The results of this study demonstrate that V. parahaemolyticus causes activation of MAPK in human intestinal epithelial cells and that this activation is linked to the cellular responses elicited by this bacterium. V. parahaemolyticus induced activation of each of the MAPK - JNK, p38 and ERK - in Caco-2 and HeLa cells (Figure 1 and 2). A mutant strain with a non-functional TTSS1 (ΔvscN1) did not cause MAPK activation, providing the first evidence that TTSS1 is responsible for the activation of MAPK in epithelial cells in response to infection with V. parahaemolyticus (Figure 2). While the role of TTSS1 in ERK activation was difficult to observe in Caco-2 cells, differences in the activation of ERK in HeLa cells co-incubated with WT compared to ΔvscN1 bacteria were clearly evident. V. parahaemolyticus therefore now joins a select group of gram-negative pathogens that use TTSS effectors to activate MAPK signalling to promote pathogen infection. Given the important role MAPK play in controlling host innate immune responses and cell growth, differentiation and death, they are commendable targets for pathogenic effectors. While several pathogens use their TTSS to inhibit MAPK activation [34, 35, 42, 43], others activate them. For example, the inflammatory responses induced by the TTSS effectors of Salmonella typhimurium are related to activation of all MAPK, especially p38 which induces IL-8 secretion from epithelial cells , and Burkholderia pseudomallei utilizes its TTSS to induce IL-8 secretion and to increase bacterial internalization via activation of p38 and JNK in epithelial cells .
Several Vibrio spp. manipulate MAPK signalling pathways to induce host cell death or disturb the host response to infection [40, 45–49]. Vibrio vulnificus triggers phosphorylation of p38 and ERK via Reactive Oxygen Species in peripheral blood mononuclear cells thereby inducing host cell death . The CtxB cholera toxin from Vibrio cholerae down-regulates p38 and JNK activation in macrophages leading to suppression of production of TNFα and other pro-inflammatory cytokines [40, 47]. Additionally Flagellin A from V. cholerae contributes to IL-8 secretion from epithelial cells through TLR5 and activation of p38, ERK and JNK . Despite the fact that V. parahaemolyticus possesses flagellin proteins similar to those of V. cholerae , cells co-incubated with heat-killed V. parahaemolyticus did not exhibit MAPK phosphorylation (Figure 1), suggesting an absence of TLR5 recognition of flagellin. TLR5 is activated by dissociated flagellin monomers and the sheathed Vibrio flagella present on intact bacteria have a limited ability to trigger host innate immunity . In our studies bacteria were washed before addition to the cells and were treated at a temperature unlikely to dissociate flagellin monomers , thereby minimising the amounts of flagellin monomers present to trigger TLR5.
The results obtained from LDH assays, MTT assays and fluorochrome staining confirmed that the TTSS1 of V. parahaemolyticus is essential for the cytotoxicity of this bacterium towards epithelial cells (Figure 3). Furthermore these results show that there was no cell death detected prior to the 2 h time point, by which time MAPK activation was observed. It has been reported that undifferentiated Caco-2 cells are more susceptible than other cell types (e.g. HeLa cells) to a TTSS2-mediated delayed cytotoxicity [15, 51]. While TTSS1 was required for cytotoxicity during the first 4 h of co-incubation, there was little difference in the levels of cytotoxicity observed with ΔTTSS1 bacteria compared to WT V. parahaemolyticus when co-incubations were performed for 6 h . This delayed cell death was attributed to the VopT TTSS2 effector . Delayed cytotoxicity was also observed by Burdette et al. in HeLa cells infected with ΔTTSS2/Δvp1680 bacteria . The mechanism of this delayed cytotoxicity is unknown. With extended co-incubations of 8 h we too saw delayed TTSS1- and VP1680-independent cytotoxicity with differentiated Caco-2 cells (unpublished data Finn and Boyd). The delayed cytotoxicity was the not the subject of this study.
The VP1680 effector protein is responsible for the TTSS1-dependent autophagic cytotoxicity against HeLa cells [25, 29]. Our results demonstrated that VP1680 is required for the induction of JNK and p38 phosphorylation in Caco-2 cells (Figure 2) and that JNK and ERK, but not p38, are involved in the TTSS1-dependent cytotoxicity (Figure 4). Each of the 3 MAPK has been proposed to regulate autophagy and/or autophagic cell death, though the role and relative importance of each one seems to be dependent on cell type and on the induction stimulus [52–54]. The activation of JNK and ERK by VP1680 seems to be important for the cytotoxicity of V. parahaemolyticus towards epithelial cells, whereas phosphorylation of p38 by this effector protein plays a different role in modification of host cell behaviour that remains to be defined. In HeLa cells VP1680 is responsible for the activation of ERK, but plays a lesser role in the activation of JNK and p38 than it does in Caco-2 cells (Figure 2). As activation of all three MAPK in HeLa cells in response to V. parahaemolyticus is TTSS1-dependent, but not VP1680-dependent, this points to the existence of an additional MAPK-activating TTSS1 effector that acts in this cell line. Since VP1680 is the principal TTSS1 effector activating MAPK in Caco-2 cells, this would suggest differing sensitivities of cell lines to the TTSS effectors.
The observation that VP1680 induces phosphorylation of all 3 MAPK raises the possibility that this protein may not target the MAPK directly, but may trigger an upstream kinase. In contrast to VP1680, the VopA TTSS2 effector has been found to inhibit MAPK in macrophages by acetylating the upstream MAPK Kinase (MKK) [18, 30]. It is important to note that the VopA studies were performed with transfected eukaryotic cells that expressed VopA heterologously, whereas the current study assessed MAPK activation by intact V. parahaemolyticus. From our studies during co-incubation of V. parahaemolyticus with Caco-2 cells it appears that the MAPK activation of VP1680 is dominant over the inhibitory effect of VopA. V. parahaemolyticus may co-ordinately regulate both TTSS to achieve appropriate control of host responses.
V. parahaemolyticus induced IL-8 secretion in an active manner as a result of delivery of the TTSS effector proteins into host cells (Figure 5). It appears that there may be a balance between TTSS1 and TTSS2 of V. parahaemolyticus where TTSS1 is involved in the activation of IL-8 production by the host while TTSS2 is involved in its inhibition. This correlates with the opposing functions of the TTSS1 effector VP1680 and the TTSS2 effector VopA in activating and inhibiting MAPK phosphorylation. Interestingly, the TTSS1 effector VP1680 mutant (Δvp1680) induced intermediate amounts of IL-8, suggesting an involvement of this protein in stimulating production of this chemokine, but not an absolute requirement (Figure 5). Similarly the inhibitory studies revealed that V. parahaemolyticus induces secretion of IL-8 partly via modulation of the ERK signalling pathway (Figure 6). The complex effect of both TTSS of V. parahaemolyticus on the host immune defence machinery illustrates the powerful tools the bacteria possess to gain maximum advantage from the host environment.