Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Dec 5.
Published in final edited form as: Cell Microbiol. 2014 Jan 8;16(6):938–947. doi: 10.1111/cmi.12252

The Vibrio parahaemolyticus effector VopC mediates Cdc42 dependent invasion of cultured cells but is not required for pathogenicity in an animal model of infection

Ryu Okada 1, Xiaohui Zhou 2, Hirotaka Hiyoshi 1, Shigeaki Matsuda 1, Xiang Chen 2, Yukihiro Akeda 3, Takashige Kashimoto 4, Brigid M Davis 2, Tetsuya Iida 1, Matthew K Waldor 2,✉, Toshio Kodama 5,✉
PMCID: PMC4670550  NIHMSID: NIHMS733594  PMID: 24345190

Summary

Vibrio parahaemolyticus is a Gram-negative marine bacterium that causes acute gastroenteritis in humans. The virulence of V. parahaemolyticus is dependent upon a type III secretion system (T3SS2). One effector for T3SS2, VopC, is a homolog of the catalytic domain of cytotoxic necrotizing factor (CNF), and was recently reported to be a Rho family GTPase activator and to be linked to internalization of V. parahaemolyticus by nonphagocytic cultured cells. Here, we provide direct evidence that VopC deamidates Rac1 and CDC42, but not RhoA, in vivo. Our results also suggest that VopC, through its activation of Rac1, contributes to formation of actin stress fibers in infected cells. Invasion of host cells, which occurs at a low frequency, does not seem linked to Rac1 activation, but instead appears to require CDC42. Finally, using an infant rabbit model of V. parahaemolyticus infection, we show that the virulence of V. parahaemolyticus is not dependent upon VopC-mediated invasion. Genetic inactivation of VopC did not impair intestinal colonization nor reduce signs of disease, including fluid accumulation, diarrhea, and tissue destruction. Thus, although VopC can promote host cell invasion, such internalization is not a critical step of the disease process, consistent with the traditional view of V. parahaemolyticus as an extracellular pathogen.

Introduction

Vibrio parahaemolyticus is a Gram-negative marine bacterium that causes acute gastroenteritis in humans. Intestinal tissue from infected humans, as well as from animal models of V. parahaemolyticus infection, displays marked inflammation and disruption of the intestinal epithelium (Ritchie et al., 2012). Infection is also associated with accumulation of intestinal fluid in infant rabbits and ligated ileal loops from adult rabbits (Hiyoshi et al., 2010, Hiyoshi et al., 2011, Ritchie et al., 2012). Both epidemiological studies and animal-based studies indicate that a key factor in V. parahaemolyticus virulence is one of its two type III secretion systems, termed T3SS2. To date, six effectors for this secretion apparatus have been identified, although only two have been found to be essential for either intestinal colonization or intestinal pathology (Hiyoshi et al., 2011, Zhang et al., 2013, Zhou et al., 2013). A second type III secretion system, termed T3SS1, has a pronounced effect of the cytotoxicity of V. parahaemolyticus in vitro; however, analyses to date suggest that it plays a relatively minor role in intestinal disease (Park et al., 2004, Hiyoshi et al., 2010).

Tissue culture analyses have revealed that V. parahaemolyticus induces T3SS2-dependent cytoskeletal changes in infected eukaryotic cells, including extensive formation of actin stress fibers and formation of focal actin clusters underneath bacterial colonies (Liverman et al., 2007, Hiyoshi et al., 2011). Actin rearrangement has also been observed in infected intestinal tissue, and may be related to loss of the intestinal permeability barrier (Zhou 2013, Ritchie 2012). The pathways underlying such rearrangements have not been fully characterized; however, at least 3 T3SS2 effectors appear to contribute. VopL, which contains 3 WH2 (actin-binding) domains, induces formation of actin filaments following translocation or transfection into host cells, as well as in conjunction with purified actin (Liverman et al., 2007). VopV, which also binds actin directly, is linked to formation of actin foci underneath bacterial microcolonies in tissue culture assays. VopV is required for fluid accumulation and pathology in the ileal loop loop assay; in contrast, the absence of vopL does not impair V. parahaemolyticus virulence (Hiyoshi et al., 2011). Recently, an additional effector, VopC, which is homologous to the catalytic domain of cytotoxic necrotizing factor (CNF), was linked to formation of actin ruffles and filipodia (Zhang et al., 2012). Unlike VopL and VopV, VopC does not bind directly to actin; instead, it targets members of the Rho family of small GTPases, which have long been recognized to regulate actin-based cytoskeletal structures (Ridley et al., 1992). VopC deaminates (and thereby activates) Rac and CDC42, but in contrast to CNF, it lacks activity upon RhoA (Zhang et al., 2012).

VopC was also reported to promote internalization of V. parahaemolyticus by host cells in vitro, although the precise mechanism of internalization was not characterized (Zhang et al., 2012). Host cell invasion by V. parahaemolyticus has occasionally been observed in vivo (e.g., (Boutin et al., 1979, Chattarjee et al., 1982), but this organism is generally considered to be a non-invasive pathogen. VopC’s contribution to intestinal colonization and intestinal pathology, and the importance of VopC-mediated invasion in vivo, have not been reported. However, CNF and other proteins that modulate the Rho GTPase cycle are thought to contribute to the virulence of a variety of other bacterial pathogens (Doye et al., 2002, Munro et al., 2004).

In this study, we have further characterized the relationships between VopC-mediated GTPase activation, cytoskeletal rearrangement, and invasion. We provide direct evidence that Rac1 and CDC42 deamidation occurs in infected cells, and that it correlates with GTPase activation. We also demonstrate that V. parahaemolyticus infection of cultured cells results in VopC-dependent formation of actin stress fibers, which appears linked to activation of Rac1 but not CDC42. In contrast, genetic analyses suggest that activation of CDC42 plays a more central role in VopC-mediated invasion of eukaryotic cells. Finally, we show that inactivation of VopC does not reduce V. parahaemolyticus colonization or virulence in the infant rabbit model of infection. Thus, VopC-dependent invasion of host cells does not appear to make a significant contribution to the pathogenicity of V. parahaemolyticus.

Results

VopC deamidates Rac1 and CDC42 Q61, but not RhoA Q63, in infected cells

A previous analysis of VopC activity revealed that this enzyme could deamidate and transglutaminate Rac1 and CDC42 in vitro (both of which have been linked to GTPase activation); however, it did not directly assess the affect of VopC upon all these GTPases in eukaryotic cells (Zheng et al., 2012). To monitor host GTPase activation directly, we used an antibody against glutamate 63/61 in the switch II region of Rho GTPases (anti-Q63E), which recognizes proteins containing deamidated glutamine 63/61 (Horiguchi et al., 1997). We first confirmed that the antibody recognizes proteins deamidated in vitro by either recombinant CNF1 (which targets RhoA, Rac1, and CDC42) or VopC (which targets Rac1 and CDC42 but not RhoA), and that deamidated GTPases were not produced by a non-catalytic VopC mutant, VopCC220S (Fig. S1A and S1B).

Subsequently, we monitored deamidation of ectopically expressed FLAG-tagged GTPases in V. parahaemolyticus-infected HeLa cells. Bacterial strains used for infection were all derived from a hemolysin and T3SS1-deficient strain, POR-2, in order to avoid the cytotoxicity associated with T3SS1 (Matsuda et al., 2012). Notably, deamidated Rac1 and CDC42 were apparent in immunoprecipitated (IP; anti-FLAG) protein isolated from cells infected with POR-2, but not with POR-2 vopC or the T3SS2-deficient strain POR-2 vcrD2 (Fig. 1A and 1B). Rac1 and CDC42 deamidase activity was restored to the POR-2 vopC mutant by expression of wt VopC (POR-2ΔvopC/vopCWT), but not by catalytically inactive protein (POR-2ΔvopC/vopCC220S) (Fig. 1A and 1B). Deamidated RhoA was not detected in the IP fraction of POR-2-infected cells, consistent with VopC’s lack of activity against RhoA in assays using purified protein (Fig. 1C and S1A). Collectively, these experiments validate the conclusions presented in Zheng et al., and provide direct evidence that VopC-mediated selective deamidation of Rho family GTPases occurs in V. parahaemolyticus-infected cells.

Fig. 1.

Fig. 1

Open in a new tab

In infected cells, VopC deamidates Rac1 and Cdc42 Q61, but not RhoA Q63. Transfected HeLa cells that expressed FLAG-tagged Rho GTPases (panel A for Rac1, panel B for Cdc42, and panel C for RhoA) were infected with V. parahaemolyticus POR-2 or isogenic derivatives for 3 h. FLAG-tagged GTPases were immunoprecipitated (IP) from cell lysates, separated by SDS-PAGE, and probed with anti-FLAG and anti-Q63E antibodies. Whole-cell lysates (input) were analyzed in parallel.

Studies of CNF have demonstrated that deamidation of Rho proteins destroys their GTPase activity, causing them to be constitutively active (Schmidt et al., 1997). Such activation increases their GTP-dependent binding to downstream targets (e.g., P21-activated kinase 3), which can be detected with co-immunoprecipitation assays. As previously reported (Zhang et al., 2012), we observed that infection of cultured cells with V. parahemolyticus resulted in activation of Rac1 and CDC42, and that this activation was dependent on production of catalytically active VopC as well as on T3SS2 (Fig. S2A). Activation also occurred in response to transfection of the C terminus of catalytically active VopC fused to GFP, but not GFP alone, although activation was less marked in the absence of V. parahaemolyticus infection (Fig. S2B). Under the conditions assayed, neither infection nor VopC tranfection appeared to alter the total level of either Rac1 or CDC42. Thus, although activation has been linked to GTPase degradation (Doye et al., 2002), such degradation is not apparent under the conditions tested here.

VopC modulates formation of T3SS2-dependent actin stress fibers

We noticed that cells infected with V. parahaemolyticus also displayed marked changes in shape and in cytoskeletal structures, and that these changes were in part dependent upon VopC. Caco-2 cells infected with POR-2 contained pronounced actin stress fibers that spanned the cell, rather than the diffusely distributed actin near the cell periphery that was observed in uninfected cells (Fig. 2). Stress fibers were anchored at vinculin foci (as in focal adhesions (FA)), and vinculin also redistributed in response to infection: nuclear staining disappeared, and foci became more numerous and more widely distributed through the cell body. These cytoskeletal changes were largely dependent upon T3SS2, and not apparent in cells infected with POR-2 vcrD2 (Fig. 2). The absence of catalytically active VopC did not eliminate rearrangement of actin in response to infection, consistent with V. parahaemolyticus’ possession of additional T3SS2 effectors that modulate the actin cytoskeleton (e.g., VopV and VopL); however, it did markedly alter the cytoskeletal changes associated with infection. Cells infected with POR-2ΔvopC contained long, branched and curved F-actin filaments with a reticular appearance, rather than actin stress fibers. Infection with POR-2 ΔvopC also did not induce redistribution of vinculin foci throughout the cell; instead, vinculin foci were largely peripheral and less intense. Ectopic expression of wt VopC in the POR-2 vopC mutant (Fig. 2; POR-2ΔvopC/vopCWT) restored the strain’s capacity to induce formation of normal stress fibers, while expression of catalytically inactive VopC did not (Fig. 2; POR-2ΔvopC/vopCC220S), thereby confirming the importance of VopC’s deamidase activity in control of actin localization.

Fig. 2.

Fig. 2

Open in a new tab

VopC contributes to T3SS2-dependent formation of actin stress fibers. Caco-2 cells were infected with V. parahaemolyticus isogenic mutant strains for 3 h. F-actin (green), vinculin (red), and cellular and bacterial DNAs (blue) were detected with Alexa Fluor 488-phalloidin, hVIN-1 antibody, Hoechst 33258 respectively. Bar = 20 μm.

Constitutively active Rac1, but not Cdc42, rescues stress fiber formation in POR-2ΔvopC-infected cells

To better define the pathway by which VopC modulates actin dynamics, we investigated the consequences of selectively activating its GTPase targets. Caco-2 cells were transfected with constitutively active (CA) forms of FLAG-tagged Rac1 or Cdc42, then infected with the POR-2ΔvopC strain and examined using immunofluorescence microscopy. Actin stress fibers were evident in ~ 85% of Caco-2 cells transfected with Rac1 prior to infection with VopC-deficient bacteria, a frequency comparable to that seen with untransfected cells infected with POR-2 (81%; Fig. 3A and 3B). In contrast, the frequency of stress fiber formation in CDC42-transfected cells infected with VopC-deficient bacteria was comparable to that seen with untransfected cells and with uninfected Caco-2 cells (Fig. 3A and 3B). Thus, activation of Rac1, rather than CDC42, appears to be the critical factor in VopC-mediated formation of actin stress fibers in response to V. parahaemolyticus infection. Notably, however, Rac1 activation was not sufficient to induce formation of stress fibers in Caco-2 cells in the absence of a bacterial infection (Fig. 3A and 3B), consistent with the idea that VopC modulates actin dynamics in conjunction with other bacterial factors. Comparable results to those in Fig. 3 were obtained using HeLa cells (Fig. S3, providing further confirmation of the importance of Rac1 for cytoskeletal regulation by VopC.

Fig. 3.

Fig. 3

Open in a new tab

Ectopic expression of constitutively active (CA) Rac1, but not Cdc42, enables actin stress fiber formation in POR-2ΔvopC-infected cells. A. Transfected (or mock transfected) Caco-2 cells expressing FLAG-tagged CA Rac1 or Cdc42 were infected with POR-2 or POR-2ΔvopC. After 3 h, F-actin (green), Rho GTPases (red), and cellular and bacterial DNAs (blue) were visualized with Alexa Fluor 488-palloidin, anti-FLAG antibody, and Hoechst 33258, respectively. Asterisks mark cells expressing FLAG-tagged CA forms of each GTPase. Bar = 20 μm. B. Percentage of cells with actin stress fibers among the cells expressing CA GTPase, uninfected or infected with POR-2ΔvopC. Data for untransfected cells +/− POR-2 or POR-2ΔvopC is shown for comparison. Over 40 cells expressing CA GTPase (FLAG-positive cells) were analyzed per experiment, and the results show means ± standard deviation (SD) from experiments of double blind in triplicate. The asterisks indicate statistically significant differences (P < 0.01).

Cdc42 appears critical for V. parahaemolyticus invasion

It has been reported that V. parahaemolyticus strains invade nonphagocytic eukaryotic cells and that VopC’s catalytic activity promotes the invasive phenotype (Akeda et al., 1997, Zhang et al., 2012). We confirmed that a low percentage of V. parahaemolyticus penetrate into HeLa cells using a gentamicin killing assay. V. parahaemolyticus invasion was T3SS2-dependent, and it occurred with greater frequency when bacteria were pre-cultured in the presence of crude bile, which stimulates T3SS2 gene expression (Figure S4A; (Gotoh et al., 2010). Invasion was reduced, but not entirely abolished, for the vopC deletion mutant (Fig. S4B). Notably, altering the relative levels of Rho family GTPases in host cells also significantly altered the frequency of V. parahameolyticus invasion. In particular, reduction of CDC42 or of the downstream factor Arp3 (via siRNA-mediated knockdown; Fig. 4A and 4B) diminished invasion. In contrast, reduction of RhoA levels significantly promoted invasion, while modulation of Rac1 had no effect. Cell-permeable C3 transferase, which inhibits all Rho isoforms (RhoA, B, and C), also had no effect on V. parahaemolyticus invasion (Fig. S4C). Collectively, these findings suggest that VopC- and Rac1-mediated formation of actin stress fibers is unlikely to contribute to bacterial invasion, while activation of CDC42 appears to be important for V. parahaemolyticus internalization by host cells. Additional studies to further delineate the steps of invasion are clearly warranted.

Fig. 4.

Fig. 4

Open in a new tab

Cdc42, but not Rac1 nor RhoA, is necessary for V. parahaemolyticus invasion. A. Knock-down of RhoA, Rac1, Cdc42, and Arp3 by siRNAs. HeLa cells were transfected with RhoA, Rac1, Cdc42, or Arp3 siRNA. The lysates from transfected cells were separated by SDS-PAGE and probed with anti-RhoA, anti-Rac1, anti-Cdc42, anti-Arp3, or anti-Actin antibodies. B. Knockdown of Cdc42 attenuates V. parahaemolyticus invasion. Cells transfected with siRNAs against RhoA, Rac1, Cdc42, or Arp3 were infected with V. parahaemolyticus strain POR-2. Frequencies of invasion are expressed relative to invasion of the control siRNA-transfected cells, which was set at 1.0. The data represent mean ± SD for three different experiments. The asterisks indicate results that differed significantly from the results obtained with the control cells (*, P < 0.01; **, P < 0.05). NS indicates no significant difference compared with the control cells (P > 0.05).

VopC catalytic activity does not contribute to V. parahaemolyticus pathogenicity in infant rabbits

Although internalization of V. parahaemolyticus by eukaryotic cells has occasionally been detected, the relationship between host cell invasion and pathogenesis has not been directly explored. To evaluate whether VopC-mediated invasion contributes to disease, we compared intestinal colonization and induction of intestinal fluid accumulation in infant rabbits in response to wt V. parahaemolyticus (initially reported in (Ritchie et al., 2012)), and a mutant that produces catalytically inactive VopC (VopC C220S). Strikingly, survival and replication of the mutant strain within the small intestines of infected rabbits was equivalent to that of wt V. parahaemolyticus, based on enumeration of colony forming units present ~38 hr post infection (Fig. 5A). Furthermore, there was no significant difference between the amount of intestinal fluid accumulation (a precursor to diarrhea) in response to the wt and VopC C220S strains (Fig. 5B). It is important to note that the wild-type control data came from previously published experiments, which represents a potential source of error. In our prior studies with wt V. parahaemolyticus carried out in 10 animals, the mean and standard error were .28 and .04 respectively. In the current experiments with the VopC C220S mutant strain, all rabbits had fluid accumulation ratios (mean of 0.24 +/− .02) within the range obtained with the historical wt control. A related analysis, performed using ligated ileal loops from adult rabbits, also suggested that VopC does not influence intestinal fluid accumulation, at least under these conditions, when V. parahaemolyticus is forcibly retained within the ligated intestinal loops (Hiyoshi et al., 2011). Finally, as in rabbits infected with the wt strain (Ritchie et al., 2012), VopC C220S-infected rabbits exhibited marked disruption of the intestinal epithelium and extensive epithelial sloughing (Fig. S5). Collectively, these data demonstrate that VopC-mediated invasion of host cells is not required for or a significant factor in the pathogenicity of V. parahaemolyticus, and affirm the long standing view that this organism is an extracellular pathogen.

Fig 5.

Fig 5

Open in a new tab

Inactivation of the VopC active site does not attenuate intestinal colonization or virulence of V. parahaemolyticus in infant rabbits. Wild type or VopC C220S V. parahaemolyticus were intragastrically inoculated in infant rabbits and the number of V. parahaemolyticus colony forming units (cfu/g) (A) and fluid accumulation in the small intestine (B) were determined ~38 hr post infection. The lines in A show geometric means; means and SEM are shown in B. The data for the wt strain were previously published in (Ritchie et al., 2012). There is no significant difference between colonization or fluid accumulation for the wt and VopC(C220S) strains.

Discussion

VopC, an effector of V. parahaemolyticus T3SS2, deamidates the Rho family GTPases Rac1 and CDC42 and thereby activates them. In cultured cells infected with V. parahaemolyticus, our analyses suggest that VopC promotes formation of actin stress fibers that are anchored in vinculin foci via deamidation and activation of Rac1. VopC also promotes invasion of V. parahaemolyticus into eukaryotic cells, which appears to be independent of Rac1 and dependent on CDC42. In vitro, only a small proportion (~1%) of infecting bacteria invade, and the clinical significance of this observation was unknown. Here, we demonstrate that VopC’s deamidase activity is not required for V. parahaemolyticus virulence: genetic inactivation of the enzyme did not reduce colonization of the intestines of infant rabbits, nor did it limit intestinal fluid accumulation or intestinal pathology in response to the infection. These data suggest that internalization of V. parahaemolyticus by host cells is not a critical step in disease pathology and support the longstanding paradigm that V. parahaemolyticus is predominantly an extracellular pathogen.

Our results, as well as previous studies, show that V. parahaemolyticus has a profound effect upon the cytoskeleton of infected host cells. In particular, infection induces marked rearrangement of actin, including formation of stress fibers. At least three T3SS2 effectors influence actin localization: VopL, VopV, and VopC. Both VopL and VopC are associated with formation of actin stress fibers, although only VopL interacts directly with actin. VopC appears to promote stress fiber assembly via activation of Rac1, as constitutive activation of Rac1 compensates for the absence of VopC in formation of actin fibers in infected HeLa and Caco-2 cells. In the absence of VopC, V. parahaemolyticus induces T3SS2-dependent formation of curved and branched actin fibers that are not reliably anchored in vinculin foci. Although Rac1 is known to modulate actin polymerization in several ways (Heasman et al., 2008), the precise pathway by which VopC modulates actin assembly remains to be defined. It should also be noted that VopC might have cell-type specific effects upon the cytoskeleton, as seen for CNF1, which has been linked to formation of stress fibers, membrane ruffles, and filapodia, depending on which cell line is intoxicated (Lemonnier et al., 2007). The formation of stress fibers observed in our study likely ensues from the combined effects of a variety of bacterial factors, including but not necessarily limited to VopC and VopL. V. parahaemolyticus’ diverse mechanisms for altering the host cytoskeleton likely contribute to the extensive disruption of intestinal tissue associated with V. parahaemolyticus infection.

VopC also promotes internalization of V. parahaemolyticus by eukaryotic cells, as has been observed for the related proteins CNF1 and DNT, which also modify and activate Rho family GTPases (Aktories et al., 2005). Our analyses suggest that VopC promotes internalization via its deamidation of CDC42, as we observed that knockdown of CDC42 levels significantly reduced invasion of HeLa cells; however, we have not investigated the extent to which silencing of a single GTPase alters the activity of related factors. It has been hypothesized that VopC-dependent invasion might play an important role in the virulence of V. parahaemolyticus (Zhang et al., 2012), but we found that inactivation of VopC does not reduce either colonization or gross indicators of intestinal infection (e.g., diarrhea, fluid accumulation, and tissue disruption) in an infant rabbit model of V. parahaemolyticus infection. Our data suggest that bacterial internalization by non-phagocytic epithelial cells is not a critical step in the pathogenic process. Nonetheless, given the importance of Rho family GTPases in regulation of numerous cellular processes, including NF-kB and MAP kinase signal transduction, cell cycle progression and apoptosis (Lemonnier et al., 2007), it is possible that VopC does have a subtle effect upon the V. parahaemolyticus infection process. The importance of GTPase-mediated signaling during infection is underscored by the abundance of bacterial factors that have been found to modulate it (Boquet et al., 2003, Aktories et al., 2005).

It is interesting that the known effects of Rho family GTPase activation include activation of NF-KB and the MAP kinases JNK and p38 (Lemonnier et al., 2007), in that we recently reported that another T3SS2 effector, VopZ, downregulates all three pathways (Zhou et al., 2013). Similarly, there is some conflict between the VopC’s (indirect) effects on actin and the effects of VopV, which appears to promote formation of a distinct set of actin structures in host cells. It seems likely that V. parahaemolyticus’ use of multiple effectors targeting overlapping pathways allows this pathogen to finely tune the host environment and maximize its ability to survive in and exploit distinct intestinal niches.

Experimental procedures

Bacterial strains and plasmids

V. parahaemolyticus strain RIMD2210633 (KP-positive, serotype O3:K6) was obtained from the Pathogenic Microbes Repository Unit, International Research Center for Infectious Diseases, Research Institute for Microbial Diseases (Osaka University, Osaka, Japan). The bacterial strains and plasmids used in this study are described in the Supporting Information (Table S1). VopC C220S was created on the wild-type background by allele exchange using suicide vector pDM4 as previously described (Zhou et al., 2008). The bacteria were grown in Luria Bertani (LB) broth containing 0.5% NaCl in a shaking incubator at 37 °C. The culture medium was supplemented with chloramphenicol (12.5 μg ml−1), as appropriate.

Immunofluorescence microscopy

Caco-2 cells were transfected with vectors that expressed constitutively active (CA) forms of FLAG-tagged GTPases (RacV12 or Cdc42V12) using Lipofectamine LTX (Invitrogen) 24 h before infection. The cells were infected with V. parahaemolyticus for 3 h at a multiplicity of infection (MOI) of 10. After infection, the cells were washed three times with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 15 min. Actin was detected with Alexa Fluor® 488-phalloidin. Cellular and bacterial DNAs were stained with Hoechst 33258. FLAG-tagged GTPases were detected with an anti-FLAG M2 monoclonal antibody (Sigma). Vinculin was stained with monoclonal anti-human vinculin clone hVIN-1 antibody (Sigma).

Detection of deamidated Rho GTPases in infected cells

HeLa cells were transfected with vectors that expressed amino-terminal FLAG-tagged Rho GTPases using Lipofectamine LTX (Invitrogen) 24 h before infection. The cells were infected with V. parahaemolyticus for 3 h at an MOI of 10. After three washes with PBS, the cells were lysed with RIPA buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.1% SDS, 0.5% deoxycholate, 1% NP-40), which contained a protease inhibitor cocktail (Sigma). The lysates were centrifuged at 16,000 g for 15 min at 4 °C. Anti-FLAG M2 Affinity Gel (Sigma) was added to the supernatants, which were then incubated overnight at 4 °C. The gels were washed twice with Tris-buffered saline, and resuspended in 0.1 M glycine (pH 3.5) to elute the FLAG-tagged GTPases. The eluates were neutralized by the addition of 1 M Tris-HCl (pH 7.5) and mixed with SDS loading buffer. The samples were separated by SDS-PAGE, and transferred to membranes (Millipore, Bedford, MA), which were then probed with anti-FLAG M2 monoclonal antibody (Sigma), and anti-Q63E antibody (courtesy of Dr. Horiguchi).

siRNA knockdown

FlexiTube siRNAs (Qiagen) were used to knockdown RhoA, Rac1, Cdc42, and Arp3. HeLa cells were transfected with the siRNAs using HiPerFect Transfection Reagent (Qiagen). At 72 h post-transfection, the cells were infected with POR-2 to evaluate the invasion efficiency. The knockdown efficiency was determined by immunoblot analysis using anti-RhoA, anti-Rac1, anti-Cdc42, and anti-Arp3 antibodies.

Gentamicin killing assay

The invasion assay was performed according to a published method, with a slight modification. Briefly, bacteria were grown in LB medium in the presence or absence of 0.04% crude bile for 3 h and used to infect HeLa cells at a MOI of 2. At 2 h after infection, the cells were incubated for 1 h in DMEM containing 10μg ml−1 gentamicin. The cells were washed twice with PBS and lysed with PBS containing 0.1% Triton X-100. The lysates were plated onto LB agar to determine the number of intracellular bacteria. The results are expressed as the percentages of internalized bacteria, which were calculated based on the mean number of colony-forming units of bacterial isolates internalized in HeLa cells relative to the number of bacteria in the original inoculum.

Statistical analysis

All of the data are expressed as the mean and standard deviation for three independent replicates per set of experimental conditions. The statistical significance of differences was determined using the Student’s t-test, and P < 0.05 was considered statistically significant.

Infant rabbit studies

Rantidine-treated infant rabbits were orogastrically inoculated with ~109 cfu of the VopC C220S V. parahaemolyticus mutant and bacterial colonization (cfu/g inteinstinal tissue) and fluid accumulation ratios at ~38 hr postinfection were determined as described in (Ritchie et al., 2012) and (Zhou et al., 2013) respectively. Tissue from the distal small intestines of rabbits infected with VopC C220S V. parahaemolyticus strain for 38 hr was stained with H&E. Animal experiments were performed in accordance with a protocol approved by the Harvard IACUC.

Supplementary Material

Supp FigureS1-S5
Supp TableS1

Acknowledgments

We thank Dr. Horiguchi for providing the anti-Q63E antibody. The V. parahaemolyticus strains were gifted by the National Bio Resource Project (NIG, Japan): Pathogenic Microbes. This work was supported by a Grant-in-Aid for Young Scientists (B) (23790474) and by a Grant-in-Aid for Scientific Research (C) (25460531) from the Japan Society for the Promotion of Science (JSPS). This work was also supported by a Grant-in-Aid from the Institute for Fermentation, Osaka (IFO), from the NAITO Foundation, and from the Takeda Science Foundation, Japan. The Waldor lab acknowledges support from R37 AI-42347, HHMI, and a fellowship from NERCE (U54 AI 057159) to Dr. Zhou.

References

  1. Akeda Y, Nagayama K, Yamamoto K, Honda T. Invasive phenotype of Vibrio parahaemolyticus. The Journal of infectious diseases. 1997;176:822–824. doi: 10.1086/517312. [DOI] [PubMed] [Google Scholar]
  2. Aktories K, Barbieri JT. Bacterial cytotoxins: targeting eukaryotic switches. Nature reviews Microbiology. 2005;3:397–410. doi: 10.1038/nrmicro1150. [DOI] [PubMed] [Google Scholar]
  3. Boquet P, Lemichez E. Bacterial virulence factors targeting Rho GTPases: parasitism or symbiosis? Trends in cell biology. 2003;13:238–246. doi: 10.1016/s0962-8924(03)00037-0. [DOI] [PubMed] [Google Scholar]
  4. Boutin BK, Townsend SF, Scarpino PV, Twedt RM. Demonstration of invasiveness of Vibrio parahaemolyticus in adult rabbits by immunofluorescence. Applied and environmental microbiology. 1979;37:647–653. doi: 10.1128/aem.37.3.647-653.1979. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chattarjee BD, Mukherjee A, Sanyal SN. Rabbit ileal loop invasion of Vibrio parahaemolyticus. Indian journal of pathology & microbiology. 1982;25:213–218. [PubMed] [Google Scholar]
  6. Doye A, Mettouchi A, Bossis G, Clement R, Buisson-Touati C, Flatau G, et al. CNF1 exploits the ubiquitin-proteasome machinery to restrict Rho GTPase activation for bacterial host cell invasion. Cell. 2002;111:553–564. doi: 10.1016/s0092-8674(02)01132-7. [DOI] [PubMed] [Google Scholar]
  7. Gotoh K, Kodama T, Hiyoshi H, Izutsu K, Park KS, Dryselius R, et al. Bile acid-induced virulence gene expression of Vibrio parahaemolyticus reveals a novel therapeutic potential for bile acid sequestrants. PloS one. 2010;5:e13365. doi: 10.1371/journal.pone.0013365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Heasman SJ, Ridley AJ. Mammalian Rho GTPases: new insights into their functions from in vivo studies. Nature reviews Molecular cell biology. 2008;9:690–701. doi: 10.1038/nrm2476. [DOI] [PubMed] [Google Scholar]
  9. Hiyoshi H, Kodama T, Iida T, Honda T. Contribution of Vibrio parahaemolyticus virulence factors to cytotoxicity, enterotoxicity, and lethality in mice. Infection and immunity. 2010;78:1772–1780. doi: 10.1128/IAI.01051-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hiyoshi H, Kodama T, Saito K, Gotoh K, Matsuda S, Akeda Y, et al. VopV, an F-actin-binding type III secretion effector, is required for Vibrio parahaemolyticus-induced enterotoxicity. Cell host & microbe. 2011;10:401–409. doi: 10.1016/j.chom.2011.08.014. [DOI] [PubMed] [Google Scholar]
  11. Horiguchi Y, Inoue N, Masuda M, Kashimoto T, Katahira J, Sugimoto N, Matsuda M. Bordetella bronchiseptica dermonecrotizing toxin induces reorganization of actin stress fibers through deamidation of Gln-63 of the GTP-binding protein Rho. Proceedings of the National Academy of Sciences of the United States of America. 1997;94:11623–11626. doi: 10.1073/pnas.94.21.11623. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Lemonnier M, Landraud L, Lemichez E. Rho GTPase-activating bacterial toxins: from bacterial virulence regulation to eukaryotic cell biology. FEMS microbiology reviews. 2007;31:515–534. doi: 10.1111/j.1574-6976.2007.00078.x. [DOI] [PubMed] [Google Scholar]
  13. Liverman AD, Cheng HC, Trosky JE, Leung DW, Yarbrough ML, Burdette DL, et al. Arp2/3-independent assembly of actin by Vibrio type III effector VopL. Proceedings of the National Academy of Sciences of the United States of America. 2007;104:17117–17122. doi: 10.1073/pnas.0703196104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Matsuda S, Okada N, Kodama T, Honda T, Iida T. A cytotoxic type III secretion effector of Vibrio parahaemolyticus targets vacuolar H+-ATPase subunit c and ruptures host cell lysosomes. PLoS pathogens. 2012;8:e1002803. doi: 10.1371/journal.ppat.1002803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Munro P, Flatau G, Doye A, Boyer L, Oregioni O, Mege JL, et al. Activation and proteasomal degradation of rho GTPases by cytotoxic necrotizing factor-1 elicit a controlled inflammatory response. The Journal of biological chemistry. 2004;279:35849–35857. doi: 10.1074/jbc.M401580200. [DOI] [PubMed] [Google Scholar]
  16. Park KS, Ono T, Rokuda M, Jang MH, Okada K, Iida T, Honda T. Functional characterization of two type III secretion systems of Vibrio parahaemolyticus. Infection and immunity. 2004;72:6659–6665. doi: 10.1128/IAI.72.11.6659-6665.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ridley AJ, Hall A. The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell. 1992;70:389–399. doi: 10.1016/0092-8674(92)90163-7. [DOI] [PubMed] [Google Scholar]
  18. Ritchie JM, Rui H, Zhou X, Iida T, Kodoma T, Ito S, et al. Inflammation and disintegration of intestinal villi in an experimental model for Vibrio parahaemolyticus-induced diarrhea. PLoS pathogens. 2012;8:e1002593. doi: 10.1371/journal.ppat.1002593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Schmidt G, Sehr P, Wilm M, Selzer J, Mann M, Aktories K. Gln 63 of Rho is deamidated by Escherichia coli cytotoxic necrotizing factor-1. Nature. 1997;387:725–729. doi: 10.1038/42735. [DOI] [PubMed] [Google Scholar]
  20. Zhang L, Krachler AM, Broberg CA, Li Y, Mirzaei H, Gilpin CJ, Orth K. Type III effector VopC mediates invasion for Vibrio species. Cell reports. 2012;1:453–460. doi: 10.1016/j.celrep.2012.04.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Zhang L, Orth K. Virulence determinants for Vibrio parahaemolyticus infection. Current opinion in microbiology. 2013;16:70–77. doi: 10.1016/j.mib.2013.02.002. [DOI] [PubMed] [Google Scholar]
  22. Zhou X, Gewurz BE, Ritchie JM, Takasaki K, Greenfeld H, Kieff E, et al. A Vibrio parahaemolyticus T3SS effector mediates pathogenesis by independently enabling intestinal colonization and inhibiting TAK1 activation. Cell reports. 2013;3:1690–1702. doi: 10.1016/j.celrep.2013.03.039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Zhou X, Shah DH, Konkel ME, Call DR. Type III secretion system 1 genes in Vibrio parahaemolyticus are positively regulated by ExsA and negatively regulated by ExsD. Molecular microbiology. 2008;69:747–764. doi: 10.1111/j.1365-2958.2008.06326.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp FigureS1-S5
NIHMS733594-supplement-Supp_FigureS1-S5.pdf (703KB, pdf)
Supp TableS1
NIHMS733594-supplement-Supp_TableS1.docx (91.4KB, docx)

RESOURCES