Proinflammatory signalling stimulated by the type III translocation factor YopB is counteracted by multiple effectors in epithelial cells infected with Yersinia pseudotuberculosis

Authors

  • Gloria I. Viboud,

    1. Department of Molecular Genetics and Microbiology, Center for Infectious Diseases, School of Medicine, State University of New York at Stony Brook, Stony Brook, NY 11794-5222, USA.
    Search for more papers by this author
  • Stephane Shu Kin So,

    1. Department of Molecular Genetics and Microbiology, Center for Infectious Diseases, School of Medicine, State University of New York at Stony Brook, Stony Brook, NY 11794-5222, USA.
    Search for more papers by this author
  • Michelle B. Ryndak,

    1. Department of Molecular Genetics and Microbiology, Center for Infectious Diseases, School of Medicine, State University of New York at Stony Brook, Stony Brook, NY 11794-5222, USA.
    Search for more papers by this author
  • James B. Bliska

    Corresponding authorSearch for more papers by this author

Summary

Type III secretion systems are used by several pathogens to translocate effector proteins into host cells. Yersinia pseudotuberculosis delivers several Yop effectors (e.g. YopH, YopE and YopJ) to counteract signalling responses during infection. YopB, YopD and LcrV are components of the translocation machinery. Here, we demonstrate that a type III translocation protein stimulates proinflammatory signalling in host cells, and that multiple effector Yops counteract this response. To examine proinflammatory signalling by the type III translocation machinery, HeLa cells infected with wild-type or YopY. pseudotuberculosis strains were assayed for interleukin (IL)-8 production. HeLa cells infected with a YopEHJ triple mutant released significantly more IL-8 than HeLa cells infected with isogenic wild-type, YopE, YopH or YopJ bacteria. Complementation analysis demonstrated that YopE, YopH or YopJ are sufficient to counteract IL-8 production. IL-8 production required YopB, but did not require YopD, pore formation or invasin-mediated adhesion. In addition, YopB was required for activation of nuclear factor kappa B, the mitogen-activated protein kinases ERK and JNK and the small GTPase Ras in HeLa cells infected with the YopEHJ mutant. We conclude that interaction of the Yersinia type III translocator factor YopB with the host cell triggers a proinflammatory signalling response that is counteracted by multiple effectors in host cells.

Introduction

Type III secretion systems are used by several Gram-negative pathogens to deliver virulence proteins inside host cells (Galán and Bliska, 1996;Hueck, 1998; Plano et al., 2001; Buttner and Bonas, 2002). Yersinia pestis, Y. pseudotuberculosis and Y. enterocolitica contain a plasmid-encoded type III system that functions to counteract host signalling responses during infection (reviewed by Cornelis et al., 1998). This type III system is composed of a secretion channel encoded by the ysc operons that secretes proteins across both bacterial membranes. YopB, YopD and LcrV are thought to comprise the translocation machinery that delivers virulence proteins across the host cell membrane (Persson et al., 1995; Håkansson et al., 1996; Sarker et al., 1998; Pettersson et al., 1999; Holmström et al., 2001). Both YopB and YopD have been shown to insert into liposomes after secretion, and YopB is required for the formation of a transmembrane channel under these conditions (Tardy et al., 1999). At least six effector proteins have been shown to be delivered into host cells. These so-called effector Yops are YopE, YopH, YopJ (YopP), YopM, YopO (YpkA) and YopT (Juris et al., 2002).

YopE, YopT and YopH inhibit phagocytosis by disrupting cytoskeletal-organizing machinery (Bliska, 2000). YopH is a protein tyrosine phosphatase that dephosphorylates scaffolding proteins involved in the formation of focal adhesions, such as p130Cas, focal adhesion kinase (FAK) and paxillin (Black and Bliska, 1997; Persson et al., 1997; Black et al., 1998). Dephosphorylation of these targets antagonizes integrin-mediated bacterial phagocytosis by host cells. YopH has also been shown to inhibit the activation of T cells and B cells (Yao et al., 1999). More recently, it has been reported that YopH has an inhibitory effect on a PI3-kinase-dependent signalling pathway that promotes proliferation of T cells and monocyte chemoattractant protein 1 expression in macrophages (Sauvonnet et al., 2002).

YopE and YopT inhibit actin rearrangements by inactivating host Rho GTPases. YopE is known to act as a GTPase-activating protein (GAP) (Black and Bliska, 2000; Von Pawel-Rammingen et al., 2000). YopE inhibits RhoA, Rac-1 and Cdc42 by accelerating the conversion of the GTP-bound form of the Rho GTPase to the GDP-bound inactive form. The GAP activity of YopE is also needed to prevent the formation of pores generated by insertion of the translocation machinery in the host cell plasma membrane (Viboud and Bliska, 2001). YopT has been found previously to inhibit Rho GTPases by releasing them from the membrane (Zumbihl et al., 1999). It was reported recently that YopT acts as a cysteine protease that cleaves the prenyl group of lipid-modified Rho GTPases (Shao et al., 2002). One unique feature of YopT is that it is not universally produced by pathogenic Y. pseudotuberculosis species (our unpublished data).

YopJ is known to trigger apoptosis in macrophages and to prevent cytokine synthesis in host cells (Juris et al., 2002; Orth, 2002). Inhibition of cytokine production by YopJ is accomplished by the downregulation of mitogen-activated protein kinase (MAPK) and nuclear factor kappa B (NFκB) pathways. YopJ can bind to and inhibit activation of multiple MAPK kinases (MAPKK). A recent report indicated that YopJ acts as a ubiquitin-like cysteine protease (Orth et al., 2000). It is still not known which proteins are targets of YopJ, or how ubiquitin-like protein modifications regulate MAPK signalling.

Two well-studied MAPK cascades are the ERK and JNK pathways. Activation of ERK is mediated by the Ras–Raf pathway, whereas activation of JNK is controlled by Rho GTPases (Coso et al., 1995; Kolch, 2000). Activation of the MAPK and NFκB pathways mediates the induction of transcription factors that activate interleukin (IL)-8 synthesis (Mukaida et al., 1994). IL-8 is an important proinflammatory chemokine that is produced by epithelial cells.

Lipopolysaccharide (LPS) appears to be the main stimulus that causes MAPK and NFκB activation leading to proinflammatory responses in macrophages during Yersinia infection (Palmer et al., 1998; Ruckdeschel et al., 1998). Unlike macrophages, epithelial cells typically respond poorly to LPS. This is thought to protect the intestinal epithelium from the inflammation that would result from the large number of Gram-negative bacteria in the normal flora (Didierlaurent et al., 2002). Thus, in this cell type, proinflammatory responses are typically triggered by stimuli other than LPS. Previous studies have shown that Yersinia infection of epithelial cells stimu-lates IL-8 production (Schulte et al., 1996; Schulte and Autenrieth, 1998). The invasin protein has been shown to play a role in stimulating IL-8 production during Yersinia infection (Schulte et al., 1998; 2000). However, these findings were largely based on studies using plasmid-cured strains, and would not have uncovered a role for plasmid-encoded factors in stimulating IL-8 production.

We have shown previously that infection of epithelial cells with a Y. pseudotuberculosis YopE mutant triggers YopB-dependent actin polymerization at the site of bacterial–host cell interaction, which was correlated with the formation of pores in the host membrane (Viboud and Bliska, 2001). Overproduction of activated Rho GTPases caused pore formation in cells infected with wild-type bacteria. Based on these results, we hypothesized that association of the translocation machinery with the host membrane triggers a signalling response that activates Rho GTPases, actin polymerization and pore formation. As Rho GTPases can regulate proinflammatory signalling responses as well as actin polymerization, we considered the possibility that interaction of the translocation machinery with host cell membranes could trigger activation of proinflammatory signalling that results in IL-8 production. With this aim in mind, we examined the production of IL-8 and the activation of the NFκB and the ERK and JNK MAPK pathways in epithelial cells infected with Yersinia Yop mutants. We obtained evidence that the type III translocation factor YopB activates proinflammatory signalling in epithelial cells, and that YopE, YopH and YopJ function to counteract this host response.

Results

Epithelial cells infected with a Y. pseudotuberculosis strain deficient in YopE, YopH and YopJ produce high levels of IL-8

To investigate whether the type III translocation machinery can activate proinflammatory signalling during Yersinia infection, we assayed the production of IL-8 in HeLa cells infected with a YopEHJOKM mutant. This strain (YP37) is deficient in the known effector Yops, but produces and secretes wild-type levels of the translocation factors YopB, YopD and LcrV (Fig. 1A). After 1 h of infection, bacteria were killed by the addition of gentamicin and, after an additional incubation of 4 h, IL-8 production from the HeLa cells was measured using an enzyme-linked immunosorbent assay (ELISA). HeLa cells infected with the YopEHJOKM mutant released high levels of IL-8 (Fig. 1B). In contrast, very low levels of IL-8 were produced when HeLa cells were infected with the wild-type strain (Fig. 1B). These results indicate that proinflammatory signalling is stimulated during infection with the YopEHJOKM mutant, and that one or more Yop effectors counteract this signalling in HeLa cells infected with the wild-type strain.

Figure 1.

IL-8 production by HeLa cells infected with Y. pseudotuberculosis .

A. SDS-PAGE analysis of secreted Yop proteins. The wild type, YopEHJKOM , YopEHJ , YopEHJB and YopEHJB /pYopB Y. pseudotuberculosis strains were grown under low-calcium conditions for 3 h. Proteins in the culture supernatants were TCA precipitated and analysed by SDS-PAGE and Coomassie blue staining. Molecular weights of reference proteins (in kDa) are shown on the left. Positions of Yops and LcrV are indicated by arrows on the right. A truncated form of YopE that is secreted by YopE strains runs near the bottom of the gel.

B. Culture supernatants were collected from HeLa cells infected with the wild type, YopE , YopH , YopJ , YopEH , YopEHJ or YopEHJOKM Y. pseudotuberculosis strains. Supernatants from triplicate wells were assayed using an IL-8 ELISA. Two samples were analysed from each well. Error bars represent the standard deviation of the mean of the six values obtained for each strain.

To investigate which effector Yops play a part in counteracting IL-8 production under these conditions, HeLa cells were infected with Y. pseudotuberculosis mutants defective for just one or two Yops. YopJ has been shown previously to inhibit IL-8 production in HeLa cells (Schesser et al., 1998). In accordance with the previous observation, HeLa cells infected with a YopJ mutant released more IL-8 than HeLa cells infected with the wild-type strain (Fig. 1B). However, HeLa cells infected with the YopJ mutant secreted much less IL-8 than HeLa cells infected with the YopEHJOKM mutant (Fig. 1B), indicating that Yops other than YopJ play a role in counteracting infection-induced signalling in HeLa cells. Although infection with a YopH mutant or a YopE mutant did not stimulate more IL-8 production than infection with wild type, a small but significant increase in IL-8 production was observed when HeLa cells were infected with a YopEH mutant (Fig. 1B). In addition, HeLa cells infected with a YopEHJ mutant released almost as much IL-8 as HeLa cells infected with the YopEHJOKM mutant (Fig. 1B). These data suggest that YopE and YopH function in concert with YopJ to counteract proinflammatory signalling that is stimulated during Y. pseudotuberculosis infection of HeLa cells. The other effectors appear to play little if any role in counteracting IL-8 production under these conditions.

Role of YopE, YopH and YopJ in counteracting IL-8 production

To explore further the role of YopH and YopE in counteracting proinflammatory signalling in HeLa cells, we asked whether the catalytic activities of these proteins are required to inhibit IL-8 production. Expression vectors encoding wild-type or catalytically inactive forms of YopH or YopE were introduced into the YopEHJ mutant. The resulting strains were used to infect HeLa cells, and the production of IL-8 was measured by an ELISA. An expression vector encoding YopJ was also introduced into the YopEHJ mutant, and this strain was analysed in parallel as a control. As shown in Fig. 2, IL-8 production was significantly reduced when YopH, YopE or YopJ were expressed in the YopEHJ mutant. In contrast, IL-8 production was not diminished when the catalytically inactive forms of YopE (YopER144A) or YopH (YopHC403S) were expressed in the YopEHJ mutant (Fig. 2). These results indicate that the catalytic activities of YopE or YopH are necessary for counteracting IL-8 production in infected HeLa cells. Although not specifically tested here, it has been shown previously that the catalytic activity of YopJ is required to inhibit tumour necrosis factor (TNF)α release from macrophages infected with Y. pseudotuberculosis (Orth et al., 2000).

Figure 2.

YopE, YopH or YopJ inhibit IL-8 production. HeLa cells were infected with the YopEHJ , YopEHJ /pYopE, YopEHJ /pYopH, YopEHJ /pYopJ, YopEHJ /pYopER144A or YopEHJ /pYopHC 403S Y. pseudotuberculosis strains. Culture supernatants were assayed for IL-8 as described in the legend to Fig. 1 .

YopB is required for IL-8 production

We postulate that association of the translocation machinery with the plasma membrane of HeLa cells stimulates a signalling response that culminates in the synthesis and release of IL-8. If this is the case, then IL-8 production should require YopB, which is an essential component of the translocation machinery (Håkansson et al., 1996). To test this hypothesis, we infected HeLa cells with a YopEHJB mutant. HeLa cells infected with the YopEHJB mutant released very small amounts of IL-8 (Fig. 3A). IL-8 production was restored when a plasmid encoding wild-type YopB (pYopB) was introduced into the YopEHJB mutant (Fig. 3A), indicating that YopB is required for induction of proinflammatory signalling under these conditions.

Figure 3.

YopB is required for IL-8 production. HeLa cells were infected with (A) YopEHJ and YopEHJB /pYopB or (B) YopEHJ /Inv , YopEHJB and YopEHJ Y. pseudotuberculosis strains. Culture supernatants were assayed for IL-8 as described in the legend to Fig. 1 .

Previous studies have shown that the invasin (Inv) protein can stimulate IL-8 production in epithelial cells infected with Y. enterocolitica (Schulte et al., 1998; 2000). To examine the role of invasin in our infection system, HeLa cells were infected with a YopEHJ/Inv mutant. The amount of IL-8 released by cells infected with the YopEHJ/Inv mutant was slightly higher than that secreted by cells infected with the YopEHJ mutant (Fig. 3B). These results indicate that invasin does not play an essential role in stimulating IL-8 secretion in HeLa cells infected with a Y. pseudotuberculosis YopEHJ mutant.

Pore formation is not required for IL-8 production

We have shown previously that infection with YopE mutant Y. pseudotuberculosis stimulates the formation of pores in HeLa cells, and that YopB is required for this activity (Viboud and Bliska, 2001). It is known that pore-forming toxins induce IL-8 production (Dragneva et al., 2001; Walev et al., 2002). We then considered the possibility that pore formation in infected HeLa cells was responsible for stimulating IL-8 production. The role of pore formation in stimulating IL-8 production was examined by treating infected cells with the actin polymerization inhibitor cytochalasin D, which we have shown previously to block YopB-dependent pore formation effectively (Viboud and Bliska, 2001). HeLa cells were treated with 3.9 mM cytochalasin D or vehicle alone (DMSO) for 2 h before and during infection with wild-type Y. pseudotuberculosis or the YopEHJ mutant. Lactate dehydrogenase (LDH) release was assayed 3 h after infection to quantify osmotic lysis of the HeLa cells. Osmotic lysis is a downstream consequence of pore formation. As shown in Fig. 4, although cytochalasin D efficiently blocked LDH released from cells infected with the YopEHJ mutant (Fig. 4A), it did not significantly reduce IL-8 production (Fig. 4B). These data argue that pore formation is not required to stimulate IL-8 production by infected HeLa cells.

Figure 4.

Pore formation is not required for IL-8 production in HeLa cells.

A and B. HeLa cells were exposed to cytochalasin D dissolved in DMSO (CD+) or to DMSO only (CD–) for 2 h before and during infection. HeLa cells were infected with the wild-type or YopEHJ mutant Y. pseudotuberculosis strains.

A. Culture supernatants were collected from wells containing HeLa cells 3 h after infection, and the amount of LDH was measured using a CytoTox 96 assay kit (Promega). Percentage (%) LDH release was calculated by dividing the amount of LDH released from infected cells by the amount of LDH released from uninfected cells that were lysed by a freeze–thaw cycle. Error bars represent the mean and standard deviation of the values obtained from triplicate samples.

B. Supernatants were collected and tested for IL-8 production as described in the legend to Fig. 1 .

C and D. HeLa cells were infected with YopEHJ, YopEHJD or YopEHJD/pYopD mutant strains, and supernatants were analysed as described above for pore-forming activity by the LDH assay(C) or for IL-8 production (D).

YopD, like YopB and LcrV, is thought to be a component of the translocation channel. YopD mutant strains are defective for translocation and are unable to cause pore formation (Neyt and Cornelis, 1999). To determine whether YopD is required for IL-8 production, we infected HeLa cells with the YopEHJ strain or a YopEHJD mutant. HeLa cells infected with the YopEHJD mutant released lower levels of LDH than HeLa cells infected with the YopEHJ mutant (Fig. 4C). Pore-forming activity was restored when a plasmid encoding YopD was introduced into the YopEHJD mutant (YopEHJD/pYopD) (Fig. 4C), demonstrating that YopD is required for pore formation. Analysis of IL-8 levels in supernatants from infected cells showed that the amount of IL-8 production stimulated by the YopEHJD mutant was comparable with that induced by the YopEHJ strain. These data indicate that neither YopD nor pore formation was required for IL-8 production.

YopB is required for activation of MAPKs

Many of the extracellular signals that lead to the activation of IL-8 converge on two main transcription factors AP-1 and NFκB. The AP-1 family of transcriptional activators is controlled by the MAPK signalling cascades. We investigated whether the MAPKs JNK or ERK are activated in HeLa cells in response to infection with YopEHJ mutant bacteria. HeLa cells were infected with the YopEHJ or YopEHJB mutant strains and, at different time points after infection, the HeLa cells were lysed, and samples of the lysates were analysed by immunoblotting using phosphospecific antibodies to JNK or ERK. Stripped blots were reprobed using standard antibodies specific for JNK or ERK to control for loading. Robust MAPK activation was observed when HeLa cells were infected with the YopEHJ mutant, whereas little or no MAPK activation was detected in cells infected with the YopEHJB mutant or in uninfected cells (Fig. 5). Interestingly, the kinetics of activation for the two MAPKs was notably different. Although ERK activation was detected as early as 15 min after infection (Fig. 5, top, lane 2), activation of JNK was not observed until 45 min after infection (Fig. 5 bottom, lane 6). Thus, Y. pseudotuberculosis infection of HeLa cells triggers YopB-dependent activation of MAPK cascades.

Figure 5.

YopB-dependent MAPK activation. HeLa cells were left uninfected or were infected with the YopEHJ or YopEHJB mutant strains for 15, 30, 45 or 60 min. Cells were lysed, and soluble fractions of the lysates were separated by centrifugation. Samples of protein equivalent to 1 × 10 7 HeLa cells were separated by SDS-PAGE and analysed by immunoblotting with antiphospho ERK antibodies (top) or antiphospho JNK antibodies (bottom). Loading controls were obtained by stripping and reprobing the blots with standard anti-ERK and anti-JNK antibodies. The different isoforms of JNK or ERK expressed in HeLa are indicated by p54/p46 or p44/p42 respectively.

YopB is required for activation of NFκB

The transcription factor NFκB is required for inducible expression of IL-8 and other proinflammatory mediators (reviewed by Li and Stark, 2002). Here, we examined the ability of the YopEHJ mutant strain to activate NFκB. HeLa cells infected for 1 h with the YopEHJ or YopEHJB mutant strains were lysed, and samples of the lysates were analysed using an NFκB ELISA. In this assay, activated NFκB specifically binds to an oligonucleotide containing the NFκB binding site attached at the bottom of the ELISA plate. An antibody against the NFκB p65 subunit detects the NFκB complex bound to the oligonucleotide. This complex is, in turn, recognized by a secondary antibody conjugated to horseradish peroxidase. Lysates from non-infected cells or cells infected with the YopEHJB mutant yielded small amounts of active NFκB (Fig. 6). Interestingly, infection with YopEHJ induced high levels of the active transcription factor (Fig. 6). These data indicate that infection of HeLa cells with a YopEHJ strain induces a strong YopB-mediated activation of NFκB.

Figure 6.

YopB-inducible activation of NFκB. HeLa cells left uninfected or infected with the YopEHJ or YopEHJB mutant strains for 1 h were lysed, and samples of the lysates were analysed using an NFκB ELISA, as described in the Experimental procedures . Results shown are the averages obtained from triplicate samples.

Ras is activated in a YopB-dependent manner

Activation of the JNK cascade is mainly controlled by Rho GTPases (Coso et al., 1995). It is therefore likely that the activation of JNK that we observed upon infection with the YopEHJ mutant results from the YopB-dependent activation of Rho GTPases that we have already reported (Viboud and Bliska, 2001). On the other hand, activation of the ERK cascade is controlled by Ras GTPases. Both Ras-dependent and -independent pathways have been reported to activate NFκB (Li and Stark, 2002). To determine whether infection with the YopEHJ mutant could activate Ras, we used a Ras activation assay. In this assay, the Ras-binding domain of Raf (Raf-RBD) was used to bind selectively the activated form of Ras in lysates of infected cells. Subsequently, an anti-Ras monoclonal antibody was used to detect the bound Ras by immunoblot analysis. Lysates of cells infected for 15 or 60 min with either YopEHJ or the YopEHJB mutant were subjected to affinity purification with beads bound to Raf-RBD. As shown in Fig. 7, higher levels of activated Ras were present in cells infected with the YopEHJ mutant than in those infected with the YopEHJB mutant (compare lanes 2 and 3, and 4 and 5). These results suggest that infection of HeLa cells with the YopEHJ mutant activates Ras in a YopB-dependent manner.

Figure 7.

YopB-dependent activation of Ras. Uninfected HeLa cells or HeLa cells infected for 15 or 60 min with the YopEHJ or YopEHJB mutant strains were lysed, and samples of soluble protein were incubated with GST-Raf-1 RBD agarose beads. The beads were washed, boiled in sample buffer to elute bound proteins, and the eluted proteins were separated by SDS-PAGE. Immunoblotting was performed with an anti-Ras monoclonal antibody (top) or an anti-GST monoclonal antibody (bottom) to control for recovery of GST-Raf-1 RBD.

Discussion

In a previous study, we showed that infection of epithelial cells with YopE mutant Y. pseudotuberculosis induced strong actin polymerization at the site of bacterial–host contact (Viboud and Bliska, 2001). As actin polymerization required YopB, we hypothesized that interaction of the type III translocation machinery with host membranes triggers signalling that leads to the activation of the Rho family of GTPases. These data prompted us to explore whether signalling induced by the association of the translocation machinery with the host cell might additionally lead to transcriptional events resulting in a proinflammatory response. With this aim, we examined the proinflammatory signalling responses of HeLa cells to infection with Yop mutant Y. pseudotuberculosis strains. We obtained evidence that the Yersinia type III translocator protein YopB is required to induce a proinflammatory response that engages the Ras, ERK, JNK and NFκB pathways, and the production of IL-8. In addition, we show that the YopB-dependent signalling response is counteracted by YopE, YopH and YopJ in host cells infected by wild-type Y. pseudotuberculosis.

Using HeLa cells infected with wild-type Y. pseudotuberculosis and various Yop mutants, we found that, in contrast to the wild type, a YopEHJ mutant induces high levels of IL-8 production. We present evidence that YopE and YopH, together with YopJ, are able to inhibit IL-8 production. Until now, YopJ was the only effector Yop recognized as having an anti-inflammatory role. We have found previously that YopJ is sufficient to inhibit production of TNFα by macrophages infected with Y. pseudotuberculosis (Palmer et al., 1999). TNFα production in macrophages infected with Y. pseudotuberculosis is thought to result from LPS stimulation. In this study, in which we evaluate an LPS-independent proinflammatory signal, we show that YopE and YopH can function to inhibit production of the proinflammatory cytokine IL-8. Using inactive forms of YopE and YopH, we show that the catalytic activities of YopE and YopH are essential for their inhibitory effects.

How can proteins that function as a GAP (YopE) or PTPase (YopH) inhibit IL-8 production? The action of YopE on the JNK pathway is relatively simple to envisage. As JNK activation is controlled by members of the Rho family of GTPases such as Cdc42 and Rac-1 (Coso et al., 1995), YopE GAP activity directed towards these GTPases could result in the downregulation of the JNK pathway. Preliminary experiments indicate that YopE strongly inhibits activation of JNK and weakly inhibits activation of ERK (unpublished data). In this context, it is worth mentioning that the JNK and ERK cascades are not totally independent. In fact, it is known that Rac and Cdc42, which activate the JNK pathway, also activate p21-activated protein kinase (PAK). PAK co-operates with Raf-1 to activate the ERK cascade (Frost et al., 1997). In addition, it has also been found that Rac- or Cdc42-activated PAK collaborates with Ras to activate Raf. Although necessary for Raf activation, the Rho GTPase family is not sufficient to upregulate Raf (Li et al., 2001). A role for Rho GTPases in co-operating with Ras to regulate the ERK pathway is consistent with the observed partial inhibitory effect of YopE on ERK activation.

How YopH may inhibit proinflammatory signalling is less evident. Preliminary experiments show that YopH inhibits activation of ERK and JNK in infected HeLa cells (unpublished data). It has been reported recently that YopH has an inhibitory effect on a PI3-kinase signalling pathway (Sauvonnet et al., 2002). It is therefore possible that YopH also inhibits IL-8 production by interfering with a PI3-kinase-dependent signalling pathway that leads to NFκB activation (Sizemore et al., 1999).

The interaction between the Yersinia invasin protein and the β1 integrin receptor permits bacterial attachment to the host cell and triggers internalization (Isberg and Barnes, 2001). Previous studies uncovered a role for invasin-mediated adhesion in the production of IL-8 in epithelial cells infected with Y. enterocolitica (Schulte et al., 1998; 2000). In our study, the invasin protein was not required for IL-8 production in HeLa cells infected with YopEHJ mutant Y. pseudotuberculosis. This discrepancy could be explained by differences in the strains used for infection. Although invasin is chromosomally encoded, the type III system is encoded on the virulence plasmid. In the earlier studies, strains of Y. enterocolitica that were cured of the virulence plasmid were used (Schulte et al., 1998; 2000). We have been able to reproduce the earlier results using plasmid-cured Inv+ or Inv variants of Y. pseudotuberculosis. A low level of invasin-dependent IL-8 production (e.g. 500 pg ml−1) is observed when HeLa cells are infected with a plasmid-cured strain of Y. pseudotuberculosis (unpublished results). In contrast, much higher levels of IL-8 are produced (e.g. 1500 pg ml−1) when HeLa cells are infected with a YopEHJ mutant. Thus, the strength of the YopB-dependent proinflammatory signal may mask that of the invasin protein in our experiments.

We have observed previously that a YopE mutant has YopB-dependent pore-forming activity that leads to lysis of the infected cell (Viboud and Bliska, 2001). As pore-forming toxins are known to activate IL-8 production (Dragneva et al., 2001; Walev et al., 2002), we were concerned about the possibility that pore formation itself might trigger IL-8 induction. However, we found that treatment of cells with cytochalasin D, which completely blocks pore formation in cells infected with a YopEHJ mutant, does not affect the levels of IL-8 induced. Moreover, a YopEHJD mutant that is unable to cause pore formation stimulates IL-8 production as well as a YopEHJ strain. These results indicate that IL-8 production in cells infected with a YopEHJ mutant does not require YopD function or pore formation. In contrast, our data suggest that proinflammatory signalling and pore formation are two independent downstream effects of a process that requires YopB.

We hypothesize that IL-8 production in HeLa cells infected with the YopEHJ mutant is initiated by association of the translocation factor YopB with the surface of the host cell. It is possible that YopB interacts directly with the plasma membrane to trigger this response. For example, YopB has been shown to insert into liposomes incubated with Y. enterocolitica (Tardy et al., 1999). Insertion of YopB into liposomes under these conditions is associated with the formation of an ion-conducting channel (Tardy et al., 1999). YopD does not appear to be required for insertion of YopB into liposomes, although YopD is required for the formation of an ion-conducting channel (Tardy et al., 1999). We envisage that insertion of YopB into the membranes of live cells initiates the formation of a channel that is composed of YopB, YopD and LcrV and is required for Yop translocation as well as pore formation. The initial step of YopB insertion into host cell membranes could elicit a stress response stimulus that activates Ras and Rho GTPases. An alternative hypothesis is that, after the insertion process, a particular region of YopB is exposed to the cytoplasmic compartment of the host cell, and that this domain engages a specific signalling pathway. Further studies, in which the individual functional domains of YopB are analysed, should provide insight into the nature and biological role of this signalling response.

In conclusion, we obtained evidence that the Yersinia type III translocation factor YopB is involved in the induction of a signalling response in epithelial cells. This type III-mediated signalling, which involves Ras and Rho GTPases, is probably not restricted to Yersinia and might be taking place during infection with other pathogens that encode homologues of YopB. For example, two recent reports show that enteropathogenic Escherichia coli (EPEC), which encodes the YopB homologue EspD, is able to induce a type III-dependent activation of MAPKs (Czerucka et al., 2001; de Grado et al., 2001). In contrast to EPEC, wild-type Yersinia prevents the consequence of this signalling by the action of several effector Yops.

Experimental procedures

Plasmids

The plasmids pYopE, pYopER144A (Black and Bliska, 2000), pYopH and pYopHC403S (Black et al., 1998) and pYopJ (pLP11) (Palmer et al., 1998) have been described previously and are derived from the low-copy expression vector pMMB67HE. The plasmid pYopB was constructed by inserting a DNA sequence coding for yopB into pMMB67HE, placing yopB expression under the control of the tac promoter. The yopB coding sequence used for this construction was obtained as a polymerase chain reaction (PCR) product, using the primers B1 and B4 (Palmer et al., 1998), and YP126 (see below) DNA as template. The plasmid pYopD was constructed in a similar fashion using primers D1 (5′-CGGATCCCATATGACAATAAATATCAAGACAGACAGCC-3′) and D2 (5′-CGGATCCGAATTCTCAGACAACACCAAAAGC GGCTT-3′) to amplify a PCR product using YP126 DNA as a template. The yopD coding sequence was inserted into pMMB67HE, placing yopD expression under the control of the tac promoter. All constructs generated using PCR were verified by sequencing. Plasmids were introduced into Y. pseudotuberculosis by conjugation as described previously (Bliska and Black, 1995).

The plasmids pLP4 and pLP13 used to inactivate yopM or yopJ, respectively, were constructed using the suicide plasmid pSB890 as described previously (Palmer et al., 1998). The plasmids pYopOD1-2199 and pYopDΔStyI used to inactivate yopO(ypkA) or yopD, respectively, were constructed using pSB890 as follows. A DNA fragment containing a precise deletion of the yopO(ypkA) coding region (2199 nucleotide deletion) was produced using recombinant PCR and the primers A3 (5′-CCGGATCCATCACTAAAAATCAGTG GCTGGAAG-3′), A4 (5′-GGGCACTT GAGATCTGCTTTACT CATCCCCATTTAACCG-3′), A5 (5′-GAGTAAAGCAGATCTC CAAGTGCCCCCTAAGCCTTGAG-3′) and J4 (Palmer et al., 1998). PCRs were carried out using A3 and A4 or A5 and J4 and YP126 DNA as template. The resulting products were mixed and used as template for a second PCR using primers A3 and J4. The resulting product was inserted into the BamHI site of pSB890, yielding pYopOΔ1-2199. To construct pYopDΔStyI, the yopD coding region was inserted into the BamHI site of pSB890. The resulting plasmid was treated with StyI, which cuts once near the centre of the yopD coding sequence. After treatment with Klenow fragment to generate blunt ends, the DNA was recircularized by ligation to create pYopDΔStyI, which contains a frameshift mutation in yopD.

Bacterial strains and infection conditions

The parental serogroup III Y. pseudotuberculosis strain YP126 (Bölin et al., 1982) and the mutants derived thereof [YP15 (YopH), YP6 (YopE), YP26 (YopJ), YP27 (YopEHJ), YP29 (YopEHJB) and YP22 (YopEHK)] have been described previously (Black and Bliska, 1997; Palmer et al., 1998; 1999; Viboud and Bliska, 2001). The YopEHJ/Inv strain (YP202/pYV27) was constructed using the virulence plasmid isolated from YP27. Electroporation was used to introduce pYV27 into an Inv, plasmid-cured strain (YP202) (Simonet and Falkow, 1992). YP37 (YopEHJKOM) was derived from YP22 by sequential inactivation of yopM, yopO(ypkA) and yopJ genes. The genes were inactivated using the plasmids pLP4, pYopOΔ1-2199 and pLP13 and an allelic replacement strategy described previously (Palmer et al., 1998). YP42 (YopEHJD) was derived from YP27 in a similar fashion using allelic replacement (Palmer et al., 1998) and the plasmid pYopDΔStyI.

YP126 and its derivatives carry a naturally occurring deletion in the virulence plasmid, which inactivates the yopT gene, and are thus devoid of YopT activity (unpublished data). Before infection, the bacteria were grown in Luria–Bertani (LB) broth at 37°C under low-calcium conditions to stimulate maximal Yop expression (Palmer et al., 1998). Yops secreted from Y. pseudotuberculosis strains grown under these conditions were analysed by SDS-PAGE and Coomassie blue staining as described previously (Palmer et al., 1998).

Cell culture

HeLa cells were cultured in DMEM (Gibco BRL) supplemented with 10% heat-inactivated fetal calf serum (FCS; Gibco BRL) and 1 mM sodium pyruvate in a 5% CO2 humidified incubator at 37°C. For infection experiments performed to determine IL-8 production and LDH release, 1 × 105 HeLa cells were seeded into wells of a 24-well tissue culture plate 24 h before assay.

Infection conditions

For experiments carried out in the presence of inhibitor, HeLa cells were preincubated for 2 h in the presence of 3.9 mM cytochalasin D (Sigma). HeLa cells were also exposed to the appropriate concentration of vehicle alone (DMSO) as a control. Bacterial infections were performed in fresh media containing the same concentration of inhibitors or vehicle. When necessary, IPTG (100 µM) was added to the cell culture medium to induce expression of the genes cloned under the tac promoter. The plates containing the infected cells were centrifuged for 5 min at 700 r.p.m. and incubated at 37°C with 5% CO2 for different periods of time to allow bacterial–host cell interaction.

IL-8 assay

HeLa cells were infected at a multiplicity of infection (MOI) of 50 as described above. After a 1 h infection, gentamicin was added to a final concentration of 100 mg ml−1 to kill extracellular bacteria. After an additional incubation of 4 h, culture supernatants were collected, centrifuged for 10 min at 12 000 r.p.m., and the supernatants were stored at −20°C until measurement of IL-8 by ELISA (Antigenix America). Values obtained from triplicate wells were assayed in duplicate and averaged.

LDH assay

The LDH assay has been described previously (Viboud and Bliska, 2001). Samples of culture media from infected cells were collected at 3 h after infection. Levels of LDH were assayed using the CytoTox 96 assay kit (Promega) according to the manufacturer's instructions. After 30 min incubation with the substrate, the reaction was stopped, and absorbance at 490 nm was determined using a MRX microplate reader (Dynatech Laboratories). Total LDH was determined by assaying supernatants from uninfected cells that had been lysed by a freeze–thaw cycle. Percentage LDH release was calculated by dividing the amount of LDH released from infected cells by the amount of total LDH.

NFκB assay

NFκB activation was measured using the TransAM NFκB kit (Active Motif). HeLa cells seeded in 60 mm tissue culture dishes were infected for 1 h. After infection, cells were washed twice with ice-cold PBS. To each well, 3 ml of cold PBS was added to collect the cells by scraping with a rubber policeman. Cells were centrifuged for 10 min at 1000 r.p.m. at 4°C. Supernatants were discarded, and the cells were resuspended in 100 ml of lysis buffer provided by the kit and incubated for 10 min at 4°C. Soluble fractions were separated by centrifugation at 14 000 r.p.m. for 20 min at 4°C, and aliquots of each fraction were stored at −80°C. Samples of the fractions containing 5 mg of protein were analysed by the NFκB ELISA. In this assay, activated NFκB specifically binds to the NFκB DNA binding site that is attached at the bottom of the ELISA plate. An antibody against the p65 subunit of NFκB detects the NFκB complex bound to the oligonucleotide that, in turn, is recognized by a secondary antibody conjugated to horseradish peroxidase (HRP). The colorimetric reaction is read at 450 nm. A TNFα-stimulated HeLa cells extract, provided by the kit, is used as a positive control for NFκB activation.

MAPK immunoblotting

HeLa cells were seeded in 60 mm tissue culture dishes. Cells were infected for different periods of time and washed three times with ice-cold PBS containing 1 mM Na3VO4 and 10 mM NaF. To each well, 0.5 ml of cold lysis buffer (10 mM Tris, pH 7.6, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM Na3VO4, 10 mM NaF, 200 µM AEBSF, 20 µM leupeptin and 1 µM pepstatin) was added, and the plates were incubated for 15 min on ice with occasional rocking. The cells from each dish were scraped into a single microcentrifuge tube, and the tube was centrifuged for 10 min at 12 000 r.p.m. at 4°C. Supernatants (soluble fractions) were transferred to new tubes. Samples of the fractions containing equivalent amounts of protein were separated by SDS-PAGE under reducing conditions and transferred onto nitrocellulose. Immunoblotting was performed using polyclonal antiphospho ERK and phospho JNK antibody (Cell Signaling) according to the manufacturer's procedures. Loading controls were assayed using ERK and JNK antibodies (Cell Signaling). Secondary anti-rabbit IgG and anti-mouse IgG antibodies conjugated to HRP were purchased from Sigma.

Ras activation assay

HeLa cells were grown to ≈ 75–80% confluence in 100 mm tissue culture dishes. Cell culture medium was replaced with serum-free medium 12 h before infection. Infected cells were lysed using Magnesium lysis buffer (MLB; Upstate Biotechnology). Samples of soluble protein (1 mg) were incubated with 5 µl of a 50% slurry of the glutathione S-transferase (GST)-Raf-1 RBD agarose (GST fused to the Ras-binding domain of Raf-1) for 30 min at 4°C. The beads were washed three times with MLB buffer and boiled in Laemmli sample buffer. After centrifugation, the supernatant was separated by SDS-PAGE and transferred to a nitrocellulose membrane. The blot was probed with 1:1000 dilution of an anti-Ras monoclonal antibody or anti-GST antibody (Cell Signaling) in TBS containing 5% non-fat dry milk and 0.05% Tween, overnight at 4°C. After incubation with a goat anti-mouse HRP-conjugated IgG, the blot was developed using the enhanced chemiluminescent method (Perkin-Elmer Life Sciences).

Acknowledgements

We thank Maya Ivanov for constructing pYopD, Lance Palmer for constructing pYopOΔ1-2199, and Yue Zhang for reviewing the manuscript. This research was funded by a grant from the National Institute of Health (AI43389) to J.B.B.

Ancillary