Letters in Applied Microbiology 2005, 40, 278–282 doi:10.1111/j.1472-765X.2005.01670.

b-D-glucuronidase activity assay to assess viable

Escherichia coli abundance in freshwaters

T. Garcia-Armisen1, P. Lebaron2 and P. Servais1

1
Ecologie des Systèmes Aquatiques, Université Libre de Bruxelles, Campus de la Plaine, Bruxelles, Belgium, and 2Observatoire
Océanologique, Université Pierre et Marie Curie, UMR 7621-7628 CNRS-INSU, Banyuls-sur-Mer cedex, France

2004/0811: received 13 July 2004, revised 22 October 2004 and accepted 10 December 2004

ABSTRACT
T . G A R C I A - A R M I S E N , P . L E B A R O N A N D P . S E R V A I S . 2005.
Aims: The relationships between the b-D-glucuronidase (GLUase) activity, the abundance of culturable
Escherichia coli and the number of viable E. coli were investigated in river and wastewater samples.
Methods and Results: GLUase activity was measured as the rate of hydrolysis of 4-methylumbelliferyl-b-D-
glucuronide. Culturable E. coli were enumerated by the most probale number (MPN) microplate method. Viable
E. coli were estimated by fluorescent in situ hybridization (FISH) coupled with a procedure of viability testing
(DVC-FISH procedure). Significant correlations were found between the log of GLUase activity and both, the log
culturable E. coli and the log of viable E. coli.
Conclusions: GLUase activity per viable E. coli gave a broadly constant value from low to highly contaminated
waters while GLUase activity per culturable E. coli strongly increased at low contaminated waters because of an
underestimation of the number of active E. coli by the culture-based method.
Significance and Impact of the Study: GLUase activity is a reliable parameter for the rapid quantification of
viable E. coli in waters.

Keywords: active but nonculturable bacteria, direct viable count, E. coli enumeration, fluorescent in situ
hybridization, river water, wastewater.

Bartram 2001). Although culture-based tests to enumerate

INTRODUCTION
Natural freshwaters can contain a large variety of
pathogenic micro-organisms released in the environment
through wastewaters and soil leaching. The routine
examination of water for the presence of intestinal
pathogens is currently a tedious, difficult and time-
consuming task. Thus surrogates such as bacterial indi-
cators of faecal contamination including total coliforms
(TC), faecal coliforms (FC), Escherichia coli and intesti-
nalis enterococci are commonly analysed as their presence
suggests that enteric pathogens may also be present.
Today, there is more and more evidence that E. coli is the
best bacterial indicator to predict the sanitary risk
associated with waters (Edberg et al. 2000; Fewtrell and
Correspondence to: Servais Pierre, Ecologie des Systèmes Aquatiques, Université
Libre de Bruxelles, Campus de la Plaine, CP 221, Boulevard du Triomphe, B-1050
Bruxelles, Belgium (e-mail: pservais@ulb.ac.be).
E. coli have been improved in terms of specificity and
rapidity by incorporating chromogenic and fluorogenic
substrates (Manafi 2000), they still require 18–24 h to
complete and are thus unable to rapidly detect any faecal
pollution.
More recently, some authors proposed to estimate the
b-D-glucuronidase (GLUase) activity of E. coli in rapid
assays performed without any cultivation step. Direct
measurements of GLUase activity have been successfully
applied to seawaters (Fiksdal et al. 1994; Caruso et al. 2002),
freshwaters (George et al. 2000, 2004; Farnleitner et al.
2001, 2002) and wastewaters (George et al. 2001, 2002).
These direct measurements of GLUase activity were
performed using the substrate 4-methylumbelliferyl-b-D-
glucuronide (MUGLU); the activity was measured as the
rate of production of fluorescent methylumbelliferone
(MU), resulting from the hydrolysis of the substrate,
ª 2005 The Society for Applied Microbiology

estimated by fluorometry. This type of direct activity
measurement can be performed in a period as short as half
an hour. Several studies showed good correlation’s in log–
log plot between GLUase activity and FC enumerated on
plate counts in different types of aquatic systems (Fiksdal
et al. 1994; George et al. 2000, 2001, 2004; Caruso et al.
2002; Farnleitner et al. 2002). More recently, a significant
correlation was found between GLUase activity and E. coli
abundance estimated by plate counts (Farnleitner et al.
2001). This rapid enzymatic assay can thus be proposed as a
possible surrogate to culture-based methods to rapidly
detect faecal pollution in river waters.
The regression straight lines in log–log plot observed by
the different authors quoted here above between GLUase
activity and FC abundance, on one hand, and between
GLUase activity and E. coli numbers, on the other, have
always slope lower than 1. George et al. (2000) hypothes-
ized that this was because of an underestimation of active
FC or E. coli by culture-based methods in low contam-
inated waters because of the existence in these environ-
ments of many active but nonculturable (ABNC)
coliforms, i.e. cells presenting a detectable GLUase
activity but unable to multiply in liquid or on solid
culture media. In this paper, we compared GLUase
activity measurements with culturable E. coli numbers
determined by the most probable number (MPN) tech-
nique using microplates and with viable E. coli numbers
estimated by fluorescent in situ hybridization (FISH)
coupled with a procedure of viability testing. A DVC-
FISH approach combining the direct viable count (DVC)
procedure (Kogure et al. 1979) with FISH procedures was
recently developed to enumerate viable E. coli in river
waters and wastewaters (Garcia-Armisen and Servais
2004). The DVC procedure involves exposing bacterial
cells to a revival medium containing antibiotics preventing
cellular division; elongated cells are then enumerated as
viable cells.
The comparison was performed on samples collected in
differently contaminated rivers and in several wastewater
treatment plants (WWTP) (raw and treated waters) in order
to cover a very large range of E. coli abundance.
MATERIALS AND METHODS
Samples collection
River water samples were collected in the Seine river
hydrographical network (France). Samples were harvested
from small rivers upstream to any domestic wastewater
discharge up to the highly contaminated Seine river
downstream to the Parisian area. Moreover, mean daily
wastewater samples were collected at the entrance and at the
outlet of various WWTP located in France and differing by
ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 40, 278–282, doi:10.1111/j.1472-765X.2005.01670.x
GLUCURONIDASE ACTIVITY OF E. COLI 279
their size and type of treatment. All samples were collected
in sterile 2-l bottles, kept at 4
C and analysed within 12 h.
b-D-glucuronidase activity measurements
b-D-glucuronidase (GLUase) activity measurements were
performed following the protocol proposed by George
et al. (2000) slightly modified. River water samples
(100 ml) and wastewater samples (1–20 ml) were filtered
through 0Æ2 lm pore-sized, 47 mm-diameter polycarbon-
ate filters (Nuclepore, Whatman, Maidstone, UK). The
filters were placed in 200-ml sterile Erlenmeyer flasks
containing 17 ml of sterile phosphate buffer (pH 6Æ9) and
3 ml of MUGLU solution (55 mg of MUGLU; Biosynth,
Staad, Switzerland) and 20 ll of Triton X-100 in 50 ml
of sterile water) was added to each flask (final concentra-
tion: 165 mg l)1). The flasks were incubated in a shaking
water bath at 44
C. Every 5 min for 30 min, a 2Æ9-ml
aliquot of the 20 ml was poured in a quartz cell with
110 ll of 1 M NaOH solution to increase the pH to 10Æ7
(corresponding to the maximum of fluorescence of the
MU). The fluorescence intensity of the aliquot was
measured with a SFM 25 spectrofluorometer (Kontron
AG, Zürich, Switzerland) at an excitation wavelength of
362 nm and emission wavelength of 445 nm. The 100%
of fluorescence intensity of the fluorometer was calibrated
using standards of known MU concentrations from 50 to
12 500 nM; this procedure allowed to study a wide range
of enzymatic activities by changing only the fluorometer
calibration. The production rate of MU (pmol of MU
liberated per minute for 100 ml of sample filtered),
expressing the enzymatic activity, was determined by
least-squares linear regression when plotting MU concen-
tration vs incubation time. Data of GLUase activities
presented in this paper are expressed in pmol of MU
liberated per minute for 100 ml of sample filtered.
E. coli enumeration by miniaturized MPN method
Standardized miniaturized MPN method (ISO 9308-3)
using microplates (Bio-Rad, Hercules, CA, USA) is based
on the defined substrate approach (Edberg and Edberg
1988). Briefly, 200 ll of several decimal dilutions (from
1/2 to 1/200 000) of the sample were added in each of the
96 wells of the microplate containing the substrate
(MUGLU) in dehydrated form. The number of dilutions
used (2, 4 or 6) depends on the abundance of E. coli in the
sample. The microplates were incubated for 36–48 h at
44
C. The hydrolysis of MUGLU by the b-D-glucuroni-
dase releases the fluorescent MU compound which can be
detected under UV light. The number of positive wells
(fluorescent under UV light) after incubation allows to
calculate the E. coli abundance using a statistical table
280 T . G A R C I A - A R M I S E N ET AL.

provided by the microplates manufacturer (Bio-Rad) based also performed by detecting autofluorescent micro-organ-
on Poisson’s law. isms.

E. coli enumeration by DVC-FISH procedure RESULTS

The DVC-FISH procedure described by Garcia-Armisen
and Servais (2004) was used in the present study for
enumeration of viable E. coli. Water samples were filtered
through 25 mm 0Æ4 lm pore-sized polycarbonate filters
(Nucleopore; Whatman). One to 100 ml were filtered
depending on the origin of the sample and on its suspended
matter content. For the DVC procedure, filters were
incubated for 4 h at 30
C on absorbent pad (Millipore,
Billerica, MA, USA) previously soaked into culture plates
(Greiner Bio-One, Frickenhausen, Germany) with 0Æ6 ml of
revival medium consisting in TCS broth (Bio-Rad) supple-
mented with yeast extract (0Æ6% w/v), nalidixic acid
(10 lg ml)1) (Sigma) and ciprofloxacin (1 lg ml)1) (ICN,
Irvine, CA, USA). After the 4 h incubation, bacteria were
fixed for 15 min with 3% (w/v) paraformaldehyde.
For hybridization, filters issued from the DVC stage
were transferred onto a new pad soaked with 300 ll of
hybridization buffer consisting of 0Æ9 M NaCl, 20 mM
Tris-HCl (pH 7Æ2), 0Æ1% (w/v) sodium dodecyl sulfate
(SDS), 0Æ2% (w/v) bovine serum albumin (BSA, fraction
V; Sigma), 0Æ1 mg ml)1 polyadenylic acid (ICN) and 22%
During this study, river water and wastewater samples
covering a large range of faecal contamination were analysed
using three techniques: GLUase activity, E. coli enumer-
ation by MPN microplate method and DVC-FISH tech-
nique. This procedure involves exposing bacterial cells to a
revival medium containing antibiotics preventing cellular
division; elongated cells are then enumerated as viable cells
after hybridization with a fluorescent probe specific for the
16S rRNA of E. coli.
In Fig. 1a, log-transformed GLUase activities were plot-
ted against log-transformed E. coli abundance. Some samples
for which the E. coli number was below the detection limit of
the MPN method (15 E. coli per 100 ml) were not
considered in this comparison. The linear correlation
between both variables in log units was highly significant
(r2 ¼ 0Æ92, n ¼ 41, P < 0Æ001); the slope of the regression
straight line was 0Æ731. In Fig. 1b, log-transformed GLUase
activities were plotted against log-transformed viable E. coli
(a) 6
pmol min–1 (100 ml)–1)

deionized formamide (ACROS, Geel, Belgium). The filters 5

Log (GLUase act

carrying the fixed bacteria were covered with 300 ll of 4

hybridization buffer containing 10 pmol of CY3-labelled 3
‘Colinsitu’ probe and 20 pmol of FITC-labelled eubacterial
2
probe EUB338. The ‘Colinsitu’ probe (sequence: 5¢-GAG
ACT CAA GAT TGC CAG TAT CAG-3¢) used for the 1
in situ E. coli hybridization was shown to be E. coli specific 0
0 2 4 6 8 10
(Regnault et al. 2000); it was labelled at the 5¢-end with the Log(E. coli (100 ml)–1)
CY3 dye. Eubacterial probe EUB338 labelled with FITC (b)
was used as a control in all in situ hybridization experi- 6
ments. The hybridization was performed for 2 h at 42
C.
pmol min–1 (100 ml)–1)

5
Log (GLUase act

Filters were then washed in new multiwell tissue culture

4
plates with 2 ml of prewarmed washing buffer [0Æ1 M
NaCl, 20 mM Tris-HCl (pH 7Æ2), 0Æ01% SDS, 5 mM 3
EDTA] and incubated at 51
C for 20 min. The filters 2
were then deposited on microscopic slides for mounting. 1
Filters with hybridized cells were observed by epifluores-
0
cence microscopy (Leica, Wetzlar, Germany) equipped with 0 2 4 6 8 10
a HBO 100-W mercury lamp (Osram, Munchen, Germany) Log(E. coli (100 ml)–1)
at a magnification of 1000·. For river water samples and
Fig. 1 (a) Log–log linear regression between GLUase activity and
wastewater samples 100 and 50 microscopic fields, respect-
Escherichia coli in river water and wastewater samples: log (GLUase
ively, were observed, elongated and hybridized cells were
activity pmol min)1 100 ml)1) ¼ 0Æ731 log (E. coli per
enumerated and the E. coli abundance (expressed in E. coli 100 ml) ) 0Æ663 (r2 ¼ 0Æ92, n ¼ 41, P < 0Æ001). (b) Log–log linear
per 100 ml) calculated. A positive control was performed by regression between GLUase activity and viable E. coli enumerated by
verifying a fluorescent event labelled with CY3 as being a the DVC-FISH procedure in river water and wastewater samples: log
bacteria; for this, bacterial cells were visualized using the (GLUase activity pmol min)1 100 ml)1) ¼ 0Æ994 log (E. coli per
EUB 338 probe labelled with FITC. A negative control was 100 ml) ) 3Æ053 (r2 ¼ 0Æ93, n ¼ 41, P < 0Æ001)

ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 40, 278–282, doi:10.1111/j.1472-765X.2005.01670.x
 GLUCURONIDASE ACTIVITY OF E. COLI 281

100 activity and E. coli abundance estimated by plate counts

GLUase act. per E. coli

using Chromocult medium in river waters (Farnleitner et al.

(fmol min–1 cell–1)

10 2001) and between GLUase activity and culturable FC (see

1
0·1
100 < EC < 1000 1000 < EC < 10 000 10 000 < EC < 1 00 000 EC > 1 00 000
Fig. 2 Average GLUase activity per culturable (black bars) and per
viable Escherichia coli (grey bars) (expressed in fmol min)1 per E. coli
cell) for various ranges of culturable E. coli (EC estimated by the MPN
technique) abundances (expressed in EC per 100 ml)
numbers estimated by the DVC-FISH procedure. The
water samples used for this regression analysis were similar
to those used for the regression analysis between GLUase
activity and culturable E. coli (Fig. 1a). The linear correla-
tion between GLUase activity and viable E. coli in log units
was highly significant (r2 ¼ 0Æ93, n ¼ 41, P < 0Æ001); the
slope of this regression straight line was 0Æ994.
Figure 2 presents the GLUase activity per culturable E. coli
(GLUase activity divided by the E. coli number estimated by
the MPN method) and per viable E. coli (GLUase activity
divided by E. coli count estimated by the DVC-FISH
procedure) for different ranges of culturable E. coli repre-
senting different ranges of faecal contamination of waters.
Data from Fig. 2 show a rather constant GLUase activity per
viable E. coli (c. 1 · 10)15 mol of MU produced per minute
per E. coli cell) whereas the GLUase activity per culturable
E. coli strongly decreased from the low contaminated
environments to highly contaminated systems (from
c. 100 · 10)15 mol min)1 per E. coli for waters containing
between 100 and 1000 culturable E. coli per 100 ml to
c. 6 · 10)15 mol min)1 per E. coli for highly contaminated
waters containing more than 105 culturable E. coli per 100 ml).
DISCUSSION
Data from Fig. 1a are congruent with previous reports
showing good correlations in log–log plot between GLUase
Table 1 for references). The slope of the regression straight
line between GLUase activity and E. coli abundance
obtained in this study was lower than 1 in agreement with
previous studies (Table 1). Table 1 shows that the slopes of
the regression straight lines in log–log plot between GLUase
activity and FC or E. coli abundance vary depending on the
type of water samples analysed (marine waters, rivers,
wastewaters), on the culture-based method used to enumer-
ate FC or E. coli and on the protocol used to determine
GLUase activities (for example, substrate concentration or
incubation temperature). Nevertheless and interestingly, in
all of these studies the slopes were found to be lower than 1,
ranging from 0Æ31 to 0Æ83. This indicates that the GLUase
activity per cultured E. coli or per FC decreased as the
number of culturable faecal bacteria increased; in other
words, the ratio GLUase activity per culturable faecal
bacteria decreased in more polluted environments. A
possible explanation for this observation was suggested by
George et al. (2000). These authors suggested that higher
enzymatic activities per culturable cells in less contaminated
natural waters may be due to an underestimation of the
number of faecal bacteria when enumerated by culture-
based methods. This underestimation may be explained by a
higher proportion of ABNC cells (cells presenting a
detectable GLUase activity but unable to multiply in or
on the specific media used in culture-based methods). The
higher proportion of ABNC faecal bacteria in less polluted
environments could be the result of more severe and/or
longer environmental stress factors such as nutrient
limitation and enhanced solar radiation effects as a result
of deeper light penetration. George et al. (2000) and Petit
et al. (2000) demonstrated in microcosm experiments the
increasing proportion of nonculturable E. coli cells main-
taining a GLUase activity when cells were exposed to light
stress.
In this study, the potential underestimation of E. coli cells
at low concentrations when detected by culture-based
methods was tested by comparing GLUase activities with

Table 1 Slope of the regression straight lines

Methods Type of water Slope References
between log-transformed GLUase activity
and log-transformed abundance of culturable FC MacConkey agar Estuarine waters and 0Æ68 Fiksdal et al. (1994)
FC or Escherichia coli published in the litera- wastewaters
ture. Table also mentions the methods used to FC Tergitol TTC agar Rivers 0Æ61 George et al. (2000)
enumerate culturable FC and E. coli as well as FC Tergitol TTC agar Wastewaters 0Æ83 George et al. (2001)
the type of waters investigated in the different FC Tergitol TTC agar Small rivers 0Æ55 George et al. (2004)
studies FC m-FC-agar Seawaters 0Æ31 Caruso et al. (2002)
FC MFC-agar Rivers 0Æ82 Farnleitner et al. (2002)
E. coli Chromocult coliform agar Rivers 0Æ83 Farnleitner et al. (2001)
E. coli MPN microplates Rivers and wastewaters 0Æ73 This study

ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 40, 278–282, doi:10.1111/j.1472-765X.2005.01670.x
282 T . G A R C I A - A R M I S E N ET AL.
viable E. coli counts using the DVC-FISH procedure.
Fig. 1b presents log-transformed GLUase activities plot-
ted against log-transformed viable E. coli numbers
estimated by the DVC-FISH procedure. The linear
correlation between GLUase activity and viable E. coli
in log units was highly significant but more interestingly,
the slope of the regression straight line was very close to
1. This indicates a rather constant GLUase activity per
viable E. coli detected by the DVC-FISH procedure. Data
from Fig. 2 show broadly constant values of GLUase
activity per viable E. coli whatever the level of contam-
ination of the aquatic system while GLUase activity per
culturable E. coli strongly decreased from the low
contaminated environments to highly contaminated sys-
tems. These data sustain the hypothesis that low
contaminated waters contain higher proportion of ABNC
cells which are not detected by culture-based methods but
maintain with GLUase activity detectable by the fluoro-
metric method. This underestimation of active cells in low
contaminated environments by the culture-based method
may explain the slope lower than 1 usually found in
regression straight lines of log-transformed GLUase
activity versus log-transformed culturable FC or E. coli
numbers (Table 1).
The data presented here show the capability of the
GLUase activity measurement to detect viable but
nonculturable E. coli in waters very differently contam-
inated by faecal material as it can be performed by the
DVC-FISH procedure. GLUase activity seems thus to be
an easy and rapid way to estimate the abundance of viable
E. coli in waters. This enzymatic assay offers the
advantage of being rapid and easy to perform but also
to be better adapted than any other method to the
detection of E. coli in low contaminated waters such as
recreational waters for which guidelines concentrations
(100–2000 E. coli per 100 ml) are in the range of
concentrations that are detectable by the enzymatic
method.
ACKNOWLEDGEMENTS
Tamara Garcia-Armisen benefits from a doctoral grant from
the ‘Fonds pour la Formation à la Recherche dans
l’Industrie et l’Agriculture’ (FRIA) (Belgium). This study
was a part of the PIREN Seine program of the Centre
National de la Recherche Scientifique (France). The authors
thank Adriana Anzil and Philippe Mercier for their help
during the field work.
ª 2005 The Society for Applied Microbiology, Letters in Applied Microbiology, 40, 278–282, doi:10.1111/j.1472-765X.2005.01670.x
REFERENCES
Caruso, G., Crisafi, E. and Mancuso, M. (2002) Development of an
enzyme assay for rapid assessment of Escherichia coli in seawaters.
J Appl Microbiol 93, 548–556.
Edberg, S.C. and Edberg, M.M. (1988) A defined substrate technology
for the enumeration of microbial indicators of environmental
pollution. Yale J Biol Med 61, 389–399.
Edberg, S.C., Rice, E.W., Karlin, R.J. and Allen, M.J. (2000)
Escherichia coli: the best biological drinking water indicator for
public health protection. J Appl Microbiol 88, 106S–116S.
Farnleitner, A.H., Hocke, L, Beiwl, C., Kavka, G.G., Zechmeister, T.
Kirschner, A.K.T. and Mach, R.L. (2001) Rapid enzymatic
detection of Escherichia coli contamination in polluted river water.
Lett Appl Microbiol 33, 246–250.
Farnleitner, A.H., Hocke, L, Beiwl, C., Kavka, G.G. and Mach, R.L.
(2002) Hydrolysis of 4-methylumbelliferyl-b-D-glucuronide in
different sample fractions of river waters and its implication for
the detection of fecal pollution. Water Res 36, 975–981.
Fewtrell, L. and Bartram, J. (2001) Water Quality: Guidelines,
Standards and Health. London, UK: World Health Organization
Water Series, IWA Publishing.
Fiksdal, L., Pommepuy, M., Caprais, M-P. and Midttun, I. (1994)
Monitoring of fecal pollution in coastal waters by use of rapid
enzymatic techniques. Appl Environ Microbiol 60, 1580–1584.
Garcia-Armisen, T. and Servais, P. (2004) Enumeration of viable
E. coli in rivers and wastewaters by fluorescent in situ hybridization.
J Microbiol Methods 58, 269–279.
George, I., Petit, M. and Servais, P. (2000) Use of enzymatic methods
for rapid enumeration of coliforms in freshwaters. J Appl Microbiol
88, 404–413.
George I., Crop P. and Servais P. (2001) Use of b-D-galactosidase and
b-D-glucuronidase activities for quantitative detection of total and
fecal coliforms in wastewater. Can J Microbiol 47, 670–675.
George I., Crop P. and Servais P. (2002) Fecal coliform removal in
wastewater treatment plants studied by plate counts and enzymatic
methods. Water Res 36, 2607–2617.
George, I., Anzil, A. and Servais, P. (2004) Quantification of fecal
coliform inputs to aquatic systems through soil leaching. Water Res
38, 611–618.
Kogure, K., Simidu, U. and Taga, N. (1979) A tentative direct
microscopic method for counting living marine bacteria. Can J
Microbiol 25, 415–420.
Manafi, M. (2000) New development in chromogenic and fluorogenic
culture media. Int J Food Microbiol 60, 2005–2218.
Petit, M., George, I. and Servais, P. (2000) Survival of Escherichia coli
in freshwaters: b-D glucuronidase activity measurements and
characterization of cellular states. Can J Microbiol 46, 679–684.
Regnault, B., Martin-Delautre, S., Lejay-Collin, M., Lefèvre, M. and
Grimont, P.A.D. (2000) Oligonucleotide probe for the visualization
of Escherichia coli/Escherichia fergusonii cells by in situ hybridization:
specificity and potential application. Res Microbiol 151, 521–533.