Protocol for Specific Isolation of Virulent Strains of Vibrio vulnificus Serovar E (Biotype 2) from Environmental Samples

Abstract

The eel pathogen Vibrio vulnificus biotype 2 comprises at least three serovars, with serovar E being the only one involved in both epizootics of eel vibriosis and sporadic cases of human infections. The virulent strains of this serovar (VSE) have only been recovered from clinical (mainly eel tissue) sources. The main objective of this work was to design and validate a new protocol for VSE-specific isolation from environmental samples. The key element of the new protocol is the broth used for the first step (saline eel serum broth [SEB]), which contains eel serum as a nutritive and selective component. This approach takes advantage of the ability of VSE cells to grow in eel serum and thus to separate themselves from the pool of competitors. The growth yield in SEB after 8 h of incubation was 1,000 times higher for VSE strains than for their putative competitors (including biotype 1 strains of the species). The selective and differential agar Vibrio vulnificus medium (VVM) was selected from five selective media for the second step because it gave the highest plating efficiency not only for the VSE group but also for other V. vulnificus groups, including biotype 3. The entire protocol was validated by field studies, with alkaline peptone water plus VVM as a control. V. vulnificus was isolated by both protocols, but serovar E was only recovered by the new method described here. All selected serovar E isolates were identified as VSE since they were virulent for both eels and iron-overloaded mice and resisted the bactericidal action of eel and iron-overloaded human sera. In conclusion, this new protocol is a suitable method for the isolation of VSE strains from environmental samples and is recommended for epidemiological studies of the pathogenic serovar E.

Vibrio vulnificus is an aquatic bacterium from tropical and warm brackish waters with pathogenic potential for humans and aquatic animals (1, 8, 14, 17, 23, 38, 43, 45, 46, 47). The species is currently subdivided into three biotypes (14, 45), with biotype 2 being recognized worldwide as an eel pathogen. This biotype causes hemorrhagic septicemia (vibriosis) in cultured eels (8), and the first cases were recorded in Japan in the mid 1970s (38). From Japan, the disease arrived in Europe in the late 1980s and spread to eel farms in Mediterranean and Atlantic countries such as Spain, Sweden, The Netherlands, and Denmark (9, 10, 18, 23, 27, 28). Now the disease is the main cause of economic losses in brackish water eel culture in Europe. Biotype 2 strains are subdivided into at least three different O-antigenic serovars (11, 13, 23, 28). Serovar E corresponds to the original isolates that gave rise to the subdivision of the species into biotypes (45). This serovar is the only one that is related to sporadic human infections as well as epizootics or outbreaks with high mortality in eel farms (1, 9, 10, 11, 17, 47). Eels were the vehicle for transmission to humans in at least four clinical cases (zoonoses) (17, 47).

At present, the hazard that virulent serovar E (VSE) strains present to public health is underestimated. The main reason for this is that no VSE strain has been recovered from seafood, fish, or water during routine environmental sampling to monitor V. vulnificus (7, 9, 16, 30, 32). Current isolation procedures involve an enrichment step in alkaline peptone water (APW; 1% peptone, 1% NaCl [pH 8.6]), supplemented or not with antibiotics (19), followed by plating onto selective differential media such as thiosulfate-citrate-bile salts-sucrose (TCBS) agar, cellobiose-polymyxin-colistin (CPC) agar, and their derivatives, modified CPC (mCPC) agar, colistin-cellobiose (CC) agar, and V. vulnificus medium (VVM) (15, 29, 36, 40, 42, 44). According to several studies, the recovery of serovar E is difficult because these strains are present in very small numbers in the aquatic environment (7, 16, 30, 32). However, this contention is not supported by experimental data, as this serovar is not recovered even from eel farm water during epizootics or outbreaks, when high numbers of viable cells are released into the water (7, 16). Our hypothesis is that the isolation methodology fails because APW favors the growth of bacterial competitors (including biotype 1 strains) that can be deleterious for VSE cells and thus inhibit their isolation. This hypothesis is based on results obtained by Marco-Noales et al. (34) for survival experiments performed with water microcosms that were coinoculated with VSE strains and selected competitors. In those experiments, competition phenomena which were detrimental to the recovery of VSE strains from mixed populations were observed in both the presence and absence of nutrients (34). If our hypothesis is correct, then the relevance of this serovar from an epidemiological point of view is higher than previously thought.

The main objective of the present study was to develop a new two-step protocol for the specific isolation of VSE strains from asymptomatic eel carriers and water samples. The new protocol should favor the growth of VSE strains and inhibit that of putative competitors, especially those of biotype 1, which can be coisolated from aquatic and fish samples and which have been observed to induce VSE strains to enter into a viable but nonculturable state (34). Since the main recognized trait that distinguishes biotype 2 from biotype 1 strains is the ability of the former to grow in nonimmune eel serum (5), we designed a selective broth containing eel serum for the first isolation step. The efficacy of this new method was compared with that of the currently employed APW method, and the optimal incubation conditions were established and tested by use of a large collection of putative competitors belonging to V. vulnificus (biotypes 1 and 3) and other aquatic bacterial species. In parallel, several selective and differential media were compared for the ability to recover VSE strains, and the most efficient medium was selected. The efficacy of the entire protocol was tested by the use of natural environmental samples, including seawater, freshwater, eel tank water, and healthy eel tissues. The results obtained clearly indicate the suitability of the new protocol for successfully isolating VSE strains from natural mixed populations. Finally, the new serovar E isolates were characterized and compared with strains collected from clinical origins.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

A total of 37 V. vulnificus strains of different biotypes, sources, and origins, as well as 23 strains of other species, were used for this study (Tables 1 and 2). The strains were maintained both as lyophilized stocks at room temperature (25°C) and as frozen stocks at −80°C in marine broth (Difco) plus 20% (vol/vol) glycerol. Strains were routinely grown in Trypticase soy broth (TSB) or on Trypticase soy agar (TSA) (Difco) supplemented with 0.5% (wt/vol) NaCl (TSB-1 and TSA-1, respectively) at 28°C for 24 h.

TABLE 1.

Origins biotypes and serovars of the reference V. vulnificus strains used for this study

Straina,m	Origin	Source	Biotype-serovarb	Virulence in eels/virulence in micec	GRd in SEB
CECT 529T (ATCC 27562)*	Human blood	USA	BT1	−/+	0
CECT 5164 (374)e	Human blood	USA	BT1	−/+	0
CECT 5167 (L-180)e	Human blood	Japan	BT1	−/+	0
CECT 5168 (CDC7184)**	Human blood	USA	BT1	−/+	0
CECT 5169 (H3308)e	Human blood	USA	BT1	−/+	0
ATCC 33816	Human blood	USA	BT1	−/+	0
94385f	Leg wound	Spain	BT1	−/ND	0
V4g	Human blood	Australia	BT1	−/ND	0
CECT 5165 (UNCC 890)e	Sea water	USA	BT1	−/ND	0
CECT 4608*,**	Eel farm water	Spain	BT1	−/−	0
JEh	Oyster	USA	BT1	−/ND	0
VV425h	Oyster	USA	BT1	−/ND	0
CG100i	Oyster	Taiwan	BT1	−/+	0
CG106i	Oyster	Taiwan	BT1	−/ND	0
CG110i	Seawater	Taiwan	BT1	−/ND	0
A2	Eel farm water	Spain	BT1	−/−	0
CECT 4601	Diseased eel	Spain	BT2-VSE/E	+/+	4
CECT 4604*,**	Diseased eel	Spain	BT2-VSE/E	+/+	4
CECT 4605	Diseased eel	Spain	BT2-VSE/E	+/+	4
CECT 4917	Diseased eel	Spain	BT2-VSE/E	+/+	4
CECT 4864	Diseased eel	Spain	BT2-VSE/E	+/+	4
CECT 4865	Diseased shrimp	Taiwan	BT2-VSE/E	+/ND	4
Ö122j	Diseased eel	Sweden	BT2-VSE/E	+/ND	4
CECT 5198	Diseased eel	Spain	BT2-VNSE/A	+/ND	4
CECT 5343	Diseased eel	Spain	BT2-VNSE/A	+/ND	3
CECT 5768	Diseased eel	Spain	BT2-VNSE/A	+/ND	1
95-8-6k	Diseased eel	Denmark	BT2-VNSE/O3	+/ND	1
95-8-7k	Diseased eel	Denmark	BT2-VNSE/O3	+/ND	1
95-8-161k	Diseased eel	Denmark	BT2-VNSE/O3:O4	+/ND	2
95-8-162k	Diseased eel	Denmark	BT2-VNSE/O3:O4	+/ND	4
97l	Human	Israel	BT3	−/ND	1
12l	Human	Israel	BT3	−/ND	0
11028l	Human	Israel	BT3	−/ND	1
58l	Human	Israel	BT3	−/ND	ND
1033l	Human	Israel	BT3	−/ND	2
32l	Human	Israel	BT3	−/ND	1
162l	Human	Israel	BT3	−/ND	0

^aCECT, Spanish Type Culture Collection; ATCC, American Type Culture Collection; CDC, Centers for Disease Control and Prevention. T, type strain.

^bBT1, biotype 1; BT2-VSE, biotype 2, virulent serovar E; BT2-VNSE, biotype 2, virulent, non-serovar E; BT3, biotype 3. Biotypes were determined on the basis of eel virulence and cellobiose fermentation. Biotype 2 serovars: A and E are according to Biosca et al. (13) and Fouz and Amaro (23); serovars O3 and O3/O4 are according to Martin and Siebeling (35).

^cData on virulence are from Biosca (9), Fouz and Amaro (23), and Amaro et al. (unpublished results). ND, not done.

^dThe final GR was calculated for each strain as the log (final count/initial count) and was coded as 0 (GR ≤ 1), 1 (1 < GR ≤ 2), 2 (2 < GR ≤ 3), 3 (3 ≤ GR ≤ 4), or 4 (4 ≤ GR).

^eSupplied by J. D. Oliver, Department of Biology, University of North Carolina, Charlotte.

^fSupplied by L. Torres, Department of Microbiology, Miguel Servet University Hospital, Zaragoza, Spain.

^gSupplied by L. Gibson, Department of Cell and Molecular Biology, The University of Technology, Sydney, Australia.

^hSupplied by M. L. Tamplin, USDA-ARS-ERRC, Wyndmoor, Pa.

ⁱSupplied by L. Hor, Department of Microbiology and Immunology, National Cheng Kung University, Tainan, Taiwan, Republic of China.

^jSupplied by T. Hongslo, Department of Fish, National Veterinary Institute, Uppsala, Sweden.

^kSupplied by I. Dalsgaard and J. L. Larsen, Department of Veterinary Microbiology, Royal Veterinary and Agricultural University, Frederiksberg, Denmark.

^lBiotype 3 strains were provided by N. Bisharat from Nuffield Department of Clinical Laboratory Science, University of Oxford, United Kingdom, through J. D. Oliver (University of North Carolina), J. C. Piffaretti (Istituto Cantonale di Microbiologia, Bellinzona, Switzerland), or S. Miyoshi (Faculty of Pharmaceutical Science, Okoyama University, Okoyama, Japan)

^m*, strains selected for growth curves of single populations in selective enrichment broth; **, strains selected for growth curves of mixed populations in selective enrichment broth.

TABLE 2.

Origins and virulence of reference strains other than V. vulnificus used for this study

Species	Straina,g	Origin	Source	Virulence for eelsb	GR in SEBc	Growth in VVM
V. harveyi	CECT 604	Seawater	Spain	−	0	−
	CECT 605	Marine plankton	Spain	ND	0	−
	CECT 606	Marine plankton	Spain	−	0	−
	CECT 607	Seawater	Spain	ND	4	−
	CECT 608	Seawater	Spain	−	0	−
		Diseased horse	Japan	−
V. alginolyticus	CECT 521T	Mackerel			0	−
	CECT 600	Seawater	Spain	ND	0	−
	CECT 601	Seawater	Spain	ND	0	−
	CECT 603	Marine plankton	Spain	ND	0	−
	CECT 610	Marine plankton	Spain	ND	0	−
	RA3d	Diseased amberjack	Spain	ND	0	−
V. mediterranei	CECT 615	Marine plankton	Spain	−	0	+(G)f
	CECT 621T	Marine sediment	Spain	−	0	+(G)f
V. splendidus	CECT 528	Marine fish	Spain	−	0	−
	CECT 4204	Seawater	USA	−	0	−
V. parahaemolyticus	CECT 611	Seawater	Spain	−	0	−
	CECT 612	Seawater	Spain	ND	0	−
	CECT 613	Marine plankton	Spain	ND	0	−
Vibrio sp.	PD-4*	Eel tank water	Spain	ND	0	−
Edwardsiella tarda	CECT 886	Eel tank water	USA	+	0	−
Pseudomonas anguilliseptica	CECT 899T	Diseased eastern fish	Japan	+	0	−
Aeromonas hydrophila	E37e	Diseased eel	Spain	+	1
A. jandaei	S345e	Healthy eel	Spain	ND	4	−

^aCECT, Spanish Type Culture Collection. T, type strain.

^bData on virulence are from Biosca et al. (9) and Esteve et al. (20). ND, not done.

^cThe final GR was calculated for each strain as the log (final count/initial count) and was coded as 0 (GR ≤ 1), 1 (1 < GR ≤ 2), 2 (2 < GR ≤ 3), 3 (3 ≤ GR ≤ 4), or 4 (4 ≤ GR).

^dSupplied by E. Alcaide, Universidad de Valencia, Valencia, Spain.

^eSupplied by C. Esteve, Universidad de Valencia, Valencia, Spain.

^fG, green colonies.

^g*, strain selected for growth curves of single and mixed populations in selective enrichment broth.

Efficiency of the new selective enrichment broth. (i) SEB.

Pooled eel sera were obtained from eels as previously described (22) and then stored at −80°C. Phosphate-buffered saline (PBS; 0.02% KCl, 0.15% Na₂HPO₄, 0.20% KH₂PO₄) supplemented with NaCl to a final concentration of 1% (PBS-1), 2% (PBS-2), or 3% (PBS-3) was adjusted to the pH of eel blood (pH 8.3). Saline eel serum broth (SEB) was prepared by diluting eel serum 1:5 in PBS-1 (SEB-1), PBS-2 (SEB-2), or PBS-3 (SEB-3) immediately before use.

(ii) Bacteria.

Sterile microcosms of artificial seawater (50) were prepared in screw-cap 30-ml glass tubes as previously described (33) and were inoculated with bacterial suspensions from TSA-1 plates in artificial seawater to a final concentration of 10⁶ to 10⁷ CFU per ml. All tubes were incubated in the dark in a static state at room temperature for at least 1 month to simulate the starvation conditions typical of natural environments. Microcosms were sampled weekly for cultivable counts by the drop plate method (26) using TSA-1 plates (33).

(iii) Growth curves.

The effectiveness of SEB and APW in inhibiting the growth of competitors and enhancing the growth of VSE strains was tested with pure cultures (Tables 1 and 2). Three independent experiments were performed with each strain and combination of strains. For each experiment, the enrichment broth was inoculated with starved cells (10² to 10³ CFU per ml) and incubated with shaking at 28°C for 10 h. Plate counts for single cultures were done with TSA-1 plates by the drop plate method (26), and samples were taken after 0, 1, 2, 3, 4, 6, 8, and 10 h of incubation. The average and standard deviation of bacterial counts per incubation time were calculated, and the significance of the differences between data was determined by variance analysis performed with SPSS, release 11.0. The most effective combination of broth and time of incubation was tested with all of the strains listed in Tables 1 and 2 and with mixed cultures (VSE strains plus one competitor). For this purpose, microcosms of these bacteria were prepared as described above and were used to inoculate the enrichment broth. Plate counts for single cultures were done with TSA-1 plates by the drop plate method (26), and for mixed cultures, plate counts were done with the differential medium MSWYE-BTB agar plus 1% mannitol (34) (on this medium, VSE strains develop green colonies while colonies of the selected competitors are yellow). Experiments were performed in duplicate. The final growth rate (GR) was calculated for each strain as the log of the final count divided by the initial count and was coded as follows: 0, GR ≤ 1; 1, 1 < GR ≤ 2; 2, 2 < GR ≤ 3; 3, 3 ≤ GR ≤ 4; or 4, 4 ≤ GR.

Efficiency of selective media.

Five selective media (TCBS agar [Difco], CC agar, CPC agar, mCPC agar, and VVM) were tested for their efficiencies of recovery of V. vulnificus serovar E strains with respect to the general medium TSA-1. These selective media were prepared according to their original descriptions (15, 29, 36, 44). Selected V. vulnificus strains (Table 1) were grown overnight in MSWYE at 28°C, and bacterial counting was done with the selective and general media by the drop plate method (26). Plates were incubated for 24 to 48 h at the following recommended temperatures: TSA-1 and TCBS agar, 28°C; VVM, 37°C (15); and CC, mCPC, and CPC agar, 40°C (29, 36, 44). To test the influence of the incubation temperature on VSE recovery, we also incubated CC, mCPC, and CPC agar plates at 37°C. The efficiency of recovery was expressed as a recovery rate (RR), which was calculated as the percentage of CFU recovered on each of the selective media compared to the CFU obtained on the corresponding TSA-1 plate and was expressed as the mean value ± the standard deviation. A Mann-Whitney nonparametric analysis of variance was performed to evaluate the statistical significance of differences in the efficiency values (SPSS for Windows, release 11.0). The most effective selective medium was tested for its effectiveness in inhibiting the growth of putative competitors. For this purpose, all of the strains listed in Table 2 were grown in MSWYE at 28°C for 18 h, and a loopful of each was streaked onto the selected agar medium. The plates were incubated at the recommended temperatures for 24 to 48 h and then examined for growth.

Field sampling.

A total of eight samples of water from a Spanish coastal site (Delta del Ebro), a freshwater lake (Albufera Lake), and two eel farms, as well as nine samples of tissues taken from wild and cultured eels, were tested (Table 3). One of the farms had experienced recurrent outbreaks of vibriosis 2 years before (Table 3). In the case of water samples, 250-ml samples were collected in sterile flasks and filtered through sterilized 0.22-μm-pore-size membrane filters (Millipore). In the case of adult eels, samples of mucus, spleens, gills, and livers from four eels per sampling were taken aseptically, pooled, and homogenized in PBS-1. In the case of glass eels, the whole animals were homogenized in PBS-1. Filters and tissue homogenates were incubated with shaking in flasks containing 250 ml of APW or in tubes containing 25 ml of the selected SEB at 28°C for 18 and 8 h (see Results), respectively. Tenfold dilutions in PBS-1 of all enrichments were performed, and aliquots of 100 μl were streaked onto the selective medium selected in prior experiments. An average of 30 suspected colonies from each sampling were purified on TSA-1 for further studies.

TABLE 3.

Isolation of V. vulnificus serovar E from water and fish samples by the APW + VVM and SEB-1 + VVM procedures

Type of samplea (n)	Eel body wt (g)	Physicochemical parameters of waterb		No. of positive samples for V. vulnificus/no. of V. vulnificus serovar E-positive samples (%)
		T (°C)	Salinity (%)	APW+VVM	SEB+VVM
Water (8)
Ebro delta (1)		25	2.5	1/0	1/1 (50)
Albufera Lake (1)		25	0.5	0/0	0/0
Freshwater eel farm (3)		25-28	0.1	3/0	3/0
Brackish-water eel farm (3)		25-28	1.5-1.7	3/0	3/2 (10/50)
Tissues (9)
Wild glass eels (3)	0.1			0/0	0/0
Wild eels (2)	27.5			1/0	1/0
Cultured elvers (4)	10	25-28	0.1	2/0	2/1 (100)

^aAll samplings of water and eels from eel farms were performed in the absence of epizootics or outbreaks. The cultured elvers came from the freshwater eel farm, which had experienced epizootics of vibriosis 2 years before the sampling. Wild glass eels came from the Mediterrancan Sea and adult eels came from Albufera Lake and were bought alive in a public market in Valencia (Spain).

^bThe temperature and salinity of tank water from which cultured eels were captured are also indicated.

Identification and characterization of environmental isolates. (i) Identification of species.

For identification to the species level, colony hybridization with a V. vulnificus-specific alkaline phosphatase-labeled DNA probe directed against a cytolysin-hemolysin gene was used (51).

(ii) Serological identification and biochemical characterization.

Confirmed V. vulnificus isolates were serologically identified by slide agglutination (4) with previously obtained rabbit anti-serovar E, anti-serovar A, anti-serovar O3, and anti-serovar O3/O4 polyclonal antibodies (4, 23; also our unpublished results). A total of nine randomly selected serovar E strains were seeded in API 20E and API 20NE strips (Biomerieux, Madrid, Spain), and the API profiles were compared with the API database (Apilab Plus, version 3.3.3; Biomerieux). Additionally, the following biochemical tests were performed: Simmons citrate test, Thornley's arginine dihydrolase test, and decarboxylation of lysine and ornithine (ODC) in Mueller broth (decarboxylase medium base [Difco]) (12).

(iii) Serum resistance.

The sensitivities of the selected V. vulnificus serovar E strains to eel and human sera were evaluated with bacteria grown on TSA-1 in microtiter plates (6). In each well, 50 μl of eel serum or iron-overloaded human serum (supplemented with 10 μM desferoxamine mesylate [Desferal; Sigma]) (3) was mixed with 50 μl of a suspension of bacteria (10³ to 10⁴ CFU/ml) in PBS-1. The assays were performed in triplicate, and samples were taken after 0, 1, 2, 3, and 4 h of incubation at room temperature. Viable counts were determined by drop plating on TSA-1.

(iv)Virulence assays.

The potential virulence of the V. vulnificus serovar E isolates for humans and eels was tested by using BALB/c mice with an average weight of 20 g (5 to 6 weeks old) and juvenile European eels with an average weight of 10 g (elvers), respectively. Groups of six animals were used for each experiment, which lasted 1 week. Mice were pretreated for 2 h with Desferal (250 μg/g of body weight; iron-overloaded mice) because the virulence of V. vulnificus is markedly increased when the serum contains this iron chelator (a hydroxymate-type siderophore that promotes growth in serum) (3). Eels were maintained in aerated tanks containing 6 liters of saline water (1% NaCl) at 25°C (2). The V. vulnificus serovar E environmental isolates were grown on TSA-1 at 28°C for 24 h, and bacterial suspensions in PBS-1 were prepared. Approximately, 10⁴ CFU/fish and 10³ CFU/mouse were injected into elvers (0.1 ml per fish) and mice (0.2 ml per mouse). Appropriate controls for each experiment were also included (mice and elvers inoculated with PBS-1) (2, 3). Mice and elver mortalities were recorded daily and were only considered if bacteria were isolated in pure culture from internal organs.

RESULTS

Efficiency of selective enrichment broths.

Growth curves for starved cells of selected VSE, V. vulnificus biotype 1, and Vibrio sp. strains (Tables 1 and 2) in SEB are shown in Fig. 1. The Vibrio sp. strain was selected because it was coisolated with V. vulnificus biotype 1 on VVM from water of an eel tank in a routine control study. V. vulnificus biotype 1, VSE, and Vibrio sp. strains showed similar growth patterns in APW, without significant differences between the curves (Fig. 1). In contrast, the growth of the VSE strain in SEB-1 and SEB-2 was significantly higher than that of the nontarget bacteria throughout the incubation period (Fig. 1). Significant differences were also observed with SEB-3, but only in the fourth hour of incubation (Fig. 1). The generation times (g) in SEB-1 and SEB-2 of the nontarget bacteria were very similar (about 1 h in SEB-1 and 0.7 h in SEB-2) and were longer than those of the target bacteria (about 0.4 h in both media) (Fig. 1). In contrast, growth rates of both target and nontarget bacteria were very similar in SEB-3 (about 0.75 h) (Fig. 1). The largest differences in the final GRs between the target and nontarget bacteria were found in SEB-1 after 8 h of incubation (3 × 10⁷ to 2.5 × 10⁹ versus 1 × 10³ to 2 × 10⁴). Thus, SEB-1 was coinoculated with one VSE strain plus one competitor (Tables 1 and 2) in proportions of 1:1. After 8 h of incubation, only green colonies could be counted in the experiments, regardless of the strain that was coinoculated with the serovar E strain. Finally, SEB-1 was tested with biotype 2 strains of other serovars (Table 1), with biotypes 1 and 3 (Table 1), and with selected competitor species (Table 2). After 8 h of incubation, only four strains, two of biotype 2, one of Vibrio harveyi, and one of Aeromonas jandaei, achieved the same GR as the VSE isolates. Five biotype 2 strains and four biotype 3 isolates grew with GR values between 1 and 3, and the rest of the strains did not grow (Tables 1 and 2).

FIG. 1.

Growth curves of V. vulnificus CECT 4604 (•), CECT 529^T (○), and Vibrio sp. strain PD-4 (▾) in APW (A), SEB-1 (B), SEB-2 (C), and SEB-3 (D). Each point represents the average ± standard deviation of values from at least three different experiments. Significant statistical differences (α = 0.05) between CECT 4604 and CECT 529^T (#) and between CECT 4604 and Vibrio sp. strain PD-4 (*) are indicated.

Efficiency of selective media.

The strains of V. vulnificus were classified into four groups, as follows: BT1 (biotype 1), BT2-VSE (biotype 2, virulent serovar E), BT2-VNSE (biotype 2, virulent, non-serovar E), and BT3 (biotype 3) (Table 1). The recovery rates (RRs) obtained for each group and for each selective medium are shown in Fig. 2. Although high standard deviations within each group were observed, statistical analysis revealed significant differences between the groups and media. No significant differences in colony counts were registered regarding the incubation time (24 or 48 h) or, in the case of CPC, mCPC, and CC agars, the incubation temperature (40 or 37°C) (P ≤ 0.05). In all cases, the lowest RRs were seen for CPC agar (from 0.02 to 1.75%), and the highest were seen for VVM (from 50.42 to 71.23%) (Fig. 2). BT2-VSE and BT3 strains did not grow well on mCPC agar and CC agar, which does not contain polymyxin B (29), and BT3 strains did not grow on TCBS agar either. In addition, statistical differences between groups regarding the RRs from several media were detected. Thus, (i) the BT1 group showed significantly higher RRs on TCBS, CC, and mCPC agars than the BT3 group and on CC and mCPC agars than the BT2-VSE group; (ii) the BT2-VNSE group had significantly higher RRs on TCBS and CC agars than the BT2-VSE and BT3 groups; and (iii) the BT2-VSE group yielded significantly higher RRs on TCBS agar than the BT3 group. Since the highest RRs of VSE strains were achieved with VVM, this medium was selected for further experiments. On this medium, the three biotypes of V. vulnificus developed bright colonies of about 2 mm, with biotypes 1 and 2 being yellow with a yellow diffusion halo and biotype 3 being green without a halo. The effectiveness of VVM at inhibiting the growth of putative competitors was tested with the strains listed in Table 2. Only Vibrio mediterraneii strains were able to grow on VVM, giving green colonies with a similar morphology to those of biotype 3.

FIG. 2.

Recovery rates (%) of the different groups of V. vulnificus on the selective media tested. Vertical bars represent standard deviations.

Field sampling.

V. vulnificus was isolated from seven of eight water samples and from three of nine tissue samples by both isolation procedures (Table 3). Regarding water samples, V. vulnificus was isolated from seawater (2.5% salinity) and tank water from fresh- and brackish-water eel farms, but not from lake water (Table 3). In the case of eel tissue samples, V. vulnificus was recovered from both cultured and wild eels (Table 3). After enrichment in APW, >95% of the presumptive yellow colonies on VVM were identified as belonging to V. vulnificus by colony hybridization with the V. vulnificus-specific alkaline phosphatase-labeled DNA probe, but none of the isolates agglutinated with serovar E-specific antisera. After enrichment in SEB-1, 100% of the presumptive colonies were identified as V. vulnificus. Serovar E was recovered from three of seven V. vulnificus-positive (VV⁺) water samples (seawater and eel farm water) and from one of three VV⁺ tissue samples (cultured eels) (Table 3). The percentages of confirmed serovar E colonies ranged from 10 to 50% in the case of water samples to 100% in the case of eel samples (Table 3). About 20 randomly selected non-serovar E isolates from water and eel samples were tested with serovar A-, serovar O3-, and serovar O3/O4-specific antisera. Two isolates (10%) from tank water agglutinated with anti-serovar A sera, and the rest (50%) did not agglutinate with these sera or were autoagglutinating (40%).

Characterization of environmental VSE strains.

Nine randomly selected serovar E isolates were further characterized. The origins of these strains and the results of the API 20E and NE systems as well as some additional conventional taxonomic tests are summarized in Table 4. Eight of these strains gave the same API 20E profile as the control clinical VSE strain, while one differed only in citrate utilization. According to the API database, the major profile corresponded to V. vulnificus with a probability of 54.4%, while the minor profile had a probability of 10.4% (Table 4). Because of previously reported discrepancies between conventional and commercial assays (12), several tests were reexamined. More than 80% of the isolates were positive for citrate by the conventional test, and >60% were positive for ODC, a phenotypic trait that had a negative result in the API 20E system. Regarding the API 20NE system, seven isolates gave the same profile as the control strain, which, according to the API database, did not correspond to V. vulnificus (Table 4). The other two strains showed slight differences, and only one was identified as V. vulnificus with a probability of 75% (Table 4).

TABLE 4.

Biochemical characteristics of selected V. vulnificus serovar E strains isolated from water and asymptomatic carriers

Strain	Source	API 20E profile	API 20NE profile	Result of conventional test
				CIT	ADH	LDC	ODC
PD-8	Eel tank water	5006005b	5473745d	−	−	+	−
AnS1	Gills	5006005b	5472645e	+	−	+	+
C1	Liver	5006005b	5472745f	+	−	+	−
Riu-2	Seawater	5006005b	5472745f	+	−	+	−
PD-2-47	Tank water	5006005b	5472745f	+	−	+	+
PD-2-50	Tank water	5006005b	5472745f	+	−	+	+
PD-2-51	Tank water	5006005b	5472745f	+	−	+	+
PD-2-55	Tank water	5006005b	5472745f	+	−	+	−
PD-2-56	Tank water	5006005c	5472745f	+	−	+	+
CECT 4604a	Diseased eel	5006005c	5472745f	+	−	+	−

^aVSE control strain.

^bIdentification according to the API data base: V. vulnificus (54.4%) or Burkholderia cepacia (45.3%).

^cIdentification according to the API data base: B. cepacia (89.3%) or V. vulnificus (10.4%).

^dIdentification according to the API data base: Aeromonas hydrophila/caviae (94.5%).

^eIdentification according to the API data base: V. vulnificus (75%), V. cholerae (17.4%), or V. alginolyticus (4.6%)

^fIdentification according to the API data base: V. cholerae (45%) or A. hydrophila/caviae (41.7%).

All serovar E isolates could survive and grow in undiluted fresh eel and iron-overloaded human sera, giving bacterial yields similar to or even higher than those of the control strain (Fig. 3). In addition, all strains were virulent for eels and iron-overloaded mice, giving a mortality rate of >75% after the injection of 10⁴ CFU/fish or 10³ CFU/mouse. Eel and mouse mortality occurred before 48 h, and bacteria were recovered as pure cultures from internal organs. The infected eels showed redness on their bodies, particularly on the head and tail. Internally, the liver and kidneys appeared hemorrhagic, as did the muscle wall on occasion. No external pathological signs were observed for moribund mice except for occasional small ulcers on the tail.

FIG. 3.

Growth curves of environmental serovar E strains VSE PD-8 (•), AnS1 (○), and C1 (▾) in eel (A) and iron-overloaded human (B) sera with respect to the control strain CECT 4604 (▿). Each point represents the average of three different experiments.

DISCUSSION

For the present study, a two-step protocol for the isolation of VSE strains from asymptomatic carriers and environmental samples was developed and validated in the field. Selected target and nontarget bacteria grew equally well in the enrichment broth recommended by the U.S. Food and Drug Administration (FDA), which confirmed that APW is not adequate for the selection of VSE strains from a pool of V. vulnificus strains. Fresh eel serum diluted in alkaline PBS (SEB) was chosen because eel virulence in V. vulnificus is related to resistance to serum complement (5) as well as to the growth rate in eel serum (5, 21; also our unpublished results). Until now, animal sera had only been used after heat inactivation as a nutrient supplement for the growth of some fastidious pathogenic bacteria (31, 48). The present approach takes advantage of the ability of these bacteria to cause infection (grow in eel serum) as a means of selecting them from a pool of competitors. The results obtained with pure and mixed cultures confirmed that SEB acted as a culture broth for VSE strains, while it was bacteriostatic or bactericidal for competitors, including biotype 1 and 3 strains of this species. The optimal selective conditions were obtained by incubating bacteria in SEB-1 for 8 h at 28°C. Under these conditions, VSE cells could be selected, even from a pool of biotype 2 cells, on the basis of the higher growth rates of the VSE strains in eel serum.

With regard to the evaluation of the efficacies of the selective media, the results clearly demonstrated that CPC agar should not be used for VSE recovery (RR < 0.5%). The inability of VSE strains to grow in CPC agar had previously been reported by Macián et al. (32) and suggested by Høi et al. (29). Furthermore, this inability was generalized for all of the V. vulnificus groups, including the biotype 1 group (RR < 2%), which contrasts with data reported by other authors (15, 27). This discrepancy may be due to differences in the physiological state of the cells since they were previously grown in a low-nutrient medium (MSWYE) instead of a high-nutrient one (TSA and others) as usual. mCPC and CC agars, which contain less antibiotic, are recommended instead of CPC agar by many investigators, including the FDA, which recommends mCPC agar. Nevertheless, as shown in this study, these CPC derivatives are also inadequate for the recovery of VSE (RR below 1.5%), VNSE (RR below 6%), and biotype 3 (RR below 0.02%) strains, although the RR of biotype 1 strains increased significantly (around 20%). These results may explain by themselves why no biotype 2 and 3 strains have been isolated from the environment. The efficacies of recovery of all groups except one (biotype 3) increased significantly when TCBS agar was employed. This medium is widely used for the isolation of the genus Vibrio (7, 25, 37, 41, 49) and is the second medium recommended by the FDA for the isolation of V. vulnificus. However, its efficacy has been questioned by several authors (15, 29, 49), primarily due to the difficulty of reproducing results. In fact, we obtained RRs that were much higher than those reported by other laboratories (29). In addition, the RRs on TCBS agar of VSE strains were significantly lower than those of VNSE strains. This result may be related to the etiological characteristics of the diseases since VNSE strains colonize the intestines and mostly produce intestinal hemorrhages and feces with blood (23). Oddly enough, VSE strains were more similar to biotype 3 strains, and VNSE isolates were more similar to biotype 1 isolates, regardless of the strains' origins. This apparent subdivision of the eel pathogenic group supports the polyphyletic origin proposed by Gutacker et al. for the biotype 2 strains of the species (24). Fortunately, VVM significantly increased the efficacies of recovery of all groups, without differences between eel pathogenic subgroups. This medium contains electrolytes (MgCl₂ · 6 H₂O and KCl) that stimulate the growth of pathogenic vibrios (19). The binding of these cations to the bacterial membrane may alter its ionic potential and make the more sensitive strains (VSE and biotype 3) more resistant to polymyxins. VVM was tested with a wide collection of strains belonging to putative competitor species, and only V. mediterranei grew, developing green colonies. Nevertheless, since this species was inhibited by SEB-1, its presence in a sample should not cause problems.

In the second part of this study, the protocol was validated in the field, with APW plus VVM used as a control. Firstly, V. vulnificus was isolated by both protocols from water and healthy tissue samples. However, the percentage of positive isolations was considerably higher than that previously reported for other sites on the Mediterranean Sea with higher salinities (7, 32) and was similar to that found for other habitats with similar salinities (16, 29). Serovar E strains were only recovered after SEB-1 enrichment, which confirmed that the competition phenomena present in APW can be abolished by using this selective enrichment broth. In fact, 43 and 33.33% of the samples that were positive for V. vulnificus from water and tissues, respectively, were also positive for serovar E recovery, which constituted between 10 and 100% of the V. vulnificus colonies, depending on the sample. In addition, after SEB enrichment, serovar A strains, together with other nontypeable strains, were also isolated. Interestingly, serovar E was not detected in wild eels. Although more samplings are needed, this result supports the hypothesis that wild European eels are not the natural host for this serovar, which underscores the highly virulent potential of this pathogen for nonimmunized eels (5, 9). All of the selected environmental isolates were clearly identified as VSE strains since they were virulent for eels and iron-overloaded mice and resisted the bactericidal action of eel and iron-overloaded human sera. The biochemical profiles of the VSE isolates were quite homogeneous and similar to that of the clinical control VSE strain included in this study. In contrast to the report of O'Hara et al. (39), we found that all VSE strains were negative for the indole test in the API 20E system, which is in accordance with previous studies (4, 10, 11, 12, 13, 38, 45). Some discrepancies in the results of Simmons citrate and ODC tests which have already been described (12) were detected, as most of the environmental isolates were positive for both characteristics in conventional tests but not in commercial ones. In addition, none of the isolates was correctly identified by the API 20E and NE systems, which shows the limited value of these systems for the identification of V. vulnificus unless these profiles are included in the API database.

In conclusion, a new two-step protocol for VSE-specific isolation from environmental samples has been developed and validated in the present work. The key element of this protocol is the enrichment broth used for the first step, which yielded differences in bacterial counts between VSE strains and competitors that were large enough to allow the isolation of this serovar from natural mixed populations after only 8 h of incubation. For the second step, the selective and differential agar VVM was selected because it gave the highest plating efficiencies, not only for the VSE group, but also for the rest of the V. vulnificus groups, including biotype 3. The entire protocol was validated in the field since VSE strains were isolated for the first time from seawater and eel farms in the absence of epizootics or outbreaks. Finally, the overall results demonstrate that serovar E strains are present in the aquatic environment and that only with adequate procedures of isolation can their true epidemiological relevance be revealed.