Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Immunol. Author manuscript; available in PMC 2011 January 1.
Published in final edited form as:
Published online 2009 November 8. doi:  10.1038/ni.1819
PMCID: PMC2872790

Different modes of TRAF3 ubiquitination selectively activate type I interferon and proinflammatory cytokine expression


Balanced production of type I interferons (IFN) and proinflammatory cytokines upon engagement of Toll-like receptors (TLRs), which signal via adaptors containing a Toll-IL-1-Receptor (TIR) domain, such as MyD88 and TRIF, has been proposed to control the pathogenesis of autoimmune disease and tumor responses to inflammation. Here we show that TRAF3, a ubiquitin ligase that interacts with both MyD88 and TRIF, differentially regulated production of IFN and proinflammatory cytokines. Degradative TRAF3 ubiquitination during MyD88-dependent TLR signaling was essential for activation of mitogen-activated protein kinases (MAPKs) and inflammatory cytokine production. By contrast, TRIF-dependent signaling triggered non-canonical TRAF3 self-ubiquitination that activated the IFN response. Inhibition of degradative TRAF3 ubiquitination prevented expression of all proinflammatory cytokines without impacting the IFN response.


Balanced production of type I interferons and proinflammatory cytokines, such as tumor necrosis factor (TNF), is proposed to play a key role in the pathogenesis of autoimmune diseases1. Furthermore, IFN production can repress tumors, while TNF and other inflammatory cytokines can promote tumor growth2, 3. Yet, the mechanisms that balance the production of type I IFN and proinflammatory cytokines are poorly understood. The main receptors capable of inducing both cytokine classes are Toll-like receptors (TLRs), which respond to ligands of microbial, fungal, viral and mammalian origins46. Despite deployment of common signaling pathways, such as MAPK cascades and IκB kinase (IKK)-dependent NF-κB signaling, different TLRs elicit distinct biological responses, some being more potent inducers of proinflammatory cytokines, whereas others mainly induce IFNs and IFN-related genes. The biochemical basis for response specificity is poorly understood, although it was attributed to differential deployment of adaptor proteins7 and selective activation of interferon regulatory factors (IRF), such as IRF3, by TLRs that trigger the IFN response8.

TLRs recruit four TIR domain containing adaptors, which include MyD88 [], TRIF [], TRAM and TIRAP, to their cytoplasmic TIR domains915. These adaptors control distinct responses classified as either MyD88-dependent or TRIF-dependent4, 16. Whereas the MyD88-dependent response mediates induction of proinflammatory cytokines, the TRIF-dependent response is critical for IFN and IFN-related gene induction10, 11. How the two responses are differentially activated is unknown, but previous studies highlighted a critical role for the signaling protein TRAF3 [] in induction of IFN-related genes and inhibition of inflammatory cytokines17, 18. However, TRAF3, which is necessary for IRF3 activation, interacts with both MyD88 and TRIF. Whereas TRAF3 positively regulates IRF3 and the type I IFN response18, it negatively regulates MAPK signaling by CD40 ligand and BAFF, members of the TNF family19. By contrast, the related protein TRAF6 positively controls MAPK signaling by TNF receptors (TNFR) and TLRs20. What makes TRAF3 function negatively in one response and positively in another is unknown. It is also unclear why MyD88, which interacts with TRAF3, does not lead to IRF3 activation upon TLR4 engagement.

Using TLR4 as a prototypical TLR that elicits both MyD88- and TRIF-dependent responses, we now show that differential TRAF3 ubiquitination is the key to selective production of type I IFNs versus proinflammatory cytokines. TRIF-mediated signaling triggered TRAF3 self-ubiquitination via non-canonical K63-linked polyubiquitination, which was essential for activation of IRF3 and the IFN response. By contrast, MyD88-dependent signaling through TRAF6 and the ubiquitin ligases cIAP1 and cIAP2 resulted in degradative TRAF3 ubiquitination, which was required for MAPK activation and induction of proinflammatory cytokines and chemokines. Elimination of cIAP1 and cIAP2 resulted in highly specific inhibition of proinflammatory genes without any effect on the anti-inflammatory and tumor-suppressive IFN response.


cIAP1/2 and MyD88-dependent MAPK signaling

The ubiquitin ligases cIAP1 and cIAP2 (subsequently referred to as cIAP1/2) are redundant E3s that direct degradative (K48-linked) TRAF3 ubiquitination and are critical for CD40-induced two-stage MAPK signaling, in which assembly of the receptor-associated signaling complex is followed by translocation of the multiprotein complex to the cytosol, the site in which MAPK cascades are activated19. Using a small molecule Smac-mimetic (SM), which triggers rapid cIAP1/2 degradation21, 22, we found that cIAP1/2 were also involved in TLR signaling. In bone marrow derived macrophages (BMDMs) and the mouse macrophage line RAW264.7, SM pretreatment inhibited activation of the MAPK kinase kinase (MAP3K) TAK1, but not IκB kinase (IKK), by TLR4 and TLR2 ligation (Fig. 1a, Supplementary Fig. 1). SM had no effect on TAK1–MAPK activation by TLR3, which signals exclusively via TRIF. Congruently, TRIF-defective TrifLps2/Lps2 BMDMs exhibited relatively intact lipopolysaccharide (LPS)-induced TAK1–MAPK activation via TLR4 that remained sensitive to SM treatment, but residual TAK1–MAPK activation in My88−/− BMDM was barely affected by SM (Fig. 1b). cIAP1/2 were not involved in TRIF-mediated signaling necessary for IFN expression, as SM pretreatment did not prevent IRF3 dimerization or nuclear translocation (Fig. 1c). The effects of SM were specific, as RAW264.7 cells silenced by siRNA for cIAP1/2 expression, cIAP1/2-deficient multiple myeloma cells or RAW264.7 cells incubated with a proteasome inhibitor also exhibited defective LPS-induced TAK1–MAPK activation (Supplementary Fig. 2a–c). Furthermore, silencing of TRAF3 rendered RAW264.7 macrophages resistant to SM treatment (Supplementary Fig. 2d). Hence, the cIAP1/2 E3 ligases, which trigger K48-specific degradative TRAF3 ubiquitination19, 23, are important for MyD88-dependent MAPK activation, but dispensable for TRIF-dependent IFN induction.

Figure 1
Role of cIAP1/2 in TLR-mediated MAPK signaling. a, TLR4-, TLR2-, but not TLR3-, induced TAK1/MAPK activation depend on cIAP1/2. BMDMs were stimulated with TLR agonists: LPS (100 ng/ml; TLR4), Pam3CSK4 (1 µg/ml; TLR2) and Poly I:C (30 µg/ml; ...

TRAF3 and cytosolic translocation of MyD88 signaling complexes

To study formation of TLR4-associated signaling complexes, BMDMs were divided into membrane (which contains plasma and endosomal membranes) and cytosolic fractions after LPS stimulation and analyzed by immunochemistry. LPS induced rapid, but transient, recruitment of MyD88, TRAF6, TRAF3, IKKγ (also known as NEMO), cIAP1/2, Ubc13 and TAK1 to TLR4 and more persistent TRIF recruitment, lasting at least 30 min (Fig. 2a). TLR4 was not detected in the cytosolic fraction, but immunoprecipitation with TAK1 antibody revealed LPS-induced formation of a large cytosolic complex persisting for at least 30 min after receptor stimulation that contained MyD88, TRAF6, IKKγ, cIAP1/2, Ubc13, MKK4 (which was not part of the receptor-associated complex) and TAK1, but not TRAF3 or TRIF (Fig. 2a). SM pretreatment stabilized the receptor-associated complex and prevented cytosolic translocation of the TAK1-associated complex (Fig. 2a). These results suggest that following assembly on the cytoplasmic face of TLR4, the MyD88-nucleated signaling complex, containing TRAF6, IKKγ, cIAP1/2, Ubc13 and TAK1, translocates to the cytosol, leaving behind TLR4 and TRAF3, and incorporates MKK4. Complex translocation required cIAP1/2 and was therefore inhibited by SM pretreatment, which also blocked MKK4 recruitment and TAK1–MAPK phosphorylation, which occurred in the cytosol and not at the receptor (Fig. 2a).

Figure 2
TLR4 engagement induces an MyD88-associated signaling complex that undergoes cIAP1/2- and TRAF6-dependent cytosolic translocation upon TRAF3 degradation. a, TLR4 activation induces formation and subsequent cytosolic translocation of a multi-protein signaling ...

Cytosolic translocation of the CD40-assembled signaling complex requires cIAP1/2 E3 activity and correlates with TRAF3 degradation19. We examined the fate of TRAF3, which was present in the TLR4-anchored complex but was not part of the cytosolic TAK1-associated complex. Total TRAF3 protein abundance rapidly, but incompletely, declined within 10 min of LPS stimulation, whereas total TRAF6 abundance remained constant (Fig. 2b). TRAF3 degradation was inhibited by SM pretreatment. Similarly, TLR4-associated TRAF3 rapidly declined at 10 min post-stimulation and this degradation was inhibited by SM (Fig. 2b). TLR4-associated TRAF6, however, was unchanged between 5 and 10 min after LPS addition, but after 10 min was undetectable unless the cells were pretreated with SM (Fig. 2b). Notably, small amounts of TRAF3 remained associated with TLR4 even at 30 min post-stimulation (Fig. 2b). This residual TRAF3 is likely to be engaged in MyD88-independent signaling. Silencing of TRAF6 in RAW264.7 macrophages prevented TAK1 recruitment to TLR4 but had no effect on recruitment of MyD88, TRIF or cIAP2, whereas TRAF3 silencing did not affect the recruitment of any of these proteins (Fig. 2c). Importantly, TRAF6 silencing slowed down the disassociation of the MyD88-assembled complex from the receptor.

Unlike TRAF2, however, TRAF6 does not interact directly with cIAP1/2 (ref. 24) and data not shown). Since recruitment of cIAP2 to TLR4 was MyD88- but not TRIF-dependent (Supplementary Fig. 3a), we examined whether MyD88 and TRIF can interact with cIAP1/2. Consistent with the genetic analysis, pulldown experiments using glutathione S transferase (GST)-MyD88 or -TRIF fusion proteins revealed an interaction between cIAP2 and MyD88, but not with TRIF (Supplementary Fig. 3b). However, it remains to be determined whether this protein interaction is direct.

It was proposed that MyD88 and TRIF recruitment to the TIR domain of TLR4 is sequential and mutually exclusively25, 26. Consistent with cIAP2 recruitment to TLR4 being MyD88-dependent, immunoprecipitation of the membrane fraction with an cIAP2 antibody resulted in isolation of TLR4 and MyD88, but not TRIF (Supplementary Fig. 4a). Furthermore, inhibition of TLR4 endocytosis with the dynamin inhibitor dynasore27 had no effect on MyD88 or cIAP2 recruitment to the receptor, but blocked TRIF recruitment (Supplementary Fig. 4b). We conclude that TRIF and MyD88 are recruited to separate pools of receptors. As SM inhibits the dissociation of MyD88 from the receptor without affecting TRIF recruitment, whereas dynasore inhibits TRIF recruitment without affecting MyD88 recruitment, it appears that each adaptor is independently recruited to TLR4.

TRAF6 is required for cIAP2 and TRAF3 ubiquitination

We examined the effect of TRAF3 and TRAF6 silencing on TLR4-induced signaling responses. LPS-induced TAK1-MAPK activation were barely detected in TRAF6-deficient cells, whereas TRAF3-depletion accelerated their activation (Fig. 3a,b). TRAF3, however, and not TRAF6, was required for IRF3 activation (Supplementary Fig. 5). LPS triggered cIAP2 and TRAF3 polyubiquitination (Fig 3c,d). Total, K48-linked and K63-linked cIAP2 ubiquitination were TRAF6-dependent but TRAF3-independent (Fig. 3c). TRAF6-depletion diminished total and K48-linked, but not K63-linked, TRAF3 polyubiquitination (Fig. 3d). Congruently, TRAF6 ablation inhibited LPS-induced TRAF3 degradation (Fig. 3a). Likewise, SM treatment inhibited total, but not K63-linked, TRAF3 ubiquitination (Fig. 4a), suggesting that cIAP1/2 are responsible for K48-linked TRAF3 ubiquitination, as observed during CD40 signaling19. Akin to TRAF2 during CD40 signaling23, TRAF6 may mediate TLR4-induced cIAP1/2 activation through their K63-linked ubiquitination and is therefore needed for TRAF3 degradation. Since TRAF6 is a K63-specific E3 ligase, the K48-linked ubiquitination of cIAP2 that shows TRAF6-dependence, is most likely due to self-ubiquitination by cIAP2 or cIAP1.

Figure 3
TRAF6 is required for LPS-induced TAK1 activation, and cIAP2 and TRAF3 ubiquitination. a, TRAF6-dependent TAK1 activation and TRAF3 degradation. Control, TRAF3- or TRAF6-silenced RAW264.7 cells were LPS stimulated. TAK1 and MAPK phosphorylation and TRAF3 ...
Figure 4
LPS-induced K63-linked TRAF3 self-ubiquitination depends on TLR endocytosis. a, K63-linked TRAF3 ubiquitination requires endocytosis but not cIAP1/2 activity. RAW264.7 cells pretreated with or without SM were stimulated with LPS in the absence or presence ...

TLR3 also triggered K63-linked TRAF3 ubiquitination (Supplementary Fig. 6a). Interestingly, the ratio between K63-linked and total TRAF3 ubiquitination was higher for TLR3, which signals exclusively via TRIF. Indeed, TLR4-induced K63-linked ubiquitination of TRAF3 was TRIF-dependent and MyD88-independent (Supplementary Fig. 6b). By contrast, SM-sensitive TRAF3 ubiquitination was MyD88-dependent, consistent with its reliance on cIAP1/2, which are recruited to TLR4 via MyD88.

TRAF3 K63-linked ubiqutination depends on endocytosis

After activation, TLR4 undergoes dynamin-dependent endocytosis, which is required for TRIF-dependent IFN signaling but not for MyD88-mediated signaling25. Since TRAF3 is a positive effector of the TRIF-dependent IFN response, we examined whether its non-canonical K63-linked ubiquitination is linked to TLR4 endocytosis. Inhibition of TLR4 endocytosis with dynasore modestly reduced total LPS-induced TRAF3 ubiquitination but strongly inhibited K63-linked TRAF3 ubiquitination (Fig. 4a). SM pretreatment diminished total TRAF3 ubiquitination but had no effect on its K63-linked ubiquitination, whereas treatment with both SM and dynasore abolished TRAF3 ubiquitination altogether (Fig. 4a). Dynasore treatment did not block activation of MAPKs and IKK (Supplementary Fig. 7).

We isolated the endosomal compartment (Supplementary Fig. 8) at different points after TLR4 activation and analyzed its composition. LPS induced association of TLR4, TRIF, TRAF6, TRAF3, Ubc13, TBK1 and TAK1, but not MyD88 or cIAP2, with endosomes (Fig. 4b). Endosomal TRAF3 was K63 polyubiquitinated and did not undergo LPS-induced degradation. Treatment with dynasore, but not SM, prevented LPS-induced endocytosis of TLR4 and its associated proteins.

TRIF-dependent K63-linked TRAF3 ubiquitination is associated with IRF3 activation and is akin to K63-linked ubiquitination of TRAF6, thought to be due to RING finger-mediated self-ubiquitination28. To examine whether K63-linked TRAF3 ubiquitination is also RING-dependent, we introduced C68A,H70A substitutions, analogous to TRAF6 inactivating mutations28, into the TRAF3 RING finger. TRAF3-silenced cells were reconstituted with either wild-type TRAF3 or the RING finger mutant (RM). Both TRAF3 forms underwent LPS-induced polyubiquitination, but K63-linked polyubiquitination of RM-TRAF3 was greatly reduced relative to that of wild-type TRAF3 (Fig. 5a). Congruently, reconstitution of TRAF3-silenced cells with either wild-type or RM-TRAF3 delayed TAK1–MAPK activation, but only wild-type–TRAF3 supported IRF3 activation (Fig. 5b). Furthermore, both TRAF3 isoforms reduced IL-6 induction, but only wild-type–TRAF3 supported type I IFN induction (Fig. 5c).

Figure 5
K63- and K48-linked ubiquitination have different and distinct roles in TRAF3 function. a, K63-linked TRAF3 ubiquitination is RING finger-dependent. TRAF3-silenced RAW264.7 cells were reconstituted with empty vector or Flag-tagged WT- or RM-TRAF3. Upon ...

We systematically substituted lysines with arginines throughout TRAF3 to identify acceptors for K48-linked polyubiquitin chains. Single mutants exhibiting reduced ubiquitination were combined to generate double mutants, amongst which (K107,156)R-TRAF3 exhibited the most substantial, but still incomplete, reduction in LPS-induced K48-linked ubiquitination with little effect if any on K63-linked ubiquitination (Fig. 5d). When expressed in TRAF3-silenced macrophages, (K107,156R)-TRAF3 supported LPS-induced IRF3 activation and IFN mRNA induction, but led to lower TAK1 activation and reduced IL-6 mRNA induction relative to wild-type TRAF3 (Fig. 5d–f).

SM inhibits inflammatory cytokines but not IFN-related gene expression

To determine the role of the two different modes of TRAF3 ubiquitination in TLR signaling, we downregulated cIAP1/2, which are responsible for degradative TRAF3 ubiquitination23, 29, 30, by treating BMDMs with SM. This treatment inhibited induction of inflammatory cytokine and chemokine genes, including Tnf, Il6, Il12b, Il12a, Cxcl2 and Cxcl1 by LPS (TLR4) and Pam3CSK4 (TLR2), but had no effect on their induction by Poly I:C (TLR3) (Fig. 6a, Supplementary Fig. 9a). SM, however, had no effect on induction of Ifnα, Ifnβ and IFN-related genes, including Il10, Cxcl10, Ccl5 and Ccl2 in response to any TLR agonist (Fig. 6a, Supplementary Fig. 9a). Similar effects on cytokine gene induction were seen in cIAP1/2-deficient RAW264.7 cells (Supplementary Fig. 9b) and multiple myeloma cells (Supplementary Fig. 9c). cIAP1/2-dependent inflammatory cytokine induction by TLR4 was unique to the MyD88-dependent response, as SM pretreatment inhibited LPS-induced inflammatory cytokines and chemokines in TrifLps2/Lps2 BMDMs, which were impaired in induction of IFN-related genes (Fig. 6b, Supplementary Fig. 9b). By contrast, induction of IFN-related genes and residual inflammatory cytokine gene expression in LPS-stimulated Myd88−/− BMDMs were not affected by SM pretreatment. Hence, the two responses, one entailing induction of inflammatory cytokines and the other encompassing type I IFN and IFN-related genes, are separately regulated and display differential sensitivity to SM.

Figure 6
Differential regulation of TLR4-induced inflammatory cytokines and IFN-related genes. a, WT BMDMs were stimulated with different TLR agonists for 2 h for Il6, Tnf, Il12b or for 6 h for Ifna4, Ifnb and Il10 with or without SM pretreatment. RNA was extracted ...


TLRs detect microbes, viruses and endogenous ligands to mediate induction of inflammatory cytokines, chemokines, IFNs and IFN-related genes16. In general, TLRs that recognize bacteria induce pro-inflammatory cytokines, chemokines and anti-microbial peptides, whereas those that detect viruses trigger the IFN response31. How these two responses, which depend on engagement of MyD88 and TRIF, are balanced to control autoimmunity1 and pro-tumorigenic vs anti-tumorigenic inflammation2 was heretofore unknown32. TRAF3 is uniquely required for the TRIF-dependent IFN response17, 18, but it is also a negative regulator of MAPK activation19. We therefore explored the basis for the different activities of TRAF3 and examined whether TRAF3 is involved in determining the inflammatory cytokines: type I IFN balance. We found that although TRAF3 is incorporated into both MyD88- and TRIF-assembled multiprotein complexes, its signaling function is differentially regulated by alternative ubiquitination modes. Within the MyD88-assembled signaling complex, TRAF3 undergoes degradative K48-linked ubiquitination dependent on its relative TRAF6 and cIAP1/2, the latter being direct K48-specific TRAF3 ubiquitin ligases19, 23. Importantly, cIAP1/2 were only present within the MyD88-assembled, but not within the TRIF-assembled, signaling complex. Degradative TRAF3 ubiquitination within the MyD88 complex precludes IRF3 activation and instead promotes cytosolic translocation of the entire signaling complex. This allows MAPK activation and induction of inflammatory genes. By contrast, association of TRAF3 with the cIAP1/2-devoid, endosomal TRIF signaling complex results in its K63-linked self-polyubiquitination, a modification that is required for IRF3 activation and induction of the IFN response. It should be noted that MyD88 and TRIF are not part of the same signaling complex and their differential signaling potential correlates with their ability to selectively engage cIAP1/2 and thereby dictate the nature of TRAF3 ubiquitination. Despite the absence of cIAP1/2, the TRIF-assembled signaling complex can also activate TAK1 to some extent and this may account for the weak induction of inflammatory cytokines that is SM-resistant, seen in MyD88-deficient cells.

The TRAF3 relatives, TRAF2 and TRAF6, are E3 ubiquitin ligases that selectively catalyze K63-linked polyubiquitination of themselves33 and other proteins, such as cIAP2 (ref.23). Their activity depends on Ubc13, a K63-specific ubiquitin conjugating enzyme33 that is essential for TNFR-and TLR-induced MAPK actiavtion34. We now demonstrate that, like its relatives, the TRAF3 RING finger is required for its K63-linked ubiquitination within the TRIF signaling complex, but unlike TRAF2 or TRAF6, K63-linked TRAF3 ubiquitination is not totally dependent on Ubc13 (unpublished results), thereby explaining activation of the IFN response in Ubc13-deficient cells34. Importantly, during MyD88 or CD40 signaling19, TRAF3 does not undergo K63-linked self-ubiquitination and instead acts as an inhibitor of MAPK activation and inflammatory cytokine induction. This inhibitory activity does not require the RING finger of TRAF3 and is eliminated upon its proteasomal degradation, which is promoted by its decoration with canonical, K48-linked, polyubiquitin chains. The extent of K63-linked TRAF3 ubiquitination correlates with IFN induction, being the highest for TLR3-stimulated macrophages. In the case of TLR4, K63-linked TRAF3 ubiquitination, just like IRF3 activation, depends on receptor endocytosis and TRIF rather than MyD88.

Our results demonstrate that MyD88-dependent MAPK signaling proceeds via a two stage mechanism, similar to that previously described for CD40 and other TNFRs19. This mechanism involves receptor-induced assembly of a multiprotein complex containing MyD88, TRAF6, Ubc13, IKKγ, cIAP1/2, TAK1 and TRAF3. Complex assembly results in TRAF6 activation, which leads to K63-linked ubiquitination of cIAP1/2 and enhancement of their activity as TRAF3 K48-specific E3 ligases19, 23. TRAF3 degradation allows translocation of the MyD88-associated signaling complex to the cytosol, where TAK1 and its subordinate MAPKs are activated. Interference with this process through SM-induced cIAP1/2 elimination selectively inhibits induction of inflammatory cytokines and chemokines without any deleterious effect on the IFN response, which includes induction of the anti-inflammatory cytokine IL-10. Curiously, IKK activation by TLR4, which also depends on TAK135 (and data not shown), is not affected by SM-induced inhibition of TAK1 phosphorylation. This suggests that unlike MAPK signaling, IKK activation depends on TAK1 but not on its protein kinase activity, an important concept that merits further investigation. Although not preventing NF-κB activation, interference with two-stage TLR signaling through SM-induced elimination of cIAP1/2 is sufficient for selective inhibition of inflammatory cytokine and chemokine production without any deleterious effect on the IFN response, which includes induction of the anti-inflammatory cytokine IL-10. We therefore propose that SM and similar cIAP1/2 antagonists may serve as superior anti-inflammatory drugs that will not compromise anti-viral immunity. This may be of importance in inflammatory diseases that respond to type I IFN1 as well as cancer whose growth is stimulated by proinflammatory cytokines, such as TNF, but inhibited by type I IFN2. Furthermore, selective inhibition of TNF and other proinflammatory cytokines without a concomitant reduction in IFN production may be useful in treatment of autoimmune disease caused by increased TNF and reduced type I IFN1.


Mice and cells

Myd88−/− and TrifLps2/Lps2 mice11, 36 were obtained from S. Akira (Osaka University) and B. Beutler (Scripps Research Institute), respectively. Control C57BL/6 mice were from the Jackson Laboratory. All mice were housed in a specific pathogen-free facility according to UCSD and NIH guidelines, and mouse protocols were approved by the UCSD Institutional Animal Care Committee. Bone marrow was collected from femurs and tibia of mice (8–10 weeks of age) and used to prepare BMDMs17, that were cultured in DMEM supplemented with 10 ng/ml M-CSF in addition to 10% FBS. KMS-28BM (WT) and KMS-28PE (cIAP1, cIAP2 doubly deficient) multiple myeloma cells were a generous gift from R. Fonseca, Mayo Clinic37. RAW264.7 cells were cultured as described38.

Subcellular fractionation

Subcellular fractions were prepared as described39, 40. Cells were resuspended in a buffer containing 250 mM sucrose, 20 mM Tris, pH 7.4, 5 mM MgCl2, 1 mM phenylmethylsulfonyl fluoride (PMSF) and 20 µg/ml of aprotinin for 20 min on ice and disrupted in a dounce homogenizer (15 strokes). After removing nuclei by centrifugation at 1000 × g for 10 min at 4 °C, the supernatants were centrifuged at 10,000 × g for 1 h at 4 °C, and the cytosol fraction was collected. The pellets containing cellular membranes were resuspended in 10 mM Tris, pH 7.4, 150 mM NaCl, and 0.2% Nonidet P-40. The nuclear fraction was solubilized in a nuclear lysis buffer containing 1% Triton X-100, 150 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EGTA, 1 mM EDTA, 1 mM PMSF and 20 µg/ml of aprotinin, centrifuged at 15,000 × g for 30 min at 4 °C, and the supernatant referred to as nuclear extract was collected. The endosomal fraction was isolated as described41. Cell pellets were resuspended in 5 volumes of a hypo-osmotic buffer (15 mM KCl, 1.5 mM Mg-acetate, 1 mM dithiothreitol (DTT), and 10 mM HEPES, pH 7.5), and homogenized in a dounce homogenizer (20 strokes). 0.1 volume of hyper-osmotic buffer (700 mM KCl, 40 mM Mg-acetate, 1 mM DTT, and 10 mM HEPES, pH 7.5) was added and the mixture centrifuged for 5 min at 800 × g. The supernatant was collected and treated with 1 µg/ml trypsin at 37 °C for 3 min. Proteolysis was stopped with soybean trypsin inhibitor (1.5 µg/ml), and the mixture was centrifuged for 20 min at 145,000 × g. The membrane pellet was resuspended in 1 ml homogenization buffer (0.25 M sucrose, 1 mM EDTA, and 10 mM Tris, pH 8.0) and centrifuged through a discontinuous sucrose gradient42 for 2 h at 100,000 × g. Fractions (1 ml/each) were collected from the bottom of the tube. Subcellular fractions were analyzed by immunoblotting with antibodies (see below) against syndecan (membrane), α-tubulin (cytosol), HDAC1 (nuclear), and transferrin and EEA1 (endosome) as markers. The transferrin and EEA1 containing fractions were pooled.

Immunoblotting and immunoprecipitation

Total cell lysates were prepared by ice-cold lysis buffer containing 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, 1% deoxycholate, 1 mM PMSF and 20 µg/ml of aprotinin, and immunoprecipitated with the indicated antibodies overnight at 4 °C. Whenever protein ubiquitination was analyzed, 20 mM N-ethylmaleimide (NEM, Sigma) was added to the lysis buffer. For complex co-immunoprecipitation, the indicated antibodies and cell lysates were incubated in 10 mM Tris, pH 7.4, 150 mM NaCl, and 0.2% Nonidet P-40. Antibodies used are described in Supplementary Methods.

IRF3 dimerization assay

The assay was performed as described43. Cells were lysed in a buffer containing 50 mM Tris, pH 8.0, 1% NP40, 150 mM NaCl, 1 mM PMSF and 20 µg/ml of aprotinin, and supplemented with native PAGE sample buffer (125 mM Tris, pH 6.8, and 30% glycerol). The samples were separated by native PAGE and analyzed by immunoblotting.

Short hairpin RNA (shRNA) constructs, lentiviral packaging and transduction

shRNA-endoding lentiviral vectors were constructed and packaged as described19. In short, 293T cells were transfected with pLSLPw-shRNA constructs along with packaging plasmids, pVSVG (Clontech) and pLV-CMV-delta 8.2 (I. Verma, Salk Institute) using Lipofectamine 2000 (Invitrogen). Virus-containing supernatants were collected at 48 to 96 h post-transfection and used to infect cells in the presence of 5 mg/ml polybrene (Sigma). After 24 h, the virus-containing medium was replaced with selection medium containing 5 mg/ml puromycin (EMD). After cell growth was stable, the cells were used in the described experiments. The oligonucleotide sequences used for shRNA expression were: m-TRAF3: 5’-GCAAGAGAGAGATTCTGGC; m-TRAF6: 5’-CGTCCTTTCCAGAAGTGCC; m-cIAP1: 5’-GGAGTAGTTCAATGTCAT and m-cIAP2: 5’-GCACCATGCCTTTGAGCTT.

Q-PCR analysis

Total cellular RNA from 1 × 105 cells was isolated with TRIzol (Invitrogen), and used to synthesize first-strand cDNA with iScript cDNA synthesis kit (Bio-Rad). mRNA amounts were analyzed by real-time quantitative polymerase chain reaction (Q-PCR)17. Sequences of the Q-PCR primers are described in Supplemental Table 1. All values were normalized to the level of cyclophilin mRNA expression, and represent means of duplicates.

Statistical analysis

Data are presented as averages ± s.d.. Differences between averages were analyzed by Student’s t test. P values < 0.05 were considered significant.

Supplementary Material

Supplementary Information


We thank H. Ichijo (University of Tokyo) for providing A.M. with space and support to carry out some of the work described above, S. Akira (Osaka University) and B. Beutler (Scripps Research Institute) for Myd88−/− and TrifLps2/Lps2 mice, respectively, R. Fonseca (Mayo Clinic) for multiple myeloma cells, X. Wang (University of Texas Southwestern) for Smac mimetic, H. Wang (St. Jude Children’s Research Hospital) for generating the K63-Ub specific monoclonal antibody, HWA4C4, and Millipore Corporate (Billerica, MA) for the K48-linked polyubiquitin antibody. Work was supported by NIH grant AI043477 to M.K. who is an American Cancer Society Research Professor. P.-H.T., A.M., and T.M. were supported in parts by American Lung Association of California, Global Center of Excellence (GCOE) program to H. I., and TOYOBO Biotechnology Foundation, respectively. D.A.A.V. was supported by NIH (AI52199), a Cancer Center Support CORE grant (CA21765), and the American Lebanese Syrian Associated Charities (ALSAC).



P.-H.T. and M.K. planned and designed all experiments and wrote the manuscript. P.-H.T. and A.M. performed most experiments. W.Z. and T.M. helped with the cell cultures, TRAF3 mutants and immunoprecipitation experiments. D.A.A.V. provided L y s 63-specific anti-ubiquitin antibody (HWA4C4).


1. Banchereau J, Pascual V. Type I interferon in systemic lupus erythematosus and other autoimmune diseases. Immunity. 2006;25:383–392. [PubMed]
2. Luo JL, Maeda S, Hsu LC, Yagita H, Karin M. Inhibition of NF-kappaB in cancer cells converts inflammation- induced tumor growth mediated by TNFalpha to TRAIL-mediated tumor regression. Cancer Cell. 2004;6:297–305. [PubMed]
3. Lin WW, Karin M. A cytokine-mediated link between innate immunity, inflammation, and cancer. J Clin Invest. 2007;117:1175–1183. [PMC free article] [PubMed]
4. Akira S, Takeda K. Toll-like receptor signalling. Nat Rev Immunol. 2004;4:499–511. [PubMed]
5. Karin M, Lawrence T, Nizet V. Innate immunity gone awry: linking microbial infections to chronic inflammation and cancer. Cell. 2006;124:823–835. [PubMed]
6. Pasare C, Medzhitov R. Toll-like receptors: linking innate and adaptive immunity. Adv Exp Med Biol. 2005;560:11–18. [PubMed]
7. Vogel SN, Fitzgerald KA, Fenton MJ. TLRs: differential adapter utilization by toll-like receptors mediates TLR-specific patterns of gene expression. Mol Interv. 2003;3:466–477. [PubMed]
8. Doyle S, et al. IRF3 mediates a TLR3/TLR4-specific antiviral gene program. Immunity. 2002;17:251–263. [PubMed]
9. Medzhitov R, et al. MyD88 is an adaptor protein in the hToll/IL-1 receptor family signaling pathways. Mol Cell. 1998;2:253–258. [PubMed]
10. Yamamoto M, et al. Cutting edge: a novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFN-beta promoter in the Toll-like receptor signaling. J Immunol. 2002;169:6668–6672. [PubMed]
11. Hoebe K, et al. Identification of Lps2 as a key transducer of MyD88-independent TIR signalling. Nature. 2003;424:743–748. [PubMed]
12. Bin LH, Xu LG, Shu HB. TIRP, a novel Toll/interleukin-1 receptor (TIR) domain-containing adapter protein involved in TIR signaling. J Biol Chem. 2003;278:24526–24532. [PubMed]
13. Yamamoto M, et al. TRAM is specifically involved in the Toll-like receptor 4-mediated MyD88-independent signaling pathway. Nat Immunol. 2003;4:1144–1150. [PubMed]
14. Fitzgerald KA, et al. Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal transduction. Nature. 2001;413:78–83. [PubMed]
15. Horng T, Barton GM, Medzhitov R. TIRAP: an adapter molecule in the Toll signaling pathway. Nat Immunol. 2001;2:835–841. [PubMed]
16. Kawai T, Akira S. TLR signaling. Semin Immunol. 2007;19:24–32. [PubMed]
17. Hacker H, et al. Specificity in Toll-like receptor signalling through distinct effector functions of TRAF3 and TRAF6. Nature. 2006;439:204–207. [PubMed]
18. Oganesyan G, et al. Critical role of TRAF3 in the Toll-like receptor-dependent and -independent antiviral response. Nature. 2006;439:208–211. [PubMed]
19. Matsuzawa A, et al. Essential cytoplasmic translocation of a cytokine receptor-assembled signaling complex. Science. 2008;321:663–668. [PMC free article] [PubMed]
20. Inoue J, Gohda J, Akiyama T. Characteristics and biological functions of TRAF6. Adv Exp Med Biol. 2007;597:72–79. [PubMed]
21. Petersen SL, et al. Autocrine TNFalpha signaling renders human cancer cells susceptible to Smac-mimetic-induced apoptosis. Cancer Cell. 2007;12:445–456. [PMC free article] [PubMed]
22. Li L, et al. A small molecule Smac mimic potentiates TRAIL- and TNFalpha-mediated cell death. Science. 2004;305:1471–1474. [PubMed]
23. Vallabhapurapu S, et al. Nonredundant and complementary functions of TRAF2 and TRAF3 in a ubiquitination cascade that activates NIK-dependent alternative NF-kappaB signaling. Nat Immunol. 2008;9:1364–1370. [PMC free article] [PubMed]
24. Werneburg BG, Zoog SJ, Dang TT, Kehry MR, Crute JJ. Molecular characterization of CD40 signaling intermediates. J Biol Chem. 2001;276:43334–43342. [PubMed]
25. Kagan JC, et al. TRAM couples endocytosis of Toll-like receptor 4 to the induction of interferon-beta. Nat Immunol. 2008;9:361–368. [PubMed]
26. Nunez Miguel R, et al. A dimer of the Toll-like receptor 4 cytoplasmic domain provides a specific scaffold for the recruitment of signalling adaptor proteins. PLoS One. 2007;2:e788. [PMC free article] [PubMed]
27. Macia E, et al. Dynasore, a cell-permeable inhibitor of dynamin. Dev Cell. 2006;10:839–850. [PubMed]
28. Lamothe B, et al. Site-specific Lys-63-linked tumor necrosis factor receptor-associated factor 6 auto-ubiquitination is a critical determinant of I kappa B kinase activation. J Biol Chem. 2007;282:4102–4112. [PMC free article] [PubMed]
29. Vaux DL, Silke J. IAPs, RINGs and ubiquitylation. Nat Rev Mol Cell Biol. 2005;6:287–297. [PubMed]
30. Li X, Yang Y, Ashwell JD. TNF-RII and c-IAP1 mediate ubiquitination and degradation of TRAF2. Nature. 2002;416:345–347. [PubMed]
31. Uematsu S, Akira S. Toll-like receptors and Type I interferons. J Biol Chem. 2007;282:15319–15323. [PubMed]
32. O'Neill LA, Bowie AG. The family of five: TIR-domain-containing adaptors in Toll-like receptor signalling. Nat Rev Immunol. 2007;7:353–364. [PubMed]
33. Pineda G, Ea CK, Chen ZJ. Ubiquitination and TRAF signaling. Adv Exp Med Biol. 2007;597:80–92. [PubMed]
34. Yamamoto M, et al. Key function for the Ubc13 E2 ubiquitin-conjugating enzyme in immune receptor signaling. Nat Immunol. 2006;7:962–970. [PubMed]
35. Sato S, et al. Essential function for the kinase TAK1 in innate and adaptive immune responses. Nat Immunol. 2005;6:1087–1095. [PubMed]
36. Kawai T, Adachi O, Ogawa T, Takeda K, Akira S. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity. 1999;11:115–122. [PubMed]
37. Keats JJ, et al. Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell. 2007;12:131–144. [PMC free article] [PubMed]
38. Park JM, et al. Signaling pathways and genes that inhibit pathogen-induced macrophage apoptosis--CREB and NF-kappaB as key regulators. Immunity. 2005;23:319–329. [PubMed]
39. Micheau O, Tschopp J. Induction of TNF receptor I-mediated apoptosis via two sequential signaling complexes. Cell. 2003;114:181–190. [PubMed]
40. Lawrence T, Bebien M, Liu GY, Nizet V, Karin M. IKKalpha limits macrophage NF-kappaB activation and contributes to the resolution of inflammation. Nature. 2005;434:1138–1143. [PubMed]
41. Beaumelle BD, Gibson A, Hopkins CR. Isolation and preliminary characterization of the major membrane boundaries of the endocytic pathway in lymphocytes. J Cell Biol. 1990;111:1811–1823. [PMC free article] [PubMed]
42. Johnson GL, Bourne HR. Influence of cholera toxin on the regulation of adenylate cyclase by GTP. Biochem Biophys Res Commun. 1977;78:792–798. [PubMed]
43. Iwamura T, et al. Induction of IRF-3/-7 kinase and NF-kappaB in response to double-stranded RNA and virus infection: common and unique pathways. Genes Cells. 2001;6:375–388. [PubMed]