Vibrio pathogenicity island and phage CTX genes in Vibrio alginolyticus isolated from different aquatic environments

Bacteria of the genus Vibrio are indigenous to the marine environment and temporarily in abundance in the ocean, being able to cause infections in humans and commercially important species of crustaceans, bivalves and fish. Severe outbreaks of diseases caused by Vibrio have large economic consequences for aquaculture and other economic sectors (Liu et al. 2004). Vibrio alginoliticus is a marine bacterium, and some strains can cause gastroenteritis in humans through the consumption of contaminated seafood. The transmission of V. alginolyticus strains from the aquatic environment, fishes and shellfish to humans has been well documented (Xie et al. 2005; Zhang & Austin 2005; Khouadja et al. 2012). This bacterium is frequently associated with mass mortality of Sparus aurata and Dicentrarchus labrax larvae and older fish in many Tunisian hatcheries installed along the Mediterranean seacoasts (Ben Kahla-Nakbi et al. 2009; Snoussi et al. 2009), and this pathogen was also reported in many other countries in Mediterranean coasts (Zanetti et al. 2000; Croci et al. 2001; Balcazar et al. 2009).

ctxAB, encoding for cholerae toxin (CT) and the V. cholerae pathogenicity island (VPI) is associated with the epidemic strains in V. cholerae (Waldor & Mekalanos 1996). The ctxAB genes are carried in the genome of a filamentous, single-stranded DNA phage-designated CTX, and their dissemination to non-pathogenic strains may, therefore, occur via phage-mediated horizontal gene transfer (Waldor & Mekalanos 1996). The VPI has recently been proposed to also be a filamentous phage VPIØ (Li et al. 2003). Unlike for the CTXØ, convincing data are lacking for the existence or the horizontal transfer of VPIØ. At present, none of the 29 genes in the VPI (other than tcpA) has been assigned, based on experimental data, any function in the proposed phage's life cycle or phage transduction (Faruque et al. 2003; Krebs & Taylor 2011). As it has been suggested that Vibrio strains may arise from toxigenic V. cholerae TCP and ToxR by infection with CTXØ and virulence genes reported in other Vibrio strains, a need was identified for an investigation into the dissemination of these genes among other Vibrio species.

In the present study, we investigated the presence of four Vibrio cholerae virulence genes (ctxA, VPI, Zot and ace) in 36 Vibrio alginolyticus isolates obtained from different seawater, sediments and varied aquatic organisms. In addition, we tested the virulence of 13 V. alginolyticus strains to see if there is a correlation between the presence of certain virulence genes and the pathogenicity.

The samples were taken from different sites along the Tunisian seacoasts (Table 1) and were transported on ice to the laboratory within 2 h of sampling and processed for bacterial identification.

Samples were analyzed according to a standard procedure (Ottaviani et al. 2003). Briefly, 25 g of blended sample was inoculated in 225 ml of alkaline peptone-water (Difco, Spain) with 3% NaCl and incubated at 30 °C for 18–24 h. Liquid samples were filtered (0.45 μm, Millipore, Sartorius Minisart CE 0297, Germany) and membranes were enriched on alkaline peptone-water (3% NaCl) and then incubated at 30 °C for 18–24 h. A 10 μl of the enrichment culture was streaked onto Thiosulphate-citrate-bile salt-sucrose-modified agar (TCBS agar) used for the selective isolation of Vibrio strains (Scharlau Microbiology, Spain). After 18–24 h of incubation at 30 °C, only the cultures giving pure yellow colonies (2–3 colonies from each sample) were randomly selected and then subcultured on Tryptic soy agar (TSA, Difco, Spain) supplemented with 3% NaCl (Hara-Kudo et al. 2001). The isolated bacteria were frozen at −80 °C with 20% (v/v) glycerol for further analysis.

Yellow colonies on the TCBS (2–3 colonies from each sample) were randomly selected and subcultured on TSA supplemented with 3% NaCl (TSA, Difco, Spain) and standard procedures were used for the determination of Gram, cell morphology, the oxidase, catalase, indole production, O–F test, motility (Mannitol-Motility agar, Pronadisa, Madrid, Spain), susceptibility to the vibriostatic compound O/129 (10 and 150 μg/disc), were the first tests employed to identify the organisms belonging to Vibrio genus. The API 20NE (bioMerieux) procedure was modified in order to incorporate a 2.5% NaCl concentration in all microtubes. An initial bacterial suspension was therefore prepared in 5 ml of a 2.5% NaCl solution instead of the recommended 0.85% NaCl medium. Incubation time and temperature were maintained within the limits prescribed (37 °C for 24 h) (Croci et al. 2007). Identification was obtained through the APILAB PLUS software (bioMerieux) and was considered acceptable when giving a probability equal to or greater than 80% (Croci et al. 2007).

Bacteria cultured on 3% NaCl tryptone soya broth (TSB-S, Oxoid) at 30 ± 1 °C for 24 h were extracted by boiling. V. alginolyticus was tested for collagenase gene according to Di Pinto et al. (2005). Primer sequences and region amplified are summarized in Table 2. All reactions were performed in a volume of 25 μl on a 9600 Applied Biosystems thermalcycler. V. alginolyticus ATCC 17749 was used as the positive control for collagenase gene detection; a negative control (sterile distilled water) was processed for every 10 strains. Polymerase chain reaction (PCR) products were visualized by electrophoresis on 1.5% agarose gel (Kodak, New Haven, CT, USA) (run at 90 V for 50 min) and photographed using a Bio-Rad Gel Doc 2000 (Bio-Rad Laboratories, Hercules, CA, USA).

In order to exclude the misidentification of other species like V. alginolyticus, the results of all screening and biochemical tests were compared with those available in the literature for the species with closest phylogenetic relations according to previously published papers (Thompson et al. 2005). Misidentification with certain species such as V. cholerae (Nandi et al. 2000), V. vulnificus (Brauns et al. 1991), V. parahaemolyticus (Kim et al. 1999) and V. harveyi (Dalmasso et al. 2009) was also excluded by subjecting the isolates to the PCR assays for the identification of these species.

Protease activities were analysed by spot inoculations on TSA 3% NaCl supplemented with skimmed milk (2% wt/vol), (Alcaide et al. 1999; Hormansdorfer et al. 2000). Haemolysin potency was evaluated using a modification technique of the plate assay described previously (Quindos et al. 1994). In brief, 10 μl of suspension (10⁸cells/ml) were spotted onto human blood and fish blood agar made by mixing 70 ml of each blood with 1,000 ml (TSA) supplemented with 3% NaCl (Khouadja et al. 2012). The plates were incubated at 30 °C for 24 h. Positive haemolytic potency was recorded by the presence of a distinct translucent halo around the inoculum area. The diameters of the zones of lysis and the colony were measured and the ratio (equal to or larger than 1) was used as a haemolytic index to represent the intensity of haemolysin production by the tested strains. All the tests were repeated three times.

All isolated identified V. alginolyticus strains were grown overnight at 37 °C on a TSA supplemented with 3% NaCl. All primers used in this study are summarized in Table 2. Bacterial DNA for PCR analysis was extracted with a Wizard Genomic DNA Purification kit (Promega). Amplification reactions were performed in a 25-μl reaction mixture containing 3 μl of genomic DNA, 5 μl of Green GO Taq buffer (5×), 200 mM of each deoxynucleoside triphosphates (dNTP), 25 μM of each primer and 1 U of GO Taq DNA polymerase (Promega, USA). The mixtures were incubated for 5 min at 94 °C, followed by 35 cycles of amplifications. Each cycle of amplifications consisted of a denaturation step at 94 °C for 40 s, annealing for 40 s and primer extension for 1 min at 72 °C. The mixtures were kept at 72 °C for 10 min for the final extension. The annealing temperatures were as follows: 60 °C for zot and ctxA and 62 °C for ace and VPI genes. CTX- and VPI-positive amplicons to be sequenced were directly purified from PCR tubes or extracted from agarose gel using Wizard SV Gel and PCR Clean-up System (Promega) according to the manufacturer's instructions. DNA sequences were determined by Bio-fab (Rome, Italy). CLC Genomics Workbench 3 software was used to align bidirectional DNA sequences, which were subsequently blasted and analyzed against the GenBank database.

The median infective dose (ID₅₀) test was conducted by intraperitoneal (i.p.) injection as previously described (Alcaide et al. 1999). V. alginolyticus strains showing the presence of CTX gene (n = 4), the five strains with VPI genes and four strains which did not carry any of the investigated genes were grown overnight in TSA 3% NaCl at 30 °C, from each one, a colony was subcultured in 40 ml fresh medium (TSB 3% NaCl) at 30 °C for 16 h.

The cells were harvested by centrifugation (5,000 rpm, 10 min), washed and resuspended in phosphate buffer saline (PBS; 0.01 M) to OD₆₀₀ of 0.2–0.9, so that the bacterial concentrations were 10²–10⁸ cfu ml⁻¹ determined by the dilution-plate method. Sterile PBS was injected i.p. into fish as a control.

Healthy juveniles of S. aurata (weight 5 g, length 8 cm, 20 individuals per testing dose) from a commercial fish farm were randomly sampled for the experiment and acclimated in the water at 25 °C for 4–5 days before testing; water was kept at salinity of 37‰ and under continuous aeration. Before infection, the juveniles were fed with commercial food (INVE Aquaculture Nutrition), while feeding was suspended during the virulence test. Each bacterial dilution containing from 10² to 10⁸ cfu ml⁻¹ was tested by i.p. injection of 50 μl of suspension (20 individuals per testing dose). Sterile PBS was injected i.p. as a negative control. Mortalities were recorded daily for 7 days and were only considered infected if V. alginolyticus was recovered from assayed fish.

The ID₅₀ was calculated by a simple method for estimating 50% endpoints (Reed & Muench 1938). The percentage of infected fish is calculated as follows: A/(A + B) × 100 with A corresponding to infected fish and (B) to not infected fish. The Reed Muench formula is used to calculate the index: index = (% infected fish at dilution immediately above 50%–50%)/(% infected at dilution immediately above 50%–% infected at dilution immediately below 50%). Applying the index calculated using this formula to the dilution that produced the infection rate immediately above 50%, we obtained the dilution of inoculum producing 50% infection. The entire experiment was under ethical approval, and fish were subject to independent health checks during the work.

V. alginolyticus is present in a broad range of aquatic environments and marine organisms (Table 1). All 36 strains collected, including the reference strain, were Gram-negative, motile, pleomorphic, mostly coccobacillary rods. They grew as unpigmented, moist, swarming colonies. All the strains grew in peptone water prepared, respectively, with 3%, 8%, and 10% of NaCl.

Api 20NE strips demonstrated the heterogeneity of V.alginolyticus populations studied, out of 36 V. alginolyticus strains tested, 22 biotypes were identified. The collagenase-based PCR (Di Pinto et al. 2005) allowed confirmation of the phenotypic identification for the 36 strains isolated. The isolates identified as V. alginolyticus by conventional procedures produced an amplicon of 737 bp (Figure 1), characteristic of this species.

V.alginolyticus is present in a broad range of aquatic environments and marine organisms (Table 1). Several studies have demonstrated that these miniaturized biochemical tests, especially API 20 E and 20 NE, were used with success to identify bacteria belonging to the Vibrionaceae family (O'Hara et al. 2003; Croci et al. 2007). But certain environmental strains can produce atypical reactions which will give unacceptable profiles. The molecular approach allows researchers to overcome the drawbacks of traditional biochemical methods because those can be distorted by atypical reactions or by the underestimation of the number of viable cells.

Prominent virulence factors of V. alginolyticus are extracellular, proteolytic enzymes and haemolysines (Zhang & Austin 2005). In addition to caseinase, gelatinase and alkaline serine protease, tissue-damaging collagenase as well as amylase, lecithinase and lipase are produced. 80% of food isolates of V. alginolyticus produce extracellular components toxic to animal cells, fish, crustaceans and corals (Wong et al. 1992; Hormansdorfer et al. 2000; Ben Kahla-Nakbi et al. 2009). Certain isolates were reported to exhibit variation in one or two biochemical reactions and exhibited deviation from the standard methodology in giving few biochemical reactions (Dileep et al. 2003). However, due to the presence of both false-positive or false-negative results in all the biochemical identification methods proposed, some authors (O'Hara et al. 2003; Croci et al. 2007) suggested caution in the interpretation of identifications and advised additional confirmatory testing, such as PCR. Conventional culture-based methods for the detection of V. alginolyticus are laborious and time-consuming (>2 days). Recently, PCR-based methods were developed, which can be completed within 1 day (Zhou et al. 2007). The PCR assay based upon the detection of the gene coding for collagenase could be used as an alternative molecular target for V. alginolyticus detection.

V. cholerae is known to be responsible for the severe diarrhoeic disease that continues to be a global threat to human health. Different genes play a role in the expression of the potent CT (Fields et al. 1992), which is encoded by the ctxAB genes on the filamentous phage CTX along with genes encoding other virulence factors such as the Zonula occludens toxin (Zot), accessory cholera enterotoxin (Ace) and a core-encoded pilin (cep) (Colombo et al. 1994; Lipp et al. 2003). Different studies demonstrated that V. parahaemolyticus and V. cholerae pathogenicity-associated genes may be transferred to different Vibrio spp. strains isolated from the coastal waters (Sechi et al. 2000; Zhou et al. 2007; Ben Kahla-Nakbi et al. 2009). It is now clear that in aquatic biotopes, V. alginolyticus might be continually undergoing genetic change by the acquisition of DNA originating from V. cholerae strains.

Our results show that nontoxigenic strains do not present any of the virulence genes. Some, including ctxA and tdh, may be horizontally transferred, leading to new pathogenic strains. V. alginolyticus strains often possess homologues of the V. parahaemolyticus and V. cholerae virulence genes such as toxR, tlh and VPI (Xie et al. 2005), which suggests that V. alginolyticus may be an important reservoir of many known virulence genes of other Vibrio species in the aquatic environment. It is probable that the aquatic environment harbours different virulence-associated genes scattered among environmental Vibrio.

In the present study, the five strains with VPI genes showed beta haemolytic activity, when the other isolates were found to exhibit weak haemolysis and none of the strains positive for the ctxA genes was found to exhibit haemolytic activity. Several negative strains for the different virulence genes showed a low haemolytic activity. The weak haemolysis points towards the presence of other virulence factors. This is confirmed during the determination of the ID₅₀, when all strains showing haemolytic activity are virulent and cause mortality during the infection tests. When compared with other infectivity assays by the same species in another study, the ID₅₀ in S. aurata was ranging from 2 × 10⁵ to 6.2 × 10⁵ cfu fish⁻¹ (Xie et al. 2005; Ben Kahla-Nakbi et al. 2009), which is very similar to S3 and S9 strain values, whereas others strains, such as S11, S22 and S25 are still deadly with much lower ID₅₀ values. In our study, strains with VPI genes appear to have greater infective power compared to the other strains tested.

The pathogenicity island is a large unstable chromosomal region that encodes several virulence genes, and is present in human pathogenic isolates and absent from non-pathogenic isolates (Hacker et al. 1997). The genes encoding the biosynthesis of TCP were initially shown to reside on a pathogenicity island, designated the VPI (Karaolis et al. 1998). The fact that some strains exhibit none of the sought genes is pathogenic for fish, and can be regarded as a synergistic action of V. alginolyticus enzymes. Prominent virulence factors correlated with the pathogenicity of many Vibrio species are extracellular, proteolytic enzymes and haemolysines (Zhang & Austin 2005). In addition to caseinase, gelatinase and alkaline serine protease, tissue-damaging collagenase as well as amylase, lecithinase and lipase are produced (Hormansdorfer et al. 2000). These factors could facilitate the propagation of the bacteria by causing extensive host tissue damage, thereby degrading host proteins that provide readily available nutrients for bacterial growth (Balebona et al. 1998). However, the pathogenic mechanism of V. alginolyticus is not completely understood. Analysis of the relationship between virulence-associated genes and pathogenicity of V. alginolyticus provides a possible explanation for why the pathogenic mechanism of V. alginolyticus might be different from that of V. parahaemolyticus and V. cholerae. This suggests that V. alginolyticus might have a different virulence gene system and different pathogenic mechanism compared with V. cholerae, despite adhesion and hydrolytic activities. The difference in the hydrolytic enzymes produced by these strains can affect the virulence and they differ one from the other. This can explain the difference among inter- and intra-specific mechanisms of virulence, suggesting the specificity of virulence can be strain-specific.

Different studies demonstrate that the death process induced by V. alginolyticus in mammalian cells is different from that in fish cells, including induction of autophagy, cell rounding and osmotic lysis (Zhao et al. 2011; Ren et al. 2013). V. alginolyticus requires its type III secretion system (T3SS) to cause rapid death of infected fish cells and infection with this species also led to membrane pore formation and release of cellular contents from infected fish cells, as evidenced by lactate dehydrogenase release and the uptake of a membrane-impermeable dye (Zhao et al. 2010, 2011).

In our study, 5 of 13 strains tested for virulence properties were able to cause mortality in juveniles of S. aurata. This strongly supports the fact that individual strains of V. alginolyticus are pathogenic to aquatic animals. However, our results do not show an obvious correlation between the presence of ctxA genes and virulence, whereas in all VPI-positive strains, the virulence was established.

The study performed in IRTA was financed by INIA project RTA 2007000630000 awarded to A.R.

All relevant data are included in the paper or its Supplementary Information.