Inhibition of the Mitogen-Activated Protein Kinase Kinase Superfamily by a Yersinia Effector

See allHide authors and affiliations

Science  17 Sep 1999:
Vol. 285, Issue 5435, pp. 1920-1923
DOI: 10.1126/science.285.5435.1920


The bacterial pathogen Yersinia uses a type III secretion system to inject several virulence factors into target cells. One of the Yersinia virulence factors, YopJ, was shown to bind directly to the superfamily of MAPK (mitogen-activated protein kinase) kinases (MKKs) blocking both phosphorylation and subsequent activation of the MKKs. These results explain the diverse activities of YopJ in inhibiting the extracellular signal–regulated kinase, c-Jun amino-terminal kinase, p38, and nuclear factor kappa B signaling pathways, preventing cytokine synthesis and promoting apoptosis. YopJ-related proteins that are found in a number of bacterial pathogens of animals and plants may function to block MKKs so that host signaling responses can be modulated upon infection.

The bacterial pathogenYersinia pestis is the agent of the bubonic plague (1). In addition to Y. pestis, two closely related species, Y. enterocolitica and Y. pseudotuberculosis, harbor a plasmid of ∼70 kb that encodes a contact-dependent type III secretion system. Upon infection, this system delivers virulence factors (referred to as Yersiniaouter proteins, or “Yops”) into host cells. The Yops disrupt host signaling functions to thwart the development of a cell-mediated immune response (1). A virulence factor from Y. pseudotuberculosis, YopJ (2–4) [YopP in Y. enterocolitica (5)], is a 33-kD protein that perturbs a multiplicity of signaling pathways. These include inhibition of the extracellular signal–regulated kinase (ERK), c-Jun NH2-terminal kinase (JNK), and p38 mitogen-activated protein kinase (MAPK) pathways and inhibition of the nuclear factor kappa B (NF-κB) pathway (2, 6–10). The interruption of these signaling pathways results in the disruption of tumor necrosis factor α and interleukin-8 production by the infected target cell (2, 7, 8). Furthermore, the expression of YopJ has been correlated with the induction of apoptosis by Yersinia(5, 8, 9).

The presence of YopJ-related proteins in other animal (AvrA ofSalmonella typhimurium) and plant pathogens (AvrRxv ofXanthamonas campestris pv. vesicatori) as well as in a plant symbiont (Y4LO of Rhizobium NGR234) suggests that this family of proteins plays a fundamentally important role in bacterial-host cell interactions (11, 12). The amino acid sequence identity shared among YopJ, AvrA, AvrRxv, and Y4LO suggests that these effectors function through a common mechanism. However, a search for homologous proteins (by BLAST and other search programs) offers no insight into the function of this family of proteins.

To identify mammalian binding partners of YopJ, we used a yeast two-hybrid screen based on a LexA-YopJ fusion protein and a HeLa cDNA library (13). Positive clones were obtained, encoding fusion proteins of the Gal4 activation domain with MAPK kinases (MKK1, MKK2, and MKK4/SEK1) (Fig. 1A). These interactions were confirmed by cotransforming yeast cells with pLexAde-YopJ with purified plasmids, encoding fusions of the VP16 activation domain with full-length MKK1 or fusions of the Gal4 activation domain with MKK2 (Fig. 1A) (14). YopJ did not interact in the yeast two-hybrid system with MAPKs ERK, JNK, and p38 or with the MAPK kinase kinase (MAPKKK) BRaf, suggesting that the interaction of YopJ with the MKKs is specific (Fig. 1A) (15). To determine if YopJ could bind directly to MKKs in vitro, MKK1, MKK3, MKK4, and MKK5 were expressed as glutathione S-transferase (GST) fusion proteins in bacteria and tested for interaction with in vitro transcribed and translated35S-methionine– labeled YopJ. GST alone failed to bind YopJ, whereas all four GST-MKK fusion proteins bound labeled YopJ (Fig. 1B). Thus, YopJ was able to bind specifically to multiple members of the MKK family in vitro. To demonstrate that YopJ targets the MKKs in vivo, we coimmunoprecipitated YopJ with endogenous MKK2 from 293 cells (Fig. 1C).

Figure 1

YopJ specifically binds MKKs. (A) YopJ cotransformed with positive clones for MKK1 (J98), MKK2 (J31), and MKK4/SEK1 (J27), and pVP16-MKK1, pGAL4-MKK2, pVP16-BRaf, pVP16-ERK, pVP16-p38, and pVP16-JNK [as indicated by the diagram (left)] all grew on tryptophan- and leucine-deficient plates (middle). However, only J98, J31, J27, pVP16-MKK1, and pGAL4-MKK2 grew on histidine-deficient plates, which indicates a positive interaction in the two-hybrid screen (right). (B) YopJ specifically bound recombinant GST-MKK1, GST-MKK3, GST-MKK4/SEK1, and GST-MKK5 but not GST. 35S-methionine–labeled pcDNA3-YopJ was incubated with bacterially expressed recombinant GST (lane 2), GST-MKK1 (lane 3), GST-MKK3 (lane 4), GST-MKK4/SEK1 (lane 5), or GST-MKK5 (lane 6) bound to glutathione-agarose and analyzed as previously described (29, 30). (C) YopJ coimmunoprecipitates with endogenous MKK2. 293 cells were transfected with either a control empty vector (V) (lanes 1 and 3) or full-length YopJ with a COOH-terminal FLAG tag (J) (lanes 2 and 4). Cells were lysed and immunoprecipitated with rabbit antibody to MKK2 (31). Coimmunoprecipitating YopJ was detected by immunoblot analysis with antibody to FLAG (Sigma) (lanes 1 and 2). Expression of YopJ was confirmed by immunoblot analysis of lysates with antibody to FLAG (lanes 3 and 4).

To determine the biological consequences of YopJ binding to MKKs, we performed a series of biochemical experiments to assess the effect of YopJ on the MAPK pathway (Fig. 2). 293 cells were transfected with or without HA-ERK in the presence or absence of YopJ (6). Cells were stimulated with epidermal growth factor (EGF) and then assayed for ERK kinase activity by an immune complex assay in which myelin basic protein (MBP) was used as a substrate (16). EGF-treated cells transfected with HA-ERK displayed ERK activity; however, EGF-treated cells that were cotransfected with HA-ERK and YopJ displayed no ERK kinase activity (Fig. 2A). Therefore, YopJ blocked the MAPK pathway downstream of the EGF stimulus. Next, the effects of YopJ were assayed on ERK activity stimulated by activated Ras (HRasV12) (16). Cells transfected with HRasV12 displayed increase ERK activity, whereas cells transfected in the presence of YopJ displayed no ERK activity (Fig. 2B). A similar inhibitory effect of YopJ was observed when cells were transfected with activated Raf (v-Raf) (16), suggesting that YopJ was inhibiting the MAPK pathway downstream of the MAPKKK (that is, v-Raf) (Fig. 2C). In contrast, when cells were transfected with a constitutively active form of MKK1 (MKK1-ED) (16), robust ERK kinase activity was observed in either the presence or absence of YopJ (Fig. 2D). Therefore, YopJ was inhibiting the MAPK pathway downstream of EGF, Ras, and Raf and upstream of MKK1 (at the level of the MKK activation). These results demonstrate that the biological consequences of the in vitro and in vivo binding of MKKs to YopJ results in a specific block in the ability of Raf to stimulate MKK.

Figure 2

YopJ inhibits the MAPK pathway downstream of EGF, Ras, and Raf and upstream of MKK. (A) 293 cells were transfected with or without HA-ERK (200 ng) in the presence or absence of either a control empty vector (V) (1 μg) or YopJ (J) (1 μg) (6). Cells were then stimulated with EGF (50 ng/ml) for 5 min. Kinase assays were performed on HA-ERK with MBP as a substrate as previously described (16). (B) HRasV12 (500 ng) was transfected into 293 cells with or without HA-ERK in the presence of either the control empty vector (V) or YopJ (J). Kinase assays were performed as described above. (C) v-Raf (100 ng) was transfected into 293 cells with or without HA-ERK in the presence of either the control empty vector (V) or YopJ (J). Kinase assays were performed as described above. (D) MKK1-ED (50 ng) was transfected into 293 cells with or without HA-ERK in the presence of either the control empty vector (V) or YopJ (J). Kinase assays were performed as described above. In (A) through (D), expression of YopJ was confirmed by immunoblot analysis of the lysate with monoclonal antibody to M45 (6). Asterisk indicates a nonspecific band.

Raf is known to function as a kinase that phosphorylates MKK1,2 on two serine residues, which results in MKK activation (17,18). As expected, HA-MKK2 was phosphorylated when cells were stimulated with EGF, but no HA-MKK2 phosphorylation was observed in cells that were cotransfected with YopJ (Fig. 3A). As a result, YopJ prevented MKK2 from being activated by phosphorylation. Similarly, when MKK3 is activated by ultraviolet radiation, YopJ inhibits its phosporylation (19). The addition of YopJ to an in vitro kinase reaction with activated Raf and recombinant MKK1 did not result in a detectable inhibition of MKK phosphorylation, which suggests that an additional factor present in mammalian cells is required for inhibition of MKK phosphorylation (19). To determine if YopJ interfered with phosphorylation of MKKs in host cells infected with Yersinia, we infected murine macrophages with a wild-type or yopJ-mutant strain of Y. pseudotuberculosis and phosphorylation of MKK1,2 was assessed by immunoblotting. Both strains stimulated an initial robust, but transient, phosphorylation of MKK1,2 (Fig. 3B). At 45 min after infection [by which time YopJ had entered the cell (20)] the phosphorylation of MKK1,2 was significantly lower in cells infected with the wild-type strain than in those infected with the yopJ-mutant strain (Fig. 3B, lanes 3 and 5). Similar results were obtained when infected cells were analyzed for MKK4/SEK1 phosphorylation (20). These results demonstrate that YopJ inhibits MKK phosphorylation not only when YopJ is overexpressed in cells, but also when YopJ is delivered to the host cell through the type III secretion system.

Figure 3

YopJ inhibits phosphorylation of MKK2. (A) HA-MKK2 was transfected into 293 cells in the presence or absence of YopJ. Cells were then stimulated with EGF (50 ng/ml) for 5 min. Lysates were immunoprecipitated with antibody to HA, and phosphorylated MKK2 was detected by immunoblotting with antibody to phospho-specific MKK1/2 (New England Biolabs, Beverly, Massachusetts). Identical samples were immunoblotted with antibody to MKK2 to confirm equal expression (31). Immunoblotting with antibody to M45 confirmed expression of YopJ in lysates (6). (B) J774A.1 murine macrophage cells were left untreated or infected with wild-type or yopJ-mutant Y. pseudotuberculosis (6). At the indicated times, the macrophages were lysed in NP-40 buffer. Samples of the lysates adjusted to contain equivalent amounts of total protein were analyzed by immunoblotting (as described above) with antibody to phospho-specific MKK1/2 (top row) to detect phosphorylated MKK1/2 (P-MKK1/2) or antibody to MKK2 (bottom row) to confirm equal loading of MKK2. One experiment, representative of three, is shown.

Previous studies have demonstrated that macrophages infected byYersinia are unable to elicit a proinflammatory response and subsequently die by apoptosis (1). The interaction of YopJ with MKK1, MKK3, MKK4, or MKK5 does not easily explain how this virulence factor could be affecting a proinflammatory or apoptotic response. Proinflammatory responses and the anti-apoptotic machinery are regulated by the NFκB pathway through the inhibitor of NF-κB (IκB) kinase complex, which in turn is regulated by upstream kinases (21, 22). The components of the IκB kinase complex that regulate the NFκB pathway include IKKα and IKKβ, which are activated by morphogenic signals and proinflammatory signals, respectively (23). In addition, recent studies have demonstrated that these kinases can be phosphorylated and activated by the MAPKKK, MEKK1 (24, 25). Figure 4D shows the parallel kinase cascades leading to MAPK and NFκB activation, where each step in the pathway contains a representative counterpart (that is, Raf and MEKK1, and MKK and IKKα or IKKβ). It was speculated that if YopJ recognized IKK as a MKK-like molecule and bound to it, the resulting interaction would result in inactivation of NFκB signaling. Therefore, we tested whether YopJ was able to inhibit the NFκB pathway downstream of MEKK1 by transfecting cells with an activated form of MEKK1 (ΔMEKK1) in the presence or absence of YopJ. Figure 4A shows that YopJ expression efficiently blocks the proinflammatory pathway downstream of MEKK1. It was next determined whether the IKKs, like the MKKs, could interact with YopJ. When labeled YopJ was incubated with GST, GST-IKKα, GST-IKKβ, and GST-BRaf in vitro (25, 26), YopJ did not interact with GST, GST-IKKα, or GST-BRaf but did interact with GST-IKKβ (Fig. 4B). Proinflammatory signaling machinery targets activation of IKKβ but not IKKα (23), suggesting that the IKKβ contains distinct structural features that allow it to be recognized as a target by both the upstream proinflammatory signaling machinery (that is, MEKK1) and YopJ. Not only was an interaction observed with YopJ and IKKβ in vitro, but it was also found that YopJ was able to coprecipitate with IKKβ in vivo (Fig. 4C). These observations support the theory that YopJ is inhibiting both the NFκB pathway and the MAPK pathway downstream of MKKKs and upstream of MKKs through an interaction with the MKKs (Fig. 4D).

Figure 4

YopJ interacts with IKKβ and inhibits the NFκB pathway downstream of MEKK1. (A) YopJ inhibits the NFκB pathway downstream of MEKK1. Cells were transfected with increasing concentrations of MEKK1 (as indicated) with a pCMV-lacZ (0.5 μg) plasmid and pNFκB-Luc reporter plasmid (0.1 μg) (Stratagene, La Jolla, California) in the presence or absence of YopJ (1 μg). (B) YopJ specifically binds GST-IKKβ but not GST-IKKα, GST-BRaf, or GST. pEBG-3X (lane 2), pEBG-3X-IKKα (lane 3), pEBG-3X-IKKβ (lane 4), and pEBG-3X-BRaf (lane 5) were expressed in 293 cells. Cells were lysed, and GST fusion proteins were isolated with glutathione-agarose. 35S-methionine–labeled YopJ was incubated with the equivalent amounts of GST fusion proteins and analyzed as previously described (25, 29). (C) YopJ coimmunoprecipitates with transiently expressed GST-IKKβ. pEBG-3X-IKKβ (lanes 1, 2, 5, and 6) or pEBG-3X (lanes 3, 4, 7, and 8) (25) were transfected into 293 cells in the presence of either a control empty vector or a FLAG-tagged YopJ. GST fusion proteins were isolated, and bound YopJ was detected by immunoblot analysis. Expression of YopJ was confirmed by immunoblot analysis of lysates with antibody to FLAG (lanes 5, 6, 7, and 8). (D) YopJ inhibits all pathways that utilize MKKs for signaling, including ERK, p38, and JNK pathways, and YopJ also inhibits the NFκB pathway. The inhibition results in a block of a proinflammatory response and of anti-apoptotic factor expression, thereby promoting programmed cell death.

These results explain how a single bacterial protein, YopJ, prevents the activation of multiple downstream MKK controlled signaling pathways (Fig. 4). MKKs are a highly conserved family of proteins. As essential components of the MAPK signaling pathways, they are responsible for induction of cytokine production (17,27). By targeting this conserved family of proteins, YopJ effectively shuts down the multiple kinase cascades that are required by the host cells to respond to a bacterial infection (6, 7,10). As previously observed, addition of a dominant negative (inactive) form of IKKβ to mammalian cells results in inhibition of the NFκB pathway (22). Therefore, the finding that YopJ binds to IKKβ provides an explanation for the negative effect of YopJ on the proinflammatory pathway induced by NFκB, which uses a mechanism analogous to the one used in the MAPK pathways. In addition to its role as an activator of cytokine gene expression, NFκB regulates the synthesis of anti-apoptotic factors (28). Thus, by simultaneously blocking MAPK and NFκB signaling functions, YopJ blocks synthesis of cytokines as well as anti-apoptotic factors. In the absence of critical anti-apoptotic factors, apoptotic signals dominate, and programmed cell death ensues, unabated.

Together, these results have important implications for understanding the functions of the YopJ-related proteins found in bacteria that establish pathogenic or symbiotic relations with plants. In fact, the plant YopJ-related protein, AvrRxv, has been shown to induce an apoptotic-like program in plants, called the hypersensitive response (11). In conclusion, these results support the concept that YopJ is used by both plant and animal pathogens as well as by plant symbionts to modulate host signaling responses.

  • * Present address: Howard Hughes Medical Institute and Department of Immunology, University of Washington, Seattle, WA 98195, USA.

  • Present address: Monsanto/Searle, 800 North Lindbergh Boulevard, St. Louis, MO 631667, USA.

  • To whom correspondence should be addressed. E-mail: jedixon{at}


View Abstract

Stay Connected to Science

Navigate This Article