Distribution vulnificus and Lactose- Fermenting Vibrios ...aem.asm.org/content/45/3/985.full.pdf ·...

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1983, p. 985-9980099-2240183/030985-14$02.00/0Copyright 0 1983, American Society for Microbiology

Vol. 45, No. 3

Distribution of Vibrio vulnificus and Other Lactose-Fermenting Vibrios in the Marine EnvironmentJAMES D. OLIVER,* ROBERT A. WARNER, AND DAVID R. CLELAND

Department ofBiology, University ofNorth Carolina, Charlotte, North Carolina 28223

Received 24 August 1982/Accepted 8 December 1982

During the summer of 1981, 3,887 sucrose-negative vibrios were isolated fromseawater, sediment, plankton, and animal samples taken from 80 sites fromMiami, Fla., to Portland, Maine. Of these, 4.2% were able to ferment lactose. Thelactose-positive strains isolated from the various samples correlated positivelywith pH and turbidity of the water, vibrios in the sediment and oysters, and totalbacterial counts in oysters. Negative correlations were obtained for water salinity.Numerical taxonomy was performed on 95 of the lactose-fermenting environmen-tal isolates and 23 reference strains. Five clusters resulted, with the major clustercontaining 33 of the environmental isolates and all of the Vibrio vulnificusreference strains. The 33 isolates, which produced an acid reaction in lactosebroth within hours of initial inoculation, represented 20% of all lactose-fermentingvibrios studied. These isolates were nearly identical phenotypically to clinicalstrains of V. vulnificus studied by the Centers for Disease Control, Atlanta, Ga.,and by our laboratory, and their identification was confirmed by DNA-DNAhybridization studies. V. vulnificus was isolated from all sample types and fromMiami to Cape Cod, Mass., and comparison of the environmental parameters ofthe eight subsites yielding this species with those of all 80 subsites revealed nosignificant differences. The majority of the isolates were obtained from animals,with clams providing most (84%) of these. On injection into mice, 82% of the V.vulnificus isolates resulted in death. Members of the remaining four clusterscontained strains which differed from V. vulnificus in such phenotypic traits asluminescence and in urease or H2S production. None of the other referencecultures, including nine other Vibrio species, were contained in the remainingclusters, and these isolates could not be identified. Most of these were also lethalfor mice. Phenotypic differences, potential pathogenicity, and geographic distri-bution of the five clusters were examined. It is concluded that V. vulnificus is aubiquitous organism, both geographically and in a variety of environmentalsources, although it occurs in relatively low numbers. The public health signifi-cance of this organism and of the other unidentified lactose-fermenting Vibriospecies is discussed.

Vibrio vulnificus is a lactose-fermenting, op-portunistic human pathogen capable of causingdeath in mice within 2.5 h (3, 19) and in humanswithin 2 or 3 days (2, 6). Although infectionsapparently can occur in otherwise healthy indi-viduals (11), human disease typically resultsfrom ingestion of contaminated seafood or frominfection of a wound, frequently of crab oroyster origin (2). Studies from our laboratoryand from case histories suggest that personswith elevated serum iron levels (owing to, e.g.,chronic alcoholism, hepatitis, thalassemia ma-jor, or hemochromatosis) are especially vulnera-ble to infection by this organism (2, 27). Theorganism is able to cause massive damage to theintestinal wall, which probably allows its pene-tration to the circulatory system (17; J. B. Del-

linger, M.S. thesis, University of North Carolinaat Charlotte). A toxin(s) has now been isolatedwhich shows cytolytic, cytotoxic, vascular per-meability, and lethal activities (12). Poole et al.(M. D. Poole, J. H. Bowdre, and D. Klapper,Abstr. Annu. Meet. Am. Soc. Microbiol. 1982,B155, p. 43) have demonstrated an extracellularproduct with proteolytic (but not hemolytic)activity able to degrade albumin, complementfractions C3 and C4, immunoglobulin G, andelastin. This elastase rapidly produced hemor-rhagic necrosis, edema, and muscle tissue dis-ruption when injected into mice at microgramlevels. Collagenolytic (23) and hemolytic (7)activities also have been demonstrated.

Infections caused by V. vulnificus have beenreported from Japan (14), Australia (5), Belgium

985

on June 27, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

http://aem.asm.org/

986 OLIVER, WARNER, AND CLELAND

TABLE 1. Sites sampled for lactose-fermentingVibrio spp.

Site No. ofsubsites

Portland, Maine ......................... 3Boston, Mass............................ 4Cape Cod, Mass......................... 4New London, Conn...................... 4New York City, N.Y..................... 10Atlantic City, N.J........................ 4Ocean City, Md.......................... 4Chesapeake Bay, Md..................... 4Virginia Beach, Va....................... 4Cape Hatteras, N.C...................... 8Fort Fisher, N.C......................... 5Myrtle Beach, S.C....................... 2Charleston, S.C.......................... 4Savannah, Ga............................ 4Jacksonville, Fla......................... 4Cocoa Beach, Fla........................ 4Miami Beach, Fla........................ 8

(15), and 20 states, including California, all GulfCoast states, and all East Coast states exceptNew Jersey and Connecticut (J. Farmer, person-al communication). Although isolation of V.vulnificus from the marine environment hasbeen reported, only regional studies of its distri-bution and ecology have been made (10, 17, 18;D. L. Tison, M. Nishibuchi, and R. J. Seidler,Abstr. Annu. Meet. Am. Soc. Microbiol. 1981,Q97, p. 216). We have shown (17; J. D. Oliver,D. R. Cleland, and R. A. Warner, Abstr. Annu.Meet. Am. Soc. Microbiol. 1981, N16, p. 175)that lactose fermentation is extremely commonamong vibrios (as evidenced by the ability tohydrolyze o-nitrophenyl-,-D-galactopyranoside[ONPG]) and is shared by several species. Alongwith V. vulnificus, one of these appears to be anew group of luminescent vibrios (18).We have examined 3,887 sucrose-negative

marine vibrios isolated from several environ-ments along the entire East Coast of the UnitedStates, and we report here on the distributionand ecology of Vibrio vulnificus and other lac-tose-positive, sucrose-negative vibrios fromthese sites.

MATERIALS AND METHODS

Sampling protocol. A total of 17 major sites (Table 1)along the U.S. East Coast from Miami, Fla., toPortland, Maine were sampled during the summer

months (May through August) of 1981. At each site, 2to 10 subsites, representing distinct sampling loca-tions, were studied for a total of 80 subsites (a detaileddescription of the subsites is available from J.D.O.).At each subsite, water was tested for the followingparameters: temperature, dissolved oxygen, salinity,pH, nitrate, phosphate, turbidity, fecal coliforms, totalbacteria, and total vibrios. In addition to the water

samples taken at each subsite, sediment, plankton,and animal samples were taken, homogenized in anartificial seawater solution, and plated onto a marinemedium (MSWYE; 19) for total bacteria and ontothiosulfate-citrate-bile salts-sucrose agar (TCBS; BBLMicrobiology Systems) for total vibrios. Methods em-ployed for obtaining the various samples and examin-ing the different parameters have been published (18).Through the use of a self-contained mobile microbiolo-gy laboratory, bacteriological analysis of all sampletypes (water, plankton, animal, and sediment) wasperformed within minutes of collection. All initialisolation plates were incubated at room temperature(ca. 21 to 27°C) for 16 to 18 h. From TCBS plates, allsucrose-negative colonies were picked, inoculated intophenol red lactose broth (Difco Laboratories; made to1% NaCl [wt/vol]), and incubated at 37°C. Tubes weremonitored at frequent intervals, and lactose-ferment-ing isolates were transferred onto MSWYE slants. Onreturning to our laboratory, isolates were then re-streaked onto TCBS to ensure purity, and the fermen-tation of lactose was confirmed. Strains not giving K/Areaction (6) on TSI medium were discarded, and 47phenotypic traits of the remaining 95 isolates, as wellas 23 reference strains (see Table 9), were determined.Taxonomic methodology was as previously described(18), except that penicillin and colistin sensitivitieswere determined by the disk diffusion method withMueller-Hinton agar (Difco) made up with three-salts(19). Sensitivity to pteridine 0/129 was determined byplacing crystals of the antibiotic directly onto themedium surface. Test results were encoded for numer-ical taxonomic analysis with the Jaccard coefficientand single-linkage clustering and were computed atour laboratory with the program TAXAN 6, developedby R. Colwell, Department of Microbiology, Universi-ty of Maryland at College Park.Mouse lethality. All environmental isolates were

grown in standing culture for approximately 18 h inbrain heart infusion broth (BBL) at 37°C. Intraperito-neal injections (0.5 ml, 108 to 109 cells per ml) weremade in 6- to 8-week-old ICR mice. All referencecultures were similarly tested for lethality.

RESULTSOccurrence and distribution of lactose-ferment-

ing vibrios. Little difference was observed forthe various physical, chemical, or microbiologi-cal parameters monitored at the 80 subsites(Table 2). Vibrios comprised a high percentage(26 to 40%) of the total bacterial populations ofall sample types. A large percentage (40 to 69%)of the vibrios did not ferment the sucrose pres-ent in TCBS agar. These isolates (n, 3,887) werepicked and tested for their ability to fermentlactose. Of the 3,887 sucrose-negative vibrios,163, or 4.2%, were able to ferment lactose, asindicated by acid production in phenol red lac-tose broth. The percentage of the sucrose-nega-tive vibrios which were able to ferment lactosevaried from 1.6 to 9.7%, depending on the typeof sample (Table 2). Standard (Pearson) correla-tion coefficient analysis performed between thebacterial populations and the seven chemical

APPL. ENVIRON. MICROBIOL.


.asm.org/

Dow

nloaded from

http://aem.asm.org/

VIBRIO DISTRIBUTION IN MARINE ENVIRONMENT 987

TABLE 2. Means of environmental parameters examined at 80 sites

Parameter Mean t SEM No. ofsamples'

Water temperature (CC) 25.8 ± 1.82 84

Salinity (0/,,) 25.6 ± 2.81 84

Water pH 7.6 ± 0.56 84

Dissolved oxygen (mg/liter) 6.8 ± 1.48 81

Turbidity (FTU)b 14.7 ± 4.34 81

NO3 (,g/liter) 207 ± 18.8 81

P04 (,ug/liter) 544 ± 25.5 82

WaterFecal coliforms (per 100 ml) 221 ± 28.7 80Total bacteriac 4.7 x 103 ± 9.0 x 101 82Total vibriosd 7.1 X 102 ± 3.5 x 101 81% Vibrios' 31.2 ± 5.73 81% Sucrose-negative vibriosf 51.6 ± 9.34 81% Lactose-positive vibriosg 5.3 ± 3.63 77

SedimentTotal bacteria 2.2 x 10W ± 9.0 x 102 40Total vibrios 6.3 x 10" + 4.7 x 102 42% Vibrios 35 ± 5.9 40% Sucrose-negative vibrios 41 ± 5.2 41% Lactose-positive vibrios 2.5 ± 2.34 43

PlanktonTotal bacteria 5.3 X 104 ± 3.2 x 102 24Total vibrios 5.9 x 103 ± 9.4 x 101 27% Vibrios 40.1 ± 5.88 25% Sucrose-negative vibrios 57.0 ± 5.67 29% Lactose-positive vibrios 1.6 ± 2.30 28

OystersTotal bacteria 7.6 x 106 + 4.4 x 103 7Total vibrios 4.6 x 106 + 3.5 x 103 7% Vibrios 38 ± 6.2 7% Sucrose-negative vibrios 40 + 5.4 7% Lactose-positive vibrios 9.2 ± 4.49 12

Crab and fishTotal bacteria 4.1 x 105 ± 5.9 x 102 3Total vibrios 7.2 x 104 ± 2.0 x 102 2% Vibrios 26 ± 0.8 2% Sucrose-negative vibrios 69 ± 3.6 2% Lactose-positive vibrios 9.7 ± 3.27 lha Number of samples may exceed 80 when a parameter was monitored more than once at a subsite.b FTU, Formazin turbidity units.c Total viable counts (CFU) on MSWYE agar.d Total viable counts on TCBS.e Average of individual [(counts on TCBS)/(counts on MSWYE)] x 100.f [(Sucrose-negative colonies on TCBS)/(Total number of colonies on TCBS)] x 100.g [(Lactose-fermenting sucrose-negative vibrios)/(Total number of sucrose-negative vibrios tested)] x 100.h Exceeds number of total vibrio (TCBS count) samples because of nonquantitative swab samples.

VOL. 45, 1983


.asm.org/

Dow

nloaded from

http://aem.asm.org/


TABLE 3. Correlation studies

Variable pair No. ofP value <a: Correlation

First element Second element cases

Total bacteria in:

Water

Plankton

Oysters

Crabs

Fecal coliforms

Total vibrios in:

Water

Sediment

Plankton

Oysters

% Vibrios in:

Water

Sediment

Plankton

Oysters

NitratesFecal coliforms

SalinityDissolved oxygenTotal bacteria in waterTotal bacteria in sediment

Phosphates

Water temperatureTotal bacteria in sedimentTotal bacteria in water

SalinityDissolved oxygenTurbidityNitratesPhosphatesTotal bacteria in waterTotal vibrios in water

TurbidityNitratesPhosphatesFecal coliformsTotal bacteria in waterTotal bacteria in oysters

TurbidityTotal bacteria in crabs

Water temppH

PhosphatesTotal bacteria in waterTotal bacteria in oysters

Total bacteria in oystersTotal vibrios in oysters

% Vibrios in plankton

Salinity% Vibrios in sediment

Total vibrios in oysters

79 0.04678 0.010

24 0.01522 0.00124 0.00115 0.001

6 0.001

3 0.0323 0.0323 0.013

80 0.00477 0.00277 0.00177 0.00178 0.00178 0.01077 0.001

78 0.03679 0.00179 0.00177 0.00179 0.0017 0.047

41 0.0013 0.033

27 0.00327 0.025

6 0.0017 0.0487 0.001

7 0.0017 0.001

15 0.009

24 0.04815 0.009

6 0.037

0.19040.2618

-0.44560.60980.66200.8930

0.9914

0.99480.9948

-0.9993

-0.2988-0.32390.44250.84560.69040.26180.5376

0.20470.43370.44920.53760.50870.6777

0.49210.9947

0.5188-0.3806

0.99210.67381.0000

0.97240.9713

0.6014

0.34810.6014

0.7680



.asm.org/

Dow

nloaded from

http://aem.asm.org/


TABLE 3-ContinuedVariable pair No. of P value <a: Correlation

First element Second element cases

% Sucrose-negative vibrios in:

Total vibrios in plankton

Total bacteria in crabs

Nitrates

Water tempSalinitypHTotal bacteria in waterTotal vibrios in oysters% Sucrose-negative vibrios in sediment

25 0.050 -0.3373

3 0.036 -0.9938

25 0.017 0.4528

7 0.0177 0.0117 0.0247 0.0236 0.0486 0.028

0.7910-0.82620.76050.7620

-0.73490.7988

% Lactose-positive vibrios in:

Water

Sediment

Plankton

Oysters

Crabs

pHTotal bacteria in oystersTotal vibrios in oysters% Sucrose-negative vibrios in oysters

SalinityTurbidityTotal vibrios in sediment% Sucrose-negative vibrios in oysters

Total bacteria in oystersTotal vibrios in oysters

% Vibrios in sediment% Sucrose-negative vibrios in sediment% Sucrose-negative vibrios in oysters

Salinity

77 0.0497 0.0037 0.0037 0.010

43 0.00142 0.04741 0.0047 0.020

0.18990.90000.90020.8320

-0.49150.26160.40740.7776

4 0.001 1.0004 0.001. 1.000

9 0.0079 0.0366 0.009

0.7731-0.62380.8874

11 0.008 -0.6991a Only P values of <0.05 are included.b Standard correlation coefficient (Pearson).

and physical environmental parameters mea-sured at each site indicated that several signifi-cant correlations exist with total bacterialcounts, total vibrio counts, and fecal coliformpopulations (Table 3). Lactose-fermenting vibrioisolates correlated positively with pH of thewater, total vibrios in the sediments and oysters,and the total viable counts and percentages ofsucrose-negative vibrios associated with oysters(Table 3). Negative correlations were observedbetween salinity of the water column and theincidence of lactose fermenters.

V. vulnificus (identified as described below)was isolated in low numbers (n, 33) throughoutthe eastern seaboard (Table 4, cluster III) andfrom all sample types examined (19 from animalsamples, 4 from sediment samples, 8 from watersamples, and 1 from a Miami plankton sample).The majority of the positive isolates were from

bivalves (16 from 3 different clams and 1 from anoyster), with the remaining 2 isolates comingfrom crabs. Of the four sediment samples inwhich V. vulnificus was found, two were brown-ish ocean beach sand (Jacksonville, Fla., andFlander's Beach, N.Y.), and two were fromriver beach sand (Fort Fisher, N.C.). None ofthe isolates was from an estuarine mud sample.Comparison of the environmental parameters ofthe 8 sites yielding V. vulnificus isolates to thoseof the entire study (80 sites) revealed no signifi-cant differences (Table 5).Taxonomy. All 95 environmental isolates were

Vibrio spp. which fermented lactose. Lactosefermentation (as indicated by an acid reaction inphenol red lactose broth at 37°C) was typicallyquite rapid (an average of 2 h) and frequentlyreverted to a neutral or alkaline pH after over-night incubation (R. Warner, D. Cleland, and J.

Water

Sediment

Plankton

Oysters

VOL. 45, 1983


.asm.org/

Dow

nloaded from

http://aem.asm.org/


TABLE 4. Distribution, geographically and bysample type, of lactose-fermenting vibrio isolates

contained in the five taxonomic clusters

Sample No. of samples in cluster0:sources I II III IV v

Water 4 3 8 2Sediment 4 3 4 2Plankton 3 1 1Animals 2 1 19 11 3

a Sources for clusters: Cluster I: Savannah, Ga.;North Myrtle Beach, S.C.; Fort Fisher, N.C.;Wrightsville Beach, N.C.; Virginia Beach, Va.; Atlan-tic City, N.J.; New York City, N.Y. Cluster II: Miami,Fla.; Jacksonville, Fla.; Savannah, Ga.; Charleston,S.C.; New York City, N.Y. Cluster III: Miami, Fla.;Jacksonville, Fla.; Fort Fisher, N.C.; Cape Hatteras,N.C.; Atlantic City, N.J.; New York City, N.Y.; CapeCod, Mass. (the sample source of one isolate fromCape Cod, Mass., is unknown). Cluster IV: CapeHatteras, N.C. Cluster V: Cape Hatteras, N.C.

Oliver, Abstr. Annu. Meet. Am. Soc. Microbiol.1982, N106, p. 195). Thus, numerous lactose-fermenting isolates would have been missed if asingle observation of the lactose reaction hadbeen made after overnight incubation.Numerical taxonomy performed on the iso-

lates resulted in five major clusters (defined asgroups containing five or more isolates) com-prising 70 of the environmental strains (Fig. 1).Clusters IV and V contained isolates from sever-al subsites of a single site (Cape Hatteras, N.C.),whereas members of the remaining three clus-ters were not regional (Table 4). None of theisolates, except for members of cluster III,grouped with any of the 23 reference strains. (Asixth grouping, identified on Fig. 1 as Vc + Vm,was comprised only of the four reference strains

of Vibrio cholerae and Vibrio mimicus.) (SeeTable 6 for phenotypic traits of the five clusters.)

Cluster I was comprised of 13 isolates, ob-tained from Savannah, Ga., to New York City,isolated from all sample types (four water, foursediment, three plankton, and two from animals;Table 4). Phenotypically, this cluster consistedof H2S-producing strains and was one of twogroups containing isolates which were biolumi-nescent (Table 6). Lethality studies showed only16.7% to be lethal for mice.

Cluster II contained 8 isolates obtained fromMiami, Fla., to New York City and from allsample types (three water, three sediment, oneplankton, and one animal). Members of thiscluster were similar to cluster I in being H2Spositive but contained no luminescent strains.Along with H2S production, these strains dif-fered from V. vulnificus in their penicillin resist-ance, general lack of salicin fermentation, andability to grow at higher NaCl concentrations.Despite these differences, all eight strains of thecluster showed a high similarity (80 to 90%) to aV. vulnificus type strain (ATCC 27562) appear-ing in cluster III and included in this study asone of the reference cultures. Members of clus-ter II demonstrated one of the highest degrees oflethality to mice observed in this study, with87.5% causing death on peritoneal injection.The largest grouping of environmental isolates

(n, 33) occurred in cluster III, which also con-tained all of the eight reference V. vulnificusstrains. Geographically, cluster III containedisolates from eight sites from Miami, Fla., toCape Code, Mass. (Table 4). The majority (n,19) of the cluster members were from animalsamples, with 8 from water, 4 from sediment,and 1 from a plankton sample. Some 10 of the 33strains of cluster III were selected at random

TABLE 5. Environmental data on sites yielding V. vulnificusMean values for:

Water parameterSites positive for V. vulnificus' All 80 sites

Temperature (°C) 26.3 (19-32) 25.8Salinity (O/O) 22.5 (9-30) 25.6pH 7.4 (6.8-7.8) 7.6Dissolved oxygen (mg/liter) 6.4 (2.5-9.2) 6.8Turbidity (FTU)b 6.4 (<2-22) 14.7NO3 (,ug/liter) 252 (44-1020) 267P04 (,ug/liter) 529 (170-1170) 544Fecal coliforms (per 100 ml) 155 (<1-880) 221Total bacteriac (per ml) 9.6 x 103 (3.4 x 102-4.2 x 104) 4.7 x 103Total vibriosd (per ml) 6.6 x 102 (4 x 101-2.9 x 103) 7.1 x 102% Sucrose-negative vibriose 59 (25-76) 51.6

a Values in parenthesis are ranges observed for each parameter.b Formazin turbidity units.c Total viable counts on MSWYE.d Total viable counts on TCBS.[(Sucrose-negative vibrios on TCBS)/(total colonies on TCBS)] x 100.



.asm.org/

Dow

nloaded from

http://aem.asm.org/


,10 \

,..*d\.1 * e.\

14 do2

9s o w

14

214

14I

I *'

as

I* 4 'jj

8I

210

* .

01

2'

25 '' -' \cc

'* \0

404

21 c.

101 b f6 F d 4

404

2O *.--*

25 *c* .<r-- *c-*c4 4*-i- \

74 *- * 6 S6 '-F

600

64 *74*4 **c@ ||* c.@cc@4400, ..cra.-.r^^44e9cccc

S: \

3

31*c4

34 cOPE0

*.e

' 0-44@-4w64e*tw

34 40S 04-44"-dX"ftS4ftF*P4**f0ta\

Op~~~~~~~~~~~~,,44w _*",4 * e \

.45

40

540

r *c 99'n''

4 \

"0

oo

117

414404

40Ili411

51152

19'

?1144.

*0

*t*44tt91'4?0v*

IV

~09

:99*

O.:~~~.

Vc+Vm*

Pew:

.9f

U..

S0

FIG. 1. Computer printout of numerical taxonomic analysis of 95 sucrose-negative, lactose-fermentingmarine vibrios. Five clusters of the environmental isolates are indicated. Also shown is the clustering of the fourreference strains of V. cholerae and V. mimicus (Vc + Vm). The printout symbol denotes intrastrain similaritiesof .80%o.

VOL. 45, 1983

00


.asm.org/

Dow

nloaded from

http://aem.asm.org/

992 OLIVER, WARNER, AND CLELAND APPL. ENVIRON. MICROBIOL.

TABLE 6. Phenotypic traits' of taxonomic clusters

Presence of trait by cluster no:Phenotypic trait

I (n = 13) II (n = 8) III (n = 33) IV (n = 11) V (n = 5)

Oxidase + + + + +

Catalase + + + + +

Fermentation of:Lactose (24 h)b V (46) + V (33) + +SucroseSalicin V (69) V (38) + + +

Gas from dextrose - - - - (5)

Hydrolysis of:Starch + + (88) + - +Gelatin V (38) + + (76) - (9) - (20)Agar

TSI= K/AC + (80) + + + +

H2S (on TSI agar) + + - (6) V (27) +

Voges-Proskauer

Citrate _ -(13)

Lysine decarboxylase V (69) + + + (91) +

Ornithine decarboxylase V (46) + + (79) V (36) + (80)

Arginine dihydrolase - - - - (9)

Urease - (20) - (13) - + (91) V (60)

Indole V (38) V (50) V (55)

NO3 reduction + + + + +

Growth in NaCl:0O - V (25) - - -3% + + + + +6% + (92) + V (64) + (82) V (60)10% - V (50)

Growth at:S°C. - - - - -

42°C. V (33) + + + V (40)

Luminescence V (54) - - - + (80)

Sensitivity to:Penicillin - (15) V (25) + (88)Pteridine 0/129 + (75) + (75) + V (36) + (80)Colistin V (46) - (13) - + (91)

Pigment

Spreading growth

a AIIisolates were gram-negative rods which fermented glucose. +, -: '97% of strains tested gave indicatedresult if no number is shown in parenthesis. Numbers in parentheses indicate percentage of strains positive forthe trait. +, .70% positive; -, s20% positive; V, Variable result (21 to 69%o positive).

b All strains examined in this study fermented lactose. Lactose reaction indicated here is result observed inphenol red lactose broth after 24 h of incubation at 37°C.

I Alkaline (or no reaction) slant, acid butt in triple sugar iron agar.


.asm.org/

Dow

nloaded from

http://aem.asm.org/


TABLE 7. Lethality in mice of lactose-fermentingvibrios

Group (no. of strains) % Lethality0Cluster I (13) 16.7Cluster II (8) 87.5Cluster III (33) 82.4Cluster IV (11) 18.2Cluster V (5) 0Nonclustering isolates (25) 64H2S-positive isolates (27) 40.7

a A 0.5-ml dose of a stationary-phase culture grownat 23°C in brain heart infusion broth was injectedintraperitoneally into ICR mice.

and examined by D. Tison, Oregon State Uni-versity, Corvallis, Oreg., for genetic homologyto V. vulnificus by DNA-DNA hybridizationstudies. All 10 of the strains were found to be V.vulnificus, with homologies of 77 to 95% atstringent temperatures. Of the 33 members ofcluster III, 82.4% were lethal for mice.Members of cluster IV were all collected from

the Cape Hatteras, N.C., area. Of these isolates,11 were obtained from animals, and the remain-ing 2 were from sediments (Table 4). Phenotypi-cally, these isolates differed significantly fromV. vulnificus in their production of urease, in thegeneral lack of extracellular enzymes, and intheir pattern of antibiotic sensitivity (Table 6).Several members produced H2S in TSI medium.Lethality studies showed that only 15.4% werelethal for mice.

Cluster V contained five members which wereisolated from waters from Cape Hatteras, N.C.This cluster could be characterized by the pro-duction of H2S by all members and of urease bythree of the five. In addition, most of the isolateswere observed to be luminescent. Cluster Vwas, therefore, phenotypically similar to clusterI. None of the strains was lethal for mice.Of the 25 lactose-fermenting isolates which

did not form clusters, 16 (64%) were also lethalfor mice (Table 7).

DISCUSSIONDistribution of V. vulnficus. In this study of 80

coastal sites from Miami, Fla., to Portland,Maine (Table 1), an average count (on the ma-rine medium MSWYE) of 4.7 x 103 bacteria perml of seawater was obtained (Table 2). Totalbacteria in the water column were seen to corre-late with two of the pollution indicators moni-tored, nitrates and fecal coliforms (Table 3).Whereas the number of bacteria associated withplankton increased as the total population in thewater column increased, a negative correlationwith salinity was observed (Table 3). This result

is in accord with the study reported by Kanekoand Colwell (9) on bacteria-plankton associa-tions in Chesapeake Bay. The validity and effec-tiveness of the methods used for monitoring thevarious environmental parameters and studyingtheir correlations is indicated by the correlationsobserved with fecal coliforms in the water col-umn (Table 3). As would be expected, positivecorrelations were obtained for the presence offecal coliforms and four water pollution indica-tors, namely, turbidity, nitrates, phosphates,and total bacteria. Also as would be expected,negative correlations were seen between num-bers of fecal coliforms and the salinity anddissolved oxygen of the water column. A signifi-cant portion (26 to 30%) of the total bacterialpopulation were Vibrio spp. (defined as cellsgiving rise to colonies on TCBS medium onovernight incubation). Although our taxonomicstudies have indicated relatively few nonvibriosgrowing on TCBS, we recognize the likelihoodof such bacteria appearing on this medium. Theactual number of vibrios present in the watercolumn may, therefore, be less than that report-ed here. Our observations, however, are con-sistent with previous studies (8, for example)which have shown vibrios to make up a highpercentage of the aquatic bacterial population,especially in summer. Several correlations werealso observed with total vibrios from the varioussamples (Table 3). Of note were those of vibriosin the water column and several pollution indica-tors, and of vibrios associated with oysters andthe total number of bacteria in oysters. Thecorrelation between vibrios in the water columnand water temperature reported by Kaneko andColwell (8) was not observed in the presentstudy but can be explained by the lack of varia-tion in water temperature (Table 2). The totalnumber of bacteria and the percentage of vibriosin the non-filter-feeding animals examined (twocrabs and one fish) were considerably less (Ta-ble 2).The observation that 4.2% of all sucrose-

negative vibrios tested were able to dissimilatelactose is similar to that recently reported byJ. T. Graikoski and J. E. Houser (Abstr. Annu.Meet. Am. Soc. Microbiol. 1982, Q69, p. 221),who found 2 to 5% of all bacterial isolates fromseawater, sediment, and animals to be able toferment this sugar. Of the lactose-fermenters westudied, a total of 33 strains (20%) of the su-crose-negative, lactose-positive vibrios wereidentified as V. vulnificus by numerical taxo-nomic methods. Identification of these strainswas confirmed by DNA-DNA homology stud-ies. Thus, the 33 isolates of V. vulnificus consti-tuted only 0.85% of the 3,887 sucrose-negativevibrios examined. In a previous study (18), weobserved that over 42% of all sucrose-negative

VOL. 45, 1983


.asm.org/

Dow

nloaded from

http://aem.asm.org/

994 OLIVER, WARNER, AND CLELAND APPL. ENVIRON. MICROBIOL.

TABLE 8. Comparison of environmental (n = 33) and clinical isolates of V. vulnificus% Positive

Taxonomic character Environmental Clinical strainsstrains Hollis et al. (6) Present studya

Gram-negative rods 100 100 100

Motile 100 100 100

Oxidase 100 100 100

Fermentative metabolism 100 100 100

Acid from:Lactose 100 81 l00bSucrose 0 3 0Dextrose 100 100 100Salicin 100 100 100

Gas from dextrose 0 0 0

Hydrolysis of:Starch 97 100 90Gelatin 76 97 100Agar 0 NRC 0

TSI= K/Ad 97 100 90

H2S production (TSI agar) 6 0 0

Voges-Proskauer 0 0 0

Decarboxylation of:Lysine 100 97 100Ornithine 79 66 100

Arginine dihydrolase 0 0 0

NO3- reduction 100 100 100

N02- reduction 0 NR 0

Growth in NaCl:0o 3 NR 703% 100 NR 1006% 64 100 1008% 9 8 3010%o 0 0 10

Growth at:50C 0 NR 40250C 100 NR 100370C 100 NR 100420C 100 NR 100

Catalase 100 NR 100

Urease 0 0 0

Indole 55 97 0

Citrate 0 76 0

Luminescence 0 NR 0

Pigment 0 NR 0


.asm.org/

Dow

nloaded from

http://aem.asm.org/


TABLE 8-Continued% Positive

Taxonomic character Environmental Clinical strainsstrains Hollis et al. (6) Present studya

Spreading growth 0 NR 20

Sensitivity to:Penicillin 88 100 100Pteridine 0/129 100 NR 100Colistin 0 0 0a Eight clinical strains received from the Centers for Disease Control were studied.b Some strains negative by 24 h; see text.c NR, Not reported.d Alkaline (or no reaction) slant, acid butt on triple sugar iron agar.

marine vibrios tested were positive for ONPGhydrolysis. Of these, only 3 to 4% could beidentified as V. vulnificus. It is evident, there-fore, that as we have previously emphasized (17,18), numerous lactose-fermenting Vibrio speciesother than V. vulnificus exist in the marineenvironment. A similar conclusion was reachedby Tison et al. (25; D. L. Tison, M. Nishibuchi,and R. J. Seidler, Abstr. Annu. Meet. Am. Soc.Microbiol. 1981, Q97, p. 216), who describedseveral taxonomically undefined groups of lac-tose-fermenting marine vibrios and showedthem to be genetically distinct from V. vulnifi-cus.

In the present study, V. vulnificus was isolat-ed from water, sediment, plankton, and animalsamples obtained from widespread geographicalregions of the East Coast of the United States(Table 4). The isolation of sucrose-negative,lactose-fermenting vibrios correlated with sever-al parameters (Table 3), but as observed byTamplin et al. (M. L. Tamplin, C. E. McKnight,G. E. Rodrick, and N. E. Blake, Abstr. Annu.Meet. Am. Soc. Microbiol. 1981, Q72, p. 212),no correlation with fecal coliform levels wasseen. Nor were any statistically significant dif-ferences observed between the values recordedfor the various environmental parameters ofthose subsites which yielded V. vulnificus andthe mean values of all 80 subsites (Table 5).Numerical taxonomic analysis of the 95 su-

crose-negative, lactose-positive isolates resultedin five major clusters (Fig. 1) comprising 70 ofthe 95 strains.

Cluster I, unusual in containing H2S-, urease-,and luminescence-positive strains, were < 70osimilar to all of the 23 reference strains includedin the analysis and were unlike any describedVibrio spp. However, when several strains fromclusters Ila and IIb from our previous study (18)of ONPG-hydrolyzing marine vibrios were in-cluded in the taxonomic analysis, a high similar-ity (80 to 100%) to these strains was observed.

Clusters Ila and IIb of that study also consistedof strains that were luminescent and H2S posi-tive, with some urease producers. Lethality formice was also similar for the two groupings(16.7% for cluster I of the present study and 22%of the clusters Ila and Ilb of the previous study).As previously concluded (18), these isolatesappear to represent a new taxonomic group ofvibrios.

Like cluster I strains, all members of cluster IIwere H2S positive. They differed from cluster I,however, in their gelatin, decarboxylase, salttolerance, and growth temperature responses(Table 6). None of the cluster II members wasluminescent. Also like cluster I, none of thereference strains appeared in this cluster, norwere its members similar to published descrip-tions of Vibrio spp. When taxonomic data fromstrains described as clusters hIa and lIb in aprevious study (18) were added to the presentnumerical taxonomic analysis, these H2S-pro-ducing strains appeared with this H2S-producingcluster. When potential pathogenicity is consid-ered, this cluster is quite significant, with over80%o of the member strains causing death in micewithin 24 h.The phenotypic traits of the 33 members of

cluster III were typical of V. vulnificus, and thisidentification was confirmed by DNA hybridiza-tion studies performed on 10 randomly selectedmember strains. The cluster contained all eightreference V. vulnificus strains and demonstrateda high percentage of similarity to members ofcluster lb (also identified as V. vulnificus) of ourprevious study (18). Cluster III could be easilydifferentiated from the other four clusters ob-tained in the present study (Table 6) and wasnearly identical phenotypically to the 38 clinicalstrains of V. vulnificus described by Holfis et al.(6), several of which have been characterized inour laboratory (Table 8). Tison and Seidler (25)have also demonstrated that V. vulnificus strainsisolated from a wide range of estuarine environ-

VOL. 45, 1983


.asm.org/

Dow

nloaded from

http://aem.asm.org/


ments are genetically indistinguishable fromclinical isolates. Differences between our envi-ronmental isolates and the clinical strains stud-ied by Hollis et al. (6) were observed in theability to grow on citrate as a sole carbonsource, NaCl tolerance, and indole formation.The negative citrate reaction observed for the 33cluster III isolates, however, agrees with theresult we have observed for V. vulnificus refer-ence strains using Simmons citrate medium (Ta-ble 8) and with our previous observations (18). Anegative citrate reaction was also reported byMatsuo et al. (14). The reason for this discrepan-cy is not known. We also found that strains fromthe Centers for Disease Control (Atlanta, Ga.)are able to grow over a wider range of NaClconcentrations than that reported by Hollis et al(6), although the determination of this trait isquite subjective. Differences in indole produc-tion between the clinical and environmental iso-lates were also noted, although negative reac-tions for this test have also been previouslyreported (15). Two strains in cluster III failedboth to produce indole and to decarboxylateornithine; thus, they were similar to V. vulnifi-cus biogroup 2 recently described by Tison et al.(24). These two strains differed from this newbiogroup, however, in their ability to grow at420C.Another aspect of the taxonomic identifica-

tion of V. vulnificus which deserves mentionconcerns the time required for lactose fermenta-tion. Hollis et al. (6) described their 38 clinicalisolates as usually fermenting lactose within 24h, although a few isolates required 3 to 7 days.Baumann et al. (1) have stated that in fact wild-type strains of V. vulnificus are unable to utilizelactose (20) and that lactose fermentation isdetected only after 1 to 3 days of incubationbecause of spontaneously arising mutants whichare capable of fermenting this sugar. All of our(presumably wild-type) isolates, on the otherhand, were capable of producing an acid reac-tion in phenol red lactose broth within severalhours of inoculation (R. Warner, D. Cleland,and J. Oliver, Abstr. Annu. Meet. Am. Soc.Microbiol. 1982, N106, p. 195). This finding hasnot been reconciled with that of Baumann et al.(1), although the discrepancy might conceivablybe due to differences in the medium employedby those authors to determine the ability toferment lactose.

Cluster IV also contained several H2S-pro-ducing strains but could be distinguished fromother clusters by the production of urease by itsmembers (Table 6). Urease production is highlyunusual among vibrios, having been describedonly for Vibrio damsela (13), occasional strainsof Vibrio parahaemolyticus (16), and an un-named vibrio causing infections in flounder (21).

TABLE 9. Reference strains employed in taxonomicand lethality studies

Reference strain and sourcea Lethalityb Strain no.Referencestrain ~~~~~~(Fig.1)V. vulnificus:ATCC 27562ATCC 29306; CDC A1402ATCC 29307CDC B3547CDC C2756CDC C7184CDC C8806CDC D9889CDC E2272CDC H3308J. Oliver, butcher

V. cholerae:CDC C4752CDC C6487Lab strain

V. mimicus CDC C6713V. parahaemolyticus:ATCC 27519K. Nealson, B113

Vibrio alginolyticus K.Nealson, B86

Vibriofischeri K. Nealson,B64

Vibrio harveyi:K. Nealson, B332K. Nealson, B352K. Nealson, B376

V. Neptuna, ATCC 25919Vibrio marinofulvusd ATCC

14395Vibrio algosuSd ATCC 14390Vibrio mrannagilisd ATCC

14398Photobacterium leiognathi

K. Nealson, B474Photobacterium phosphore-um K. Nealson, B404

Aeromonas hydrophilaATCC 9071

+ 103+ 115+ 102+ 99+ 100+ 944+ 104+ 101+ NRC+ NR+ NR

+ 106+ 107+ 108+ 105

+ NR+ 111- 116

- 110

- 117- NR- 118- 97- 98

NTe 114- 113

- 109

- NR

+ 112

a ATCC, American Type Culture Collection, Rock-ville, Md.; CDC, Centers for Disease Control.

b Examined as described in Table 7.c NR, Not a reference strain in the numerical taxo-

nomic study.d Not a recognized scientific name.NT, Not tested.

However, members of cluster IV differed fromV. damsela and V. parahaemolyticus in numer-ous phenotypic traits, including lactose fermen-tation, H2S production, lack of gas production incarbohydrate broth, and salicin fermentation,and from the flatfish vibrio in the reactions tosucrose, indole, and starch and in penicillinsensitivity. No similarity to the ONPG-hydro-lyzing vibrios studied previously (18) could bedemonstrated for cluster IV, and the identity ofthis group was undetermined.



.asm.org/

Dow

nloaded from

http://aem.asm.org/


Cluster V also contained H2S- and urease-producing members, most of which were lumi-nescent (Table 6). Except for our previous study(18), H2S-producing luminescent bacteria havenot been previously described (18, 20). Membersof cluster lIb of that study also contained H2S-and urease-producing strains and were lumines-cent. In neither study did our isolates clusterwith the luminescent reference strains includedfor comparison. We were not able to identifythese isolates as being members of a describedVibrio species, and we suggest that they, alongwith the luminescent strains of cluster I, repre-sent a new taxonomic group of marine vibrios.The fact that four of the five taxonomic clus-

ters described here could not be identified is notunusual. Graikoski and Houser (Abstr. Annu.Meet. Am. Soc. Microbiol. 1982, Q69, p. 221)recently reported that ca. 60% of all isolatesfrom marine waters, sediments, and animalscould not be identified. One aspect of the pres-ent study which has made specific identificationdifficult was the frequent occurrence of H2S-producing isolates among the 95 lactose-fer-menting vibrios. Although strains producingH2S from thiosulfate have been reported for V.cholerae and V. parahaemolyticus (4), this traithas not generally been observed among vibrios(22, 26; P. A. West, Ph.D. thesis, University ofKent, Canterbury, England, 1980). The possibil-ity that the H2S trait may be plasmid mediated iscurrently being investigated in our laboratory.What is especially relevant to the present studyis the extent to which these undescribed specieswere virulent for mice. When even large inoculaof the various reference strains were injectedintraperitoneally into mice, more than half of thespecies failed to cause death (Table 9). Thus, itis evident that the technique employed can dif-ferentiate the environmental isolates and refer-ence strains tested in our study regarding viru-lence for laboratory animals (Tables 7 and 9).Although mouse lethality as determined by themethod used in this study cannot be equated tohuman pathogenicity, the lethality studies re-ported here indicate that numerous lactose-fer-menting marine vibrios, in addition to V. vulnifi-cus, are potentially pathogenic for humans.Among these unidentified Vibrio species are alarge percentage (28.4%) of H2S producerswhich were, as a group, quite virulent for mice(Table 7). Tison and Seidler (25) concluded thatenvironmental isolates of V. vulnificus may be ofpublic health significance. We agree with thatconclusion and suggest that although considera-bly more work is warranted on all aspects of thebiology of V. vulnificus, the other taxonomicallyundefined clusters of lactose-fermenting vibriosdescribed here deserve particular attention aspossible public health hazards.

ACKNOWLEDGMENTSWe especially thank Dave Tison, Oregon State University,

for performing the DNA homology studies and for valuablediscussions. We also thank Jim Kaper and Ken Nealson forsome of the reference strains employed.This study was supported by Public Health Service grant Al

17059-02 from the National Institute of Allergy and InfectiousDiseases.

LITERATURE CITED

1. Baumann, P., L. Bauman, and B. G. Hall. 1981. Lactoseutilization by Vibrio vulnificus. Curr. Microbiol. 6:131-135.

2. Blake, P. A., M. H. Merson, R. E. Weaver, D. G. Hollis,and P. C. Heublein. 1979. Disease caused by a marinevibrio. N. Engl. J. Med. 300:1-5.

3. Bowdre, J. H., M. D. Poole, and J. D. Oliver. 1981. Ede-ma and hemoconcentration in mice experimentally infect-ed with Vibrio vulnificus. Infect. Immun. 32:1193-1199.

4. Colwell, R. R. 1970. Polyphasic taxonomy of the genusVibrio: numerical taxonomy of Vibrio cholerae, Vibrioparahaemolyticus, and related Vibrio species. J. Bacteri-ol. 104:410-433.

5. Ghosh, H. K., and T. E. Bowen. 1980. Halophilic vibriosfrom human tissue infections on the Pacific coast ofAustralia. Pathology 12:397-402.

6. Hoilis, D. G., R. E. Weaver, C. N. Baker, and C. Thorns-berry. 1976. Halophilic Vibrio species isolated from bloodcultures. J. Clin. Microbiol. 3:425-431.

7. Johnson, D. E., and F. M. Calia. 1981. Hemolytic reactionof clinical and environmental strains of Vibrio vulnificus.J. Clin. Microbiol. 14:457-459.

8. Kaneko, T., and R. R. Colweli. 1973. Ecology of Vibrioparahaemolyticus in Chesapeake Bay. J. Bacteriol.113:24-32.

9. Kaneko, T., and R. R. Colwell. 1975. Adsorption of Vibrioparahaemolyticus onto chitin and copepods. Appl. Micro-biol. 29:269-274.

10. Kelly, M. T., and D. M. Avery. 1980. Lactose-positiveVibrio in seawater: a cause of pneumonia and septicemiain a drowning victim. J. Clin. Microbiol. 11:278-280.

11. Kelly, M. T., and W. F. McCormick. 1981. Acute bacteri-al myositis caused by Vibrio vulnificus. J. Am. Med.Assoc. 246:72-73.

12. Kreger, A., and D. Lockwood. 1981. Detection of extracel-lular toxin(s) produced by Vibrio vulnificus. Infect. Im-mun. 33:583-590.

13. Love, M., D. Teebken-Fisher, J. E. Hose, J. J. Farmer III,and G. Richard Fanning. 1981. Vibrio damsela, a marinebacterium causes skin ulcers on the damselfish Chromispunctipinnis. Science 214:1139-1140.

14. Matsuo, T., S. Kohno, T. Ikeda, K. Saruwatari, and H.Ninomiya. 1978. Fulminating lactose-positive vibrio septi-cemia. Acta Pathol. Jpn. 28:937-948.

15. Mertens, A., J. Nagler, W. Hansen, and E. Gepts-Frieden-reich. 1979. Halophilic, lactose-positive Vibrio in a case offatal septicemia. J. Clin. Microbiol. 9:233-235.

16. Oberhofer, T. R., and J. K. Podgore. 1982. Urea-hydro-lyzing Vibrio parahaemolyticus associated with acutegastroenteritis. J. Clin. Microbiol. 16:581-583.

17. Oliver, J. D. 1981. The pathogenicity and ecology ofVibrio vulnificus. Mar. Technol. Soc. J. 15:45-52.

18. Oliver, J. D., R. A. Warner, and D. C. Cleland. 1982.Distribution and ecology of Vibrio vulnificus and otherlactose-fermenting marine vibrios in coastal waters of thesoutheastern United States. Appl. Environ. Microbiol.44:1404-1414.

19. Poole, M. D., and J. D. Oliver. 1978. Experimental patho-genicity and mortality in ligated ileal loop studies of thenewly reported halophilic lactose-positive Vibrio sp. In-fect. Immun. 20:126-129.

20. Rekchelt, J. L., and P. Baumann. 1973. Taxonomy of themarine, luminous bacteria. Arch. Mikrobiol. 94:283-330.

VOL. 45, 1983


.asm.org/

Dow

nloaded from

http://aem.asm.org/


21. Robohm, R. A., and C. Brown. 1978. A new bacterium(presumptive Vibrio species) causing ulcers in flatfish.Mar. Fish. Rev. 40:4-7.

22. Shewin, J. M., M. Veron, R. H. W. Schubert, and M. S.Hendrie. 1974. Family II. Vibrionaceae Veron 1%5, 5245,p. 340-352. In R. E. Buchanan and N. E. Gibbons (ed.),Bergey's manual of determinative bacteriology, 8th ed.The Williams & Wilkins Co., Baltimore.

23. Smith, G. C., and J. R. Merkel. 1982. Collagenolyticactivity in Vibrio vulnificus: potential contribution to itsinvasiveness. Infect. Immun. 35:1155-1156.

24. Tison, D. L., M. Nishibuchi, J. D. Greenwood, and R. J.Seidler. 1982. Vibrio vulnificus biogroup 2: new biogroup


pathogenic for eels. Appl. Environ. Microbiol. 44:640-646.

25. Tison, D. L., and R. J. Seidler. 1981. Genetic relatednessof clinical and environmental isolates of lactose-positiveVibrio vulnificus. Curr. Microbiol. 6:181-184.

26. Wachsmuth, I. K., G. K. Morris, and J. C. Feeley. 1980.Vibrio, p. 226-234. In E. H. Lennette, A. Balows, W. J.Hausler, Jr., and J. P. Truant (ed.), Manual of clinicalmicrobiology, 3rd ed. American Society for Microbiology,Washington, D.C.

27. Wright, A. C., L. S. Simpson, and J. D. Oliver. 1981. Roleof iron in the pathogenesis of Vibrio vulnificus infections.Infect. Immun. 34:503-507.


.asm.org/

Dow

nloaded from

http://aem.asm.org/

Distribution vulnificus and Lactose- Fermenting Vibrios ...aem.asm.org/content/45/3/985.full.pdf ·...

Documents

Transcript of Distribution vulnificus and Lactose- Fermenting Vibrios ...aem.asm.org/content/45/3/985.full.pdf ·...