IFN regulatory factor
IFN-stimulated response element
myeloid differentiation protein 88
TIR domain-containing adaptor protein
TIR domain-containing protein
TNF receptor-associated factor
TRIF-related adaptor molecule
TIR domain-containing adaptor inducing IFN-β
Toll-like receptor 3 (TLR3) plays an important role in antiviral responses through recognizing viral double-stranded RNA produced during viral infection and mediating induction of type I IFN. TRIF is a Toll/IL-1 receptor (TIR) domain-containing adaptor protein that is associated with TLR3 and critically involved in TLR3-mediated signaling. In yeast two-hybrid screens, we identified TNF receptor-associated factor (TRAF)1 as a TRIF-interacting protein. The TRAF-C domain of TRAF1 and the TIR domain of TRIF were responsible for their interaction. Overexpression of TRAF1 inhibited TRIF- and TLR3-mediated activation of NF-κB, IFN-stimulated response element and the IFN-β promoter. Overexpression of TRIF caused caspase-dependent cleavage of TRAF1. The cleaved N-terminal but not C-terminal fragment of TRAF1 was responsible for inhibiting TRIF signaling. Mutation of the caspase cleavage site of TRAF1 or addition of the caspase inhibitor crmA inhibited TRAF1 cleavage and abolished the ability of TRAF1 to inhibit TRIF signaling, suggesting that TRIF-induced cleavage of TRAF1 is required for its inhibition of TRIF signaling. Our findings provide a novel mechanism for negative regulation of TRIF-mediated signaling.
Toll-like receptors (TLR) recognize specific pathogen-associated molecular patterns via their extracellular leucine-rich repeats, leading to the activation of innate immune responses and the shaping of the subsequent adaptive immune responses 1, 2. TLR3 recognizes viral double-stranded RNA produced during viral infection and plays a role in innate antiviral responses 3. Engagement of TLR3 causes activation of two transcription factors, NF-κB and IFN regulatory factor (IRF)-3, which collaborate to induce transcription of type I IFN. Type I IFN activate STAT pathways, leading to transcriptional induction of a wide range of genes. The induced downstream gene products, such as IRF, PKR, IP10, ISG15 and OAS, orchestrate inhibition of viral replication and clearance of virus-infected cells, thus leading to antiviral responses 4.
Recently, a Toll/IL-1 receptor (TIR) domain-containing adaptor protein named TRIF has been shown to be associated with TLR3 and critically involved in TLR3-mediated signaling 5, 6. Gene knockout and chemical mutagenesis studies indicated that TRIF-deficient mice are defective in TLR3-mediated activation of NF-κB and expression of type I IFN 7, 8. Further studies indicated that TRIF signals NF-κB activation through recruitment of a TNF receptor-associated factor (TRAF)6-TAK1-TAB2 complex to TLR3 9, 10. TRIF also interacts with IRF-3, as well as two noncanonical IκB kinase (IKK) family members, TBK1 and IKKϵ 10–12. It has been shown that TBK1 and IKKϵ can directly phosphorylate IRF-3, a process required for translocation of IRF-3 into the nucleus and binding to the enhancer element, IFN-stimulated response element (ISRE) 13, 14. These studies have demonstrated that TRIF bifurcates TLR3-mediated NF-κB and IRF-3 activation pathways.
In this report, we identified TRAF1 as a TRIF-interacting protein. TRIF causes caspase-dependent cleavage of TRAF1. The cleaved N-terminal fragment of TRAF1 inhibits TRIF-mediated signaling. Our findings identified a novel regulatory mechanism on TLR3-mediated signaling pathways.
Identification of TRAF1 as a TRIF-interacting protein
To identify TRIF-interacting proteins, we performed yeast two-hybrid screens with full-length TRIF as bait. We screened a total of ∼8×106 clones from a human B cell library and obtained eight independent β-galactosidase-positive clones, one of which encodes the C-terminal fragment aa 114–416 of TRAF1. Protein-protein interaction assay confirmed the TRIF-TRAF1(114–416) binding in yeast cells (Fig. 1A).
To demonstrate whether TRAF1 interacts with TRIF in mammalian cells, we did co-immunoprecipitation experiments in 293 cells. The results indicated that TRAF1 could interact with TRIF (Fig. 1B), but not with another TIR domain-containing adaptor protein, TIR domain-containing protein (TIRP)/TRIF-related adaptor molecule (TRAM) 15–17 (Fig. 1C).
Domain mapping of the interaction between TRAF1 and TRIF
TRAF1 is a member of the TRAF family 18, 19. It contains a conserved TRAF-C domain at its C terminus and a coiled coil domain in the middle. In most cases, TRAF1 interacts with its binding partners through its TRAF-C domain 18, 19. TRIF contains a conserved TIR domain in the middle that is responsible for its interaction with TLR3 and IRF-3 5, 10. To determine which domains are responsible for the interaction between TRAF1 and TRIF, we constructed series of deletion mutants of TRAF1 and TRIF. Transient transfection and co-immunoprecipitation experiments indicated that the TRAF-C domain of TRAF1 (aa 164–416) and the TIR domain of TRIF (aa 394–532) are required for the interaction between TRAF1 and TRIF (Fig. 2).
TRAF1 inhibits TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter
Because TRAF1 could interact with TRIF, we determined whether TRAF1 was involved in the regulation of TRIF-mediated signaling pathways. In reporter gene assays, TRAF1 inhibited TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter in a dose-dependent manner (Fig. 3A–C). In similar experiments, TRAF1 did not inhibit TRAF6-mediated NF-κB activation and IRF-3- or IRF-7-mediated ISRE activation (Fig. 3D, E), suggesting that TRAF1 does not target these downstream signaling components.
We also determined whether TRAF1 could inhibit signaling by other TIR domain-containing adaptors. In reporter gene assays, TRAF1 significantly inhibited TIR domain-containing adaptor protein (TIRAP)- and myeloid differentiation protein 88 (MyD88)- but not TIRP/TRAM-mediated NF-κB activation (Fig. 3D and data not shown). Consistently, we found that TRAF1 could interact with TIRAP and MyD88 (data not shown) but not TIRP/TRAM (Fig. 1C). These data point to the possibility that TRAF1 is also a negative inhibitor for TIRAP- and MyD88-mediated signaling.
Since TRIF is a TLR3-associated adaptor protein, we determined whether TRAF1 could regulate TLR3-mediated signaling. As shown in Fig. 4A–C, TRAF1 significantly inhibited poly(I:C)-induced activation of NF-κB, ISRE and the IFN-β promoter in a TLR3-expressing 293 cell line. In a similar experiment, TRAF1 did not inhibit IFN-γ-induced activation of the IRF-1 promoter (Fig. 4D). These data suggest that TRAF1 specifically inhibits TLR3-mediated activation of NF-κB and ISRE.
Previously, we have shown that overexpression of TRIF causes apoptosis through a RIP-FADD-caspase-8-dependent pathway 10. Co-transfection experiments indicated that TRIF-mediated apoptosis was not inhibited by TRAF1 (data not shown), suggesting that TRAF1 specifically inhibits TRIF-mediated activation of NF-κB and ISRE but not apoptosis.
Caspase-dependent cleavage of TRAF1 is required for its inhibitory role on TRIF signaling
Previously, it has been demonstrated that TRAF1 is cleaved by activated caspase-8 following TNF stimulation or Fas engagement. The cleavage occurs after aa D163 of TRAF1 and the cleaved C-terminal fragment (aa 164–416) functions as an inhibitor of NF-κB activation 20–22. Since TRIF also induces caspase-8-dependent apoptosis 10, we determined whether TRAF1 is cleaved following overexpression of TRIF. As shown in Fig. 5, overexpression of TRIF but not TIRP/TRAM caused TRAF1 cleavage and this was inhibited by the caspase inhibitor crmA. The mass of the cleaved N-terminal fragment was ∼19 kDa. In addition, TRIF did not cause cleavage of TRAF1(D163A), a TRAF1 mutant in which aa D163 is changed to alanine (Fig. 5). These data suggest that overexpression of TRIF caused caspase-dependent cleavage of TRAF1 after D163.
We next determined the significance of TRAF1 cleavage on TRIF signaling. In reporter gene assays, the non-cleavable TRAF1 mutant, TRAF1(D163A), lost its ability to inhibit TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter (Fig. 6A–C). CrmA, which inhibited TRIF-induced cleavage of TRAF1, also abolished the ability of TRAF1 to inhibit TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter (Fig. 6A–C). These data suggest that the cleavage of TRAF1 is required for its ability to inhibit TRIF-mediated signaling. In these experiments, crmA enhanced TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter (Fig. 6A–C). The simplest explanation for this observation is that crmA inhibits TRIF-mediated apoptosis and therefore enhances TRIF-mediated signaling.
We next determined which cleaved fragment of TRAF1 is responsible for inhibition of TRIF signaling. As shown in Fig. 6D–E, TRAF1(1–163) but not TRAF1(164–416) inhibited TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter. In addition, TRAF1(1–163) but not TRAF1(164–416) inhibited TLR3-mediated activation of the IFN-β promoter (Fig. 6G). The lack of inhibitory effects of TRAF1(164–416) on TRIF- and TLR3-mediated signaling is not due to its decreased expression level because it was expressed at a level comparable with TRAF1(1–163) as shown by Western blot analysis (Fig. 2).
Several proteins have been described to negatively regulate TLR signaling, including the orphan receptor ST2/T1 23, IL-1R-associated kinase M 24, suppressor of cytokines signaling 1 25, the TIR domain protein SIGIRR 26 and A20 27, 28. In this study, we identified TRAF1 as a novel negative regulator of TLR3-mediated IFN-β signaling.
In yeast two-hybrid screens, TRAF1 was identified to interact with TRIF. This interaction was confirmed in mammalian cells (Fig. 1). Overexpression of TRAF1 inhibited TRIF- and TLR3-mediated activation of NF-κB, ISRE and the IFN-β promoter, suggesting that TRAF1 negatively regulates TRIF- and TLR3-mediated signaling (Fig. 2, 3). TRAF1 does not interact with TIRP/TRAM (Fig. 1C) and does not inhibit TLR4-mediated signaling (data not shown). Since TLR4 signals through a TRIF/TIRP complex, our data suggest that TRAF1 does not inhibit TRIF signaling when TRIF is complexed with TIRP/TRAM.
It has been demonstrated that TRIF mediates a RIP-FADD-caspase-8-dependent apoptosis pathway 10. We found that overexpression of TRIF caused cleavage of TRAF1 (Fig. 5). This cleavage was inhibited by addition of the caspase inhibitor crmA, suggesting the cleavage was dependent on caspase activity. Previously, it has been demonstrated that TRAF1 is cleaved by caspase-8 after D163 upon TNF stimulation or Fas engagement 20–22. Based on the size of the cleaved fragment, we believed that TRAF1 was also cleaved after D163 upon TRIF overexpression. This was confirmed by our observation that the TRAF1(D163A) mutant was not cleaved by overexpression of TRIF (Fig. 5). Interestingly, TRAF1(D163A) lost its ability to inhibit TRIF-mediated activation of NF-κB and ISRE, while crmA suppressed the ability of TRAF1 to inhibit TRIF-mediated activation of NF-κB and ISRE (Fig. 6). These data suggest that the cleavage of TRAF1 is required for its ability to inhibit TRIF signaling.
Previously, it has been shown that the cleaved C-terminal fragment of TRAF1 inhibits NF-κB activation and promotes apoptosis induced by TNF and Fas ligand 20–22. This was confirmed in our experiments (data not shown). In our study, we found that the cleaved N-terminal but not C-terminal fragment was responsible for inhibition of TRIF-mediated activation of NF-κB, ISRE and the IFN-β promoter (Fig. 6). One possible explanation is that the two fragments of TRAF1 target different signaling components in different pathways. How the cleaved N-terminal fragment of TRAF1 inhibits TRIF signaling needs to be further investigated.
Previous studies indicate that the expression of TRAF1 is up-regulated in an NF-κB-dependent manner 29, and TRAF1 is cleaved by activated caspase-8 after the D163 residue 20–22. Since TRIF mediates both NF-κB and caspase-8 activation 10, TRAF1 may play a negative feedback role in TRIF-mediated signaling leading to activation of NF-κB and ISRE.
Materials and methods
Reagents and cell culture
Monoclonal antibodies against Flag and HA epitopes (Sigma, St. Louis, MO), IFN-γ (R&D Systems Inc., Minneapolis, MN), poly(I:C) (Sigma), human embryonic kidney 293 cells (ATCC, Manassas, VA) and human B cell cDNA library (Clontech, Palo Alto, CA) were purchased from the indicated manufactures. TLR3-expressing 293 cells (293-T3Y cells) were a kind gift from Drs. Katherine Fitzgerald and Tom Maniatis.
Yeast two-hybrid screening and protein-protein interaction assay
pGBT9-TRIF bait vector was described previously 10 and the human B cell cDNA library was screened as described 10. β-Galactosidase-positive clones were sequenced. To validate the interactions, the in-frame clones were fused to Gal4 DNA binding domain of pGBT9 vector and co-transformed into yeast with pGADT7-TRIF, in which TRIF cDNA is fused to Gal4 transcriptional activation domain. Interactions were tested by growth of yeast cells on plates with medium lacking histidine, leucine, tryptophan and adenine and containing 20 mM 3-aminotriazole.
The IFN-β promoter, ISRE, NF-κB and IRF-1 promoter luciferase reporter constructs were previously described 10. Mammalian expression plasmids for Flag-TRAF1, Flag-TRAF6, Flag- or HA-tagged TRIF and its deletion mutants, Flag-IRF-3, HA-IRF-7 and crmA were previously described 10. Mammalian expression plasmids for HA-TRAF1 and its mutants were constructed by PCR amplification of the corresponding cDNA fragments and subsequently cloned into a CMV promoter-based vector containing a 5′ HA tag.
Cell transfection and reporter gene assays
293 cells (∼1×105) were seeded in 12-well plates and transfected the following day by standard calcium phosphate precipitation. Within the same experiment, each transfection was performed in triplicate, and where necessary, empty control plasmid was added to ensure that each transfection received the same amount of total DNA. To normalize for transfection efficiency, 0.05 μg of pRL-TK luciferase reporter plasmid was added to each transfection. Approximately 14 h after transfection, reporter assays were performed using a dual-specific luciferase assay kit (Promega, Madison, WI) by following the manufacturer's procedures. Where necessary, 14 h after transfection, cells were treated with poly(I:C) (50 μg/mL) or IFN-γ (100 ng/mL) or left untreated for 6 h before luciferase assays were performed. All reporter assays were repeated for at least three times and similar data were obtained.
Co-immunoprecipitation and Western blot analysis
For transient transfection and co-immunoprecipitation experiments, 293 cells (∼2×106) were transfected and lysed the following day in 1.5 mL of lysis buffer (20 mM Tris, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 5 μg/mL aprotinin, 5 μg/mL leupeptin, 2 mM PMSF, 30 mM NaF, 2 mM sodium pyrophosphate, pH 7.5). For each immunoprecipitation, a 0.5-mL aliquot of lysate was incubated with 0.3 µg of the indicated monoclonal antibody or control mouse IgG and 20 µl of 1:1 slurry of GammaBind A Plus-Sepharose (Amersham Pharmacia, Piscataway, NJ) for 4 h in cold room. The sepharose beads were washed three times with 1 mL of lysis buffer. The precipitates were fractionated on SDS-PAGE and subsequent Western blot analysis was performed as described. All immunoprecipitation experiments were repeated for at least three times and similar data were obtained.
This work was supported by a grant from the Chinese High Technology Program (no.2003AA221030).