TRAF5 is a downstream target of MAVS in antiviral innate immune signaling

Abstract

The recognition of nucleic acids by the innate immune system during viral infection results in the production of type I interferons and the activation of antiviral immune responses. The RNA helicases RIG-I and MDA-5 recognize distinct types of cytosolic RNA species and signal through the mitochondrial protein MAVS to stimulate the phosphorylation and activation of the transcription factors IRF3 and IRF7, thereby inducing type I interferon expression. Alternatively, the activation of NF-kappaB leads to proinflammatory cytokine production. The function of MAVS is dependent on both its C-terminal transmembrane (TM) domain and N-terminal caspase recruitment domain (CARD). The TM domain mediates MAVS dimerization in response to viral RNA, allowing the CARD to bind to and activate the downstream effector TRAF3. Notably, dimerization of the MAVS CARD alone is sufficient to activate IRF3, IRF7, and NF-kappaB. However, TRAF3-deficient cells display only a partial reduction in interferon production in response to RNA virus infection and are not defective in NF-kappaB activation. Here we find that the related ubiquitin ligase TRAF5 is a downstream target of MAVS that mediates both IRF3 and NF-kappaB activation. The TM domain of MAVS allows it to dimerize and thereby associate with TRAF5 and induce its ubiquitination in a CARD-dependent manner. Also, NEMO is recruited to the dimerized MAVS CARD domain in a TRAF3 and TRAF5-dependent manner. Thus, our findings reveal a possible function for TRAF5 in mediating the activation of IRF3 and NF-kappaB downstream of MAVS through the recruitment of NEMO. TRAF5 may be a key molecule in the innate response against viral infection.

Figure 1. Dimerization of the MAVS CARD restores activation of NF-κB.

A, C, D HEK293T cells were transfected with wild-type or mutant MAVS, and an NF-κB reporter, pLuc-PRD(II)₂. MAVS CARDmt contains a substitution of alanine for tryptophan at residue 68. Lysates were prepared and luciferase reporter assay was performed 24 hrs later. AP1510 was added 12 hrs prior to cell lysis (C). In the inset figure in (A), a western blot of lysates for FLAG-MAVS proteins was performed (A). B, E HEK293T cells were cotransfected with HA-IkBα and pcDNA3 or MAVS proteins and lysates were prepared. Anti-HA immunoprecipitates were probed for phosphorylated IkBα. Lysates were probed for HA-IkBα and FLAG-tagged MAVS proteins (E).

Figure 2. TRAF5 activates IRF3 and stimulates interferon-β production.

A Luciferase reporter assay was performed in HEK293T cells cotransfected with FLAG-tagged wild-type or trunctated TRAF3 or TRAF5 expression constructs along with pLuc-PRD(III-I)₃. In the inset figure, lysates were immunoblotted for FLAG-tagged proteins. Luciferase activity was measured 24 hrs following plasmid transfection. B HEK293T cells were transfected with FLAG-tagged TRAF3 or TRAF5 and HA-IRF3. Lysates were prepared and anti-HA immunoprecipitates were probed for phosphorylated IRF3 (Ser 396). Lysates were probed for FLAG-TRAF3 and FLAG-TRAF5. C, D, E Luciferase reporter assay was performed in HEK293T cells cotransfected with expression constructs for the indicated proteins along with pLuc-IFNβ. SeV infection was performed six hours following transfection and 18 hrs prior to luciferase assay (E). Discontinuity in graph indicates a broken Y-axis (C, D).

Figure 3. TRAF5 associates with the dimerized MAVS CARD.

A, B, C HEK293T cells were cotransfected with FLAG-tagged MAVS constructs together with HA-tagged TRAF5 and whole cell lysates were prepared. Anti-FLAG immunoprecipitates were probed for HA-TRAF5 or FLAG-MAVS. Lysates were immunoblotted for HA-TRAF5. D Yeast were transformed with the bait or prey constructs indicated as described in the Materials and Methods section and streaked onto minimal SD media lacking leucine and tryptophan (SD-LW) or leucine, tryptophan, adenine, and histidine (SD-LWAH). E HEK293T cells were transfected with FLAG-tagged RIG-I or MAVS constructs, HA-TRAF5, and AU1-Ub. Whole cell lysates were prepared and anti-HA immunoprecipitates were probed for Ub-conjugated TRAF3 and TRAF5 proteins using AU1 antibody from or total protein using HA antibody. F HEK293T cells were transfected with FLAG-MAVS CARD-FPK3, HA-TRAF3 or HA-TRAF5, and AU1-Ub. Anti-HA immunoprecipitates were probed for AU1-Ub conjugates or HA-TRAF3/5. Lysates were immunoblotted for FLAG-MAVS CARD-FPK3. AP1510 was added 12 hrs prior to cell lysis. G, H HEK293T cells transfected initially with siRNAs targeting GFP or TRAF5. 48 hrs later, cells were transfected with FLAG-RIG-IΔRD, FLAG-MAVS, or FLAG-p65, and pLuc-PRD(III-I)₃ (IRF3/7) (G) or pLuc-PRD(II)₂ (NF-κB) (H). Luciferase activity was measured 24 hrs following plasmid transfection. Discontinuity in graph indicates a broken Y-axis (H).

Figure 4. TRAF5 is required for IRF3/7 and NF-κB activation in response to RLR signaling.

A HEK293T cells were initially transfected with siRNAs targeting GFP, TRAF3, or TRAF5. 48 hrs later, cells were transfected with pLuc-IFNβ and 6 hrs later, poly I:C was transfected. Luciferase activity was measured 18 hours following poly I:C transfection. In the inset figure, HA-TRAF3 or HA-TRAF5 was cotransfected with siRNAs targeting GFP, TRAF3, or TRAF5 and anti-HA immunoblots were performed on prepared whole cell lysates. B HEK293T cells were initially transfected with siRNAs targeting GFP, TRAF3, or TRAF5. 48 hrs later, cells were transfected with HA-IRF3, lysates were prepared, and anti-HA immunoprecipitates were probed for phosphorylated IRF3 (Ser 396) or total HA-IRF3. Cells were transfected with poly I:C 8 hrs prior to cell lysis. C HEK293T cells were transfected with siRNAs targeting GFP, TRAF3, or TRAF5. Cell lysates probed for phospho-STAT1 (Y701) or total STAT1. Poly I:C was transfected 8 hrs prior to cell lysis. D, E Luciferase reporter assays were performed in HEK293T cells transfected with siRNA targeting GFP, MAVS, or TRAF5. 48 hrs following siRNA transfection, cells were transfected with pLuc-IFN-β, pLuc-PRD(III-I)₃ (IRF3/7) or pLuc-PRD(II)₂ (NF-κB). SeV infection was performed 4 hrs following transfection and 20 hrs prior to luciferase assay.

Figure 5. TRAF5 is required for the induction of gene expression downstream of RLR signaling.

A HEK293T cells were transfected with siRNAs targeting GFP or TRAF5. 48 hrs later, cells were mock infected or infected with SeV for 8 hrs and total RNA was prepared. Ifnb, IP10, and Tnfa mRNA levels were measured by real-time RT-PCR using specific primers as described in the Materials and Methods section. GADPH was used as an internal control. Asterisk, P<0.01 versus GFP siRNA + infection (n = 2), double asterisk, P<0.005 versus GFP siRNA + infection (n = 2). B, HEK293 cells were transfected with siRNAs targeting GFP, MAVS, or TRAF5. 48 hrs later, cells were mock infected or infected with SeV. Tissue culture supernatants were collected 24 hrs following infection and human IFN-β ELISA was performed. The concentration of IFN-β is shown (mean +/- sd). ND, not detectable. Asterisk, P = 0.01 versus GFP siRNA + infection (n = 2), double asterisk, P<0.005 versus GFP siRNA + infection (n = 2). C, HEK293 cells were transfected with siRNAs targeting GFP or TRAF5. 48 hrs later, cells were transfected with poly I:C for 2 or 6 hrs or mock treated for 6 hrs. Nuclear extracts were prepared and analyzed for p65 expression by immunoblotting. TBP was used as a loading control. D, HEK293 cells were transfected with siRNAs targeting GFP, MAVS, TRAF3, or TRAF5. 48 hrs later, cells were transfected with poly dA:dT or mock treated. Tissue culture supernatants were collected for human IFN-β ELISA 24 hrs after poly dA:dT transfection. The concentration of IFN-β is shown (mean +/− sd). ND, not detectable. Asterisk, P<0.025 versus GFP siRNA + poly dA:dT (n = 2), double asterisk, P<0.001 versus GFP siRNA + poly dA:dT (n = 2). E, HEK293 cells were transfected with siRNAs as in (D). 48 hrs later, cells were transfected with poly dA:dT or mock treated for 8 hrs and lysates prepared. Either 1 or 4 ug/ml of poly dA:dT was used as indicated. Lysates were probed for phospho-STAT1 (Y701) or total STAT1. F, Yeast were transformed with the bait or prey constructs indicated as described in the Materials and Methods section and streaked onto minimal SD media lacking leucine and tryptophan (SD-LW) or leucine, tryptophan, adenine, and histidine (SD-LWAH).

Figure 6. NEMO is recruited to the MAVS CARD in a TRAF3/5-dependent manner.

A HEK293T cells were transfected with AU1-NEMO and FLAG-MAVS CARD-FPK3 or FLAG-MAVS CARDmt-FPK3 where indicated. MAVS CARDmt contains a substitution of alanine for tryptophan at residue 68. Lysates were prepared and anti-FLAG immunoprecipitates were probed for AU1-NEMO. Lysates were probed for FLAG-MAVS-CARD-FPK3 protein and AU1-NEMO. AP1510 was added to media 12 hrs prior to cell lysis. B HEK293T cells were transfected with AU1-NEMO and FLAG-MAVS CARD-FPK3, HA-TRAF3 or HA-TRAF5 where indicated. Anti-FLAG-immunoprecipitates were probed for AU1-NEMO. Lysates were immunoblotted for FLAG-MAVS proteins and AU1-NEMO. AP1510 was added to media 12 hrs prior to cell lysis. C HEK293T cells first transfected with siRNAs targeting GFP, TRAF3, or TRAF5. 48 hrs later, cells were transfected with FLAG-MAVS CARD-FPK3 and AU1-NEMO. Anti-FLAG-immunoprecipitates were probed for AU1-NEMO. Lysates were immunoblotted for FLAG-MAVS proteins and AU1-NEMO. AP1510 was added to media 12 hrs prior to cell lysis.

Figure 7. Schematic representation of model for signal transduction downstream of MAVS.

MAVS homodimers, induced by binding of the activated forms of the RLRs RIG-I and MDA-5, are able to signal downstream in a TRAF3 and TRAF5-dependent manner to activate IRF3/7 and in a TRAF5-dependent manner to activate NF-κB. NEMO may be recruited to the MAVS signaling complex by ubiquitinated TRAF3 and TRAF5, and may lead to IRF3 phosphorylation through the recruitment of TANK and TBK1. Alternatively, NEMO may be recruited to a distinct MAVS signaling complex by ubiquitinated TRAF5, thereby promoting the recruitment IKKα and IKKβ, and leading to NF-κB activation. The mechanisms underlying the selective recruitment of signaling molecules to the MAVS signaling apparatus remain to be determined.