Negative Regulation of the Retinoic Acid-inducible Gene I-induced Antiviral State by the Ubiquitin-editing Protein A20*

Activation of the interferon regulatory factors (IRFs) 3 and 7 transcription factors is essential for the induction of type I interferon (IFN) and development of the innate antiviral response. Retinoic acid-inducible gene I has been shown to contribute to virus-induced IFN production independent of the Toll-like receptor pathways in response to a variety of RNA viruses and double-stranded RNA. In the present study, we demonstrate that the NF-κB-inducible, anti-apoptotic protein A20 efficiently blocks RIG-I-mediated activation of NF-κB-, IRF-3-, and IRF-7-dependent promoters but only weakly interferes with TRIF-TLR-3-mediated IFN activation. Expression of A20 completely blocked CARD domain containing ΔRIG-I-induced IRF-3 Ser-396 phosphorylation, homodimerization, and DNA binding. The level of A20 inhibition was upstream of the TBK1/IKKϵ kinases that phosphorylate IRF3 and IRF7 and paradoxically, A20 selectively degraded the TRIF protein but not RIG-I. A20 possesses two ubiquitin-editing domains, an N-terminal deubiquitination domain and a C-terminal ubiquitin ligase domain consisting of seven zinc finger domains. Deletion of the N-terminal de-ubiquitination domain had no significant effect on the inhibitory effect of A20, whereas deletion or mutation of zinc finger motif 7 ablated the inhibitory function of A20 on IRF- or NF-κB-mediated gene expression. Furthermore, cells stably expressing the active form of RIG-I induced an antiviral state that interfered with replication of vesicular stomatitis virus, an effect that was reversed by stable co-expression of A20. These results suggest that the virus-inducible, NF-κB-dependent activation of A20 functions as a negative regulator of RIG-I-mediated induction of the antiviral state.

minating in the production of cytokines and chemokines that disrupt virus replication and initiate innate and adaptive immune responses (1)(2)(3). Rapid induction of type I interferon (IFN) 4 expression is a central event in establishing the innate antiviral response and requires pathogen-inducible, activation of transcription factors that function in a synergistic fashion to induce gene expression (reviewed in Refs. 4 -9). Among the members of the interferon regulatory factor (IRF) family, IRF-3 and IRF-7 play essential roles in the virus-induced type I IFN gene activation following virus infection (10 -17). IRF-3 is activated by C-terminal phosphorylation, which promotes dimerization, cytoplasmic to nuclear translocation, DNA binding, association with CBP/p300 histone acetyltransferases, and transactivation of downstream early genes such as IFNB, IFNA1, and RANTES (regulated on activation normal T cell expressed and secreted). In contrast, IRF-7 protein is synthesized de novo upon IFN stimulation and contributes to the expression of delayed-type genes, including other IFNA subtypes. As with IRF-3, virus infection induces C-terminal phosphorylation and activation of IRF-7 (15,16). The IKK-related kinases, IKK⑀ (18) and TBK1 (19 -21), were shown to be essential signaling components required for IRF-3 and IRF-7 phosphorylation (22)(23)(24).
A separate signaling pathway utilizes the retinoic acid-inducible gene I (RIG-I) to recognize a variety of RNA viruses and trigger the innate antiviral response, independent of the TLR-dependent pathways. RIG-I contains a DEX(D/H) box RNA helicase domain and two caspase recruitment domains (CARDs); full-length RIG-I can interact with dsRNA through its DEX(D/H) box within C terminus and augment IFN production in response to viral infection in an ATPase-dependent man-ner, and the two copies of the CARD at its N terminus transduce signals leading to the activation of IRF-3 and NF-B. The constitutively active form of RIG-I (CARD domain alone) is capable of activating IRF-3 and NF-B and stimulating IFN-␤ production (36). Recent studies demonstrated that the hepatitis C virus (HCV) gene product NS3/4A protease complex efficiently blocks RIG-I signaling pathway and contributes to the establishment of HCV persistence (37)(38)(39). The generation of RIG-I-deficient mice revealed that RIG-I, but not the TLR system, plays an essential role in antiviral responses in various cells, except plasmacytoid dendritic cells. Reciprocally, the TLR system, but not RIG-I, was indispensable to IFN secretion in plasmacytoid dendritic cells (40).
Regulation of the TLR-independent, RIG-I signaling pathway leading to IRF-3 and NF-B has not been well defined, although recent experiments suggest that the NF-B-inducible ubiquitin-editing protein A20 negatively regulates IRF-3 activation (41,42). We therefore sought to define the involvement of A20 in the regulation of RIG-I signaling. Activation of the IFN antiviral state by RIG-I was completely inhibited by A20 expression, and the C-terminal zinc finger domain of the ubiquitin ligase region was required for the inhibitory function of A20. Furthermore, cells stably expressing the active form of RIG-I induced an antiviral state that significantly blocked replication of VSV, an effect that was reversed by stable co-expression of A20. These results demonstrate that virus-mediated activation of A20 functions as a negative regulator of RIG-I mediated induction of the antiviral state.
Antibody Preparation-RIG-I-  was expressed in Escherichia coli as a glutathione S-transferase fusion protein and purified by glutathione-Sepharose column chromatography. The recombinant proteins were injected into rabbits to produce antisera against RIG-I-(1-228).
Co-immunoprecipitation and Western Blot Analysis-Transient transfection, co-immunoprecipitation, and Western blot analysis were performed as previous described (10).
Electrophoretic Mobility Shift Assay-Briefly, cell pellets were treated with 10 mM HEPES, 50 mM NaCl, 10 mM EDTA, 5 mM MgCl 2 , 0.5 mM spermidine, 0.15 mM spermine, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin (10 g/ml), pepstatin (10 g/ml), aprotinin (10 g/ml), and 1 mM Na 3 VO 4 . Suspension was held on ice for 30 min and brought to 0.1% Nonidet P-40 and 10% glycerol concentration. Samples were spun for 5 min at 5,000 rpm at 4°C. Supernatant was removed, and the pellet was washed in 50 mM NaCl. Nuclear extracts were obtained in a 10 mM HEPES, 400 mM NaCl, 0.1 mM EDTA, 0.5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, leupeptin (10 g/ml), pepstatin (10 g/ml), aprotinin (10 g/ml), and 1 mM Na 3 VO 4 solution. Samples were left to rotate at 4°C for 30 min and spun at 15,000 rpm for 10 min at 4°C. Whole cell extracts were assayed for IRF-3 binding in gel shift analysis using a 32 P-labeled double-stranded oligonucleotide corresponding to the interferon-stimulated response element (ISRE) of the IFNA/B-inducible ISG15 gene (5Ј-GATCGGAAAGGGAAACCGAAACTGAA-GCC-3Ј). Complexes were formed by incubating the probe with 10 g of nuclear extract for 20 min at room temperature in 10 mM Tris-Cl (pH 7.5), 1 mM EDTA, 50 mM NaCl, 2 mM dithiothreitol, 5% glycerol, 0.5% Nonidet P-40, 1 mg of bovine serum albumin per ml, and poly(dI-dC) (1.0 mg/ml). Extracts were run on a 5% polyacrylamide gel (60:1 crosslink) prepared in 0.25ϫ Tris borate-EDTA. After running at 160 V for 3 h, the gel was dried and exposed to a Kodak film at Ϫ70°C overnight. To demonstrate the specificity of the detected signal, 1 mg of anti-IRF-3 (Santa Cruz Biotechnology FL-425) antibody was incubated for 30 min on ice prior to the addition of the probe to observe a supershift in the complex formation.

RESULTS
A20 Disrupts RIG-I Signaling-RIG-I signaling pathway leads to the activation of NF-B and IRF transcription factors, which are essential for the activation of IFNB promoter (36), and A20 is a potent inhibitor of TLR3-and Sendai virus-induced activation of ISRE promoter (42). To determine the ability of A20 to inhibit RIG-I-mediated activation of IFNB gene transcription, a constitutively active form of RIG-I-(⌬RIG-I) and A20-expressing plasmids were co-transfected into HEK293 cells together with an IFNB promoter construct and examined for their ability to activate IFNB reporter gene activity. The IFNB promoter had a low basal activity that was not affected by A20 expression (Fig. 1A). The expression of ⌬RIG-I alone resulted in a 275-fold stimulation of IFNB promoter activity, and co-expression of A20 totally inhibited ⌬RIG-I-mediated activation of IFNB promoter activity (Fig. 1A). An ISRE reporter construct together with ⌬RIG-Iand A20-expressing plasmids further demonstrated that RIG-I mediated activation of IRF-3 was blocked by A20. As shown in Fig. 1B, the expression of constitutively active form of RIG-I stimulated the ISRE luciferase reporter gene activity up to 1700-fold, whereas co-expression of A20 with ⌬RIG-I almost completely blocked ISRE luciferase reporter gene activity (Fig. 1B). A20 also blocked RIG-I-mediated activation of IRF-7 and the IFNA4 luciferase reporter gene. Expression of ⌬RIG-I activated IRF-7 and further enhanced IRF-7-mediated IFNA4 promoter activity 24-fold, whereas co-expression of A20 completely blocked IFNA4 (Fig. 1C). Similarly, A20 blocked RIG-I mediated NF-B activation (Fig. 1D); the P2 (2)-TK luciferase reporter (two copies of NF-B binding site from human IFNB promoter linked to minimal TK promoter) was induced 15-fold by ⌬RIG-I, an induction that was completely blocked by A20 co-expression. These experiments indicate that A20 is a strong inhibitor of RIG-I signaling to IRF-3 and NF-B.
A20 Blocks RIG-I-mediated Transactivation but Only Partially Inhibits TRIF-or IKK⑀/TBK1-mediated Transactivation-A20 targets the TLR-3 adapter TRIF and inhibits TLR-3-and Sendai virus-induced activation of ISRE and IFNB promoter (42). In addition, Saitoh and colleagues (41) showed that A20 physically interacted with IKK⑀/TBK1 and inhibited TLR-3-and Newcastle disease virus-mediated activation of IRF-3. To further investigate the level at which A20 inhibited ISRE activation, a dose-response curve was performed with increasing amounts of A20 and RIG-I, TRIF, IKK⑀, or TBK1 expression plasmids (Fig. 2). ⌬RIG-I (200 ng) resulted in 1500-fold induction of the ISRE promoter, whereas A20 expression plasmid (40 ng) inhibited ⌬RIG-I-  In all transfections, the pcDNA3 vector was added to bring the total plasmids to 1500 ng. Luciferase activity was analyzed at 24-h post-transfection by the Dual-Luciferase Reporter assay as described by the manufacturer (Promega). Relative luciferase activity was measured as -fold activation (relative to the basal level of reporter gene in the presence of pcDNA3 vector after normalization with co-transfected RLU activity); values are mean Ϯ S.D. for three experiments.

A20 Inhibition of RIG-I Signaling
mediated activation more than 15-fold; increasing the A20 concentration essentially reduced the ⌬RIG-I activation to background levels ( Fig.  2A). In contrast, when TRIF, IKK⑀, or TBK1 signaling components were used to activate ISRE promoter activity (2800-, 750-, and 850-fold, respectively), the inhibitory effect of A20 was significantly weaker, with only 2-to 3-fold inhibition observed at the highest concentrations of transfected A20 used (Fig. 2, B-D). This result argues that RIG-I signaling is the primary target for A20 inhibition, that TRIF-TLR-3 signaling is not significantly affected, and that the inhibitory effect occurs upstream of the IKK⑀ or TBK1 kinases.
Silencing of A20 Expression Enhances Virus-mediated Activation of ISRE Promoter-Next, to determine whether interference with endogenous A20 expression would modulate ISRE promoter activity, HEK293 cells were transfected with an small interference RNA expression construct directed against A20 (44) and subsequently infected with Sendai virus. Expression of A20 was induced by virus and effectively blocked by A20 specific small interference RNA but not a scrambled small interference RNA (Fig. 3A, lanes 2 and 3). Inhibition of A20 expression correlated with enhanced ISRE-dependent transcription induced by Sendai virus infection (Fig. 3B), thus demonstrating that endogenous A20 is involved in the regulation of ISRE-dependent promoter activity.

C-terminal Zinc Finger Domain of A20 Is Both Necessary and Sufficient to Block RIG-I-and Virus-mediated Activation of ISRE Promoter-
Wertz et al. (47) demonstrated that A20 down-regulated NF-B signaling through the cooperative activity of its two ubiquitin-editing domains. To determine which region of A20 was responsible for the inhibition of virus-and RIG-I-mediated signaling, a series of A20 dele-tion and point mutations were generated to test the inhibitory potential of each mutant on RIG-I induction; all modified proteins were expressed equivalently (Fig. 5). ⌬RIG-I activated the ISRE reporter 1200-fold, and expression of both full-length A20 and the C-terminal domain of A20 (aa373-790) inhibited ⌬RIG-I transactivation more than 500-fold. Conversely, truncation of the C terminus of A20 (aa1-380) only weakly inhibited transactivation of the ISRE promoter by ⌬RIG-I (Ͻ3-fold), indicating that the C-terminal ubiquitin ligase domain is important for A20-mediated inhibition.
Sendai virus was also used to activate the ISRE promoter and assess the inhibition mediated by A20 deletion and point mutations; essentially similar results demonstrating the requirement for the C-terminal zinc finger motifs were obtained (supplemental Fig. S1). Similar results were also obtained with an NF-B-dependent reporter construct (supplemental Fig. S2), altogether demonstrating that the C-terminal ubiquitin ligase domain is necessary and sufficient for A20-mediated inhibition of RIG-I-and virus-induced activation of IRF-3 and NF-B.
A20 Does Not Interact with ⌬RIG-I-To determine if A20 could interact with RIG-I, TRIF, IKK⑀, or TBK1, co-immunoprecipitation of A20 and RIG-I was performed with Myc-and FLAG-tagged proteins (Fig. 6). After immunoprecipitation of Myc-tagged proteins from cell extracts, immunoblot analysis revealed that FLAG-tagged A20 did not co-precipitate with Myc-tagged ⌬RIG-I (Fig. 6, lane 3) and only weakly interacted with full-length RIG-I (Fig. 6, lane 4). As a positive control, FLAG-tagged A20 co-precipitated with Myc-tagged A20 zinc finger domain (Fig. 6, lanes 2).
Because the inhibition of RIG-I-induced activation by A20 is mediated through its C-terminal ubiquitin ligase domain, we then examined whether A20 targeted RIG-I for degradation. Increasing A20 expression significantly reduced the level of TRIF (Fig. 7B, lane 4, bottom panel) and at high levels of A20, the amount of ⌬RIG-I was also reduced (Fig. 7A,   lane 4, bottom panel). However, A20 had essentially no effect on the level of IKK⑀ or TBK1 expression (Fig. 7, C and D, lane 4, bottom panel). The complete loss of TRIF protein by A20 did not, however, correlate with the 2-to 3-fold inhibition of TRIF-mediated transactivation (Fig.  2B); furthermore, because A20 inhibition of ⌬RIG-I-mediated transactivation was far greater (Ͼ100-fold) than the decrease in ⌬RIG-I protein (2-to 3-fold), we speculate that the target of A20 is an as yet unidentified adapter of the RIG-I pathway. Furthermore, the capacity of A20 to inhibit RIG-I signaling but not TRIF-TLR3 signaling and yet target TRIF for degradation, is reminiscent of the effect of the hepatitis C virus NS3/4A protease on these two pathways (37,38).
A20 Repression of the ⌬RIG-I-induced Antiviral State-To determine whether the A20 could repress the ⌬RIG-I-stimulated expression of endogenous IFN and ISG genes, HEK293 cells that stably express RIG-I, ⌬RIG-I, ⌬RIG-I, and A20 and ⌬RIG-I and HCV NS3/4A were generated (Fig. 8). The polyclonal FLAG-A20 and control HEK293 cells were infected with Sendai virus or treated with IFN␣2 to examine endogenous ISG56 and RIG-I protein expression. As shown in Fig. 8A, ISG56 and RIG-I were highly expressed in virus-infected control HEK293 cells (Fig. 8A, lane 2), whereas in virus-infected A20-expressing cells, ISG56 and RIG-I were inhibited (Fig. 8A, lane 5). These proteins were not detected in uninfected cells (Fig. 8A, lanes 1 and 4), and as a relative measure of specificity, actin expression was not altered by A20. Importantly, IFN␣2-mediated induction of ISG56 and RIG-I was not inhibited by A20 expression, indicating the specificity of A20 inhibition for the RIG-I but not the Jak-STAT pathway (Fig. 8A, lanes 2 and 5). The expression of constitutively active form of RIG-I strongly induced ISG56 and RIG-I protein expression in the absence of virus infection or IFN treatment (Fig. 8B, lanes 2 and 3), whereas co-expression of A20 or HCV protease NS3/4A completely blocked the ⌬RIG-I-mediated activation of ISG56 and RIG-I gene expression.
To further evaluate the antiviral state, the stable HEK293 cell lines were infected with VSV, and viral protein expression was measured at different times after infection. In control or RIG-I-expressing cells, VSV proteins (nucleocapsid (N), surface glycoprotein (G), and matrix (M))  were detected at 8 h post-infection, whereas in ⌬RIG-I-expressing cells, VSV replication was significantly delayed with viral proteins detected only at a low level beginning at 12 h post-infection (Fig. 9A). Interestingly, in cells that expressed either A20 or NS3/4A, a restoration of the kinetics of VSV expression was observed, with viral proteins again detectable as early as 8 h post-infection (Fig. 9B). Taken together, these results demonstrate that A20 can efficiently block the RIG-I-mediated signaling pathway and down-regulate cellular antiviral response. Although the target of A20 remains unknown, the similarity between A20-mediated inhibition and HCV NS3/4A inhibition suggests that the cellular A20 and the viral NS3/4A may be targeting related components or adapters of the RIG-I pathway.
A20 Inhibits MAVS/VISA/IPS-1/CARDIF-mediated Activation-Recently, the adaptor molecule that links RIG-I sensing of incoming viral RNA and downstream activation events was elucidated by four independent groups (48 -51). Under the name of Cardif, this protein was cleaved at its C-terminal end, adjacent to the mitochondrial targeting domain, by the NS3/4A protease of hepatitis C virus (51). To determine the ability of A20 to inhibit MAVS/VISA/IPS-1/CARDIF-mediated activation of IRF-3, an ISRE reporter construct together with MAVS/ VISA/IPS-1/CARDIF and increasing amounts of A20-or NS3/4A-expressing plasmids were co-transfected into HEK293 cells. As shown in Fig. 10A, the expression of MAVS/VISA/IPS-1/CARDIF stimulated the ISRE luciferase reporter gene activity up to 1100-fold, whereas co-expression of NS3/4A or A20 almost completely blocked ISRE luciferase reporter gene activity. Increasing levels of co-expression of NS3/4A dramatically altered the subcellular localization of MAVS/VISA/IPS-1/ CARDIF from the insoluble fraction of cytoplasmic extracts to the soluble fraction and directly targeted MAVS/VISA/IPS-1/CARDIF for proteolytic cleavage (Fig. 10B, lanes 1-4), whereas the expression of increasing amounts of the cellular protein A20 did not alter the subcellular localization or stability of MAVS/VISA/IPS-1/CARDIF (Fig. 10B,  lanes 5-7). These results indicated that A20 does not directly target MAVS/VISA/IPS-1/CARDIF protein.

DISCUSSION
The results of the present study demonstrate that cellular NF-Binduced, anti-apoptotic protein A20 efficiently blocks RIG-I-mediated signaling to the IRF and NF-B pathways. Furthermore, expression of A20 completely blocks ⌬RIG-I-induced IRF-3 Ser-396 phosphorylation, dimerization, and DNA binding. Mutational analysis of A20 demonstrated that the N-terminal de-ubiquitination domain had no effect on the inhibitory activity of A20, whereas the deletion of the C-terminal ubiquitin ligase domain almost completely ablated the inhibitory function of A20, an effect that was localized to the distal most zinc finger motifs. Finally, cells stably expressing the active form of RIG-I induced an antiviral state that delayed replication of VSV, an effect that was reversed by stable co-expression of A20 or the HCV-encoded NS3/4A protease. These results demonstrate that virus-mediated activation of A20 functions as a negative regulator of RIG-I-mediated induction of the antiviral state. A20 was originally characterized as a TNF-inducible gene in human umbilical vein endothelial cells (52). As an NF-B target gene (53), A20 is also induced in many other cell types by a wide range of stimuli, including virus infection. Overexpression of A20 has been shown to protect from TNF-␣-induced apoptosis and functions via a negativefeedback loop to block NF-B activation induced by TNF and other stimuli (54). A20-deficient mice (Tnfaip3 Ϫ/Ϫ ) were generated, and these mice developed severe multiorgan inflammation and were extremely susceptible to TNF due to the enhanced sensitivity to TNFinduced apoptosis (55). A20-deficient fibroblasts displayed prolonged NF-B activity and were unable to properly terminate TNF-induced NF-B activation; A20 was also essential for the termination of TLRinduced NF-B activation macrophages (56). The inhibition of both IRF-and NF-B-dependent pathways suggests that A20 functions in the physiological context as a negative feedback regulator of immune  response signaling. It will be of interest to determine the response of A20-deficient mice to pathogenic viruses.
It has been reported that A20 inhibits NF-B signaling through the cooperative activity of its two ubiquitin-editing domains: the N-terminal ovarian tumor (OTU) domain mediates de-ubiquitinating activity, and the C-terminal zinc finger region functions as ubiquitin ligase (47). Our results indicate that the C-terminal ubiquitin ligase domain of A20 is necessary and sufficient to block RIG-I-mediated signaling. Substitution of the indispensable cysteine residue in the catalytic OTU domain with alanine (A103) had essentially no effect on A20-mediated inhibition of RIG-I signaling (Fig. 5). More importantly, expression of the C-terminal ubiquitin ligase domain of A20 reduced ⌬RIG-I-mediated activation of ISRE promoter more than 100-fold (Fig. 5). The C-terminal ubiquitin ligase domain of A20 consists of seven novel zinc finger motifs of the type CX 2-4 CX 11 CX 2 C (57). Klinkenberg and colleagues (58) reported that the zinc finger motifs of murine A20 are functionally redundant, and A20 mutants containing a minimum of four zinc finger motifs are sufficient to inhibit TNF-induced NF-B activation to a level comparable to that obtained with the wild-type A20 protein. No strict requirement for a particular zinc finger structure was observed, because A20 mutants containing either the first four or last four zinc finger motifs had full inhibitory activity (58). In contrast, Natoli et al. (59) demonstrated that the last zinc finger of A20 was absolutely required for inhibition of TNF-induced NF-B activation. Here we show that deletion or mutation of zinc finger motif 7 reduced the inhibitory potential of human A20 more than 70-fold on ⌬RIG-I-induced ISRE activation, whereas deletion of internal zinc finger motifs 2 through 4 reduced the inhibitory potential of human A20 only 2-to 4-fold. These results demonstrate that the last zinc finger motifs of human A20 are absolutely required for inhibition of RIG-I-induced ISRE activation. A20 was shown recently to be involved in negative regulation of TLR-3-and Sendai virus-mediated activation of IRF-3 (42). Wang and colleagues reported that A20 interacted with TRIF and inhibited TRIFbut not TBK1-and IKK⑀-induced activation of ISRE and IFNB promoter (42). Saitoh and colleagues (41) demonstrated that A20 physically interacted with TBK1 and IKK⑀ and inhibited TLR-3-and Newcastle disease virus-mediated IRF-3 activation. The present studies confirm that A20-mediated inhibition of ISRE promoter activity correlated with the inhibition of IRF-3 activation; A20 completely blocked the ⌬RIG-Iinduced IRF-3 phosphorylation, dimerization, and protein-DNA complex formation. However, A20 only minimally reduced TRIF-and TBK1-mediated IRF-3 activation and did not inhibit TBK1-or IKK⑀induced gene activation (Fig. 3). The fact that A20 had no effect on the stability of either RIG-I itself or on TBK1/IKK⑀ suggested that the biological target of A20 may be an as yet unidentified adapter molecule that links sensing of virus infection by RIG-I with the downstream kinase activation. In support of this concept, recent studies demonstrated that hepatitis C virus (HCV) gene product NS3/4A protease strongly inhibited virus-and RIG-I-mediated activation of NF-B and IRF-3 (37, 39), but NS3/4A only weakly inhibited TRIF-mediated induction of NF-B and IRF-3 (37). Paradoxically, TRIF was identified as a proteolytic substrate of NS3/4A (46) despite the fact that the TRIF pathway does not appear to be a major pathway for IFN response to HCV; in vitro, RIG-I was not a proteolytic substrate for NS3/4A, and expression of NS3/4A did not alter the stability of RIG-I, again in contrast to the strong inhibition of the RIG-I pathway by NS3/4A (37,39). Recently, Meylan et al. (51) has shown that MAVS/VISA/IPS-1/CARDIF is a direct target by NS3/4A. Although A20 strongly inhibited RIG-I-and MAVS/VISA/ IPS-1/CARDIF-mediated transactivation of IRF and NF-B pathways, no strong association between RIG-I and A20 or MAVS/VISA/IPS-1/ CARDIF and A20 was detected by immunoprecipitation; A20 also only weakly inhibited TRIF-, TBK1-, or IKK⑀-mediated transactivation, further implicating a distinct target for A20 interaction and function. Studies are underway to characterize other components of the RIG-I signaling pathway as potential targets of A20 regulation.