Ubiquitination is an essential mechanism in the control of antiviral immunity upon virus infection. Here, we identify a series of ubiquitination-modulating enzymes that are modulated by vesicular stomatitis virus (VSV). Notably, TRIM24 is down-regulated through direct transcriptional suppression induced by VSV-activated IRF3. Reducing or ablating TRIM24 compromises type I IFN (IFN-I) induction upon RNA virus infection and thus renders mice more sensitive to VSV infection. Mechanistically, VSV infection induces abundant TRIM24 translocation to mitochondria, where TRIM24 binds with TRAF3 and directly mediates K63-linked TRAF3 ubiquitination at K429/K436. This modification of TRAF3 enables its association with MAVS and TBK1, which consequently activates downstream antiviral signaling. Together, these findings establish TRIM24 as a critical positive regulator in controlling the activation of antiviral signaling and describe a previously unknown mechanism of TRIM24 function.
Infection by RNA viruses, such as influenza and dengue viruses, remains a global threat to human health. Upon infection, viral RNA is recognized by host RNA sensors, such as retinoic acid inducible gene I (RIG-I), which initiates IFN-I signaling (Chiang et al., 2018; Goubau et al., 2013). After sensing viral RNA, RIG-I is recruited to interact with the downstream adaptor mitochondrial antiviral signaling protein (MAVS), which leads to the activation of TANK-binding kinase 1 (TBK1)/IFN regulatory factor 3/7 (IRF3/7), and finally the secretion of IFNα/β, members in a family of antiviral cytokines, to suppress virus propagation in vivo (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). In the battle between RNA viruses and their hosts, invading pathogens have evolved multiple strategies to counteract host antiviral immune signaling, such as inhibiting the recognition of viral RNA (Chan and Gack, 2016; Manokaran et al., 2015), preventing the binding between RIG-I and MAVS (He et al., 2016), or blocking the activation of TBK1 or IRF3 (Dalrymple et al., 2015; Zhu et al., 2019b), leading to promoted viral escape from host immune surveillance. However, the molecular events controlling the activation of host antiviral immune signaling remain poorly understood.
Ubiquitination is a type of posttranslational modification that has been found by numerous studies to play an important role in regulating host IFN-I signaling (Heaton et al., 2016; Khan et al., 2019; van Gent et al., 2018). Upon RNA virus infection, the signaling molecules in this pathway, such as RIG-I, MAVS, and TRAF3, undergo different types of ubiquitination by various E3 ubiquitin ligases and thus have different outcomes (Castanier et al., 2012; Gack et al., 2007; Mao et al., 2010; Tseng et al., 2010; Yan et al., 2014; Zhong et al., 2009). For example, K63-linked ubiquitination promotes the activation of downstream signaling and enhances the transcription of IFNα/β (Gack et al., 2007; Mao et al., 2010; Tseng et al., 2010; Yan et al., 2014), whereas K48-linked ubiquitination guides these molecules for proteasome degradation (Arimoto et al., 2007; Castanier et al., 2012; Zhong et al., 2009), generating a negative feedback loop to restrain IFN-I signaling. In addition, the ubiquitin that is added to these signaling molecules can be removed by deubiquitinases (DUBs) to counteract the effect of E3 ligase–induced ubiquitination (Cui et al., 2014; Friedman et al., 2008; Kayagaki et al., 2007; Pauli et al., 2014). Therefore, the dynamic regulation of the ubiquitination status of these signaling molecules by E3 ligases or DUBs fine-tunes the activation of IFN-I signaling.
Considering the critical fine-tuning role of ubiquitination in modulating IFN-I signaling, it is highly likely that the virus controls this cellular machinery by regulating the expression of ubiquitination-related enzymes and thus modulates the host production of antiviral IFNα/β. In this study, we found that RNA virus–activated IRF3 suppresses the expression of tripartite motif 24 (TRIM24), an E3 ubiquitin ligase that mediates virus-induced K63-linked ubiquitination of TRAF3, leading to suppressed activation of downstream IFN-I signaling and thus antagonizing host antiviral immune responses.
TRIM24 is down-regulated upon infection with vesicular stomatitis virus (VSV)
In an attempt to identify the potential ubiquitination-regulatory genes that are modulated by RNA viruses, we performed RNA-sequencing and quantitative PCR (QPCR) analysis and identified 49 up-regulated or down-regulated genes encoding E3 ligases or DUBs expressed at levels of significant difference (log2 ≤ −1.5 or log2 ≥ 1.5, P < 0.005) in murine primary macrophages infected with VSV (Fig. 1, A–C). Next, we knocked down the 49 differentially expressed E3 ligase or DUB genes in immortalized bone marrow–derived macrophages (iBMDMs) to examine VSV-induced Ifnb1 gene expression. The results revealed that TRIM24 and Mib2 were the top two ubiquitination-regulatory genes to modulate VSV-induced Ifnb1 expression (Fig. 1, D and E). Because a published study demonstrated a positive role for Mib2 in regulating antiviral immunity (Ye et al., 2014), which validated our screening data, we focused our investigation on TRIM24 in this study.
To confirm the role of TRIM24 in regulating VSV-induced Ifnb1 expression, we knocked down TRIM24 in iBMDMs by using two specific shRNAs and found that reducing TRIM24 expression indeed impaired Ifnb1 induction upon VSV infection (Fig. 1, F and G). Accordingly, knocking down TRIM24 sharply promoted virus propagation, as reflected by enhanced GFP fluorescence, GFP protein expression, and virus titer after infection with GFP-conjugated VSV (Fig. 1, H–J). In contrast, TRIM24 overexpression dramatically curtailed virus propagation (Fig. 1, K–M). Together, these results identified TRIM24 as a VSV-modulated gene that exhibits a positive function in restraining virus propagation.
Loss of TRIM24 specifically impairs RNA virus–induced IFN-I production
Next, we generated TRIM24 total knockout (hereafter termed TRIM24-KO) mice by crossing TRIM24 floxed mice with mice expressing CMV-Cre recombinase, which is ubiquitously expressed in all cells and tissues (Fig. S1, A and B). Consistent with the knockdown data, TRIM24 deficiency significantly suppressed the mRNA expression of Ifna4, Ifnb1, and IFN-stimulated genes (ISGs) in mouse primary macrophages infected with VSV or Sendai virus (SeV) and those transfected with polyI:C (Fig. 2 A and Fig. S1 C). In addition, VSV-induced IFNβ protein secretion was inhibited in TRIM24-deficient macrophages compared with WT cells (Fig. 2 B). Accordingly, loss of TRIM24 dramatically promoted virus propagation, as reflected by increased VSV mRNA expression and virus titer in VSV-infected TRIM24-deficient cells (Fig. 2, C and D). Interestingly, TRIM24 deficiency did not affect the Ifnb1 expression induced by LPS or herpes simplex virus 1 (HSV-1; Fig. 2 E), suggesting that TRIM24 specifically mediated RNA virus–induced, but was dispensable for TLR or DNA virus–induced, production of IFN-I.
We then examined the effect of TRIM24 in regulating RNA virus–induced activation of IFN-I signaling. In concert with the gene expression results, TRIM24 deficiency markedly inhibited VSV- or transfected polyI:C-induced phosphorylation of TBK1 and IRF3 (Fig. 2, F and G). Accordingly, VSV-induced IRF3 dimerization and nuclear translocation were also dramatically suppressed in TRIM24-deficient primary mouse macrophages compared with WT cells (Fig. 2, H and I). More interestingly, knocking down TRIM24 also inhibited VSV-induced IFNB1 expression and promoted virus propagation in human peripheral blood mononuclear cell (PBMC)–derived macrophages (Fig. 2 J). However, TRIM24 deficiency affected neither the VSV-induced activation of NF-κB or MAPKs nor the expression of STAT1 or proinflammatory genes (Fig. S1, D and E). These results collectively suggested that TRIM24 regulated RNA virus–induced activation of IFN-I signaling without affecting NF-κB or MAPK activation and thus specifically mediated IFN-I production.
TRIM24 deficiency impairs in vivo antiviral immunity
To study the in vivo function of TRIM24 in regulating antiviral immunity, WT and TRIM24-KO mice were infected with VSV. As expected, TRIM24 deficiency significantly increased the VSV-induced mouse mortality rate and exaggerated immune cell infiltration and injury in the lungs (Fig. 3, A and B). In addition, the production of IFNα/β cytokines in the serum and their mRNA expression in the lungs and spleens were both significantly inhibited in the VSV-infected TRIM24-KO mice compared with WT controls (Fig. 3, C–E). Accordingly, loss of TRIM24 markedly promoted virus propagation in vivo after VSV infection, as reflected by increased VSV expression at both mRNA and protein levels in the lungs and spleens of VSV-infected TRIM24-deficient mice compared with WT control mice (Fig. 3, F and G). These data suggested that TRIM24 is critical for mediating in vivo antiviral immunity against VSV infection.
TRIM24 functions in mitochondria to mediate the antiviral response
Mitochondria are central hubs that dock antiviral-related signaling molecules to initiate the activation of downstream antiviral kinase cascades (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). To investigate whether TRIM24 also functions in mitochondria to regulate the antiviral immune response, immunofluorescence analysis was performed to examine the subcellular localization of TRIM24 upon virus infection. The results revealed that most of the TRIM24 protein (∼80%) was localized in the nucleus of resting macrophages (Fig. 4, A and B). However, within 3 h of VSV infection, nuclear TRIM24 was translocated into the cytoplasm, leading to a dramatic change in subcellular localization such that ∼40% of the TRIM24 protein was mobilized from the nucleus to the cytoplasm of the virus-infected cells (Fig. 4, A and B). In addition, the results from the immunoblot analysis confirmed that VSV infection gradually decreased nuclear TRIM24 protein levels but increased cytoplasmic TRIM24 protein levels (Fig. 4 C). More interestingly, with MitoTracker probes, we observed that VSV infection induced extensive localization of TRIM24 in mitochondria of peritoneal macrophages (Fig. 4, D–G), implying that TRIM24 was translocated to mitochondria upon virus infection to regulate antiviral immune signaling.
To determine the functional importance of TRIM24 translocation from the nucleus to the cytoplasm, we generated a mutant TRIM24 fusion protein with two nuclear localization signals (NLSs; Zhou et al., 2018) to ensure its maintenance in the nucleus, and then blocked its translocation to the cytoplasm. The immunofluorescence assay confirmed that TRIM24-NLS proteins were maintained in the nucleus with or without VSV infection, whereas the WT TRIM24 proteins were detected in both the nucleus and cytoplasm of 293T cells (Fig. 4 H). As expected, the TRIM24-NLS protein failed to promote RIG-I–induced IFNβ luciferase activity in HEK293T cells, in contrast to the promotion induced by WT TRIM24 (Fig. 4 I). Published studies have suggested that nuclear export protein chromosomal maintenance 1 (CRM1) mediates the nucleocytoplasmic transportation of cargo proteins (Hutten and Kehlenbach, 2007). We found that TRIM24 was constitutively associated with CRM1 with or without VSV infection, and the TRIM24–CRM1 interaction was confirmed by using a transient transfection experiment in which TRIM24 overexpression was strongly associated with CRM1 (Fig. 4 J and Fig. S2 A). Moreover, the CRM1 selective inhibitor Kpt330 dramatically blocked TRIM24 transportation to the cytoplasm without affecting its overall cellular level, and thus, the differences in VSV-induced Ifnb1 expression were abolished, as the mRNA levels in the WT and TRIM24-deficient mouse primary macrophages were reduced to a similar level (Fig. 4, K and L). Together, these data demonstrate that the translocation of TRIM24 from the nucleus to mitochondria is crucial for its antiviral activity.
TRIM24 mediates VSV-induced TRAF3 K63-linked ubiquitination
To figure out the direct molecular target of TRIM24 in IFN-I signaling, we performed a luciferase assay to examine IFNβ transcriptional activity after different key molecules in this signaling cascade were overexpressed with or without TRIM24 in HEK293T cells. The results revealed that TRIM24 significantly promoted RIG-I– and MAVS-induced IFNβ luciferase activity, whereas TRIM24 was dispensable for TBK1- or IRF3-induced IFNβ luciferase activity (Fig. 5 A), suggesting that TRIM24 targeted a molecule downstream of MAVS and upstream of TBK1. In concert with this result, knocking down TRIM24 dramatically suppressed the IFNB1 mRNA expression induced by overexpressed RIG-I and MAVS, whereas TRIM24was found to be dispensable for the IFNB1 mRNA expression induced by overexpressed TBK1 and IRF3 (Fig. 5 B). In addition, the results from the coimmunoprecipitation experiment confirmed that TRIM24 did not bind with overexpressed RIG-I, MAVS, TBK1, or IRF3 in HEK293T cells (Fig. 5 C). Moreover, TRIM24 did not associate with TRAF6 (Fig. 5 C), a MAVS downstream adaptor, to activate NF-κB or MAPK (Yoshida et al., 2008), which explained why TRIM24 was dispensable for VSV-induced proinflammatory gene expression. Interestingly, we found that TRIM24 coimmunoprecipitated with TRAF3, another MAVS downstream adaptor, to transduce antiviral signaling that mediates IFN-I transcription in HEK293T cells, in which these two proteins were overexpressed (Fig. 5 D). The TRIM24–TRAF3 association was confirmed in mouse primary macrophages, in which the two molecules were found to bind to each other in resting cells and VSV-infected cells (Fig. 5 E).
Next, plasmids encoding full-length or truncated TRIM24 were generated to investigate the functional domain that contributes to TRIM24-mediated antiviral immunity (Fig. 5 F). The results revealed that deleting the C-terminal of TRIM24, which lacks the pleckstrin homology domain and bromodomain, did not affect its ability to promote MAVS-induced IFNβ luciferase activity, whereas an N-terminal–truncated TRIM24 construct that lacked the ring domain failed to boost MAVS-mediated IFNB1 transcription (Fig. 5 G). In addition, reconstitution of the N-terminal–deleted TRIM24 failed to restore VSV-induced Ifnb1 expression in TRIM24-KO macrophages to the same level as that induced by full-length TRIM24 (Fig. 5 H). These results suggest that ring domain–related E3 ligase activity is critical for TRIM24-mediated antiviral activity.
Considering the interaction between TRIM24 and TRAF3, we speculated that TRIM24 functions as an E3 ligase of TRAF3 to mediate antiviral immunity. Indeed, overexpression of TRIM24 specifically promoted the K63-linked ubiquitination, but not other types (K6-, K11-, K27-, K29-, K33-, or K48-linked), of TRAF3 (Fig. 6 A and Fig. S2, B and C). In addition, overexpression of full-length TRIM24, but not a ring domain–deleted mutant (TRIM24ΔN), markedly enhanced TRAF3 ubiquitination (Fig. 6 B). Moreover, overexpression of TRIM24 also promoted K63-linked ubiquitination of the C68A/H70A mutant TRAF3 (Fig. 6 C), indicating that TRIM24-mediated TRAF3 ubiquitination is independent of its autoubiquitination activity. Consistently, TRIM24 deficiency dramatically suppressed VSV-induced endogenous K63-linked TRAF3 ubiquitination in mouse primary macrophages, as determined through separate pull-down assays of TRAF3 and ubiquitin (Fig. 6, D and E). Furthermore, an in vitro ubiquitination assay confirmed that the TRIM24 protein could directly and specifically add to the polyubiquitin chains in the in vitro cell-free translated TRAF3 protein but not to those in the TRAF2 protein (Fig. 6, F and G; and Fig. S2 D), suggesting that TRIM24 is a direct E3 ligase that mediates the K63-linked ubiquitination of TRAF3.
Because K63-linked ubiquitination is known to serve as a docking site for signaling molecules (Wang et al., 2012), it is reasonable to speculate that K63-linked ubiquitination of TRAF3 enhances its binding with antiviral-related signaling molecules. Indeed, we found that overexpression of TRIM24 promoted the binding of TRAF3 with MAVS or TBK1 in a dose-dependent manner in HEK293T cells that transiently expressed these proteins (Fig. S2, E and F). Additionally, loss of TRIM24 dramatically suppressed the VSV-induced endogenous association of TRAF3 with MAVS or TBK1 (Fig. 6 H), which explains why TRIM24 docks at mitochondria to mediate the activation of antiviral signaling upon VSV infection. Therefore, these results collectively suggest that TRIM24-mediated K63-linked ubiquitination of TRAF3 facilitates the downstream signal transduction of the IFN-I pathway.
TRAF3 K429/436 ubiquitination is critical for TRIM24-mediated antiviral immunity
We next performed mass spectrometry analysis of Flag-tagged TRAF3 in the presence of TRIM24 overexpression and identified K369, K429, and K436 as potential ubiquitination sites targeted by TRIM24 (Fig. 7 A and Fig. S3). We subsequently generated a series of corresponding TRAF3 point mutations and found that TRIM24 failed to promote the ubiquitination of the K429R or K436R single- or double-point mutated TRAF3 (Fig. S4, A and B), indicating that K429 and K436 are the actual ubiquitination sites targeted by TRIM24. Accordingly, the K429R and K436R TRIM24 mutants were less efficient than WT TRAF3 in promoting RIG-I–induced IFNβ luciferase activity, and the K429/436R double-point mutation impaired the ability of TRAF3 to boost RIG-I activity to a greater extent than did the single-point mutation (Fig. S4 C). In addition, the reconstitution of the K429/436R TRAF3 mutant in TRAF3-deficient mouse embryonic fibroblasts (MEFs) failed to recover VSV-induced Ifna4 and Ifnb1 mRNA expression to the level that WT TRAF3 did (Fig. S4, D and E).
To confirm the importance of endogenous TRAF3 K429/436 ubiquitination, we generated K429/436R double knock-in HEK293T cells with both copies of the mutant TRAF3 gene by using the CRISPR/Cas9 system and verified the mutations by sequencing the genomic DNA (Fig. 7 B). K429/436R knock-in did not affect the endogenous protein expression levels of TRAF3, MAVS, or IRF3 (Fig. S4 F). However, this knock-in mutation dramatically suppressed RIG-I overexpression–induced IFNβ luciferase activity, whereas it was dispensable for TIR domain-containing adapter-inducing IFNβ (TRIF)– and stimulator of IFN genes (STING)–induced IFNβ luciferase activity (Fig. 7 C), confirming that TRIM24-induced TRAF3 K429/436 ubiquitination specifically regulated VSV-induced IFN-I expression. In addition, TRIM24 overexpression failed to promote RIG-I–induced IFNβ luciferase activity in the K429/436R knock-in HEK293T cells to the level it was expressed in the WT cells (Fig. 7 D). We next examined the ubiquitination status of TRAF3 and found that the TRAF3 K429/436R knock-in nearly abolished the endogenous ubiquitination induced by VSV infection (Fig. 7 E). Accordingly, the VSV-induced interaction of TRAF3 with MAVS, TBK1, and IRF3 phosphorylation, dimerization, and nuclear translocation were greatly inhibited in the TRAF3 K429/436R knock-in cells (Fig. 7, F–H). As a consequence, VSV-induced IFNβ luciferase activity and IFNB1 mRNA expression were dramatically suppressed, and virus propagation was significantly enhanced in the TRAF3 K429/436R knock-in cells (Fig. 7, I–K). These results collectively established a critical role for TRIM24-induced TRAF3 K429/436 ubiquitination in mediating antiviral immunity against RNA viruses.
RNA virus–activated IRF3 suppresses TRIM24 transcription
Because our data established a key role of TRIM24 in fighting against RNA viruses, we next sought to investigate how TRIM24 is modulated upon virus infection. We found that VSV infection gradually decreased both the mRNA and protein levels of TRIM24 in a time-dependent manner in mouse primary macrophages, suggesting that VSV suppressed TRIM24 expression at the transcriptional level (Fig. 8, A and B). VSV infection also inhibited TRIM24 mRNA expression in human PBMC-derived primary macrophages (Fig. 8 C). In addition, TRIM24 mRNA expression was significantly suppressed in different types of human cells upon infection with various RNA viruses, such as respiratory syncytial virus (RSV) and H1N1 influenza virus (Fig. 8, D and E), implying that the suppression of TRIM24 expression was not specific to VSV infection.
RNA virus infection induces the activation and nuclear translocation of IRF3, NF-κB, and activator protein 1 (AP-1) to modulate gene transcription (Chiang et al., 2018; Goubau et al., 2013; Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005), which prompted us to examine whether these transcription factors directly contribute to TRIM24 suppression. The results revealed that inhibition of NF-κB– or AP-1–related kinases (ERK, JNK, or p38) through corresponding selective inhibitors did not restore TRIM24 expression upon VSV infection (Fig. 8 F). Interestingly, we found that infection by the HSV-1 DNA virus also suppressed TRIM24 mRNA expression in a time-dependent manner in mouse primary macrophages (Fig. 8 G), which prompted us to speculate that IRF3, a transcription factor that is activated by both DNA and RNA viruses, may mediate TRIM24 transcriptional suppression. Indeed, we found that VSV infection failed to inhibit TRIM24 expression and completely abolished IFNB1 induction in IRF3-deficient HEK293T cells, and reconstitution of IRF3 restored both VSV-induced TRIM24 suppression and IFNB1 expression (Fig. 8, H and I), suggesting that virus-activated IRF3 contributed to the suppression of TRIM24 in the host cells.
To confirm the direct transcriptional inhibitory function of IRF3 on TRIM24 expression, we generated luciferase reporter plasmids by introducing different mouse TRIM24 promoter fragments into a pGL4-basic plasmid (Fig. 8 J) and found that the promoter region between −3,813 and −2,256 contributed to the basal expression of TRIM24 (Fig. 8 K). In addition, we also determined that the promoter region between −3,813 and −2,804 effectively responded to VSV infection to suppress TRIM24 transcription (Fig. 8 L), and IRF3 deficiency abolished VSV-induced transcriptional suppression of this promoter region (Fig. 8 M). As expected, a conserved IRF3-binding motif was found in the promoter region between −3,813 and −2,804 (Fig. 8 N), and the chromatin immunoprecipitation (ChIP)–QPCR experiment results also confirmed that VSV infection induced abundant IRF3 binding to the TRIM24 promoter region and to the IFNB1 gene promoter (Fig. 8 O). Moreover, IRF3 failed to suppress TRIM24 transcription and could not efficiently bind to the DNA sequence when four bases of the IRF3-binding motif were mutated (Fig. 8, P and Q). Together, these data demonstrated that the RNA virus suppressed TRIM24 transcription directly by activated IRF3.
Although RNA virus infection induces swift activation of host innate immunity to produce IFN-I, long-term infection impairs the expression of host antiviral cytokines, leading to uncontrolled virus propagation (Chan and Gack, 2016; Dalrymple et al., 2015; He et al., 2016; Manokaran et al., 2015; Zhu et al., 2019b). TRIM family proteins play crucial roles in modulating innate antiviral immunity and are thus essential for fighting against viral infection (van Gent et al., 2018). Interestingly, a screening study of all TRIM family proteins demonstrated that TRIM24 positively regulates RIG-I–induced IFNβ luciferase activity by using an overexpression system (Versteeg et al., 2013), the findings of which were confirmed by our current study at the genetic level: TRIM24 deficiency indeed impaired RNA virus–induced IFN-I induction and in vivo antiviral immunity. Our findings suggested that TRIM24 functions as a positive regulator to mediate the activation of IRF3 for antiviral immunity, and that it can also be transcriptionally suppressed by activated IRF3 directly, which forms a classic negative feedback loop to restrain the overactivation of antiviral IFN signaling. We speculated that IRF3-induced down-regulation of TRIM24 is mediated through its recruitment of transcriptional coinhibitors to the promoter region of the TRIM24 gene, a supposition that needs further investigation in future studies.
A previous study revealed that TRIM24 negatively regulates STAT1 expression and then further inhibits ISG expression in liver tissues through direct transcriptional suppression in the nucleus (Tisserand et al., 2011). However, our study demonstrated that TRIM24 in macrophages does not affect STAT1 expression but functions in the mitochondria to promote IFN-I and ISG expression. This functional difference may be due to the cell-specific mechanism of TRIM24 to regulate ISG expression. In addition to these findings on TRIM24, our RNA-sequencing experiment also enabled us to identify a number of differentially expressed ubiquitination-regulatory genes that were modulated upon virus infection. The overall effect of these virus-modulated ubiquitination-regulator genes may indicate an important mechanism through which antiviral immunity is controlled.
Mitochondria are the cytoplasmic organelles where MAVS docks (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). On the one hand, MAVS in mitochondria receives signals from viral RNA-primed RIG-I; on the other hand, it activates antiviral kinase cascades to initiate IFN-I transcription through the TRAF3 adaptor (Kawai et al., 2005; Meylan et al., 2005; Seth et al., 2005; Xu et al., 2005). Hence, mitochondria serve as central platforms to transmit the IFN induction signal by tethering related signaling molecules together. Interestingly, our data showed that VSV infection induced CRM1-dependent TRIM24 translocation from the nucleus to mitochondria, where TRIM24 is associated with TRAF3 for downstream signaling activation. Because TRIM24 is constitutively bound with CRM1, viral infection did not affect its association with CRM1 but induced the translocation of TRIM24 to the cytoplasm through CRM1, a finding that needs further investigation to elucidate the mechanism controlling this action.
TRAF3 is a MAVS adaptor protein that undergoes K63-linked ubiquitination in mitochondria upon RNA virus infection, and this posttranslational modification of TRAF3 is essential in the virus-triggered induction of IFN-I (Häcker et al., 2011; Mao et al., 2010; Oganesyan et al., 2006). In contrast, the deubiquitination of TRAF3 induced by DUBA, OTUB1/2, or UCHL1 negatively regulates virus-induced IFN-I production (Karim et al., 2013; Kayagaki et al., 2007; Li et al., 2010). In contrast to the cellular apoptosis inhibitor cIAP1/2, which mediates both K48- and K63-linked ubiquitination of TRAF3 under specific circumstances, TRIM24 was found to be a direct E3 ligase that specifically induces the K63-linked ubiquitination of TRAF3 upon virus infection, and this action is independent of TRAF3 autoubiquitination activity. In addition, we also found that TRIM24-induced TRAF3 ubiquitination occurs at K429 and K436, and the knock-in data also confirmed the functional importance of ubiquitination at these two lysine residues in mediating virus-induced activation of IFN-I signaling. Interestingly, a published study showed that TRAF3 Y440 and Q442 are critical for its association with the TIM pocket of MAVS, and the Y440A/Q442A TRAF3 mutant blocked this interaction and thus impaired RNA virus–induced IFN-I production (Saha et al., 2006). Considering that K429 and K436 are physically close to Y440 and Q442 in the TRAF domain of TRAF3, it is highly likely that the K63-linked ubiquitination at these two lysine residues also facilitates the interaction of TRAF3 with MAVS, an association that was confirmed by our knock-in data showing that the K429/436R TRAF3 mutant failed to interact with MAVS and thus suppressed downstream antiviral signaling.
In summary, we found that RNA virus infection induced CRM1-dependent TRIM24 translocation from the nucleus to the mitochondria, where TRIM24 mediated the K63-linked ubiquitination of TRAF3 and thus activated the downstream antiviral transcription factor IRF3, which was translocated into the nucleus to initiate IFN-I gene transcription. In the meantime, activated IRF3 also bound to the promoter of TRIM24 and directly inhibited its gene transcription, forming a negative feedback loop to restrain overactivation of antiviral signaling (Fig. 9). Therefore, targeting TRIM24 may have a beneficial effect by boosting the host antiviral immune response and thus antagonizing RNA virus infection.
Materials and methods
TRIM24 floxed mice (previously described in Yu et al., 2019) were backcrossed with C57BL/6 mice for ≥6 generations. C57BL/6 background TRIM24 floxed mice were crossed with B6.C-Tg (CMV-Cre) 1Cgn/J mice (J006054; purchased from GemPharmatech Co.) to produce TRIM24 total knockout mice (termed TRIM24-KO). In all experiments, WT littermate controls were used. All mice were maintained in a specific pathogen–free facility, and all animal experiments were in accordance with protocols approved by the institutional Biomedical Research Ethics Committee, Shanghai Institute of Nutrition and Health, Chinese Academy of Sciences, Shanghai, China.
Cells and viruses
TRAF3-KO MEFs were kindly provided by Dr. Shao-Cong Sun (MD Anderson Cancer Center, Houston, TX). IRF3-KO HEK293T were kindly provided by F. Hou (Shanghai Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences). 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. SeV, VSV, VSV-GFP, and HSV-1 were also kindly provided by F. Hou (Zhu et al., 2019a).
Viral infection in vitro and in vivo
Cells were infected with VSV, HSV-1, or SeV at multiplicity of infection of 1 for the indicated times. For in vivo viral infection studies, age- and sex-matched WT and TRIM24-KO mice were infected with VSV (1 × 108 PFU/mouse) via tail vein injection and monitored for survival status. Lungs from control or virus-infected mice were dissected, fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, stained with H&E solution, and examined by light microscopy for histological changes. Serum cytokine production was measured by ELISA. mRNA expression of Ifna4, Ifnb1, and VSV in the lung and spleen was detected via real-time QPCR. VSVG protein expression in the lung and spleen was determined by immunoblot assay.
Plasmids, antibodies, and reagents
The cDNA encoding TRIM24 and its truncations were cloned from THP-1 cells and constructed into pcDNA vector. The expression vector encoding Flag-TRAF3, Flag-MAVS, Flag-TBK1, Flag-TRAF6, HA-tagged different types of ubiquitin, IFNβ luciferase reporter, and pRL-TK plasmids were kindly provided by Dr. Shao-Cong Sun. The expression vector encoding mutant Flag-TRAF3 C68A/H70A was kindly provided by Dr. Xiang He (Beijing Institute of Biotechnology, Beijing, China). For the generation of TRAF3 K369R, K429R, K436R, and K429/436R mutants, point mutations were constructed by site-directed mutagenesis. The expression vectors encoding Flag-RIGI-N and Flag-IRF3-5D were kindly provided by Dr. Chengjiang Gao (Shandong University School of Basic Medical Sciences, Shandong, China). The expression vector encoding CRM1 was kindly provided by Dr. Wenjun Liu (Institute of Microbiology, Chinese Academy of Sciences). All homemade and requested constructs were confirmed by DNA sequencing.
Antibodies for TRIM24 (14208-1-AP), GAPDH (60004-1-1G), TRAF3 (18099-1-AP) and MAVS (14341-1-AP) were from Proteintech. Antibodies for TRIM24 (C-4, sc-271266), Hsp60 (H-1, sc-13115), Lamin B (C-20, sc-6216), p65 (C-20, sc-372), IκBα (C-21, sc-371), p38 (H-147, sc-7149), Erk1 (K-23, sc-94), P-Erk (E-4, sc-7383), TRAF2 (C-20, SC-876), TRAF3 (H-122, sc-1828), CRM1 (H-7, sc-74455), IRF3 (FL-425, sc-9082), STAT1 (M-22, sc-592), ubiquitin (Ub; P4D1, sc-8017), c-Myc (9E10, sc-40), c-Myc-HRP (9E10, sc-40-HRP), and donkey anti-goat IgG (HRP, sc-2020) were from Santa Cruz. Antibodies for tubulin (2125S), VDAC (4661S), IRF3 (4302), Jnk (9252), TBK1 (3013), K63-linkage specific polyubiquitin (5621), MAVS (4983), p-IRF3 (29047), normal rabbit IgG (2729), p-Jnk (4668), p-p65 (3033), p-IκBα (2859), p-p38 (9215), and p-TBK1 (5483) were from Cell Signaling. Antibodies for β-actin (A2228), Flag (A8592), and VSVG (SAB4200695) were from Sigma-Aldrich. Antibody for GFP was from Abcam. Antibody for HA (2013819) was from Roche. Antibodies for Alexa Fluor Plus 488 conjugated mouse IgG (A32723) was from Thermo Fisher Scientific.
LPS (L3129) was purchased from Sigma-Aldrich; polyI:C was from InvivoGen; protein A/G magnetic beads (HY-K0202), Selinexor (KPT-330, HY-17536), and pyrrolidinedithiocarbamate ammonium (PDTC, HY-18738) were from MedChemExpress; PD98059 (S1177), SB203580 (S1076), and SP600125 (S1460) were from SelleckChem; and FastStart universal SYBR Green master mix (4913914001) was from Roche. LipoFiterTM Liposomal Transfection Reagent (HB-LF10001) was from Hanbio. HE staining kit (E607318-0200), Anti-Fade Mounting Medium (E675011), BSA (A602440), puromycin (A610593), proteinase K (A300491), and DAPI dihydrochloride (A606584) were from Sangon Biotech. FBS (10270), 2-mercaptoethanol (21985023), penicillin-streptomycin (15140-122), and GlutaMAX Supplement (35050-061) were from Gibco. DMEM/High Glucose (SH30243.01) and RPMI Medium Modified (SH30809.01) were from Hyclone. Protease inhibitor cocktail (B14001) and phosphatase inhibitor cocktail (B15001) were from Bimake. pMD18-T Vector (6011) and PrimeScript RT reagent kit (RR037A) were from Takara. TRIzol reagent (15596018), RNase A (8003089), Lipofectamine RNAiMAX (13778075), and Lipofectamine 3000 (L3000015) were from Thermo Fisher Scientific. BBL Thioglycollate Medium Brewer Modified (211716) was from BD. EZ-ChIP kit (17-371) and immobilon Western chemiluminescent HRP substrate (WBKLS0500) were from Millipore. Dual-luciferase reporter assay system (E1960) and TNT quick coupled transcription/translation systems (L1170) for in vitro protein expression were from Promega. Recombinant human UbcH5a/UBE2D1-ubiquitin charged (E2-800), recombinant human UbcH5a (uncharged, E2-616), recombinant human ubiquitin-activating Enzyme/UBE1 (E-305), recombinant human ubiquitin N-Terminal Biotin (UB-560), and Mg2+-ATP (B-20) for in vitro ubiquitination assay were from Boston Biochem. ClonExpress II one step cloning kit (C112) and AxyPrep PCR cleanup kit (AP-PCR-250) were from Vazyme and Axygen, respectively. Recombinant human macrophage colony-stimulating factor (C417) and human IL-10 (CD04) were from Novoprotein. Recombinant mouse IFN Alpha 4 (12115) and mouse IFNβ ELISA Kit (42400) were from PBL. The mouse Macrophage Nucleofector Kit (VPA-1009) was purchased from Lonza.
Mouse and human macrophage preparation
Mouse BMDMs were prepared as previously described (Zhang et al., 2018). In brief, bone marrow cells isolated from mouse tibia and femur were cultured in 10-cm dishes with DMEM containing 20% FBS and L929 conditional medium. After 4–5 d of culture, the adherent macrophages were detached and seeded into culture plates for further experiments. For the preparation of peritoneal macrophages, 4% thioglycolate (BD) was i.p. injected into 6–8-wk-old WT and TRIM24 KO mice. After 4–5 d, mice were sacrificed, and the peritoneal cavity was lavaged with DMEM medium. The peritoneal cells were collected by centrifugation and seeded in the dish. Macrophages were allowed to adhere for 4 h, washed with fresh medium to remove unattached cells, and incubated overnight.
Human PBMC-derived macrophages were prepared as previously described (Jin and Kruth, 2016). Briefly, PBMCs isolated from human peripheral blood of healthy donors were cultured in RPMI 1640 with 2 mM l-glutamine, 50 ng/ml M-CSF, 25 ng/ml IL-10, and 10% FBS. After monocytes differentiated and proliferated sufficiently to become confluent, which required ∼6 d of culture, macrophages were detached for further experiments. All human blood samples were collected after informed consent was obtained, and the related experiments were in accordance with protocols approved by the Institutional Biomedical Research Ethics Committee, Shanghai Institute of Nutrition and Health, Chinese Academy of Sciences.
Virus titer assay
Viral titer was determined by using a plaque assay as previously described (Zhu et al., 2019a). In brief, HEK293T cells in 6-well plates were infected with serial dilutions of the recovered VSV for 1 h. The infected cells were overlaid with 1% soft agar dissolved in DMEM and incubated for 48 h. Plates were stained with 0.1% crystal violet in DMEM to display plaques, which were then quantitated.
Macrophages and HEK293T cells that were left uninfected or infected with VSV were washed and then harvested with 50 µl ice-cold lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, and 1% NP-40 containing protease inhibitors). After centrifugation at 13,000 g for 5 min, supernatant protein was quantified and diluted with 2× native PAGE sample buffer (125 mM Tris-HCl, pH 6.8, 30% glycerol, and 0.01% bromophenol blue), and then applied to a prerun 6% native gel for separation. After electrophoresis, proteins were transferred onto a nitrocellulose membrane for immunoblotting.
The expression of mRNA was examined by using real-time QPCR as previously described (Fan et al., 2019). In brief, total RNA was extracted by TRIzol reagent according to the manufacturer’s protocol. cDNA was synthesized using the PrimeScript RT Reagent Kit (Takara). QPCR was performed by using SYBR Green Master mix (Roche). The relative expression of genes was calculated by a standard curve method and normalized to the expression level of Actb. Gene-specific PCR primers are listed in Table S1.
Gene knockdown in macrophages and HEK293T cells
For knockdown of TRIM24 genes in mouse iBMDMs, the pLKO.1 vectors containing shRNA sequences targeting specific genes along with lentiviral packaging vectors, psPAX2 and pMD2, were transfected into HEK293T cells with Lipofectamine 3000. After 48 h, the lentiviral supernatants were collected for the infection of mouse iBMDM. The infected cells were selected with puromycin (8 µg/ml) for 48 h and examined for knockdown efficiency by immunoblotting. For knockdown of TRIM24 in human PBMC-derived macrophages and HEK293T cells, siRNA targeting TRIM24 (siTRIM24) or negative control (siControl) were transfected into cells with Lipofectamine RNAiMAX. For knockdown of E3 ligase or DUB genes in iBMDM, siRNA targeting the indicated genes or negative control were transfected into cells.
Cells were fixed for 10 min with 4% cold paraformaldehyde and then permeabilized for 5 min with 0.2% Triton X-100. After blocking with 2% BSA in PBS containing 0.5% Tween-20, cells were stained with specific primary antibodies, followed by blotting with fluorescent conjugated secondary antibody. Nuclei were labeled with DAPI (Sangon Biotech). The stained cells were visualized and photographed with ZEISS Cell Observer.
Total RNA was extracted from VSV-infected or uninfected BMDMs and subjected to RNA-sequencing analysis by BGI Tech Solutions. The raw transcriptomic reads were mapped to a reference genome (GRCm38/mm10) with Bowtie. Gene expression levels were quantified with the RSEM software package. Significantly affected E3 ligase or DUB genes were acquired by setting a threshold with corrected P < 0.005 and log2 ≤ −1.5 or log2 ≥ 1.5, and are presented as heat maps.
Luciferase reporter assay
IFNβ luciferase reporter was cotransfected with pRL-TK, and other expression vectors where indicated, into HEK293T cells by using LipoFiterTM Transfection Reagent (HanBio). IFNβ transcriptional activity was measured with Dual-Luciferase Reporter Assay System (Promega), and the relative light units of chemiluminescence were measured with LB 9508 Lumat3 (Berthold Technologies).
Coimmunoprecipitation and immunoblot analysis
For coimmunoprecipitation assays, cells were lysed with radioimmunoprecipitation assay (RIPA) buffer containing protease/phosphatase inhibitors. The whole-cell lysates were incubated with desired antibodies, and the target protein was then pulled down with protein A/G magnetic beads. For immunoblot analysis, immunoprecipitates or whole-cell lysates were resolved using SDS-PAGE, transferred to nitrocellulose membranes (Millipore), and then blotted with specific primary and secondary antibodies. Immunoblots were visualized using the Immobilon Western chemiluminescent HRP substrate (Millipore) with luminescent imaging workstation (Tanon). In some experiments, the mitochondria proteins were isolated by using Qproteome Mitochondria Isolation Kit (Qiagen) according to the manufacturer’s instructions, and then applied for immunoblot assay.
The in vivo ubiquitination assay was performed as previously described (Liu et al., 2018; Zhang et al., 2019). In brief, VSV infected or uninfected macrophages or HEK293T cells transfected with the desired plasmids were lysed with cell lysis RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 1 mM EDTA) containing protease inhibitor and N-ethylmaleimide (Sigma-Aldrich). After saving some cell extracts for input analysis, the remaining cell extracts were added to SDS to a final concentration of 1% and then boiled at 100°C for 5 min, which dissociated all the potential protein complexes under such denaturing conditions. The boiled cell extracts were diluted with RIPA buffer until the SDS concentration was 0.1%, precleaned with protein A/B–coupled agarose beads, and then incubated with specific immunoprecipitation antibody on a shaker under denaturing conditions (0.1% SDS) at 4°C overnight. The next day, the immunoprecipitated proteins were collected by incubation with protein A/B–coupled agarose beads on a shaker at 4°C for 2 h, washed with RIPA buffer containing protease inhibitors, PMSF, and N-ethylmaleimide, boiled at 100°C for 5 min, and then loaded to run SDS-PAGE. The immunoprecipitates were immunoblotted with anti-ubiquitin or indicated antibodies.
For in vitro ubiquitination assay, TRAF3 and TRIM24 proteins were expressed in vitro with the TNT Quick Coupled Transcription/Translation Systems (Promega). In vitro ubiquitination assays were performed with the Ubiquitination Kit (Boston Biochem) according to the manufacturer’s instructions.
Flag-TRAF3, HA-Ub, and Myc-TRIM24 expression plasmids were cotransfected into HEK293T cells. Cells were harvested 48 h after transfection, and the lysates were immunoprecipitated with anti-Flag antibody. After washing, the eluted samples were resolved with SDS-PAGE, followed by Coomassie brilliant blue staining. The sample of TRAF3 band was cut and sent to process with mass spectrum analysis using a QE1 system at the National Facility for Protein Science in Shanghai, Zhangjiang Lab.
The ChIP assay procedure was modified from the manufacturer’s instructions (EZ-ChIP, Millipore). Briefly, isolated peritoneal macrophages (∼1 × 107 cells) were fixed with 1% formaldehyde (Sigma-Aldrich) at room temperature for 10 min in 10 ml medium, followed by quenching with 125 mM glycine. Nuclear extracts were sonicated with Covaris E220 for 660 s. After preclearing with normal IgG for 1 h, the sonicated cell lysates were immunoprecipitated with the IRF3 antibody overnight on a nutator at 4°C. The next day, protein A/G magnetic beads were added, and cell lysates were incubated on a nutator for another 2 h. After washing with buffers, chromatin was eluted from the protein/DNA complex and digested with proteinase K and RNaseA at 65°C overnight to reverse cross-links. The freed DNA was purified with AxyPrep PCR cleanup kit (Axygen) and subjected to QPCR analysis with SYGR Green master mix. All sequences of primers for ChIP-QPCR are shown in Table S1.
CRISPR-Cas9–mediated genome editing
To create TRAF3-K429/436R gene-targeted alleles in HEK293T cells, the small guide RNA (sgRNA) sequences near the codon encoding Lys429/436 were chosen on the basis of their specificity scores (https://crispr.mit.edu/). The sgRNA sequences were then cloned into the lentiCRISPRv2 plasmid (Addgene). The repair template harboring ∼1-kb homology arms flanking the TRAF3 K429/436 codon was amplified from the genomic DNA of 293T cells and cloned into pMD18-T Vector (Takara). The mutation encoding K429/436R (AAG→AGG) was then introduced into the repair template. To avoid the cleavage of the repair template by Cas9, an additional synonymous mutation was designed to make the repair template sequence different from the sgRNA sequences. The lentiCRISPRv2 plasmids and repair template were cotransfected into HEK293T cells. 48 h later, the cells were treated with 5 µg/ml puromycin for 3 d to remove the cells without transfection, and single cells were seeded into separate wells of 96-well plates. After clonal expansion, genomic DNA was amplified by PCR using primers flanking the TRAF3 K429/436 codon. The PCR products were then sequenced to validate the TRAF3 K429/436R mutation in both alleles. The sgRNA sequence is human TRAF3 K429/436R sgRNA: 5′-GGAAGCAGGAGGCCGTCATG-3′.
The data are shown as mean ± SEM, and unless otherwise indicated, all the presented data are representative results of at least three independent repeats. Statistical analysis was performed with Prism 6 (GraphPad), and the statistics were analyzed by two-tailed Student’s t test or one-way or two-way ANOVA as indicated. Differences considered to be significant at P < 0.05 are indicated by *, those at P < 0.01 are indicated by **.
The RNA-sequencing data have been deposited into the Gene Expression Omnibus (accession code GSE136363). All other data supporting the findings of this study are available from the corresponding author on reasonable request.
Online supplemental material
Fig. S1 shows the identification of TRIM24 deletion efficiency in different cells of TRIM24-KO mice, and that TRIM24 positively regulates VSV-induced ISG genes and is dispensable for the induction of proinflammatory genes Fig. S2 shows that TRIM24 facilitates the association of TRAF3 with MAVS and TBK1. Fig. S3 shows potential ubiquitination sites of TRAF3 identified by mass spectrum. Fig. S4 shows that TRAF3 K429/K436 are critical for the VSV-induced IFN-I.
We thank Dr. Shao-Cong Sun (MD Anderson Cancer Center, Houston, TX), Dr. Xiang He (Beijing Institute of Biotechnology, Beijing, China), and Dr. Chengjiang Gao (Shandong University School of Basic Medical Sciences, Shandong, China) for the generous gift of plasmids.
This research was supported by the grants from the National Key R&D Program of China (2018YFA0107201, 2018YFA0902703), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB39000000 and XDB19000000), the Key Research Program of the Chinese Academy of Sciences (KFZD-SW-216), the National Natural Science Foundation of China (81770567 and 81825018), the Thousand Young Talents Plan of China, and Chinese Academy of Sciences Key Laboratory of Tissue Microenvironment and Tumor.
Author contributions: Q. Zhu designed and performed the experiments, prepared the figures, and wrote part of the manuscript; T. Yu, S. Gan, Y. Wang, Y. Pei, Q. Zhao, S. Pei, S. Hao, J. Yuan, and J. Xu contributed to the experiments; C. Peng and P. Wu contributed to mass spectrum analysis; F. Hou provided the DNA and RNA virus and IRF3-KO HEK293T cells and contributed to virus titer experiments; X. Wu provided TRAF3-KO MEFs; J. Qin contributed to plasmids and critical editing; Y. Xiao designed and supervised this study, prepared the figures, and wrote the manuscript.
Disclosures: The authors declare no competing interests exist.