Identification of Two Translocon Proteins of Vibrio parahaemolyticus Type III Secretion System 2

Bacterial strains and plasmids. V. parahaemolyticus strain RIMD2210633 (KP positive, serotype O3:K6) (23) was used for construction of deletion mutants and functional studies. E. coli DH5α and SM10λpir (25) were used for general manipulation of plasmids and for mobilization of plasmids into V. parahaemolyticus. Strains and plasmids used in this study are listed in Table 1.

Analysis of secreted proteins.Secreted proteins were prepared as described previously (21). Briefly, bacterial strains were grown for 6 h in LB medium supplemented with 0.5% NaCl. Secreted proteins were harvested by precipitation with cold trichloroacetic acid at a final concentration of 10% (vol/vol) on ice for 60 min, followed by centrifugation at 48,000 × g for 60 min. The pellets were rinsed in cold 100% acetone and solubilized in Laemmli buffer.

Immunoblot analysis.Samples used for Western blot analysis were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (12% polyacrylamide). Transferred membranes were probed with anti-VopB2, anti-VopD2, or anti-VopP polyclonal antibodies (32) and then probed with horseradish peroxidase-conjugated goat anti-rabbit antibody (Zymed). The blots were developed by using an ECL Western blotting kit (Amersham).

Cytotoxicity assays.The cytotoxicity assays were performed as described previously (21). Briefly, eukaryotic cells were cocultured for 6 h with phosphate-buffered saline-washed bacteria at a multiplicity of infection (MOI) of 10. The release of lactate dehydrogenase (LDH) into the medium was quantified using CytoTox96, a nonradioactive cytotoxicity kit (Promega), according to the manufacturer's instructions. The LDH release (percent cytotoxicity) was calculated using the following equation: (optical density at 490 nm [OD₄₉₀] for experimental release − OD₄₉₀ for spontaneous release)/(OD₄₉₀ for maximum release − OD₄₉₀ for spontaneous release) × 100. The spontaneous release was the amount of LDH released from the cytoplasm of uninfected cells, whereas the maximum release was the amount of LDH released by total lysis of uninfected cells.

Adenylate cyclase reporter gene.Caco-2 cells were infected with V. parahaemolyticus harboring a pSA-VopT-CyaA plasmid at an MOI of 10 for 6 h. After infection, the cyclic AMP (cAMP) levels in infected cells were determined using a cAMP Biotrak EIA kit (Amersham Biosciences) according to the manufacturer's instructions.

T3SS-dependent pore formation assay.The T3SS-dependent permeability assay was performed as described previously (34). Briefly, Caco-2 cells were infected with V. parahaemolyticus for 2 h at an MOI of 100, which was followed by exposure to 25 μg/ml ethidium bromide and 5 μg/ml acridine orange.

Fractionation of infected host cells.Fractionation of infected host cells was performed as described previously (10, 11, 20). Briefly, Caco-2 cells were infected with V. parahaemolyticus for 3 h at an MOI of 10. Infected cells were washed with ice-cold phosphate-buffered saline and then suspended in homogenization buffer (3 mM imidazole, 250 mM sucrose, 0.5 mM EDTA; pH 7.4) supplemented with protease inhibitor cocktail (Sigma) and mechanically disrupted by passage through a 22-gauge needle. The homogenate was centrifuged at 3,000 × g for 15 min to pellet unbroken cells, bacteria, nuclei, and cytoskeletal components. The supernatant was subjected to ultracentrifugation at 41,000 × g for 20 min to separate the host cell membrane (pellet) from the cytoplasm (supernatant).

Rabbit ileal loop test.Rabbit ileal loop tests were performed as previously described (31, 32). The bacteria tested (10⁹ CFU) were suspended in 1 ml LB broth with 0.5% NaCl and injected into ligated ileal loops of rabbits, and then the fluid accumulation in each loop was measured 15 h after challenge. The fluid accumulation was expressed as the amount of accumulated fluid (in milliliters) per centimeter of ligated rabbit small intestine.

Statistical analysis.All data were expressed as means and standard deviations of three determinations per experimental condition. The statistical significance was determined by a one-way analysis of variance, followed by Dunnett's multiple-comparison test. A P value of <0.05 was considered statistically significant.

Identification of T3SS2-dependent secreted proteins.Although type III secretion is generally considered to be contact dependent, V. parahaemolyticus can secrete T3SS-dependent secreted proteins (translocons and effectors) into culture medium (32). Therefore, to identify T3SS2-dependent secreted proteins, supernatants from cultures of POR-2, which is a strain with an inactivated T3SS1 apparatus, and POR-2 ΔvcrD2, which is a strain deficient in both T3SS1 and T3SS2, were separated by SDS-PAGE (Fig. 1A). Most bands were obtained for both strains, but three abundant proteins with molecular masses of approximately 40, 30, and 27 kDa were found to be unique to the POR-2 strain. Analysis of their N-terminal amino acid sequences revealed that two of these three proteins, VPA1362 (40-kDa band) and VPA1361 (30-kDa band), are encoded in tandem in the T3SS2 region. VPA1362 showed a high degree of identity (94%) to the predicted secreted protein VspD from non-O1, non-O139 V. cholerae AM-19226 (6) and 22% identity with Aeromonas salmonicida subsp. salmonicida A449 AopB protein (Fig. 1B). It has been reported that AopB has 42% identity with Yersinia enterocolitica translocon protein YopB and is predicted to be involved in the translocation function of the A. salmonicida subsp. salmonicida T3SS (2, 3, 12, 28). VPA1362 also shows weak sequence similarity to EspD from enteropathogenic E. coli (27% identity and 44% positive), which is one of the translocon proteins of the locus of enterocyte effacement-encoded T3SS (7, 9). By contrast, VPA1361 did not show any motif or sequence homology with known proteins, except for the conserved hypothetical protein NT01VC2344, which is encoded adjacent to VspD in the T3SS gene cluster of non-O1, non-O139 V. cholerae AM-19226 (6). In many cases, translocon-associated proteins are encoded in tandem in the T3SS region and are secreted abundantly in culture supernatants under suitable conditions. We designated the VPA1362 and VPA1361 proteins VopB2 and VopD2, respectively, and their possible role as translocons was examined in the experiments described below. The 27-kDa protein, which is secreted in a T3SS2-dependent manner, was also detected in the supernatants from the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains and was not recognized by anti-VopB2 or anti-VopD2 antibodies (data not shown). Therefore, we considered this protein a T3SS2-dependent secreted protein distinct from VopB2 and VopD2.

• Open in new tab
• Download powerpoint

FIG. 1.

Proteomic analysis of T3SS2-dependent secreted proteins. (A) SDS-PAGE of total secreted proteins from the POR-2 and POR-2 ΔvcrD2 strains. The gel was stained with Coomassie brilliant blue G-250. Protein bands that were excised and analyzed by N-terminal amino acid sequencing are indicated by arrows. The positions of molecular weight markers are indicated on the left. (B) Genomic organization of the VPA1362 (VopB2) and VPA1361 (VopD2) genes and sequence identities and similarities of the proteins with other putative T3SS translocon proteins.

VopB2 and VopD2 are specifically secreted via T3SS2. V. parahaemolyticus strain RIMD2210633 possesses sets of genes for two T3SS, both of which are able to secrete proteins but which recognize distinct proteins (23, 32). Western blot analysis was used to determine whether VopB2 and VopD2 are secreted only by T3SS2 or also by T3SS1. As shown in Fig. 2A, both proteins were specifically secreted by T3SS2. Both protein bands were absent when either a vopB2 or vopD2 deletion mutant (POR-2 ΔvopB2 or POR-2 ΔvopD2) was examined, but deletion of either gene did not influence the secretion of VopP, which is another protein secreted by T3SS2 (32), and deletion of one gene did not affect secretion of the product of the other gene. The production of the proteins in supernatants of strains POR-2 ΔvopB2 and POR-2 ΔvopD2 was fully restored by in trans complementation with the vopB2 and vopD2 genes, respectively (Fig. 2B). These results indicate that VopB2 and VopD2 seem to be T3SS2-specific substrates and are not required for T3SS2 secretion.

• Open in new tab
• Download powerpoint

FIG. 2.

VopB2 and VopD2 are specifically secreted via T3SS2. (A) Western blot analysis of bacterial pellets and supernatants of isogenic POR-1 strains. The lanes contained POR-1 (tdhAS deletion mutant) (lane 1), POR-2 (vcrD1 deletion mutant) (lane 2), POR-3 (vcrD2 deletion mutant) (lane 3), POR-2 ΔvopB2 (vcrD1 vopB2 mutant) (lane 4), and POR-2 ΔvopD2 (vcrD1 vopD2 mutant) (lane 5). Blots were probed with anti-VopB2 (αVopB2) (top), anti-VopD2 (αVopD2) (middle), and anti-VopP (αVopP) (bottom) antibodies. (B) Western blot analysis of supernatants of the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains expressing VopB2 or VopD2. Blots were probed with anti-VopB2 (top) and anti-VopD2 (bottom) antibodies.

VopB2 and VopD2 are necessary for T3SS2-dependent cytotoxicity.We have reported previously that one of the characteristic effects of T3SS2 is cytotoxicity to Caco-2 cells in vitro (21). Therefore, we tried to determine whether VopB2 and VopD2 are involved in this cytotoxicity. The cytotoxic activity of the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains in infected cells was dramatically decreased to a level similar to that in cells infected with the POR-2 ΔvcrD2 strain (T3SS1 and T3SS2 deficient) (Fig. 3). The cytotoxicity of the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains was fully restored by in trans complementation with the vopB2 and vopD2 genes, respectively. These results indicated that VopB2 and VopD2 are necessary for T3SS2-dependent cytotoxicity.

• Open in new tab
• Download powerpoint

FIG. 3.

VopB2 and VopD2 are necessary for T3SS2-dependent cytotoxicity. Caco-2 cells were infected for 6 h with V. parahaemolyticus mutant and complemented strains. Cytotoxicity was evaluated by determining the amount of LDH released. The error bars indicate standard deviations for the results of triplicate experiments. An asterisk indicates that the results are significantly different from the results obtained with the parent strain (P < 0.05).

VopB2 and VopD2 are necessary for T3SS2-dependent cell permeability.Some T3SSs can form pores in the plasma membranes of host eukaryotic cells or synthetic lipid bilayers. This process is presumably mediated by translocons that are essential for effector translocation (4). Therefore, the pore-forming activity of T3SS2 with Caco-2 cells was evaluated by examining cellular uptake of two fluorescent molecules, ethidium bromide (red) and acridine orange (green). Acridine orange penetrates cells having an intact membrane, whereas ethidium bromide selectively enters cells having perforated membranes (34). As shown in Fig. 4, POR-2, a strain which is deficient in both TDH and the T3SS1 apparatus genes, exhibited extensive pore-forming activity with the Caco-2 cells, whereas no pore-forming activity was seen with either POR-2 ΔvcrD2 or the noninfected control. Similar to infection of cells with POR-2 ΔvcrD2, infection of Caco-2 cells with the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains did not induce any detectable ethidium bromide uptake. The pore-forming activity of the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains was fully restored by in trans complementation with the vopB2 and vopD2 genes, respectively. Furthermore, the pore-forming activity of VopB2 and VopD2 was also confirmed by infection with T3SS2 effector deletion mutant strains, such as POR-2 ΔvopP, POR-2 ΔvopC, POR-2 ΔvopT, and POR-2 ΔvopL, indicating that none of the known T3SS2 effectors is involved in the pore-forming process. To exclude the possibility that this phenomenon was caused by membrane disruption due to cytotoxic activity of the strains, the levels of LDH in the supernatants were also determined. The low levels of LDH released from cells exposed to each of the V. parahaemolyticus strains indicated that no extensive cell disruption occurred during the period investigated (data not shown). Taken together, these results suggest that T3SS2 exhibits pore-forming activity in the host cell membrane and that both VopB2 and VopD2 are necessary for this activity.

• Open in new tab
• Download powerpoint

FIG. 4.

VopB2 and VopD2 are necessary for T3SS2-dependent cell permeability. Caco-2 cells were incubated with V. parahaemolyticus mutant strains for 2 h, and this was followed by exposure to acridine orange (green) and ethidium bromide (red).

VopB2 and VopD2 associated with the host cell membrane.The association of VopB2 and VopD2 with the host cell membrane was then determined. Caco-2 cells infected with V. parahaemolyticus mutant strains were mechanically disrupted, and the postnuclear supernatant was fractionated into cytosolic (host cytosol) and membrane (host membrane) fractions. Each of the fractions was then analyzed by Western blotting with anti-VopB2 and anti-VopD2 antibodies (Fig. 5). Both VopB2 and VopD2 were detected in the host cell membrane fraction when cells were infected with the POR-2, POR-2 ΔvopB2(pvopB2), and POR-2 ΔvopD2(pvopD2) strains (Fig. 5, lanes 7, 9, and 11). No signal was detected in the host cell membrane fraction during infection of cells with the POR-2 ΔvopB2, POR-2 ΔvopD2, and POR-2 ΔvcrD2 strains (Fig. 5, lanes 8, 10, and 12). VopD2 could be translocated into the cytosol of host cells, but the translocation of VopB2 into host cytosol could not be determined because a nonspecific band was detected in this fraction. The host cytosolic protein β-tubulin was absent from the host membrane fractions, and the endoplasmic reticulum integral membrane protein calnexin was absent from the host cytosol fractions. These results indicated that both VopB2 and VopD2 are at least associated with the host cell membrane and that translocation requires a functional T3SS2 and both VopB2 and VopD2.

• Open in new tab
• Download powerpoint

FIG. 5.

VopB2 and VopD2 localize to the host cell membrane. Caco-2 cells were infected with V. parahaemolyticus mutant strains for 3 h and subjected to subcellular fractionation by differential centrifugation. The fractions analyzed were the host cell cytosol (Host cytosol) and host cell membranes (Host membrane). The strains used were POR-2 (vcrD1 deletion mutant) (lanes 1 and 7), POR-2 ΔvopB2 (vcrD1 vopB2 mutant) (lanes 2 and 8), POR-2 ΔvopB2(pvopB2) (vcrD1 vopB2 strain expressing VopB2) (lanes 3 and 9), POR-2 ΔvopD2 (vcrD1 vopD2 mutant) (lanes 4 and 10), POR-2 ΔvopD2(pvopD2) (vcrD1 vopD2 strain expressing VopD2) (lanes 5 and 11), and POR-2 ΔvcrD2 (vcrD1 vcrD2 deletion mutant) (lanes 6 and 12). Blots were probed with anti-VopB2 (αVopB2), anti-VopD2 (αVopD2), antitubulin (αTubulin) (host cell cytosol marker), and anticalnexin (αCalnexin) (host cell membrane marker) antibodies.

VopB2 and VopD2 are necessary for translocation of T3SS2 effector protein.To determine whether vopB2 and vopD2 are necessary for the translocation of VopT by T3SS2, pSA-vopT-cyaA, a plasmid expressing VopT C-terminally fused with the catalytic domain of Bordetella adenylate cyclase toxin, was transformed into V. parahaemolyticus strains (21). Caco-2 cells were infected with the resulting transformants, and the level of intracellular cAMP was determined by a specific enzyme-linked immunosorbent assay (Fig. 6). The level of intracellular cAMP induced in infected cells by the POR-2 strains expressing VopT-cyaA was higher than the detection limit of the assay, whereas the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains induced low levels of cAMP production similar to the level induced by POR-2 ΔvcrD2, suggesting that vopB2 and vopD2 are necessary for translocation of the T3SS2 effector protein.

• Open in new tab
• Download powerpoint

FIG. 6.

VopB2 and VopD2 are necessary for VopT translocation. Caco-2 cells were infected with V. parahaemolyticus mutant strains expressing VopT C-terminally fused with the catalytic domain of adenylate cyclase toxin, and the intracellular cAMP concentrations in infected cells were determined by an enzyme-linked immunosorbent assay. The error bars indicate the standard deviations for the results of triplicate experiments. An asterisk indicates that the results are significantly different from the results obtained with the parental strain (P < 0.01).

VopB2 and VopD2 are necessary for T3SS2-dependent enterotoxicity.To investigate the contribution of vopB2 and vopD2 to the T3SS2-dependent enterotoxicity of V. parahaemolyticus, we examined the enterotoxic activity of vopB2 and vopD2 mutant strains in rabbit ileal loops. The POR-2 strain caused a high level of fluid accumulation (Fig. 7), in accordance with a previous report (32). The POR-2 ΔvopB2 and POR-2 ΔvopD2 strains induced little fluid accumulation, and the levels were similar to the levels induced by the POR-2 ΔvcrD2 strain and to noninfected control levels (negative control). The enterotoxicity of the POR-2 ΔvopB2 and POR-2 ΔvopD2 strains was fully restored by in trans complementation with the vopB2 and vopD2 genes, respectively. These results strongly indicate that vopB2 and vopD2 have a critical role in T3SS2-dependent enterotoxicity.

• Open in new tab
• Download powerpoint

FIG. 7.

VopB2 and VopD2 are necessary for T3SS2-dependent enterotoxicity. The enterotoxic activities of V. parahaemolyticus mutant and complemented strains in rabbit ileal loops were examined. Fluid accumulation is the amount of fluid accumulated (in milliliters) per centimeter of ligated rabbit small intestine. The error bars indicate the standard deviations for the results of triplicate experiments. An asterisk indicates that the results are significantly different from the results obtained with the parental strain (P < 0.05).

In this study, we identified VopB2 (VPA1362) and VopD2 (VPA1361) as T3SS2-dependent secreted proteins. VopB2 was similar to the predicted translocon protein AopB from A. salmonicida subsp. salmonicida A449 and was weakly similar to EspD from enteropathogenic E. coli. EspD is well known as an locus of enterocyte effacement-encoded translocon protein (7, 9). Accordingly, it was expected that VopB2 would be a T3SS2 translocon-associated protein. By contrast, VopD2 did not show any motif or sequence homologies with any known protein except NT01VC2344, a conserved hypothetical protein of non-O1, non-O139 V. cholerae (6). Because VopB2 and VopD2 were secreted abundantly via T3SS2 under the conditions used and were encoded in tandem in the T3SS2 region, we tried to examine their possible roles in the translocon of T3SS2.

Several pieces of evidence supported the hypothesis that the two proteins that were identified in this study (VopB2 and VopD2) were functional as T3SS2 translocon-associated proteins. Both proteins were found to be (i) secreted specifically by the T3SS2 (Fig. 2); (ii) not necessary for T3SS2 effector secretion but necessary for effector translocation (Fig. 2A and 6); (iii) required for pore formation in host cell membranes (Fig. 4); (iv) localized to the host cell membrane (Fig. 5); and (v) essential for T3SS2-dependent biological activities, such as cytotoxicity in Caco-2 cells and enterotoxicity in a rabbit ileal loop model (Fig. 3 and 7).

The predicted structural features of VopB2 and VopD2 also support our hypothesis. Most translocon proteins reported so far contain one or two predicted transmembrane domains (4). The TM-PRED program (http://www.ch.embnet.org/cgi-bin/TMPRED form parser; 17 to 33 residues; scores, >1,000) predicts that VopB2 and VopD2 contain two transmembrane regions and one transmembrane region, respectively (in VopB2, amino acids 119 to 142 and 175 to 196; in VopD2, amino acids 99 to 118), supporting the idea that these proteins may function as translocons. Taken together, these findings strongly suggest that both VopB2 and VopD2 are functional T3SS2 translocon-associated proteins and are molecules that play a key role in the T3SS2-dependent enterotoxicity of V. parahaemolyticus.

The non-O1, non-O139 V. cholerae AM-19226 strain, which can cause gastroenteritis in humans, also carries functional T3SS genes, which are related to the T3SS2 genes of V. parahaemolyticus. The T3SS of AM-19226 is known to contribute to its virulence (6). The T3SS gene set of this strain contains a pair of genes that are highly similar to the VopB2 and VopD2 genes (Fig. 1B). In this regard, it is possible that not only VopB2 and VopD2 from V. parahaemolyticus but also the gene products from V. cholerae may be functional translocons of T3SS and may contribute to the virulence of these bacteria.

There are many things that are unclear regarding the detailed mechanism of action of T3SS2 of V. parahaemolyticus and the T3SS2-related T3SS gene cluster in non-O1, non-O139 V. cholerae, especially because there are a number of hypothetical genes in this region. Identification and functional analysis of proteins secreted by T3SS2 should lead to a better understanding of the pathogenic mechanisms not only of V. parahaemolyticus but also of non-O1, non-O139 V. cholerae. The results presented here for VopB2 and VopD2 could also increase our understanding of the pathogenic mechanism of non-O1, non-O139 V. cholerae.