Comparison of three different media for the detection

Water Science & Technology Vol 54 No 3 pp 141–145 Q IWA Publishing 2006

of E. coli and coliforms in water
C. Bernasconi*, G. Volponi and L. Bonadonna**
*European Commission, Joint Research Center, Institute for Environment and Sustainability, Via E. Fermi 1,
21020 Ispra (Varese), Italy (E-mail: camilla.bernasconi@jrc.it )
**Istituto Superiore di Sanita’, Rome, Italy

Abstract The European Drinking Water Directive defines reference methods for the enumeration of
microbiological parameters in drinking water. The method to be used for Escherichia coli and coliforms is
the membrane filtration technique on Lactose TTC agar with Tergitol 7. Many technical drawbacks of the
procedure, as well as its limitations regarding the recent taxonomy of coliforms, make it necessary to
evaluate alternative methods. Two alternative assays, a chromogenic media (m-ColiBlu24w) and a defined
substrate technology-DST test (Colilert 18/Quanty Traye) were compared with the ISO standard with
attention to the phenotypic characteristic of the isolates. Results showed that the ISO method failed to
detect an important percentage of coliforms and E. coli while m-ColiBlu24w and Colilert 18 provided results
in a shorter time allowing the simultaneous detection of E. coli and coliforms with no further confirmation
steps.
Keywords Coliforms; E. coli; microbiological methods; water

Introduction
Coliform bacteria and Escherichia coli are recommended as hygienic indicator organisms.
The EU Directive (EC, 1998) specified the reference method for their enumeration in
drinking waters (ISO 9308-1) based on membrane filtration and growth on Lactose TTC
agar with Tergitol 7 (ISO, 2000). Readability of results and time to obtain the response
are the main shortcomings of the method. Based on the traditional coliform definition
(oxidase-negative, facultative anaerobic rod-shaped bacteria that ferment lactose to
produce acid and gas within 48 h at 35 8C (Breed and Norton, 1937)), coliforms weakly
reduce TTC giving rise to yellow-orange colonies. However, from our experience, other
organisms can often be present in water. Serratia, Shigella, Proteus, Yersinia and Hafnia
belonging to the Enterobacteriaceae but not all species are able to ferment lactose. Further-
more, some E. coli strains neither ferment lactose nor produce indole (Leclerc et al., 2001).
Advances in molecular methods and typing have recently redesigned coliform
taxonomy: the group, including E. coli, possesses b-galactosidase which hydrolyses ortho-
nitrophenil-b-galactopyranoside (ONPG) to the yellow-coloured o-nitrophenol; unlike
other coliforms, E. coli also possesses b-glucuronidase which catalyses the hydrolysis of
the 4-methyl-umbelliferyl-b-D -glucuronide (MUG). New substrate and methods have been
designed on the basis of the new coliform taxonomic scheme. Generally, during the
process of substrate utilisation, a chromogen or fluorochrome is released from the substrate
indicating the presence of the target microorganism (Bonadonna, 2003).
For environmental monitoring the use of chromogenic media is increasing. They allow
a better differentiation and readability of the analysis results with the additional advan-
tage of providing response in a shorter time compared with other methods (Edberg and
Kontnick, 1986; Sartory and Howard, 1992; Frampton and Restaino, 1993).
doi: 10.2166/wst.2006.460 141

Downloaded from http://iwaponline.com/wst/article-pdf/54/3/141/431713/141.pdf

by guest
 The present study compared m-ColiBlu24w (Millipore, MA, USA) and Colilert
18/Quanti-Traye (IDEXX, ME, USA) to the ISO standard for the enumeration of
coliforms and E. coli in different water samples, taking into account practical aspects
such as readability of results and user friendliness.

Materials and methods

Samples
C. Bernasconi et al.

Simulated samples were prepared by spiking drinking water with chlorine-injured organ-
isms derived from surface water according to DWI procedures (DWI, 2000). The same
volumes of the same samples (presumptive levels of 1– 50 target organisms/100 mL)
were processed simultaneously and in parallel with the three methods.

ISO 9308-1 (Lactose-TTC)

Analysis was performed according to ISO 9308-1 (ISO, 2000). Samples were filtered and
the membrane was incubated at 36 ^ 2 8C for 21 ^ 3 h. Confirmation steps (oxidase
activity and indole production from tryptophane) were necessary in order to discriminate
coliforms and E. coli. Oxidase (OX)-negative colonies were considered coliforms whilst
those that were oxidase-negative and indole-positive were counted as E. coli. Results
were available within 72 h.

Colilert 18/Quanti-Traye (Colilert 18)

Colilert 18 was used according to the manufacture’s instructions. It is based on the most
probable number (MPN) technique and is an enzyme-based, semi-automatic assay
reduced to multi-wells. It can analyse coliforms and E. coli in one phase with incubation
of 18 h at 36 ^ 1 8C. Yellow-coloured wells were considered total coliforms whereas
E. coli-positive wells were yellow and fluorescent under UV (365 nm).

M-ColiBlu24 (CB24)
The method uses the traditional membrane filtration technique and a chromogenic
substrate. Samples were filtered and incubated at 36 ^ 2 8C for 22 ^ 2 h. Blue colonies
were considered E. coli whereas coliforms appeared as red colonies.

Confirmation tests
The isolated colonies from all three media were considered as presumptive and subjected
to confirmatory tests. Following incubation, presumptive colonies were counted and
isolated, in a percentage of 70– 100%, for further confirmation. m-ColiBlu24 and TTC-
lactose isolates were sub-cultured onto plate count agar (PCA) medium and those from
Colilert 18 were plated on MacConkey agar. Presumptive coliform isolates were analysed
for cytochrome oxidase (negative), the ability to ferment lactose (LAC) (positive) and the
hydrolysis of ortho-nitrophenyl-b-D -galactopyranoside (ONPG) (positive). Presumptive
E. coli isolates were analysed for the same biochemical reactions (OX negative, LAC
positive, ONPG hydrolysis positive) and indole production (positive). In the case of
absence of lactose fermenting colonies, a well isolated negative colony was isolated and
characterised. Some colonies, negative for lactose fermentation but positive for ONPG
reaction or with an unusual biochemical profile, were identified at species level with the
VITEK System (Biomérieux, France).

Results and discussion

The number of coliforms isolated with each method and subjected to the OX test is
142 reported in Table 1.

Downloaded from http://iwaponline.com/wst/article-pdf/54/3/141/431713/141.pdf

by guest
 Table 1 Number of presumptive coliforms isolated with each method and subjected to the OX test

ISO 9308-1 Colilert 18 m-ColiBlu24 Totals

No presumptive 103 188 224 515

OX positive (%) 54 (53) 19 (10) 18 (8) 91 (18)

In this first screening, 54 colonies (52%) from Lactose TTC were OX positive and

C. Bernasconi et al.
thus not belonging to the coliform group. From the same samples analysed using the
Colilert 18 and the CB24, 19 (10%) and 18 (8%), respectively, had the same character-
istic. Among the OX-positive colonies, 22 (24%) were identified. Plesiomonas shigel-
loides and Aeromonas hydrophila were the predominant isolated species (27%).
Results obtained from OX-negative colonies, combining in pairs the biochemical
characteristics LAC and ONPG, are reported in Table 2. Different yields were obtained
on the colonies isolated with the three methods. From the same sample, each method was
able to sustain the growth of fractions of bacterial populations, even if those belonging to
the group of coliforms, were at least in part, dissimilar. Particularly, Lactose TTC agar
did not recover some typical coliforms (LAC and ONPG positive) in comparison with the
other methods. Furthermore, when compared with the Lactose-TTC, Colilert 18 and
CB24 were able to detect a higher number of coliforms unable to ferment lactose (and
thus atypical on the base of the traditional coliform definition).
From colonies with an atypical biochemical profile in at least one of the confirmation
tests, 32 were also biochemically identified. Results of the identification tests are reported
in Table 3. Citrobacter freundii complex, Enterobacter cloacae and Serratia marcescens
were the predominant species (13%). Other species of coliforms showed atypical profiles;
most of them (41%) were LAC negative vs. those (25%) showing a negative reaction to
ONPG. The verification of ONPG hydrolysis therefore seemed, based on obtained data,
to be the more appropriate test for the identification of the target microorganisms with
respect to the more traditional lactose fermentation.
In Table 4 the number of E. coli isolated with each method and subjected to the OX
test is reported. In this first screening, 17 colonies (21%) from Lactose TTC were OX
positive. From the same samples analysed using Colilert 18 and CB24, five (3%) and
four (3%), respectively, had the same characteristic. All the OX-negative colonies were
subcultivated at 44 8C. Different yields were obtained on the colonies isolated with the
three methods: 34 (41%), 168 (97%) and 164 (98%) colonies from Lactose TTC, Colilert
18 and CB24, respectively, were able to grow at this temperature.
Results obtained by grouping in pairs the biochemical characteristics LAC and ONPG
for E. coli, are reported in Table 5. Also in this case, the lowest recovery of the typical
target organism was found when Lactose TTC was used.
From colonies with an atypical biochemical profile in at least one of the confirmation
tests, 30 isolates were also biochemically identified (Table 6). Klebsiella pneumonia/oxytoca
Table 2 Results obtained from OX-negative colonies, combining in pairs the biochemical characteristics
LAC and ONPG

Method LAC/ONPG Totals

1 / 1 (%) 1 /2 (%) 2 /1(%) 2/2 (%)

ISO 9308-1 34 (33) 13 (13) 0 2 (1) 49

Colilert 18 132 (70) 12 (7) 19 (10) 6 (3) 169
m-ColiBlu24 188 (84) 0 18 (8) 0 206
Totals 354 (69) 25 (5) 37 (7) 8 (1) 424
Note: The percentage was calculated on the total number of presumptive coliforms 143

Downloaded from http://iwaponline.com/wst/article-pdf/54/3/141/431713/141.pdf

by guest
 Table 3 Number of species identified with the corresponding biochemical characteristics lactose (LAC) and
ortho-nitrophenil-b-galactopyranoside (ONPG)

LAC fermentation ONPG hydrolysis Identified species Number (%)

þ 2 Citrobacter farneri 1 (3)

þ 2 Citrobacter freundii complex 4 (13)
þ 2 Citrobacter braaki 1 (3)
þ
C. Bernasconi et al.

2 Enterobacter aerogenes 1 (3)

2 þ Enterobacter amnigenus 3 (9)
þ 2 Enterobacter intermedius 1 (3)
2 þ Enterobacter cloacae 4 (13)
2 þ Escherichia coli 1 (3)
2 þ Escherichia hermannii 1 (3)
þ 2 Klebsiella pneumoniae 2 (6)
2 þ Klebsiella pneumoniae 1 (3)
2 þ Klebsiella ornithinolytica 1 (3)
2 þ Leclercia adecarboxylata 1 (3)
2 þ Morganella morganii 2 (6)
2 þ Pantoea agglomerans 1 (3)
2 2 Providencia retgeri 1 (3)
2 2 Providencia stuarti 1 (3)
2 þ Serratia marcescens 4 (13)
þ 2 Yersinia intermedia 1 (3)

Table 4 Number of presumptive E. coli isolated with each method and subjected to OX test

ISO 9308-1 Colilert 18 m-ColiBlu24 Totals

No presumptive 82 173 168 423

OX positive (%) 17 (21) 5 (3) 4 (3) 26 (6)

Table 5 Results obtained from OX-negative colonies, E. coli presumptive, combining in pairs the
biochemical characteristics LAC and IND

Method LAC/IND

1/1(%) 1 /2 (%) 2/1(%) 2/2 (%) Totals

ISO 9308-1 20 (23) 9 (10) 0 5 (1) 34

Colilert 18 128 (74) 17 (10) 4 (2) 19 (11) 168
m-ColiBlu24 148 (88) 14 (8) 1 (0.6) 1 (0.7) 164
Totals 296 (70) 40 (9) 5 (1) 25 (6) 366
Note: The percentage was calculated on the total number of presumptive E. coli

Table 6 Number of species identified with the corresponding biochemical characteristics lactose (LAC) and
indole (IND)

LAC fermentation IND production Identified species Number (%)

þ 2 Citrobacter farneri 6 (20)

þ 2 Citrobacter freundii 3 (10)
2 þ Citrobacter freundii 2 (7)
2 þ Citrobacter koseri 1 (3)
2 þ Enterobacter cloacae 4 (13)
2 þ Escherichia coli 2 (7)
þ 2 Escherichia coli 3 (10)
þ 2 Klebsiella pneumoniae 9 (30)

(30%) and Citrobacter farmeri (20%) were the prevalent species. Among the colonies with
typical reactions for LAC (þ) and IND (þ), 30 confirmed E. coli were also identified: 73%
confirmed as E. coli, C. farmeri (17%), K. pneumoniae/oxytoca (7%) and C. freundii (3%)
144 were also isolated.

Downloaded from http://iwaponline.com/wst/article-pdf/54/3/141/431713/141.pdf

by guest
 Under the specific test conditions and for the analysed water samples, the reference
method allowed a large variety of interfering bacteria to grow. This was, presumably,
due to both the composition of the growth substrate and to the non-selective incubation
temperature. In fact, the calculated confirmation rate for coliforms was low (27%) for
Lactose TTC against 85% for Colilert 18 and 73% for CB24. A similar confirmation rate
was observed for E. coli: Lactose TTC 32%, Colilert 87% and CB24 94%.

C. Bernasconi et al.
Conclusions
Two different systems (Colilert 18/Quanti Tray and m-ColiBlu24) for the detection and
enumeration of coliforms and E. coli in water samples were compared with ISO 9308-1
(Lactose-TTC). This study demonstrated that the presence of background flora affected the
counts on Lactose TTC. Other authors have reached similar conclusions (Schets et al., 2002;
Bonadonna et al., 2005). Interfering organisms, cultivated as orange to brown colonies on
the medium, made it extremely difficult to select and read results. As a consequence, most of
the presumptive isolates failed to be confirmed. On the other hand, results from Colilert
18 and m-ColiBlu24 were easier to read. Colilert 18 and m-ColiBlu24 were more sensitive
than the ISO reference method and were also able to detect coliforms not fermenting lactose.
In addition, the two alternative methods had the advantage of requiring no confirmatory
steps. For coliform analysis, Colilert 18 had the highest confirmation rate (85%) whereas for
E. coli, presumptive target organisms were confirmed at 94% with m-ColiBlu24. Moreover,
Colilert 18 provided results faster than the other methods (18 h). The use of procedures that
provide quicker results are of fundamental importance when an environmental intervention
decision is needed, especially when there is a potential public health risk.

References
Bonadonna, L. (2003). Rapid analysis of microbial contamination of water. In: Rapid and On-line
Instrumentation for Food Quality Assurance. Tothill, I.E. (ed.), Woodhead Publishing, pp. 161 – 182,
ISBN 0-8493-1759-2.
Bonadonna, L., Cataldo, C., Chiaretti, G., Coccia, A. and Semproni, M. (2005). Il nuovo metodo rapido per
la determinazione di coliformi ed Escherichia coli nelle acque destinate al consumo umano. Igiene e
Sanità Pubblica (in press).
Breed, R.S. and Norton, J.F. (1937). Nomenclature for the colon group. Am. J. Publ. Hlth., 27, 560 – 563.
Drinking Water Inspectorate (DWI) (2000). The Microbiology of Drinking Water. The Comparison of
Methods for Drinking Water Bacteriology – Cultural Techniques. Report of the Drinking Water
Inspectorate, Department of the Environment, Transport and Regions.
EC (1998). Council Directive 98/83/EC. On the Quality of Water Intended for Human Consumption. Off J
Eur Comm, 5 December, L330/32 L330/53.
Edberg, S.C. and Kontnick, C.M. (1986). Comparison of b-glucuronidase-based substrate systems for
identification of Escherichia coli. J. Clin. Microbiol., 24, 368 – 371.
Frampton, E.W. and Restaino, L. (1993). Methods for Escherichia coli in food, water and clinical samples
based on ß-glucuronidase detection. J. Appl. Bacteriol., 74, 223 –233.
ISO (2000). ISO 9308-1: 2000 Water quality-detection and enumeration of Escherichia coli and Coliform
bacteria – Part 1: Membrane filtration method.
Leclerc, H., Mossel, D.A., Edberg, S.C. and Struijk, C.B. (2001). Advances in the bacteriology of the
coliform group: their suitability as markers of microbial water safety. Ann. Rev. Microbiol., 55, 201 – 234.
Sartory, D.P. and Howard, L. (1992). A medium detecting ß-glucuronidase for the simultaneous membrane filtration
enumeration of Escherichia coli and coliforms from drinking water. Lett. Appl. Microbiol., 15, 273–276.
Schets, F.M., Nobel, P.J., Strating, S., Mooijman, K.A., Engels, G.B. and Brouwner, A. (2002). EU Drinking
Water Directive reference methods for enumeration of total coliforms and methods. Lett. Appl. Microbiol.,
34, 227 – 231.

145

Downloaded from http://iwaponline.com/wst/article-pdf/54/3/141/431713/141.pdf

by guest