APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 1979, p. 704-714 0099-2240/79/04-0704/11$02.00/0
Vol. 37, No.4
Comparative Study of the Aerobic, Heterotrophic Bacterial Flora of Chesapeake Bay and Tokyo Bay B. AUSTIN,' S. GARGES,' B. CONRAD,' E. E. HARDING,' R. R. COLWELL,`* U. SIMIDU,2 AND N. TAGA2 Department of Microbiology, University of Maryland, College Park, Maryland 207421 and Ocean Research Institute, University of Tokyo, Nakano, Tokyo 164, Japan2 Received for publication 24 January 1979
A comparative study of the bacterial flora of the water of Chesapeake Bay and Tokyo Bay was undertaken to assess similarities and differences between the autochthonous flora of the two geographical sites and to test the hypothesis that, given similarities in environmental parameters, similar bacterial populations will be found, despite extreme geographic distance between locations. A total of 195 aerobic, heterotrophic bacterial strains isolated from Chesapeake Bay and Tokyo Bay water were examined for 115 biochemical, cultural, morphological, nutritional, and physiological characters. The data were analyzed by the methods of numerical taxonomy. From sorted similarity matrices, 77% of the isolates could be grouped into 30 phena and presumptively identified as Acinetobacter-Moraxella, Caulobacter, coryneforms, Pseudomonas, and Vibrio spp. Vibrio and Acinetobacter species were found to be common in the estuarine waters of Chesapeake Bay, whereas Acinetobacter-Moraxella and Caulobacter predominated in Tokyo Bay waters, at the sites sampled in the study.
With the increased interest in the microbial ecology of aquatic environments in recent years, it is important to have an understanding of the natural, or autochthonous, microbial flora, not only in terms of biomass and potential activity, but also community structure and species composition. Since the bacteria are well-known agents of mineralization and transformation of organic and inorganic matter in bays and estuaries, it was considered useful to determine whether bacteria in the water column in two of the major bays of the world supported a similar bacterial flora. Under the auspices of a cooperative research program between the University of Maryland and the University of Tokyo, a study was initiated in 1974 in which water samples were collected using standard methods at sites selected for similarity in environmental parameters, and in which the methods of bacteriological analysis were rigidly conforming to permit minimization of variation ascribable to other than natural floristic differences. The pure cultures obtained in the two parallel studies were analyzed by previously agreed-upon procedures in the two laboratories, with exchange of reference cultures, as well as fresh isolates, so that internal quality control could be maintained in the analyses. Although there has been rapid improvement 704
in the taxonomy of aquatic bacteria since the advent of numerical taxonomy and subsequent pioneering studies of Pfister and Burkholder (31) and Quigley and Colwell (34) and the more recent work of Bauman and associates (3, 4), Delabr6 et al. (12), and Reichelt and Baumann (35), this information has been of limited value because of the relatively few organisms examined, the restricted number of tests used in identification of isolates, and the limited geographical distribution of the organisms studied. Pfister and Burkholder (31) attempted to compare the bacterial microflora of tropical and antarctic seawater, a type of broadly defined study that is essential for proper assessment of the distribution of bacteria in the marine environment. Baumann et al. (3, 4) restricted their studies to a set of selected marine isolates, for the most part from culture collections of various investigators, although some fresh isolates were included in their analyses. In the study described here, aerobic, heterotrophic, estuarine bacteria representative of the microflora of Chesapeake Bay and Tokyo Bay waters were compared by numerical taxonomy and molecular genetic methods, providing new information concerning the species composition of microbial communities present in the water column at selected sites in two major estuaries of the world.
705
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 MATERIALS AND METHODS
umes
Isolation and maintenance of strains. Stations selected from track lines routinely surveyed by the University of Maryland in Chesapeake Bay and by the Ocean Research Institute, University of Tokyo. The stations selected were the two that were the most similar in physical and chemical characteristics. Selection was on the basis of information available to the respective laboratories for the stations over the past 15 years. Field studies designed specifically to select locations with the same salinity, temperature, etc., would have been preferable, but neither time nor funds permitted such an exhaustive search for sites exhibiting precisely identical parameters. In any event, an estuarine system is dynamic, and salinity, temperature, and other parameters are not constant throughout the year and from year to year. The stations selected, on average, were the most similar, over time, based on respective laboratory records. Thus, the locations, time of year, salinity, depth, amount of pollution, tidal cycle, time of day, etc., were carefully considered, and these factors were as closely similar as circumstances permitted. Water samples were collected on 19 June 1975 from Cape Charles in Chesapeake Bay and on 4 July 1975 from Tokyo Bay during cruises aboard research vessels that had been scheduled a year in advance, the dates being the closest in time that the respective schedules permitted. Cape Charles is a high-salinity station in Chesapeake Bay, with a surface-to-sediment depth of 29 m, situated approximately 32 km north of Norfolk, Va. and 26 km northeast of the James River. The sampling station in Tokyo Bay, Japan, with a depth of 17 m, is situated 2 km west of the Obitsu River. Physical and chemical characteristics of the stations in Chesapeake Bay and Tokyo Bay are listed in Table 1. Water samples were collected aseptically, using a Niskin Bag Sampler (General Oceanics Inc., Miami, Fla.), from approximately 2 m above the underlying sediment at Cape Charles, since the Chesapeake Bay is a typical "salt water wedge" estuary and the most saline water was sought, and from 1 m below the surface of the water in Tokyo Bay. Triplicate Niskin casts were made, each collecting 2 liters of water to ensure statistical validity of the sampling procedure. Serial dilutions to 10' of the water samples were prepared immediately after collection, using 9-ml vol-
were
of sterile marine salts solution (0.7% [wt/vol]
MgSO4.7H20, 2.38% [wt/vol] NaCl, and 0.07% [wt/ vol] KCI; pH 7.0), the composition of which had been agreed upon by the two laboratory groups. For estimation of total counts of Chesapeake Bay water, 0.1ml aliquots were pipetted onto Simidu medium (41), thiosulfate citrate bile salts agar (TCBS; BioQuest, Cockeysville, Md.), marine 2216 agar (50; Difco), chitin agar, and fish protein medium (0.1% [wt/vol] XM-1 fish protein hydrolysate [Zapata Haynie Corp., Baltimore, Md.], 0.1% [wt/vol] Difco yeast extract, 1.5% [wt/vol] Difco agar, and 1 liter of marine salts solution). The Tokyo Bay water samples were plated on all of the above except the fish protein medium which was included for comparative purposes and comprised part of another study carried out by the University of Maryland laboratory. All plates were spread using a sterile glass spreader. Three replicates of each dilution were plated, and inoculated plates were incubated at 25°C, approximating the in situ temperature, for 14 days. Cultures were randomly selected from platings of each of the three water samples on the media used in the study. Thus, from each of the water samples from the three separate casts made at the sampling site, approximately the same number of cultures were randomly picked from plates of each of the media employed and from the same dilution, containing between 30 and 300 colonies. The cultures were subsequently purified, after incubation at 25°C, by streaking and restreaking three times on plates of 2216 agar. Heatfixed smears from 24-h cultures were stained using Hucker's modification of the Gram stain (16) and examined microscopically. When the cultures were considered to be pure, they were inoculated onto slopes of marine 2216 agar. Stock cultures were maintained under sterile mineral oil on 2216 agar slopes at 25°C and subcultured every 6 to 8 weeks. After all of the isolation and purification steps had been followed, a total of 195 isolates, approximately equally representing the Japanese and U.S. samples, were examined for the biochemical, cultural, morphological, nutritional, and physiological characteristics included in the analyses. Reference strains. In addition to the fresh isolates from Chesapeake Bay and Tokyo Bay, 18 reference cultures were also included in the study (Table 2). Characterization of the strains. Each strain was
TABLE 1. Physical, chemical, and microbiological characteristics of the sampling stations Total counts' ency of Fish Chitin water, Temp Marin Simdu measured (0C) Marine FCSprtish medium from surmedium agar agar ga face (in)agrmeim 3.3 25.7 2.0 x 104 8.4 x 101 6.3 x 102 4.2 x 103 8.3 x 102
TransparDissolved
Depth of water Salinity Station location Sample type sample M%O (m)
oxygen (mg/ liter)
Water
27
25.8
7.4
2 km west of Ob- Water itsu River, Tokyo Bay
11
30.1
7.5
Cape Charles, Chesapeake Bay
2.2
21.3 9.1 X 104 3.3 x
10' 1.8 x 10'
_b
6.7 x 101
Colony-forming units, i.e., total viable, aerobic, heterotrophic bacterial counts on the respective media employed.
bNot done.
706
AUSTIN ET AL.
APPL. ENVIRON. MICROBIOL.
TABLE 2. Reference cultures included in the numerical taxonomy analyses Laboratory Culture reference Reference culture collection Comment/isolated from no. no. Acinetobacter lwoffia Acinetobacter iwoffi
R12 R13
Aeromonas salmonicidaa Arthrobacter crystallopoietes Arthrobacter marinus
R8 R6 Rl
ATCC 15309 Univ. of Maryland Culture Collection ATCC 14174 ATCC 15481 ATCC 25374
Corynebacterium poinsettiae Erwinia herbicola Pseudomonas aureofaciens' Pseudomonas cepacia
R9 R18 R24 R25
ATCC 9069 ATCC 12887 ATCC 13985 ATCC 17759
Pseudomonas maltophiliaa
R20
ATCC 13637
Pseudomonas mendocina' Pseudomonas putida" Pseudomonas stutzeria Staphylococcus saprophyticus" Vibrio anguillarum Vibrio anguillarum" Vibrio parahaemolyticus' Xanthomonas begoniae
R26 R7 R27 R17 R15 R14 R16 R21
ATCC 25411 ATCC 12633 ATCC 17588 ATCC 15305 ATCC 14181 ATCC 19264 ATCC 17802 ATCC 11725
Diseased brook trout Soil Renamed Pseudomonas marina a; seawater
a Denotes type strain, obtained from the American Type Culture Collection
examined for a total of 115 unit characters. Whenever possible, marine 2216 agar was used as the basal medium, and, in all cases, media were prepared using marine salts solution as the diluent. Unless otherwise stated, all inoculated media were incubated at 25°C for 14 days. To assess possible test error, 20 randomly chosen strains were examined in duplicate; otherwise, tests were repeated only in the event of inconclusive results. For inter-laboratory quality control, a set of fresh isolates was exchanged and included along with the reference cultures in both laboratory testing regimes. Most of the tests have been described elsewhere (9, 14, 19, 25, 40); details of the remainder are given below. Utilization of substrates as the sole source of carbon for energy and growth. Carbon utilization tests were carried out using the medium devised by Stevenson (47). The carbon compounds included aDL-alanine, L-arginine hydrochloride, asparagine, Lhistidine hydrochloride, L-leucine, L-lysine, L-methionine, DL-phenylalanine, L-proline, sodium acetate, sodium adipate, sodium citrate, sodium formate, sodium gluconate, sodium glutamate, sodium glutarate, sodium lactate, sodium malonate, sodium oxalate, sodium tartrate, and L-threonine. These were sterilized as 20% (wt/vol) solutions by tyndallizing for 1 h on 3 consecutive days at 100°C, or by filtration, and added to the supporting medium to give a final concentration of 0.2% (wt/vol). This medium was dispensed into divided replidishes (44) and inoculated by means of a multipoint inoculator (23). After incubation, the medium was examined visually at 14 days, recording the presence or absence of growth. Determination of the G+C content of DNA. The guanine plus cytosine content (G+C content, moles percent) of purified deoxyribonucleic acid (DNA), obtained by the method of Marmur (26), was
River clay Renamed Pseudomonas multivoransa; forest soil Oropharynx of patient with mouth cancer Soil Degrades aromatic acid Spinal fluid Urine Ulcerous lesion in cod Food poisoning, Japan Semituberous Begonia spp.
(Rockville, Md.).
determined from the thermal denaturation temperature (Tm) of the DNA using a Gilford 2400-S recording spectrophotometer at 260 nm (27). The DNA preparations were dialyzed to lx SSC (0.15 M NaCl, 0.015 M sodium citrate) and heated at 0.5°C min-', during which the optical density was monitored at 260 nm. Molar percentage of G+C was calculated from the Tm by the equation of De Ley (13). Coding of data. The characters were coded 1 for positive or present, 0 for negative or absent, and 9 for noncomparable or not applicable. The final n x t matrix contained 213 strains, including reference strains, and 115 characters recorded for all strains. Computer analyses. The data were analyzed using SsM, the simple matching coefficient (45), which included positive and negative matches, and the Jaccard coefficient, Sj, which excludes negative matches (43). Clustering was by unweighted average linkage (45), and sorted similarities and dendrograms were constructed. The hypothetical median organism (22) for each cluster was also calculated. Programs employed included UMDTAXON3 and IGPS3 program packages available on the University of Maryland UNIVAC 1108 computer.
RESULTS Total viable bacterial counts. Four media were employed in the Chesapeake Bay-Tokyo Bay comparative study: marine 2216 agar (50); a medium designed to enhance growth of marine vibrios (41); chitin agar; and TCBS, originally developed for isolation of Vibrio cholerae and subsequently employed for isolation of Vibrio parahaemolyticus. Counts obtained on these
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979
media are given in Table 1. Analysis of the results obtained for each of the three casts showed no statistically significant differences in total viable aerobic, heterotrophic bacterial counts. That is, there was no significant difference in counts for the three water samples that were collected at the same time to test for sample variation. However, it can be seen that higher counts on all media except 2216 agar were obtained for the Chesapeake Bay water samples. The count obtained on TCBS agar was 10-fold greater for the Chesapeake Bay water sample and 2-fold greater for the Simidu marine vibrio medium. This difference is important in view of the results (see below) showing a greater preponderance of vibrios in Chesapeake Bay water samples. Clustering of the strains. Results of analyses employing the SSM and SJ coefficients included 154 of the environmental isolates, representing 77% of the total set, and the six reference cultures of Acinetobacter Iwoffi, Pseudomonas maltophilia, Pseudomonas marina, Vibrio anguillarum (two strains), and V. parahaemolyticus in 30 clusters at 270% similarity (S). Results PERCENTAGE SIMILARITY
707
of the analysis using the SJ coefficient are given in Fig. 1. Four clusters (phena 1, 2, 21, and 25) comprised 65 strains, including the reference culture of V. parahaemolyticus, with the remaining 26 groups containing between 2 and 9 strains each. Identification and description of the isolates. Presumptive identification of the 30 clusters was accomplished either by inclusion of a reference culture within the cluster or by consultation of diagnostic keys in Bergey's Manual of Determinative Bacteriology (8) and specialist keys and tables including those of Shewan et al. (38), Stanier et al. (46), Bousfield (7), Jones (17), Bianchi (M. Bianchi, these de doctorat des sciences naturelles, Universite d'Aix-Marseille, 1976), Otto and Pickett (29), and Pagel and Seyfried (30). Six generic groups comprised the major taxa represented, including Vibrio spp., Acinetobacter-Moraxella spp., Pseudomonas spp., Caulobacter spp., coryneforms, and groupings of as yet unidentified gram-negative rods (see Fig. 1). A summary of selected characteristics of the six generic groupings is given in Table 3. PHENON
NO. OF STRAINS
SOURCE
IDENTITY
16 17 18 19 20
3 3 2 5 5 2 4 2 9
CB CB CB CB CB CB CB CB CB CB CB CB CB C,B C.B/TB. CB/TB TB. TB TB. TB
21 22 23 24
22 2 2 3
TB TB TB TB
25 26 27 28 29 30
21
TB
Coulobocter sp. Gram negotive rods Grom negative rods Grom negative rods Gram negotive rods
5 2 2
TB TB TB TB. TB
Grom negotive rods P pu/bdo Gram negative rods Gram negative rods Gram negotive rods
2 3 4 5 6 7 8 9 10 I1 12 13 14 15
12 10 6 2 2 2 3 5 9 4 7
4
6
Vibrio porohoerno/yt/cus V f,scheri V ongul//orum Vibrio sp. Vibr,o sp V,br,o sp
Aclnetobocter Gram negotive Gram negative Gram negative P morino P mo/ltophi/io Grom negative
co/coocet/cus rods rods rods rods
Coryneforms" Gram negative rods Gram negative rods Aclnetobocter - Aoroxe/lo sp. Aclnetobocter - Moroxe/lo sp Pseudolonos sp. Grom negative rods
FIG. 1. Simplified dendrogram prepared with the SJ coefficient and unweighted average linkage clustering technique. C.B., Chesapeake Bay; T.B., Tokyo Bay.
708
APPL. ENVIRON. MICROBIOL.
AUSTIN ET AL. 3
C14
cq
3
>+1I
3
>+I
3
CD
I1
I
la00
-I + +-I
la00 +0
+
+0
3
+0
cq
3
eq
CD
cs
10
3
10
+
>
C4
3
0
-r
>> +r +
eq
IIIII II
IlIlIl
+
+ I +++I
III
IIIIII
+
II
11-111 II II I I
II
+
I,I I IIIII
I I + +-I
+
1+-ol
II
IlIlIlIl II
+
+I++II
II
I+
+
I1I1I+
I+1111 I1 I1
+
II
III ++I
III+I+I++
+
I
I
+II0II1
+
+a
.-Is
cs
3
l
3+
+0
0
_14
I-o +-I
1
+a
.!
3 -4
.S
I I++I
+ I
3
oq
-4
+0 +a
+
I I++I +I + + I
I0
3
C13
+
IIII I-III+
+ I+
o
+ If
+ ++ +
C CD
+0
+
2
3:
+
+ + + +la
+I++I
IIIlII
I
+
+0 a -01+ ++
+0
+ + +l
0
++
+1I
+11
lll+ lr
I++ I
III
0
+
_t
3 CD
0.0
4
tII
to
+
^ eq
3
u-i
Io~
0
-4
+ rX
+
+ 44
-0l++-01 ~~0 IlIlIllaIll + + II +I+-o+I 0+0
IO +
+
+I +10 I01 l+ 1001 10
+
0
+I
+
C4
+
+ I++ II
+0
+
+ I ++ II
Pk
+
+I+1
+ +
+0
3
+ Pk
CIA
+
+I+-o+++++0I++I o-
+0
3 cli
+
+
+
Ct
I I I I I I I I ++
+
3 3
0+1I1I1I
I + I
+
+
+
+
I
II I + I++ II +
I o
3
++I II
I
+
I1+ I1I 01 I5 3
eq
0LO& -4
3
I+1I
3
ral
o
z
10r
Pk
+I+ +l++
+ rZ
II 1
+IC
+01I -o +
+
+++++I
0++++++I 1+
+
+
+- + + +
Ia 0+ ++ +
+
+
0
0 ++ + +I
I1X+0+ +0
I+
+
l
l
+ Pk
3
I1+I 01+
3
+
3
+
+ Pk
I+ 1
+
I
+ + + + +
l
+ +
+ + ++ ++
.0_
0 03
O0 0
2u
cu+ ZI
Qo
c
S=is t f W t2~~~~~~~~~m E
-
00
c
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 II I'X do
I X X I
+
I1+1I1
+
++ I
+
+'0I.X
1+1I +
+'0+'0
+
+ I +.
I I I'
I'0'0' X1I1 '0
'0'
+
I'0+0I+
+
0
+ + +.
._Y0
0. IrII
+0
I++II
eq 0
.
oo~
@0 v
X
X0
I'X I1
l0
l'll1
0. -*
VI
X
-
X~
c) M~; E
X
Al
4 0
0
4,XE
0 b
709
Phena 1, 3, 7, 11, and 12 were identified as V. parahaemolyticus, V. anguillarum, Acinetobacter calcoaceticus, P. marina, and P. maltophilia, respectively, by the grouping of the reference strains with the isolates. In general, the characteristics of the Vibrio strains matched the species descriptions in Bergey's Manual of Determinative Bacteriology (8) and those given by Bianchi (these de doctorat), although it was observed that the strains identified as V. anguillarum were negative for the indole reaction. Although the reference strain of A. Iwoffi was recovered in phenon 7, A. Iwoffi is now regarded as a synonym of A. calcoaceticus (8, 30). The characteristics of A. calcoaceticus (phenon 7) were in close agreement with the description provided in Bergey's Manual of Determinative Bacteriology (8). Phenon 2 was identified as Vibrio fischeri by micromorphology, biochemical, degradative, and physiological reactions and overall DNA content of 44.8% G+C (Table 3). In contrast to the species description, none of the strains exhibited luminescence (39). However, luminescence is a characteristic that is highly dependent on medium, temperature, and other variables. "Dark" strains are very commonly encountered among luminescent species (8, 39). Phena 4, 5, and 6 are classified as Vibrio spp., viz., gram-negative, curved, fernentative, oxidase- and catalase-positive rods, possessing overall DNA base compositions of 46 and 49% G+C. The strains were unable to grow at temperatures of 37°C or above. The strains isolated in this study and identified as Vibrio spp. did not degrade casein or Tween 80 or grow at temperatures below 100C. In contrast to the genus description (39), it should be noted that none of the strains demonstrated a marked sensitivity to pteridine. Thus, phena 4, 5, and 6 were not identified as any of the species described and named in the literature and should be considered new species. Phena 10, 16, 20, 25, 29, and 30 possessed some ofthe characteristics of Acinetobacterspp. However, the range of overall DNA base composition, 50 to 66% G+C, is higher than the range of values usually associated with this genus, i.e., 40 to 47% (21), and closer to that of Pseudomonas. The oxidase-negative reaction, for the present, excludes these organisms from the genus Pseudomonas. Therefore, these phena are regarded as unidentified until further work can be done on the taxonomy of these organisms. Phena 17 and 18 were tentatively identified as belonging to the Acinetobacter-Moraxella group. Although all strains possessed the general characteristics of Acinetobacter-Moraxella,
710
AUSTIN ET AL.
they showed specific properties of Moraxella, i.e., they were oxidase positive and sensitive to penicillin. It has been suggested that oxidasenegative strains should be reclassified as Acinetobacter (5); this would exclude the strains clustered in these phena. However, with the exception of Moraxella phenylpyruvica, Moraxella species, including M. lacunata, M. bovis, M. nonliquefaciens, and M. osloensis, are normally associated with pathological conditions and have not been reported to occur in the aquatic environment. More importantly, the description of the strains isolated in this study could not be matched with any of the described species of Moraxella or, for that matter, of Acinetobacter. Thus, it was decided to regard the phena as groups intermediate between Acinetobacter and Moraxella until further studies on the classification of these organisms can be completed. The overall DNA base composition was not determined for these organisms despite several attempts, because of problems with harvesting cells and extracting the DNA. It is possible that these two phena are similar to phena 10, 16, 19, 20, 22, 23, 24, 25, 26, 28, 29, and 30, which also exhibited characteristics of Acinetobacter-Moraxella but could not be classified in either of these groups due to the high G+C content of the DNA, the latter being more representative of Pseudomonas spp. Some of the clusters of gramnegative rods that remain unidentified may also be Pseudomonas or Alteromonas spp. However, they are indeed new species, if not new genera, and further, more precise study of the classification, identification, and nomenclature of these strains is in progress. Phenon 14 was identified as coryneform bacteria largely on the basis of the micromorphology of the strains. Unfortunately, the organisms could not be identified further, even after consultation of the comprehensive descriptions of Bousfield (7) and Jones (17), and may also represent a new species. Phena 8, 9, 13, and 15 were not related to bacterial species described in the available literature, i.e., in diagnostic keys and/or in Bergey's Manual of Determinative Bacteriology (8), and will require further analysis for precise classification. Strains clustered in phenon 21 produced stalks and formed "rosettes," hence were identified as Caulobacter spp. The strains were short, oxidative, gram-negative rods when observed in wet mounts under phase-contrast microscopy. They could not be identified to the species level and represent, in our view, a new species of Caulobacter. Phenon 27 was identified as Pseudomonas putida (8, 11). Diagnostic features included
APPL. ENVIRON. MICROBIOL.
gram-negative, oxidative, oxidase-positive rods that were fluorescent under ultraviolet light, produced arginine dihydrolase but not lysine decarboxylase, and utilized L-arginiie hydrochloride and sodium citrate as the sole source of carbon for energy and growth, but did not hydrolyze casein gelatin, starch, or Tween 80. The strains were nonmotile when observed in wet mount under phase-contrast microscopy. DNA/ DNA hybridization studies of these strains with reference strains are in progress. Test error. Twenty strains examined in duplicate were tested on different occasions, but the same methods were used throughout. The majority of tests, i.e., a total of 85, gave identical results; these included the Gram staining reaction, colonial morphology, and the majority of biochemical degradative and nutritional characters. The error was calculated to be approximately 2%, most of which could be attributed to a few characters, notably the pleomorphic nature of many of the isolates and the H25, methyl red, oxidase, Voges-Proskauer, and gelatin tests, which either gave weakly positive, equivocal results or were, in the case of the biochemical test results, contradictory on retesting. The H2S test can be variable, depending on the presence or loss of plasmids (variability due to plasmids has been noted in earlier studies [S. A. Orndorff, B. Austin, L. A. McNicol, and R. R. Colwell, submitted for publication]), and, in the case of the methyl red, Voges-Proskauer, and gelatin tests, there is a strain variability in the intensity of the reactions that has been observed in our own work with strains freshly isolated from the aquatic environment. Variation in test results between laboratories was not significant and therefore permitted combining the data from both laboratories for a comprehensive analysis. Selectivity of the media employed. Only Vibrio spp. were isolated on TCBS agar, including V. parahaemolyticus, V. fischeri, and unidentified Vibrio spp. The Simidu medium was also highly selective for Vibrio spp., with V. parahaemolyticus, V. fischeri, and unidentified Vibrio spp. being isolated. However, two isolates, most probably of the Acinetobacter-Moraxella groups, were also isolated from the Simidu medium. The 2216 medium yielded a variety of organisms, selecting neither for nor against Vibrio spp. It should be pointed out that the fish protein agar was by far the most useful medium in providing growth for a very wide variety of bacterial genera and species, including Vibrio, Moraxella, Acinetobacter, and Pseudomonas. DISCUSSION The objective of this study was to assess sim-
VOL. 37, 1979
DISTRIBUTION OF ESTUARINE BACTERIA
ilarities and differences between microflora of two environmentally similar sites at widely separated geographical locations. That is, the question to be answered is whether the water of estuaries in different parts of the world will support a relatively similar species composition in the microbial community structure. In spite of the usual restraints of time, budget, cruise schedules, and related logistical problems, it was possible to select two sites for comparison that were reasonably similar and for which significant data had been gathered in other studies. Cruise schedules for the research vessels available to the cooperating laboratories fortuitously overlapped so that samples could be collected at the same time of year. It is recognized that a more precise matching of sites could be achieved, but the matching of sites was the best possible considering the logistics of the situation. Furthermore, it is also recognized that biomass measurements should have been made, i.e., direct measurement of the total bacterial population using epifluorescence, Limulus lysate assay, adenosine triphosphate, or others of the more recently developed methods for biomass estimation. Tests of potential activity, using 14C heterotrophic uptake measurements, would also be useful. Nevertheless, the question was that of community structure and species composition, and for this reason the comparative study is considered to provide useful information. It must also be recognized that the isolates examined in this study represent only that portion of the microbial community capable of growth on the media employed. For example, for oligotrophs, low-nutrient media must be employed for isolation and characterization (25). Nevertheless, for the purposes of this study, namely to compare microbial populations under standardized conditions of plating, media, etc., this point does not negate the conclusions drawn from the results obtained. The results show that, from the counts obtained on the media employed and given in Table 1 and from the numerical taxonomy analyses, the predominant aerobic, heterotrophic bacterial flora present in the water column of Chesapeake Bay water, at the site included in this study, consisted of Vibrio, Acinetobacter, Pseudomonas, and coryneforms, whereas the microflora of the water column in Tokyo Bay, at a comparable site, was predominately Acinetobacter-Moraxella-like species, Caulobacter spp., and groups of gram-negative, rodlike bacteria resembling Pseudomonas spp. but sufficiently dissimilar to be considered new species. Variation in species composition among the individual water samples was not observed. Thus the microflora of the two bodies of water appear
711
to be dissimilar at the sites from which samples examined in this study were collected. Vibrio was the numerically dominant genus in Chesapeake Bay, regardless of the medium employed, a finding concurring with conclusions of Lovelace et al. (24), Kaneko and Colwell (18), and Cook and Goldman (10). Vibrios were not predominant in the microbial populations found in the water of Tokyo Bay at the stations included in the study. It has been suggested previously by Simidu et al. (42) that the absence of vibrios in Tokyo Bay is caused by antagonistic interactions between Vibrio spp. and phytoplankton and by organic nutrients in Tokyo Bay. The presence of V. parahaemolyticus in Chesapeake Bay is known to be inversely related to salinity, in that extremes of salinity affect viability (18, 36), and warmer water temperatures are favorable for their growth and distribution during the warmer seasons of the year (18). In this connection, the water at Cape Charles demonstrated comparatively high salinity and temperature, and relatively few V. parahaemolyticus were isolated. Acinetobacter spp., which accounted for the third most predominant taxon at Cape Charles, were tentatively identified as the predominating organisms in Tokyo Bay. Large numbers of Acinetobacter have been isolated from the aquatic environment by other investigators, notably in Lake Ontario (37) and Lake Superior (6). However, organisms placed in this genus no doubt have been misclassified because of the relatively loose description available for the genus. Thus, it can be concluded that Acinetobacter isolated from the aquatic environment are not sufficiently well studied to be readily separated from the other gram-negative, nonmotile, nonpigmented bacteria, a situation also true for Flavobacterium (14). Many of the strains recovered from Tokyo Bay possessed characteristics more appropriate for groups intermediate between Acinetobacter and Moraxella and, in some cases, related to Pseudomonas. Pseudomonas spp., which have been isolated from a wide range of habitats within the aquatic environment (3, 10, 12, 31, 34), were recovered in moderately large numbers from Chesapeake Bay and also from Tokyo Bay. Several of the clusters of gram-negative, rodlike bacteria were very similar to Pseudomonas, except for a tendency to pleomorphism and differences in certain characteristics, such as sensitivity to penicillin, and may ultimately prove to be new spe-
cies of Pseudomonas. It would be very useful to assess the ability of these strains to degrade petroleum and to measure their resistance to heavy metals and other pollutants, since the pseudomonads, together with the coryneforms (which were surprisingly absent from the Tokyo
712
AUSTIN ET AL.
Bay water samples), are associated with petroleum degradation (1, 2) and mobilization of heavy metals (28). The coryneform strains recovered in this study could not be identified to the genus level, reflective of the complexity of the taxonomy of the coryneform group of bacteria and uncertainty of the validity of the classification of those coryneforms found in the marine environment (7, 17, 49). Caulobacter spp. generally occur in waters containing low concentrations of organic nutrients (32, 33). Slow-growing strains of stalked bacteria, identified as Hyphomonas polymorpha, have been isolated in large numbers in other, more polluted areas of Chesapeake Bay (25). The occurrence of stalked bacteria in the estuarine environment would suggest that such organisms may have important ecological functions, and further studies are in progress. The lower Chesapeake Bay, at the station included in this study, receives much less allochthonous input than the upper Chesapeake Bay. It would appear that the differences in species composition noted in this study provide an index of pollution, since the Tokyo Bay station can be considered to be eutrophic and more polluted, in general, than the Chesapeake Bay. Pfister and Burkholder (31) observed a degree of specificity between tropical and antarctic waters in the distribution of bacterial taxa in the aquatic environment. Recent work in our laboratory has demonstrated a geographical distribution among petroleum-degrading bacteria and a specificity of these bacteria for water or sediment in polluted and unpolluted sites in Chesapeake Bay (1, 2). However, the distribution was not as striking for those genera of bacteria capable of degrading petroleum as were the differences for the bacterial taxa comprising the microflora of Chesapeake Bay and Tokyo Bay. An interesting observation arising from this study is that the choice of medium will exert a significant influence on the recovery of bacterial taxa in the marine environment, an observation in agreement with earlier studies of Goulder (15) and Vaatiinen (48). In this study, fish protein agar was used for comparative purposes only in the Chesapeake Bay portion of the study. Although the counts on the fish protein agar were one log less than on 2216 agar, the fish protein agar medium permitted the initial isolation of the widest range of taxa, that is, a greater number of genera and species than the more commonly used marine 2216 agar of ZoBell (50). This observation merits further study. In comparison, the TCBS medium was highly selective for vibrios, as was the Simidu medium, both of which were designed for isolation of Vibrio spp. After isolation and culture in the laboratory,
APPL. ENVIRON. MICROBIOL.
those strains isolated on media other than 2216 agar were able, subsequently, to grow on this medium, indicating an adaptation of the isolates to the richer medium. In a previous study of slow-growing oligotrophic bacteria recovered from the estuarine environment (25), the composition of the isolation medium was found to exert a significant effect upon the recovery of bacterial taxa. In particular, it was shown that Streptothrix and other sheathed bacteria could be recovered only on a medium containing a very low concentration of nutrients, consisting of Chesapeake Bay water to which agar was added. No variation in genera and species was noted between casts. That is, each of the separate casts yielded the same representation of genera and species on each of the media, with significant differences noted only between media. The TCBS and Simidu media were selective for Vibrio spp., and the other media, i.e., chitin, 2216, and fish protein agar media, selected neither for nor against Vibrio spp. Of all the media, however, the fish protein concentrate medium yielded the widest variety of genera and species in the case of the Chesapeake Bay water samples. Clearly, the effects of media composition are important in the recovery of bacterial taxa from the natural environment, and this phenomenon should be carefully considered in future studies of microbial community structure. In conclusion, intrinsic differences were observed in the microflora of Chesapeake Bay and Tokyo Bay water in a given location and season. Space and time dependency of microbial community structure, of course, is well recognized. Nevertheless, the results of this study indicate that it cannot be assumed that estuaries will contain microbial communities similarly comprised of representative taxa, but, rather, it should be recognized that geographical differences may be strongly influential, with the type and concentration of nutrients present, the indigenous plankton populations, and the nature, source, quality, and quantity of allochthonous material entering an estuary acting selectively on the microbial populations. ACKNOWLEDGMEENTS This work was supported by National Science Foundation grants OIP 7417540 and BMS 72-02227-A03. The computer time for the project was supplied, in full, through the facilities of the Computer Science Center at the University of Maryland. The Japanese cooperative research was supported by grant no. 5R064 from the Japan Society for the Promotion of Science. The assistance of the captains and crews of the R/V Ridgely Warfifld and of the R/V Tanseimaru is gratefully acknowledged. The authors gratefully acknowledge the assistance of Jayne Carney and Gary Sayler in the initial stages of this project.
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 LITERATURE CITED
1. Austin, B., R. R. Colwell, J. D. Walker, and J. J. Calomiris. 1977. Numerical taxonomy and ecology of petroleum-degrading bacteria. Appl. Environ. Microbiol. 33:60-68. 2. Austin, B., R. R. Colweli J. D. Walker, and J. J. Calomiris. 1977. The application of numerical taxonomy to the study of petroleum degrading bacteria isolated from the aquatic environment. Dev. Ind. Micro-
biol. 18:685695. 3. Baumann, L, P. Baumann, M. Mandel, and R. D. Allen. 1972. Taxonomy of aerobic marine eubacteria.
J. Bacteriol. 110:402-429. 4. Baumann, P., L. Baumann, and M. Mandel. 1971. Taxonomy of marine bacteria: the genus Beneckea. J.
Bacteriol. 107:268-293. 5. Baumann. P., AL Doudoroff, and R. Y. Stanier. 1968.
A study of the Moraxella group. II. Oxidase-negative species (genus Acinetobacter). J. Bacteriol. 95:15201541. 6. Bennett, E. A. 1969. A study of the distribution of heterotrophic bacteria in the Great Lakes. 1. The heterotrophs of lake water. Bacteriology Branch, Division of Laboratories, Ontario Water Resources Commission, Ontario, Canada. 7. Bousfield, I. J. 1972. A taxonomic study of some coryneform bacteria. J. Gen. Microbiol. 71:441-455. 8. Buchanan, R. E., and N. E. Gibbons (ed.). 1974. Bergey's manual of determinative bacteriology, 8th ed. Williams & Wilkins Co., Baltimore. 9. Colwell, R. R., and W. J. Weibe. 1970. "Core" characteristics for use in classifying aerobic, heterotrophic bacteria by numerical taxonomy. Bull. Ga. Acad. Sci. 18:165-185. 10. Cook, T. M., and C. K. Goldman. 1976. Bacteriology of Chesapeake Bay surface waters. Chesapeake Sci. 17: 40-49. 11. Cowan, S. T. 1974. Cowan and Steel's manual for the identification of medical bacteria. Cambridge University Press, Cambridge. 12. Delabre, AL, A. Bianchi, and M. Veron. 1973. g#tude critique de methodes de taxonomie numerique. Application a une classification de bacteries aquicoles. Ann. Microbiol. 124:489-506. 13. De Ley, J. 1970. Re-examination of the association between melting point, buoyant density, and chemical base composition of deoxyribonucleic acid. J. Bacteriol. 101:738-754. 14. Goodfellow, M., B. Austin, and C. H. Dickinson. 1976. Numerical taxonomy of some yellow pigmented bacteria isolated from plants. J. Gen. Microbiol. 97:219-233. 15. Goulder, R. 1976. Evaluation of media for counting viable bacteria in estuaries. J. Appl. Bacteriol. 41:351-355. 16. Hucker, G. J., and H. J. Conn. 1923. Methods of Gram staining. In Technical Bulletin of the New York State Agricultural Experiment Station, no. 93. 17. Jones, D. 1975. A numerical taxonomic study of coryneforms and related organisms. J. Gen. Microbiol. 87:5796. 18. Kaneko, T., and R. R. Colwell. 1974. Distribution of Vibrio parahaemolyticus and related organisms in the Atlantic Ocean off South Carolina and Georgia. Appl. Microbiol. 28:1009-1017. 19. King, E. O., M. K. Ward, and D. E. Raney. 1954. Two simple media for the demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44:301-307. 20. Lautrop, H. 1974. Moraxella Lwoff 1939, 173 Nom. cons. Opin. 41, Jud. Commn. 1971, 106, p. 433-436. In, R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 21. Lautrop, H. 1974. Acinetobacter Brisou and Prevot 1954,
713
717, p. 436-438. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 22. Liston, J., W. Weibe, and R. R. Colwell. 1963. Quantitative approach to the study of bacterial species. J. Bacteriol. 85:1061-1070. 23. Lovelace, T. E., and R. R. Colwell. 1968. A multipoint inoculator for petri dishes. Appl. Microbiol. 16:944-945. 24. Lovelace, T. E., H. Tubiash, and R. R. Colwell. 1968. Quantitative and qualitative commensal bacterial flora of Crassostrea virginica in Chesapeake Bay. Proc. Natl. Shellfish. Assoc. 58:82-87. 25. Mallory, L. M., B. Austin, and R. I. Colwell. 1977. Numerical taxonomy and ecology of oligotrophic bacteria isolated from the estuarine environment. Can. J. Microbiol. 23:733-750. 26. Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid from microorganisms. J. Mol. Biol. 3: 208-218. 27. Marmur, J., and P. Doty. 1962. Determination of the base composition of deoxyribonucleic acid from its thermal denaturation temperature. J. Mol. Biol. 5:109-118. 28. Nelson, J. D., and R. R. Colwell. 1975. The ecology of mercury resistant bacteria in Chesapeake Bay. Microb. Ecol. 1:191-218. 29. Otto, L A., and M. J. Pickett. 1976. Rapid method for identification of gram-negative, nonfermentative bacilli. J. Clin. Microbiol. 3:566-575. 30. Pagel, J. E., and P. L. Seyfried. 1976. Numerical taxonomy of aquatic Acinetobacter isolates. J. Gen. Microbiol. 95:220-232. 31. Pfister, R. M., and P. R. Burkholder. 1965. Numerical taxonomy of some bacteria isolated from antarctic and tropical seawaters. J. Bacteriol. 90:863-872. 32. Poindexter, J. S. 1964. Biological properties and classification of the Caulobacter group. Bacteriol. Rev. 28: 231-295. 33. Poindexter, J. S. 1974. Caulobacter Henrici and Johnson 1935, 83, p. 153-155. In R. E. Buchanan and R. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 34. Quigley, M. M., and R. R. Colwell. 1968. Properties of bacteria isolated from deep-sea sediments. J. Bacteriol. 95:211-220. 35. Reichelt, J. L., and P. Baumann. 1973. Taxonomy of the marine luminous bacteria. Arch. Mikrobiol. 94:283330. 36. Sayler, G. S., J. D. Nelson, A. Justice, and R. R. Colwell. 1976. Incidence of Salmonella spp., Clostridium botulinum, and Vibrio parahaemolyticus in an estuary. Appl. Environ. Microbiol. 31:723-730. 37. Seyfried, P. L. 1973. Sampling bacteria in Lake Ontario and the Toronto Harbour, In Proceedings of the 16th Congress of Great Lakes Research. p. 163-182. International Association for Great Lakes Research, Ann Arbor, Mich. 38. Shewan, J. M., G. Hobbs, and W. Hodgkiss. 1960. A determinative scheme for the identification of certain genera of Gram-negative bacteria, with special reference to the Pseudomonadaceae. J. Appl. Bacteriol. 23: 379-390. 39. Shewan, J. M., and M. Veron. 1974. Vibrio Pancini 1854, 411, p. 340-345. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 40. Sierra, G. 1957. A simple method for the detection of lipolytic activity of microorganisms and some observations on the influence of the contact between cells and fatty substrates. Antonie van Leeuwenhoek J. Microbiol. Serol. 23:15-22. 41. Simidu, U., and T. Kaneko. 1973. A numerical taxonomy
714
42. 43. 44.
45. 46.
APPL. ENVIRON. MICROBIOL.
AUSTIN ET AL.
of Vibrio and Aeromonas from normal and diseased marine fish. Bull. Japanese Soc. Sci. Fisheries 39:689703. Simidu, U., E. Kaneko, and N. Taga. 1977. Microbiological studies of Tokyo Bay. Microb. Ecol. 3:173-191. Sneath, P. H. A. 1957. The application of computers to taxonomy. J. Gen. Microbiol. 17:201-226. Sneath, P. H. A., and M. Stevens. 1967. A divided petri dish for use with multipoint inoculatorsa J. Appl. Bacteriol. 30:495-497. Sokal, R. R., and C. D. Michener. 1958. A statistical method for evaluating systematic relationships. Univ. Kansa Sci. Bull. 38:1409-1438. Stanier, R. Y., N. J. Palleroni, and M. Doudoroff.
47. 48. 49. 50.
1966. The aerobic pseudomonads: a taxonomic study. J. Gen. Microbiol. 43:159-271. Stevenson, I. L. 1967. Utilization of aromatic hydrocarbons by Arthrobacter spp. Can. J. Microbiol. 13:205211. Vaitiinen, P. 1977. Effects of composition of substrate and inoculation technique on plate counts of bacteria in the northern Baltic Sea. J. Appl. Bacteriol. 42:437-443. Yamada, K., and K. Komagata. 1972. Taxonomic studies on coryneform bacteria. V. Classification of coryneform bacteria. J. Gen. Appl. Microbiol. 18:417431. ZoBell, C. E. 1941. Studies on marine bacteria. 1. The cultural requirements of heterotrophic aerobes. J. Mar. Res. 4:42-75.
Vol. 37, No.4
Comparative Study of the Aerobic, Heterotrophic Bacterial Flora of Chesapeake Bay and Tokyo Bay B. AUSTIN,' S. GARGES,' B. CONRAD,' E. E. HARDING,' R. R. COLWELL,`* U. SIMIDU,2 AND N. TAGA2 Department of Microbiology, University of Maryland, College Park, Maryland 207421 and Ocean Research Institute, University of Tokyo, Nakano, Tokyo 164, Japan2 Received for publication 24 January 1979
A comparative study of the bacterial flora of the water of Chesapeake Bay and Tokyo Bay was undertaken to assess similarities and differences between the autochthonous flora of the two geographical sites and to test the hypothesis that, given similarities in environmental parameters, similar bacterial populations will be found, despite extreme geographic distance between locations. A total of 195 aerobic, heterotrophic bacterial strains isolated from Chesapeake Bay and Tokyo Bay water were examined for 115 biochemical, cultural, morphological, nutritional, and physiological characters. The data were analyzed by the methods of numerical taxonomy. From sorted similarity matrices, 77% of the isolates could be grouped into 30 phena and presumptively identified as Acinetobacter-Moraxella, Caulobacter, coryneforms, Pseudomonas, and Vibrio spp. Vibrio and Acinetobacter species were found to be common in the estuarine waters of Chesapeake Bay, whereas Acinetobacter-Moraxella and Caulobacter predominated in Tokyo Bay waters, at the sites sampled in the study.
With the increased interest in the microbial ecology of aquatic environments in recent years, it is important to have an understanding of the natural, or autochthonous, microbial flora, not only in terms of biomass and potential activity, but also community structure and species composition. Since the bacteria are well-known agents of mineralization and transformation of organic and inorganic matter in bays and estuaries, it was considered useful to determine whether bacteria in the water column in two of the major bays of the world supported a similar bacterial flora. Under the auspices of a cooperative research program between the University of Maryland and the University of Tokyo, a study was initiated in 1974 in which water samples were collected using standard methods at sites selected for similarity in environmental parameters, and in which the methods of bacteriological analysis were rigidly conforming to permit minimization of variation ascribable to other than natural floristic differences. The pure cultures obtained in the two parallel studies were analyzed by previously agreed-upon procedures in the two laboratories, with exchange of reference cultures, as well as fresh isolates, so that internal quality control could be maintained in the analyses. Although there has been rapid improvement 704
in the taxonomy of aquatic bacteria since the advent of numerical taxonomy and subsequent pioneering studies of Pfister and Burkholder (31) and Quigley and Colwell (34) and the more recent work of Bauman and associates (3, 4), Delabr6 et al. (12), and Reichelt and Baumann (35), this information has been of limited value because of the relatively few organisms examined, the restricted number of tests used in identification of isolates, and the limited geographical distribution of the organisms studied. Pfister and Burkholder (31) attempted to compare the bacterial microflora of tropical and antarctic seawater, a type of broadly defined study that is essential for proper assessment of the distribution of bacteria in the marine environment. Baumann et al. (3, 4) restricted their studies to a set of selected marine isolates, for the most part from culture collections of various investigators, although some fresh isolates were included in their analyses. In the study described here, aerobic, heterotrophic, estuarine bacteria representative of the microflora of Chesapeake Bay and Tokyo Bay waters were compared by numerical taxonomy and molecular genetic methods, providing new information concerning the species composition of microbial communities present in the water column at selected sites in two major estuaries of the world.
705
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 MATERIALS AND METHODS
umes
Isolation and maintenance of strains. Stations selected from track lines routinely surveyed by the University of Maryland in Chesapeake Bay and by the Ocean Research Institute, University of Tokyo. The stations selected were the two that were the most similar in physical and chemical characteristics. Selection was on the basis of information available to the respective laboratories for the stations over the past 15 years. Field studies designed specifically to select locations with the same salinity, temperature, etc., would have been preferable, but neither time nor funds permitted such an exhaustive search for sites exhibiting precisely identical parameters. In any event, an estuarine system is dynamic, and salinity, temperature, and other parameters are not constant throughout the year and from year to year. The stations selected, on average, were the most similar, over time, based on respective laboratory records. Thus, the locations, time of year, salinity, depth, amount of pollution, tidal cycle, time of day, etc., were carefully considered, and these factors were as closely similar as circumstances permitted. Water samples were collected on 19 June 1975 from Cape Charles in Chesapeake Bay and on 4 July 1975 from Tokyo Bay during cruises aboard research vessels that had been scheduled a year in advance, the dates being the closest in time that the respective schedules permitted. Cape Charles is a high-salinity station in Chesapeake Bay, with a surface-to-sediment depth of 29 m, situated approximately 32 km north of Norfolk, Va. and 26 km northeast of the James River. The sampling station in Tokyo Bay, Japan, with a depth of 17 m, is situated 2 km west of the Obitsu River. Physical and chemical characteristics of the stations in Chesapeake Bay and Tokyo Bay are listed in Table 1. Water samples were collected aseptically, using a Niskin Bag Sampler (General Oceanics Inc., Miami, Fla.), from approximately 2 m above the underlying sediment at Cape Charles, since the Chesapeake Bay is a typical "salt water wedge" estuary and the most saline water was sought, and from 1 m below the surface of the water in Tokyo Bay. Triplicate Niskin casts were made, each collecting 2 liters of water to ensure statistical validity of the sampling procedure. Serial dilutions to 10' of the water samples were prepared immediately after collection, using 9-ml vol-
were
of sterile marine salts solution (0.7% [wt/vol]
MgSO4.7H20, 2.38% [wt/vol] NaCl, and 0.07% [wt/ vol] KCI; pH 7.0), the composition of which had been agreed upon by the two laboratory groups. For estimation of total counts of Chesapeake Bay water, 0.1ml aliquots were pipetted onto Simidu medium (41), thiosulfate citrate bile salts agar (TCBS; BioQuest, Cockeysville, Md.), marine 2216 agar (50; Difco), chitin agar, and fish protein medium (0.1% [wt/vol] XM-1 fish protein hydrolysate [Zapata Haynie Corp., Baltimore, Md.], 0.1% [wt/vol] Difco yeast extract, 1.5% [wt/vol] Difco agar, and 1 liter of marine salts solution). The Tokyo Bay water samples were plated on all of the above except the fish protein medium which was included for comparative purposes and comprised part of another study carried out by the University of Maryland laboratory. All plates were spread using a sterile glass spreader. Three replicates of each dilution were plated, and inoculated plates were incubated at 25°C, approximating the in situ temperature, for 14 days. Cultures were randomly selected from platings of each of the three water samples on the media used in the study. Thus, from each of the water samples from the three separate casts made at the sampling site, approximately the same number of cultures were randomly picked from plates of each of the media employed and from the same dilution, containing between 30 and 300 colonies. The cultures were subsequently purified, after incubation at 25°C, by streaking and restreaking three times on plates of 2216 agar. Heatfixed smears from 24-h cultures were stained using Hucker's modification of the Gram stain (16) and examined microscopically. When the cultures were considered to be pure, they were inoculated onto slopes of marine 2216 agar. Stock cultures were maintained under sterile mineral oil on 2216 agar slopes at 25°C and subcultured every 6 to 8 weeks. After all of the isolation and purification steps had been followed, a total of 195 isolates, approximately equally representing the Japanese and U.S. samples, were examined for the biochemical, cultural, morphological, nutritional, and physiological characteristics included in the analyses. Reference strains. In addition to the fresh isolates from Chesapeake Bay and Tokyo Bay, 18 reference cultures were also included in the study (Table 2). Characterization of the strains. Each strain was
TABLE 1. Physical, chemical, and microbiological characteristics of the sampling stations Total counts' ency of Fish Chitin water, Temp Marin Simdu measured (0C) Marine FCSprtish medium from surmedium agar agar ga face (in)agrmeim 3.3 25.7 2.0 x 104 8.4 x 101 6.3 x 102 4.2 x 103 8.3 x 102
TransparDissolved
Depth of water Salinity Station location Sample type sample M%O (m)
oxygen (mg/ liter)
Water
27
25.8
7.4
2 km west of Ob- Water itsu River, Tokyo Bay
11
30.1
7.5
Cape Charles, Chesapeake Bay
2.2
21.3 9.1 X 104 3.3 x
10' 1.8 x 10'
_b
6.7 x 101
Colony-forming units, i.e., total viable, aerobic, heterotrophic bacterial counts on the respective media employed.
bNot done.
706
AUSTIN ET AL.
APPL. ENVIRON. MICROBIOL.
TABLE 2. Reference cultures included in the numerical taxonomy analyses Laboratory Culture reference Reference culture collection Comment/isolated from no. no. Acinetobacter lwoffia Acinetobacter iwoffi
R12 R13
Aeromonas salmonicidaa Arthrobacter crystallopoietes Arthrobacter marinus
R8 R6 Rl
ATCC 15309 Univ. of Maryland Culture Collection ATCC 14174 ATCC 15481 ATCC 25374
Corynebacterium poinsettiae Erwinia herbicola Pseudomonas aureofaciens' Pseudomonas cepacia
R9 R18 R24 R25
ATCC 9069 ATCC 12887 ATCC 13985 ATCC 17759
Pseudomonas maltophiliaa
R20
ATCC 13637
Pseudomonas mendocina' Pseudomonas putida" Pseudomonas stutzeria Staphylococcus saprophyticus" Vibrio anguillarum Vibrio anguillarum" Vibrio parahaemolyticus' Xanthomonas begoniae
R26 R7 R27 R17 R15 R14 R16 R21
ATCC 25411 ATCC 12633 ATCC 17588 ATCC 15305 ATCC 14181 ATCC 19264 ATCC 17802 ATCC 11725
Diseased brook trout Soil Renamed Pseudomonas marina a; seawater
a Denotes type strain, obtained from the American Type Culture Collection
examined for a total of 115 unit characters. Whenever possible, marine 2216 agar was used as the basal medium, and, in all cases, media were prepared using marine salts solution as the diluent. Unless otherwise stated, all inoculated media were incubated at 25°C for 14 days. To assess possible test error, 20 randomly chosen strains were examined in duplicate; otherwise, tests were repeated only in the event of inconclusive results. For inter-laboratory quality control, a set of fresh isolates was exchanged and included along with the reference cultures in both laboratory testing regimes. Most of the tests have been described elsewhere (9, 14, 19, 25, 40); details of the remainder are given below. Utilization of substrates as the sole source of carbon for energy and growth. Carbon utilization tests were carried out using the medium devised by Stevenson (47). The carbon compounds included aDL-alanine, L-arginine hydrochloride, asparagine, Lhistidine hydrochloride, L-leucine, L-lysine, L-methionine, DL-phenylalanine, L-proline, sodium acetate, sodium adipate, sodium citrate, sodium formate, sodium gluconate, sodium glutamate, sodium glutarate, sodium lactate, sodium malonate, sodium oxalate, sodium tartrate, and L-threonine. These were sterilized as 20% (wt/vol) solutions by tyndallizing for 1 h on 3 consecutive days at 100°C, or by filtration, and added to the supporting medium to give a final concentration of 0.2% (wt/vol). This medium was dispensed into divided replidishes (44) and inoculated by means of a multipoint inoculator (23). After incubation, the medium was examined visually at 14 days, recording the presence or absence of growth. Determination of the G+C content of DNA. The guanine plus cytosine content (G+C content, moles percent) of purified deoxyribonucleic acid (DNA), obtained by the method of Marmur (26), was
River clay Renamed Pseudomonas multivoransa; forest soil Oropharynx of patient with mouth cancer Soil Degrades aromatic acid Spinal fluid Urine Ulcerous lesion in cod Food poisoning, Japan Semituberous Begonia spp.
(Rockville, Md.).
determined from the thermal denaturation temperature (Tm) of the DNA using a Gilford 2400-S recording spectrophotometer at 260 nm (27). The DNA preparations were dialyzed to lx SSC (0.15 M NaCl, 0.015 M sodium citrate) and heated at 0.5°C min-', during which the optical density was monitored at 260 nm. Molar percentage of G+C was calculated from the Tm by the equation of De Ley (13). Coding of data. The characters were coded 1 for positive or present, 0 for negative or absent, and 9 for noncomparable or not applicable. The final n x t matrix contained 213 strains, including reference strains, and 115 characters recorded for all strains. Computer analyses. The data were analyzed using SsM, the simple matching coefficient (45), which included positive and negative matches, and the Jaccard coefficient, Sj, which excludes negative matches (43). Clustering was by unweighted average linkage (45), and sorted similarities and dendrograms were constructed. The hypothetical median organism (22) for each cluster was also calculated. Programs employed included UMDTAXON3 and IGPS3 program packages available on the University of Maryland UNIVAC 1108 computer.
RESULTS Total viable bacterial counts. Four media were employed in the Chesapeake Bay-Tokyo Bay comparative study: marine 2216 agar (50); a medium designed to enhance growth of marine vibrios (41); chitin agar; and TCBS, originally developed for isolation of Vibrio cholerae and subsequently employed for isolation of Vibrio parahaemolyticus. Counts obtained on these
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979
media are given in Table 1. Analysis of the results obtained for each of the three casts showed no statistically significant differences in total viable aerobic, heterotrophic bacterial counts. That is, there was no significant difference in counts for the three water samples that were collected at the same time to test for sample variation. However, it can be seen that higher counts on all media except 2216 agar were obtained for the Chesapeake Bay water samples. The count obtained on TCBS agar was 10-fold greater for the Chesapeake Bay water sample and 2-fold greater for the Simidu marine vibrio medium. This difference is important in view of the results (see below) showing a greater preponderance of vibrios in Chesapeake Bay water samples. Clustering of the strains. Results of analyses employing the SSM and SJ coefficients included 154 of the environmental isolates, representing 77% of the total set, and the six reference cultures of Acinetobacter Iwoffi, Pseudomonas maltophilia, Pseudomonas marina, Vibrio anguillarum (two strains), and V. parahaemolyticus in 30 clusters at 270% similarity (S). Results PERCENTAGE SIMILARITY
707
of the analysis using the SJ coefficient are given in Fig. 1. Four clusters (phena 1, 2, 21, and 25) comprised 65 strains, including the reference culture of V. parahaemolyticus, with the remaining 26 groups containing between 2 and 9 strains each. Identification and description of the isolates. Presumptive identification of the 30 clusters was accomplished either by inclusion of a reference culture within the cluster or by consultation of diagnostic keys in Bergey's Manual of Determinative Bacteriology (8) and specialist keys and tables including those of Shewan et al. (38), Stanier et al. (46), Bousfield (7), Jones (17), Bianchi (M. Bianchi, these de doctorat des sciences naturelles, Universite d'Aix-Marseille, 1976), Otto and Pickett (29), and Pagel and Seyfried (30). Six generic groups comprised the major taxa represented, including Vibrio spp., Acinetobacter-Moraxella spp., Pseudomonas spp., Caulobacter spp., coryneforms, and groupings of as yet unidentified gram-negative rods (see Fig. 1). A summary of selected characteristics of the six generic groupings is given in Table 3. PHENON
NO. OF STRAINS
SOURCE
IDENTITY
16 17 18 19 20
3 3 2 5 5 2 4 2 9
CB CB CB CB CB CB CB CB CB CB CB CB CB C,B C.B/TB. CB/TB TB. TB TB. TB
21 22 23 24
22 2 2 3
TB TB TB TB
25 26 27 28 29 30
21
TB
Coulobocter sp. Gram negotive rods Grom negative rods Grom negative rods Gram negotive rods
5 2 2
TB TB TB TB. TB
Grom negotive rods P pu/bdo Gram negative rods Gram negative rods Gram negotive rods
2 3 4 5 6 7 8 9 10 I1 12 13 14 15
12 10 6 2 2 2 3 5 9 4 7
4
6
Vibrio porohoerno/yt/cus V f,scheri V ongul//orum Vibrio sp. Vibr,o sp V,br,o sp
Aclnetobocter Gram negotive Gram negative Gram negative P morino P mo/ltophi/io Grom negative
co/coocet/cus rods rods rods rods
Coryneforms" Gram negative rods Gram negative rods Aclnetobocter - Aoroxe/lo sp. Aclnetobocter - Moroxe/lo sp Pseudolonos sp. Grom negative rods
FIG. 1. Simplified dendrogram prepared with the SJ coefficient and unweighted average linkage clustering technique. C.B., Chesapeake Bay; T.B., Tokyo Bay.
708
APPL. ENVIRON. MICROBIOL.
AUSTIN ET AL. 3
C14
cq
3
>+1I
3
>+I
3
CD
I1
I
la00
-I + +-I
la00 +0
+
+0
3
+0
cq
3
eq
CD
cs
10
3
10
+
>
C4
3
0
-r
>> +r +
eq
IIIII II
IlIlIl
+
+ I +++I
III
IIIIII
+
II
11-111 II II I I
II
+
I,I I IIIII
I I + +-I
+
1+-ol
II
IlIlIlIl II
+
+I++II
II
I+
+
I1I1I+
I+1111 I1 I1
+
II
III ++I
III+I+I++
+
I
I
+II0II1
+
+a
.-Is
cs
3
l
3+
+0
0
_14
I-o +-I
1
+a
.!
3 -4
.S
I I++I
+ I
3
oq
-4
+0 +a
+
I I++I +I + + I
I0
3
C13
+
IIII I-III+
+ I+
o
+ If
+ ++ +
C CD
+0
+
2
3:
+
+ + + +la
+I++I
IIIlII
I
+
+0 a -01+ ++
+0
+ + +l
0
++
+1I
+11
lll+ lr
I++ I
III
0
+
_t
3 CD
0.0
4
tII
to
+
^ eq
3
u-i
Io~
0
-4
+ rX
+
+ 44
-0l++-01 ~~0 IlIlIllaIll + + II +I+-o+I 0+0
IO +
+
+I +10 I01 l+ 1001 10
+
0
+I
+
C4
+
+ I++ II
+0
+
+ I ++ II
Pk
+
+I+1
+ +
+0
3
+ Pk
CIA
+
+I+-o+++++0I++I o-
+0
3 cli
+
+
+
Ct
I I I I I I I I ++
+
3 3
0+1I1I1I
I + I
+
+
+
+
I
II I + I++ II +
I o
3
++I II
I
+
I1+ I1I 01 I5 3
eq
0LO& -4
3
I+1I
3
ral
o
z
10r
Pk
+I+ +l++
+ rZ
II 1
+IC
+01I -o +
+
+++++I
0++++++I 1+
+
+
+- + + +
Ia 0+ ++ +
+
+
0
0 ++ + +I
I1X+0+ +0
I+
+
l
l
+ Pk
3
I1+I 01+
3
+
3
+
+ Pk
I+ 1
+
I
+ + + + +
l
+ +
+ + ++ ++
.0_
0 03
O0 0
2u
cu+ ZI
Qo
c
S=is t f W t2~~~~~~~~~m E
-
00
c
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 II I'X do
I X X I
+
I1+1I1
+
++ I
+
+'0I.X
1+1I +
+'0+'0
+
+ I +.
I I I'
I'0'0' X1I1 '0
'0'
+
I'0+0I+
+
0
+ + +.
._Y0
0. IrII
+0
I++II
eq 0
.
oo~
@0 v
X
X0
I'X I1
l0
l'll1
0. -*
VI
X
-
X~
c) M~; E
X
Al
4 0
0
4,XE
0 b
709
Phena 1, 3, 7, 11, and 12 were identified as V. parahaemolyticus, V. anguillarum, Acinetobacter calcoaceticus, P. marina, and P. maltophilia, respectively, by the grouping of the reference strains with the isolates. In general, the characteristics of the Vibrio strains matched the species descriptions in Bergey's Manual of Determinative Bacteriology (8) and those given by Bianchi (these de doctorat), although it was observed that the strains identified as V. anguillarum were negative for the indole reaction. Although the reference strain of A. Iwoffi was recovered in phenon 7, A. Iwoffi is now regarded as a synonym of A. calcoaceticus (8, 30). The characteristics of A. calcoaceticus (phenon 7) were in close agreement with the description provided in Bergey's Manual of Determinative Bacteriology (8). Phenon 2 was identified as Vibrio fischeri by micromorphology, biochemical, degradative, and physiological reactions and overall DNA content of 44.8% G+C (Table 3). In contrast to the species description, none of the strains exhibited luminescence (39). However, luminescence is a characteristic that is highly dependent on medium, temperature, and other variables. "Dark" strains are very commonly encountered among luminescent species (8, 39). Phena 4, 5, and 6 are classified as Vibrio spp., viz., gram-negative, curved, fernentative, oxidase- and catalase-positive rods, possessing overall DNA base compositions of 46 and 49% G+C. The strains were unable to grow at temperatures of 37°C or above. The strains isolated in this study and identified as Vibrio spp. did not degrade casein or Tween 80 or grow at temperatures below 100C. In contrast to the genus description (39), it should be noted that none of the strains demonstrated a marked sensitivity to pteridine. Thus, phena 4, 5, and 6 were not identified as any of the species described and named in the literature and should be considered new species. Phena 10, 16, 20, 25, 29, and 30 possessed some ofthe characteristics of Acinetobacterspp. However, the range of overall DNA base composition, 50 to 66% G+C, is higher than the range of values usually associated with this genus, i.e., 40 to 47% (21), and closer to that of Pseudomonas. The oxidase-negative reaction, for the present, excludes these organisms from the genus Pseudomonas. Therefore, these phena are regarded as unidentified until further work can be done on the taxonomy of these organisms. Phena 17 and 18 were tentatively identified as belonging to the Acinetobacter-Moraxella group. Although all strains possessed the general characteristics of Acinetobacter-Moraxella,
710
AUSTIN ET AL.
they showed specific properties of Moraxella, i.e., they were oxidase positive and sensitive to penicillin. It has been suggested that oxidasenegative strains should be reclassified as Acinetobacter (5); this would exclude the strains clustered in these phena. However, with the exception of Moraxella phenylpyruvica, Moraxella species, including M. lacunata, M. bovis, M. nonliquefaciens, and M. osloensis, are normally associated with pathological conditions and have not been reported to occur in the aquatic environment. More importantly, the description of the strains isolated in this study could not be matched with any of the described species of Moraxella or, for that matter, of Acinetobacter. Thus, it was decided to regard the phena as groups intermediate between Acinetobacter and Moraxella until further studies on the classification of these organisms can be completed. The overall DNA base composition was not determined for these organisms despite several attempts, because of problems with harvesting cells and extracting the DNA. It is possible that these two phena are similar to phena 10, 16, 19, 20, 22, 23, 24, 25, 26, 28, 29, and 30, which also exhibited characteristics of Acinetobacter-Moraxella but could not be classified in either of these groups due to the high G+C content of the DNA, the latter being more representative of Pseudomonas spp. Some of the clusters of gramnegative rods that remain unidentified may also be Pseudomonas or Alteromonas spp. However, they are indeed new species, if not new genera, and further, more precise study of the classification, identification, and nomenclature of these strains is in progress. Phenon 14 was identified as coryneform bacteria largely on the basis of the micromorphology of the strains. Unfortunately, the organisms could not be identified further, even after consultation of the comprehensive descriptions of Bousfield (7) and Jones (17), and may also represent a new species. Phena 8, 9, 13, and 15 were not related to bacterial species described in the available literature, i.e., in diagnostic keys and/or in Bergey's Manual of Determinative Bacteriology (8), and will require further analysis for precise classification. Strains clustered in phenon 21 produced stalks and formed "rosettes," hence were identified as Caulobacter spp. The strains were short, oxidative, gram-negative rods when observed in wet mounts under phase-contrast microscopy. They could not be identified to the species level and represent, in our view, a new species of Caulobacter. Phenon 27 was identified as Pseudomonas putida (8, 11). Diagnostic features included
APPL. ENVIRON. MICROBIOL.
gram-negative, oxidative, oxidase-positive rods that were fluorescent under ultraviolet light, produced arginine dihydrolase but not lysine decarboxylase, and utilized L-arginiie hydrochloride and sodium citrate as the sole source of carbon for energy and growth, but did not hydrolyze casein gelatin, starch, or Tween 80. The strains were nonmotile when observed in wet mount under phase-contrast microscopy. DNA/ DNA hybridization studies of these strains with reference strains are in progress. Test error. Twenty strains examined in duplicate were tested on different occasions, but the same methods were used throughout. The majority of tests, i.e., a total of 85, gave identical results; these included the Gram staining reaction, colonial morphology, and the majority of biochemical degradative and nutritional characters. The error was calculated to be approximately 2%, most of which could be attributed to a few characters, notably the pleomorphic nature of many of the isolates and the H25, methyl red, oxidase, Voges-Proskauer, and gelatin tests, which either gave weakly positive, equivocal results or were, in the case of the biochemical test results, contradictory on retesting. The H2S test can be variable, depending on the presence or loss of plasmids (variability due to plasmids has been noted in earlier studies [S. A. Orndorff, B. Austin, L. A. McNicol, and R. R. Colwell, submitted for publication]), and, in the case of the methyl red, Voges-Proskauer, and gelatin tests, there is a strain variability in the intensity of the reactions that has been observed in our own work with strains freshly isolated from the aquatic environment. Variation in test results between laboratories was not significant and therefore permitted combining the data from both laboratories for a comprehensive analysis. Selectivity of the media employed. Only Vibrio spp. were isolated on TCBS agar, including V. parahaemolyticus, V. fischeri, and unidentified Vibrio spp. The Simidu medium was also highly selective for Vibrio spp., with V. parahaemolyticus, V. fischeri, and unidentified Vibrio spp. being isolated. However, two isolates, most probably of the Acinetobacter-Moraxella groups, were also isolated from the Simidu medium. The 2216 medium yielded a variety of organisms, selecting neither for nor against Vibrio spp. It should be pointed out that the fish protein agar was by far the most useful medium in providing growth for a very wide variety of bacterial genera and species, including Vibrio, Moraxella, Acinetobacter, and Pseudomonas. DISCUSSION The objective of this study was to assess sim-
VOL. 37, 1979
DISTRIBUTION OF ESTUARINE BACTERIA
ilarities and differences between microflora of two environmentally similar sites at widely separated geographical locations. That is, the question to be answered is whether the water of estuaries in different parts of the world will support a relatively similar species composition in the microbial community structure. In spite of the usual restraints of time, budget, cruise schedules, and related logistical problems, it was possible to select two sites for comparison that were reasonably similar and for which significant data had been gathered in other studies. Cruise schedules for the research vessels available to the cooperating laboratories fortuitously overlapped so that samples could be collected at the same time of year. It is recognized that a more precise matching of sites could be achieved, but the matching of sites was the best possible considering the logistics of the situation. Furthermore, it is also recognized that biomass measurements should have been made, i.e., direct measurement of the total bacterial population using epifluorescence, Limulus lysate assay, adenosine triphosphate, or others of the more recently developed methods for biomass estimation. Tests of potential activity, using 14C heterotrophic uptake measurements, would also be useful. Nevertheless, the question was that of community structure and species composition, and for this reason the comparative study is considered to provide useful information. It must also be recognized that the isolates examined in this study represent only that portion of the microbial community capable of growth on the media employed. For example, for oligotrophs, low-nutrient media must be employed for isolation and characterization (25). Nevertheless, for the purposes of this study, namely to compare microbial populations under standardized conditions of plating, media, etc., this point does not negate the conclusions drawn from the results obtained. The results show that, from the counts obtained on the media employed and given in Table 1 and from the numerical taxonomy analyses, the predominant aerobic, heterotrophic bacterial flora present in the water column of Chesapeake Bay water, at the site included in this study, consisted of Vibrio, Acinetobacter, Pseudomonas, and coryneforms, whereas the microflora of the water column in Tokyo Bay, at a comparable site, was predominately Acinetobacter-Moraxella-like species, Caulobacter spp., and groups of gram-negative, rodlike bacteria resembling Pseudomonas spp. but sufficiently dissimilar to be considered new species. Variation in species composition among the individual water samples was not observed. Thus the microflora of the two bodies of water appear
711
to be dissimilar at the sites from which samples examined in this study were collected. Vibrio was the numerically dominant genus in Chesapeake Bay, regardless of the medium employed, a finding concurring with conclusions of Lovelace et al. (24), Kaneko and Colwell (18), and Cook and Goldman (10). Vibrios were not predominant in the microbial populations found in the water of Tokyo Bay at the stations included in the study. It has been suggested previously by Simidu et al. (42) that the absence of vibrios in Tokyo Bay is caused by antagonistic interactions between Vibrio spp. and phytoplankton and by organic nutrients in Tokyo Bay. The presence of V. parahaemolyticus in Chesapeake Bay is known to be inversely related to salinity, in that extremes of salinity affect viability (18, 36), and warmer water temperatures are favorable for their growth and distribution during the warmer seasons of the year (18). In this connection, the water at Cape Charles demonstrated comparatively high salinity and temperature, and relatively few V. parahaemolyticus were isolated. Acinetobacter spp., which accounted for the third most predominant taxon at Cape Charles, were tentatively identified as the predominating organisms in Tokyo Bay. Large numbers of Acinetobacter have been isolated from the aquatic environment by other investigators, notably in Lake Ontario (37) and Lake Superior (6). However, organisms placed in this genus no doubt have been misclassified because of the relatively loose description available for the genus. Thus, it can be concluded that Acinetobacter isolated from the aquatic environment are not sufficiently well studied to be readily separated from the other gram-negative, nonmotile, nonpigmented bacteria, a situation also true for Flavobacterium (14). Many of the strains recovered from Tokyo Bay possessed characteristics more appropriate for groups intermediate between Acinetobacter and Moraxella and, in some cases, related to Pseudomonas. Pseudomonas spp., which have been isolated from a wide range of habitats within the aquatic environment (3, 10, 12, 31, 34), were recovered in moderately large numbers from Chesapeake Bay and also from Tokyo Bay. Several of the clusters of gram-negative, rodlike bacteria were very similar to Pseudomonas, except for a tendency to pleomorphism and differences in certain characteristics, such as sensitivity to penicillin, and may ultimately prove to be new spe-
cies of Pseudomonas. It would be very useful to assess the ability of these strains to degrade petroleum and to measure their resistance to heavy metals and other pollutants, since the pseudomonads, together with the coryneforms (which were surprisingly absent from the Tokyo
712
AUSTIN ET AL.
Bay water samples), are associated with petroleum degradation (1, 2) and mobilization of heavy metals (28). The coryneform strains recovered in this study could not be identified to the genus level, reflective of the complexity of the taxonomy of the coryneform group of bacteria and uncertainty of the validity of the classification of those coryneforms found in the marine environment (7, 17, 49). Caulobacter spp. generally occur in waters containing low concentrations of organic nutrients (32, 33). Slow-growing strains of stalked bacteria, identified as Hyphomonas polymorpha, have been isolated in large numbers in other, more polluted areas of Chesapeake Bay (25). The occurrence of stalked bacteria in the estuarine environment would suggest that such organisms may have important ecological functions, and further studies are in progress. The lower Chesapeake Bay, at the station included in this study, receives much less allochthonous input than the upper Chesapeake Bay. It would appear that the differences in species composition noted in this study provide an index of pollution, since the Tokyo Bay station can be considered to be eutrophic and more polluted, in general, than the Chesapeake Bay. Pfister and Burkholder (31) observed a degree of specificity between tropical and antarctic waters in the distribution of bacterial taxa in the aquatic environment. Recent work in our laboratory has demonstrated a geographical distribution among petroleum-degrading bacteria and a specificity of these bacteria for water or sediment in polluted and unpolluted sites in Chesapeake Bay (1, 2). However, the distribution was not as striking for those genera of bacteria capable of degrading petroleum as were the differences for the bacterial taxa comprising the microflora of Chesapeake Bay and Tokyo Bay. An interesting observation arising from this study is that the choice of medium will exert a significant influence on the recovery of bacterial taxa in the marine environment, an observation in agreement with earlier studies of Goulder (15) and Vaatiinen (48). In this study, fish protein agar was used for comparative purposes only in the Chesapeake Bay portion of the study. Although the counts on the fish protein agar were one log less than on 2216 agar, the fish protein agar medium permitted the initial isolation of the widest range of taxa, that is, a greater number of genera and species than the more commonly used marine 2216 agar of ZoBell (50). This observation merits further study. In comparison, the TCBS medium was highly selective for vibrios, as was the Simidu medium, both of which were designed for isolation of Vibrio spp. After isolation and culture in the laboratory,
APPL. ENVIRON. MICROBIOL.
those strains isolated on media other than 2216 agar were able, subsequently, to grow on this medium, indicating an adaptation of the isolates to the richer medium. In a previous study of slow-growing oligotrophic bacteria recovered from the estuarine environment (25), the composition of the isolation medium was found to exert a significant effect upon the recovery of bacterial taxa. In particular, it was shown that Streptothrix and other sheathed bacteria could be recovered only on a medium containing a very low concentration of nutrients, consisting of Chesapeake Bay water to which agar was added. No variation in genera and species was noted between casts. That is, each of the separate casts yielded the same representation of genera and species on each of the media, with significant differences noted only between media. The TCBS and Simidu media were selective for Vibrio spp., and the other media, i.e., chitin, 2216, and fish protein agar media, selected neither for nor against Vibrio spp. Of all the media, however, the fish protein concentrate medium yielded the widest variety of genera and species in the case of the Chesapeake Bay water samples. Clearly, the effects of media composition are important in the recovery of bacterial taxa from the natural environment, and this phenomenon should be carefully considered in future studies of microbial community structure. In conclusion, intrinsic differences were observed in the microflora of Chesapeake Bay and Tokyo Bay water in a given location and season. Space and time dependency of microbial community structure, of course, is well recognized. Nevertheless, the results of this study indicate that it cannot be assumed that estuaries will contain microbial communities similarly comprised of representative taxa, but, rather, it should be recognized that geographical differences may be strongly influential, with the type and concentration of nutrients present, the indigenous plankton populations, and the nature, source, quality, and quantity of allochthonous material entering an estuary acting selectively on the microbial populations. ACKNOWLEDGMEENTS This work was supported by National Science Foundation grants OIP 7417540 and BMS 72-02227-A03. The computer time for the project was supplied, in full, through the facilities of the Computer Science Center at the University of Maryland. The Japanese cooperative research was supported by grant no. 5R064 from the Japan Society for the Promotion of Science. The assistance of the captains and crews of the R/V Ridgely Warfifld and of the R/V Tanseimaru is gratefully acknowledged. The authors gratefully acknowledge the assistance of Jayne Carney and Gary Sayler in the initial stages of this project.
DISTRIBUTION OF ESTUARINE BACTERIA
VOL. 37, 1979 LITERATURE CITED
1. Austin, B., R. R. Colwell, J. D. Walker, and J. J. Calomiris. 1977. Numerical taxonomy and ecology of petroleum-degrading bacteria. Appl. Environ. Microbiol. 33:60-68. 2. Austin, B., R. R. Colweli J. D. Walker, and J. J. Calomiris. 1977. The application of numerical taxonomy to the study of petroleum degrading bacteria isolated from the aquatic environment. Dev. Ind. Micro-
biol. 18:685695. 3. Baumann, L, P. Baumann, M. Mandel, and R. D. Allen. 1972. Taxonomy of aerobic marine eubacteria.
J. Bacteriol. 110:402-429. 4. Baumann, P., L. Baumann, and M. Mandel. 1971. Taxonomy of marine bacteria: the genus Beneckea. J.
Bacteriol. 107:268-293. 5. Baumann. P., AL Doudoroff, and R. Y. Stanier. 1968.
A study of the Moraxella group. II. Oxidase-negative species (genus Acinetobacter). J. Bacteriol. 95:15201541. 6. Bennett, E. A. 1969. A study of the distribution of heterotrophic bacteria in the Great Lakes. 1. The heterotrophs of lake water. Bacteriology Branch, Division of Laboratories, Ontario Water Resources Commission, Ontario, Canada. 7. Bousfield, I. J. 1972. A taxonomic study of some coryneform bacteria. J. Gen. Microbiol. 71:441-455. 8. Buchanan, R. E., and N. E. Gibbons (ed.). 1974. Bergey's manual of determinative bacteriology, 8th ed. Williams & Wilkins Co., Baltimore. 9. Colwell, R. R., and W. J. Weibe. 1970. "Core" characteristics for use in classifying aerobic, heterotrophic bacteria by numerical taxonomy. Bull. Ga. Acad. Sci. 18:165-185. 10. Cook, T. M., and C. K. Goldman. 1976. Bacteriology of Chesapeake Bay surface waters. Chesapeake Sci. 17: 40-49. 11. Cowan, S. T. 1974. Cowan and Steel's manual for the identification of medical bacteria. Cambridge University Press, Cambridge. 12. Delabre, AL, A. Bianchi, and M. Veron. 1973. g#tude critique de methodes de taxonomie numerique. Application a une classification de bacteries aquicoles. Ann. Microbiol. 124:489-506. 13. De Ley, J. 1970. Re-examination of the association between melting point, buoyant density, and chemical base composition of deoxyribonucleic acid. J. Bacteriol. 101:738-754. 14. Goodfellow, M., B. Austin, and C. H. Dickinson. 1976. Numerical taxonomy of some yellow pigmented bacteria isolated from plants. J. Gen. Microbiol. 97:219-233. 15. Goulder, R. 1976. Evaluation of media for counting viable bacteria in estuaries. J. Appl. Bacteriol. 41:351-355. 16. Hucker, G. J., and H. J. Conn. 1923. Methods of Gram staining. In Technical Bulletin of the New York State Agricultural Experiment Station, no. 93. 17. Jones, D. 1975. A numerical taxonomic study of coryneforms and related organisms. J. Gen. Microbiol. 87:5796. 18. Kaneko, T., and R. R. Colwell. 1974. Distribution of Vibrio parahaemolyticus and related organisms in the Atlantic Ocean off South Carolina and Georgia. Appl. Microbiol. 28:1009-1017. 19. King, E. O., M. K. Ward, and D. E. Raney. 1954. Two simple media for the demonstration of pyocyanin and fluorescein. J. Lab. Clin. Med. 44:301-307. 20. Lautrop, H. 1974. Moraxella Lwoff 1939, 173 Nom. cons. Opin. 41, Jud. Commn. 1971, 106, p. 433-436. In, R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 21. Lautrop, H. 1974. Acinetobacter Brisou and Prevot 1954,
713
717, p. 436-438. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 22. Liston, J., W. Weibe, and R. R. Colwell. 1963. Quantitative approach to the study of bacterial species. J. Bacteriol. 85:1061-1070. 23. Lovelace, T. E., and R. R. Colwell. 1968. A multipoint inoculator for petri dishes. Appl. Microbiol. 16:944-945. 24. Lovelace, T. E., H. Tubiash, and R. R. Colwell. 1968. Quantitative and qualitative commensal bacterial flora of Crassostrea virginica in Chesapeake Bay. Proc. Natl. Shellfish. Assoc. 58:82-87. 25. Mallory, L. M., B. Austin, and R. I. Colwell. 1977. Numerical taxonomy and ecology of oligotrophic bacteria isolated from the estuarine environment. Can. J. Microbiol. 23:733-750. 26. Marmur, J. 1961. A procedure for the isolation of deoxyribonucleic acid from microorganisms. J. Mol. Biol. 3: 208-218. 27. Marmur, J., and P. Doty. 1962. Determination of the base composition of deoxyribonucleic acid from its thermal denaturation temperature. J. Mol. Biol. 5:109-118. 28. Nelson, J. D., and R. R. Colwell. 1975. The ecology of mercury resistant bacteria in Chesapeake Bay. Microb. Ecol. 1:191-218. 29. Otto, L A., and M. J. Pickett. 1976. Rapid method for identification of gram-negative, nonfermentative bacilli. J. Clin. Microbiol. 3:566-575. 30. Pagel, J. E., and P. L. Seyfried. 1976. Numerical taxonomy of aquatic Acinetobacter isolates. J. Gen. Microbiol. 95:220-232. 31. Pfister, R. M., and P. R. Burkholder. 1965. Numerical taxonomy of some bacteria isolated from antarctic and tropical seawaters. J. Bacteriol. 90:863-872. 32. Poindexter, J. S. 1964. Biological properties and classification of the Caulobacter group. Bacteriol. Rev. 28: 231-295. 33. Poindexter, J. S. 1974. Caulobacter Henrici and Johnson 1935, 83, p. 153-155. In R. E. Buchanan and R. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 34. Quigley, M. M., and R. R. Colwell. 1968. Properties of bacteria isolated from deep-sea sediments. J. Bacteriol. 95:211-220. 35. Reichelt, J. L., and P. Baumann. 1973. Taxonomy of the marine luminous bacteria. Arch. Mikrobiol. 94:283330. 36. Sayler, G. S., J. D. Nelson, A. Justice, and R. R. Colwell. 1976. Incidence of Salmonella spp., Clostridium botulinum, and Vibrio parahaemolyticus in an estuary. Appl. Environ. Microbiol. 31:723-730. 37. Seyfried, P. L. 1973. Sampling bacteria in Lake Ontario and the Toronto Harbour, In Proceedings of the 16th Congress of Great Lakes Research. p. 163-182. International Association for Great Lakes Research, Ann Arbor, Mich. 38. Shewan, J. M., G. Hobbs, and W. Hodgkiss. 1960. A determinative scheme for the identification of certain genera of Gram-negative bacteria, with special reference to the Pseudomonadaceae. J. Appl. Bacteriol. 23: 379-390. 39. Shewan, J. M., and M. Veron. 1974. Vibrio Pancini 1854, 411, p. 340-345. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of determinative bacteriology, 8th ed. The Williams & Wilkins Co., Baltimore. 40. Sierra, G. 1957. A simple method for the detection of lipolytic activity of microorganisms and some observations on the influence of the contact between cells and fatty substrates. Antonie van Leeuwenhoek J. Microbiol. Serol. 23:15-22. 41. Simidu, U., and T. Kaneko. 1973. A numerical taxonomy
714
42. 43. 44.
45. 46.
APPL. ENVIRON. MICROBIOL.
AUSTIN ET AL.
of Vibrio and Aeromonas from normal and diseased marine fish. Bull. Japanese Soc. Sci. Fisheries 39:689703. Simidu, U., E. Kaneko, and N. Taga. 1977. Microbiological studies of Tokyo Bay. Microb. Ecol. 3:173-191. Sneath, P. H. A. 1957. The application of computers to taxonomy. J. Gen. Microbiol. 17:201-226. Sneath, P. H. A., and M. Stevens. 1967. A divided petri dish for use with multipoint inoculatorsa J. Appl. Bacteriol. 30:495-497. Sokal, R. R., and C. D. Michener. 1958. A statistical method for evaluating systematic relationships. Univ. Kansa Sci. Bull. 38:1409-1438. Stanier, R. Y., N. J. Palleroni, and M. Doudoroff.
47. 48. 49. 50.
1966. The aerobic pseudomonads: a taxonomic study. J. Gen. Microbiol. 43:159-271. Stevenson, I. L. 1967. Utilization of aromatic hydrocarbons by Arthrobacter spp. Can. J. Microbiol. 13:205211. Vaitiinen, P. 1977. Effects of composition of substrate and inoculation technique on plate counts of bacteria in the northern Baltic Sea. J. Appl. Bacteriol. 42:437-443. Yamada, K., and K. Komagata. 1972. Taxonomic studies on coryneform bacteria. V. Classification of coryneform bacteria. J. Gen. Appl. Microbiol. 18:417431. ZoBell, C. E. 1941. Studies on marine bacteria. 1. The cultural requirements of heterotrophic aerobes. J. Mar. Res. 4:42-75.
