NS5 Conservative Site Is Required for Zika Virus to Restrict the RIG-I Signaling

During host–virus co-evolution, cells develop innate immune systems to inhibit virus invasion, while viruses employ strategies to suppress immune responses and maintain infection. Here, we reveal that Zika virus (ZIKV), a re-emerging arbovirus causing public concerns and devastating complications, restricts host immune responses through a distinct mechanism. ZIKV nonstructural protein 5 (NS5) interacts with the host retinoic acid-inducible gene I (RIG-I), an essential signaling molecule for defending pathogen infections. NS5 subsequently represses K63-linked polyubiquitination of RIG-I, attenuates the phosphorylation and nuclear translocation of interferon regulatory factor 3 (IRF3), and inhibits the expression and production of interferon-β (IFN-β), thereby restricting the RIG-I signaling pathway. Interestingly, we demonstrate that the methyltransferase (MTase) domain of NS5 is required for the repression of RIG-I ubiquitination, IRF3 activation, and IFN-β production. Detailed studies further reveal that the conservative active site D146 of NS5 is critical for the suppression of the RIG-I signaling. Therefore, we uncover an essential role of NS5 conservative site D146 in ZIKV-mediated repression of innate immune system, illustrate a distinct mechanism by which ZIKV evades host immune responses, and discover a potential target for anti-viral infection.


INTRODUCTION
Zika virus (ZIKV), a re-emerging arbovirus, has raised public concerns due to its global spread and clinical symptoms. Since 2007, ZIKV infection has caused a series of epidemics in Micronesia, the South Pacific, and most recently the Americas (1). The viral infection is responsible for the development of devastating complications, including Guillain-Barre Syndrome (2), meningoencephalitis in adults, and microcephaly in fetuses (3), as well as testis damage and infertility in male mice (4). ZIKV is a single-stranded positive-sense RNA virus (5), and its genome contains a 5 ′ -untranslated region (UTR), an open reading frame (ORF), and a 3 ′ -UTR (6). The ORF encodes a single polyprotein that is processed into capsid protein (C), precursor membrane protein (prM), envelope protein (E), and nonstructural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) (7).
During infection, ZIKV RNA is recognized by RIG-I (20). ZIKV has developed several strategies to limit host immune responses and to successful replicate and spread (21). ZIKV NS5 suppresses type I IFN by targeting STAT2 for degradation (22). An antagonistic system employing multiple ZIKV NS proteins restricts antiviral responses by limiting the JAK-STAT signaling (23). ZIKV NS5 suppresses IFN-β production by targeting IRF3 and TBK1 (24,25). Here, we reveal that ZIKV NS5 antagonizes IFN-β production by targeting RIG-I. NS5 interacts with RIG-I to inhibit RIG-I K63-linked polyubiquitination, IRF3 phosphorylation and nuclear translocation, and IFNβ production, thereby repressing the RIG-I signaling. More interestingly, NS5 conservative site D146 is required for NS5 in the suppression of RIG-I. Thus, this work uncovers an essential function of conservative site D146 in the regulation of IFN-β production and RIG-I signaling, and reveals a distinct mechanism by which ZIKV restricts antiviral responses.

Viruses
The ZIKV isolate z16006 (GenBank accession number, KU955589.1) isolated by the Institute of Pathogenic Microbiology, Center for Disease Control and Prevention of Guangdong (Guangzhou, Guangdong, China) was used in this study. C6/36 cells were maintained at 30 • C in DMEM (Gibco) (Grand Island, NY, USA) supplemented with 10% heat-inactivated FBS with penicillin and streptomycin (Gibco) (Grand Island, NY, USA) and 1% tryptose phosphate broth (Sigma) (St. Louis, MO, USA). To free the ZIKV stocks used in this study of mycoplasma contamination, it was tested by using the MycoTest Kit (ChanGEnome, Chian) and by using TEM. SeV was propagated in embrocated eggs and titrated by blood coagulation test.

Nuclear and Cytoplasmic Extraction
In a 12-well plate, 60-70% confluent HeLa cells were transfected with the indicated plasmids for 24 h, then disposed by using by nuclear and cytoplasmic extraction reagents (Thermo scientific, 78833, USA). Cytosol or nuclear lysate concentration was determined by Bradford assay (Bio-Rad, Hercules, CA, USA).

Confocal Microscopy
In a 24-well plate, 40%−50% HEK293T or HeLa cells were transfected with the indicated plasmids (500 ng) for 24 h, then cells were washed twice with PBS and fixed in 4% paraformaldehyde at room temperature for 10 min, then permeabilized with wash buffer (PBS containing 0.1% Triton X-100) for 5 min, washed three times with PBS, and finally blocked with PBS containing 5% BSA for 1 h. The cells were then incubated with the primary antibody overnight at 4 • C, followed by incubated with FITC-conjugate donkey anti-mouse IgG and Dylight 649-conjugate donkey anti-rabbit IgG (Abbkine) for 1 h. Using wash buffer three times, cells were incubated with DAPI solution for 5 min, and then washed three more times with PBS. Finally, the cells were analyzed using a confocal laser scanning microscope (Fluo View FV1000; Olympus, Tokyo, Japan).

Statistical Analyses
All experiments were reproducible and repeated at least three times with similar results. Samples were analyzed by one-way analysis of variance with Tukey's post-hoc test. Abnormal values were eliminated using a follow-up Grubbs test. A Levene's test for equality of variances was performed, which provided information for Student's t-tests to distinguish the equality of means. Means were illustrated using histograms, with error bars representing standard error of the mean (s.e.m); values of P < 0.05 were considered to indicate statistical significance ( * P < 0.05, * * P < 0.01, and * * * P < 0.001).

ZIKV NS5 Represses IFN-β Production by Targeting the RIG-I Pathway
IFN-β plays an important role in activating immune cells and suppressing virus replication (26)(27)(28)(29)(30)(31), and ZIKV infection leads to low levels of type I IFNs (32). Here, we initially showed that IFN-β mRNA was significantly induced by poly(I:C), but such induction was suppressed by ZIKV infection ( Figure 1A). Additionally, IFN-β-Luc activity was induced upon Sendai virus (SeV) infection, but the induction was suppressed by ZIKV in A549 cells ( Figure 1B) or Hela cells ( Figure 1C). These results demonstrate that ZIKV suppresses IFN-β expression by the stimulation of poly(I:C) or SeV. IFN-β-Luc activity was induced upon Sendai virus (SeV) infection ( Figure 1D) or by poly(I:C) treatment ( Figure 1E), but the induction was suppressed by NS5 in HEK293T cells (Figures 1D,E). Moreover, endogenous IFN-β mRNA was induced upon SeV infection ( Figure 1F) and by poly(I:C) treatment ( Figure 1G), but such induction was attenuated by NS5 (Figures 1F,G). These results demonstrate that NS5 suppresses IFN-β expression upon the infections of SeV or by the stimulation of poly(I:C). Since ZIKV genome is recognized by RIG-I, we investigated whether NS5 affects RIG-I function.
Overexpression of NS5 in HEK293 cells attenuated the activation of IFN-β promoter luciferase reporter activity by RIG-I and MAVS (Figures 1H,I). Taken together, we demonstrate that ZIKV suppresses IFN-β production by repressing the RIG-I signaling through NS5.

NS5 Restricts IRF3 Phosphorylation and Nuclear Translocation
IRF3 is an important component of the RIG-I pathway and activation of IRF3 depends on phosphorylation, which leads to IRF3 nuclear localization and IFN-β production (18). Here, we investigated the effect of NS5 on IRF3 phosphorylation and nuclear translocation. IRF3 phosphorylation was induced upon SeV infection (Figure 2A, lane 2 vs. 1), whereas NS5 significantly repressed SeV-induced IRF3 phosphorylation but not IRF3 production ( Figure 2C) and Hela cells (Figure 2D), IRF3 alone was mainly distributed in the cytoplasm, a small proportion of NS5 was located in the cytoplasm, and a large proportion of NS5 was distributed in the nucleus; however, in SeV-infected HEK293T cells ( Figure 2C) and Hela cells (Figure 2D), IRF3 was translocated from the cytoplasm to the nucleus in the absence of NS5; and interestingly, most of IRF3 remained in the cytoplasm in the presence of NS5. Taken together, the results reveal that NS5 restricts IRF3 phosphorylation and nuclear translocation, but not IRF3 production, thereby repressing IRF3 activation.

NS5 Binds to the CARD Domain of RIG-I
The mechanism by which NS5 represses the RIG-I signaling was elucidated. Initially, we determined whether NS5 interacts with the RIG-I signaling components. Interestingly, NS5 interacted with RIG-I, TBK1, and IRF3 ( Figure 3A, lanes 2, 4, and 6), but failed to interact with MAVS or IKKε ( Figure 3A, lanes 3 and 5). Co-IP further confirmed that NS5 associated with RIG-I ( Figure 3B) and RIG-I interacted with NS5 ( Figure 3C). Additionally, NS5 was co-immunoprecipitated with endogenous RIG-I in ZIKV-infected Hela cells and IFNAR −/− MEF cells (Figures 3D,E). Moreover, in HEK293T cells ( Figure 3F) and Hela cells (Figure 3G), RIG-I was diffusely distributed in the cytoplasm, a small proportion of NS5 was distributed in the cytoplasm, and a large proportion of NS5 was distributed in the nucleus; however, most of RIG-I was internalized in the cytoplasm and a proportion of NS5 was co-localized with RIG-I in the cytoplasm (Figures 3F,G). Thus, these results demonstrate that NS5 interacts with RIG-I. Since RIG-I contains three domains, CARD, DExD/H, and RD (14), we determined which domain is involved in the interaction with NS5 and revealed that NS5 interacted with RIG-I and CARD domain (Figure 3H, lanes 2 and 5), but not with DExD/H domain or RD domain ( Figure 3H, lanes 3 and 4). Therefore, the results demonstrate that NS5 binds to the CARD domain of RIG-I.  (Figure 4B, lane 4). Moreover, RIG-I polyubiquitination was weakly catalyzed by K63R (lysine 63 of ubiquitin was mutated into arginine) ( Figure 4C, lane 1), which was not affected by NS5 (Figure 4C, lane 2); however, RIG-I polyubiquitination was strongly catalyzed by K48R (lysine 48 of ubiquitin was mutated into arginine) (Figure 4C, lane 3), which was attenuated by NS5 ( Figure 4C, lane 4). The effects of NS5 on the regulation of RIG-I polyubiquitination upon virus infection or RIG-I signaling activation was then investigated. RIG-I polyubiquitination catalyzed by HA-Ub ( Figure 4D) or HA-K63 ( Figure 4E) was induced upon SeV infection, but such induction was repressed by NS5 (Figures 4D,E). Interestingly, RIG-I polyubiquitination catalyzed by Ub (Figure 4F, upper) or K63 (Figure 4F, lower) was induced by SeV, but such induction was repressed by ZIKV infection (Figure 4F). RIG-I possesses two caspase activation and recruitment domains (CARDs), a DExD/H-box helicase domain, and a repressor domain (RD), and RIG-I undergoes robust ubiquitination at its N-terminal CARD domain. We showed that RIG-I(2CARD) polyubiquitination was catalyzed by HA-Ub (Figure 4G) or by HA-K63 (Figure 4H), but attenuated by NS5 (Figures 4G,H). Taken together, we demonstrate that NS5 impairs K63-linked polyubiquitination of RIG-I CARD domain.
Conservative Site D146 Is Essential for NS5 in the Suppression of IFN-β NS5 plays a key role in the replication of viral genome and contains a methyltransferase (MTase) domain and an RNAdependent RNA polymerase (RdRp) domain (35,36). Here,   Figure 5D, lane 4 vs. 2). Additionally, IFN-β-Luc activity was induced by SeV ( Figure 5E, lane 2 vs. 1), and such induction was suppressed by NS5 (Figure 5E, lane 3 vs. 2) but not by NS5-D146A (Figure 5E, lane 4 vs. 2). These results reveal that the conservative site D146 is essential for NS5 in the repression of IFN-β. These results indicate that D146 is involved in the repression of virus-induced phosphorylation of IRF3. Moreover, in mockinfected cells, IRF3 was diffusely distributed in the cytoplasm, a small proportion of NS5 and NS5-D146A was located in the cytoplasm, and a large proportion of NS5 and NS5-D146A was distributed in the nucleus (Figure 6C, top); however, in SeVinfected cells, IRF3 was translocated from the cytoplasm to the nucleus in the absence of NS5 or in the presence of NS5-D146A, but remained in the cytoplasm in the presence of NS5 ( Figure 6C); revealing that D146 is essential for NS5 in the repression of IRF3 nucleus translocation.
MTase Activity Does Not Contribute to NS5 Repressing RIG-I Signaling Because we found that MTase site D146 is required for the repression of RIG-I signaling, we asked whether NS5 mediates methylation of RIG-I to affect its activity. We found that overexpression of the NS5 did not lead to increased monomethylation of RIG-I in HEK293T cells (Figure 7A, lane 4 vs. 3). Moreover, we also determined whether MTase activity contributed to NS5 repressing RIG-I signaling. It has been reported that MTase inhibitors SAH can inhibit ZIKV NS5 MTase activity (38). In HEK293T cells, IFN-β-Luc activity induced by RIG-I(2CARD) was attenuated by NS5. Similarly, NS5 also attenuated the IFN-β-Luc activity in HEK293T cells adding SAH (Figure 7B). In addition, NS5 attenuates K63-linked polyubiquitination of RIG-I. We also found that adding SAH did not influence NS5 suppressing polyubiquitination or K63linked polyubiquitination of RIG-I (Figures 7C,D). These data suggest that NS5 did not increase mono-methylation of RIG-I and its MTase activity did not contribute to NS5 repressing RIG-I signaling.

DISCUSSION
Host cells have developed innate immune systems to inhibit the virus invasion and replication (39), while viruses must employ strategies to evade host immune systems to maintain infection and replication (40). We have recently demonstrated that ZIKV NS5 induces host inflammatory responses by facilitating the NLRP3 inflammasome assembly and interleukin-1β secretion (41). Here, we show a distinct mechanism by which ZIKV restricts host antiviral immune responses by targeting the RIG-I signaling. ZIKV suppresses IFN-β expression mediated by SeV infection and poly(I:C) stimulation, which is consistent with a previous study showing that ZIKV infection leads to low levels of type I IFN production (32). NS5 represses IFN-β activation induced by SeV infection and poly(I:C) stimulation, suggesting that ZIKV represses type I IFN through NS5. Previous studies reported that ZIKV NS5 interferes with type I IFN signaling by targeting STAT2 for proteasomal degradation (22), while NS1 and NS4B prevent RLR pathway activation by targeting TBK1 to inhibit IFN-β, NS2B3 impairs the JAK-STAT pathway through degrading JAK1 (23) and attenuates the cGAS/STING pathway by cleaving STING (42), and ZIKV suppresses IFN-β expression mediated by SeV infection and poly(I:C) stimulation (32). Two recent studies demonstrated that NS5 limits IFN-β by interacting with IRF3 and TBKI (24,25). Therefore, our results are consistent with the conclusions of previous reports, and further illustrate a distinct mechanism that NS5 antagonizes IFN-β production by interacting with RIG-I. IFN-β plays important roles in the activation of immune cells and the suppression of virus replication (26)(27)(28)(29)(30)(31). During this process, IRF3 phosphorylation and nuclear translocation are essential for IFN-β production upon virus infection (18). We reveal that NS5 restricts IRF3 phosphorylation and nuclear translocation, thereby repressing IRF3 activation and IFN-β production.
The molecular mechanism by which NS5 suppresses IFNβ production and RIG-I signaling is revealed. Two studies have shown that NS5 blocks dsRNA-stimulated IFN response by interacting with IRF3 or TBK1 (24,25). Interestingly, our results are consistent with their data, and we also reveal that NS5 interacts with RIG-I, TBK1, and IRF3, and fails to interact with MAVS or IKKε. Here we mainly focus on the study  of the molecular mechanism by which NS5 suppresses IFNβ production by interacting with RIG-I. We further reveal that NS5 binds to RIG-I through the CARD domain. RIG-I K63-linked polyubiquitination is crucial for the pathway to elicit host antiviral immune responses. West Nile virus (WNV) and influenza A virus (IAV) have developed diverse strategies to minimize IFN by decreasing RIG-I K63-linked polyubiquitination (43,44). Interestingly, we reveal that ZIKV represses RIG-I K63-linked polyubiquitination through NS5. These results suggest that NS5 may play a role in the regulation of viral infection through repressing RIG-I ubiquitination, since RIG-I ubiquitination plays a key role in the regulation of viral infection (34).
Functional analyses of the domains of ZIKV NS5 reveal that the MTase domain, but not the RdRp domain, represses IFN-β activation by targeting RIG-I. The conservative catalytic tetrad of K61-D146-K182-E218 is positioned in the MTase domain to form the active site of MTase (45). Our results demonstrate that the conservative site D146 is essential for NS5 in repressing IFN-β production, IRF3 activation, and RIG-I K63-linked polyubiquitination. Since D146 is an NS5 MTase catalytic site, we supposed that MTase activity may contribute to NS5 suppressing RIG-I signaling. But we found that ZIKV NS5 does not lead to increased methylation of RIG-I and NS5 MTase activity does not contribute to NS5-attenuated IFN-β activation and RIG-I K63-linked polyubiquitination. Therefore, we suggest that the conservative site D146 is important for NS5 repressing IFN-β production, IRF3 activation, and RIG-I K63-linked polyubiquitination, but it has no relation with the NS5 MTase activation. The conservative site D146 may play an important role in maintaining the NS5 space structure; this hypothesis needs to be further investigated.
In conclusion, this study reveals a distinct mechanism by which ZIKV restricts antiviral response by targeting RIG-I. NS5 binds to RIG-I through interacting with the CARD domain, resulting in the restriction of RIG-I polyubiquitination, IRF3 activation, and IFN-β production, thereby inhibiting the RIG-I signaling. These results provide evidence that ZIKV NS5 interferes with the host immune system by targeting RIG-I. More interestingly, MTase active site D146 is essential for the repression of RIG-I signaling, but MTase activity does not contribute to NS5 suppressing RIG-I signaling.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.