
cells? What is the disadvantage of applying a PDMS device for a spotted
sensor surface?
5. A new trend is given in this chapter in kinetic evaluation regarding the
simple monophasic interaction model by applying distribution analysis of
the rate constants. Microarrays with dierent ligands and ligand surface
coverages show dierent binding kinetics, e.g. changed mass transport
limitation conditions. Why is the distribution analysis model of interest
for microarrays and can be regarded as a new trend?
6. Spotting and immobilizing ligands on sensor chips can be carried out
using microuidic devices. Explain how via a self-assembly process
thousands of protein ligands can be immobilized in an ultra-high-density
microarray with squares of 25 25 mm.
References
1. T. Masadome, Y. Asano, T. Imato, S. Ohkubo, T. Tobita, H. Tabei,
Y. Iwasaki, O. Niwa and Y. Fushinuki, Anal. Bioanal. Chem., 2002, 373, 222.
2. C. Thirstrup, W. Zong, M. Borre, H. Ne, H.C. Pedersen and
G. Holzhueter, Sens. Actuators B, 2004, 100, 298308.
3. J.R. Krone, R.W. Nelson, D. Dogruel and P. Williams, P., R. Granzow,
Anal. Biochem., 1997, 244, 124132.
4. C.P. Sonksen, E. Nordho, O. Jansson, M. Malmqvist and P. Roepstor,
Anal. Chem., 1998, 70, 2731.
5. T. Natsume, H. Nakayama, O. Jansson, T. Isobe, K. Takio and
K. Mikoshiba, Anal. Chem., 2000, 72, 4193.
6. J. Grote, N. Dankbar, E. Gedig and S. Koenig, Anal. Chem., 2005, 77,
1157.
7. E. Bouartigues, H. Leh, M. Anger-Leroy, S. Rimsky and M. Buckle,
Nucleic Acids Res., 2007, 35, e39.
8. N.F.C. Visser, A. Scholten, R.H.H. van den Heuvel and A.J.R. Heck,
ChemBioChem, 2007, 8, 298305.
9. J.J. Gilligan, P. Schuck and A. Yergey, Anal. Chem., 2002, 74, 2041.
10. M. Abrantes, M.T. Magone, L.F. Boyd and P. Schuck, Anal. Chem., 2001,
73, 2828.
11. N.R. Gonzales, P. Schuck, J. Schlom and S.V. Kashmiri, J. Immunol.
Methods, 2002, 268, 197210.
12. E. Melles, T. Bergman, M. Sta hlberg, C. Thirstrup, J. Wahren, H. Jo rnvall
and J. Shafqat, J. Biomol. Tech., 2005, 16, 392.
13. J.W. Attridge, P.B. Daniels, J.K. Deacon, G.A. Robinson and G.P.
Davidson, Biosens. Bioelectron., 1991, 6, 201214.
14. K.M. Shakeshe, X. Chen, M.C. Davies, A. Domb, C.J. Roberts, S.J.B.
Tendler and P.M. Williams, Langmuir, 1995, 11, 39213927.
15. J. Davies, C.J. Roberts, A.C. Dawkes, J. Sefton, J.C. Edwards,
T.O. Glasbey, A.G. Haymes and P.M. Williams, Langmuir, 1994, 10,
26542661.
16. A. Zangwill, Physics at Surfaces, Cambridge University Press, Cambridge,
1988, p. 186.
17. R.P. van Duyne, Science, 2004, 306, 985986.
18. C.L. Haynes and R.P. Van Duyne, J. Phys. Chem. B, 2001, 105, 5599.
19. A.J. Haes and R.P. Van Duyne, J. Am. Chem. Soc., 2002, 124,
1059610604.
20. X. Hong and F.J. Kao, Appl. Opt., 2004, 43, 28682873.
392 Chapter 12
21. J. Daughton, IEEE Trans. Magn., 1994, 30, 4608.
22. G. MacBeath and S.L. Schreiber, Science, 2000, 289, 17601763.
23. V. Kanda, J.K. Kariuki, J.D. Harrison and M.T. McDermott, Anal.
Chem., 2004, 76, 72577262.
24. P. Connoly, Biosens. Bioelectron., 1995, 10, 16.
25. M.P. Hudson, R.H. Christenson, L.K. Newby, A.L. Kaplan and E.M.
Ohman, Clin. Chim. Acta, 1999, 284, 223237.
26. B.J. Tortella, R.F. Lavery, J.V. Doran and J.H. Siegel, Am. J. Clin. Pathol.,
1996, 106, 124127.
27. J.H. Nichols, Blood Gas News, 1999, 8, 414.
28. A. Manz, N. Graber and H.M. Widmer, Sens. Actuators B, 1990, 1, 244
248.
29. P.S. Dittrich, K. Tachikawa and A. Manz, Anal. Chem., 2006, 78, 3887.
30. P.S. Dittrich and A. Manz, Nature, 2006, 5, 210218.
31. J.W. Jorgenson and K.D. Lukacs, J. High Resolut. Chromatogr., 1981, 4,
230.
32. R.B.M. Schasfoort, S. Schlautmann, J. Hendrikse and A. van den Berg,
Science, 1999, 286, 942945.
33. M.A. Hayes, I. Khetarpal and A.G. Ewing, Anal. Chem., 1993, 65, 27.
34. M.J. Jory, G.W. Bradberry, P.S. Cann and J.R. Sambles, Sensors and
Actuators B, 1996, 35, 197201.
35. V. Lioubimov, A. Kolomenskii, A. Mershin, D.V. Nanopoulos and H.A.
Schuessler, Applied Optics, 2004, 43, 3426.
36. A. Tadjeddine, Electrochim. Acta, 1989, 34, 2933.
37. R.G.H. Lammertink, S. Schlautmann, G.A.J. Besselink and R.B.M.
Schasfoort, Anal. Chem., 2004, 76, 11.
38. D. Kohlheyer, G. Besselink, S. Schlautmann and R.B.M. Schasfoort, Lab
Chip, 2006, 6, 374380.
39. D.E. Raymond, A. Manz and H.M. Widmer, Anal. Chem., 1994, 66, 2858
2865.
40. D. Kohlheyer and R.B.M. Schasfoort, Eur. Pat. Appl. EP 05076569.2,
2005.
41. D. Kohlheyer, J.C.T. Eijkel, S. Schlautmann, A. van den Berg and R.B.M.
Schasfoort, Anal. Chem., 2007, 79, 81908198.
42. F. Mugele and J.-C. Baret, J. Phys.: Condens. Matter, 2005, 17, 705.
43. M.K. Chaudhury and G.M. Whitesides, Science, 1992, 256, 1539.
44. M.G. Pollack, A.D. Shenderov and R.B. Fari, Lab Chip, 2002, 2, 96.
45. W. Qian, D. Yao, F. Yu, B. Xu, R. Zhou, X. Bao and Z. Lu, Clin. Chem.,
2000, 46, 14561463.
46. B. Johnsson, S. Lo fas and G. Lindquist, Anal. Biochem., 1991, 198, 268
277.
47. B. Zhao and W.J. Brittain, Prog. Polym. Sci., 2000, 25, 677710.
48. S. Edmondson, V.L. Osborne and W.T.S. Huck, Chem. Soc. Rev., 2004, 33,
1422.
49. D.M. Jones, J.R. Smith, W.T.S. Huck and C. Alexander, Adv. Mater.,
2002, 14, 11301134.
50. M. Ejaz, S. Yamamoto, K. Ohno, Y. Tsujii and T. Fukuda, Macromole-
cules, 1998, 31, 59345936.
51. G. Dorman and G.D. Prestwich, Trends Biotechnol., 2000, 18, 6477.
52. S.A. McMahan and R.R. Burgess, Biochemistry, 1994, 33, 12092.
53. K.E. Nelson, J.O. Foley and P. Yager, Anal. Chem., 2007, 79, 35423548.
54. G. Binnig and H. Roehrer, Sci. Am., 1985, 253, 5056.
55. F.S. Ligler and T. Cass, Immobilized Biomolecules in Analysis, a Practical
Approach, Oxford University Press, Oxford, 1998.
56. M. Abrantes, M.T. Magone, L.F. Boyd and P. Schuck, Anal. Chem., 2001,
73, 28282835.
57. J.J. Gilligan, P. Schuck and A.L. Yergey, Anal. Chem., 2002, 74, 2041.
58. E.J. Sundberg, P.S. Andersen, I.I. Gorshkova, P. Schuck, in Protein
Interactions, Biophysical Approaches for the Study of Complex Reversible
Systems, P. Schuck (Ed.), Springer, New York, 2007, Protein Reviews, vol.
5, 97142 (Review).
59. P.G. de Gennes, Macromolecules, 1980, 13, 10691075.
60. A. Chakrabarti and R. Toral, Phys. Rev. B, 1990, 42, 2445.
61. T. Zacher and E. Wischerho, Langmuir, 2002, 18, 17481759.
62. S. Haber-Pohlmeier and A. Pohlmeier, J. Colloid Interface Sci., 1997, 188,
377386.
63. A.M. Puziy, Langmuir, 1999, 15, 60166025.
64. V.M. Gunko, R. Leboda, V.V. Turov, F. Villieras, J. Skubiszewska-Zieba,
S. Chodorowski and M. Marciniak, J. Colloid Interface Sci., 2001, 238,
340356.
65. J. Svitel, A. Balbo, R.A. Mariuzza, N.R. Gonzales and P. Schuck, Biophys.
J., 2003, 84, 40624077.
66. J. Svitel, H. Boukari, D. Van Ryk, R.C. Willson and P. Schuck, Biophys.
J., 2007, 92, 17421758.
68. U. Amato and W. Hughes, Inverse Problems, 1991, 7, 793808.
69. P. Schuck, Curr. Opin. Biotechnol., 1997, 8, 498502.
70. T. Vorup-Jensen, C.V. Carman, M. Shimaoka, P. Schuck, J. Svitel and
T.A. Springer, Proc. Natl. Acad. Sci. USA, 2005, 102, 16141619.
71. Z. Chen, P. Earl, J. Americo, I. Damon, S.K. Smith, Y.H. Zhou, F. Yu,
A. Sebrell, S. Emerson, G. Cohen, R.J. Eisenberg, J. Svitel, P. Schuck, W.
Sattereld, B. Moss and R. Purcell, Proc. Natl. Acad. Sci. USA, 2006, 103,
1882.
72. Z. Chen, M. Moayeri, Y.H. Zhou, S. Leppla, S. Emerson, A. Sebrell,
F. Yu, J. Svitel, P. Schuck, M. St. Claire and R. Purcell, J. Infect. Dis.,
2006, 193, 625.
73. M. Steukers, J.M. Schaus, R. van Gool, A. Hoyoux, P. Richalet,
D.J. Sexton, A.E. Nixon and M. Vanhove, J. Immunol. Methods, 2006,
310, 126135.
394 Chapter 12

Subject Index
Note: Page numbers in italics refer to figures or tables.
absorbance, enhancement of 289
adhesion linking layers 173, 174, 1813
adsorbate 82
adsorption
modes 1025
multilayer 105, 11315
non-specific 1745, 1835
role in life processes 82
adsorption kinetics 81122, 2826, 3802
basics and terminology 816
distribution analysis model 38390
mass transfer 908, 2845
solute transport and interaction 8790
mechanisms 98-9
effect of competing reactions 1012
idealized partition process 99101
surface functions 1025, 105
multilayer growth 105, 11315
saturation 10412, 105
optical quantification 867
reaction rate 8790
reaction- vs transfer-limited rate 8790
affinity 8
amine coupling
reductive amination 2023
via reactive esters 199202
aminopropyltriethoxysilane 183
analysis cycle 5, 2214
analyte
calibration curves 67
correction for depletion 1289
defined 4
small molecules 7, 193, 277, 3247
angle scanning instruments 401, 5960
microarray analysis 2302
miniaturization 315
angle shift 36, 37
units 38
angle tuning 315
antibiotics 340, 342, 346, 347
antibodies
microarrays see immunoassays
immobilization of 213
Salmonella 350
AOAC certification 3445
APTES 183
assay 36, 2268
inhibition 2278, 336, 337
surface competition 337, 338
see also immunoassays
association 5, 6
constants 78
and mass transport limitation 1256,
2845
atomic force microscopy 360
attenuated total reflection 20
autoimmune diseases see
immunoassays

baseline 45, 5, 6, 2245
-agonists 3423, 347
-lactams 3478
396 Subject Index
Biacore instruments 11, 12, 53, 6978
A100 70, 71, 735
basic systems reviewed 6970
FLEXChip 61, 69, 70, 757
Q sensor 334, 3389
T100 703
binding
competing ligand 13741
intermolecular bivalent 14752
models 1315, 14352, 1524
rebinding 13741, 226
sites 824
thermodynamics 15461
bioaffinity 8
bioinert/immobilization matrices 173,
174, 181
bioinert hydrogels 1857
non-specific adsorption 1835
biomolecular interactions 1747
kinetics/thermodynamics see adsorp-
tion kinetics; ligandreceptor
interactions
biosensors
definition 9
historical development 912
see also sensor chip
biotin 344
biotinylation 21213, 2323
bivalent binding 14752
bovine milk proteins 349
Brewster angle 18
buffer solution 45, 2235
and electrostatic interaction 1767
bulk effect 222

calibration curve 67
calorimetry 1613
capture arrays 363
carbodiimides 199200
see also EDC
carboxymethylated dextran see under
hydrogels
cell diagnosis 3756
cephalosporins 347
ceramic substrates 1823
Charm tests 348
chip see sensor chip
chloramphenicol 340, 342, 346
chromophores 278, 2812, 359
CLAMP software 1334, 147
clenbuterol 342, 347
clinical diagnostics 31332, 3556
advantages of SPR 31314
bulk refractive index compensation
3234
immunoassay
complex samples 3279
concentration gradient 27980
disposable card 328, 329
small molecules 3247
instrument miniaturization
optical system 316
wavelength and angle tuning 315
optimizing imager performance 31721
temperature fluctuations 3213
see also lab-on-a-chip; microfluidic
devices
coherence length 234
color multiplexing 3068
competition reactions 1012
competitive assay 2267
complex samples 3279
concentration gradient 37980
conformational change 1447, 184
convection 178
counter-ion evaporation 185, 196
coupling reactions 20312
cuvette-based instruments 11, 489, 612
correction for analyte depletion 1289
mass transfer process 901, 92, 95,
143, 178
Cytop 251, 253, 254

desorption 84
dextran brush 277, 298
dextran hydrogels see under hydrogels
diagnostic card 328
diffraction grating 201
instrumentation 412, 301
diffusion and diffusion layer 17881, 192
mass transfer kinetics 945
Subject Index 397
digital microfluidics 3736
dip 24, 223
fluorescence and absorbance
enhancement 29
presentation in sensorgrams 368
silver vs gold layer 27
direct assay 226, 227
direct detection 4, 174
directed immobilization 21213
dispersion relation 1719, 20
disposable diagnostic card 328, 329
dissociation 5, 6
constants 78
and mass transfer limitation 1256
rebinding model 1379
distribution analysis 180, 181, 3839
dithiothreitol 204
DNA
amplified detection
gold nanoparticles 2602
microRNA detection 2649
RNase H 2578
coding technology 3646
immobilization 211
microarrays see microarray imaging
drug development 1234, 1657
dry immobilization 1979
dynamic range of scans 2302

EDCNHS chemistry 125, 175, 192,
199200
coupling procedure 201
hydrazide activation 2078
EDC/(sulfo) coupling 2001
electrochemical SPR 369
electroosmotic microfluidic devices 3679
electrostatic immobilization 210, 211
electrostatic interactions 1757, 185
electrostatic preconcentration 1957, 196
electrowetting 3746
entropic stabilization 185, 186
enzymatic enhancement 247, 254, 2645
epoxy-mediated coupling 20810
equilibrium constants 78, 84
model for distribution analysis 3835
equilibrium signal 1268, 286
1-ethyl-3-(3-dimethylaminopropyl)-
carbodiimide 199-200, see also
EDCNHS
evanescent field 3, 17, 86, 1778, 191
and hydrogel thickness 186, 192
and optical requirements 41
evanescent wave 1617
EVILFIT software 385
excitation 1921
Eyring plots 1635

fan-shaped beam instruments 39, 504, 55
field enhancement 212
field of view 3201
filter layers 193, 194
fire blight 342
fixed angle instruments 41, 549
and microarray analysis 22830
FLEXchip 61, 69, 70, 757, 355
flow cells 458
mass transfer process 908, 143
microfluidic see microfluidic devices
flow rates 947, 178
fluidics see microfluidic devices
fluorescence 234, 362
enhancement of 289
fluorescence spectroscopy 2758, 309,
35960
color multiplexing 3068
fluorescence imaging 3068
grating coupling 3013
kinetic studies 286
hybridization of oligonucleotides
28698
protein binding 298300
Kretschmann configuration 27881,
3001, 307
long-range surface plasmons 3036
protein binding 298300
wavelength-resolved spectra 299300
folic acid 344
food analysis 3334
assay formats/steps 3348
genetically modified organisms 3501
honey 342, 343, 3456
398 Subject Index
instrumentation 334, 335, 3389
milk 34750
Qflex kits 334, 340
AOAC certification 3445
Salmonella antibodies 350
veterinary drug residues 333, 340
antibiotics 342, 346, 347
-agonists 3423
sulfonamides 339, 340, 3412, 346
Tylosin 343
vitamins 340, 3434
foulbrood 342, 346
Frankvan der Merwe growth 113, 114
free flow electrophoresis 36973
Fresnel equation 278

gauche defects 182
genetically modified organisms 3501
genotyping 247, 2624
glass substrates 1823
global kinetic analysis 1524
bimolecular models 1315
complex binding models 14352
glycidylpropyltrimethoxysilane 183
glycoproteins 2078
gold layer 267, 44
adhesion linking layers 1812
and coherence length 234
and electroosmotic flow devices 3679
sputtered surface 190, 191
gold nanoparticles see nanoparticles
GPTMS 183
gradient chemistries 37980
grating coupler 21
fluorescence spectroscopy 301
instrumentation 412
growth promoters 333, 343

His
6
-tagged ligands 211
honey 342, 343, 3456
hybridization reactions 28697
hydrazide activation/coupling 2078
hydrodynamic addressing 478, 73, 362
hydrogels
dextran 10, 12, 175, 187, 1913
epoxy activation 20811
for different sensor applications 188
film density 1923
film thickness 186, 192
filter layers 193, 194
polymer brushes 277, 298, 3768
protein-compatible polymers/
functionalities 186, 187
surface chemistry 1857
synthetic polycarboxylates 192
three-dimensional nanoarchitecture
1914, 277
two-dimensional surfaces 18991,
2767
hydrophilic interactions 181
protein-compatible polymers 187
hydrophobic interactions 175, 176, 177,
184
hyphenation technology 3567, 35960
mass spectrometry 3579

image averaging 318
imaging instruments 43, 639, 356
IBIS iSPR for multiplex analysis
22834
see also microarray imaging
immobilization see ligand
immobilization
immobilization matrices see bioinert
matrices
immunoassays 222, 2258
analysis cycle 2224, 22832
assay formats 2258
buffer solutions 2245
dynamic range/reliability 2301
experimental setup 22830
limit of detection 2324
monoclonal antibody screening
3756
non-milk proteins 3489
serum antibodies of rheumatoid
arthritis 23542
SPR and mass spectrometry 3579
see also clinical diagnostics;
microfluidics
inhibition assays 2278, 336, 337
inorganic dielectrics 1823
Subject Index 399
instrumentation 3580
basic principles 89, 358
future trends 3546
general optical requirements 445
history of 1112
imaging instruments 43
liquid handling systems 45
cuvettes 489, 232
flow cells 458, 229, 232
miniaturization 31417
optical systems 389
angle scanning 401, 315
fan-shaped beam 39
fixed angle 41
grating coupler 21, 412
interferometers 43
resonant mirror 42, 62
wavelength interrogation 42
instruments reviewed 50, 51, 52
angle-scanning
ESPRIT (Eco Chemie) 40, 5960
SPRINGLE (Eco Chemie) 40, 5960
Biacore instruments see Biacore
fan-shaped beam
BI-SPR (Biosensing Instruments)
50, 54
DKK-TOA SPR-20 (Tacadanobaba)
52, 55
Plasmonic Biosensor (Wallenfels)
53, 55
SensiQ (Nomadics) 52, 54
SR7000 DC (Reichert) 52, 55
fixed-angle
-SPR (Sensia) 56, 57
BIOSUPLAR-321 (Sinzing) 56, 57
K-MAC SPRi/SPR
LAB
(Korea
Materials) 578
Moritex (Myutron) 54, 56
Multiskcop of Optrel GBR
(Kleinmachnow) 56
Nanofilm EP
3
(Gttingen) 589
Resonant Probes SPTM (Goslar) 54
imaging instruments
GenOptics 64
IBIS iSPR (IBIS Technologies)
678, 69, 22830
LFIRE (Maven Biotechnologies)
64, 66
MultiSPRinter (Toyobo) 64, 66, 67
Plasmon Imager (Graffinity
Pharmaceuticals) 67, 68
Proteomic Processor (Lumera)
65, 66
ProteOn XPR36 (BioRad
Laboratories) 64, 66
SPRi-Lab+ (GenOptics) 64, 65
SPRi-Plex (GenOptics) 64, 65
SPRimager II (GWC Technologies)
63
other systems
IAsys Neosensors (Sedgefield) 612
SPR 100 module (Thermo Electron
Corp.) 601
integrated microfluidic cartridge 465
interaction arrays 363
interference filter 315
interferometers 43
intermolecular bivalent binding 14752
ionic immobilisation 210, 211

k
obs
kinetic analysis 1301
kinetics
determination of parameters 56, 78
instrumentation for 43, 45
kinetic models see adsorption kinetics;
ligandreceptor interactions
surface enzyme 2547
KramersKronig relation 29
Kretschmann configuration 2, 9, 10, 20,
279
fluorescence spectroscopy 279, 304,
305, 307

lab-on-a-chip 354, 362, 367, 378
biomarker imaging 36973
electroosmotic microfluidic devices
3679
labeling 1745, 225
Lactobacillus plantarum 344
Langmuir absorption model 2824
see also adsorption kinetics;
ligandreceptor interactions
400 Subject Index
lateral resolution 360
ligand immobilization 1256, 173, 174,
175, 1945, 2212
adsorptive methods 195
covalent coupling 195
amine coupling through reductive
amination 2023
amine coupling via reactive esters
199201
epoxy-mediated 20811
hydrazide activated aldehydes 2078
thiol coupling 2037
directed (biotinylated) methods 21213
electrostatic methods 21011
preconcentration 1957
and kinetic analysis 1801
membrane protein immobilization
21315
photoactivation of surfaces 3789
preconcentration procedures 195
dry immobilization 1979
electrostatic 1957
summary of strategies 21415, 21617
ligandreceptor interactions 45,
12371, 1801
affinity constants
from equilibrium signals 1268
from kinetic analysis 12934
in solution vs at the surface 1548
complex binding models 143
conformational change 1447
intermolecular bivalent bonding
14752
drug research 1234, 1657
mass transport limitation see
transport-limited interactions
rate constants 12931
thermodynamics
binding constants in solution 1548
Eyring transition states 1635
SPR compared to calorimetry 1613
vant Hoff analysis 15861
ligation chemistry 25960
light harvesting complex 299300
limit of detection 2325
line spotter 365
lipid bilayers 214
liquid handling systems see under
instrumentation
long-range surface plasmons 251, 253,
254
and fluorescence spectroscopy 3036

maleimide coupling 2056
manufacturers see instruments reviewed
mass spectrometry 3579
mass transport/transfer 8790, 17880,
184
transfer in biosensors 908, 295
see also transport-limited interactions
matrix sites 82
meat 350
membrane protein immobilization 21315
mercaptoalkyls 190
metal substrates 182
micelles, mixed 213
microarray imaging 43, 67, 269, 3623
fluorescence spectroscopy 3068
instrumentation and surface chemistry
24754
kinetic and thermodynamic
parameters 251
long-range surface plasmons 251,
253, 254
microchannel flow cell 251, 252
surface probes/target molecules
summarized 249
nanoparticle-amplified sensing 2602
SNP genotyping 2624
optimizing clinical performance
31723
spotting on gold 3624
DNA coding technology 3646
surface enzymatic enhancement 247,
254, 2649
enzyme kinetics 2547
RNA microarrays with ligation
25960
RNase H-amplification of DNA
257
see also immunoassays
Subject Index 401
microfluidic devices 251, 252, 359, 3612
concentration gradient immunoassay
37980
digital microfluidics/electrowetting
3736
electroosmotic flow 3679
free flow electrophoresis 36973
integrated cartridge 465
line spotter 366
milk 3479
miniaturization 31417
see also lab-on-a-chip
molecular interactions 1747
monoclonal antibody screening 3756
multi-layered systems, analysis of 278
multilayer adsorption 105, 11315
multiplex analysis cycles 22834

N-hydroxysuccinimide see EDCNHS
nanoarchitecture 174, 1878
three-dimensional hydrogels 1914
two-dimensional surfaces 18991
nanoparticles 2931, 3601
amplified DNA detection 2602
NHS see EDCNHS
nitrilotriacetic acid 210, 211
non-specific binding 174, 175, 1835, 327
NTA 210, 211

oligonucleotides 210, 28698, 364
and gene-modified organisms 351
optical systems see under
instrumentation
optogels 10, 44, 356
oscillatory flow 362

p-polarized light 2, 1718, 21, 44
pantothenic acid 344
partition processes 99
passivation layer see adhesion linking
penicillin 340, 342, 347
pharmaceutical research 1234, 1657
phase jump 223
phenytoin 325, 326
photoactivation 3789
physics of SPR 1533, 867
planar flow cell 467
plasma deposition 183
plasmons see surface plasmons
plastic substrates 183, 195
platinum 182
point-of-care diagnostics see clinical
diagnostics
polarization control 317
polycarboxylate hydrogel 192
poly(dimethylsiloxane) 329
flow cell 252, 252
line spotter 365
microarray fabrication 364
portable SPR see clinical diagnostics
preconcentration methods 1959
protein A-modified surfaces 213
protein binding
fluorescence spectroscopy 297,
298301
see also immunoassays; ligand
receptor interactions
protein microarrays 3623
see also microarray imaging
proteins
compatible matrix polymers 1867
disulfide reduction 204
immobilization 21315
milk assays 3489
non-specific absorption 1845
proteomics 357
lab-on-a-chip 36973

Qflex kit 334, 335
quantum dots 306, 308

ractopamine 343
Raman spectroscopy 30, 31
rate constants 78, 84
model for distribution analysis 3835
reaction rate 8790
REBIND software 1378
rebinding 13741, 226
402 Subject Index
receptorligand interactions see
ligandreceptor
reductive amination 2023
reflectance 18, 29
reflectivity change 36, 38, 45
refraction
and evanescent wave 1617, 867
SPR principles 23, 2223
refractive index
bulk compensation 3234
optical matching 44, 356
resolution 31720
temperature dependence 445, 3212
regeneration 5, 6, 175, 176
solution 225, 338
resolution
lateral 360
refractive index 31721
resonance 2, 18, 27980, 281
resonant mirror measurements 42, 62
rheumatoid arthritis 23543
auto-immune antibody complexes
3579
riboflavin 344
RNA microarrays 25960
RNADNA heteroduplexes 25460
RNase H
amplified detection of DNA 2578
and surface enzyme kinetics 2547

Salmonella 350
SAMs 182, 190, 191, 277, 300
sandwich assay 227, 228
saturation, surface 10412, 105
scanning probe microscopy 360
scanning SPR see microarray imaging
Scheimpflug condition 319, 3201
Schiff bases 202
selective capture 3
self assembled monolayers 182, 190,
191, 277, 300
sensitivity, optimization of 267
sensor chips
applications and surface structure 188
development/history 1, 912
elements of 1734, 181
nanoarchitecture 174, 1879
three-dimensional hydrogels 1914
two-dimensional surfaces 18991
optimal surface selection 17781
surface interactions 1747
see also adhesion linking layers;
bioinert matrices; ligand
immobilization
sensorgrams 3, 5, 6, 223
types/presentation of dip 368
SERS 30
serum antibodies 23542, 3579
silanes 1823
silver layer 267, 44
adhesion linking layer 182
and coherence length 234
single nucleotide polymorphisms 2624
small molecules 7, 193, 3247
smart polymer brushes 3768
spotting, microfluidic 3626
SPREETA chip 314, 366
SPRI see microarray imaging
stagnant layer 957, 178
streptavidin-modified surfaces 212
streptomycin 340, 342, 346
substrate, chip 173, 1823, 195
photoactivatable 3789
sulfadiazine 340, 3412
sulfamethazine 340, 3412
sulfathiazole 340, 346
sulfonamides 340, 3412, 346
surface chemistry see sensor chip
surface competition assay 337, 338
surface conditioning 4
surface enzymatic enhancement 247,
254, 2645
surface plasmon fluorescence
spectroscopy see fluorescence
spectroscopy
surface plasmon resonance (overview)
113, 2214, 2756
assay 36
calibration curve 67
dip to real time measurement 34
experimental parameters 267
Subject Index 403
history and biosensor development
1, 912
instrumentation 89
kinetic parameters 78
multi-layered systems 278
surface plasmons 1516
analysis of multi-layered systems 278
coherence length 234
dispersion relation 1719, 20
enhanced fluorescence and
absorbance 289
field enhancement 212
long-range 251, 253, 254
and nanoparticles 2931
phase jump 223
resonance 18
surface-enhanced Raman spectroscopy
30
symbols 323, 11819, 168

T4 DNA ligase 25960
temperature stabilization/compensation
445, 3223
thermodynamics
affinity in solution vs at the surface
1548
compared to calorimetry 1613
Eyring plots 1635
thioethers 182
thiols 182
tilted image plane 3201
transition state analysis 1635
transport see mass transport
transport-limited interactions 382
adsorption kinetics 8790, 284
detection and modeling 135
assay of high off-rates 1412
association and viscosity 1357
dissociation rebinding model 1379
quantitative considerations 1423
ligandreceptor affinity/kinetics 12535
oligonucleotide hybridization 2967
Tylosin 343

veterinary drugs 333, 33946
viscosity and association 1357
vitamins 340, 3434
VolmerWeber growth 113, 114
Vroman sequence 184

wall-jet flow cell 47
wavelength interrogation 42
wavelength tuning 315
wavevectors/wave equations 1617

Handbook of SPR

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handbook of SPR

Uploaded by

Copyright:

Available Formats

Handbook of Surface Plasmon Resonance

To Femke and Nick, Margot and Niels

Figure 1.1 Schematic experimental set-up of surface plasmon resonance excitation. A

. Jansson, D. Areskoug and J. Buijs,

You might also like