Introduction

Vibrio parahaemolyticus is a natural inhabitant of the marine environment, a halophilic bacterium which frequently causes food borne illness where raw or undercooked seafood is consumed [1]. However, it is believed that only certain environmental forms of the species carry virulence factors, which enable the bacterium to cause disease in humans. Currently, isolates of V. parahaemolyticus are considered virulent strains if they carry either thermostable direct haemolysin (TDH) coded by the tdh gene or the TDH-related haemolysin (TRH) coded by the trh gene or both [2, 3]. The presence of the tdh gene that correlates with Kanagawa phenomenon (KP)-positive phenotype has been reported in 9% of environmental V. parahaemolyticus isolates in and around Mexico [4] and only 6% of isolates found in Molluscan shellfish [5] on the Atlantic and Gulf Coast. Only 0.33% (5 out of 1500) isolates from the aquatic environment around Bangladesh carried the tdh sequence [6]. In Italy, Caburlotto et al. [7] have demonstrated that 6% of the environmental V. parahaemolyticus strains presented the trh gene. In France, 5% of environmental V. parahaemolyticus strains were found to be trh-positive [8]. In Japan, Shirai et al. [2] found 7% of environmental strains of V. parahaemolyticus carried the trh gene. Furthermore, Miyamoto et al. [9] reported that almost all clinical V. parahaemolyticus isolates exhibit the KP beta-type haemolysis produced by TDH on Wagatsuma agar but only 1–2% found in environmental isolates did. It has been recognized as a marker for discriminating pathogenic from non-pathogenic strains [10] but does not do so completely. More recently, Park et al. [11] proposed a new concept in the pathogenic mechanism of V. parahaemolyticus that, as well as TDH and TRH, a type three secretion system (T3SS2) also plays a role in the pathogenicity to humans. Other workers have suggested that there are other unknown effectors involved in the pathogenic mechanism of this organism [12].

There are simple, validated molecular methods for identification of the species V. parahaemolyticus. The TOX R transmembrane protein encoded by the toxR gene is unique to V. parahaemolyticus. A toxR targeted polymerase chain reaction (PCR) for specific detection of V. parahaemolyticus producing a 368 bp amplicon has been developed by Kim et al. [13]. Another PCR target is the pR72H fragment, which contains a phosphatidylserine synthetase gene and a non-coding region, and the primers used amplify a fragment of either 320 bp or 387 bp in length, that are considered specific for identifying V. parahaemolyticus [14]. Moreover, Nasu et al. [15] have used the orf8 gene as a genetic marker for pandemic strains. However, Okura et al. [16] have suggested that orf8 and toxRS genes are not a reliable genetic marker for identification of pandemic strains.

Molecular typing methods can be used to identify markers of virulence/pathogenicity for detection purposes, but before their use, it is necessary to have gleaned which isolates are likely to be virulent or avirulent. This information is likely to be related to the source of the isolates (i.e. clinical or environmental). Currently, this information is corroborated by performing tests for the presence of putative virulence factors such as tdh, trh and T3SS2. Molecular typing tests such as pulsed-field gel electrophoresis (PFGE), ribotyping, tDNA, enterobacterial repetitive intergenic consensus (ERIC) or RAPD-PCR can then be used to identify markers which differentiate virulent and avirulent forms of the species. These methods can also be used for the characterization of pandemic clones, for example, Okura et al. [17] have used RAPD-PCR to differentiate pandemic and non-pandemic strains of V. parahaemolyticus. In addition, Khan et al. [18] have used ERIC-PCR to identify the V. parahaemolyticus O3:K6-specific DNA segment.

Previous work done by Miah [19] used RAPD-PCR to identify a unique DNA fragment produced in the majority of clinical isolates of V. parahaemolyticus. Here, we have used this fragment to create a PCR-based method for differentiating many clinical isolates from environmental isolates of V. parahaemolyticus.

The aim of this study was to develop a reliable, rapid and simple method for differentiating virulent clinical and mainly avirulent environmental isolates of V. parahaemolyticus that were collected from different geographical regions, having different serotypes and genotypes. Initially, the isolates were subjected to RAPD-PCR analysis, which allowed the recognition of a DNA sequence unique to virulent isolates and the development a PCR using a single set of primers to detect them. We have also compared the ability of this new PCR based test to differentiate clinical and environmental isolates with the use of PCR tests for tdh, trh and T3SS2 and a test for cytotoxicity.

Materials and methods

Bacterial isolates and growth media

Bacterial isolates used in this study are listed in Tables 1 and 2. All isolates were maintained on slants of marine salts agar (MSA, Difco) at room temperature (21°C) in the dark. All broth cultures were grown at 37°C with aeration in Tryptone Soya Broth (TSB, Oxoid). These cultures were used to extract template DNA for PCR analysis and concentrate extra-cellular products (ECPs).

Table 1 Vibrio parahaemolyticus isolates used in this study
Full size table
Table 2 Vibrio species used to test the specificity of the assay
Full size table

DNA extraction

Genomic DNA from bacteria was extracted by using a genomic DNA extraction kit (Qiagen, Ltd., UK) according to the manufacturer’s instructions and quantified by comparison with lambda DNA standards (Sigma Aldrich, Gillingham, UK) on agarose gels (0.8%). The DNA was stained with ethidium bromide (0.5 μg/ml), and photographed under UV light (309 nm) with a Gel Documentation system (Uvi-Tech, UK), and the DNA concentration was determined for each isolate by comparison with a range of the following known amounts of standard bacteriophage λ: 0.552 μg, 0.276 μg and 0.138 μg and using ‘UV Photo’ software (UVi-Tech). Alternatively, DNA was quantified with a NanoDrop 100 spectrophotometer (Lab-Tech, UK). Optical density was measured at 260 and 280 nm.

RAPD-PCR

RAPD analysis was performed using 55 isolates of V. parahaemolyticus using RAPD Analysis Beads (Amersham) and a set of P1 (5’-d [GGTGCGGGAA]-3’) and P5 (5’-d [AACGCGCAAC]-3’) 10-mer primers. For each reaction, a bead was resuspended in a 25-μl volume to contain 1 unit Amplitaq DNA polymerase and Stoffel fragment, 0.4 mM of each dNTP, 2.5 μg BSA and buffer (3 mM MgCl2, 30 mM KCl and 10 mM Tris, [pH 8.3]). To each reaction 25 pmol (5 μl) of the appropriate primer and DNA template (10 ng) were added to generate optimum banding patterns for agarose gel electrophoresis. All RAPD-PCR reactions were performed in a Primus 96 plus thermal cycler (MWG Biotech) programmed for one cycle consisting of: 95°C for 5 min; and 45 cycles consisting of 95°C for 1 min, 36°C for 1 min and 72°C for 2 min.

PCR-OMP

PCR primers to detect the outer membrane protein sequence (OMP) of virulent isolates were constructed. VPOMP1 and VPOMP2 were designed by using primer design software Primer3 (http://biotools.umassmed.edu/bioapps/primer3_www.cgi). A lack of secondary structure and primer-dimer formation was assessed using a DNA calculator (Sigma-Genosys). Primer sequences were then run through BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Primers were synthesised by MWG Biotech Ltd., Germany. Each reaction contained 1 unit (U) of Taq DNA polymerase, 1x reaction buffer containing 1.5 mM MgCl2 (Roche Applied Science, UK), 50 pmol of each primer, each deoxynucleoside triphosphate at a concentration of 0.2 mM, 20 ng of bacterial DNA and the reaction volume was made up to 50 μl using PCR grade water. PCR conditions started with 1 cycle of pre-denaturation at 94°C for 5 min, followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and elongation at 70°C for 2 min. The amplification ended with a final elongation at 70°C for the 5 min. Next, 15 μl of amplified PCR products were electrophoresed in 2% agarose gels submerged in 1x TBE buffer (pH 8), with 100 bp DNA ladder (Invitrogen, UK). Gel electrophoresis was performed at 90 V for 6 h by using Pharmacia tanks (Biotech, Sweden). Following electrophoresis, the DNA bands were stained by immersing the gel for 20 min in a solution containing ethidium bromide (EtBr, 0.5 μg/ml) in TBE buffer followed by 10 min destaining in distilled water to remove any unbound EtBr. The stained bands were visualized with UV light (309 nm) using a trans-illuminator and gels were recorded as digital TIFF images using a gel documentation system (UVI-Tech). A negative control without DNA was included in each run to monitor for contamination.

Cloning and sequencing of DNA fragment

PCR fragments were excised and extracted from electrophoresis gel by using an agarose gel extraction kit (Roche) and the amplified PCR product was purified by using the ‘High Pure PCR Product Purification Kit ’ (Roche) according to the manufacturers’ instruction and ligated into pGEM-T Easy vector (Promega, UK). The ligation product was transformed into E. coli JM109 and cultured on LB agar media including ampicillin (100 μg/ml), IPTG (0.5 mM), and X-Gal (80 μg/ml). Next, plasmid DNA was isolated from white colonies showing inserts using the Miniprep Plasmid DNA Extraction Kit (BioRad, Hemel Hempstead, UK), and cloning was verified by digestion with restriction enzymes. Both strands of cloned DNA were sequenced directly using a cycle sequencing method by MWG-Biotech Ltd., Germany. Computer analysis of the DNA sequence data was performed via the National Centre for Biotechnology Information (NCBI) using GenBank databases search and BLAST programs (NCBI). The nucleotide sequences were submitted to GenBank (accession number BA000031.2).

PCR detection of tdh, trh and T3SS2 genes

PCR assay used primers (5’-CCACTACCACTCTCATATGC-3’ and 5’-GGTACTAAATGGCTGACATC-3’) for tdh and (5’-GGCTCAAAATGGTTAAGCG-3’ and 5’ CATTTCCGCTCTCATATGC-3’) for trh, as previously described [20]. T3SS2 specific primers were designed by using sequence information from isolate RIMD2210633 [21], three pairs of PCR primers targeted to two genes on T3SS2 as well as a region flanking the T3SS2 genes representing the V. parahaemolyticus pathogenicity island (Vp-PA1) [22]. The sequences of these primers are as follows: (VPA1321F 5’-GGTTAGTGAATCCAACCAAACCGC-3’ and VPA1321R 5’-TTGCCGTGCATGTCATACAACCAG-3’), (VPA1339F 5’-GACACTCGCTGTTGTTCTCAGGTA-3’ and VPA1339R 5’-GTAAGCGCGTGATGTTAGCTCTTC) and (VPA1355F 5’-GGCATGTGGTGTCTATTTGACACG-3’ and VPA1355R 5’-TACGACAACTGCAGGTAGTCCAAG-3’) were used in the PCR assay to detect T3SS2. The reaction mixture contained 1 U of Taq polymerase (Roche Applied Science, UK), reaction buffer containing 1.5 mM MgCl2 (Roche), 24 pmol of each primer, 0.2 mM for each deoxynucleoside triphosphate (dNTP), 10 ng of bacterial DNA and the reaction volume was made up to 25 μl using molecular biology grade water (Sigma, UK). Optimised PCR cycling conditions were 1 cycle of 2 min at 94°C followed by 30 cycles of 1 min at 94°C, 45 s at 65°C and 90 s at 72°C followed by a final extention time of 5 min at 72°C.

Cytotoxicity assay

Human colorectal epithelial cells (Caco-2) were grown and maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, UK) supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin, 100 U/ml penicillin and 5 mM glutamine and incubated in an atmosphere containing 5% CO2 at 37°C for 21 days for cell confluence. The medium was replaced every 2–3 days to avoid the cytotoxicity of dead cells. After confluent growth of the cells in a 25-cm2 flask, the cells were washed twice with 0.53 mM Versene solution and treated with 0.25% trypsin-EDTA solution. Before infection, cells were washed with DMEM without phenol red.

Cultures of each bacterial isolate were grown in TSB (10 ml) at 37°C for 48 h with shaking. After centrifugation at 2,800 x g for 15 min at 4°C, the culture supernatants were obtained and re-centrifuged at 2,800 x g for 15 min and concentrated by using Minicon B15 Clinical Sample Concentrators (Millipore, USA) according to the manufacturer’s instructions. Cytotoxicity of the concentrated extracellular products (ECPs) of all clinical and environmental isolates of V. parahaemolyticus was determined in the Caco-2 cell cultures by release of lactate dehydrogenase (LDH) using the CytoTox96® Non-Radioactive Cytotoxicity Assay kit (Promega, UK) based on the manufacturer’s instructions. Approximately 5x103 trypsinized cells in 100 μl of the appropriate medium were dispensed in each well of 96-well microtiter plates (Fisher Scientific International, Inc., UK), and 15 μl of concentrated ECPs was added to the confluent monolayers of tissue cells grown in 96-well plates. After 4 h, the supernatants were collected after centrifugation at 500 x g for 5 min. LDH concentration was measured by reading absorbance at 490 nm using the Fusion Universal microplate analyzer (Optimax, Molecular Devices UK Ltd.). Cytotoxicity calculations were based on the formula: % cytotoxicity = 100 x ([optical density at 490 nm ‘OD490’ for experimental release ‘exp’ – OD490 for spontaneous release ‘spon’]/ [OD490 for maximum release ‘max’ – OD490 for spontaneous release]), where the spontaneous release was the amount of LDH released from the cytoplasm of untreated cells, while the maximum release was the amount of LDH release from toxin-untreated cells that were totally lysed with the lysis buffer included in the kit.

Statistical analysis

Statistical analysis was performed using t test by SPSS 18.0 software (SPSS; Plymouth University, UK). A p < 0.05 was considered statistically significant.

Results

This study builds upon the initial observation by Miah [19] that RAPD performed with Amersham P1 and P5 primer set produces a unique band in some isolates of V. parahaemolyticus. This band was named band Y. In the present study, band Y was commonly observed as an approximately 600 bp fragment on electrophoresis of PCR products produced from clinical isolates and rarely in environmental isolates tested (Fig. 1). Patterns generated by RAPD-PCR were reproducible in three different assays. The 600 bp amplified DNA fragment was purified from agarose gels and cloned into the vector pGEM using E. coli JM109 as the host organism and then plasmid DNA was extracted, after blue/white screening, from white colonies. Derived sequences of the fragment are shown in (Fig. 2). BLAST analysis in comparison with sequences available at GenBank and EMBL [23] revealed that the 600 bp sequence of the obtained RAPD-PCR fragment was identical to the sequences occupying 984020 bp to 984595 bp on chromosome 1 of V. parahaemolyticus RIMD 2210633 DNA (accession no.BA000031.2). The nucleotide sequence characterized as a gene encoding a putative outer membrane protein (accession no.BA000031.2), with 100% homology for V. parahaemolyticus, 93% homology to a putative outer membrane protein of V. alginolyticus (accession no. ZP_01260526.1), and 90% homology to a conserved hypothetical protein of V. harveyi (accession no. ZP_06175389.1). Primer pair VPOMP1 (5’-GTCACGCGGCCAAACAAAGAGA-3’) and VPOMP2 (5’ACCGCATATCACTGTTGGCTGGG-3’) was designed by using the sequence information of V. parahaemolyticus (accession no. BA000031.2). Using this primer pair, PCR was performed on genomic DNAs of a collection of V. parahaemolyticus isolates of various serotypes as described in Table 1, together with 22 isolates of various other Vibrio species as shown in Table 2. The expected 200 bp product was found in the 21 out of 23 clinical isolates of V. parahaemolyticus, including NCIMB reference strain R1 from clinical source and only 1 out of 32 environmental isolates (Table 1). No band was seen with any of the other species of Vibrio tested (Table 2). Out of 22 V. parahaemolyticus isolates that produced the 200 bp fragment, 18 (81%) were tdh-positive, 10 (40%) were trh-positive, and 19 (86.4%) were positive for T3SS2 (Table 1). In comparison, of the 33 V. parahaemolyticus isolates that did not produce the 200 bp PCR product, 5 (15%) were tdh-positive, 7 (21%) were trh-positive and 10 (30%) were positive for T3SS2.

Fig. 1
figure 1

Randomly amplified polymorphic DNA polymerase chain reaction (RAPD-PCR) fragment patterns for genomic DNAs of clinical and environmental V. parahaemolyticus strains with primer pair P1 and P5. Lane 1, negative control; Lane 2, E. coli BL21 (control DNA); Lane 3, C2; Lane 4, C4; Lane 5, C5; Lane 6, C7; Lane 7, C8; Lane 8, E16; Lane 9, E22; Lane M, molecular size marker (100 bp DNA ladder). Arrowhead indicates approximately 600-bp fragments commonly found in the virulent strains

Full size image
Fig. 2
figure 2

Nucleotide sequence of the randomly amplified polymorphic DNA polymerase chain reaction (RAPD-PCR) of clinical V. parahaemolyticus product and locations of PCR primers VPOMP1 and VPOMP2

Full size image

To investigate the cytotoxicity of representative strains, the internal enzyme, lactate dehydrogenase (LDH) release, was measured from Caco-2 cells treated separately with ECPs of each isolate. Table 1 shows that strains of V. parahaemolyticus that were positive for the 200 bp PCR products caused significantly (p < 0.05) more (mean 72.88%) LDH release than environmental isolates (mean 15.3%), with the exception of one environmental isolate (E9) that produced the 200 bp amplicon.

Discussion

There are several molecular typing systems for investigating the differences between the bacterial isolates; however, the main criteria to choose the typing methods are their reproducibility, typeability, constancy, cost and simplicity of performance. PFGE is a gold standard technique that has been used for molecular typing of V. parahaemolyticus [19, 24], but it is expensive and time-consuming. RAPD-PCR is quick, cheap, relatively simple to achieve and was used in this study to initially differentiate V. parahaemolyticus clinical (virulent) isolates from environment (mainly avirulent) isolates. RAPD-PCR with random primers P1 and P5 of most clinical isolates DNA showed a unique band Y at approximately 600 bp, which was purified from agarose gels, cloned and sequenced and used to design a specific primer set in this study.

Using data from the sequence information for band Y, primer pair VPOMP1 and VPOMP2 was designed to amplify a 200 bp section of a gene encoding a yet uncharacterized outer membrane protein. This PCR band was produced by 21 out of 23 (91%) clinical (virulent) isolates tested and only 1 out of 32 environmental (mainly avirulent) isolates, the latter rate being in-line with previous studies to determine the likely capacity to cause disease amongst environmental strains [5, 6, 8]. While not 100% for all clinical isolates, this new PCR detected known virulent isolates better than any of the current individual PCR tests for tdh positives, trh positives or T3SS2 positives (Table 1). Baffone et al. [25] have found that a limited number of environmental strains of V. parahaemolyticus are mostly responsible for V. parahaemolyticus food-borne diseases. While the tdh and trh genes have been considered as virulence marker genes of V. parahaemolyticus for some time [20], type three secretion systems 2 (T3SS2) have emerged as another marker. Makino et al. [21] and Ono et al. [26] have reported that T3SS1 was found on chromosome 1 of all V. parahaemolyticus strains while T3SS2 was additionally found on chromosome 2 of most KP-positive strains. The present study shows that of clinically recognized isolates of V. parahaemolyticus tested 91% produced the 200 bp PCR-OMP product, which is better than the 84% recognized as tdh-positive, 40% as trh-positive, and 86% as T3SS2-positive isolates, suggesting that this new PCR test would be of value to detect strains potentially pathogenic to humans. These results support the notion that tdh and trh genes are not always present in pathogenic forms of V. parahaemolyticus [27]. And it reinforces the idea that the presence of T3SS2 is related to the pathogenicity of a V. parahaemolyticus isolate [11].

Unfortunately, detailed information on the sources of the clinical isolates or indeed the immune status of the patients suffering from infection with V. parahaemolyticus used in this study was not available. However, each was isolated from human cases of gastroenteritis. A possible explanation for the existence of clinical isolates that lack the 200 bp PCR product (or any other putative virulence factor) is that some patients may have been immunocompromised and therefore more susceptible to less virulent strains. Conversely, we should be aware that a positive detection of a gene sequence by PCR does not necessarily mean that gene products will be produced phenotypically. Moreover, it should be recognized that virulence in any bacterium will rarely be defined by a single gene.

By using human Caco-2 intestinal cells, it was shown that the mean cytotoxicity of V. parahaemolyticus ECPs was significantly greater for isolates that produced the 200 bp PCR-OMP product than for those that lacked this fragment. Importantly, these results found quantitative evidence for enhanced virulence-associated characteristics of OMP group isolates compared to those of other V. parahaemolyticus strains, since the quantity of LDH released is relative to the degree of the host cell membrane damage [28]. Based on cytotoxicity to Caco-2 cell lines, Yeung et al. [29] also differentiated virulent and avirulent strains and the cytotoxicity of clinical isolates was significantly different (p ≤ 0.05) from food isolates. Raimondi et al. [30] reported that the high TDH concentration had been demonstrated to induce cytotoxicity in Caco-2 cells, but in the present study there is a high cytotoxic effect of ECPs in isolates that produced the 200 bp fragment even though in some instances, one or more of the known virulence genes (tdh, trh and T3SS2) may be absent.

Okura et al. [17] have similarly used arbitrarily primed-PCR with random primer P2 (5’-d[GTTTCGCTCC]-3’) to investigate a unique band (approximately 930 bp) in pandemic strains of V. parahaemolyticus (O3:K6, 04:K68, O1:KUT, O1:K25 and O1:K26) corresponding to a hypothetical protein, which was found to be 80% homologous to the Mn2+ and Fe2+ transporter of the NRAMP family of V. vulnificus. Okura et al. [17] then designed primers to detect a specific amplicon at 235 bp, which was a useful diagnostic tool to distinguish the pandemic group from other V. parahaemolyticus strains. In the present study typing of V. parahaemolyticus isolates with primer pair VPOMP1 and VPOMP2 found that a 200 bp PCR product was found in clinical O3:K6, O4:K12, O1:K69, O5:K68 and O1:KUT (untypeable) serotypes and four clinical O3:KUT serotypes. These isolates were sourced from different geographical regions. Other workers in the United States have developed a PCR to distinguish V. parahaemolyticus O3:K6 from non-O3:K6 isolates [18].

While the present study developed a PCR which detects nearly all virulent V. parahaemolyticus isolates from various global sources of serotype, it is realized that the number of isolates used is relatively small and thus represents a preliminary, but informative, study. A larger number of clinical V. parahaemolyticus isolates and other species should be examined in future studies, perhaps with a number of collaborators. The 200 bp PCR-OMP offers the possibility of screening environments or seafoods and other foodstuffs for the presence of potentially pathogenic forms of V. parahaemolyticus. Furthermore, the nature and role of this OMP in this bacterium and its contribution to virulence needs to be studied more.

In conclusion, this study has shown that a simple PCR assay shows promise for the identification of virulent forms of V. parahaemolyticus. Further studies are currently in progress for the future development of a multiplex PCR that can be used to reliably identify all V. parahaemolyticus species isolates and distinguish between virulent and avirulent strains, simultaneously.