Article13 December 2007free access A critical role of RICK/RIP2 polyubiquitination in Nod-induced NF-κB activation Mizuho Hasegawa Mizuho Hasegawa Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Yukari Fujimoto Yukari Fujimoto Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Search for more papers by this author Peter C Lucas Peter C Lucas Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Hiroyasu Nakano Hiroyasu Nakano Department of Immunology, Juntendo University School of Medicine, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Koichi Fukase Koichi Fukase Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Search for more papers by this author Gabriel Núñez Gabriel Núñez Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Comprehensive Cancer Center, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Naohiro Inohara Corresponding Author Naohiro Inohara Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Department of Biochemistry 2nd, Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Chuou, Yamanashi, Japan Search for more papers by this author Mizuho Hasegawa Mizuho Hasegawa Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Yukari Fujimoto Yukari Fujimoto Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Search for more papers by this author Peter C Lucas Peter C Lucas Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Hiroyasu Nakano Hiroyasu Nakano Department of Immunology, Juntendo University School of Medicine, Bunkyo-ku, Tokyo, Japan Search for more papers by this author Koichi Fukase Koichi Fukase Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan Search for more papers by this author Gabriel Núñez Gabriel Núñez Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Comprehensive Cancer Center, The University of Michigan Medical School, Ann Arbor, MI, USA Search for more papers by this author Naohiro Inohara Corresponding Author Naohiro Inohara Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA Department of Biochemistry 2nd, Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Chuou, Yamanashi, Japan Search for more papers by this author Author Information Mizuho Hasegawa1, Yukari Fujimoto2, Peter C Lucas1, Hiroyasu Nakano3, Koichi Fukase2, Gabriel Núñez1,4 and Naohiro Inohara 1,5 1Department of Pathology, The University of Michigan Medical School, Ann Arbor, MI, USA 2Department of Chemistry, Graduate School of Science, Osaka University, Toyonaka, Osaka, Japan 3Department of Immunology, Juntendo University School of Medicine, Bunkyo-ku, Tokyo, Japan 4Comprehensive Cancer Center, The University of Michigan Medical School, Ann Arbor, MI, USA 5Department of Biochemistry 2nd, Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Chuou, Yamanashi, Japan *Corresponding author. Department of Pathology, Comprehensive Cancer Center, University of Michigan Medical School, 1500 E. Medical Center Dr., C574 MSRB2, Ann Arbor, MI 48109, USA. Tel.: +1 734 936 3317; Fax: +1 734 647 9654; E-mail: [email protected] The EMBO Journal (2008)27:373-383https://doi.org/10.1038/sj.emboj.7601962 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Nod1 and Nod2 are intracellular proteins that are involved in host recognition of specific bacterial molecules and are genetically associated with several inflammatory diseases. Nod1 and Nod2 stimulation activates NF-κB through RICK, a caspase-recruitment domain-containing kinase. However, the mechanism by which RICK activates NF-κB in response to Nod1 and Nod2 stimulation is unknown. Here we show that RICK is conjugated with lysine-63-linked polyubiquitin chains at lysine 209 (K209) located in its kinase domain upon Nod1 or Nod2 stimulation and by induced oligomerization of RICK. Polyubiquitination of RICK at K209 was essential for RICK-mediated IKK activation and cytokine/chemokine secretion. However, RICK polyubiquitination did not require the kinase activity of RICK or alter the interaction of RICK with NEMO, a regulatory subunit of IκB kinase (IKK). Instead, polyubiquitination of RICK was found to mediate the recruitment of TAK1, a kinase that was found to be essential for Nod1-induced signaling. Thus, RICK polyubiquitination links TAK1 to IKK complexes, a critical step in Nod1/Nod2-mediated NF-κB activation. Introduction Members of the nucleotide-binding domain-like receptor (NLR) protein family are found in both mammals and plants (Inohara et al, 2005; Fritz et al, 2006). Several mammalian NLRs, like plant NLRs, are involved in host resistance against microbial pathogens and genetically associated with human immune diseases (Inohara et al, 2005; Fritz et al, 2006). Two NLR family members, Nod1 and Nod2, play an essential role in the recognition of specific bacterial peptidoglycan (PGN)-related molecules (Chamaillard et al, 2003; Inohara et al, 2003; Girardin et al, 2003a, 2003b). Nod1 and Nod2 mediate transcriptional activation of innate immune genes including antimicrobial peptides, proinflammatory cytokines and chemokines, that recruit immune cells to the sites of microbial stimulation (Chamaillard et al, 2003; Kobayashi et al, 2005; Boughan et al, 2006; Marks et al, 2006; Masumoto et al, 2006; Voss et al, 2006). Although the precise structure of the bacterial molecules that stimulate NLR signaling remains largely unknown, the core recognition moieties of Nod1 and Nod2 have been identified as γ-glutamyl-meso-diaminopimelic acid (iE-DAP), and muramyl dipeptide (MDP), respectively (Chamaillard et al, 2003; Girardin et al, 2003a, 2003b; Inohara et al, 2003). Previous studies using synthetic and natural PGN-related molecules containing MDP and/or iE-DAP revealed that these molecules induce/enhance innate and adaptive immune responses (Goto and Aoki, 1987; Fritz et al, 2007). The induction of immune response genes through Nod1 and Nod2 signaling largely depends on NF-κB, a proinflammatory transcriptional factor (Inohara et al, 1999, 2000; Ogura et al, 2001; Masumoto et al, 2006). Transcriptional activation of proinflammatory genes is initiated by nuclear translocation of NF-κB following its release from the IκB proteins (Karin and Ben-Neriah, 2000; Hoffmann and Baltimore, 2006). Previous studies demonstrated that RICK (RIP2/CARDIAK/RIPK2) is a critical downstream mediator of Nod1 and Nod2 signaling (Inohara et al, 1999; Ogura et al, 2001; Chin et al, 2002; Kobayashi et al, 2002; Park et al, 2007). RICK is composed of N-terminal kinase and C-terminal caspase-recruitment domains (CARD) linked via an intermediate (IM) region (Inohara et al, 1998; McCarthy et al, 1998; Thome et al, 1998). RICK physically associates with Nod1 and Nod2 through CARD–CARD interaction and activates IκB kinase (IKK), leading to phosphorylation and degradation of IκBs (Inohara et al, 1999, 2000; Ogura et al, 2001). Functional studies revealed that close proximity of RICK molecules by self-oligomerization of Nod1 and Nod2 is important for NF-κB activation (Inohara et al, 2000). RICK is known to interact with NEMO (IKKγ/IKBKG) and this interaction is essential for NF-κB activation in RICK-mediated signaling (Inohara et al, 2000). Furthermore, Nod2 stimulation with MDP induces NEMO polyubiquitination and this event is required for optimal NF-κB activation (Abbott et al, 2004). However, the molecular mechanism by which the IKK complex is activated via RICK to induce NF-κB activation in response to NLR stimulation is unknown. Here we show that RICK is conjugated with K63-linked polyubiquitin chains upon Nod1 and Nod2 stimulation. RICK polyubiquitination was essential for NF-κB activation and chemokine/cytokine secretion in response to Nod1 stimulation. Furthermore, we provide evidence for a bridging function of RICK ubiquitination by linking TAK1 to the IKK complex that is critical for NF-κB activation during Nod signaling. Results Nod1 and Nod2 stimulation induce lysine 63-linked polyubiquitination of RICK In order to determine the mechanism by which Nod1 stimulation induces NF-κB activation, we used mouse embryonic fibroblasts (MEFs) because we previously found that non-immune cells are much more sensitive to Nod1-stimulatory molecules than hematopoietic cells (Hasegawa et al, 2006; Masumoto et al, 2006). Stimulation of MEFs with KF1B, a synthetic specific Nod1-stimulatory compound, induced secretion of the chemokine CCL2 (Figure 1A) and phosphorylation of IκBα (Figure 1B). Such responses were dependent on RICK in that the induction of CCL2 production induced by KF1B, but not TNFα, was abolished in RICK-deficient MEFs (Figure 1A). Figure 1.Nod1 stimulation induces ubiquitination of RICK. (A) MEFs from WT and RICK-deficient mice were stimulated with 5 μg/ml KF1B (synthetic Nod1-stimulatory compound), 10 ng/ml TNFα or left alone. Twenty-four hours post-stimulation, the levels of CCL2 in culture supernatant were determined by ELISA. The results shown are given as mean±s.d. of triplicate cultures and are representative of three experiments. (B) WT or RICK-deficient MEFs were stimulated with 5 μg/ml KF1B for the indicated times. Post-stimulation, cells were lysed and immunoblotted with anti-phospho-IκBα, anti-IκBα or IKKβ Ab. (C) MEFs were stimulated with or without KF1B for the indicated times. Post-stimulation, cells were lysed and RICK was immunoprecipitated with rabbit anti-RICK Ab. Ubiquitinated (upper panel) and total (lower panel). RICK proteins in the immunoprecipitates were immunodetected by mouse monoclonal anti-RICK and anti-Ub Abs, respectively. Nonspecific signals detected with the Ab are indicated by asterisks. Download figure Download PowerPoint Recently, ubiquitination of RIP, a mediator of TNFα signaling, was found to be important for IKK activation (Ea et al, 2006; Li et al, 2006). Because our previous studies suggest functional and structural similarities between RIP and RICK, we first hypothesized that Nod1 stimulation might result in ubiquitination of RICK. As shown in Figure 1, RICK from MEFs stimulated with KF1B co-immunoprecipitated with high-molecular-weight ubiquitinated proteins (Figure 1C, upper panel). To further assess a role of ubiquitination in Nod1 and Nod2 signaling, we tested if RICK is ubiquitinated upon Nod1 and Nod2 stimulation, using human embryonic kidney (HEK) 293 cells constitutively expressing Nod1 and Nod2. Nod1- and Nod2-expressing HEK293 cells were transfected with plasmids producing Myc-RICK and HA-Ub and stimulated with specific synthetic Nod1 and Nod2 agonists, KF1B or MDP, respectively. Immunoblotting analysis revealed that RICK was polyubiquitinated after KF1B and MDP stimulation (Figure 2A). RICK polyubiquitination was detected 1 h after stimulation with Nod1 and Nod2 agonists. The polyubiquitination of RICK by Nod1 and Nod2 stimulation was specific in that it was not detected in parental HEK293 cells stimulated with TNFα (Figure 2A). The ubiquitination of RICK is not due to overexpression of proteins because we also found significant ubiquitination of expressed RICK whose levels were similar to those of endogenous RICK in MEFs (see below, Figures 5D and 7D). To determine the linkage type of polyubiquitin chains on RICK, Nod1- and Nod2-expressing cells were transfected with RICK-Myc and HA-tagged wild-type (WT) or lysine-mutant Ub proteins containing single lysine mutations at position 48 (K48R) or 63 (K63R). Upon Nod1 and Nod2 stimulation with KF1B and MDP, respectively, RICK was strongly polyubiquitinated with K48R Ub, but much less with K63R Ub (Figure 2A). Thus, the polyubiquitin chains on RICK were linked primarily through lysine 63 of ubiquitin. Figure 2.Nod1 and Nod2 stimulation induce K63-linked polyubiquitination of RICK. (A) HEK293 cells constitutively expressing Nod1 (for KF1B stimulation), Nod2 (for MDP stimulation) and parental HEK293 cells (for TNFα stimulation) were transfected with expression plasmid of Myc-RICK in the presence of expression plasmids of HA-tagged WT, K48R, K63R mutant Ub proteins. Cells were stimulated with 100 ng/ml KF1B, 100 ng/ml MDP (synthetic Nod2-stimulatory compound) or 10 ng/ml TNFα, or left alone (0 h control) for the indicated times. Myc-RICK proteins in transfected cells were immunoprecipitated with rabbit anti-Myc Ab and ubiquitinated RICK proteins were then detected by immunoblotting with anti-HA Ab (upper panels labeled as RICK-Ubn). As a control, RICK proteins in the same samples and Ub proteins in total lysate were detected by immunoblotting with the indicated Abs. (B) Oligomerization-dependent RICK ubiquitination. HEK293T cells were transfected with control pcDNA3-RICK-Fpk3-Myc (−) or pcDNA3-RICK-Fpk3-Myc (RICK) in the presence of HA-Ub expression plasmids. Eighteen hours post-transfection, cells were treated with or without 200 nM AP1510. Twenty-four hours post-transfection, proteins were immunoprecipitated and ubiquitinated RICK was analyzed as described in panel A. Download figure Download PowerPoint Previous studies showed that enforced oligomerization of RICK by Nod1 and Nod2 induces NF-κB activation (Inohara et al, 2000). Similarly, oligomerization of RICK-ΔCARD-Fpk3, a RICK fusion protein in which the CARD is replaced with three tandem FKBP-related dimerization domains, results in NF-κB activation that is dependent on the FKBP-specific dimerizer AP1510 (Inohara et al, 2000; Figure 2B). Using this approach, we tested if oligomerization of RICK-ΔCARD-Fpk3 results in RICK polyubiquitination in HEK293T cells that coexpress HA-Ub. Immunoblotting analysis revealed that RICK-ΔCARD-Fpk3 was polyubiquitinated in an AP1510-dependent manner whereas the control Fpk3 protein was not (Figure 2B). These results suggest that induced proximity of RICK (1–435) lacking the CARD domain mimics the interaction between RICK and Nod proteins and induces RICK signaling. The kinase domain of RICK is essential for RICK ubiquitination To determine which region of RICK is required for polyubiquitination, we constructed RICK-ΔCARD-Fpk3-mutant proteins lacking various regions of RICK and tested their ability to be ubiquitinated in the absence and presence of the dimerizer AP1510 (Figure 3A). Immunoblotting analysis showed that all RICK-ΔCARD-Fpk3 proteins containing residues 1–292 of RICK were ubiquitinated in an AP1510-dependent manner, whereas that encompassing residues 293–435 was not (Figure 3B). These results indicate that the region composed of amino-acid residues 1–292 located within the kinase region of RICK contains the polyubiquitinated site(s). To determine the role of RICK polyubiquitination in NF-κB activation, the ability of RICK fusion proteins to activate NF-κB was tested using a reporter assay. Transfection of HEK293T cells with the different constructs revealed that all RICK proteins carrying the region between amino acid 1 and 319 induced NF-κB activation in an AP1510-dependent manner (Figure 3C). In contrast, enforced oligomerization of the fusion proteins carrying the region between 1 and 292 and that between 293 and 435 did not result in NF-κB activation (Figure 3C). These results indicate that both the kinase domain (residues 1–292), which contains the polyubiquitinated site(s), and the IM region (residues 293–319) are critical for NF-κB activation. Figure 3.The kinase domain of RICK is essential for RICK ubiquitination, but is not sufficient to activate NF-κB. (A) Schematic diagram of the structural domains of RICK and different deletion mutants. The results of ubiquitination and NF-κB activation experiments are summarized on the right. The region, which requires NF-κB activation in the IM domain is indicated by a gray box. (B) HEK293T cells were transfected with pcDNA3-RICK-(1–435)-Fpk3-Myc (ΔC), RICK-(1–292)-Fpk3-Myc, RICK-(1–319)-Fpk3-Myc, RICK-(1–328)-Fpk3-Myc, RICK-(1–356)-Fpk3-Myc, RICK-(1–390)-Fpk3-Myc RICK-(293–435)-Fpk3-Myc and control vector in the presence of HA-Ub expression plasmid and treated with AP1510 as in Figure 2B. Twenty-four hours post-transfection, proteins were immunoprecipitated and ubiquitinated RICK were analyzed as described in Figure 2. (C) HEK293T cells were transfected with pcDNA3-RICK-(1–435)-Fpk3-Myc (ΔC), RICK-(1–292)-Fpk3-Myc, RICK-(1–319)-Fpk3-Myc, RICK-(1–328)-Fpk3-Myc, RICK-(1–356)-Fpk3-Myc, RICK-(1–390)-Fpk3-Myc RICK-(293–435)-Fpk3-Myc and control vector in the presence of reporter plasmids. Eight hours post-transfection, cells were treated with medium containing the indicated amounts of AP1510. Twenty-four hours post-transfection, ligand-dependent NF-κB activation was determined with reporter assay. The level of NF-κB-dependent transcription activity in cells transfected with control vector is given as 1. The expression levels of Myc-RICK proteins are shown in an inset. (D) HEK293T cells were transfected with control vector (−), pcDNA3-RICK-(1–435)-Fpk3-Myc, RICK-(1–319)-Fpk3-Myc and RICK-(1–292)-Fpk3-Myc in the presence of reporter plasmids, and treated with AP1510 as in Figure 2B. Twenty-four hours post-transfection, RICK proteins were immunoprecipitated with rabbit polyclonal anti-Myc Ab, and then Flag-NEMO (upper panel) and Myc-RICK (middle panel) were immunodetected with mouse monoclonal anti-Flag and anti-Myc Abs, respectively. As control, Flag-NEMO in total lysate was immunodetected with anti-Flag Ab and shown in the lower panel. Download figure Download PowerPoint Previous studies showed that NF-κB activation mediated by RICK requires NEMO (Inohara et al, 2000). Therefore, we hypothesized that at least one of the regions required for NF-κB activation interacts with NEMO. To explore this hypothesis, we tested the ability of RICK-Fpk3-Myc constructs carrying different regions of RICK to interact with NEMO by co-immunoprecipitation assay. Consistent with a previous study (Inohara et al, 2000), a RICK truncation mutant expressing a region between amino-acid residues 1–319 interacted with NEMO (Figure 3D). Significantly, a RICK mutant composed of residues 1–292 did not associate with NEMO (Figure 3D). These results indicate that the interaction between RICK and NEMO requires the IM region-spanning residues 293–319 of RICK. Lysine 209 in the kinase domain of RICK is polyubiquitinated in a signal-dependent manner To map the polyubiquitinated lysine residue(s) in the kinase domain of RICK, we first performed close inspection of the amino-acid sequence of RICK. Alignment of amino-acid sequences from the kinase domain and IM region of RICK revealed 10 lysine residues that are evolutionarily conserved in 11 animal species (Supplementary Figure 1). Therefore, we introduced site-directed mutations in the 10 absolutely conserved and seven additional highly conserved lysine residues and tested if the mutations affect the ability of RICK-ΔCARD-Fpk3 to be ubiquitinated in an AP1510-dependent manner (Figure 4A). As shown in Figure 4B, the point mutant K209R, but not the remaining 16 lysine mutations, abolished the polyubiquitination of RICK. Notably, the K47M mutation, which replaces the essential lysine residue in the conserved ATP-binding site and results in loss of RICK kinase activity (Inohara et al, 1998), did not impair RICK polyubiquitination (Figure 4B). Thus, the conserved K209, but not the kinase activity of RICK, is essential for polyubiquitination. To determine if K209 is responsible for polyubiquitination of RICK during Nod1 signaling, we introduced the K209R mutation into a full-length RICK construct and tested the ability of WT and K209R-mutant RICK proteins to be ubiquitinated in Nod1-expressing HEK293 cells upon Nod1 stimulation with KF1B. Consistent with the results presented in Figure 4B, the K209R mutation abolished KF1B-induced polyubiquitination of RICK (Figure 4C). These results indicate K209 is essential for RICK polyubiquitination in Nod1 signaling. Figure 4.Lysine 209 is ubiquitinated in the kinase domain of RICK. (A) Location of lysine residues, which were replaced by arginines in RICK site-directed mutants. (B) HEK293T cells were transfected with control vector (−), pcDNA3-RICK-ΔCARD-Fpk3-Myc (WT) and various lysine-to-arginine mutants in the presence HA-Ub expression plasmid and treated with AP1510 as in Figure 2B. Twenty-four hours post-transfection, proteins were immunoprecipitated and ubiquitinated RICK were immunodetected as described in Figure 1. (C) HEK293 cells constitutively expressing Nod1 were transfected with the expression plasmids of WT and K209R-mutant Myc-RICK in the presence of HA-Ub expression plasmid. Eighteen hours post-transfection, cells were stimulated with or without 100 ng/ml KF1B. Twenty-four hours post-transfection, proteins were immunoprecipitated and ubiquitinated RICK were immunodetected as described in Figure 1. Download figure Download PowerPoint Polyubiquitination of RICK is essential for Nod1 and Nod2-mediated NF-κB activation To test if polyubiquitination at K209 of RICK is required for NF-κB activation, we assessed the ability of lysine-mutant RICK-ΔCARD-Fpk3 proteins to activate NF-κB after stimulation with different doses of AP1510. All the point mutant RICK proteins were expressed at levels comparable with that of WT protein, except the kinase-dead K47M mutant that was expressed at significantly lower levels (Figure 5A, inset). Consistent with its lower expression, the K47M mutant showed reduced levels of NF-κB activation when compared with WT protein (Figure 5A). However, the K47 mutant did show a dose-dependent increase in NF-κB and since K47 is essential for the kinase activity of RICK (Inohara et al, 1998), we conclude that the kinase activity of RICK is required for neither NF-κB activation nor polyubiquitination. Notably, the K209R mutation, but not those of the remaining conserved lysines, abolished AP1510-dependent NF-κB activation (Figure 5A). These results indicate that polyubiquitination of RICK at K209 is essential for NF-κB activation. Figure 5.Polyubiqutination at K209 is essential for RICK-mediated signaling. (A) HEK293T cells were transfected with pcDNA3-RICK-ΔCARD-Fpk3-Myc (WT) or the mutants in the presence of reporter plasmids. Eight hours post-transfection, cells were treated with medium containing the indicated amount of AP1510. Twenty-four hours post-transfection, ligand-dependent NF-κB activation was determined with reporter assay. The level of NF-κB-dependent transcription activity in cells transfected with control vector is given as 1. Myc-tagged RICK proteins detected with anti-Myc Ab are shown in the inset. (B) RICK-deficient MEFs were infected with retrovirus vectors expressing WT or K209R HA-RICK or control vector and selected by hygromycine. RICK proteins detected with anti-RICK Ab. A nonspecific signal detected with the Ab is indicated by an asterisk. (C) RICK-deficient MEFs were infected with control (−) or retrovirus vectors expressing WT or K209R HA-RICK. Infected MEFs were selected by hygromycin and stimulated with 5 μg/ml KF1B, 10 ng/ml TNFα or left alone. Twelve hours post-stimulation, secretion levels of CCL2 and CXCL1 were determined by ELISA. The results shown are given as mean±s.d. of triplicate cultures and are representative of three experiments. (D) MEFs stably expressing RICK or RICK K209R were stimulated with 5 μg/ml KF1B for the indicated time. Post-stimulation, cells were lysed, RICK was immunoprecipitated and ubiquitinated RICK was immunodetected by anti-Ub Ab (upper panel). Cell lysate was also immunoblotted with anti-phospho-IκBα, anti-IκBα or IKKβ Ab (bottom three panels). A nonspecific signal detected with the Ab is indicated by an asterisk. (E) MEFs stably expressing RICK or RICK K209R were stimulated with 5 μg/ml KF1B or TNFα for indicated times, and then anti-NEMO Ab was used to immunoprecipitate the IKK complex. The IKK assay was carried out by using GST-IκBα and γ-32P-ATP as substrates (upper). The amount of IKKβ in the immunoprecipitates was detected by immunoblotting. Download figure Download PowerPoint To test if K209 polyubiquitination on RICK is important for Nod1 signaling, we assessed the ability of WT and K209R-mutant RICK proteins to complement RICK function in embryonic fibroblasts deficient in RICK. In these experiments, MEFs from RICK-deficient mice were infected with a retroviral vector producing WT or K209R-mutant RICK proteins. The expression levels of both WT and K209R RICK proteins in RICK−/− MEFs were similar to that of endogenous RICK in WT MEFs (Figure 5B). Notably, expression of WT RICK conferred responsiveness of MEFs deficient in RICK to KF1B, and resulted in secretion of chemokines CCL2 and CXCL1 (Figure 5C). In contrast, the K209R RICK mutant failed to rescue responsiveness to the Nod1 agonist (Figure 5C). Immunoblotting showed that WT but not K209R-mutant RICK was polyubiquitinated upon Nod1 stimulation and resulted in IKK activation (Figure 5D and E). Because control TNFα induced similar levels of CCL2 and CXCL1 secretion from cells infected with WT and K209R-mutant RICK constructs, these findings indicate that polyubiquitination of RICK at K209 is essential for Nod1 signaling but dispensable for that induced by TNFα stimulation. A20 inhibits Nod1- and Nod2-mediated NF-κB activation and polyubiquitination of RICK A20 is a deubiquitinase that removes K63-linked polyubiquitin chains and functions as a negative feedback regulator of inflammatory responses (Lee et al, 2000; Boone et al, 2004; Wertz et al, 2004). Previous studies showed that A20 expression is induced by NF-κB activation (Krikos et al, 1992) and Nod1 stimulation (Masumoto et al, 2006). To test if A20 negatively regulates Nod signaling through deubiquitination of RICK, we tested the ability of A20 to affect RICK polyubiquitination and NF-κB activation in response to Nod1 or Nod2 stimulation. To assess the latter, A20 was coexpressed with Myc-RICK and HA-ubiquitin in HEK293T cells and the level of polyubiquitinated RICK was determined by immunoblotting. We found that expression of A20 decreased RICK polyubiquitination (Figure 6A). HEK293T cells express low, but significant, levels of endogenous Nod1 and Nod2 (Viala et al, 2004). Consistent with the latter, treatment of HEK293T cells with high amount of lipophilic Nod1- and Nod2-stimulatory molecules KF1B (5 μg/ml) and AcMDP (acetyl-(6-O-stearoyl)-muramyl-Ala-D-Glu-NH2) (5 μg/ml) induced NF-κB activation (Figure 6B). Using this system, we tested the ability of A20 to affect the NF-κB activation induced by these Nod1- and Nod2-stimulatory molecules in HEK293T cells, and found that NF-κB activation induced by KF1B and AcMDP, as well as that induced by TNFα and IL-1β, was reduced by A20 overexpression (Figure 6B). We next determined whether endogenous A20 mediates negative feedback of NF-κB activation after Nod1 stimulation. To test this, we used A20 small-RNA interference (RNAi) construct, which suppresses specifically the expression of A20 (Saitoh et al, 2005; also see inset in Figure 6C). Like that induced by TNFα and IL-1β, NF-κB activation induced by KF1B and AcMDP were enhanced by transfection with the A20 RNAi plasmid but not with control plasmid (Figure 6C). In addition, the ubiquitination level of endogenous RICK induced by Nod1 stimulation was upregulated by expression of A20 RNAi (Figure 6D). These findings suggest that A20 is a negative feedback regulator of Nod1 and Nod2 signaling. Figure 6.A20 functions as a negative regulator of Nod1 and Nod2 signaling via deubiquitination of RICK. (A) HEK293T cells were transfected with Flag-tagged A20 expression plasmids or control vector (−) in the presence of pcDNA3-RICK-ΔCARD-Fpk3-Myc (WT) and HA-Ub expression plasmid. Eighteen hours post-transfection, cells were treated with 200 nM AP1510 or left alone. Twenty-four hours post-transfection, proteins immunoprecipitated with rabbit anti-Myc Ab were subjected to SDS–PAGE and analyzed by western blot analysis using anti-HA, Myc and Flag Abs to detect ubiquitinated RICK proteins, RICK proteins, and A20 protein, respectively. (B) HEK293T cells were transfected with expression plasmids of A20 or control vector in the presence of reporter plasmids. Eight post-transfection, cells were treated with medium containing 1 μg/ml KF1B, 1 μg/ml AcMDP, 10 ng/ml TNF or 10 ng/ml IL-1β. Twenty-four hours post-transfection, ligand-dependent NF-κB activation was determined with reporter assay. The level of NF-κB-dependent transcription activity in the absence of A20 is given as 100%. The results shown are given as mean±s.d. of triplicate cultures and are representative of three experiments. (C) HEK293T cells were transfected with pcDNA3 (−), A20 interference RNA (RNAi) plasmid or control RNAi plasmid (Ctli) in the presence of reporter plasmids. Thirty-six hours after transfection, cells were treated with medium containing 1 μg/ml KF1B, 1 μg/ml AcMDP, 10 ng/ml TNFα or IL-1β. Forty-eight hours post-transfection, ligand-dependent NF-κB activation was determined with reporter assay. The level of NF-κB-dependent transcripti