Figure 1.
TRAF6 associates with the TAK1-TAB1-TAB2 complex in the absence of a functional TRAF6 RING finger domain.
(A) 293T cells were transfected with NFκB-luciferase, 40 ng TAB1 and TAK1, and 400 ng of empty vector (EV), TRAF6, or TRAF6 C70A as indicated. Cell lysates were analyzed by luciferase reporter assay. Values indicate fold increase over background and are normalized against a β-galactosidase internal standard. (B) 293T cells were transfected with equal amounts (1 µg) of epitope-tagged TRAF6 (FLAG) or TAK1 (Myc) expression plasmids and subjected to immunoprecipitation (IP) with anti-FLAG agarose. IP and whole cell extracts (WCE) were immunoblotted with anti-Myc or anti-FLAG as indicated. (C), (D), (E) 293T cells were transfected with WT or C70A FLAG-TRAF6(1-358)-Gyrase B as indicated, plus an equal amount of Myc-tagged TAB2 (C), TAB1 (D) or TAK1 (E). Cell lysates were processed and analyzed as in (B).
Figure 2.
A lysine-deficient TRAF6 N-terminus-Gyrase B fusion protein interacts with and activates TAK1, and induces NFκB and AP-1 reporter activity.
(A) Generation of mutant versions of murine TRAF6 containing lysine (K) to arginine (R) mutations in the RING finger, zinc finger, and coiled-coil domains (TRAF6 K32-348R) (top), or in the RING finger, zinc finger, and coiled-coil domains, but with the TRAF6-C domain replaced with the Gyrase B artificial oligomerization domain (TRAF6(1-358)ΔK-Gyrase B) (bottom). (B) 293T cells were transfected with the indicated full-length versions of FLAG-TRAF6, then lysates were immunoprecipitated with FLAG and subjected to in vitro ubiquitination in the presence or absence of ATP using recombinant WT, K48-only, or K63-only ubiquitin. Unmodified and modified TRAF6 were detected by immunoblotting with anti-TRAF6. (C) 293T cells were transfected with equal amounts (1 µg) of epitope-tagged TAK1 (Myc) in combination with either wild-type (WT), RING mutant (C70A), or lysine-deficient (ΔK) TRAF6(1-358)-Gyrase B (FLAG) expression plasmids, and subjected to immunoprecipitation (IP) with anti-FLAG agarose. IP and whole cell extracts (WCE) were immunoblotted with anti-Myc or anti-FLAG as indicated. (D) 293T cells were transfected with 10 ng of TAK1 in combination with either wild-type (WT), RING mutant (C70A), or lysine-deficient (ΔK) TRAF6(1-358)-Gyrase B (FLAG) expression plasmids, and left untreated or treated for 5 minutes with Coumermycin A before being harvested and subjected to IP with anti-FLAG agarose. IP and whole cell extracts (WCE) were immunoblotted with anti-phospho TAK1 (T187), anti-TAK1, or anti-FLAG as indicated. (E), (F) 293T cells were transfected with NFκB-luciferase (E) or AP-1-luciferase (F) plus WT, C70A, or K32-348R TRAF6(1-358)-Gyrase B as indicated. Cell lysates were analyzed by luciferase reporter assay. Values indicate fold increase over background and are normalized against a β-galactosidase internal standard.
Figure 3.
A lysine-deficient mutant TRAF6 rescues IL-1-mediated NFκB and MAPK activation, as well as IL-6 production, in TRAF6-deficient fibroblasts.
(A) A mutant version of full-length TRAF6 containing lysine (K) to arginine (R) mutations at all lysine residues (referred to as TRAF6 K32-518R or TRAF6 ΔK) was generated, and mutations are depicted in relation to TRAF6 domain location. (B) TRAF6-deficient fibroblasts were retrovirally-rescued with the indicated full-length versions of FLAG-TRAF6, then lysates immunoprecipitated with FLAG and subjected to in vitro ubiquitination in the presence or absence of ATP using recombinant K63-only ubiquitin. The last two lanes represent C70A and K32-518R lysates mixed 1∶1 prior to immunoprecipitation to assess trans-ubiquitination potential of TRAF6 monomers. Unmodified and modified TRAF6 were detected by immunoblotting with anti-TRAF6. (C) Cell lines used in (B) were treated for 5 minutes with IL-1β, then lysed in the presence of N-ethyl maleimide (NEM) and subjected to immunoblotting with anti-TRAF6 to detect both unmodified and high molecular weight forms of TRAF6. (D), (E) TRAF6-deficient fibroblasts retrovirally-rescued with the indicated full-length versions of FLAG-TRAF6 were treated as indicated with IL-1β, then lysed and subjected to immunoblotting against the activated phosphorylated forms of TAK1 (D), IKKα/β (D) IκBα (E), JNK (E), and p38 (E). (F) TRAF6-deficient fibroblasts retrovirally-rescued with the indicated full-length versions of FLAG-TRAF6 were left untreated or treated for 12 hours with IL-1β. Supernatants were collected and assayed by ELISA for IL-6 production. Values are normalized to crystal violet assays of the cell culture plates.
Figure 4.
A lysine-deficient mutant TRAF6 rescues RANKL-mediated NFκB and MAPK activation, as well as osteoclastogenesis in TRAF6-deficient BMM.
(A) Wild-type or TRAF6-deficient bone marrow macrophages (BMM) retrovirally-rescued with wild-type (WT), RING mutant (C70A), or lysine-deficient (ΔK) full-length versions of FLAG-TRAF6 were treated as indicated with RANKL, then lysed and subjected to immunoblotting against the activated phosphorylated forms of IκBα, JNK, and p38. B, BMM described in (A) were replated and cultured with M-CSF and RANKL for 5 days to induce osteoclast differentiation. (C) Osteoclasts depicted in (B) were fixed and subjected to TRAP solution assay and quantified at 405 nm absorbance. (D) Total cell counts per well of retrovirally-rescued osteoclasts depicted in (B) as defined by cells containing at least 3 nuclei and being at least 100 µM in diameter.
Figure 5.
A lysine-deficient TRAF6 N-terminus-Gyrase B fusion protein remains competent to mediate a TRAF6-specific ubiquitin modification of NEMO.
(A) Wild-type (WT) and TRAF6-deficient (KO) fibroblasts were stimulated for 5 minutes with IL-1β, then lysed with buffer containing NEM. To determine specificity of in vivo ubiquitination, lysates were dissociated and denatured before being immunoprecipitated with anti-NEMO. WCE and IP extracts were subjected to immunoblotting with anti-Ubiquitin. (B) 293T cells were transfected with equal amounts (1 µg) of Myc-tagged NEMO and expression plasmids for TRAF6, TRAF2, TRAF5, constitutively active IKKβ (IKKβEE), TAB1 and TAK1, MEKK3, or ASK1. Cells were lysed in the presence of phosphatase inhibitors and NEM, and WCE were then immunoblotted with anti-NEMO to detect modification of NEMO. (C) 293T cells were transfected with equal amounts (1 µg) of Myc-tagged NEMO and expression plasmids for either wild-type, RING mutant (C70A), or lysine-deficient FLAG-TRAF6(1-358)-Gyrase B. Cells were lysed in the presence of phosphatase inhibitors and NEM, and WCE were then immunoblotted with anti-NEMO to detect modification of NEMO. (D) 293T cells were transfected with 10 µg Myc-NEMO and co-transfected with wild-type or C70A TRAF6(1-358)-Gyrase B (10 µg). These plates, and an untransfected plate, were lysed and immunoprecipitated with anti-Myc agarose. IP extracts were dissociated from anti-Myc agarose by heating (65°C) in non-reducing SDS-PAGE buffer, and separated from agarose by centrifugation. Reducing agent was added to IP extracts, and 10% of each was used for immunoblotting with anti-Myc to confirm the presence of NEMO modification (not shown), while 90% was run on a preparative gel and stained with colloidal coomassie blue (not shown). The NEMO modification band found in the lane co-transfected with wild-type, but not with C70A TRAF6(1-358)-Gyrase B, was excised, digested with trypsin, and analyzed by mass spectrometry. Depicted is the m/z profile of the NEMO peptide QEVIDKLKEEAEQHK, for which all detected ions gave predicted values except for the fragment encompassing lysine 285 between y7 and y8, which exhibits a residue mass of 242.2. This profile is consistent with one lysine and two glycines, and is indicative of an ubiquitin modification at NEMO K285. (E) 293T cells were transfected with equal amounts (1 µg) of Myc-tagged wild-type or K285R NEMO in combination with expression plasmids for either wild-type, RING mutant (C70A), or lysine-deficient FLAG-TRAF6(1-358)-Gyrase B. Cells were lysed in the presence of phosphatase inhibitors and NEM, and WCE were then immunoblotted with anti-NEMO to detect modification of NEMO.
Figure 6.
Ubiquitination of NEMO by TRAF6 is partially required for IL-1-mediated activation of NFκB.
(A) NEMO-deficient fibroblasts retrovirally-rescued with wild-type or K285R NEMO were treated as indicated with IL-1β, then lysed and subjected to immunoblotting against the activated phosphorylated forms of TAK1, IKKα/β, and IκBα. (B) NEMO-deficient fibroblasts retrovirally-rescued with WT or K285R NEMO were left untreated or treated for 12 hours with IL-1β. Supernatants were collected and assayed by ELISA for IL-6 production. Values are normalized to crystal violet assays of the cell culture plates.
Figure 7.
A revised model for TRAF6-dependent activation of the MAPK and IKK signaling pathways in the absence of TRAF6 autoubiquitination.
Upon receptor engagement, K63-linked ubiquitin chains are affixed to TRAF6, at least in part through the E3 ubiquitin ligase activity of the TRAF6 RING finger domain. Both TRAF6 autoubiquitination and the TRAF6 RING finger domain appear dispensable for recruitment of the TAB1-TAB2-TAK1 complex. Previous reports indicate this interaction to be dependent on the ability of the TAB2 ubiquitin-binding domain to attach to K63-linked ubiquitin chains [17], suggesting an unidentified factor “X”, which interacts with TRAF6 independently of autoubiquitination, may serve as the binding partner for TAB2. At the same time, recruitment of TAK1 to TRAF6 is insufficient for TAK1 activation, which specifically requires TRAF6 RING finger activity. An additional unknown factor “Y” may be indicated, in this case to serve as a substrate(s) for TRAF6-mediated K63-linked ubiquitination. Upon being ubiquitinated, factor “Y” would promote TAK1 phosphorylation, leading to MAPK and IKKα/β activation. TRAF6-mediated activation of the IKKα/β complex is only partially dependent on TAK1 [31], but exhibits an absolute requirement for the regulatory component IKKγ (NEMO) [32], [33]. NEMO ubiquitination on K285 by TRAF6 is required for optimal activation of IKKα/β. RING finger-dependent activation of TAK1 and IKKα/β by TRAF6 might involve K63-linked ubiquitination of numerous targets, which together form a signaling complex sufficient to meet activation thresholds. TRAF6 autoubiquitination may, in fact, be required in the absence of alternative ubiquitin substrates, or in as yet untested signaling systems lacking robust redundancy. This schematic underscores the point that in contrast to the absolute requirement for TRAF6 RING finger E3 ligase activity in mediating downstream signaling, K63-linked ubiquitination of TRAF6 should be viewed as a marker of activation.