United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-84-166 Dec. 1984
<&ER& Project Summary
Injury and the Improved
Recovery of Coliform Bacteria
in Drinking Water
Gordon A. McFeters, Mark W. LeChevallier, and Matthew J. Domek
A study was conducted to answer
basic questions about the fate and han-
dling of stressed indicator bacteria in
aquatic environments. The importance
of stressed indicator bacteria in asses-
sing water quality has been gaining at-
tention as microbiologists have discov-
ered the inaccuracy of present methods
and media for recovering indicator bac-
teria in water. The study examined the
effects of laboratory manipulations on
recovery of injured indicator bacteria.
These manipulations include exposure
to diluents, choice of enumeration
medium, and membrane fitter surface
pore morphology. Results indicated
that diluent composition was of little
importance, provided the diluent was
cold (4°C) and exposure was brief (<30
min.), and that the most commonly
used media recovered less than 30% of
an injured cell population.
Based on these results, a new mem-
brane fitter medium, m-T7, was de-
veloped that recovered 86% to 99%
more laboratory-injured coliforms and
3 times more coliforms from drinking
water than did m-Endo medium. The
m-T7 medium also recovered 43%
more verified coliforms from 67 surface
and drinking water samples than did
the standard m-Endo membrane fitter
technique. The medium was also
evaluated for its effectiveness as a fecal
coliform medium. Recoveries of fecal
coliforms on m-T7 were 3.1 times
greater than the standard m-FC
method and 1.7 times greater than the
two-layer enrichment, temperature ac-
climation procedure.
Comparisons of verification tech-
niques for the membrane fitter total
coliform method suggest that standard
procedures may result in significant un-
derestimations of indicator densities,
particularly in waters where coliforms
are stressed or injured.
Laboratory and field data indicated
that copper at levels normally found in
drinking water could injure coliform
bacteria. Study results showed that
90% injury could occur at copper con-
centrations of 0.050 mg/L.
Studies were also initiated to investi-
gate the interactions between standard
plate count (SPC) bacteria and coliform
organisms. A significant correlation
was observed between the initial SPC
level and the rate of coliform decline.
This Project Summary was devel-
oped by EPA's Water Engineering Re-
search Laboratory, Cincinnati, OH, to
announce key findings of the research
profect that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Indicator bacteria have been success-
fully used in this country for nearly 100
years to identify water that may contain
infectious agents. During this time, seri-
ous disease outbreaks associated with
water have decreased significantly be-
cause of disinfection, the acceptance of
improved engineering practices, waste-
water treatment facilities, and micro-
biological surveillance. However, dur-
ing the last decade a pronounced in-
crease has occurred in both the number
of cases and outbreaks of waterborne
disease reported by the Center for Dis-
ease Control (CDC) in Atlanta, Georgia.
-------
At the same time, some experts believe
that as many as 90% of the outbreaks
are unreported, since many people fail
to associate their illness with contami-
nated drinking water. This increase
could simply be the result of improved
reporting, or it could be a real
phenomenon resulting from the over-
loading of our treatment plants with
source water of increasingly lower qual-
ity.
The ultimate solution to this large and
complex problem depends on the cor-
rectly directed efforts of federal, state,
and local authorities charged with the
responsibility of providing drinking
water that is free from contamination.
The microbiological facet of this task
largely depends on meaningful epi-
demiological and surveillance informa-
tion, which in turn relies on the accurate
detection and enumeration of water-
borne indicator bacteria within a rea-
sonable time. The basic microbiological
tools presently being used for this pur-
pose are inaccurate, however, because
many of the procedures recover only a
portion of the total population of indi-
cator bacteria present in the water.
One of the basic problems underlying
this assessment dilemma concerns the
lack of understanding of what happens
to indicator bacteria in aquatic environ-
ments. It is now apparent that injured
indicator bacteria occur in almost every
aquatic environment. The proportion of
injured coliforms can range up to 90%
and more. Injury has been defined as
the sublethal physiological conse-
quence of exposure to stresses which
cause a loss in the ability of microor-
ganisms to grow normally under condi-
tions satisfactory for untreated cells. In-
jury has also been described as reversi-
ble; that is, under proper conditions of
temperature and nutrients, injured or-
ganisms can repair the cellular lesion
and become indistinguishable from un-
stressed cells. As a result, it is possible
to resuscitate and enumerate injured
coliform bacteria.
This study addresses some of these
fundamental questions regarding stress-
ed indicator organisms in aquatic sys-
tems, including the following: how to
collect and handle samples that contain
injured coliform bacteria, how to enum-
erate injured total and fecal coliforms,
how verification procedures influence
the assessment of potable water
supplies, and how various physio-
chemical and biological parameters
cause injury to coliform bacteria in
drinking water systems.
Results and Discussion
Sample Collection and Handling
The way a water sample is collected
and handled before enumeration great-
ly influences the viability of injured bac-
teria in that sample. Sodium thiosulfate
is usually added to drinking water sam-
ples to neutralize any chlorine residual.
In addition, the results suggest that a
chelator such as EDTA should be rou-
tinely added to remove heavy metals,
particularly if there is an extended
period between sample collection and
enumeration.
Laboratory manipulations involving
exposure to diluents may cause further
underestimations of bacterial densities
in surface water. If diluents are main-
tained at approximately refrigerator
temperatures (about 4°C), their com-
positions and exposure times have min-
imal impacts on enumeration effi-
ciency. However, substantially lower
enumeration recoveries of injured bac-
teria have been associated with room
temperature and extended exposure
time (Figure 1). The enrichment of di-
luents with low concentrations (0.05%
to 0.1%) of organic material such as
peptone, gelatin, tryptone, or milk has
been demonstrated to be of value in the
enumeration of stressed aquatic bac-
teria.
Enumeration of Injured Total
and Fecal Coliform Bacteria
New membrane filter techniques
have been proposed for enumerating
stressed total and fecal coliforms. Steps
taken in the formulation of a new selec-
tive agar medium (m-T7) have ensured
that it is selective and that it minimizes
the inhibition of stressed coliforms. The
formula and preparation instructions
for m-T7 agar appear in Table 1. The
effectiveness of m-T7 agar was dem-
onstrated when it recovered an average
of 43% more verified coliforms from 67
surface and drinking water samples
than did the standard m-Endo mem-
brane filter technique. Overall, m-T7
agar recovered significantly more (p <
0.05) coliforms than either the standard
m-Endo or m-Endo with lauryl tryptose
broth (LTB) resuscitation techniques. In
one drinking water sample, the m-T7
agar enumerated 17 confirmed coli-
forms, whereas the m-Endo + LTB
method recovered none. From the 44
drinking water samples analyzed, m-T7
agar detected 2.7 times more coliforms
(a 63% increase) than the m-Endo re-
suscitation technique and nearly 3
Minutes Exposed
Figure 1. Effect of diluent composition
and exposure time on the re-
covery of injured (90%) and
uninjured control^., colt suspen-
sions at 24°C. The control line is
an average of all data obtained
when uninjured cells were ex-
posed in each diluent. Diluents
are 1.0% milk, 0.1% peptone.
reagent-grade water, phosphate
buffer, and phosphate buffer
amended with 0.1% peptone.
Values are a mean of three
experiment replications.
times more coliforms than the standard
m-Endo method (Table 2). We also
found that m-T7 agar recovered more
verified coliforms from drinking water
samples than did the m-Endo; m-Endo
+ LTB pad; m-Endo + lactose agar (LA)
overlay; m-T7 + proteose peptone #3,
lactose, and yeast extract (PLY) pad; or
m-T7 anaerobic techniques. In addition,
five collaborating laboratories reported
an average increase of 31% in coliform
recoveries from a variety of contami-
nated surface and sewage samples
using m-T7 agar compared with the
standard m-Endo method. Note that our
results showed that m-T7 agar per-
formed best for the isolation of coli-
forms from drinking water. Since m-T7
agar was specifically designed to re-
cover injured coliforms from drinking
water, further studies are needed to
evaluate its potential fully. Other inves-
tigators interested in coliform injury
will, it is hoped, use m-T7 and evaluate
the results in relation to other coliform
media.
-------
Tahiti. Formulation of m-T7 Medium
Ingredient?
Amount per Liter of
Distilled Water
Proteose peptone #3
Yeast extract
Lactose
Tergitol 7 (25% solutionJ
Polyoxyethylene ether W1
Brom thymol blue
Brom cresol purple
Agar
Final pH*
59
3g
20 g
Ami
5g
0.1g
O.lg
15 g
7.4
'All ingredients were manufactured by Difco Laboratories except for polyoxyethylene ether
W1 and brom cresol purple, which were manufactured by Sigma Chemicals, and tergitol 7,
which was obtained from Baker Chemical Company.
"The medium was autoclaved at 121°C for 15 min and finalpH was adjusted aseptically to 7.4
with 0.1 N NaOH. Additional selectivity may be obtained by aseptically adding 0.1 \t.g/ml
penicillin G to the medium after autoclaving. Media prepared with penicillin G should be used
within 1 week when stored at 4°C.
Table 2. Comparison of m-Endo and m- 77 Techniques
Relative Coliform Counts
Medium
m-Endo
m-Endo + LTB pad"
m-Endo + LA overlay*
m-T7
m-T7 + PLY pad'
m-T7 — anaerobic
Drinking
Water
n' = 44
1.41
1.48
1.27
3.96*
3.85*
0.57
Surface
Water
n= 11
27.8
32.3**
32.8*
39.9*
40.3*
37.4*
Chlorinated
Surface Water
n = 12
13.6
14.2
22.6*
20.6*
25.2*
22.8*
'Number of samples.
bLauryl tryptose broth pad.
"Asterisk indicates a significantly greater value compared with m-Endo (p < .05).
dLactose agar overlay.
'Base composition of m- T7 containing proteose peptone #3, lactose and yeast extract.
The problem of enumerating injured
fecal coliform bacteria has been recog-
nized by the 15th edition of Standard
Methods for the Examination of Water
and Wastewater in a new section en-
titled "Stressed Organisms" (section
no. 920). Included in this section are
various resuscitation techniques for
fecal coliforms. We compared the effec-
tiveness of m-T7 agar with some of the
proposed resuscitation techniques for
the recovery of injured fecal coliforms
from chlorinated sewage effluents.
Though a limited number of samples
were examined, m-T7 agar preincu-
bated at 37°C for 8 h before transfer to
44.5°C for an additional 12 h was
superior for the recovery of verified
fecal coliforms compared with any of
the resuscitation techniques tested.
This method (abbreviated m-T7 8 h,
37°C) produced counts 3.1 times greater
than the standard m-FC technique and
1.7 times greater than the two-layer en-
richment, temperature acclimation pro-
cedure.
Influence of Verification
Procedures
When routine coliform analyses of
drinking water samples yielded a high
percentage of typical presumptive coli-
forms that did not produce gas in LTB,
studies were initiated to evaluate a
more reliable verification procedure
and to investigate whether injury to
coliform organisms affected their ability
to produce gas in LTB. Comparison of
verification techniques for the mem-
brane filter total coliform method were
made using the standard LTB proce-
dure and the B-galactosidase and cyto-
chrome oxidase (B-gal/CO) method
using coliforms isolated from 21 sur-
face and drinking water samples. Of the
682 presumptive coliforms tested from
all types of water samples, 577 (84.6%)
were verified using the B-gal/CO
method, whereas only 402 (58.9%) pro-
duced gas in LTB (Table 3). When verifi-
cation procedures included the confir-
matory brilliant green-lactose-2% bile
broth (BGLB) step, the verification rate
dropped to 50.4%. Of the 146 presump-
tive coliforms tested from drinking
water samples, 119 (81.5%) were ver-
ified using the B-gal/CO method, and
only 15 (10.3%) produced gas in LTB
(Table 3). More than 90% of the anaero-
genic lactose fermenters verified by the
B-gal/CO method were representative
of typical coliform genera (Escherichia,
Enterobacter, Klebsiella, or Citrobacter).
A significant correlation existed be-
tween the level of injury and the percent
of verification (using LTB) of coliforms
from m-T7 agar (r = -0.63; p <0.001).
This correlation may be related to some
irreversible cellular injury involved in
gas production from lactose, or it may
reflect an association between an-
aerogenic coliforms and an increased
susceptibility to aquatic injury.
Sources of Injury to Coliform
Bacteria
Injury in the aquatic environment
may be related to a number of factors,
including time and temperature of ex-
posure, disinfection levels, strain of or-
ganism, concentration of nutrients, and
possibly other undefined chemical and
physical parameters. In addition, this
study establishes the role of copper and
heterotrophic plate count bacteria in the
injury process of coliforms in drinking
water.
A combination of laboratory studies
and field data indicated that copper
present at concentrations normally
found in drinking water could cause sig-
nificant injury to coliform bacteria.
Though other metals such as cadmium,
lead, and zinc were also detected in
drinking water samples, these metals or
combinations of them were not as im-
portant in the injury process as copper
alone. Copper has been detected in
65.2% of 380 finished drinking water
supplies in the United States with a
mean concentration of 0.043 mg/L and
a range of 0.001 to 1.060 mg/L Results
indicate that 90% injury could occur in
coliform populations within 2 days at
copper concentrations near the national
mean (Figure 2). Statistical analyses of
30 drinking water samples suggested
that pH, alkalinity, and temperature also
influenced the amount of injury, pre-
sumably by altering the chemical
species and activity of the dissolved
-------
Table 3.
Comparison of Verification Procedures in Drinking Water, Surface Water, and
Chlorinated Surface Water for Total Conforms.
V-gal/CO Method"-"
L TB Procedure0
Source
Drinking
water
Chlorinated
surface water
Surface
water
Total
No. of No. of
Samples Presumptives
12 146
4 280
5 256
21 682
No. ONPG +
Co-
r/9
232
226
577
% ONPG +
Co~
81.5
82.8
88.3
84.6
No. Gas +
inLTB
15
190
197
402
% Gas +
inLTB
10.3
69.8
76.9
58.9
"fi-galactasidase and cytochrome oxidase method.
bONPG = o-nitrophenyl-$-D-galactopyranoside; CO = cytochrome oxidase.
CL TB = lauryl tryptose broth.
copper. Other factors such as total or-
ganic carbon levels and the species of
coliform present are thought to influ-
ence injury resulting from copper.
Physiologically, copper-injured cells
show impaired respiratory activity and
are unable to use oxygen during the re-
99.9 r
99
90
10
0.05 mg Cu/L
0 1
345
Days
Figure 2. Injury induced by copper at con-
centrations within the national
range for drinking water. Cells
were injured by placing a
washed suspension of E. coli
/10a cfu/ml) in carbonate-buf-
fered, reagent-grade water (pH
7.0). This suspension was
amended with various levels of
copper fas CuSOfSH^O) and
incubated at 4°C for periods up
to 7 days.
pair process. Sources of copper in
drinking water include copper pipe ma-
terial, brass, faucets, and other attach-
ments. Copper may also be introduced
as copper sulfate (which is used to treat
algal problems in reservoirs) and as
trace contaminants added with floccula-
tion material. In addition, copper may
also enter water from geological
sources. This process may be especially
important in some areas (i.e., Montana)
where copper is naturally abundant.
Another source of coliform injury in
drinking water results from biological
interactions. Some strains of hetero-
trophic or standard plate count (SPC)
bacteria in drinking water have been
shown to suppress coliform detection.
In addition, interactions of SPC bacteria
with coliforms may also cause injury.
For example, experiments showed that
the interaction between Pseudomonas
maltophilia and f. coli produced little
decrease in viable counts over control
levels, but it produced 55% injury over
a 7-day period (Figure 3). The decline in
viable bacteria for the control and
mixed coliform cultures was identical,
but injury was observed only in the
mixed culture. Injury did not occur until
P. maltophilia levels reached 104to 105
colony forming units (cfu)/ml. The ratio
of coliforms to P. maltophilia (1:104 to
105) in the above experiments where in-
jurious interactions occurred are similar
to the ratios for coliforms and SPC bac-
teria in drinking water (coliforms = 1 to
10 cfu/100 ml; SPC > 500 cfu/ml) where
coliform suppression has been re-
ported. Attempts to detect bacteriocins
or antibiotic production by overlaying
irradiated SPC colonies with cultures of
four different coliforms (E. coli, Kleb-
siella pneumoniae, Citrobacter freundii
and Enterobacter agglomerans) were
unsuccessful. Although bacteriophage
could not be ruled out, a nutritional in-
teraction between SPC and coliform
bacteria was thought to be more likely.
Such interactions are important, par-
ticularly when samples are transported
or stored or both.
Conclusions
The following conclusions can be
drawn from data obtained from this
study:
1. Injury to indicator bacteria can re-
sult from a variety of sources, including
chlorine and other disinfectants, heat,
freezing, sunlight, pH, transition metals,
antagonistic plate count organisms,
and possibly other undefined chemical
and physical parameters. Injured bac-
teria exist in every aquatic environment
examined thus far.
2. A number of factors can influence
the extent or severity of injury, includ-
ing time and temperature of exposure,
strain of organism, physiological state,
concentration of nutrients, neutraliza-
tion technique (e.g., EDTA to neutralize
copper effects), diluent, and choice of
sample methodology.
3. The use of m-T7 medium provides
an effective alternative method that is
10
8 '•
a
10*
s
v*.
o
10
P. maltophilia
in Mixed Culture
'. maltophilia
/ 3 5 7
Days
Figure 3. Interaction between Pseudo-
monas maltophilia and E. coli.
Number represents percent
injury in E. coli population. No
injury was observed in the
control culture.
-------
specifically designed for the enumera-
tion of injured total and fecal conforms.
Recommendations
1. Water operators and managers
should become aware of the existence
[reality] of injured conforms. Injured in-
dicator bacteria exist in every aquatic
environment studied thus far, and these
organisms may represent about half of
the coliforms present.
2. Laboratories examining potable
water should perform parallel conform
analyses using m-Endo and m-T7 agars
to assess the occurrence and extent of
injured coliforms in their systems.
3. Epidemiological investigations
should be initiated to evaluate further
the significance of stressed indicator
bacteria associated with waterborne
outbreaks.
Other Reports Based On This
Research
McFeters, G.A., S.C. Cameron and
M.W. LeChevallier. 1982. Influence of
diluents, media and membrane filters
on the detection of injured waterborne
coliform bacteria. Appl. Environ. Micro-
biol. 43:97-103.
LeChevallier, M.W., S.C. Cameron
and G.A. McFeters. 1983. New medium
for the improved recovery of coliform
bacteria from drinking water. Appl. En-
viron. Microbiol. 45:484-492.
LeChevallier, M.W., S.C. Cameron
and G.A. McFeters. 1982. A new
medium for the improved recovery of
injured coliforms from drinking water.
Abstracts of the meeting of the Amer.
Soc. for Microbiol., p. 218.
LeChevallier, M.W., S.C. Cameron
and G.A. McFeters. 1983. Comparison
of verification procedures for the mem-
brane filter total coliform technique.
Appl. Environ. Microbiol. 45:1126-1128.
Domek, M.J., M.W. LeChevallier and
G.A. McFeters. 1983. Evidence for the
role of heavy metals in the injury pro-
cess of coliforms in drinking water.
Abstracts of the meeting of the Amer.
Soc. for Microbiol., p. 261.
LeChevallier, M.W., S.C. Cameron
and G.A. McFeters. 1983. Comparison
of verification procedures for the mem-
brane filter total coliform technique.
Abstracts of the meeting of the Amer.
Soc. for Microbiol., p. 236.
LeChevallier, M.W., P.E. Jakanoski,
A.K. Camper and G.A. McFeters. 1984.
Evaluation of m-T7 agar as a fecal coli-
form medium. Appl. Environ. Micro-
biol., submitted.
Domek, M.J., M.W. LeChevallier, S.C.
Cameron and G.A. McFeters. 1984. Evi-
dence for the role of copper in the injury
process of coliforms in drinking water.
Appl. Environ. Microbiol., submitted.
LeChevallier, M.W. and G.A. McFet-
ers. 1984. Causes, implications and
methods for enumeration of injured
coliforms in drinking water. J. Amer.
Water Works Assoc., submitted.
The full report was submitted in ful-
fillment of Grant No. R807092 by Mon-
tana State University under the spon-
sorship of the U.S. Environmental Pro-
tection Agency.
Gordon A. McFeters, Mark W. LeChevallier. and Matthew J. Domek are with
Montana State University. Bozeman. MT 59717.
Donald J. Reasoner is the EPA Project Officer (see below).
The complete report, entitled "Injury and the Improved Recovery of Coliform
Bacteria in Drinking Water." (Order No. PB 85-117 240; Cost $13.00, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
* U.S.OOVBWUCNTPnNTMQOn'ICeiMB- 559-111/10757
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAI
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
PS 0000329
U i» tNVlR PROTECTION
KEblUN 5 LJbRARr
230 i> UEAR00RN STKEET
CMiCAGO 1U 60604
------- |