|Home | About | Journals | Submit | Contact Us | Français|
Vibrio parahaemolyticus is a bacterial pathogen causative of food-borne gastroenteritis. Whole-genome sequencing of V. parahaemolyticus strain RIMD2210633, which exhibits Kanagawa phenomenon (KP), revealed the presence of two sets of the genes for the type III secretion system (T3SS) on chromosomes 1 and 2, T3SS1 and T3SS2, respectively. Although T3SS2 of the RIMD2210633 strain is thought to be involved in human pathogenicity, i.e., enterotoxicity, the genes for T3SS2 have not been found in trh-positive (KP-negative) V. parahaemolyticus strains, which are also pathogenic for humans. In the study described here, the DNA region of approximately 100 kb that surrounds the trh gene of a trh-positive V. parahaemolyticus strain, TH3996, was sequenced and its genetic organization determined. This revealed the presence of the genes for a novel T3SS in this region. Animal experiments using the deletion mutant strains of a gene (vscC2) for the novel T3SS apparatus indicated that the T3SS is essential for the enterotoxicity of the TH3996 strain. PCR analysis showed that all the trh-positive V. parahaemolyticus strains tested possess the novel T3SS-related genes. Phylogenetic analysis demonstrated that although the novel T3SS is closely related to T3SS2 of KP-positive V. parahaemolyticus, it belongs to a distinctly different lineage. Furthermore, the two types of T3SS2 lineage are also found among pathogenic Vibrio cholerae non-O1/non-O139 strains. Our findings demonstrate that these two distinct types are distributed not only within a species but also beyond the species level and provide a new insight into the pathogenicity and evolution of Vibrio species.
Vibrio parahaemolyticus is a gram-negative halophilic marine and estuarine bacterium which is an important pathogen causative of food-borne gastroenteritis and traveler's diarrhea (1). Although most V. parahaemolyticus strains are nonpathogenic for humans, a limited population of these organisms causes human diseases. Almost all clinical V. parahaemolyticus isolates produce the thermostable direct hemolysin (TDH) and/or the TDH-related hemolysin (TRH), which are encoded by the tdh and trh genes, respectively (5, 21). The Kanagawa phenomenon (KP), a beta-type hemolysis on a special blood agar (Wagatsuma agar) (28), is known as a good marker of pathogenic strains (5, 21). V. parahaemolyticus strains which exhibit KP possess the two tdh genes tdhA (tdh2) and tdhS (tdh1) but not the trh gene (6, 19, 21). In contrast, KP-negative clinical V. parahaemolyticus strains possess the trh gene only or both the trh and tdh genes, while the majority of the nonpathogenic strains possess neither tdh nor trh.
TDH and TRH, which have several biological activities in common (5, 20, 30, 33), are considered to be the major virulence factors in clinical V. parahaemolyticus strains (5, 30). However, several studies have demonstrated that although the enterotoxicity was reduced in tdh- or trh-deleted mutant strains from that in the parent strains, the enterotoxic activity of these mutant strains partially remained (24, 25, 34). These results suggest that in addition to TDH and TRH, extra virulence factors are likely to exist in the organisms.
Whole-genome sequencing of a KP-positive V. parahaemolyticus strain, RIMD2210633, has disclosed the presence of two sets of the genes for the type III secretion system (T3SS) on chromosomes 1 and 2, T3SS1 and T3SS2, respectively (16). T3SS is possessed by gram-negative bacteria, especially in animal and plant pathogens, and is thought to contribute to the virulence of these pathogens. T3SS delivers bacterial virulence effectors directly into the host cells, which means that this system contributes to virulence against the host.
Our previous studies demonstrated that the T3SSs of V. parahaemolyticus RIMD2210633 are important for virulence of the organism (13, 23, 26). The genes for T3SS1 are present in all V. parahaemolyticus strains examined (10, 16, 26). T3SS1 of the strain RIMD2210633 is involved in its cytotoxicity (23, 26). In contrast, deletion of the genes for T3SS2 of the RIMD2210633 strain partially eliminated fluid-accumulating activity in rabbit ileal loops, indicating that T3SS2 is involved in enterotoxicity of this strain (26). So far, the genes for T3SS2 have been found only in KP-positive strains (10, 16, 26).
The KP-positive strain-specific T3SS2 genes are present on a pathogenicity island (Vp-PAI) consisting of a ca. 80-kb DNA region on its chromosome 2, and this region was found to contain the genes for TDH as well (16, 31). Pathogenicity islands (PAIs) are large genomic regions (ca. 10 to 200 kb) which are acquired by horizontal gene transfer. PAIs often possess mobile genetic elements and the genes involved in virulence (3, 22).
It has not been made clear whether trh-positive clinical V. parahaemolyticus strains include any PAIs. Several reports have suggested, however, that a PAI may be present in the region surrounding the trh gene in trh-positive V. parahaemolyticus strains (8, 31, 32).
To examine whether a PAI is present in trh-positive V. parahaemolyticus strains, in this study we sequenced the region surrounding the trh gene on chromosome 2 in V. parahaemolyticus TH3996 (a trh-positive strain). This disclosed the presence of an approximately 100-kb DNA region which is considered to be a PAI. Our findings also demonstrated the presence of a set of T3SS genes, which are related to but of a distinctly different lineage from the T3SS2 genes in RIMD2210633.
Table Table11 shows the bacterial strains and plasmids used in this study. All of the V. parahaemolyticus strains were obtained from the Laboratory for Culture Collection, Research Institute for Microbial Diseases, Osaka University. Clinical V. parahaemolyticus strains, including TH3996, were isolated at the Osaka and Kansai International Airport quarantine stations from patients with traveler's diarrhea. The bacteria were cultured at 37°C with shaking in Luria-Bertani (LB) broth (tryptone, 1%; yeast extract, 0.5%) with 3% NaCl. The Escherichia coli DH5α and SM10λpir (17) strains were used for general manipulation of plasmids and their mobilization into V. parahaemolyticus, respectively. The E. coli strains were grown in LB broth or on LB agar. Thiosulfate-citrate-bile-sucrose agar (Nissui, Tokyo, Japan) was used for the screening of mutant strains, and LB agar with 3% NaCl was used for colony hybridization. Antibiotics were used at the following concentrations: ampicillin, 100 μg/ml; kanamycin, 50 μg/ml; and chloramphenicol, 5 μg/ml.
Chromosomal DNA from V. parahaemolyticus strains was extracted from overnight culture of the organism in LB broth with 3% NaCl by means of the QIAamp DNA mini kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. DNA used for subcloning or nucleotide sequence analysis was extracted from E. coli by using the QIAprep Spin miniprep kit (Qiagen) according to the manufacturer's instructions. Cloning, restriction endonuclease procedures, DNA ligation, and transformation of E. coli by plasmids were carried out with previously described standard protocols (29). All of the restriction enzymes and DNA ligation kits were purchased from Takara Shuzo (Otsu, Japan).
A fosmid library was constructed using the CopyControl Fosmid Library Production Kit (Epicentre Biotechnologies, Madison, WI) according to the manufacturer's instructions. Purified genomic DNA of V. parahaemolyticus TH3996 was digested with sonication, and the DNA fragments were ligated into the Cloning-Ready CopyControl pCC1Fos vector (Epicentre). To screen for clones containing a portion of the target region among the 768 fosmid clones, colony hybridization at 40°C was performed as described previously (7). Four probes for the colony hybridization analysis, probe-left, -right, -trh, and -vopC, were prepared by PCR using oligonucleotide primers (see the supplemental material) that were synthesized based on the sequence of two genes of Vp-PAIRIMD2210633 and the trh and vopC genes in strain TH3996, with genomic DNA of strain TH3996 as the template. Each probe was labeled with the PCR DIG probe synthesis kit (Boehringer Mannheim, Mannheim, Germany) with specific primers.
PCR-amplified DNA fragments used for constructing the in-frame deletion mutation of vscC2 were generated by means of overlap PCR as described previously (26) with the PCR primers vscC-1, vscC-2, vscC-3, and vscC-4 (see the supplemental material). Two DNA fragments were amplified by PCR with V. parahaemolyticus TH3996 chromosomal DNA as the template and with the primer pair vscC-1 and vscC-2 and primer pair vscC-3 and vscC-4, respectively. The primer vscC-2 included a complementary 15-bp sequence at its 3′ end and vscC-3 at its 5′ end. The two fragments were then used as templates for a second PCR with the primers vscC-1 and vscC-4, resulting in the construction of a fragment with a deletion in the vscC2 gene. The fragment containing the deletion was purified and cloned into the pT7Blue T-vector (Novagen, Inc., Madison, WI). This fragment was then removed from the pT7Blue T-vector by digestion with BamHI and PstI and cloned into a suicide vector, pYAK1, which contains the sacB gene, conferring sensitivity to sucrose. This plasmid was introduced into E. coli SM10λpir and then mated with V. parahaemolyticus strain TH3996. Thiosulfate-citrate-bile-sucrose agar containing chloramphenicol at a concentration of 5 μg/ml was used to screen vscC deletion mutants, then the mutants were selected on LB plates supplemented with 10% sucrose. We compared the growth rates of the parent and mutant strains in LB medium with 3% NaCl, but we could not detect any significant difference in growth rates between the parent and the mutants.
The vscC2 complementation study was performed as described previously (14, 23, 26). The vscC2 gene was amplified by PCR using the V. parahaemolyticus strain TH3996 chromosomal DNA as the template and the primers comple-F and comple-R (see the supplemental material). The amplicon was cloned downstream from the tdhA promoter in pSA-tdhP (14) by insertion into the BamHI and SalI sites. The plasmid was introduced into the vscC2 mutant strain by electroporation.
To sequence the Vp-PAITH3996 region (approximately 100 kb), we digested the insert fragments of fosmid clones containing Vp-PAITH3996. The fosmid clones were cut into smaller fragments by using the appropriate restriction enzymes and then ligated into the pUC119 vector. Gap closure was achieved with PCR direct sequencing, with primers that were designed to anneal to each end of neighboring contigs. Nucleotide sequencing was performed with the ABI PRISM 3100 genetic analyzer (Applied Biosystems, Foster City, CA) and the BigDye v3.1 cycle sequencing kit (Applied Biosystems). The Genetyx sequence analysis program (Software Development, Tokyo, Japan) was used for computer analysis of DNA sequences. Homology searches against deposit sequences were performed via the National Center for Biotechnology Information using the BLAST network service (http://www.ncbi.nlm.nih.gov) and the BLAST service at the Genome Information Research Center (http://genome.naist.jp/bacteria/vpara/). Sequence information was obtained from the NCBI. The computer program CLUSTAL W was used for the amino acid sequence alignment and phylogenetic analysis.
Secreted proteins were prepared as described previously (26). Secreted proteins from the parent and mutant strains were isolated from the supernatants of bacterial cell cultures grown for 6 h at 37°C in LB medium. Secreted proteins were precipitated by the addition of trichloroacetic acid to a final concentration of 10% (vol/vol). The proteins were collected by centrifugation at 17,500 × g for 30 min at 4°C, and the resultant pellets were washed in ice-cold 100% acetone and suspended in sodium dodecyl sulfate sample buffer.
The secreted proteins used for Western blot analysis were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis with 12% polyacrylamide. The transferred membrane was first probed with anti-VopD2 polyclonal antibody (14) and then with horseradish peroxidase-conjugated goat antirabbit antibody (Zymed Laboratories, South San Francisco, CA). The blots were developed by using the ECL Western blotting kit (Amersham, Piscataway, NJ) according to the manufacturer's instructions.
The enterotoxic activities of the wild-type and mutant strains were assessed with a rabbit ileal loop test (26). The two strains were cultured at 37°C with shaking in LB broth (3% NaCl), diluted 100 times with LB broth (0.5% NaCl), and cultured overnight at 37°C. Next, the cultures were diluted 100 times with LB broth (0.5% NaCl) and cultured at 37°C for 6 h. The organisms were harvested by centrifugation at 3,000 × g for 10 min and suspended in LB broth (0.5% NaCl). The rabbit ileal loop test used 1.5-kg female New Zealand White rabbits, whose small intestine was used to make 5 or 10 ligated loops per rabbit. Bacterial cells (109, 108, or 107 CFU) of the wild-type or mutant strains and negative control were injected into the loops as described previously (26), followed by measurement of the fluid accumulation in each loop 16 h after challenge. We performed 6 (109 CFU) or 10 (108 and 107 CFU) experiments for each sample using different rabbits.
The supplemental material shows the oligonucleotide primers used in this study. PCR conditions for the construction of mutant strains and probes were as follows: after 2 min of denaturation at 94°C, a cycle of 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min was repeated 30 times. To detect the presence of the T3SS genes, PCR was performed using the EX-PCR kit (Takara Shuzo, Kyoto, Japan). The PCR conditions were as follows: after initial denaturation at 94°C for 3 min, a cycle of 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min or 2 min was repeated 30 times. PCR scanning of Vp-PAITH3996 was performed using genomic DNA as a template and a long accurate PCR kit (Takara Shuzo). The PCR conditions were as follows: after initial denaturation at 94°C for 3 min, a cycle of 94°C for 30 s and 65°C for 10 min was repeated 30 times. Custom-synthesized oligonucleotides for the PCR were purchased from Gene Design (Osaka, Japan).
Statistical significance was determined using the t test. A P value of <0.05 was considered statistically significant.
The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL, and GenBank nucleotide sequence database under accession number AB455531.
For the cloning of the DNA region flanking the trh gene on the small chromosome of V. parahaemolyticus strain TH3996, genomic DNA of the strain was isolated and sonicated to yield approximately 40-kb fragments. The DNA fragments obtained were inserted into the Cloning-Ready CopyControl pCC1Fos vector (Epicentre Biotechnologies, Madison, WI) to construct a fosmid library. To screen the clones containing the trh gene and surrounding regions from among a total of 768 fosmid clones, colony hybridization was carried out using three probes (probe-left, -trh, and -right) (Fig. (Fig.1a)1a) (see the supplemental material). Two probes, probe-left and -right, were generated by amplifying genomic DNA from strain TH3996 with primers (see the supplemental material) based on sequences from the KP-positive strain RIMD2210633. This yielded three clones, which hybridized with one each of the probes (Fig. (Fig.1a,1a, clones 1 to 3) and were digested with several restriction enzymes and subcloned into the corresponding sites of the pUC119 vector. By sequencing the subclones of clone 3, we identified the vopC gene at the 5′ end of this clone. To screen new clones containing the vopC gene, which encodes a homologue of the cytotoxic necrotizing factor (13), from the fosmid library, we prepared a new probe (Fig. (Fig.1a,1a, probe-vopC) and performed colony hybridization using this probe. The clone thus obtained (Fig. (Fig.1a,1a, clone 4) was subcloned and sequenced as were the other three. By sequencing these subclones, we could get two contigs, contigs A (ca. 55.1 kb) and B (ca. 36.8 kb), which almost cover the approximately 100-kb region surrounding the trh gene (Fig. (Fig.1b1b).
The sequences of the 3′ end of contig A (approximately 570 bp) and the 5′ end of contig B (approximately 560 bp) showed high homology with the 5′ and 3′ ends, respectively, of an open reading frame (ORF), VPA1357, of the RIMD2210633 strain. Comparison with the nucleotide sequences of the RIMD2210633 genome suggested that the region between contigs A and B corresponds to VPA1357 of RIMD2210633 (Fig. (Fig.1b)1b) (16). VPA1357, 4,869 bp in size, encodes a hypothetical protein which has numerous repeated sequences, spanning nearly two-thirds of the sequence of the gene (16). It was difficult to determine the sequence of this region accurately due to the numerous repeats. Instead of completing the sequencing of the region, we therefore estimated its size by both PCR amplification of the region and construction of restriction maps of the fosmid clone 4. We estimated the gap region between contigs A and B to be approximately 6.1 kb (data not shown) (Fig. (Fig.1b).1b). On the basis of this estimate, we speculated that an approximately 7.2-kb ORF which is homologous to VPA1357 of strain RIMD2210633 exists in this region of the small chromosome in the TH3996 strain. The estimated size of this ORF is obviously larger than that of VPA1357 of the RIMD2210633 strain (4,869 bp).
In order to confirm that the four fosmid clones cover the whole region flanking the trh gene, PCR scanning with 10 PCR primer pairs (see the supplemental material) was used for the genomic DNA of TH3996 (Fig. (Fig.1c,1c, PCR-A1 to -A6 and -B1 to -B4). The expected sizes of the amplicons were obtained for all the primer pairs, suggesting that the fosmid clones cover the whole target region (data not shown).
The G+C content of the ca. 100-kb region we sequenced accounted for approximately 39.8%, which is notably lower than the average G+C content of the small chromosome of RIMD2210633, which is 45.4% (16).
Annotation of the DNA region revealed that the region possesses a total of 100 ORFs: contigs A and B possess 53 and 46 ORFs, respectively, and 1 unsequenced ORF was homologous to VPA1357; and among these ORFs, we identified a set of genes for the T3SS. Although most of the ORFs encode hypothetical proteins, this 100-kb region contained known possible virulence factor genes, including the trh gene and the urease gene cluster (24), in addition to T3SS-related genes, as well as mobile elements, such as transposases (Fig. (Fig.1b).1b). Furthermore, as mentioned above, its G+C content was significantly lower than the genome average. From this series of findings, we hypothesized that the ca. 100-kb region on the small chromosome of the TH3996 strain is a PAI, newly identified for this strain, and tentatively named it Vp-PAITH3996. The PAI of RIMD2210633, which was referred to as Vp-PAI in previous studies (10, 31), was tentatively named Vp-PAIRIMD2210633 for this report.
At least 14 ORFs showed significant homology with T3SS-related genes that have been reported to date (9, 14, 15, 26). These genes were predicted to encode the apparatus proteins of T3SS (vscCJQRSTU and vcrD), an ATPase (vscN), translocons (vopBD), and effectors (vopCLP) (vopP is also known as vopA) (Fig. (Fig.2)2) (9, 14, 15, 26). A recent study of ours found that the TH3996 strain possesses the genes for T3SS1 on its genome, as do other V. parahaemolyticus strains (10). The novel T3SS-related genes found in Vp-PAITH3996 were obviously different from the T3SS1 genes of the RIMD2210633 strain. The genetic organization of the former was similar to that of T3SS2 of the RIMD2210633 strain but not identical. The T3SS-related genes found in Vp-PAITH3996 had 34.5% to 89.5% homology with the corresponding T3SS2 genes in Vp-PAIRIMD2210633. These results demonstrated that the novel T3SS-related gene set present on the small chromosome of the trh-positive strain TH3996 can be considered to be T3SS2-related T3SS.
In Vp-PAITH3996, at least 14 T3SS-related genes were found, as mentioned above, and all of these genes were conserved in T3SS2 on Vp-PAIRIMD2210633 as well. We therefore hypothesized that the T3SS genes found in Vp-PAITH3996 may express a functional secretion system like T3SS2 in Vp-PAIRIMD2210633. To confirm the expression of a secretion system of T3SS in Vp-PAITH3996, T3SS-dependent protein secretion was analyzed. A T3SS-deficient mutant was constructed by disruption of the homologue of the vscC2 gene, which encodes an outer membrane protein of T3SS2 (26), in the TH3996 strain (resulting in TH3996 ΔvscC2). The secretion of VopD2, a translocon protein of T3SS2, by the parent and mutant strains was examined by means of Western blot analysis using the anti-VopD2 antibody of strain RIMD2210633 (14). As shown in Fig. Fig.3,3, VopD2 was detected in the supernatant of the parent strain but not in the vscC2 deletion mutant strain. The secretion of VopD2 was restored by complementation with the vscC2 gene (Fig. (Fig.3).3). These results suggested that the genes for T3SS in Vp-PAITH3996 express a functional secretion system.
Previous studies have demonstrated that T3SS2 of V. parahaemolyticus RIMD2210633 contributes to the enterotoxicity of the organism (26). To determine the possible contribution of the T3SS-related genes encoded in Vp-PAITH3996 to enterotoxicity of the strain, we examined the enterotoxic activity of the wild-type and mutant strains in the rabbit ileal loop test. For this test, we used TH3996 ΔvscC2 and a mutant strain with a deletion in the trh gene (TH3996 Δtrh), the contribution of which to the enterotoxicity of the strain was previously reported (34). Furthermore, we constructed a double deletion mutant strain with deletions in both the trh and vscC2 genes (TH3996 Δtrh ΔvscC2). The wild-type strain or the mutants (109 CFU [each]) or LB broth as a negative control was injected into the ligated ileal loop of rabbits. After 16 h, the small intestines of the rabbits were removed and the fluid accumulation in the ligated ileal loops was measured (Fig. (Fig.4a).4a). The vscC2 deletion mutant strain was associated with little fluid accumulation, at levels similar to that with the negative control, LB. There was no significant difference in fluid accumulation levels between results for the trh gene-disrupted mutant strain and the wild-type strain (Fig. (Fig.4a).4a). Results obtained under these experimental conditions indicated that the T3SS encoded in Vp-PAITH3996 is an important factor in the enterotoxicity of the organism. These results also suggested that the T3SS is functionally expressed under in vivo conditions.
A previously reported significant reduction of enterotoxicity of trh-deleted mutant strains (34) was not observed during the experiment (Fig. (Fig.4a).4a). However, the dose of the bacteria (109 CFU) injected into ligated ileal loops of the rabbits in our study differed from the one used in a previous study (107 or 108 CFU) (34). To confirm the effect of the bacterial inoculation dose, we performed additional rabbit ileal loop tests (Fig. (Fig.4b).4b). The wild-type strain and the trh deletion mutant were injected into the ligated ileal loop of rabbits at doses of 107 CFU and 108 CFU for each, and the fluid accumulation in each loop was measured 16 h after challenge. At both doses, fluid accumulation with the trh deletion mutant strain was significantly lower than that with the wild-type strain (Fig. (Fig.4b),4b), thus confirming the findings of Xu et al. (34). We can therefore conclude that the discrepancy was due to the difference in the dose of the bacterial inoculum.
To analyze the distribution of the T3SS-related genes found in Vp-PAITH3996 in other V. parahaemolyticus strains, a PCR assay was performed using oligonucleotide primer pairs (see the supplemental material) targeting the genes present in Vp-PAITH3996, i.e., trh, ureC, vscC2N2R2S2T2U2, vcrD2, vopB2D2, or vopCLP, for 33 V. parahaemolyticus strains. The PCR primer pairs were designed based on the sequences of genes in strain TH3996 (see the supplemental material). Since they did not amplify the T3SS2 genes of the RIMD2210633 strain (data not shown), these primer pairs were specific to the genes of TH3996. The 33 strains of V. parahaemolyticus included strains of various serotypes and had been isolated for years (Table (Table2).2). Of these strains, 27 were trh positive, 3 KP-positive (tdh positive and trh negative), and 3 tdh and trh negative.
Most of the genes were amplified by PCR in the 27 trh-positive strains tested (Table (Table2);2); however, none of the genes tested in any of the KP-positive strains could be amplified. Furthermore, no amplicons were obtained in the tdh- and trh-negative strains. Although the amplicons for vscT2, vopB2, or vopP were not obtained in a few of the trh-positive strains (Table (Table2),2), most of the T3SS-related genes seemed to be conserved in these strains. Although the trh genes can be divided into two groups, trh1 and trh2, based on the sequences (11), there was no difference in the presence of the T3SS-related genes between trh1-positive and trh2-positive strains (Table (Table2).2). These findings suggest that all the trh-positive V. parahaemolyticus strains tested possess the novel T3SS-related genes, which were found in Vp-PAITH3996. Furthermore, these genes were not detected in KP-positive V. parahaemolyticus strains or in tdh- and trh-negative V. parahaemolyticus strains.
The results presented here suggest that trh-positive V. parahaemolyticus strains possess T3SS-related genes which are closely related to the ones present in Vp-PAITH3996. Although genetic organization of the T3SS-related genes in Vp-PAITH3996 is similar to that of T3SS2 in Vp-PAIRIMD2210633, there are clear distinctions, including the different locations of vopP (Fig. (Fig.2).2). Furthermore, PCR analysis findings suggest that the sequences of the T3SS-related genes in Vp-PAITH3996 are rather dissimilar to those of the T3SS2 genes in Vp-PAIRIMD2210633 (Table (Table2),2), suggesting that these two gene sets belong to different lineages.
Previous studies reported that the V. cholerae non-O1/non-O139 serogroup strains AM-19226, 1587, and 623-39, which do not have the cholera toxin gene, possess a set of T3SS genes (4, 18). In their report, Dziejman et al. pointed out that the gene organization of the T3SS gene cluster of V. cholerae AM-19226 and that of the T3SS2 of V. parahaemolyticus RIMD2210633 are similar (Fig. (Fig.2)2) (4). We therefore performed a phylogenetic analysis of the T3SS genes from the V. parahaemolyticus and V. cholerae strains reported to date.
For the phylogenetic analysis, we used amino acid sequences of five T3SS-related genes (i.e., vscCNRT and vcrD) of six strains of Vibrio species, V. parahaemolyticus TH3996, RIMD2210633 (T3SS1 and T3SS2), and AQ3810 (tdh positive and trh negative), and V. cholerae strains AM-19226, 1587, and 623-39. In addition, six pathogenic species that are known to possess the T3SS genes, namely, those from the genera Yersinia, Shigella, Salmonella, Pseudomonas, and Escherichia, were included in the analysis. By using the neighbor-joining method, we constructed phylogenetic trees for each gene.
The analysis clearly demonstrated that the T3SS-related genes in Vp-PAITH3996 are only distantly related to the genes of T3SS1 of RIMD2210633 and are more likely to belong to the cluster containing T3SS2 of RIMD2210633 and T3SSs of V. cholerae (here referred to as the T3SS2 family) (Fig. (Fig.55).
Unexpectedly, the T3SS-related genes in Vp-PAITH3996 were found to be more closely related to the T3SS genes of V. cholerae strains 1587 and 623-39 than were the T3SS2 genes in Vp-PAIRIMD2210633. Furthermore, the T3SS genes of another V. cholerae strain, AM-19226, were more closely related to the T3SS2 genes in Vp-PAIRIMD2210633 than were those of V. cholerae strain 1587 and strain 623-39 (Fig. (Fig.5).5). These results prompted us to classify the T3SSs of V. parahaemolyticus and V. cholerae into two phylogroups, one comprising T3SS2 in Vp-PAIRIMD2210633 and the T3SS of V. cholerae strain AM-19226 and the other comprising the T3SS in Vp-PAITH3996 and the T3SSs of V. cholerae strains 1587 and 623-39. We tentatively designated the former phylogroup T3SS2α and the latter T3SS2β.
A recent study reported that clinical V. cholerae non-O1/non-O139 serogroup strains V51 and NRT36S also possess the T3SS-related genes in a pathogenicity island (VPI-2) on those chromosomes and that the genes for T3SS in strains V51, NRT36S, and AM-19226 have only ~90% homology with those of 1587 and 623-39 (18). On the basis of this report and our findings, we suggest that the T3SS-related genes in V51 and NRT36S belong to T3SS2α.
Vibrio parahaemolyticus is an important human pathogen. Strains isolated from diarrheal patients produce TDH or TRH or both, but the strains isolated from the environment do not have these properties. TDH, which is produced by KP-positive V. parahaemolyticus strains, is known as a major virulence determinant of these strains. However, whole-genome sequencing of a KP-positive strain, RIMD2210633, revealed the presence of two sets of the T3SS genes, T3SS1 and T3SS2, on its chromosomes 1 and 2, respectively (16). Although the genes for T3SS1 are present in all V. parahaemolyticus strains examined (10, 16, 24), those for T3SS2 are found only in KP-positive strains (10). The genes for T3SS2 of the RIMD2210633 strain are located on a pathogenicity island (Vp-PAIRIMD2210633) (16).
In our study, the DNA region of approximately 100 kb that surrounds the trh gene of trh-positive V. parahaemolyticus TH3996 was sequenced and its gene organization identified. We detected a PAI-like structure (Vp-PAITH3996) in the genome of the strain and demonstrated the presence of the genes for novel T3SS in the Vp-PAITH3996 region (Fig. (Fig.1b1b and and2).2). Gene organization and phylogenetic analysis indicated that the newly discovered T3SS genes in TH3996 are more closely related to the T3SS2 than the T3SS1 genes of RIMD2210633 (Fig. (Fig.55).
In the KP-positive V. parahaemolyticus strain RIMD2210633, the presence of a number of T3SS2-related genes that encode the structural components or effectors has been reported (9, 13, 15, 23, 24). Construction of the mutant strains of those T3SS2-related genes indicated that T3SS2 is important for enteropathogenicity of the KP-positive V. parahaemolyticus strain (13, 23, 24). Since the presence of the novel T3SS-related genes in trh-positive V. parahaemolyticus strains was proven, we investigated whether the T3SS is also involved in the enteropathogenicity of the TH3996 strain. Animal experiments using the deletion mutant strains of a gene (vscC2) for the T3SS apparatus indicated that the T3SS is essential for the enterotoxicity of the TH3996 strain. A previous report showed that the deletion of the trh gene in a trh-positive V. parahaemolyticus strain, TH3996, significantly decreased the fluid accumulation in rabbit ileal loop tests, but the mutant partially retained its enterotoxic activity (34). This suggests that the virulence factor(s) of the TH3996 strain is not limited to TRH alone, but to date no other virulence factor has been reported in trh-positive strains. In our study, we could demonstrate that the T3SS-related genes present in Vp-PAITH3996 are involved in enterotoxicity of the trh-positive strains, making them a strong candidate for this previously unidentified virulence factor.
Although the genetic organization of the T3SS2-related gene cluster of Vp-PAITH3996 was similar to that of Vp-PAIRIMD2210633, the homologies of individual genes of TH3996 with those of RIMD2210633 varied widely, from 34.5% to 89.5%. The PCR with several primer pairs that can amplify the genes for the T3SS of trh-positive V. parahaemolyticus strains could not amplify the genes of KP-positive V. parahaemolyticus strains. This seems to indicate, therefore, that the nucleotide sequences of the T3SS-related genes are different for KP-positive and trh-positive strains. Our findings are supported by those of a study using comparative genomic hybridization analysis of five trh-positive strains, which could not detect the presence of the T3SS2-related genes (10). From these results, we conclude that the trh-positive strain-specific T3SS-related gene cluster does not occur in KP-positive strains. This indicates that a distinct lineage of T3SS2-related genes (T3SS2α and T3SS2β) must exist in KP-positive and trh-positive V. parahaemolyticus strains.
The phylogenetic analysis showed that the two types of T3SS2, T3SS2α and T3SS2β, are also distributed among pathogenic V. cholerae non-O1/non-O139 serogroup strains (Fig. (Fig.5).5). However, the gene compositions of VPI-2 of V. cholerae strains (18), except for the T3SS gene cluster, were not similar to those of Vp-PAITH3996 and Vp-PAIRIMD2210633. Thus, the gene compositions of the PAI cassettes in V. parahaemoloyticus and V. cholerae, except for the genes for T3SS, were clearly different, implying that the evolutionary history of the PAIs of the two species is also different.
The genetic organization of Vp-PAITH3996 was found to have features in common with that of Vp-PAIRIMD2210633, with both PAIs containing the T3SS and hemolysin genes, i.e., tdh or trh (Fig. (Fig.6).6). Recently the three ORFs on Vp-PAIRIMD2210633, VPA1394, VPA1395, and VPA1396, were identified as genes for the Tn7 superfamily of transposons (31). Tn7 is a bacterial transposon that is widespread in diverse species and is involved in the formation of genomic islands (27). In the aforementioned study by Sugiyama et al., it was found that on chromosome 2 of the KP-positive strain RIMD2210633, Vp-PAIRIMD2210633 was flanked by 5-bp direct repeats (DRs) and was inserted between VPA1309 and VPA1397 (31). It was thus speculated by the authors that this Tn7 superfamily-like genetic element is involved in the formation of Vp-PAIRIMD2210633 on chromosome 2 of strain RIMD2210633. In the trh-positive strain TH3996, we detected the presence of three ORFs in Vp-PAITH3996 that showed high homology with VPA1394, VPA1395, and VPA1396, which, as mentioned above, are components of the Tn7 superfamily (Fig. (Fig.6).6). In addition, we found that Vp-PAITH3996 was flanked by 5-bp DRs (Fig. (Fig.6).6). Since these structural features of Vp-PAITH3996 were similar to those of Vp-PAIRIMD2210633, it was speculated that the Tn7 superfamily transposons found in Vp-PAITH3996 also might be involved in the initial formation of the PAI cassettes. However, since the entire set of Tn7 transposon genes was not conserved in Vp-PAITH3996, as was reported in the case of Vp-PAIRIMD2210633, the Tn7 superfamily transposons in Vp-PAITH3996 and Vp-PAIRIMD2210633 might no longer be functional as transposable elements.
Vp-PAITH3996 and Vp-PAIRIMD2210633, large gene clusters of more than 80 kb, were acquired as a result of horizontal gene transfer in V. parahaemolyticus. However, it is unclear how V. parahaemolyticus integrated the PAI cassette into its chromosome after acquisition of the foreign PAI cassette into its cytoplasm. As mentioned above, the Tn7 superfamily in Vp-PAIRIMD2210633 and Vp-PAITH3996 may no longer be functional, and there are no reports of such a large DNA region being transferred horizontally by the Tn7 superfamily, which might thus not be involved in insertion of the PAI cassette in the V. parahaemolyticus chromosome. It is considered likely that the exchange of O-antigen-encoding cassettes, which are large DNA fragments (more than 32 kb in size), in V. cholerae is mediated by homologous recombination, because the regions flanking the cassettes show high homology (2). Similarly, the sequences flanking Vp-PAITH3996 and Vp-PAIRIMD2210633 in V. parahaemolyticus were highly conserved. This implies that the large PAI cassettes may be integrated into the chromosome by means of homologous recombination in V. parahaemolyticus.
It is interesting that the distribution of T3SS2α and T3SS2β is not limited to being found within a species but goes beyond the species level (Fig. (Fig.7).7). The presence of common secretion system genes in organisms beyond the species level suggests that the possession of such secretion systems may confer some common beneficial effect(s) on the organisms. Although the nature of such benefit(s) is as yet unknown, attempts to identify the role of those secretion systems in aquatic environments may lead to a better understanding of the life cycles of human pathogens in nature. This points to the significance of the results of our phylogenetic analysis of the T3SS genes, because we believe they provide a new insight into the pathogenicity and evolution of Vibrio species.
This work was supported by Grants-in-Aid for Scientific Research on Priority Areas and for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
We thank the staff of the Kansai International Airport Quarantine Station for supplying the V. parahaemolyticus strains.
Editor: J. B. Bliska
Published ahead of print on 15 December 2008.
†Supplemental material for this article may be found at http://iai.asm.org/.