Vol. 57, No. 9APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1991, p. 2750-27570099-2240/91/092750-08$02.00/0Copyright © 1991, American Society for Microbiology

Characterization of Vibrio anguillarum and Closely Related SpeciesIsolated from Farmed Fish in Norway

EGIL MYHR,l* JENS LAURITS LARSEN,2 ATLE LILLEHAUG,l ROAR GUDDING,lMARIANNE HEUM,' AND TORE HASTEIN'

National Veterinary Institute, P.O. Box 8156 Dep, N-0033 Oslo 1, Norway,' and Department of VeterinaryMicrobiology, Royal Veterinary and Agricultural University, DK-1870 Fredriksberg, Denmark2

Received 27 February 1991/Accepted 10 July 1991

A total of 264 bacterial strains tentatively or definitely classified as Vibrio anguillarum were examined. Thestrains were isolated from diseased or healthy Norwegian fish after routine autopsy. With the exception of fiveisolates from wild saithe (Pollachius virens), the strains originated from nine different species of farmed fish.The bacteria were subjected to morphological, physiological, and biochemical studies, numerical taxonomicalanalyses, serotyping by slide agglutination and enzyme-linked immunosorbent assay, DNA-plasmid profiling,and in vitro antimicrobial drug susceptibility testing. The results of the microbiological studies were correlatedto anamnestic information. The bacterial strains were identified as V. anguillarum serovar 01 (n = 132),serovar 02 (n = 89), serovar 04 (n = 2), serovar 08 (n = 1), and not typeable (n = 1) as well as Vibriospkndidus biovar I (n = 36) and biovar II (n = 1), Vibrio tubiashii (n = 1), and Vibrio fischerii (n = 1). V.anguillarum serovar 01 or 02 was isolated in 176 out of 179 cases of clinical vibriosis in Atlantic salmon (Salmosalar). V. anguillarum serovar 01 was the only serovar isolated from salmonid fish species other than Atlanticsalmon, while V. anguillarum serovar 02 was isolated from all marine fish suffering from vibriosis. A 48-Mdaplasmid was isolated from all V. anguillarum serovar 01 isolates examined. Serovar 02 isolates did not harborany plasmids. Resistance against commonly used antibiotic compounds was not demonstrated among V.anguillarum isolates. Neither V. splendidus biovar I nor other V. anguillarum-related species appeared to be ofclinical importance among salmonid fish. However, such bacteria were isolated from diseased turbot(Scopthalmus maximus) and sea bass (Dicentrarchus labrax). The precise role of these bacteria as fish pathogenshas to be elucidated. Cell and culture morphology, cell motility, hemolysis patterns, and especiallyalginate-degrading ability were found to be suitable characteristics for the differentiation of V. splendidusbiovar I within the arginine decarboxylase-positive group of Vibrio species. Luminescence could not bedemonstrated among V. splendidus isolates.

Vibriosis due to Vibrio anguillarum is one of the mostimportant bacterial infections in fish throughout the world (1,19). The disease has great importance, particularly for ma-rine fish farming (4, 15, 55). Infections have been reportedalso among farmed bivalve mollusks and crustaceans (11,12). V. anguillarum and closely related bacterial species arecommonly found in estuarine and coastal marine habitatsand can readily be isolated from different environmentalsources (31-34, 52, 65, 67). These bacteria constitute part ofthe normal microflora of healthy marine fish (34, 41, 45).On the basis of biochemical and serological differences,

two distinct biotypes of V. anguillarum were described (6, 8,19). Biotype 2 was later differentiated as a new species,Vibrio ordalii (54). A total of 10 different 0 serovars (01 to010) and additional 0 subgroups and K serotypes have beendescribed among V. anguillarum strains (38, 46-50, 58). V.anguillarum 01 and 02 are the most common serovarsassociated with vibriosis in farmed and feral fish worldwide(57, 58, 63, 64). Environmental strains of V. angutillaruimusually belong to 0 serovars 03 to 010 or are not typeable(34, 58).Other Vibrio species such as V. tubiashii (22), V. splendi-

dus, V. pelagius (9), and so-called V. angiillarum-likebacteria or unidentified V. anguillarum-related (VAR) spe-cies (13, 31, 32), commonly considered environmentalstrains without pathogenic importance, have recently been

* Corresponding author.

associated with disease in marine fish and shellfish (21, 28,30, 37, 37a, 63).

In Norway, V. anguillarum was isolated for the first timein 1964 from diseased farmed rainbow trout (27). Since then,vibriosis has been a significant problem in Norwegian aqua-culture (16, 17, 24, 25, 43, 68). In general, vaccination is aneffective prophylactic measure to control vibriosis (19, 55),and it was introduced in Norway in 1977 (23). Vibriosisvaccines employed domestically contain antigen prepara-tions of V. anguillarum serovars 01 and 02. Despite anintensive vaccination campaign, some farms do not vacci-nate and vibriosis is still a common diagnosis. Moreover,some VAR species have been isolated from vaccinated fish.The aim of the present study was to obtain comprehensive

information on vibriosis in Norway by characterization ofisolates of V. anguillarum and closely related bacterialspecies from clinical cases. It was of particular interest toevaluate whether other bacterial strains should be includedin the vibriosis vaccines used. A further purpose was toestablish a taxonomic relationship between V. anguillarumand VAR species.

MATERIALS AND METHODS

Bacterial strains. A total of 264 bacterial strains, provision-ally identified as V. anguillarum, were included in the study.Of these, 198 were collected as part of a monitoring programcovering extensive parts of the Norwegian coastline during a12-month period in 1987 and 1988. The strains and available

2750

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHARACTERIZATION OF V. ANGUILLARUM AND VAR STRAINS 2751

TABLE 1. Occurrence of V. anlgiuillai-iru serovars amongdifferent fish species with vibriosis in Norway in 1975 and 1982

Occurrence of V'. anguillarumV. anguillartinz serovars among:V.fItl<gl(lilbirlttl ~~~~~~~Totalserovar Atlantic Other

salmon salmonids" Nonsalmonids"

01 20 19 1 4002 24 2404 22

Total 46 19 1 66

Rainbow trout (n = 17) and sca trout (n = 2).h Eel.

anamnestic information and data such as time and site oforigin, water type, fish species and age, and mortality ratesand/or pathological findings were collected and recordedeither in conjunction with diagnostic work during diseaseoutbreaks or as part of health monitoring programs atcommercial fish farms. Fourteen and 52 strains isolatedduring 1975 and 1982, respectively, were included in parts ofthe study. These 66 isolates represented the complete V.angiuillarlim bacterial strain collection at the National Vet-erinary Institute, Oslo, Norway, during these years.Of the 264 bacterial strains, 246 had been isolated from

anadromous salmonid fish: Atlantic salmon (Salumno salar)(n = 214), rainbow trout (On orhvnclhus nvykiss) (n = 25).sea trout (Salno trutta) (n = 4), and arctic char (Salvelinaslalpinius) (n = 3). Altogether, 17 strains from marine fish wereincluded in the study: turbot (Scopthlalhus n(aximiiis) (n =

7), saithe (Pollachius 'irens) (n = 5), sea bass (Di(entrtar-chlus labrax) (n = 3), cod (Gadius morhimia) (n = 1). and sole(Solea solea) (n = 1). One strain was isolated from acatadromous fish species, eel (Anguil/la atngiilal) (n = 1).Apart from the five isolates from wild saithe, all strainsoriginated from farmed fish. The origins of the strains aregiven in Tables 1 and 2.

Bacterial isolation and identification. Isolation of bacteriafrom diseased or apparently healthy fish was carried out atlocal fish health laboratories by standard procedures. Sam-ples were taken aseptically from the anterior kidney or othertissues (gills, skin mucus, or intestines), streaked onto bloodagar plates (blood agar base [Difco Laboratories. Detroit.

TABLE 2. Distribution of serovars of V. anguillarumn and VARspecies isolated from various fish species in 1987 and 1988

Distribution of serovars among fish species"Serovar of Other Non-

V. anguillarum Atlantic salmon salmo- salmo- Totaland VAR nids" nids"strains

P H P H P H

01 75 4 12 1 9202 57 1 (02a) 7" 65Other 1(08) 1 (NG)' 2VAR 29 9 1 39

Total 133 35 12 1 16 1 198

P, isolated from diseased fish or fish with pathological lesions: H. isolatedfrom healthy fish.

b Rainbow trout (n = 8). brown trout (n = 2). and arctic char (n = 3).Saithe (n = 5), turbot (n = 7). cod (n = 1). sea bass (n = 3). and sole

(n = 1)."One strain was subtyped as 02b.NG, not groupable.

Mich.] containing 5% citrated bovine blood and 2% NaCl),and incubated aerobically at 22°C for 48 h. Bacterial strainsprovisionally identified as V. aingiuillariimn were submitted tothe National Veterinary Institute. The strains were analyzedas described below. Cell material for the listed assays was

taken from fresh pure cultures grown at 22°C on 2% NaCl-supplemented blood agar plates.

Cell and culture characteristics. Cell morphology and mo-

tility were studied in Gram-stained smears and by dark-fieldmicroscopy. Culture morphology and hemolysis patternswere recorded after incubation on 2% NaCl-supplementedblood agar plates for 2 and 5 days, respectively.

Biochemical and physiological tests: numerical taxonomy.All 264 strains were examined for oxidase and catalaseproduction, reduction of nitrate to nitrite, tolerance to 6%NaCl. oxidative and fermentative degradation of D-glucosewith and without the formation of gas, and dehydrolyzationand decarboxylation of L-arginine and decarboxylation ofL-lysine and L-ornithine according to standard procedures(32, 66).

Forty-one strains that could not directly be identified as V.angiillarutma. i.e., strains showing only narrow or no vibrio-static compound 2,4-diamino-6,7-diisopropyl-pteridine (0/129) inhibition zones on blood agar plates, which did not

agglutinate with V. anguillarim 0 antisera 01, 02, and 03,and which did exhibit cell and culture morphology patternsdifferent from those of V. anigiillaruimnl (34), were subjectedto numerical taxonomic analyses. These 41 strains were

consequently labelled VAR strains throughout the study. Atotal of 25 physiological and biochemical tests were applied(Table 3). Selections of tests were based on a recommendedminimum test set for distinguishing Vibrii phenons proposedby Bryant and coworkers (14). Tests directed at distinguish-ing V. an1giuillaruim1 from closely related Vibrio species were

chosen and performed according to information given byBryant et al. (13, 14), Lee and Donovan (35), West andColwell (66). and Baumann et al. (9). Six type and referencestrains were bioassayed and statistically analyzed in parallel.Tests were scored as - or + and coded 1 = -, 2 = +, andblank = NT (not tested). Statistical analyses and identifica-tion were performed according to the methods of Bryant andcoworkers (14). The concept of the program employed is theprobablistic identification of an unknown with presence-absence data against an established Vibrio identificationmatrix of percent positive characters (14, 56). Identificationis given as the "best identification."

Susceptibility to the vibriostatic agent 0/129. The test was

performed by the disk diffusion method by using Neo-Sensitabs (vibriostatic compound 0/129, 150 ,ug) (A/SRosco, Taastrup, Denmark). The test strains were floodinoculated onto blood agar plates containing 2% NaCl. Forbacterial strains exhibiting narrow or no inhibition zones on

blood agar, i.e., a majority of the VAR strains, the test was

repeated with Mueller-Hinton agar containing 2% NaCl,0.02% MgCl, and 5% lysed horse blood agar. The formermedium consisted of 10 g of tryptone, 5 g of yeast, 20 g ofNaCl, 15 g of agar, and 60 ml of sterile lysed horse blood in950 ml of aqua destilata. The plates were read after 1 to 5days of incubation at 22°C.

Serogrouping. All isolates were grouped by slide aggluti-nation by using 0 antisera produced in rabbits against V.angitillaI-rumn serovars 01, 02, and 03. Strains used for 0antiserum production were as follows: 01, VI/FT 1197(Culture Collection, National Veterinary Institute); 02,VI/FT 1445 (Culture Collection, National Veterinary Insti-tute): and 03, ATCC 43307 (American Type Culture Collec-

VOL. 57, 1991

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2752 MYHR ET AL.

TABLE 3. Biochemical characterization of 41 VAR strains and 6 type and reference strains of related Vibrio speciesa

Results for:

VAR strains Type and reference strains

Characteristic V. splen- . . V. splen- V. anguil- V. splen- V. splen- V. pe- V. pe-didus V. anguil- V. tubi- V. fischeri didus bv. V angu ar

didus didus lagius lagiusbvVI larum ashl .n larumnl rovarum4e v. I bv. II bv. I bv.aIdid36)s (n = 2) (n = 1) (n6= ATCC NCMB 1T NCMB NCMB NCMB(n =36) (n1 30 CB 1251 19j0T 2253

Arginine (M0ller) + + + - + + + + - -

Lysine (M0ller) - - - - - - - - - - -Ornithine (M0ller)

Nitrate to nitrite + + + + + + + + + + +Voges-Proskauer 4/36 1/2 - - - + + - -Hemolyses sheep + + + + + + + + + + +RBC

Sucrose, acid 9/36 1/2 + + - + + - - + +L-arabinose, acid - - - - - + + - - +

Growth on:L-Arabinose - - - - - + +Gluconate 8/36 1/2 + - - + + - - + +D-Glucuronate 4/36 1/2Sucrose 9/36 + + - - + + - - + +Xylose -

Inositol

Motility + + + + + + + + + + +SwarmingLuminescence - - - - - - - +Oxidase + + + + + + + + + + +Glucose, gasO/129 (150 ,ug/ml) S S S S S S S S S S S

Hydrolysis of:Alginate 34/36 - - - - - - + - - +Chitin + + - - + + + + + - +Gelatin + + + - + + + + + - +Starch + + + + + + + + + + +Tween 80 + + + + + + + + + + +

a Abbreviations: T, type strain; RBC, erythrocytes; S, sensitive.

tion). The 0 serogroups of the Norwegian V. anguillarumisolates VI/FT 1197 and VI/FT 1445 had previously beendetermined by the use of reference antisera (58). The prep-aration of antigens for immunization, the production ofantisera, and the slide agglutination test were performed asdescribed by S0rensen and Larsen (58). The bacterial cellswere not pretreated prior to the agglutination assay. Allisolates serogrouped as V. anguillarum 02 by slide aggluti-nation were serologically retested in an enzyme-linked im-munosorbent assay with a V. anguillarum serovar 02-specific monoclonal antibody, 02-T2.2.178 (60). Bacterialisolates which could not be grouped by the 01, 02, and 03antisera and isolates showing cross-reactions in the slideagglutination assay or unexpected results in the numericaltaxonomy analysis were retested serologically by slide ag-glutination by using V. anguillarum 0 antisera 03 to 010 orV. anguillarum-absorbed antisera raised against the sub-types 02a and 02b (47, 48, 50).

Pathogenicity. The pathogenic importance of the bacterialisolates was evaluated on the basis of systematic informationgiven in the anamnestic reports. Isolates were designatedpathogenic if they were isolated in large numbers and pureculture from the pronephros of diseased fish from epizooticsand/or from individuals exhibiting pathological findings char-

acteristic of hemorrhagic septicemia. Bacteria isolated inlow numbers from scantily growing mixed cultures from thekidneys, integument, or mucosal surfaces from healthyindividuals investigated as part of a health monitoring pro-gram or isolates from diseased or dead fish with a differentmain diagnosis were designated casual isolates.Plasmid screening. DNA-plasmid patterns were deter-

mined in 60 strains: 10 V. anguillarum 01 strains, 9 V.anguillarum 02 strains, and all 41 VAR strains (2 of whichwere later designated V. anguillarum serovars 02a and 08).All selected strains were isolated in 1987 and 1988 andoriginated from five different coastal counties. The V. an-guillarum serovar 01 strains were isolated from diseasedAtlantic salmon (n = 7), rainbow trout (n = 2), and arcticchar (n = 1). The serovar 02 isolates were collected fromdiseased Atlantic salmon (n = 7) and wild saithe (n = 2).The alkaline lysis method described by Maniatis et al. (39)

was employed, with minor modifications. Bacteria weregrown in 5 ml of brain heart infusion broth (Difco) containing2% NaCl at 22°C for 48 h on a shaker. Plasmid-DNAelectrophoresis was run on a vertical apparatus with a 1%agarose gel and Tris-borate-EDTA buffer. Plasmids fromEscherichia coli V517, PDK9, Sa, and Rl were used as size

APPL. ENVIRON. MICROBIOL.

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHARACTERIZATION OF V. ANGUILLARUM AND VAR STRAINS 2753

markers. Plasmid sizes were calculated according to themethod of Rochelle et al. (53).

In vitro antimicrobial susceptibility. MICs of five antimi-crobial drugs for 58 of the original 60 bacterial strainsexamined for plasmid-DNA profiles were determined. TwoVAR strains, later designated V. angiiillarlim serovars 02aand 08, were not included in this part of the study. Theantimicrobial compounds tested were enrofloxacin, flume-quine, oxolinic acid, nitrofurazolidone, oxytetracycline, andtrimethoprim-sulfadiazine. MICs were determined by a drugmicrodilution method as described previously (40).

RESULTS

Cell and culture characteristics, biochemical tests, and0/129 resistance. All 264 strains showed identical key phe-notypical features, i.e., motile, oxydase- and catalase-posi-tive, gram-negative straight and curved rods which degradedD-glucose fermentatively without gas production. All grew in1% peptone broth containing 6% NaCl and showed, with oneexception, a positive L-arginine dihydrolase-decarboxylasereaction and negative reactions for L-lysine and L-ornithinedecarboxylation. All isolates reduced nitrate to nitrite andwere sensitive to the vibriostatic agent 0/129.

Regarding 0/129 susceptibility, inconclusive results wereinitially obtained for the majority of VAR strains. With bloodagar or Mueller-Hinton agar containing NaCl and MgCl2,narrow or no inhibition zones were demonstrated. Withlysed horse blood agar, comparatively distinct and stableinhibition zones could be observed.

Significant differences between two main groups of bacte-ria were observed concerning cell and culture morphology,motility patterns, and agglutination against V. angiuillariin 0antisera. Bacterial isolates with cell and culture morphologyand hemolysis patterns resembling those of V. anguiillarirn(34) and which also agglutinated with V. angutillarium antise-rum 01 or 02 were designated V. anguillaruim. Theseorganisms were short, straight or curved rods showing rapidmotility. The colonies were circular, regular, medium sized(2 mm at 48 h), low convex, shiny, and semitransluscent,with an unguent consistency. Young colonies had a yellow-ish, golden color which changed to brownish in older cul-tures. Two different hemolysis patterns could be observed:(i) a narrow, largely subcolonial, clear P3-hemolytic zone and(ii) a broader semitransparent x-hemolytic zone which be-came clear.

Altogether, 41 isolates showing no or only some similar-ities with V. anguillarium in terms of cell and culture mor-phology, patterns of hemolysis, and agglutinability by V.angiuillarium antisera 01 to 03 were classified as VARstrains. In most cases, these strains exhibited long, straightor curved, plump sausagelike cells, single or in short chainsand with slow, rotating movements. Short, coccoid rodswith moderate motility were also displayed. The colonieswere mostly circular, convex, shiny, opaque, and large (3 to5 mm at 48 h), with a mucoid consistency and a broad andclear 3-hemolytic zone.

Numerical taxonomy analysis. Results of the numericaltaxonomy analysis of VAR strains are presented in Table 3.Of the 41 strains examined, 36 were identified as V. splen-didius I, one as V. splendidus II, one as V. tubiashii, one asVibrio fischeri, and two as V. anguillarrum. The two V.angiuillarum isolates were retested serologically and found tobelong to serovars 02a and 08.

Serogrouping of V. anguillarum. The results of the sero-typing are shown in Tables 1 and 2. Of the 225 V. angiil-

TABLE 4. Distribution of V. anguillaritin serovars 01 and 02 inAtlantic salmon with vibriosis in different years

Distribution of serovari-s per yearSerovar

1975 1982 1987 1988

01 6 14 26 4902 1 23 2 55

laruin strains, 221 (98%) were serogrouped as serovar 01 or02. 132 strains being serovar 01 and 89 strains being serovar02. Two isolates were grouped as serovar 04. A single strainwas identified as V. an1giuill(arumitl serovar 08 through numer-ical taxonomy and serological retesting. One isolate was notgroupable. By using absorbed polyclonal 02 serosubtypeantisera, two initially crossagglutinating isolates were iden-tified as serovars 02a and 02b, respectively. The remainingserovar 02 isolates without crossagglutinating propertieswere not subtyped. All 89 V. aCngiuillariniii serovar 02 iso-lates, as grouped by the slide agglutination assay. showedidentical positive results on retesting in an enzyme-linkedimmunosorbent assay with a V. anIgiuillairiium-specific mono-clonal antibody.

Host specificity and pathogenicity among V. anguillarumserovars. As seen from Tables 1 and 2, 176 (98.3%) of 179 V.aingiuillaruim isolates from diseased Atlantic salmon belongedto serovar 01 or 02. Two isolates determined to be patho-genic were grouped as serovar 04, and a single strain wasclassified as serovar 08. The distribution of serovars 01 and02 among V. angiuillarulzm isolated from diseased Atlanticsalmon in different years is presented in Table 4. Six V.angiuillaruim isolated from Atlantic salmon without a vibrio-sis diagnosis were grouped as serovar 01 (n = 4) and 02a(n = 1), while one casual isolate was not groupable (Table 2).

All V. angiuillaruem isolates, clinical or casual, from othersalmonid fish (rainbow trout, brown trout, and arctic char)were serogrouped as 01 (Table 2). V. an,£giillariin isolatesoriginating from diseased marine fish were all serotyped asserovar 02 or subserovar 02b with the following fish speciesdistribution: wild saithe (n = 4), farmed coastal cod (n = 1),and farmed turbot (n = 1). The serovar 02b isolate wasrecorded among diseased wild saithe. A single V. anigiiil-laruitm serovar 01 isolate was identified as the causativeagent of vibriosis in a catadromous fish species, eel.

Host specificity and pathogenicity among VAR strains. Asshown in Table 2, clinical disease caused by VAR strainscould not be demonstrated among salmonid fish. Such bac-teria were, according to the clinical information obtained,collected as isolates from skin mucus, gills, and intestinalcontents of healthy fish or as casual isolates in low numbersfrom mixed cultures from the kidneys of salmonid individu-als suffering another main bacterial or viral infectious dis-ease, such as cold water vibriosis, vibriosis, furunculosis,yersiniosis, infectious salmon anemia (62), or infectiouspancreas necrosis. VAR strains were isolated also fromcachectic individuals and from dead and autolytic individu-als.On the other hand, possible clinical importance of such

strains among nonsalmonid marine fish species was indi-cated in 9 of 10 recorded individual cases. Six V. splendidusstrains, biovar I (n = 5) and biovar II (n = 1), were isolatedfrom turbot at two different seawater hatcheries duringdisease outbreaks, adults and fry being affected. Three V.splendidiis I isolates with potential clinical implications wereobtained from sea bass at one hatchery. One clinical isolate

Vot- 57. 1991

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

2754 MYHR ET AL.

TABLE 5. Plasmid profiles of 36 V. splendidus biovar I strains

No. of Plasmid size (MDa)strains

3 1351 63 37 253 552 55 37 251 55 372 371 251 3.41 6.0 1.7 1.421a

Total 3 1 6 6 4

a No plasmids detected.

originated from sole, and a single casual V. splendidus Iisolate was obtained from turbot.

Plasmid profiles. All 10 V. anguillarum serovar 01 strainshad one plasmid of 48 MDa. The nine 02 strains did notcarry any plasmids. The serovar 02a isolate had two plas-mids (140 and 37 MDa). No plasmids were found in theserovar 08 isolate. Of the 36 strains of V. splendidus biovarI, 15 strains contained plasmids, as shown in Table 5. Onlytwo of the plasmid-containing V. splendidus strains appearedto be of clinical importance, both isolated from sea bass atthe same hatchery. Thirteen plasmid-containing strains werecasual isolates from Atlantic salmon isolated at differentcoastal sites. Isolates with plasmids of 37 and 55 MDa wereisolated from Atlantic salmon and sea bass individuals inseawater farming facilities located in the same fjord system.The remaining 21 V. splendidus strains, including six isolatesfrom marine nonsalmonid fish species with assumed clinicalimportance, did not harbor any plasmids.

Antimicrobial susceptibility. The V. anguillarum serovar01 and 02 strains were highly susceptible to the antibioticstested in vitro (Table 6). Nitrofurazolidone was slightly lesseffective than the other antibacterial compounds tested.VAR strains as a group showed particularly high resistanceto nitrofurazolidine and, to a lesser degree, also to oxytet-racycline, flumequine, and trimethoprim-sulfadiazine.

DISCUSSION

The results of the present study show that serovars 01 and02 of V. anguillarum are of paramount importance ascausative agents of vibriosis among cultured salmonid fish,particularly Atlantic salmon, in Norway. This is in accor-

TABLE 6. MICs of different antibiotics for 90% ofV. anguillarum and VAR strains tested

MICg)" (,Lg/ml) for:Antibiotic V. anguillarum 01 V. anguillarum 02 VAR

(n = 10) (n = 9) (n = 39)

Trimethoprim- 0.5 0.5 2.0sulfadiazine

Nitrofurazolidone 1.0 4.0 32.0Oxytetracycline <0.5 <0.5 4.0Oxolinic acid <0.5 <0.5 <0.5Flumequine <0.5 <0.5 2.0Enrofloxacin <0.5 <0.5 <0.5

a MIC90, MIC for 90% of the isolates tested.

dance with previous reports (57, 58, 61, 68). However, V.anguillarum serovar 02 could not be demonstrated to be acausative agent of vibriosis among salmonid fish other thanAtlantic salmon, whereas serovar 02 was the only serovarisolated from marine fishes. Our findings diverge from theresults published by Toranzo et al. (64), who found that mostof the V. anguillarum strains isolated from turbot farmed onthe Atlantic coast of Spain belonged to serotype 01, only afew isolates being serotyped as 02.

Regarding outbreaks of vibriosis among Atlantic salmon,seasonal distribution patterns of serovar 01 versus 02 of V.anguillarum were apparent. Serovar 02 was predominant in1982 and 1988, and serovar 01 was predominant in 1975 and1987 (Table 4). Epizootiological data also reveal certainserovar geographical distribution patterns, V. anguillarumserovar 02 being frequently isolated along the southerncoastline of Norway and serovar 01 being predominant innorthern Norway.

Plasmid profiles for V. anguillarum serovars 01 and 02strains correspond to the results of Wiik et al. (68), i.e., a48-MDa pJM1 virulence plasmid analog was demonstrated inall V. anguillarum serovar 01 isolates, whereas V. anguil-larlum serovar 02 strains did not harbor any plasmids.However, the V. anguillarum serovar 02 isolates examinedin our study had all shown pathogenic properties, suggestingthe existence of alternate virulence mechanisms, such as thechromosomal DNA-encoded iron-sequestering system (36).In our study, the 10 plasmid-containing V. anguillarumserovar 01 strains were collected from several salmonid fishspecies from a wide geographic area, thus indicating anextensive distribution of this specific plasmid among Norwe-gian V. anguillarum serovar 01 strains. According to thepresent results, plasmid profiles of V. anguillarum serovar02 isolates are of no particular significance in epizootiolog-ical studies.Though VAR strains seem to be of no clinical importance

among farmed salmonid fish, they might be so among farmedmarine fish species. According to the present investigation,V. splendidus biotype I in particular seems to be associatedwith disease in turbot. Disease outbreaks of mixed viral andbacteriological etiology have recently been described forfarmed turbot (10, 37). Lupiani et al. (37) reported a newsyndrome of viral and bacterial etiology in turbot, in whichan isolated reovirus was considered to be of pathologicalsignificance in conjunction with V. splendidus. Bloch et al.(10) described a case of encephalomyelitis associated with apicornaviruslike agent among reared turbot. Bacteriologicalexamination of the diseased fish revealed sparse growth ofmixed cultures of various Vibrio spp: V. splendidus biovarsI and II, Vibrio pelagius, Vibrio alginolyticus, V. fischeri,and bacteria closely related to V. anguillarum. The authorsconcluded, in light of the high diversity of the presentbacteria, that a bacterial infection could not be regarded theprimary cause of mortality in the fish. A key question iswhether V. splendidus or other VAR bacteria should beconsidered primary pathogens or secondary or opportunisticbacteria acting concomitantly with other infective agents ofviral, bacterial, or parasitic origin under unfavorable breed-ing or environmental conditions.

Regarding the taxonomical differentiation of V. anguil-larum versus VAR strains investigated, striking differencescould be observed concerning cell and culture morphology,as well as cell motility and hemolysis patterns. However, nodifferences could be observed between the bacterial groupsor species on the basis of the eight key biochemical testsinitially chosen. Morphological and physiological features

APPL. ENVIRON. MICROBIOL.

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHARACTERIZATION OF V. ANGUILLARUM AND VAR STRAINS 2755

are thus of primary importance in distinguishing V. anguil-larum from other closely related species.Most V. splendidus strains showed reduced 0/129 sensi-

tivity when tested on blood agar plates or Mueller-Hintonagar plates. This can be explained by spontaneous in vitrodevelopment of thymine- and thymidine-dependent 0/129-resistant clones of V. splendidus. In vivo and in vitrodevelopment and selection of trimethoprim-resistant bacte-ria unable to synthesize thymidine are well known (2, 29,42). A correlation of resistance between the chemicallyrelated drug trimethoprim and the vibriostatic compound0/129 has been reported by several authors (3, 18, 26, 59).Such trimethoprim- and 0/129-cross-resistant mutants aredependent on the exogenous supply of thymine and thymi-dine for DNA synthesis and growth. The in vitro resistanceof V. splendidus to the vibriostatic compound 0/129 was notexpressed on lysed horse blood agar. This might be ex-plained by the relatively high content of the thymidaseenzyme in this particular medium and the inability of thymi-dine-dependent bacteria to multiply under these conditions(5, 42). In order to confirm the findings of the present studythat most V. splendidus strains seem actually to be thymi-dine-dependent 0/129-resistant clones, their growth on de-fined glucose-supplemented minimal salts agar, with andwithout thymine, would have to be tested. The apparent invitro 0/129 resistance among V. splendidus strains as re-corded on blood agar, creating substantial initial identifica-tion problems, might be regarded as a bacterial speciesmarker and thus a characteristic which could be exploiteddiagnostically.Other prominent V. splendidus biovar I phenotypical

traits, as revealed by the numerical taxonomical part of thestudy, are the ability to degrade alginate and the negativereactions regarding luminescence and L-arabinose utiliza-tion. The alginate-degrading capacity seems to be an excel-lent trait to distinguish V. splendidus biovar I from otherarginine decarboxylase-positive Vibrio species. None of theV. splendidus biovar I isolates tested showed luminescence.This finding is in contrast to previous reports emphasizingthis trait as being bacterial biovar specific (6, 7, 9). Similarphenomena of luminescent and nonluminescent bacterialclusters or subgroups of Vibrio cholerae, Vibrio harveij, andVibrio vulnificus have been described previously (44, 51, 65).The taxonomic conclusion is that luminescence is a propertyof different aquatic vibrios which may or may not beexpressed. A parallel conclusion seems appropriate for theV. splendidus group. The inability to produce acid fromL-arabinose might serve as a suitable characteristic to differ-entiate V. splendidus from specific serotypes of V. anguil-larum, especially serovar 01.

In our study, no serological cross-reactions were observedbetween 0-antigen preparations of V. splendidus strainsinvestigated and V. anguillarum 0 antisera 01, 02, and 03.However, in preliminary agglutination assays with V. splen-didus 0-antigen preparations and V. anguillarum antiseraother than 01, 02, and 03, cross-reactions could be dem-onstrated.

In conclusion, V. anguillarum 01 and 02 are by far themost common serovars associated with vibriosis in Norwe-gian farmed fish. Whereas serovars 01 and 02 both causedvibriosis in Atlantic salmon, serovar 01 was the only serovarcausing disease among other salmonid fish. Other serovarsof V. anguillarum as well as VAR species were also isolatedfrom Atlantic salmon, but only rarely in association withdisease. In marine fish species, vibriosis was caused only byV. anguillarum serovar 02. V. splendidus was associated

with disease in these fish species. The antigenic compositionof current vibriosis vaccines for salmonid fish in Norway isconsequently in accordance with the V. anguillarum sero-vars causing disease. The data from the present study alsogive a basis for the selection of bacterial strains for vaccinesagainst vibriosis in marine species.Concerning the taxonomic relations between V. anguil-

larum and its related species, V. splendidus biovar I was themost frequent species isolated from clinical specimens in thisstudy. In addition to cell and culture characteristics, thealginate-degrading ability was found to be a good parameterfor differentiation of V. splendidus biovar I and V. anguil-larum.

ACKNOWLEDGMENTS

We thank all who provided strains for this study, particularlyAnne Berit Olsen, Christian W. R. Koren, Aud Skrudland, IngerMette Hogstad, Brit T0rud, and Trude Bakke J0sund. We aregrateful to Maj-Britt H0jgard, The Royal Veterinary and Agricul-tural University, Copenhagen, Denmark, and Vera Sundvold, Na-tional Veterinary Institute, for their important and skillful technicalassistance.

This research program has been supported by the NorwegianVESO project 1020 and The Danish Agricultural and VeterinaryResearch Council grant 13-4508-INMP.

REFERENCES1. Anderson, J. W., and D. A. Conroy. 1970. Vibrio diseases in

fishes, p. 226-272. In S. F. Snieszko (ed.), A symposium ondiseases of fishes and shellfishes, special publication no. 5.American Fisheries Society, Washington, D.C.

2. Andrew, M. H. E. 1973. Use of trimethoprim to obtain thymine-requiring mutants of Streptococcus faecalis. J. Gen. Microbiol.74:195.

3. Aoki, T., T. Kitao, and K. Kawano. 1981. Changes in drugresistance of Vibrio anguillarum in cultured ayu, Plecoglossusaltivelis, Temminck and Schegel, in Japan. J. Fish Dis. 4:223-230.

4. Austin, B., and D. A. Austin. 1987. Vibrios, p. 263-296. InBacterial fish pathogens. Disease in farmed and wild fish. EllisHorwood Publishers, Chichester, United Kingdom.

5. Barker, J., D. Healing, and J. G. P. Hutchinson. 1972. Charac-teristics of some cotrimoxazole-resistant Enterobacteriaceaefrom infected patients. J. Clin. Pathol. 25:1086.

6. Baumann, P., S. S. Bang, and L. Baumann. 1978. Phenotypiccharacterization of Beneckea anguillarum biotypes I and II.Curr. Microbiol. 1:85-88.

7. Baumann, P., and L. Baumann. 1981. The marine Gram-nega-tive eubacteria: genera Photobacterium, Beneckea, Alteromo-nas, Pseudomonas and Alcaligenes, p. 1302-1331. In M. P.Starr, H. Stolp, H. G. Truper, A. Balows, and H. G. Schlegel(ed.), The procaryotes: a handbook of habitats, isolation andidentification of bacteria. Springer-Verlag, Berlin.

8. Baumann, P., L. Baumann, S. S. Bang, and M. J. Woolkalis.1980. Reevaluation of the taxonomy of Vibrio, Beneckea, andPhotobacterium. Abolition of the genus Beneckea. Curr. Bac-teriol. 4:127-132.

9. Baumann, P., A. L. Furniss, and J. V. Lee. 1984. Genus VibrioPacini 1854, 411AL, p. 518-538. In N. R. Krieg and J. G. Holt(ed.), Bergey's manual of systematic bacteriology, vol. 1. TheWilliams & Wilkins Co., Baltimore.

10. Bloch, B., K. Gravningen, and J. L. Larsen. 1991. Encephalo-myelitis among turbot associated with a picornavirus-like agent.Dis. Aquat. Org. 10:65-70.

11. Bolinches, J., A. E. Toranzo, A. Silva, and J. L. Barja. 1986.Vibriosis as the main causative factor of heavy mortalities in theoyster culture industry in northwestern Spain. Bull. Eur. Assoc.Fish Pathol. 6:1-4.

12. Bowser, P. R., R. Rosemark, and C. R. Reiner. 1981. Apreliminary report of vibriosis in cultured American lobsters,Homarus americanus. J. Invertebr. Pathol. 37:80-85.

VOL. 57, 1991

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

APPL. ENVIRON. MICROBIOL.

13. Bryant, T. N., J. V. Lee, P. A. West, and R. R. Colwell. 1986.Numerical classification of species of Vibrio and related genera.J. Appl. Bacteriol. 61:437-496.

14. Bryant, T. N., J. V. Lee, P. A. West, and R. R. Colwell. 1986. Aprobability matrix for the identification of species of Vibrio andrelated genera. J. Appl. Bacteriol. 61:469-480.

15. Colwell, R. R., and D. J. Grimes. 1984. Vibrio diseases of marinefish populations. Helgol. Meeresunters. 37:265-287.

16. Egidius, E., and K. Andersen. 1978. Host-specific pathogenicityof strains of Vibrio anguillarum isolated from rainbow troutSalmo gairdneri Richardson and saithe, Pollachius virens (L.).J. Fish Dis. 1:45-50.

17. Egidius, E., and K. Andersen. 1984. Disease problems in codrearing. In E. Dahl, D. S. Danielsen, E. Moksnes, and P.Solemdal (ed.), The propagation of cod Gadus morhuca L.Fl0devigen Rapporter 1:761-769.

18. Escande, F., G. Gerbaud, J. L. Martel, and P. Courvalin. 1991.Resistance to trimethoprim and 2,4-diamino-6,7-diisopropyl-pteridine (0/129) in Pasteurella haemolvtica. Vet. Microbiol.26:107-114.

19. Evelyn, T. P. T. 1971. First records of vibriosis in Pacific salmoncultured in Canada, and taxonomic status of the responsiblebacterium, Vibrio anguillarum. J. Fish. Res. Board Can. 28:517-525.

20. Evelyn, T. P. T. 1984. Immunization against pathogenic vibrios,p. 121-150. In P. de Kinkelin and C. Michel (ed.), Symposiumon fish vaccination. Office International des Epizooties, Paris.

21. Fouz, B., R. F. Conchas, J. Bolinches, J. L. Romalde, J. L.Barja, and A. E. Toranzo. 1990. Relationship among pathogenicVibrio anguillarum and Vibrio tubiashii with environmentalvibrios, p. 77-90. In T. C. Cheng and F. 0. Perkins (ed.),Pathology in marine science. Academic Press, Inc., New York.

22. Hada, H. S., P. A. West, J. V. Lee, J. Stemmler, and R. R.Colwell. 1984. Vibrio tubiashii sp. nov., a pathogen of bivalvemollusks. Int. J. Syst. Bacteriol. 34:1-A.

23. Hastein, T., F. Hallingstad, T. Refstie, and S. 0. Roald. 1980.Recent experiences of field vaccination trials against vibriosis inrainbow trout (Salmo gairdneri), p. 53-59. In W. Ahne (ed.),Fish diseases, third COPRAQ session. Springer-Verlag, Berlin.

24. Hastein, T., and G. Holt. 1972. The occurrence of Vibrio diseasein wild Norwegian fish. J. Fish Biol. 4:33-37.

25. Hastein, T., and J. E. Smith. 1977. A study of Vibrio anguil-larum from farmed and wild fish using principal component

analyses. J. Fish Biol. 11:69-75.26. Hjeltnes, B., K. Andersen, and E. Egidius. 1987. Multiple

resistance in Vibrio salmonicida. Bull. Eur. Assoc. Fish Pathol.7:85.

27. Holt, G. 1970. Vibriosis (Vibrio anguillarum) as an epizooticdisease in rainbow trout (Salmo gairdneri). Acta Vet. Scand.11:600-603.

28. Jeifries, V. E. 1982. Three Vibrio strains pathogenic to larvae ofCrassostrea gigas and Ostrea edulis. Aquaculture 29:201-226.

29. Jobanputra, R. S., and N. Datta. 1974. Trimethoprim R factorsin Enterobacteria from clinical specimens. J. Med. Microbiol.7:169-177.

30. Langdon, J. S. 1988. Diseases of introduced Australian fish, p.

225-259. In Fish diseases, proceedings, refresher course forveterinarians. University of Sydney, Sydney, Australia.

31. Larsen, J. L. 1982. Vibrio anguillarum: prevalence in threecarbohydrate-loaded marine recipients and a control. Zentralbl.Bakteriol. Parasitenkd. Infektionskr. Hyg. Abt. 1 Orig. Reihe C3:519-530.

32. Larsen, J. L. 1983. Vibrio anguillarum: a comparative study offish pathogenic, environmental, and reference strains. Acta Vet.Scand. 24:456-476.

33. Larsen, J. L. 1985. Vibrio anguillarum: prevalence of typicaland atypical strains in marine recipients with special referenceto carbohydrate pollution. Acta Vet. Scand. 26:449-460.

34. Larsen, J. L., H. B. Rasmussen, and I. Dalsgaard. 1988. Study ofVibrio anguillarum strains of different sources with emphasis on

ecological and pathobiological properties. Appl. Environ. Mi-crobiol. 54:2264-2267.

35. Lee, J. V., and T. J. Donovan. 1985. Vibrio, Aeromonas and

Pleisiomonas, p. 13-33. In D. H. Collins and J. M. Grange (ed.),Isolation and identification of microorganisms of medical andveterinary importance. Society for Applied Bacteriology tech-nical ser., vol. 21. Academic Press Ltd., London.

36. Lemos, M. L., P. Salinas, A. E. Toranzo, J. L. Barja, and J. H.Crosa. 1988. Chromosome-mediated iron uptake system inpathogenic strains of Vibrio anguillarum. J. Bacteriol. 170:1920-1925.

37. Lupiani, B., C. P. Dopazo, A. Ledo, B. Fouz, and J. L. Barja.1989. New syndrome of mixed bacterial and viral etiology incultured turbot Scophtalmus maximus. J. Aquat. Anim. Health1:197-204.

37a.Lupiani, B., A. Ledo, C. P. Dopazo, A. Baja, A. E. Toranzo, andJ. L. Barja. 1989. Virulence of Vibrio splendidus and turbotreovirus (TRV) for three different fish species, abstr. EAEP IV,p. 103. Abstr. Int. Conf. Dis. Fish Shellfish, Santiago deCompostela, Spain.

38. Mackie, C., and T. H. Birkbeck. 1987. Antigenic components inthe cell envelope in Vibrio anguillarum. Aquaculture 67:249-251.

39. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual, p. 368-369. Cold Spring HarborLaboratory, Cold Spring Harbor, N.Y.

40. Martinsen, B., E. Myhr, E. Reed, and T. Hastein. In vitroantimicrobial activity of sarafloxacin against clinical isolates offish pathogenic bacteria. J. Aquat. Anim. Health, in press.

41. Mattheis, T. 1964. Das Vorkommen von Vibrio anguillarum inOstseefischen. Z. Fisch. N. F. 12:259-263.

42. Okubadejo, 0. A., and R. M. Maskell. 1973. Thymine-requiringmutants of Proteus mirabilis selected by co-trimoxazole in vivo.J. Gen. Microbiol. 77:533.

43. Olafsen, J. A., M. Christie, and J. Raa. 1981. Biochemicalecology of psychrotrophic strains of Vibrio anguillarum isolatedfrom outbreaks of vibriosis at low temperature. Zentralbl.Bakteriol. Mikrobiol. Hyg. Abt. 1 Orig. C 2:339-348.

44. Oliver, J. D., D. M. Roberts, V. K. White, M. A. Dry, and L. M.Simpson. 1986. Bioluminescence in a strain of the humanpathogenic bacterium Vibrio vulnificus. Appl. Environ. Micro-biol. 52:1209-1211.

45. Oppenheimer, C. H. 1962. Marine fish diseases, p. 541-572. InG. Borgstrom (ed.), Fish as food, vol. 2. Academic Press, Inc.,New York.

46. Rasmussen, H. B. 1987. Evidence for the presence of a Kantigen in strains of Vibrio anguillarum. Curr. Microbiol. 14:275-278.

47. Rasmussen, H. B. 1987. Immunoelectrophoretic study of lipo-polysaccharide antigens from Vibrio anguillarum strains, sero-group 02. Curr. Microbiol. 15:219-222.

48. Rasmussen, H. B. 1987. Subgrouping of lipopolysaccharide 0antigens from Vibrio anguillarum serogroup 02 by immunoelec-trophoretic analyses. Curr. Microbiol. 16:39-42.

49. Rasmussen, H. B. 1987. Evidence for two new Vibrio anguil-larum K antigens. Curr. Microbiol. 16:105-107.

50. Rasmussen, H. B., and J. L. Larsen. 1987. Further antigenicanalyses of the fish pathogenic bacterium Vibrio anguillarum.Curr. Microbiol. 16:145-148.

51. Reichelt, J. L., P. Baumann, and L. Baumann. 1976. Study ofgenetic relationships among marine species of the genera Be-necka and Photobacterium by means of in vitro DNA/DNAhybridization. Arch. Microbiol. 110:101-120.

52. Roberts, R. J. 1989. Fish pathology, p. 302-312. Baillere Tin-dall, London.

53. Rochelle, P. A., J. C. Fry, M. J. Day, and J. Bale. 1985. Anaccurate method for estimating sizes of small and large plasmidsand DNA fragments by gel electrophoresis. J. Gen. Microbiol.132:53-59.

54. Schiewe, M. H., T. J. Trust, and J. H. Crosa. 1981. Vibrio ordaliisp. nov.: a causative agent of vibriosis in fish. Curr. Microbiol.6:343-348.

55. Smith, P. D. 1988. Vaccination against vibriosis, p. 67-84. InA. E. Ellis (ed.), Fish vaccination. Academic Press Ltd.,London.

56. Sneath, P. H. A. 1979. BASIC program for identification of an

2756 MYHR ET AL.

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

CHARACTERIZATION OF V. ANGUILLARUM AND VAR STRAINS 2757

unknown with presence-absence data against an identificationmatrix of percent positive characters. Comput. Geosci. 5:195-213.

57. Song, Y. L., S. H. Chen, and G. H. Kou. 1988. Serotyping ofVibrio anguillarum strains isolated from fish in Taiwan. FishPathol. 23:185-189.

58. S0rensen, U. B. S., and J. L. Larsen. 1986. Serotyping of Vibrioanguillarum. Appl. Environ. Microbiol. 51:593-597.

59. S0rum, H., T. Poppe, and 0. Olsvik. 1988. Plasmids in Vibriosalmonicida isolated from salmonids with hemorrhagic syn-

drome (Hitra disease). J. Clin. Microbiol. 26:1697-1683.60. Svendsen, I., and J. L. Larsen. 1988. Monoclonal antibodies

against surface antigens of Vibrio anguillarum serogroup 02.Acta Vet. Scand. 29:363-368.

61. Tajima, K., Y. Ezura, and T. Kimura. 1985. Studies on thetaxonomy and serology of causative organisms of fish vibriosis.Fish Pathol. 20:131-142.

62. Thorud, K., and H. 0. Djupvik. 1988. Infectious anaemia inAtlantic salmon (Salmo salar L.). Bull. Eur. Assoc. Fish Pathol.8:109-111.

63. Toranzo, A. E., and J. L. Barja. 1990. A review of the taxonomyand seroepidemiology of Vibrio anguillarum with special refer-

ence to the aquaculture in northwest of Spain. Dis. Aquat. Org.9:73-82.

64. Toranzo, A. E., Y. Santos, M. L. Lemos, A. Ledo, and J.Bolinches. 1987. Homology of Vibrio anguillarum strains caus-ing epizootics in turbot, salmon, and trout reared on the Atlanticcoast of Spain. Aquaculture 67:41-52.

65. West, P. A., P. R. Brayton, T. N. Bryant, and R. R. Colwell.1986. Numerical taxonomy of vibrios isolated from aquaticenvironments. Int. J. Syst. Bacteriol. 36:531-543.

66. West, P. A., and R. R. Colwell. 1984. Identification and classi-fication of Vibrionaceae-an overview, p. 285-363. In R. R.Colwell (ed.), Vibrios in the environment. John Wiley & Sons,New York.

67. West, P. A., J. V. Lee, and T. N. Bryant. 1983. A numericaltaxonomic study of species of Vibrio isolated from the aquaticenvironment and birds in Kent, England. J. Appl. Bacteriol.55:263-282.

68. Wiik, R., K. A. Hoff, K. Andersen, and F. L. Daae. 1989.Relationships between plasmids and phenotypes of presumptivestrains of Vibrio anguillarum isolated from different fish species.Appl. Environ. Microbiol. 55:826-831.

VOL. 57, 1991

on June 22, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from