Pathogenic Yersinia spp. use a panel of virulence proteins that antagonize signal transduction processes in infected cells to undermine host defense mechanisms. One of these proteins, Yersinia enterocolitica outer protein P (YopP), down-regulates the NF-κB and MAPK signaling pathways, which suppresses the proinflammatory host immune response. In this study, we explored the mechanism by which YopP succeeds to simultaneously disrupt several of these key signaling pathways of innate immunity. Our data show that YopP operates upstream of its characterized eukaryotic binding partner IκB kinase-β to shut down the NF-κB signaling cascade. Accordingly, YopP efficiently impaired the activities of TGF-β-activated kinase-1 (TAK1) in infected cells. TAK1 is an important activator of the IκB kinase complex in the TLR signaling cascade. The repression of TAK1 activities correlated with reduced activation of NF-κB- as well as AP-1-dependent reporter gene expression in Yersinia-infected murine macrophages. This suggests that the impairment of the TAK1 enzymatic activities by Yersinia critically contributes to down-regulate activation of NF-κB and of MAPK members in infected host cells. The inhibition of TAK1 potentially results from the blockade of signaling events that control TAK1 induction. This process could involve the attenuation of ubiquitination of the upstream signal transmitter TNFR-associated factor-6. Together, these results indicate that, by silencing the TAK1 signaling complex, Yersinia counteracts the induction of several conserved signaling pathways of innate immunity, which aids the bacterium in subverting the host immune response.
The Gram-negative bacterium Yersinia disarms phagocytic cells and disrupts their response to infection. This is an essential characteristic in the pathogenesis of diseases mediated by human pathogenic Yersinia spp. (1). These are Y. pestis, the etiological agent of bubonic plague, as well as Y. pseudotuberculosis and Y. enterocolitica, which cause gastrointestinal symptoms and lymphadenitis. One specific Yersinia virulence protein, Yersinia outer protein P (YopP)4 in Y. enterocolitica or its homologue YopJ in Y. pestis and Y. pseudotuberculosis, impairs the activation of several key signaling pathways of innate immunity (1, 2). YopP/YopJ disrupts the NF-κB and MAPK cascades, which leads to inhibition of macrophage TNF-α production and to macrophage apoptosis. These activities of YopP/YopJ aid the bacteria to disseminate within the host and to establish prolonged systemic infection (3, 4).
The mechanisms by which YopP/YopJ exerts its effects on the host cell are not yet completely understood. YopP/YopJ is translocated by the Yersinia type III secretion system into the host cell cytoplasm, where it interacts with eukaryotic signaling molecules (1, 2). YopP/YopJ appears to engage several host cell targets to suppress activation of the NF-κB and MAPK pathways. It has been shown that YopP/YopJ binds and blocks members of the MAPK kinase (MKK) family and the NF-κB-activating IκB kinase β (IKKβ) (5). However, the molecular mechanisms leading to inactivation of these signal transmitters have not yet been fully elucidated. It has been proposed that YopP/YopJ acts as a cysteine protease on cellular proteins that are modified with the ubiquitin-like molecule small ubiquitin-related modifier-1 (SUMO-1) (6). But MKKs and IKKβ appear not to be direct targets for the putative YopP/YopJ enzymatic activity. Instead, it has been speculated that YopP/J may proteolytically silence not yet characterized signaling complexes that are associated with MAPK or NF-κB activation (6, 7).
In this study, we further constricted the mechanisms of NF-κB inhibition by YopP/YopJ. NF-κB activation is initiated through cytokine signaling, innate and adaptive immune responses, and environmental stress (8). The different signaling pathways converge at the IKK complex that consists of the two catalytic subunits IKKα und IKKβ, and the regulatory subunit NF-κB essential modulator (NEMO)/IKKγ. The activated IKK complex phosphorylates the NF-κB inhibitory IκB proteins, which sequester the transcription factor NF-κB in the cytoplasm in unstimulated cells. IκB phosphorylation leads to NF-κB liberation, which can than translocate to the nucleus and activate transcription (8, 9). The transmembrane TLRs are important inducers of the NF-κB pathway in macrophages (10). They sense an invading pathogen by recognizing conserved microbial components. Our previous studies have shown that yersiniae can signal through TLR2, the receptor for bacterial lipoproteins and lipoteichoic acid, and through the LPS receptor TLR4 (11). Stimulation of these receptors triggers association with the intracellular adapter MyD88, which in turn recruits IL-1R-associated kinase (IRAK) members (10). The phosphorylation of IRAK1 by IRAK4 is followed by the interaction with TNFR-associated factor-6 (TRAF6), which subsequently forms a cytoplasmic complex that comprises TGF-β-activated kinase-1 (TAK1) and the adapter molecules TAK1-binding protein-1 (TAB1) and TAB2 or TAB3 (10, 12, 13, 14, 15, 16). The activation of these multiprotein complexes requires the ubiquitination of TRAF6, which then induces TAK1 activation (15, 17). Activated TAK1 acts as central inducer of several critical signaling pathways of innate immunity (10, 17, 18, 19, 20, 21, 22). TAK1 phosphorylates and activates the IKK complex. Furthermore, the activation of TAK1 mediates phosphorylation of MKK members and induction of the JNK and p38 MAPK pathways (10, 17, 18, 19, 20, 21, 22). By controlling the activities of these pathways, TAK1 regulates the innate immune response to bacterial infection (10, 23, 24, 25). This appears to be a conserved function of TAK1 because also the Drosophila TAK1 homologue dTAK1 essentially contributes to mount innate immune responses to bacteria (26, 27, 28).
The data presented in this work show that Y. enterocolitica efficiently interferes with the activities of TAK1 in infected mammalian cells through the action of YopP. The down-regulation of the TAK1 activity in infected macrophages leads to impairment of the induction of NF-κB- and AP-1-dependent reporter genes. Our data suggest that, by repressing the activity of TAK1, Yersinia strikes a key signal transducer of innate immunity. The inactivation of the TAK1 signaling complex contributes to suppress the proinflammatory response in Yersinia-infected host cells.
Materials and Methods
Yersinia strains, cell lines, and infection conditions
The Y. enterocolitica strains used in this study were the serotype O8 wild-type strain WA-314 and its isogenic yopP-knockout mutant WA-ΔyopP (29), as well as the serotype O9 YopP-negative mutant E40-ΔyopP (30) and its complemented derivative E40-ΔyopP/YopP (31). The strain E40-ΔyopP has been kindly provided by G. Cornelis (Division of Molecular Microbiology, University of Basel, Basel, Switzerland). E40-ΔyopP/YopP moderately overexpresses serotype O8 YopP as a fusion protein with the first 138 aa of YopE under control of the yopE promoter (29, 31). For infection, overnight cultures grown at 27°C were diluted 1/20 in fresh Luria-Bertani broth and grown for another 2 h at 37°C (29). Shift of the growth temperature to 37°C initialized activation of the Yersinia type III secretion machinery for efficient translocation of Yops into the host cell upon cellular contact. To synchronize infection, bacteria were seeded on the cells by centrifugation at 400 × g for 5 min at a ratio of 20 bacteria/cell. For incubation times longer than 90 min, bacteria were killed by the addition of gentamicin (100 μg/ml) after 90 min. The human embryonic kidney (HEK) 293 cell line was cultured in DMEM cell growth medium containing 10% heat-inactivated FCS (11). Murine J774A.1 macrophages were grown in RPMI 1640 medium supplemented with 10% heat-inactivated FCS and 5 mM l-glutamine (29).
Cell transfection and measurement of NF-κB and AP-1 activation
HEK293 cells were seeded in 24-well cell culture plates and transfected with human cDNA constructs for Flag-TRAF6 or Flag-IKKβ (50 ng of each), together with endothelial leukocyte adhesion molecule (ELAM)-1-NF-κB firefly luciferase (100 ng) and phRL-null Renilla luciferase reporter vectors (100 ng; Promega) by the calcium-phosphate transfection method, as described (11). The Flag-TRAF6 plasmid has been previously described (32). The Flag-IKKβ plasmid was provided by Tularik (33). The total amount of DNA was kept constant with empty vector for each transfection (250 ng). Twenty-four hours after transfection, HEK293 cells were infected with yersiniae for 18 h, as indicated. J774A.1 cells were transfected with ELAM-1-NF-κB or AP-1 luciferase plasmid (100 ng), the phRL-null vector (100 ng), and either empty control vector, a kinase-inactive hemagglutinin (HA)-TAK1(K63W) expression plasmid (300 ng of each), mouse TAK1, or control small-interfering RNA (siRNA) (Santa Cruz Biotechnology; 300 ng of each) using the ExGen 500 transfection reagent, according to the manufacturer’s instructions (Fermentas). The AP-1 luciferase reporter (Stratagene) and the HA-TAK1 and HA-TAK1(K63W) plasmids were kindly provided by K. Conzelmann (Genzentrum, Max von Pettenkofer-Institute, Munich, Germany), and K. Matsumoto and T. Ishitani (Department of Molecular Biology, Graduate School of Science, Nagoya University, Nagoya, Japan) (21), respectively. Twenty hours after transfection, J774A.1 cells were infected, as indicated. Cells were lysed for measurement of NF-κB- or AP-1-dependent reporter activation after 6 or 20 h. The luciferase activities were measured with a microtiter plate chemiluminometer (Berthold Technologies), according to the manufacturer’s instructions (Dual-Luciferase Reporter System; Promega) (11). The NF-κB- and AP-1-directed firefly luciferase activities were normalized to Renilla luciferase activities to compensate differences in the transfection efficiencies. Data on NF-κB and AP-1 activities are the means ± SD of five to six independent experiments. Statistical analysis was performed using Student’s t test.
Immunoprecipitation, kinase assays, and Western immunoblotting
For the detection of transiently overexpressed Flag-TRAF6, Flag-IKKβ, or HA-TAK1(K63W), the cells were transfected with 0.5 μg of DNA of the respective plasmid and lysed 20 h later with lysis buffer containing 10 mM HEPES (pH 7.8), 10 mM KCl, 2 mM MgCl2, 0.1 mM EDTA, 0.5% Nonidet P-40, 1 mM DTT, and phosphatase and protease inhibitors (Roche). In the RNA interference experiments, 1 μg of mouse TAK1 siRNA or control siRNA (Santa Cruz Biotechnology) was cotransfected together with 0.5 μg of HA-TAK1 and HA-NEMO, or 1 μg of empty control plasmid, as indicated. The lysates were separated by SDS-PAGE, transferred to polyvinylidene difluoride (PVDF) membrane, and probed with mAbs directed against the epitope tags of the respective constructs (anti-Flag, Sigma-Aldrich; anti-HA, Roche). Immunoreactive bands were visualized using appropriate secondary Abs and ECL detection reagents (Amersham Biosciences). The effects of YopP overexpression on ubiquitination of TRAF6 and NEMO were analyzed by transfection of eukaryotic YopP expression vectors together with an HA-ubiquitin plasmid and equal amounts of Flag-TRAF6 or HA-NEMO plasmids (32, 34) into HEK293 cells (0.5 μg of each plasmid). Two different types of YopP expression plasmids were used, either encoding wild-type YopP from Y. enterocolitica serotype O8 (Flag-YopP), or YopP with a point mutation in the putative YopP cystein protease motif (Flag-YopP(C172A)) (6). The ubiquitin expression vector encoding an octameric tandem fusion of HA-ubiquitin was kindly provided by M. Treier (European Molecular Biology Laboratory, Heidelberg, Germany). The total amount of DNA in control transfections without YopP and/or ubiquitin plasmid was kept constant with empty vector (1.5 μg). Cellular lysates were prepared and processed with monoclonal anti-Flag or polyclonal anti-NEMO Abs (34), as indicated above.
For the in vitro kinase assays, cells were first treated with the lysis buffer described above. The cellular lysates were then diluted 1/1 with buffer TN (20 mM Tris (pH 7.5), 200 mM NaCl, 1 mM DTT, phosphatase and protease inhibitors) and incubated with anti-Flag or anti-HA Abs to precipitate overexpressed Flag-IKKβ or HA-TAK1, respectively (29). Endogenous TAK1 from J774A.1 macrophages was precipitated by using polyclonal anti-TAK1 Abs (Santa Cruz Biotechnology). The immune complexes were collected with protein A/G-agarose (Santa Cruz Biotechnology), washed three times with precipitation and then kinase buffer, and subsequently subjected to kinase assay. The IKKβ kinase reactions were performed with 1 μg of rGST-IκΒα for 30 min at 30°C in the presence of 20 μM ATP (4 μCi of [γ-32P]ATP per sample) and IKKβ kinase buffer (20 mM HEPES (pH 8), 10 mM MgCl2, 50 mM NaCl, 2 mM DTT, phosphatase and protease inhibitors) (29). The plasmid encoding GST-IκΒα was kindly provided by U. Siebenlist (National Institute of Allergy and Infectious Diseases, Bethesda, MD). The TAK1 kinase assays were conducted on 1 μg of rHis-MKK6 for 30 min at 30°C in the presence of 20 μM ATP (4 μCi of [γ-32P]ATP per sample) and TAK1 kinase buffer (10 mM HEPES (pH 7.4), 5 mM MgCl2, 1 mM DTT, phosphatase and protease inhibitors) (35). The His-MKK6 plasmid was kindly provided by E. Nishida (Department of Cell and Developmental Biology, Graduate School of Biostudies, Kyoto University, Kyoto, Japan) (35, 36). The in vitro phosphorylation of endogenous TAK1 was assayed in the absence of His-MKK6. Proteins were separated by SDS-PAGE and electrotransferred to PVDF membrane. The membranes were first analyzed by autoradiography and subsequently immunoblotted with anti-Flag, anti-HA, or polyclonal anti-TAK1 or anti-IKKβ serine-181 Abs (Cell Signaling Technology) to determine the amounts of precipitated Flag-IKKβ, HA-TAK1, or endogenous TAK1, respectively. When required, the phosphotransferase activities were quantified by densitometry using AIDA image analyzer (Raytest). Statistical analysis was performed using Student’s t test.
To determine binding of TAK1 and TRAF6 to TAB2, HEK293 cells were transfected with TAB2 (0.3 μg), TAK1, TRAF6, and ubiquitin expression vectors (0.5 μg of each). The T7 epitope-tagged TAB2 expression vector was kindly provided by K. Matsumoto and T. Ishitani (Department of Molecular Biology, Graduate School of Science, Nagoya University, Nagoya, Japan) (12, 14). Eighteen hours after transfection, the cells were infected with yersiniae and then lysed with a buffer containing 20 mM HEPES (pH 7.4), 150 mM NaCl, 0.5% Triton X-100, 1.5 mM MgCl2, 2 mM EGTA, 2 mM DTT, and protease and phosphatase inhibitors (14). Immunoprecipitations were performed with anti-T7 mAb (Novagen-Merck), as described above. The precipitates were subjected to Western immunoblotting, using anti-HA Abs to detect HA-TAK1, and subsequently anti-Flag Abs to detect Flag-TRAF6. The precipitation of TAB2 was assessed in parallel control samples with anti-T7 Ab. The data shown are representative for one experiment of at least three performed.
YopP acts upstream of IKKβ to suppress the activation of NF-κB
To assess at which signaling step YopP initiates its inhibitory effect on the NF-κB cascade, we cotransfected HEK293 cells with a NF-κB-dependent luciferase reporter vector and expression plasmids for either Flag-IKKβ or Flag-TRAF6, which acts upstream of IKKβ. The transfection with either construct resulted in noticeable stimulation of NF-κB reporter activity, which indicates that overexpression of IKKβ as well as of TRAF6 confers constitutive NF-κB activation (Fig. 1 A). The cells were subsequently infected with the Y. enterocolitica serotype O8 wild-type strain WA-314 or the corresponding YopP-negative mutant WA-ΔyopP. The YopP-negative mutant WA-ΔyopP mediated prominent NF-κB signals in control vector-, as well as in IKKβ- and TRAF6-transfected cells. The activation of NF-κB in these conditions is predominantly conferred by the Yersinia outer surface protein invasin (11), which stimulates the NF-κB pathway in HEK293 cells through the binding of β1 integrins. The wild-type strain WA-314 suppressed infection-related activation of NF-κB that was observed for the YopP-negative mutant. This confirms the known inhibitory effect of YopP on the NF-κB pathway (1, 2, 29). Interestingly, wild-type yersiniae were unable to impair the IKKβ-dependent constitutive NF-κB response, whereas basal NF-κB activation conferred by TRAF6 overexpression was markedly reduced. This implies that YopP does not directly act on IKKβ to inhibit NF-κB activation, but it appears to operate more upstream to shut down the NF-κB signaling pathway. YopP could potentially inhibit a signaling event that mediates IKKβ activation downstream of TRAF6.
To directly correlate the results of the NF-κB luciferase reporter assays with the IKKβ catalytic activities, we performed in vitro kinase assays with overexpressed and immunoprecipitated IKKβ (Fig. 1,B). The cells were infected in these experiments with the YopP-negative mutant E40-ΔyopP or its complemented derivative E40-ΔyopP/YopP. Compared with the wild-type strain WA-314, E40-ΔyopP/YopP produces YopP under moderate overexpression conditions (29, 31). The YopP-negative mutants E40-ΔyopP and WA-ΔyopP are functionally equivalent (31). The E40-ΔyopP mutant induced an increase in the phosphotransferase activities toward rGST-IκBα by IKKβ after 45 min of stimulation, indicating IKKβ activation (Fig. 1,B, upper panel). IKKβ remained active in these conditions for at least 90 min. In contrast, the YopP-producing strain E40-ΔyopP/YopP prevented conspicuous activation of IKKβ. However, the kinase activity was still markedly elevated compared with basal levels. This supports the concept that YopP impairs the induction of IKKβ by upstream signal transmitters, but not constitutive IKKβ activities (7). This conclusion is in line with the results of previous IKK kinase assays conducted on macrophages (29, 31). Interestingly, overexpressed IKKβ was itself phosphorylated in the in vitro kinase assays (Fig. 1,B, upper panel). There, we noticed an increase in IKKβ phosphorylation after 45 min of infection, which was abolished by the YopP-producing strain within 90 min. Because it has been reported that the activation of IKKβ depends on phosphorylation of the serine residues 177 and 181 in the IKKβ activation loop (8, 9), we probed the membrane with an Ab that recognizes phosphorylated serine-181 of IKKβ. In fact, this Ab displayed reduced binding to IKKβ after infection with strain E40-ΔyopP/YopP (Fig. 1 B, lower panel). This suggests that YopP deactivates signaling events that mediate IKKβ phosphorylation. Because upstream kinases have been implicated in IKKβ activation (8, 9, 10), the observed down-regulation of IKKβ phosphorylation by YopP may result from the impairment of IKKβ-inducing enzymes.
YopP suppresses the TAK1 activity in Yersinia-infected cells
TAK1 is a member of the MAPK/ERK kinase kinase (MEKK) family, which has been shown to be essential for TLR/IL-1-induced NF-κB activation (10, 22, 23, 24, 25). TAK1 operates upstream of IKKβ, but downstream of TRAF6 in the NF-κB cascade (12, 13, 14, 15, 16, 17, 18, 19, 20, 21). Because TAK1 can directly phosphorylate IKKβ in the activation loop, thereby promoting IKKβ activation (17), we investigated the impact of Y. enterocolitica on the TAK1 activity in infected cells. Accordingly, we immunoprecipitated overexpressed TAK1 from transfected HEK293 cells and analyzed the ability of TAK1 to phosphorylate rHis-tagged MKK6. In these experiments, immunoprecipitated TAK1 displayed substantial phosphotransferase activities toward its substrate even in untreated cells, indicating a prominent basal activity of overexpressed TAK1 (Fig. 2 A). A 60-min infection with the YopP-negative strain E40-ΔyopP noticeably intensified MKK6 phosphorylation. Thus, Y. enterocolitica infection additionally induces TAK1. Interestingly, the YopP-producing strain E40-ΔyopP/YopP considerably suppressed the activity of overexpressed TAK1. The TAK1 activities were actually reduced to less than background levels in E40-ΔyopP/YopP-infected cells. This reveals a remarkable capacity of YopP to suppress the enzymatic activity of TAK1 during infection.
To find out whether yersiniae impair TAK1 also in macrophages, which are supposed to be primary target cells of the immunomodulatory action of the Yersinia type III protein secretion system (1), we analyzed the activities of endogenous TAK1 in J774A.1 macrophages infected with the YopP-producing Yersinia strains E40-ΔyopP/YopP and WA-314 in comparison with their YopP-negative counterparts E40-ΔyopP and WA-ΔyopP, respectively. Similar to overexpressed TAK1, endogenous TAK1 also displayed constitutive kinase activity (Fig. 2,B; untreated cells). Nevertheless, 15 min of infection with all of the investigated Yersinia strains slightly enhanced the ability of macrophage TAK1 to phosphorylate MKK6 (Fig. 2,B; 15 min data). This indicates that Yersinia infection can trigger TAK1 activation in macrophages. However, after additional 15 min of infection, the YopP-producing strains E40-ΔyopP/YopP and WA-314 significantly reduced the activities of TAK1 below basal levels, whereas TAK1 remained active in E40-ΔyopP- and WA-ΔyopP-infected cells (Fig. 2 B; 30 min data). This demonstrates that yersiniae expeditiously and efficiently down-regulate the basal as well as infection-induced TAK1 activities in macrophages through the action of YopP. This could imply that the repression of TAK1 is a critical event in the subversion of eukaryotic innate immune signaling pathways. The fact that the suppression of the TAK1 activities occurred only 15 min after initial induction is explained by a certain delay necessary for YopP translocation and direction to its targets (1).
The TAK1 activity is important for the induction of NF-κB and AP-1 in Yersinia-infected macrophages
We next wondered what the consequences of TAK1 impairment on proinflammatory signaling in Yersinia-infected macrophages might be. We used an expression vector encoding a kinase-inactive TAK1 version (TAK1(K63W)), whose overexpression could mimic the repression of the TAK1 catalytic activity by YopP. The TAK1(K63W) construct was cotransfected with an NF-κB- or an AP-1-dependent luciferase reporter plasmid into J774A.1 macrophages. NF-κB- and AP-1-driven reporter activation was then monitored after infection with the YopP-negative Yersinia strain E40-ΔyopP. The NF-κB and AP-1 reporters give indirect indication on activation of the IKK-NF-κB or MAPK signaling pathways, respectively (37, 38, 39). The AP-1-dependent reporter gene can respond to induction of the JNK, ERK1/2, or p38 MAPK pathways in dependence on the composition of the AP-1 dimer (37, 38). Fig. 3 A demonstrates that strain E40-ΔyopP triggered substantial activation of both the NF-κB- and AP-1-dependent reporter genes in macrophages that were transfected with the empty control vector. This reflects the ability of Yersinia to induce signals that mediate NF-κB and AP-1 activation in the absence of YopP. The YopP-producing strain E40-ΔyopP/YopP counteracted these signals, which complies with the described abilities of Yersinia to down-regulate NF-κB and MAPK activation (1, 2, 40, 41). Interestingly, transfection of enzymatically inactive TAK1 also impaired NF-κB and AP-1 activation. Accordingly, NF-κB- and AP-1-dependent reporter activities were substantially diminished in macrophages transfected with TAK1(K63W) and infected with YopP-negative yersiniae. The reduction of the NF-κB- and AP-1-related signals was not a consequence of onset of apoptosis, because >90% of the TAK1(K63W)-transfected cells were viable at the time points investigated (data not shown) (29). This indicates that the TAK1 activity plays an important role in the control of the induction of NF-κB and AP-1 in Yersinia-infected macrophages. This in turn suggests that the repression of TAK1 by YopP leads to impairment of the activities of IKKβ and MAPK members, which could concomitantly prevent NF-κB and AP-1 activation.
To provide further evidence for a role of endogenous TAK1 as a regulator of the inflammatory response in Yersinia-infected cells, we analyzed NF-κB reporter gene activation in J774A.1 macrophages that were transfected with TAK1 siRNA. The transfection of TAK1 siRNA significantly diminished the NF-κB response induced by the YopP-negative strain E40-ΔyopP (Fig. 3,B). To analyze the TAK1 protein expression levels in the transfected cells, we performed immunoblot experiments. Because of a relatively low transfection efficiency in J774A.1 macrophages (<10%) (29), we were unable to detect a clear down-regulation of the expression of endogenous TAK1 in J774A.1 cells (data not shown). However, the TAK1 siRNA preparation effectively inhibited the expression of mouse HA-TAK1 in HEK293 cells that were cotransfected with mouse HA-TAK1 and HA-NEMO expression plasmids (Fig. 3,C). The coexpression of HA-NEMO was not affected (Fig. 3 C). This suggests that the TAK1 siRNA reagent could specifically impair the expression of TAK1. The reduction of the NF-κB response by TAK1 RNA interference supports the concept that TAK1 acts as a mediator of proinflammatory signaling in Yersinia-infected macrophages.
YopP potentially disrupts signaling events involved in TAK1 activation
We next wondered how Yersinia could impair the activity of TAK1. One possibility is that Yersinia interferes with the induction of TAK1 by an upstream activator. However, the molecular mechanisms that mediate TAK1 activation are not yet completely understood. There are indications that TAK1 requires phosphorylation to be active (13, 14, 15, 16, 35, 42, 43). The TAK1 phosphorylation events appear to result from autophosphorylation (35, 42, 43), although other yet unknown kinases could also be involved (13). It has been shown that phosphorylated TAK1 displays slower electrophoretic mobility, resulting in a shift of TAK1 immunoreactive bands in SDS-PAGE (12, 13, 14, 16, 23, 35, 42, 43). When we prepared lysates from TAK1-transfected cells that were infected with the YopP-negative strain E40-ΔyopP, we similarly observed TAK1 bands with reduced electrophoretic mobilities (Fig. 4,A). The appearance of the slower migrating bands was inhibited by the YopP-producing strain E40-ΔyopP/YopP. This could indicate that YopP interferes with TAK1 phosphorylation, which may contribute to impair the activity of TAK1. Slower migrating TAK1 immunoreactive bands became visible also in the TAK1 immunoprecipitates that were subjected to the in vitro kinase assays (Fig. 4,B, lower panel). The inhibition of the TAK1 mobility shift by YopP correlated in these experiments with reduced phosphorylation of TAK1 itself (Fig. 4,B, upper panel). This suggests that the slower electrophoretic mobilities of TAK1 result from phosphorylation events that are blocked by YopP. In fact, the inhibition of TAK1 phosphorylation by YopP was not restricted to overexpressed TAK1, but occurred also on endogenous TAK1 that was immunoprecipitated from Yersinia-infected J774A.1 macrophages (Fig. 4 C). In these experiments, strain E40-ΔyopP/YopP substantially suppressed the phosphotransferase activities toward TAK1 within 30 min of infection, an effect that was not yet obvious after 15 min or after infection with strain E40-ΔyopP. This indicates that the activity of YopP shuts down a signal transduction process that controls TAK1 phosphorylation.
More recently, it was revealed that TAK1 activation upstream of phosphorylation requires the formation of a complex with TRAF6 and TAB2 or TAB3 (12, 13, 14, 15). The assembly of this complex involves TRAF6-dependent polyubiquitination reactions. TRAF6, together with the ubiquitin-conjugating enzyme Ubc13 and the Ubc-like protein Uev1A, functions as a ubiquitin ligase that links the adapters TAB2 and TAB3 with polyubiquitin chains. Furthermore, TRAF6 is itself a target of ubiquitination (14, 15, 17, 44). Polyubiquitinated TRAF6 associates with TAB2 and TAB3, which mediates recruitment of TAK1. The oligomerization of TAK1 in the resulting signaling complexes may then facilitate TAK1 phosphorylation and activation of the TAK1 kinase activity (14, 15, 19, 20). Because YopP appears to hamper a signaling event that mediates TAK1 induction, we wondered whether yersiniae could interfere with the formation of a TAK1-TAB2-TRAF6-comprising signaling complex. We overexpressed HA-TAK1, T7-TAB2, and Flag-TRAF6 in HEK293 cells and analyzed association of TAK1 and TRAF6 with immunoprecipitated TAB2 after Yersinia infection. We included an expression plasmid for ubiquitin in these experiments because the interaction between these molecules has been shown to be stabilized by polyubiquitination (15). Fig. 5 shows that TRAF6 was detected in the anti-T7 immunoprecipitates independently whether the cells were untreated or infected with YopP-producing (E40-ΔyopP/YopP) or YopP-negative yersiniae (E40-ΔyopP). This suggests that overexpressed TRAF6 binds to TAB2 and this interaction appears not to be critically affected by Yersinia. However, TAK1 accumulated in the anti-T7 immune precipitates after infection with YopP-negative E40-ΔyopP, an effect that was less pronounced after infection with the YopP-producing strain E40-ΔyopP/YopP. This suggests that YopP could counteract the recruitment of TAK1 to TAB2 and consequently to TRAF6. This may then lead to inhibition of TAK1 activation. The successful precipitation of TAB2 was evaluated in parallel control samples (Fig. 5, lower panel) because the simultaneous detection of TAB2 and TAK1 in the same immunoprecipitates was complicated by the same m.w. of TAK1 and TAB2. Together, these overexpression experiments confirm that both TRAF6 and TAK1 can interact with TAB2 (12, 13, 14, 15). Translocated YopP apparently interferes with the stable formation of this multiprotein complex.
YopP overexpression modifies TRAF6 and NEMO polyubiquitination
The polyubiquitination of TRAF6 is a critical step in TRAF6-TAB2/TAB3 assembly and TAK1 activation (14, 15, 17, 44). Thus, we additionally assessed the impact of YopP on the ubiquitination of TRAF6. To generate polyubiquitinated TRAF6 molecules, we cotransfected HEK293 cells with the TRAF6 and the ubiquitin expression plasmid. This resulted in TRAF6 polyubiquitination, as evidenced by immunoblot analysis of cellular lysates (Fig. 6,A). When the transfected cells were infected with the different Yersinia strains according to the experiments described above, however, we were unable to reveal a reliable effect of YopP on TRAF6 ubiquitination although trying different experimental conditions (data not shown). For this reason, we additionally analyzed the cellular effects of YopP overexpression using a eukaryotic YopP expression plasmid. When TRAF6 was cotransfected with the YopP expression vector, the appearance of the TRAF6 polyubiquitin bands was considerably decreased (Fig. 6,A; Flag-YopP wild type (WT)). This indicates that overexpressed YopP is able to interfere with stable formation of TRAF6 polyubiquitin chains. This effect apparently requires the proposed enzymatic activity of YopP (6), because the inhibition of TRAF6 ubiquitination was not as obvious for a YopP expression version with a point mutation in the putative YopP/YopJ cysteine protease motif (Fig. 6,A; Flag-YopP(C172A)). The C172A mutation generates an inactive YopP and yersiniae that produce C172-mutagenized YopP functionally behave like YopP-negative mutants with respect to NF-κB and MAPK inhibition and apoptosis induction (6, 45) (data not shown). However, the inhibitory effect of YopP overexpression on polyubiquitination was not restricted to TRAF6, but occurred moderately also on NEMO/IKKγ (Fig. 6 B). This suggests that overexpressed YopP could potentially modify the ubiquitination of several signal transducers that are involved in NF-κB activation. This idea is in line with the results of previous publications that have indicated that YopP/YopJ overexpression attenuates the conjugation of cellular proteins with ubiquitin-like SUMO-1 (6) and the ubiquitination of IKKβ (46). This could give reason to speculations that YopP/YopJ may impair cellular signaling by antagonizing the ubiquitin modification of several critical signal transducers.
TAK1 is a central signal transducer of innate immunity (10). This function of TAK1 has been conserved during evolution (26, 27, 28). In mammals, the activity of TAK1 links TLR/IL-1-responsive signals to activation of the NF-κB and the JNK and p38 MAPK signaling pathways, which critically determine the host immune response to bacterial infection (10, 17, 20, 21). The data presented in this work show that YopP from Y. enterocolitica efficiently down-regulates the activity of TAK1 in infected cells. The repression of TAK1 by YopP contributes to impair the NF-κB and MAPK pathways in Yersinia-infected macrophages. This suggests that Yersinia can simultaneously shut down several key signaling pathways of innate immunity by targeting TAK1. The impairment of TAK1 is in line with a number of studies that have indicated that YopP/YopJ apparently does not directly act on its hitherto characterized eukaryotic binding partners (6, 7). These studies and our data imply that YopP/YopJ may use IKKβ, and possibly MKK members, as shuttles to be directed to its actual target molecules. TAK1 could be such preferred target, whose inactivation may suppress the induction of several downstream kinases that are essential for innate immunity. In fact, IKKβ is organized in a large multiprotein complex and association with TAK1 becomes possible through the interaction with upstream signaling complexes upon receptor stimulation (21, 22, 47, 48). Thus, YopP may prevent the generation of active signaling complexes by binding to IKKβ and by subsequently suppressing the TAK1 activity.
Although TAK1 is a major IKK kinase in the TLR pathway that also plays essential role in the activation of JNK and p38, it is important to note that several other MEKK family members have been implicated in IKK or MAPK activation upon TLR induction (8, 9, 10). This includes the kinases MEKK1, MEKK3, and tumor progression locus-2 (TPL-2/Cot) (49, 50, 51). TPL-2/Cot confers activation of the ERK MAPK pathway in LPS-treated macrophages (51). The ERK pathway is apparently not responsive to TAK1 induction (17, 24), but is sensitive to inhibition by Yersinia infection similar to the p38 and JNK pathways (1, 2, 40). This suggests that other members of the MEKK family besides TAK1 may similarly be affected by YopP/YopJ. TPL-2/Cot could be a potential target molecule, whose inactivation would cause inhibition of ERK induction. However, TAK1 is not only of critical importance for the activation of TLR/IL-1-responsive immune effector pathways, but also pivotally participates in TNF-α-, receptor activator of NF-κB ligand-, TCR-, and TGF-β superfamily member-dependent signal transduction (22, 52, 53, 54). Our previous studies have shown that TNF-α-mediated NF-κB activation in HeLa cells is also suppressed by Yersinia infection (41). This suggests that the inhibition of TAK1 by YopP may be relevant also in other cell types and could be a general mechanism that helps Yersinia to block the proinflammatory host response.
The mechanism, by which YopP suppresses TAK1 activities, appears less clear. Our data indicate that YopP interferes with TAK1 phosphorylation, mediated either by TAK1 itself or by another yet unidentified kinase. Because TAK1 phosphorylation appears to be an essential modification step for the induction of the TAK1 enzymatic activity (13, 14, 15, 16, 35, 42, 43), YopP may block an upstream signaling event that regulates TAK1 activation. In fact, our experiments indicate that YopP can interfere with the assembly of a TRAF6-TAB2-TAK1 multiprotein complex. The formation of this complex is controlled by polyubiquitination reactions that involve the TRAF6 ubiquitin ligase (14, 15, 17, 44). The modification of TRAF6 and TAB2/TAB3 by polyubiquitin chains mediates the oligomerization and stabilization of the TRAF6-TAB2/TAB3-TAK1 complexes, which eventually triggers TAK1 activation (14, 15, 19, 20). Our data suggest that YopP could block the formation of this activator complex by preventing the effective recruitment of TAK1. This could then impair TAK1 induction. Thus, it is tempting to speculate that YopP interferes with TRAF6-dependent polyubiquitination reactions to suppress TAK1 induction. In fact, the overexpression of YopP was able to remarkably decrease the cellular concentration of TRAF6- and also of NEMO-ubiquitin conjugates. A similar effect of YopP/YopJ has been previously reported on the ubiquitination of IKKβ (46) and on nonspecified cellular proteins modified with the ubiquitin-like molecule SUMO-1 (6). These results were ascribed to the putative cysteine protease activity of YopP/YopJ, which could mediate cleavage of ubiquitin-like molecules (6, 46). However, we could not reveal a reliable effect of YopP on the ubiquitination of TRAF6, NEMO, or TAB2 in Yersinia-infected cells (data not shown). This could indicate that a discreet effect of YopP on the ubiquitination of these molecules is masked in infection conditions. Or alternatively, neither TRAF6, nor NEMO, nor TAB2 are the major cellular targets of the proposed YopP cysteine protease activity during infection. Other potential targets could be TAB3 or receptor-interacting protein-1, whose ubiquitination was shown to be involved in NF-κB activation (14, 15, 55). Furthermore, it was demonstrated that overexpressed YopP/YopJ not only affects the overall cellular content of SUMO-1-conjugated proteins, but also the cellular levels of free SUMO-1 (6). We similarly observed a reduction of free HA-ubiquitin in cells that overexpressed YopP (data not shown). This could indicate that the interference of overexpressed YopP/YopJ with the ubiquitination of eukaryotic molecules involves several distinct inhibitory mechanisms. Future in vitro and in vivo experiments will help to specify by which molecular activities YopP represses the induction of TAK1 to impair the innate immune response of macrophages.
We thank Martin Aepfelbacher and Jürgen Heesemann for constructive discussions and helpful advice. We furthermore thank Konrad Trülzsch for critical reading of the manuscript. Moreover, we thank G. R. Cornelis for supplying us with Yersinia strains, as well as K. Matsumoto, T. Ishitani, E. Nishida, K. K. Conzelmann, M. Treier, U. Siebenlist, and Tularik for providing us with expression vectors.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (Grants DFG Ru788/1 and 2).
Abbreviations used in this paper: Yop, Yersinia outer protein; ELAM, endothelial leukocyte adhesion molecule; HA, hemagglutinin; HEK, human embryonic kidney; IKK, IκB kinase; IRAK, IL-1R-associated kinase; MEKK, MAPK/ERK kinase kinase; MKK, MAPK kinase; NEMO, NF-κB essential modulator; PVDF, polyvinylidene difluoride; siRNA, small-interfering RNA; SUMO, small ubiquitin-related modifier; TAB, TAK1 binding protein; TAK1, TGF-β-activated kinase-1; TPL-2, tumor progression locus-2; TRAF6, TNFR-associated factor-6.