:">-
                    United States
                    Environmental Protection
                    Agency
 Water Engineering
 Research Laboratory
 Cincinnati OH 45268
•<.•,-
                    Research and Development
 EPA/600/S2-85/061  Aug. 1985
<&ER&          Project  Summary
                    Enumeration  and
                    Identification  of Heterotrophic
                    Bacteria from Drinking Water
                    James T. Staley
                     Various spread-plating enumeration
                    media and procedures were tested to
                    determine the method of choice for
                    enumerating the highest  numbers of
                    heterotrophic bacteria from chlorinated
                    drinking waters. Dilute media, includ-
                    ing a casemate peptone starch medi-
                    um, a dilute peptone medium, and R2A
                    medium provided greater recoveries
                    than the standard plate count medium
                    currently used. In addition, reduced
                    temperatures of 20°C and prolonged in-
                    cubation periods of 14 to 28 days re-,
                    suited in the highest recoveries of het-
                    erotrophic bacteria from the waters of
                    the two chlorinated distribution sys-
                    tems examined.
                     In the Seattle, Washington, water
                    treatment and distribution system,
                    unchlorinated source water had higher
                    diversities of heterotrophic bacteria
                    than did chlorinated drinking water
                    samples.  The predominant types re-
                    covered from chlorinated drinking
                    waters of that system were  Gram-
                    negative pigmented bacteria. Most of
                    these could not be readily identified to
                    known genera. Some of these could be
                    placed in the poorly described genus
                    Flavobacterium, but the great variety
                    of types of pigmented Gram-negative
                    bacteria included many that could not
                    be readily classified in this group with-
                    out further studies including DNA  hy-
                    bridization.
                     A number of representatives of
                    known genera were encountered
                   among  the isolates from the chlori-
                   nated drinking  waters. Included were
                   members of the genera Caulobacter,
                   Hyphomonas, Aquaspirillum, Vibrio,
                   Gluconobacter,  Azomonas, and
                   Aeromonas.
  This Project Summary was devel-
 oped by EPA's Water Engineering Re-
 search Laboratory, Cincinnati, OH, to
 announce key findings of the research
 project that is fully documented in a
 separate report of the same title (see
 Project Report ordering information at,
 back).


 Introduction
  Water is universally consumed in
 large quantities by the public. Thus
 from a public health perspective, it
 would be desirable to know the types
 and numbers of bacteria that are in-
 gested by drinking chlorinated water.
 Yet numerous  investigations have
 shown that the standard method for
 enumerating heterotrophic bacteria
 from drinking waters (i.e., the standard
 plate count or  SPC procedure) fre-
 quently provides lower counts of bacte-
 ria from drinking waters than other pro-
 cedures. Furthermore, no adequate
 procedure currently exists for character-
 izing and identifying heterotrophic bac-
 teria from drinking waters.
  The objectives of this investigation
 were twofold: a) to evaluate alternative
 procedures for enumerating het-
 erotrophic bacteria from chlorinated
 water supplies and  b) to develop a pro-
 cedure that could be used to identify the
 bacteria that commonly occur in drink-
 ing waters.
  The City of Seattle water distribution
system was used as a test system for
the study of enumeration procedures.
Various media and methods were com-
 pared with the SPC procedure to pro-
vide an optimum count of  the viable
 heterotrophic bacteria.

-------
  About 100 bacterial strains (50  iso-
lates each from  the  Cincinnati, Ohio,
and Seattle, Washington, water sys-
tems) were studied to develop a schema
for use in characterizing and identifying
the  predominant bacteria  from  treated
distribution  water. Morphologic  and
biochemical tests were used to  charac-
terize the isolates.

Materials and Methods
  Drinking  water samples from  the
Seattle water system  were collected in
sterile sample bags or bottles contain-
ing  sodium  thiosulfate to inactivate
residual chlorine, as recommended by
Standard Methods for the  Examination
of Water and Wastewater (15th  Ed.,
American  Public Health  Association,
1980).  Samples  were iced until pro-
cessed, and all samples were analyzed
within 8 hr of collection.
  Enumeration of heterotrophic bacte-
ria was done by the spread plate proce-
dure using several media, including
Standard Methods  agar (SMA), dilute
peptone agar (DP), caseinate peptone
starch  agar (CPS),  and R2A medium.
The pour plate procedure was also used
for some SMA plates.
  Strain diversity in water  samples  was
determined  based on colony  type.  The
Shannon index was calculated as an es-
timate  of strain  diversity from the  for-
mula:
H' =
  where
                 3.3 log10ni/N
    H' = diversity index value
    N =total number of colonies
        counted
    ni = number of colonies of one
        type (i.e., ith strain)
     s = number of strains

  Direct counts were made microscopi-
cally using acridine orange (AO) stain-
ing and an epifluorescence microscope..
  Bacterial strains used for the charac-
terization and  identification studies
were  isolated from winter and summer
water samples  from both the Seattle
and  Cincinnati distribution  systems.
Seattle isolates were obtained from DP
plates, and Cincinnati isolates were se-
lected from R2A plates. Characterization
tests  included Gram-stain, colonial and
cellular morphology, anaerobic growth,
oxidase test, growth on mineral media
with  carbon sources, oxidation/
fermentation reaction, polyhydroxybu-
tarate incorporation,  spore formation,
growth at pH 4.5, acid production from
ethanol, nitrification, cellulose degrada-
tion, pigmentation or fluorescence,
growth temperature, and other special
media tests as required. Media were in-
cubated at 20°C unless otherwise speci-
fied. Known baterial  strains were ob-
tained for comparison  from  other
investigators and from the American
Type Culture Collection.

Experimental Results

Enumeration of Viable
Heterotrophic Bacteria from
Chlorinated Drinking  Water
  Water samples were collected and ex-
amined from 22 sites  in the Seattle
water distribution system, representing
both pre-  and  post-chlorination loca-
tions.
  All four media were used to examine
samples from three surveys. The results
showed that the higest counts were ob-
tained after 14 to 28 days of incubation,
depending on  the sample. However,
counts from SMA plates  incubated at
35°C for 48 hr were  never as high as
those obtained by other means,  and
20°C SMA counts were almost always
inferior to those obtained with other
media used in the surveys. Overall, the
best medium was DP, followed by  CPS
and R2A. DP and CPS consistently  pro-
vided the highest counts for chlorinated
water samples, but DP was less satisfac-
tory than CPS for enumerating bacteria
in water samples collected before chlo-
rine treatment. Because of the superior
performance of DP  and CPS, these
media were used for subsequent viable
plate count analyses.
  The length and temperature of incu-
bation strongly affected  viable plate
counts. After 14 days  of incubation,
most chlorinated samples showed vi-
able counts at 20°C that were 50% to
100% of the counts obtained after 28 to
30 days at 20°C. Incubation beyond 28 to
30 days at 20°C infrequently showed
significant increases in plate counts.
Bacterial counts on DP  and CPS media
were compared after  plates were incu-
bated at 20°, 30°, and  35°, and at a  dual
temperature—20° for 48 hr, then at 30°C
for  up to 30 days (20° -> 30°C). Ranked
from highest to lowest counts, the re-
sults were 20° > (20° -H>  30°) > 30°C.
  Total direct counts by the AO method
were much higher than viable counts,
as expected. For  unchlorinated water
samples, only about 1% of the total di-
rect count bacteria were cultivable; and
for chlorinated samples, the percentage
of viable organism recovery was even
lower. In some  reservoirs, the ratio of
viable  to AO counts exceeded 10%
(range 11.8% to 42.4%). Such high ratios
indicate that the water was eutrophic
rather than  oligotrophic.
  Diversity  indices showed that  chlori-
nation of the water caused a dramatic
change in bacterial diversity  (Table 1).
Whereas the  unchlorinated source
water contained a variety of  colony
types, including many nonpigmented
ones, the chlorinated water contained
relatively few colony types and most
were pigmented. Since these  results
were supported  by those from an earlier
study in  which  diversity  indices were
calculated  for  both  chlorinated  and
unchlorinated waters, it was concluded
that  the  disinfection and distribution
processes in the Seattle system com-
monly  (and perhaps  always) result in
selection of relatively few  bacterial
species  compared with the source
water.

Characterization and Identi-
fication of Bacteria from
Chlorinated Drinking Waters
  The major purpose of this portion of
the study was to identify the predomi-
nant heterotrophic bacteria from two
different  drinking water systems. The
process of  identification assumes that
the  organisms  being investigated are
known organisms that have been previ-
ously characterized,  described, and
named. We anticipated that most bacte-
ria found in chlorinated drinking waters
would  be strains of known organisms
and that characterization of the isolates
and  their comparison with known  or-
ganisms would be rather straightfor-
ward. Such was not the case, however.
Because  of the  slow growth of the iso-
lates and their general nonreactivity in
biochemical characterization  media,
they posed  special problems for charac-
terization. Tests were developed  to per-
mit phenotypic characterization and de-
scription of these  bacteria,  and a
dichotomous flow chart for characteri-
zation was prepared (Figure 1). Many of
the bacterial isolates were not identifi-
able to known  species, but use of the
dichotomous key permitted grouping of
similar types of organisms.
  Most of the bacterial isolates exam-
ined were Gram-negative rods, some of
which were morphologically distinctive.

-------
Table 1.
Diversity Indices of Source  Water (CPRI) and Chlorinated Samples Collected
 September 2, 1981, and Counted after 28 and 40 days
                    After 28 Days
                                             After 40 Days
Sample
CPRI
13
K3
L3
M3
No. Bacteria/ml
3,600
210
1,700
2,900
150,000
Diversity Index*
4.14
1.85
2.81
1.23
1.08
No. Bacteria/mlt
-
570
-
3,070
160,000
Diversity Indext
-
2.45
-
1.45
1.55
*The diversity index is based on 55 colonies, except for sample M3, in which 40 colonies
 were used.
tForty-day counts were determined from 28-day plates that had been refrigerated. In-
 creased counts on plates were due to growth of Hyphomonas.
^Diversity index is based on Hyphomonas colonies. Total colonies counted for 13 were 81;
 L3 were 58; and M3 were 44.
                        such as prosthecate bacteria. Prosthe-
                        cate  bacteria  identified included
                        Caulobacter and strains  recognized as
                        belonging to the Hyphomonas—
                        Hyphomicrobium group. In some sam-
                        ples from the  Seattle system, Hy-
                        phomonas constituted from 2.8% to
                        12.8% of the viable count. Other mor-
                        phologically distinct isolates  included
                        a species of  Vibrio, three spiral-
                        shaped organisms identified as
                        Aquaspirillum strains,  and three large,
                        ovoid,  motile cells identified as
                        Azomonas.
                          The majority of the isolates (44) were
                        Gram-negative,  nonfermentative, oxi-
                        dase positive, aerobic  rods, and 24 of
                        these were pigmented. Of the nonpig-
                        mented group,  only  one  isolate  was
                        identified as a typical Pseudomonas
                                     Phase Morphology
                                     Gram Stain (102)
Gram Negath
Anaerobi
-(•
Oxii
te Rods (72)
c Growth
17)

-------
strain, although several isolates resem-
bled Pseudomonas species  in all re-
spects except for their inability to grow
on mineral medium with acetate as the
carbon source. About half of the pig-
mented  Gram-negative rods  (24) were
unable to grow at 37°C and  7 did not
grow at 30°C. Many did not grow suffi-
ciently well to permit thorough charac-
terization. All of the pigmented strains
were negative for cellulose digestion,
and none survived the  heat survival
tests for sporeformers.
  A few Gram-positive and some Gram-
variable isolates were obtained. Of the
five Gram-positive cocci, three were
identified as Staphylococcus, one was
Micrococcus luteus, and one was possi-
bly a strain of Planococcus. One of the
Gram-positive rods was  identifiable to
genus; its characteristics were typical of
the genus Gluconobacter. The remain-
ing four Gram-positive rods and two
Gram-variable rods were not identifi-
able.  All  but one were pigmented
strains.
  Because  the  pigmented Gram-
negative rods made  up an especially
large group of the isolates and were
well represented in  both distribution
systems, much effort was made to iden-
tify them. Because of the difficulty in
characterizing those isolates, however,
identification was not successful. Many
of these isolates could possibly be clas-
sified as members of the  genus
Flavobacterium, which is a broadly de-
fined genus that encompasses virtually
all pigmented Gram-negative rods. So
broadly  is it defined, however, that it is
known as a wastebasket genus. The pig-
mented  isolates  from this  study were
not identified as Flavobacterium for
several reasons. First, there was no con-
vincing  evidence that the pigmented
distribution  system isolates were re-
lated to  those that are currently placed
in the Flavobacterium genus. Second,
some  nonpigmented strains closely re-
semble  the  pigmented strains except
for the pigmentation.  Thus, until both
the pigmented  and nonpigmented
strains  are  better characterized and
studied, it seemed  inappropriate to
make  an arbitrary assignment on the
basis  of only one characteristic, pig-
mentation.  The third  reason was that
the group of pigmented isolates found
in this study is a very diverse group, and
because of the wide range of pheno-
typic characteristics, it is doubtful that
all pigmented isolates could  be placed
in a single genus.  Even the genus
Flavobacterium  as  currently defined
does not contain organisms that
demonstrate the diversity shown by the
pigmented  bacterial isolates found in
this study.

Conclusions
  A variety of media were tested to de-
termine which provided the highest vi-
able counts of heterotrophic bacteria in
chlorinated drinking waters from the
City  of Seattle water distribution  sys-
tem. The most dilute medium  tested
consistently provided the highest recov-
eries of heterotrophic bacteria from the
chlorinated water  samples examined
during this study.
  The  highest viable counts of  het-
erotrophic bacteria from the chlorinated
samples were obtained when (1) plates
were incubated at  20°C rather than at
30° or 35°C, and (2) incubation time was
longer than 14 days. Use of dual-
temperature incubation (20°C for 48 hr
followed by 30°C for the remainder of
the incubation period) provided  higher
counts than did incubation at 30°C but
lower counts than  incubation at 20°C.
Twenty-eight days of incubation  nor-
mally gave maximal recovery at 20°C,
but some bacteria did not appear  until
after about 40 days.
  Total direct microscopic  count using
AO generally decreased following chlo-
rination. Viable count results indicate
that  only about 1% of the total bacteria
can be cultivated on laboratory media.
  The diversity of heterotrophic bacte-
ria on  each enumeration medium was
determined using the Shannon diver-
sity  index  formula. Chlorination re-
sulted in a  diversity index decrease
(1.08 to 2.81) compared with values be-
fore chlorination (>4.0).
  A total of 102 bacterial isolates were
obtained from the two treated munici-
pal drinking water systems in Seattle
and  Cincinnati.  Of about 50 isolates
from each  system,  half were isolated
during the summer months and the
other half during the winter months.
  The bacterial isolates were character-
ized and identified,  where possible, to
known genera using phenotypic tests.
Seventeen of the isolates were morpho-
logically distinctive  and  included
Caulobacter (9 strains), Hyphomonas
(1 strain), Aquaspirillum (3 strains), a
black  pigmented strain of  Vibrio
(1 strain), and Azomonas (3 strains).
This is apparently  the first report of
Caulobacter and Hyphomonas in chlori-
nated drinking water. Among a group
(14  cultures) of Gram-positive and
Gram-variable nonfermentative rods.
only one was identified, a strain of Glu-
conobacter.
  Most of the isolates were Gram-
negative, obligately aerobic rods. Only
one was clearly identifiable as a strain
of Pseudomonas, although several
other isolates were motile and nonpig-
mented and shared many features of
the genus. Many of the Gram-negative
rods could be superficially  classified in
the genus Flavobacterium because of
their pigmentation and other features.
However, these  organisms showed
greater diversity than even that poorly
defined genus. Further study, including
DNA hybridization, is needed to deter-
mine whether any of the isolates resem-
ble  known  strains of the genus
Flavobacterium closely enough to be
classified as  such.
  Only a few of the Gram-negative rods
were facultative anaerobes, some of
which were  identified  as strains of
Aeromonas hydrophila.
  The full report was submitted in fulfill-
ment of Cooperative  Agreement No.
CR-807570 by the University of Wash-
ington under the sponsorship of the
U.S. Environmental Protection Agency.
                                                                                   US.QOVEHNMENT PRINTING OFFICE 1W5 559-111/20634

-------
    James T. Staley is with University of Washington. Seattle. WA 98195.
    Donald J. Reasoner is the EPA Project Officer (see below).
    The complete report, entitled "Enumeration and Identification of Heterotrophic
      Bacteria from Drinking Water," (Order No. PB 85-207 496/AS; Cost: $11.50.
      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
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

EPA/600/S2-85/061
                    0169064   HERL
                                                    60604

-------