The signaling adaptor MAVS forms prion-like aggregates to activate an innate antiviral immune response after viral infection. However, the molecular mechanisms that regulate MAVS aggregation are poorly understood. Here we identified TRIM31, an E3 ubiquitin ligase of the TRIM family of proteins, as a regulator of MAVS aggregation. TRIM31 was recruited to mitochondria after viral infection and specifically regulated antiviral signaling mediated by RLR pattern-recognition receptors. TRIM31-deficient mice were more susceptible to infection with RNA virus than were wild-type mice. TRIM31 interacted with MAVS and catalyzed the Lys63 (K63)-linked polyubiquitination of Lys10, Lys311 and Lys461 on MAVS. This modification promoted the formation of prion-like aggregates of MAVS after viral infection. Our findings reveal new insights in the molecular regulation of MAVS aggregation and the cellular antiviral response through TRIM31-mediated K63-linked polyubiquitination of MAVS.
At a glance
Innate immunity provides the first line of defense against invading pathogens. Activation of innate immunity requires the recognition of pathogen-associated molecular patterns through pattern-recognition receptors1, 2. Several types of pattern-recognition receptors have been identified that recognize viral nucleic acid to regulate viral infection. Toll-like receptor 3 (TLR3), TLR7, TLR8 and TLR9 detect viral RNA and DNA in the endosome3, whereas the RNA helicases RIG-I and MDA-5 sense viral RNA in the cytoplasm4. Several sensors that recognize cytosolic DNA, such as cGAS (cyclic GMP–AMP synthase) and IFI16, have been identified5, 6. These pattern-recognition receptors recognize invading viruses and initiate a series of signaling events that lead to the production of type I interferons (IFN-α and IFN-β, collectively called 'IFN-α/β' here). IFN-α/β further activate downstream signaling pathways that lead to the transcriptional induction of a wide range of genes encoding antiviral products, and those antiviral products act together to elicit a cellular antiviral response through various mechanisms7.
RIG-I and MDA-5 harbor two N-terminal caspase-recruitment (CARD) domains, a central DExD–H-box helicase domain and a C-terminal domain8. The DExD–H-box helicase domain and C-terminal domain are important for binding viral RNA, whereas the two CARDs are responsible for downstream signaling. Activation of RIG-I and MDA-5 by infection with an RNA virus results in the recruitment of MAVS (also known as IPS1, VISA or Cardif) to the RLR signalosome, which leads to IFN-β production via activation of the transcription factor IRF3 and of the complex of the transcription factor NF-κB and IRF3 activator TBK1 and the kinase IKKi9, 10, 11, 12. MAVS is a 540-residue protein composed of a N-terminal CARD, a proline-rich domain and a C-terminal transmembrane domain. The transmembrane domain anchors MAVS to the mitochondrial outer membrane, and the CARD is responsible for the recruitment of MAVS to RIG-I and MDA-5 for initiation of the antiviral innate immune response8.
A published study has reported that viral infection induces a conformational change in MAVS, which leads to the formation of prion-like functional aggregates that provide a sensitive trigger for antiviral signaling13. This study involved an elegant in vitro reconstitution system and demonstrated that the formation of MAVS aggregates needs RIG-I and unanchored Lys63 (K63)-linked ubiquitin chains13. In vitro structural studies have shown that the two CARDs of RIG-I bind to unanchored K63-linked ubiquitin chains to form a tetramer with four 'two-CARD' domains, which serves as a template to nucleate the MAVS aggregates in the mitochondrion14. However, the detailed molecular mechanisms of MAVS aggregation, especially in cells, are not clear.
MAVS must be localized to mitochondria to function10. We hypothesized that proteins localized to mitochondria could function as regulators for MAVS. Here we identified a mitochondria-localized protein, the E3 ubiquitin ligase TRIM31, as a positive regulator of MAVS. TRIM31 deficiency attenuated the innate antiviral response to infection with an RNA virus both in vitro and in vivo. TRIM31 interacted with MAVS and promoted K63-linked polyubiquitination on Lys10, Lys311 and Lys461 of MAVS after viral infection. Notably, TRIM31-mediated K63-linked polyubiquitination facilitated the formation of prion-like MAVS aggregates.
Regulation of RLR-induced IFN-β signaling by TRIM31
Mitochondria have emerged as important platforms for intracellular antiviral signaling. Several mitochondrial proteins, such as MAVS, NLRX1 and MARCH5, have been linked to modulating antiviral signaling10, 15, 16. TRIM31 is a member of TRIM family of proteins that is encoded in the locus encoding the major histocompatibility complex (MHC) class I proteins. It has also been reported TRIM31 localizes mainly to the cytoplasm, but a fraction of TRIM31 has been found to be associated with the mitochondria17. To investigate the possible role of TRIM31 in antiviral signaling, we studied the cellular localization of TRIM31 in HEK293T human embryonic kidney cells and mouse primary peritoneal macrophages. Fluorescence microscopy images showed that ectopically expressed TRIM31 tagged with green fluorescent protein (GFP) in HEK293T cells was distributed in the cytoplasm with very little colocalization with mitochondria before infection with Sendai virus (SeV) (Supplementary Fig. 1a,b). Infection with SeV greatly increased the colocalization of TRIM31 with mitochondria (Supplementary Fig. 1a). We found, by immunoblot analysis, that TRIM31 was localized to mitochondria in primary peritoneal macrophages after infection with SeV (Supplementary Fig. 1b); the expression of TRIM31 protein was not altered after infection with SeV (Supplementary Fig. 1c). These data indicated that TRIM31 was recruited to mitochondria after infection with SeV.
To explore the potential role of TRIM31 in antiviral signaling, we designed a small interfering RNA (siRNA) that targeted mouse Trim31 and transfected it into primary peritoneal macrophages. We found that expression of endogenous Trim31 mRNA and TRIM31 protein was much lower in cells transfected with the Trim31-specific siRNA than in those transfected with control (non-targeting)siRNA (Fig. 1a). siRNA-mediated knockdown of Trim31 expression significantly decreased the expression of Ifnb1 mRNA (encoding IFN-β) and production of IFN-β protein in macrophages after infection with SeV (Fig. 1a). As a specificity control, we knocked down expression of TRIM30-α using siRNA and found no effect on SeV-induced expression of Ifnb1 (Supplementary Fig. 1d), consistent with a published report concluding that TRIM30-α has negligible effects in RLR-mediated interferon signaling18. Conversely, we found that siRNA-mediated knockdown of TRIM26 expression increased SeV-induced expression of Ifnb1 (Supplementary Fig. 1d), as TRIM26 has been identified as a negative regulator for RLR signaling19. Similar to the data obtained with mouse macrophages, we observed that siRNA-mediated knockdown of TRIM31 expression in human THP-1 monocytic cells decreased SeV-induced expression of IFNB1 (Supplementary Fig. 1e). In contrast, overexpression of TRIM31 in HEK293T cells significantly increased IFNB1 expression after transfection of the TLR3 ligand poly(I:C) (polyinosinic-polycytidylic acid) or infection with SeV (Fig. 1b). Similar to the data obtained with HEK293T cells (Fig. 1b), we observed that overexpression of Flag-tagged TRIM31 in HeLa cells greatly enhanced the expression of IFNB1 mRNA after transfection of poly(I:C) or infection with SeV (Supplementary Fig. 1f).
To directly investigate the effect of TRIM31 on antiviral responses, we used vesicular stomatitis virus (VSV). We found that use of a Trim31-specific siRNA significantly reduced the expression of Ifnb1 in primary peritoneal macrophages after infection with VSV (Fig. 1c). Consistent with that result, we found, by quantitative PCR (qPCR) and plaque assay, that the expression of VSV-specific mRNA and VSV titers were greater in macrophages in which Trim31 was knocked down than in those transfected with control siRNA (Fig. 1c). Overexpression of TRIM31 in HEK293T cells significantly increased IFNB1 expression after infection with VSV relative to its expression in cells expressing the control vector (Fig. 1d). Consistent with those data, the expression of VSV mRNA and VSV titers were decreased after overexpression of TRIM31 in HEK293T cells (Fig. 1d). We also infected HEK293T cells that ectopically expressed TRIM31 with VSV expressing GFP (VSV–GFP). Fluorescence microscopy of HEK293T cells showed that overexpression of TRIM31 substantially inhibited viral replication relative to such replication in cells expressing the control vector (Fig. 1e). RNA viruses, such as SeV and VSV, or poly(I:C) that is transfected into cells is recognized by RIG-I and MDA-5. To directly confirm that TRIM31 was affecting RLR-mediated IFN-β signaling, we used RIG-I-deficient mouse embryonic fibroblasts (MEFs). We reconstituted RIG-I expression in RIG-I-deficient (Ddx58−/−) MEFs and found that SeV-induced IFN-β expression was restored in those cells relative to its expression MEFs controls vector, whereas overexpression of TRIM31 further increased SeV-induced Ifnb1 expression in Ddx58−/− MEFs relative to that in MEFs without overexpression of TRIM31 (Fig. 1f). TRIM31 alone was unable to inhibit VSV replication in Ddx58−/− MEFs, whereas TRIM31 greatly inhibited VSV replication in Ddx58−/− MEFs that were reconstituted with a RIG-I expression plasmid (Fig. 1f). Together these data suggested that TRIM31 positively regulated RLR-mediated IFN-β signaling.
Impaired cellular antiviral response in TRIM31 deficiency
To investigate the physiological role of TRIM31, we generated Trim31−/− mice through TALEN ('transcription-activation-like effector nuclease') technology. Three lines of deficient mice with different deletions were constructed: Trim31−/− (missing one cytosine (crossed out): GACTGTGGGCACAAC), Trim31−/− (G) (missing one guanosine: GACTGTGGGCACAAC) and Trim31−/− (5) (missing the five bases GGGCA: GACTGT GGGCACAAC (Supplementary Fig. 2a). Successful mutation of Trim31 was confirmed by sequencing of the PCR fragments in the TALEN-targeting region amplified from genomic DNA (Supplementary Fig. 2a) and by immunoblot analysis of TRIM31 in peritoneal macrophages from Trim31+/+ and Trim31−/− mice (Supplementary Fig. 2b). Trim31−/− mice were viable, normal in size and without gross physiological or behavioral abnormalities (data not shown).
We prepared peritoneal macrophages from Trim31+/+ and Trim31−/− mice; we then stimulated them with lipopolysaccharide (LPS), poly(I:C) or RNA mimics, or infected them with SeV or herpes simplex virus type 1 (HSV-1). We observed that Trim31−/− macrophages had lower Ifnb1 expression after infection with SeV or stimulation with RNA mimics (5′-pppRNA) than that of their Trim31+/+ counterparts (Fig. 2a). The expression of Ccl5 and Cxcl10, downstream genes in IFN-β signaling, was also lower in Trim31−/− macrophages than in Trim31+/+ macrophages (Fig. 2a). In contrast, Ifnb1 expression -induced by LPS, poly(I:C) or HSV-1 was not impaired in Trim31−/− macrophages (Fig. 2a). IFN-β secretion was also lower in Trim31−/− macrophages after infection with SeV than in their Trim31+/+ counterparts (Fig. 2b). SeV-induced expression of Ifnb1, Ccl5 and Cxcl10 was also much lower in peritoneal macrophages prepared from the other two TRIM31-deficient mice lines, Trim31−/− (G) and Trim31−/− (5), than in Trim31+/+ peritoneal macrophages (Supplementary Fig. 2c). Therefore, we used Trim31−/− mice in the following experiments. Infection with SeV or treatment with 5′-pppRNA resulted in much lower expression of Ifnb1, Ccl5 and Cxcl10 in bone-marrow-derived macrophages (BMDMs) prepared from Trim31−/− mice than in those from Trim31+/+ mice, whereas LPS-, poly(I:C)- or HSV-1-induced expression of Ifnb1 was not impaired (Supplementary Fig. 2d). To further confirm the function of TRIM31, we prepared MEFs from Trim31+/+ and Trim31−/− mice. Ifnb1 expression induced by SeV or 5′-pppRNA, but not that induced by LPS, poly(I:C) or HSV-1, was lower in Trim31−/− MEFs than in Trim31+/+ MEFs (Supplementary Fig. 2e).
Trim31−/− macrophages expressed less Ifnb1 in response to infection with VSV than did Trim31+/+ macrophages (Fig. 2c). VSV replication was significantly enhanced in Trim31−/− macrophages (Fig. 2c), whereas replication of the DNA virus HSV-1 in Trim31−/− macrophages was not affected (Fig. 2d). Together these data demonstrated that TRIM31 specifically regulated IFN-β signaling and antiviral innate immune responses mediated by RLRs but not those mediated by TLRs or DNA sensors.
Regulation of infection with an RNA virus by TRIM31 in vivo
To assess the relevance of TRIM31 in vivo, we challenged Trim31+/+ and Trim31−/− mice with VSV and found, by qPCR analysis, that the expression of Ifnb1 in the spleen, liver and lungs of Trim31−/− mice was significantly lower than that in those organs from Trim31+/+ mice, after infection with VSV (Fig. 3a). ELISA also showed less IFN-β in serum from Trim31−/− mice than in serum from Trim31+/+ mice (Fig. 3b). In contrast, the production of IFN-β after challenge with HSV-1 was barely affected in Trim31−/− mice relative to that in Trim31+/+ mice (Supplementary Fig. 3a). Consistent with less production of IFN-β, we found that VSV titers and replication were significantly greater in the spleen, liver and lungs of Trim31−/− mice than in those from Trim31+/+ mice (Fig. 3c–e). However, we did not observe any difference in the copy number of HSV-1 genomic DNA and viral titer of HSV-1 in the brains of Trim31−/− mice versus those of Trim31+/+ mice (Supplementary Fig. 3b). Moreover, Trim31−/− mice were less resistant than Trim31+/+ mice to infection with VSV but not to infection with HSV-1 (Fig. 3f and Supplementary Fig. 3c). Hematoxylin-and-eosin staining showed greater infiltration of immune cells and injury in the lungs of Trim31−/− mice, relative to that in the lungs of Trim31+/+ mice, after infection with VSV (Fig. 3g). These data demonstrated that Trim31−/− mice were more susceptible than wild-type mice to infection with an RNA virus.
Promotion of RLR-mediated innate antiviral signaling by TRIM31
To investigate the effect of TRIM31 on RLR-mediated signaling, we performed several experiments. We transfected a TRIM31-expression plasmid into HEK293T cells and saw enhanced SeV-induced activation of an IFNB1 promoter reporter in a TRIM31-dose-dependent manner (Fig. 4a). Notably, overexpression of TRIM31 increased SeV-induced activation of a IFNB1 reporter containing only PRD I–III elements, which represent the binding site for IRF3 (ref. 20) (Fig. 4a). SeV-induced phosphorylation of TBK1 and IRF3 was greater in HEK293T cells transfected with the TRIM31-expression plasmid than in those transfected with empty vector (Fig. 4b). We found that TRIM31 alone was unable to mediate SeV-induced activation of IRF3 in Ddx58−/− MEFs; however, expression of TRIM31 enhanced the SeV-induced activation of IRF3 in Ddx58−/− MEFs reconstituted with a RIG-I-expression plasmid (Fig. 4c).
Phosphorylation of TBK1 and IRF3 after infection with SeV was lower in Trim31−/− macrophages than in Trim31+/+ macrophages (Fig. 4d). The dimerization of IRF3 was also lower in macrophages from Trim31−/− mice infected with SeV than in their Trim31+/+ counterparts (Fig. 4e). Trim31−/− macrophages had less translocation of IRF3 to the nucleus than did Trim31+/+ macrophages in response to infection with SeV (Fig. 4f). In contrast, translocation of IRF3 to the nucleus was not affected in Trim31−/− macrophages after infection with HSV-1 (Fig. 4g). Together these data demonstrated that TRIM31 positively regulated RLR-mediated innate antiviral signaling.
Targeting of MAVS by TRIM31
To identify the molecules regulated by TRIM31, we first assessed the effects of TRIM31 on IFNB1 expression in HEK293T cells mediated by the molecules in the RLR signaling pathway. We knocked down TRIM31 expression and found that the expression of IFNB1 mRNA mediated by RIG-I, MDA-5 or MAVS was attenuated, whereas that mediated by TBK1 or IRF3-5D (a constitutively active IRF3 variant in which Ser396, Ser398, Ser402, Ser404 and Ser405 were replaced by the phosphomimetic aspartate) was not impaired (Fig. 5a). We overexpressed TRIM31 and found increased expression of IFNB1 mRNA mediated by RIG-I, MDA-5 or MAVS, but not that mediated by IKKε, TBK1 or IRF3-5D (Fig. 5b). We also observed that TRIM31 substantially increased activation of IFNB1 reporter induced by MDA-5, RIG-I or MAVS, but it had no effect on such activation induced by IKKε, TBK1 or IRF3-5D (Supplementary Fig. 4a). siRNA-mediated knockdown of TRIM31 expression in HEK293T cells inhibited activation of IFNB1 reporter and the interferon-stimulated response element (ISRE) reporter mediated by RIG-I or MAVS but not that mediated by TBK1 (Supplementary Fig. 4b). These data suggested that TRIM31 might specifically target MAVS to regulate the RLR signaling.
To confirm that TRIM31 targets MAVS, we investigated their interaction and colocalization. Co-immunoprecipitation experiments in HEK293T cells showed that TRIM31 associated with MAVS but not with RIG-I, TBK1, IRF3 or STING (Fig. 5c, left, and Supplementary Fig. 4c). Notably, MAVS interacted specifically with TRIM31 but not with TRIM26 or TRIM30-α (Fig. 5c, right). As a positive control, we found that MAVS interacted with TRIM25 (Fig. 5c, right). We performed confocal microscopy and found that TRIM31 colocalized with MAVS (Fig. 5d). To further verify that TRIM31 associated with MAVS, we expressed Flag-tagged TRIM31 and Myc-tagged MAVS in an in vitro protein-expression system, mixed the recombinant proteins together and monitored them in immunoprecipitation assays. Myc–MAVS co-imunoprecipitated with Flag–TRIM31 (Fig. 5e), which indicated that TRIM31 interacted with MAVS in vitro. We did not detect an interaction between TRIM26 and MAVS under the same conditions (Fig. 5e). Furthermore, in co-immunoprecipitation experiments, we found that endogenous TRIM31 and MAVS formed a complex in macrophages after infection with SeV (Fig. 5f).
TRIM31 is composed of a RING-finger domain (amino acids 16–57), a B-box (amino acids 90–129) and a coiled-coil (C-C) motif (amino acids 126–307) (Fig. 5g). To determine which domain of TRIM31 was necessary for interaction with MAVS, we constructed several TRIM31 deletion mutants and assessed them in co-immunoprecipitation experiments (Fig. 5g). We found that TRIM31-ΔC-C (in which the C-C motif was deleted) lost the ability to interact with MAVS (Fig. 5g, lane 6). TRIM31(C53A,C56A), in which the conserved cysteine residues at positions 53 and 56 within the RING domain were replaced with alanine, also interacted with MAVS (Fig. 5g, lane 7). MAVS is composed of a CARD (amino acids 1–90), a proline-rich domain (amino acids 91–172) and a transmembrane domain (amino acids 513–540). Deletion of the MAVS proline-rich domain ablated the interaction between MAVS and TRIM31 (Fig. 5h). Together these data demonstrated that TRIM31 physically interacted with MAVS through the C-C motif of TRIM31 and the proline-rich domain of MAVS.
TRIM31 catalyzes K63-linked polyubiquitination of MAVS
TRIM31 is a RING type E3 ubiquitin ligase reported to have auto-ubiquitination activity17, 21. To investigate whether TRIM31 regulates MAVS signaling through its E3 ligase activity, we assessed TRIM31-mediated ubiquitination of MAVS. TRIM31-mediated polyubiquitination of MAVS was readily detectable in HEK293T cells transfected with plasmids expressing Myc-tagged MAVS and hemagglutinin (HA)-tagged ubiquitin in the presence of a plasmid expressing Flag-tagged TRIM31 (Fig. 6a, lane 3). As a control, we found that polyubiquitination of RIG-I was not induced by Flag–TRIM31 but was readily induced by Flag–TRIM25, as reported previously22 (Supplementary Fig. 5a). TRIM31(C53A,C56A) lost the ability to increase the polyubiquitination of MAVS (Fig. 6a, lane 4), which indicating that the E3 ligase activity of TRIM31 was required for the polyubiquitination of MAVS.
To investigate the type of TRIM31-mediated polyubiquitination of MAVS, we used vectors expressing HA-tagged mutants ubiquitin(K48) and ubiquitin(K63), which contain substitution of arginine for all lysine residues except the lysine at position 48 or position 63, respectively. TRIM31 catalyzed polyubiquitination of MAVS in the presence of HA-tagged wild-type ubiquitin (HA–ubiquitin(WT)) and HA–ubiquitin(K63) but not in the presence of HA–ubiquitin(K48) (Fig. 6b). Furthermore, we performed in vitro ubiquitination assays to confirm that TRIM31 promoted K63-linked polyubiquitination of MAVS. TRIM31 was found to catalyze the polyubiquitination of MAVS in the presence of ubiquitin(WT) and ubiquitin(K63) but not in the presence of ubiquitin(K48) in these in vitro ubiquitination assay (Fig. 6c). We further performed in vitro ubiquitination assays using the mutants ubiquitin(K48R) and ubiquitin(K63R), which contain a single lysine-to-arginine substitution at position 48 or 63, respectively. MAVS was ubiquitinated by TRIM31 in the presence of ubiquitin(WT) and ubiquitin(K48R) but not in the presence of ubiquitin(K63R) (Fig. 6d). To confirm the TRIM31-induced ubiquitination of endogenous MAVS, we measured the polyubiquitination of MAVS in SeV-infected primary macrophages prepared from Trim31+/+ and Trim31−/− mice and found that endogenous MAVS was robustly ubiquitinated with both K48-linked and K63-linked chains (Fig. 6e). However, the total amount of K63-linked ubiquitination of MAVS was much lower in Trim31−/− macrophages than in Trim31+/+ macrophages, whereas there was little difference between Trim31−/− macrophages and Trim31+/+ macrophages in the amount of K48-linked ubiquitination of MAVS (Fig. 6e). In a control experiment, we found that SeV-induced polyubiquitination of MAVS was not impaired in Trim26−/− macrophages, as TRIM26 has been reported to target IRF3 (ref. 19) (Fig. 6e).
K63-linked polyubiquitination has often been reported to activate signaling via the innate immune system23. Consistent with that, TRIM31 expression increased SeV-induced activation of the IFNB1 and ISRE reporters in HEK293T cells, inhibited VSV replication and increased SeV-induced phosphorylation of IRF3 (Supplementary Fig. 5b–d). However, TRIM31(C53A,C56A) lost the ability to regulate IFNB1 expression or an antiviral response (Supplementary Fig. 5b–d).
MAVS has 14 lysine residues present in different domains (Fig. 6f). To identify the lysine residues responsible for TRIM31-mediated polyubiquitination, we first generated the mutant MAVS-K0, in which all of the lysine residues in MAVS were replaced with arginine. Then, we reintroduced individual lysine residues into MAVS-K0 to generate the single-lysine mutants (Fig. 6f). Cotransfection and co-immunoprecipitation analyses of HEK293T cells showed that TRIM31 induced polyubiquitination of MAVS(WT) and the mutants MAVS(K10), MAVS(K311) and MAVS(K461) (Fig. 6g). We further constructed the mutants MAVS(K10R), MAVS(K10R,K311R) and MAVS(K10R,K311R,K461R), in which the lysine residues at positions 10, 311 and 461 were replaced with arginine (Fig. 6f). Co-immunoprecipitation analysis of HEK293T cells showed that TRIM31-induced polyubiquitination of the MAVS mutants was gradually decreased with an increase in the number of lysine-to-arginine substitutions relative to that of cells expressing MAVS(WT) (Fig. 6h). Notably, the K63-linked ubiquitination of MAVS(K10R,K311R,K461R) decreased to a level similar to that of MAVS(K0) (Fig. 6h). Together these data demonstrated that TRIM31 catalyzeds the K63-linked polyubiquitination of MAVS on Lys10, Lys311 and Lys461.
Promotion of MAVS aggregate formation by TRIM31
MAVS forms prion-like aggregates that potently propagate downstream signaling after infection with an RNA virus13. Infection with SeV induced the formation of MAVS aggregates in peritoneal macrophages, as measured by semi-denaturing detergent agarose-gel electrophoresis (SDD–AGE) (Fig. 7a), whereas the amount of SeV-induced aggregation of MAVS was much lower in Trim31−/− macrophages than in Trim31+/+ macrophages (Fig. 7a). Similarly, the amount of MAVS aggregation induced by infection with SeV was lower in Trim31−/− MEFs than in Trim31+/+ MEFs (Supplementary Fig. 6a). In contrast, overexpression of TRIM31 in HEK293T cells greatly increased the formation of MAVS aggregates relative to their formation in cells transfected with control vector (Supplementary Fig. 6b).
To confirm that the decreased amount of MAVS aggregation in Trim31−/− MEFs was due to the deletion of TRIM31, we reintroduced an expression plasmid encoding Flag-tagged mouse TRIM31 (mTRIM31) into Trim31−/− MEFs. This reintroduction restored the SeV-induced formation of MAVS aggregates (Fig. 7b, lane 3). Notably, reintroduction of mTRIM31(C52A,C55A) in Trim31−/− MEFs did not restore SeV-induced formation of MAVS aggregates (Fig. 7b, lane 4). Consistent with the restoration of MAVS polymerization after the reintroduction of mTRIM31 in Trim31−/− MEFs, SeV-induced expression of Ifnb1, Ccl5 and Cxcl10 was also restored in these MEFs, whereas expression of mTRIM31(C52A,C55A) in Trim31−/− MEFs had no such effect (Fig. 7b and Supplementary Fig. 6c).
To confirm that the aggregation of MAVS was directly regulated by TRIM31-induced polyubiquitination of MAVS, we isolated crude mitochondria (P5) from macrophages that were either infected with SeV or left uninfected. The mitochondria were incubated in an in vitro ubiquitination assay buffer containing ubiquitin, the ubiquitin-activating enzyme E1, the ubiquitin-conjugating enzyme UbcH5a and TRIM31. In SDD–AGE experiments, we found no formation of MAVS aggregates in mitochondria before infection with SeV (Fig. 7c, left; lanes 1–3), whereas infection with SeV induced MAVS aggregation (Fig. 7c, left; lanes 4–6). In the presence of E1, E2 and TRIM31, SeV-induced aggregation of MAVS was greatly increased (Fig. 7c, left; lane 6). Of note, we found that TRIM31 promoted the polyubiquitination of mitochondrial MAVS with or without infection with SeV, in the in vitro ubiquitinaton system (Fig. 7c, right). Furthermore, we used ubiquitin(K63) and ubiquitin(K48) in the in vitro ubiquitination system and demonstrated that TRIM31-mediated K63-linked ubquitination, but not K48-linked ubquitination, was required for the formation of MAVS aggregates following infection with SeV (Fig. 7d, lanes 2 and 6). Furthermore, we found that there was no aggregation of MAVS in Ddx58−/− MEFs, even in the presence of TRIM31, whereas transfection of a RIG-I-expression plasmid into Ddx58−/− MEFs restored SeV-induced aggregation of MAVS (Fig. 7e). Notably, overexpression of TRIM31 further increased the SeV-induced aggregation of MAVS (Fig. 7e).
To directly address the function of TRIM31-induced ubiquitination in the activation and aggregation of MAVS, we studied Mavs−/− MEFs reconstituted with wild-type MAVS or its ubiquitination-site mutants MAVS(K10), MAVS(K311), MAVS(K461), MAVS(K10R), MAVS(K10R,K311R) and MAVS(K10R,K311R,K461R). Infection with SeV did not stimulate the production of IFN-β by Mavs−/− MEFs, whereas transfection of wild-type MAVS-expression plasmids into Mavs−/− MEFs restored the expression of Ifnb1 (Supplementary Fig. 6d). The overexpression of TRIM31 further increased the SeV-induced activation of the IFNB1 and ISRE reporters in Mavs−/− MEFs reconstituted with wild-type MAVS (Supplementary Fig. 6e). However, SeV-induced activation of the IFNB1 and ISRE reporters in Mavs−/− MEFs reconstituted with the mutant MAVS(K10R), MAVS(K10R,K311R) or MAVS(K10R,K311R,K461R) was gradually decreased with an increase in the number of lysine-to-arginine substitutions in MAVS in the presence of TRIM31 (Supplementary Fig. 6e). Notably, MAVS(K10R,K311R,K461R)-mediated activation of the IFNB1 and ISRE reporters in the presence of TRIM31 decreased to a level similar to that seen after transfection of a construct expressing MAVS alone (Supplementary Fig. 6e). Consistent with the observation that TRIM31 did not increase the MAVS(K10R,K311R,K461R)-induced activation of IFN-β, we found that expression of TRIM31 accelerated MAVS aggregation only in Mavs−/− MEFs transfected with the MAVS(WT)-expressing plasmid, not in transfected with the MAVS(K10R,K311R,K461R)-expressing plasmid (Fig. 7f). The MAVS(K10R)-, MAVS(K311R)- or MAVS(K461R)-induced expression of Ifnb1 and aggregation of MAVS were partially diminished in the presence of TRIM31 after viral infection (Supplementary Fig. 6f). Consistent with the MAVS-aggregation data, SeV-induced expression of Ifnb1, Ccl5 and Cxcl10 was enhanced only in the Mavs−/− MEFs transfected with MAVS(WT)-expressing plasmid, not in those transfected with MAVS(K10R,K311R,K461R)-expressing plasmid in the presence of TRIM31 (Fig. 7g).
To investigate whether TRIM31 interacted with MAVS independently of MAVS polymerization, we studied the polymerization-deficient mutant MAVS(W56R). As expected, aggregation of MAVS(W56R) was abolished, relative to the aggregation of MAVS(WT), when overexpressed in Mavs−/− MEFs in the presence or absence of TRIM31 (Supplementary Fig. 6g). Consistent with the loss of MAVS aggregation, MAVS(W56R) was unable to induce Ifnb1 expression after infection with SeV (Supplementary Fig. 6g), which indicated that the action of TRIM31 on MAVS was dependent exclusively on the ability of MAVS to polymerize. However, in co-immunoprecipitation experiments, we found that, similarly to MAVS(WT), MAVS(W56R) interacted with TRIM31 (Fig. 7h). Notably, we found that MAVS(W56R) was ubiquitinated by TRIM31 in a manner similar to the ubiquitination of MAVS(WT) (Fig. 7h). These data suggested that TRIM31 interacted with MAVS independently of MAVS polymerization. Together these data indicated that TRIM31 bound to inactive MAVS after viral infection and catalyzed K63-linked polyubiquitination on Lys10, Lys311 and Lys461 to further promote aggregation of MAVS and thus enhance IFN-β production and antiviral signaling.
As a mitochondrial membrane protein, MAVS functions as a critical adaptor of the RLR signaling pathway that links upstream recognition of viral RNA to downstream signal activation in the antiviral response24, 25. Here we provided several lines of evidence to demonstrate that TRIM31 served as a specific regulator of MAVS-mediated innate antiviral immunity through direct conjugation of K63-linked ubiquitin chains to MAVS on positions Lys10, Lys311 and Lys461. Therefore, the production of IFN-β and the antiviral immune responses were not impaired by a TRIM31 deficiency after innate immune signaling induced by TLRs or intracellular DNA sensors, which use the adaptors TRIF and STING, respectively, to initiate an antiviral response.
It has been reported that MAVS forms functional prion-like aggregates after viral infection, which are required for the activation of the RLR signaling13. We found that TRIM31 was essential for the formation of MAVS aggregates after viral infection. TRIM31 deficiency greatly decreased the amount of MAVS aggregation after infection with SeV. Reconstitution of TRIM31 expression in TRIM31-deficient MEFs restored SeV-induced aggregation of MAVS. Notably, cells expressing TRIM31(C52A,C55A) lost the ability to restore SeV-induced aggregation of MAVS. In vitro ubiquitination assays showed that TRIM31 promoted the formation of MAVS aggregates in mitochondria after viral infection in the presence of the ubiquitin-activating enzyme E1 and ubiquitin-conjugating enzyme E2. Furthermore, unlike MAVS(WT), MAVS(K10R,K311R,K461R) was unable to form aggregates in the presence of TRIM31 in MAVS-deficient MEFs.
The aggregation of MAVS is required for its activation and an RLR-mediated antiviral response. Consistent with that, we found that TRIM31 deficiency resulted in a substantial reduction in the RNA-virus-mediated activation of IRF3 and expression of Ifnb1. Trim31−/− mice were more susceptible to infection with an RNA virus than were their wild-type counterparts. Thus, our study has revealed a novel mechanism for the regulation of MAVS aggregation through K63-linked ubiquitination.
The mechanism by which K63-linked polyubiquitination of MAVS regulates the aggregation and activation of MAVS deserves further investigation. We found that TRIM31 promoted the polyubiquitination of mitochondrial MAVS in the absence of infection with SeV, whereas TRIM31 promoted the aggregation of MAVS only after viral infection. These data suggested that, except for the formation of K63-linked polyubiquitin chains on MAVS, TRIM31-induced MAVS aggregation required other molecules induced by infection with SeV. The most likely candidate for this would be RIG-I, as activation of RIG-I after viral infection is required for the aggregation of MAVS. Indeed, our data showed that SeV-induced MAVS aggregation required RIG-I, as we did not detect aggregation of MAVS in RIG-I-deficient (Ddx58−/−) MEFs in the presence of TRIM31. K63-linked polyubiquitin chains on MAVS might be bound by the 2CARD of RIG-I to facilitate the formation of a RIG-I tetramer, which finally leads to MAVS aggregation. Another possibility is that K63-linked polyubiquitin chains in MAVS recruit other adaptors that contain ubiquitin-binding domains, which promote the oligomerization of MAVS after viral infection. A similar mechanism has been proposed for activation of the kinase TAK1 and the adaptor TBK1 (refs. 26,27).
Polyubiquitination has been reported to regulate the activation of MAVS28. Several E3 ligases have been identified that modulate the polyubiquitination of MAVS, including AIP4 (refs. 29,30), RNF5 (ref. 31), RNF125 (ref. 32), Smurf1 and Smurf2 (refs. 33,34), TRIM25 (ref. 35) and MARCH5 (ref. 16). However, all of these E3 ligases promote K48-linked ubiquitination of MAVS. In our study, we identified TRIM31 as an E3 ligase for the K63-linked polyubiquitination of MAVS. To our knowledge, TRIM31 is the first E3 ligase to be identified as being involved in the activation of MAVS.
In summary, we have identified an E3 ubiquitin ligase as a regulator of MAVS aggregation and a host factor directed against infection with an RNA virus. After infection with an RNA virus, TRIM31 interacted with MAVS on mitochondria and then catalyzed the K63-linked polyubiquitination of MAVS on Lys10, Lys311 and Lys461, which facilitated the formation of prion-like MAVS aggregates for the activation of an innate antiviral immune response.
LPS (Escherichia coli 055:B5) was purchased from Sigma-Aldrich. Poly(I:C) and 5′-pppRNA were purchased from Invivogen. LPS, poly(I:C), and 5′-pppRNA were used at a final concentration of 100 ng/ml, 10 μg/ml, and 0.5 μg/ml, respectively. For transfection, poly(I:C) was used at a final concentration of 1 μg/ml. The antibodies specific for HA, ubiquitin, β-actin and protein G–agarose used for immunoprecipitation were from Santa Cruz Biotechnology. The antibodies specific for MAVS, TRIM31, Flag and VSV G protein was purchased from Sigma. The antibodies specific to Myc, IRF3, TBK1, phospho-TBK1, phospho-IRF3, PCNA and ubiquitin(K63) and ubiquitin(K48) were from Cell Signaling Technology. The antibody specific for TOMM20 was purchased from Abcam. The horseradish-peroxidase-conjugated secondary antibodies were purchased from Santa Cruz Biotechnology. The detailed information of the antibodies are listed in Supplementary Table 1. E1, UbcH5a, ubiquitin(WT) and its variants (K48, K63, K48R and K63R) were purchased from Boston Biochem.
The TRIM31-deficient mice were generated by Cyagen Biosciences Inc. (Guangzhou, China) using TALEN technology. The knockout mice were produced by microinjecting TALEN mRNAs into fertilized eggs from C57BL/6 mice. The genotyping of the Trim31-knockout mice was confirmed by sequencing of the PCR fragments (250 bp) in the TALEN-targeting region amplified from genomic DNA isolated from tail tips using the following primers: Forward 5′-GGCCTTGGATTTCTGTACTTTCACATC-3′ and Reverse 5′-TGGGCCTGAACGTATTCTTATTCACAG-3′. The mouse experiments were carried out following the general guidelines published by the Association for Assessment and Accreditation of Laboratory Animal Care. All of the mice were on the C57BL/6 background and were maintained under specific-pathogen-free conditions with the approval of the Scientific Investigation Board of the Medical School of Shandong University.
Isolation of mouse embryonic fibroblasts, macrophages and bone-marrow-derived macrophages.
Embryonic fibroblasts (MEFs) from WT and mutant mice were prepared from day 15 embryos and cultured in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% FBS. BMDMs (bone-marrow-derived macrophages) were isolated from the tibia and femur. In brief, cells were cultured in a 10-cm Petri dish at 37 °C for 5 d. At day 3, 5 ml medium (DMEM with 20% FBS, glutamine and 30% L929 supernatant) was added. Peritoneal macrophages were harvested from mice 4 d after thioglycollate (BD) injection and cultured in DMEM supplemented with 5% FBS.
Cells and viruses.
Human HEK293T, HeLa and THP-1 cells were obtained from American Type Culture Collection. Mavs−/− MEFs was kindly provided by Z. Jiang (Peking University, China). Ddx58−/− MEFs was kindly provided by D. Guo (Wuhan University, China). The cells were cultured at 37 °C under 5% CO2 in DMEM supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin. Sendai virus (SeV) was purchased from the China Center for Type Culture Collection (Wuhan University, China). Vesicular stomatitis virus (VSV), VSV-GFP and herpes simplex virus 1 (HSV-1) were provided by H. Meng (Institute of Basic Medicine, Shandong Academy of Medical Sciences, China).
Sequences, plasmids and transfection.
TRIM31 cDNA was amplified from THP-1 cells or peritoneal macrophages by standard PCR and cloned in pCMV2-Flag, pEGFP-C1 (Promega) or pcDNA3.1/HisB (Invitrogen) plasmids. Deletion, truncation and point mutations were generated by the QuikChange site-directed mutagenesis kits with PfuUltra as the polymerase (Stratagene) and the plasmids encoding WT protein as the template. All constructs were confirmed by DNA sequencing. The IFNB1 and PRD I–III promoter reporter plasmids were provided by K.A. Fitzgerald (University of Massachusetts Medical School, Worcester MA). The expression vector for MAVS(W56R) and TRIM30-α were provided, respectively, by F. Hou and B. Sun (Shanghai Institutes for Biological Sciences, China). Other plasmids used in this study were described previously19. For transient transfection of plasmids and siRNA duplexes into HEK293T cells, Lipofectamine 2000 reagents (Invitrogen) were used. For macrophages and THP-1 cells, plasmids and siRNA duplexes were transfected into cells with the Geneporter 2 Transfection Reagent (Genlantis). Target sequences for transient silencing were sense 5′-GGACCACAAAUCCCAUAAUTT-3′, antisense 5′-AUUAUGGGAUUUGUGGUCCTT-3′ for human TRIM31; sense 5′-GCUCACUAAAUCCUUGAAATT-3′, antisense 5′-UUUCAAGGAUUUA GUGAGCTT-3′ for mouse Trim31. 'Scrambled' control sequences were sense 5′-UUCUCCGAACGUGUCACFUTT-3′ and antisense 5′-ACGUGACACGUU CGGAGAATT-3′. All of these siRNAs were obtained from GenePharma.
Enzyme-linked immunosorbent assay (ELISA).
The concentrations of IFN-β in culture supernatants and sera were measured by ELISA Kits (R&D Systems).
Native polyacrylamide gel electrophoresis (PAGE).
The IRF3 dimerization assay was performed as described previously36.
Total RNA was extracted with TRIzol reagent, according to the manufacturer's instructions (Invitrogen). Specific primers used for RT–PCR assays are shown in Supplementary Table 2. Reverse-transcription products of different samples were amplified by an ABI 7300 Detection System (Applied Biosystems) using the SYBR Green PCR Master Mix (Toyobo), according to the manufacturer's instructions, and data were normalized to the expression of the Gapdh housekeeping gene in each individual sample. The 2−ΔΔCt method was used to calculate relative expression changes.
Co-immunoprecipitation and immunoblot analysis.
For immunoblot analysis, cells or tissues were lysed with M-PER Protein Extraction Reagent (Pierce) supplemented with a protease inhibitor 'cocktail'. Protein concentrations in the extracts were measured with a bicinchoninic acid assay (Pierce) and were made equal in different samples with extraction reagent. For immunoprecipitation (IP), whole-cell extracts were collected 36 h after transfection and lysed in IP buffer containing 1.0% (vol/vol) NP-40, 50 mM Tris-HCl, pH 7.4, 50 mM EDTA, 150 mM NaCl, and a protease inhibitor cocktail (Merck, Darmstadt, Germany). After centrifugation for 10 min at 14,000g, supernatants were collected and incubated with protein G Plus–Agrose Immunoprecipitation reagent (Santa Cruz Biotechnology) together with 1 μg of the corresponding antibodies. After 6 h of incubation, beads were washed five times with IP buffer. Immunoprecipitates were eluted by boiling with 1% (wt/vol) SDS sample buffer (60 mM Tris-HCl (pH 6.8), 1% (wt/vol) SDS, 5% (vol/vol) glycerol, 0.005% (wt/vol) bromophenol blue and 1% (vol/vol) 2-mercaptoethanol). For immunoblot analysis, immunoprecipitates or whole-cell lysates were loaded and subjected to SDS–PAGE, transferred onto nitrocellulose membranes and then blotted with specific antibodies.
For analysis of the ubiquitination of MAVS in HEK293T cells, HEK293T cells were transfected with plasmids expressing Myc–MAVS, HA–ubiquitin(WT), HA–ubiquitin(K48), HA–ubiquitin(K63) and Flag–TRIM31(WT) or its mutants, and then whole-cell extracts were immunoprecipitated with the Myc-specific antibody and analyzed by immunoblot with anti-HA. For analysis of the ubiquitination of MAVS in macrophages, macrophages were infected with SeV, then whole-cell extracts were immunoprecipitated with anti-MAVS and analyzed by immunoblot with anti-ubiquitin, anti-ubiquitin(K48) or anti-ubiquitin(K63).
Luciferase activity was measured with the Dual-Luciferase Reporter Assay system according to the manufacturer's instructions (Promega). Data were normalized for transfection efficiency by calculating the ratio between firefly luciferase activity and Renilla luciferase activity.
Lungs from control or virus-infected mice were dissected, fixed in 10% phosphate-buffered formalin, embedded into paraffin, sectioned, stained with hematoxylin and eosin solution, and examined by light microscopy for histological changes.
Immunofluorescence confocal microscopy.
The experiments were performed as described previously19. Cells grown on glass coverslips were fixed with 4% paraformaldehyde in PBS for 20 min, permeabilized with 0.1% Triton X-100 and blocked with 1% bovine serum albumin. Then the cells were stained with the indicated primary antibodies followed by incubation with fluorescent-dye-conjugated secondary antibodies. Nuclei were counterstained with DAPI (Sigma-Aldrich). For mitochondrial staining, living cells were incubated with 300 nM Mito Tracker Red (Invitrogen) for 30 min at 37 °C. Imaging of the cells was carried out using a Leica laser-scanning confocal microscope.
In vitro binding and ubiquitination assay.
TRIM31 and MAVS proteins were expressed with a TNT Quick Coupled Transcription/Translation System (Promega). In vitro interaction and ubiquitination assays were performed as described19.
Semi-denaturing detergent agarose gel electrophoresis (SDD–AGE).
Semi-denaturing detergent agarose gel electrophoresis (SDD–AGE) was performed according to a published protocol with minor modifications13. Cells were isolated in Buffer A (10 mM Tris-HCl, pH 7.5, 10 mM KCl, 1.5 mM MgCl2, 0.25 M D-mannitol, and protease inhibitor cocktail) by repeated douncing, and the homogenized cells were centrifuged at 700g for 10 min at 4 °C. The supernatant was transferred and centrifuged at 10,000g for 30 min at 4 °C to form a pellet, which was the intact crude mitochondria (P5). Crude mitochondria (P5) were resuspended in 1× sample buffer (0.5× TBE, 10% glycerol, 2% SDS and 0.0025% bromophenol blue) and loaded onto a vertical 1.5% agarose gel (Bio-Rad). After electrophoresis in the running buffer (1× TBE and 0.1% SDS) for 35 min with a constant voltage of 100 V at 4 °C, the proteins were transferred to an Immobilon membrane (Millipore) for immunoblotting. In some cases, crude mitochondria (P5) were first resuspended in the in vitro ubiquitination buffer with E1, UbcH5a and TRIM31; then SDD-AGE and SDS–PAGE were performed.
Viral infection and plaque assay.
Mouse macrophages or other cells (2 × 105) were plated 24 h before infection. Cells were infected with VSV (0.1 MOI), HSV-1 (10 MOI) or SeV for the indicated times. VSV plaque assay and VSV replication were performed as described previously19. HSV-1 plaque assay was conducted in Vero cells. HSV-1 genomic DNA copy number were determined by qPCR using the primers: Forward 5′-TGGGACACATGCCTTCTTGG-3′ and Reverse 5′-ACCCTTAGTCAGACTCTGTTACTTACCC-3′.
Viral infection in vivo.
For in vivo viral infection studies, age- and sex-matched Trim31+/+ and Trim31−/− mice were infected with VSV (4 × 107 pfu/mouse) or HSV-1 (2 × 107 pfu/mouse) by intraperitoneal injection. Serum cytokine production was measured by ELISA. The VSV titers in the lung, spleen and liver, and HSV-1 titers in the brain, were determined by standard plaque assays. For the survival experiments, mice were monitored for survival after VSV or HSV-1 infection.
Statistical significance between groups was determined by two-tailed Student's t-test and two-way ANOVA test. Differences were considered to be significant when P < 0.05. For mouse survival studies, Kaplan–Meier survival curves were generated and analyzed for statistical significance with GraphPad Prism 5.0. Pilot studies were used for estimation of the sample size to ensure adequate power. There was no exclusion of data points or mice. No randomization or blinding was used.
The data that support the findings of this study are available from the corresponding author upon request.
- The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nat. Immunol. 11, 373–384 (2010). &
- Self-regulation and cross-regulation of pattern-recognition receptor signaling in health and disease. Nat. Rev. Immunol. 16, 35–50 (2016).
- Innate immune sensing and signaling of cytosolic nucleic acids. Annu. Rev. Immunol. 32, 461–488 (2014). &
- RIG-I-like receptors: cytoplasmic sensors for non-self RNA. Immunol. Rev. 243, 91–98 (2011). , &
- IFI16 is an innate immune sensor for intracellular DNA. Nat. Immunol. 11, 997–1004 (2010). et al.
- Cyclic GMP–AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 339, 786–791 (2013). , , , &
- Pattern-recognition receptors and inflammation. Cell 140, 805–820 (2010). &
- Structural mechanism of RNA recognition by the RIG-I-like receptors. Immunity 29, 178–181 (2008). &
- Cardif is an adaptor protein in the RIG-I antiviral pathway and is targeted by hepatitis C virus. Nature 437, 1167–1172 (2005). et al.
- Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NF-κB and IRF3. Cell 122, 669–682 (2005). , , &
- VISA is an adapter protein required for virus-triggered IFN-β signaling. Mol. Cell 19, 727–740 (2005). et al.
- IPS-1, an adaptor triggering RIG-I- and Mda5-mediated type I interferon induction. Nat. Immunol. 6, 981–988 (2005). et al.
- MAVS forms functional prion-like aggregates to activate and propagate antiviral innate immune response. Cell 146, 448–461 (2011). et al.
- Structural basis for ubiquitin-mediated antiviral signal activation by RIG-I. Nature 509, 110–114 (2014). , , , &
- NLRX1 is a regulator of mitochondrial antiviral immunity. Nature 451, 573–577 (2008). et al.
- The mitochondrial ubiquitin ligase MARCH5 resolves MAVS aggregates during antiviral signaling. Nat. Commun. 6, 7910 (2015). et al.
- Characterization of TRIM31, upregulated in gastric adenocarcinoma, as a novel RBCC protein. J. Cell. Biochem. 105, 1081–1091 (2008). &
- TRIM30-α negatively regulates TLR-mediated NF-κB activation by targeting TAB2 and TAB3 for degradation. Nat. Immunol. 9, 369–377 (2008). et al.
- TRIM26 negatively regulates interferon-β production and antiviral response through polyubiquitination and degradation of nuclear IRF3. PLoS Pathog. 11, e1004726 (2015). , , , &
- Primary activation of interferon A and interferon B gene transcription by interferon regulatory factor 3. Proc. Natl. Acad. Sci. USA 95, 9837–9842 (1998). et al.
- The cellular level of TRIM31, an RBCC protein overexpressed in gastric cancer, is regulated by multiple mechanisms including the ubiquitin–proteasome system. Cell Biol. Int. 35, 657–661 (2011).
- TRIM25 RING-finger E3 ubiquitin ligase is essential for RIG-I-mediated antiviral activity. Nature 446, 916–920 (2007). et al.
- Reconstitution of the RIG-I pathway reveals a signaling role of unanchored polyubiquitin chains in innate immunity. Cell 141, 315–330 (2010). et al.
- RIG-I-like receptors: sensing and responding to RNA virus infection. Semin. Immunol. 21, 215–222 (2009). , , &
- Toll-like receptor and RIG-I-like receptor signaling. Ann. NY Acad. Sci. 1143, 1–20 (2008). &
- Direct activation of protein kinases by unanchored polyubiquitin chains. Nature 461, 114–119 (2009). et al.
- Crystal structure and mechanism of activation of TANK-binding kinase 1. Cell Rep. 3, 734–746 (2013). et al.
- Ubiquitin in the activation and attenuation of innate antiviral immunity. J. Exp. Med. 213, 1–13 (2016). , &
- PCBP2 mediates degradation of the adaptor MAVS via the HECT ubiquitin ligase AIP4. Nat. Immunol. 10, 1300–1308 (2009). et al.
- Poly(C)-binding protein 1 (PCBP1) mediates housekeeping degradation of mitochondrial antiviral signaling (MAVS). Cell Res. 22, 717–727 (2012). , , &
- The E3 ubiquitin ligase RNF5 targets virus-induced signaling adaptor for ubiquitination and degradation. J. Immunol. 184, 6249–6255 (2010). et al.
- Negative regulation of the RIG-I signaling by the ubiquitin ligase RNF125. Proc. Natl. Acad. Sci. USA 104, 7500–7505 (2007). et al.
- Ndfip1 negatively regulates RIG-I-dependent immune signaling by enhancing E3 ligase Smurf1-mediated MAVS degradation. J. Immunol. 189, 5304–5313 (2012). , &
- Smurf2 negatively modulates RIG-I-dependent antiviral response by targeting VISA/MAVS for ubiquitination and degradation. J. Immunol. 192, 4758–4764 (2014). et al.
- MAVS ubiquitination by the E3 ligase TRIM25 and degradation by the proteasome is involved in type I interferon production after activation of the antiviral RIG-I-like receptors. BMC Biol. 10, 44 (2012). et al.
- Identification of Ser386 of interferon regulatory factor 3 as critical target for inducible phosphorylation that determines activation. J. Biol. Chem. 279, 9698–9702 (2004). et al.
We thank Z. Jiang (Peking University) for Mavs−/− MEFs; D. Guo (Wuhan University) for Ddx58−/− MEFs; K.A. Fitzgerald (University of Massachusetts Medical School) for the human IFNB1 promoter reporter and its mutated reporter; and B. Sun and F. Hou (Shanghai Institutes for Biological Sciences) for the expression plasmids for TRIM30-α and MAVS(M56R). Supported by the Natural Science Foundation of China (81525012, 81273219 and 81471538 to C.G.) and the Specialized Research Fund for the Doctoral Program of Higher Education of China (20130131130010 to C.G.).
- Supplementary Figure 1: TRIM31 is recruited to mitochondria after infection with SeV. (83 KB)
(a) Confocal microscopy of TRIM31-GFP transfected into HEK293T cells for 24 h followed with SeV infection for 6 h. MitoTracker (red) were used to label mitochondria. Scale bar 10 μm. (b) Immunoblot of TRIM31 protein and TOMM20 protein in the crude mitochondrial extracts (upper) and TRIM31 protein and β-actin in the whole cell lysate (lower) prepared from peritoneal macrophages infected with SeV for indicated times. (c) Immunoblot of TRIM31 protein in mouse peritoneal macrophages infected with SeV for indicated times. (d) qPCR analysis of Trim31, Trim26, Trim30α, Ifnb1 expression in primary peritoneal macrophages transfected with mice TRIM31 siRNA, TRIM26 siRNA, TRIM30α siRNA or control siRNA for 48 h, followed with SeV infection for indicated times. mRNA results are presented relative to those of untreated cells transfected with control siRNA. (e) qPCR analysis of TRIM31 and IFNB1 mRNA in THP-1 cells transfected with human TRIM31 siRNA or control siRNA for 48 h, followed with SeV infection for indicated times. mRNA results are presented as in d. (f) qPCR analysis of IFNB1 mRNA in Hela cells transfected with Flag-TRIM31 expression plasmid or control vector for 24 h, followed with SeV infection or transfection with poly(I:C) for the indicated times. mRNA results are presented relative to those of untreated cells transfected with control plasmid. The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in d-f). *p < 0.01. (two-tailed Student’s t-test).
- Supplementary Figure 2: TRIM31 deficiency impairs RNA-virus-induced production of type I interferon and antiviral responses. (191 KB)
(a) Schematic diagram of the deletions of the Trim31 deficient mice generated by transcription activation-like effector nuclease (TALEN) technology. The deletions were confirmed by DNA sequencing analysis of genomic DNA isolated from Trim31+/+ and Trim31–/– mice tails. (b) Immunoblot of TRIM31 protein in Trim31+/+ and Trim31–/– peritoneal macrophages. (c) qPCR analysis of Ifnb1, Ccl5 and Cxcl10 mRNA in the macrophages prepared form Trim31–/– (G) mice or Trim31–/– (5) mice infected with SeV for indicated times. mRNA results are presented relative to those of untreated wild-type cells. (d,e) qPCR analysis of Ifnb1, Ccl5 and Cxcl10 mRNA in bone marrow-derived macrophages (BMDMs) (d) or MEFs (e) prepared from Trim31+/+ and Trim31–/– mice followed with infection of SeV and HSV-1 or stimulation of 5′ppp-RNA, LPS and poly(I:C) for indicated times. mRNA results are presented as in c. The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in c-e). *p < 0.05 and **p < 0.01. (two-tailed Student’s t-test).
- Supplementary Figure 3: Trim31-deficient mice are not impaired in their response to infection with a DNA virus. (158 KB)
(a) qPCR analyses of Ifnb1 mRNA in brains and ELISA analysis of IFN-β production in sera of Trim31+/+ and Trim31–/– mice intraperitoneally infected with HSV-1 (2×107 PFU/mouse）for 24 h. mRNA results are presented relative to those of untreated wild-type cells. (b) qPCR analysis of HSV-1 genomic DNA and plaque assays of HSV-1 titer in brains of Trim31+/+ and Trim31–/– mice as treated in (a). (c) Survival of Trim31+/+ and Trim31–/– mice (n=7) after intraperitoneally injection of HSV-1 (1×108 pfu/mouse). The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in a, b). *p < 0.01 (two-tailed Student’s t-test(a,b) or two-way analysis of variance (ANOVA) (c)).
- Supplementary Figure 4: TRIM31 targets MAVS. (191 KB)
(a) Luciferase activity in HEK293T cells transfected with IFNB1 luciferase reporter and expression plasmids for RIG-IN, MDA5, MAVS, TBK1, IKKε and IRF3 5D together with Flag-TRIM31 expression plasmid for 24 h. (b) Luciferase activity in HEK293T cells transfected with control siRNA or TRIM31 siRNA and IFNB1 luciferase reporter or ISRE luciferase reporter together with expression plasmids for RIG-IN, MAVS and TBK1 for 24 h. (c) Coimmunoprecipitation analysis of Flag-TRIM31 interaction with Myc-IRF3 or Myc-STING in HEK293T cells. The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in a, b). *p < 0.01. (two-tailed Student’s t-test).
- Supplementary Figure 5: TRIM31 catalyzes K63-linked polyubiquitination of MAVS. (79 KB)
(a) Coimmunoprecipitation analysis of RIG-I ubiquitination in HEK293T cells transfected with Myc-RIG-I and HA-Ubiquitin together with control plasmid, Flag-TRIM31 plasmid or Flag-TRIM25 plasmid. (b) Luciferase activity in HEK293T cells transfected with IFNB1 luciferase reporter or ISRE luciferase reporter together with Flag-TRIM31-WT plasmid or Flag-TRIM31-C53/56A plasmid for 24 h followed with SeV infection for indicated times. (c) qPCR analysis of IFNB1 mRNA and VSV RNA in HEK293T cells transfected with control plasmid, Flag-TRIM31-WT plasmid or Flag-TRIM31-C53/56A plasmid for 24 h followed with VSV infection (MOI, 0.1) for indicated times. mRNA results are presented relative to those of untreated cells transfected with control plasmid. (d) Immunoblot of phosphorylated and total IRF3 protein in lysates of HEK293T cells transfected with Flag-TRIM31-WT or Flag-TRIM31-C53/56A plasmid for 24 h followed with SeV infection for indicated times. The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in b, c). *p < 0.01. (two-tailed Student’s t-test).
- Supplementary Figure 6: TRIM31 promotes the formation of MAVS aggregates. (121 KB)
(a) SDD-AGE analysis of MAVS aggregation in Trim31+/+ and Trim31–/– MEFs infected with SeV for indicated times (upper). SDS-PAGE analysis of MAVS protein and TRIM31 protein in the whole cell lysates prepared above (lower). (b) SDD-AGE analysis of MAVS aggregation in HEK293T cells transfected with Flag-TRIM31 expression plasmid and Myc-MAVS for 24 h. (c) Immunoblot of TRIM31 protein in Trim31–/– MEFs reconstituted with mTRIM31-WT or mTRIM31-C52/55A. TRIM31+/+ MEFs were used as controls. qPCR analysis of Ifnb1, Ccl5 and Cxcl10 mRNA in TRIM31+/+ or Trim31–/– MEFs reconstituted with mTRIM31-WT or mTRIM31-C52/55A followed with SeV infection for 12 h. mRNA results are presented relative to those of untreated wide-type cells transfected with control plasmid. (d) qPCR analysis of Ifnb1 mRNA in Mavs–/– MEFs transfected with MAVS expression plasmid or control plasmid for 24 h followed with SeV infection for 12 h. mRNA results are presented relative to those of untreated cells transfected with control plasmid. (e) Luciferase activity in Mavs–/– MEFs transfected with IFN-β or ISRE luciferase reporter and expression plasmids for MAVS-WT, MAVS-K10R, MAVS-K10/311R, MAVS-K10/311/461R along with Flag-TRIM31 expression plasmid for 24 h followed with SeV infection for 12 h. (f) qPCR analysis of Ifnb1 mRNA and SDD-AGE analysis of MAVS aggregation in Mavs–/– MEFs transfected with MAVS-WT, MAVS-K10R, MAVS-K311R and MAVS-K461R along with Flag-TRIM31 expression plasmid or control plasmid for 24 h followed with SeV infection for 12 h. (g) qPCR analysis of Ifnb1 mRNA and SDD-AGE analysis of MAVS aggregation in Mavs–/– MEFs transfected with MAVS-WT or MAVS-W56R along with Flag-TRIM31 expression plasmid or control plasmid for 24 h followed with SeV infection for 6 h. mRNA results are presented as in d. The data are representative of 3 independent experiments with three biological replicates (means ± S.D. in c-g). *p < 0.01. (two-tailed Student’s t-test).
- Supplementary Text and Figures (1,837 KB)
Supplementary Figures 1–6 and Supplementary Tables 1 and 2