SARS-CoV-2 membrane glycoprotein M antagonizes the MAVS-mediated innate antiviral response

A novel SARS-related coronavirus (SARS-CoV-2) has recently emerged as a serious pathogen that causes high morbidity and substantial mortality. However, the mechanisms by which SARS-CoV-2 evades host immunity remain poorly understood. Here, we identified SARS-CoV-2 membrane glycoprotein M as a negative regulator of the innate immune response. We found that the M protein interacted with the central adaptor protein MAVS in the innate immune response pathways. This interaction impaired MAVS aggregation and its recruitment of downstream TRAF3, TBK1, and IRF3, leading to attenuation of the innate antiviral response. Our findings reveal a mechanism by which SARS-CoV-2 evades the innate immune response and suggest that the M protein of SARS-CoV-2 is a potential target for the development of SARS-CoV-2 interventions.


INTRODUCTION
The innate immune response is the first line of host defense against viral infection and is initiated by the recognition of pathogenassociated molecular patterns by cellular pattern recognition receptors (PRRs). This triggers a series of signaling events that lead to the induction of type I interferons (IFNs), proinflammatory cytokines, and other antiviral effector genes. [1][2][3][4] These downstream effector proteins mediate the innate antiviral response and facilitate adaptive immunity to suppress viral replication and eliminate viruses in vivo. 5,6 Among the PRRs, retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs), including RIG-I and melanoma differentiation associated gene 5 (MDA5), have been demonstrated to be cytosolic viral RNA sensors in mammals. [6][7][8] Although RIG-I and MDA5 have a certain specificity for sensing distinct types of viral RNAs, they utilize a common adaptor protein called MAVS (also known as VISA, IPS-1 or Cardif) to trigger downstream signal transduction. [9][10][11][12] After binding viral RNA, RLRs are recruited to the mitochondria-located MAVS via their respective CARD domains, 1 leading to the aggregation and formation of prionlike MAVS polymers. 13 The aggregated MAVS acts as a central platform for the recruitment of downstream signaling components, including TRAF2/3/5/6, TBK1 and IKK, leading to the activation of the transcription factors IRF3 and NF-κB and to the induction of downstream antiviral effector genes. [14][15][16][17] SARS-CoV-2, a member of the coronavirus family, is a typical single-stranded positive-sense RNA virus that encodes over 28 proteins, including 4 structural proteins (spike, membrane, envelope, and nucleocapsid), 16 nonstructural proteins (NSP1-NSP16), and 8 auxiliary proteins (ORF3a, ORF3b, ORF6, ORF7a, ORF7b, ORF8, ORF9b and ORF14). [18][19][20] Recently, SARS-CoV-2 has caused a global pandemic and worldwide social and economic disruption. Patients with severe SARS-CoV-2 infection usually have dyspnea after 1 week, which then rapidly manifests as acute respiratory distress syndrome and results in death. [21][22][23][24][25] Dysregulation of IFNs has been found in patients with severe COVID-19. 26,27 However, little is known about how SARS-CoV-2 evades the immune system. SARS-CoV-2 membrane protein M is a 222-amino acid glycosylated structural protein with three N-terminal membranespanning domains, which are essential for the assembly of viral particles. 18 In this study, we identified the SARS-CoV-2 M protein as an inhibitor of the MAVS-mediated innate antiviral immune response. Our results suggest that M targets MAVS and impairs its aggregation and recruitment of downstream signaling components. Our findings reveal a mechanism by which SARS-CoV-2 evades the innate immune response.

RESULTS
SARS-CoV-2 protein M antagonizes the viral RNA-triggered innate antiviral immune response It has been demonstrated that RLRs mediate the innate immune response to RNA viruses, including SARS-CoV. [28][29][30] To identify SARS-CoV-2 proteins that may inhibit the RLR-mediated induction of downstream antiviral genes, we constructed 17 SARS-CoV-2 protein expression clones and screened for candidates that inhibit the Sendai virus (SeV, an RNA virus)-induced activation of the IFNβ promoter in HEK293 cells by reporter assays (Fig. 1A). This screen identified the SARS-CoV-2 ORF3a, M, N, ORF9b and NSP1 proteins as candidates. In reporter assays, ectopic expression of the M protein dose-dependently inhibited the SeV-induced activation of We next performed ELISA experiments and found that the secretion of IFN-β and TNF-α following SeV infection or poly (I:C) transfection was also impaired in HEK293-M cells (Fig. 1G). In addition, M inhibited the phosphorylation of TBK1, IKKα/β, IRF3 and p65 induced by SeV or poly(I:C) (Fig. 1H). Consistently, confocal microscopy showed that compared with that in the control cells, the SeV-triggered nuclear translocation of IRF3 was disrupted in cells ectopically expressing M (Fig. 1I). In contrast, M had no marked effects on the IFN-β-induced transcription of the ISG56 and CXCL10 genes (Fig. 1J). Since the above experiments were performed with 293 cells stably expressing M, we further evaluated whether M affected cell viability. Cell proliferation assays confirmed that M had no effects on cell viability (Fig. 1K). Taken together, these results suggest that M acts as an inhibitor of the RNA virus-triggered innate immune response. induced by overexpression of RIG-I-CARD, MDA5 and MAVS but not p65 (Fig. 2C). These results suggest that M inhibits innate antiviral signaling at the MAVS level. Coimmunoprecipitation experiments indicated that M specifically associated with MAVS but not RIG-I, MDA5 or TBK1 in the mammalian overexpression system (Fig. 3A, B). Further experiments indicated that the M protein was constitutively associated with endogenous MAVS before and after SeV infection in HEK293 cells (Fig. 3C). In vitro GST pull-down assays with recombinant proteins indicated that the M protein directly interacted with MAVS in a dose-dependent manner (Fig. 3D). Domain-mapping experiments indicated that the transmembrane domain of MAVS was required for its association with M. Among the examined truncations, MAVS (360-540) interacted with the M protein most strongly, while MAVS (180-540) and MAVS (100-540) interacted with the M protein weakly, suggesting that the middle fragment (aa100-360) of MAVS inhibits its interaction with the M protein (Fig. 3E). All the examined M truncations, including TM1/2 (contained only the two N-terminal transmembrane domains) and ΔTM1/2 (lacked the two N-terminal transmembrane domains), interacted with MAVS (Fig. 3E), suggesting that the M protein could interact with MAVS via its different fragments. Reporter assays indicated that the two N-terminal transmembrane domains of the M protein were required for its inhibition of SeV-induced IFNβ promoter activation (Fig. 3F) as well as for the MAVSmediated transcription of the downstream IFNB1 and CXCL10 genes (Fig. 3G). A possible explanation is that although multiple motifs of M could mediate its association with VISA, only the TM1/ 2 domains inhibit the activation of VISA. Taken together, these results suggest that M inhibits the innate antiviral response by targeting MAVS, and its TM1/2 domains are essential for this inhibitory function.

M impairs MAVS aggregation and its recruitment of downstream components
The prion-like aggregation of MAVS was previously reported to be essential for activating the innate antiviral response after viral infection. 13 We investigated whether M impairs the formation of MAVS aggregates. As shown in Fig. 4A, semidenaturing detergent agarose gel electrophoresis (SDD-AGE) analysis indicated that the M protein and its TM1/2 truncation but not the ΔTM1/2 truncation impaired the aggregation of MAVS induced by poly (I:C) transfection. Previously, SNX8 was reported to be indispensable for MAVS aggregation after virus infection. 31 As shown in Fig. 4B, the M protein disrupted the self-association of MAVS and its association with SNX8. These results suggest that the M protein impairs the innate antiviral response by inhibiting MAVS aggregation.
Upon aggregation, MAVS acts as a central platform to recruit downstream signaling components such as TRAF3, TBK1 and IRF3, leading to the subsequent induction of downstream antiviral genes. 16 Consistently, coimmunoprecipitation experiments indicated that the M protein disrupted the recruitment of TRAF3, TBK1 and IRF3 to the MAVS complex but had no marked effects on the interaction between MAVS and RIG-I or MDA5 (Fig. 4C). In vitro GST pull-down assays showed that the M protein inhibited the binding of MAVS to TRAF3, TBK1 and IRF3 (Fig. 4D). These results suggest that the M protein impairs the recruitment of TRAF3, TBK1 and IRF3 to the MAVS complex, leading to the inhibition of the innate antiviral response.

DISCUSSION
During coevolution with hosts, viruses have evolved multiple mechanisms to evade host immune defense. [32][33][34] How SARS-CoV-2 evades the host immune defense is not well understood. In this study, we identified the SARS-CoV-2 M protein as a factor underlying the inhibition of host antiviral innate immunity by directly targeting the central adaptor MAVS in the RLR-mediated induction of type I IFNs.
Several lines of evidence suggest that M directly targets MAVS to inhibit the innate immune response. The M protein specifically inhibited the SARS-CoV-2-, SeV-and poly (I:C)-induced transcription of downstream antiviral genes. The M protein suppressed signaling mediated by RIG-I, MDA5 and MAVS but not their downstream components TBK1 or p65. Coimmunoprecipitation and in vitro pull-down assays indicated that the M protein directly interacted with MAVS. Further studies suggested that the M protein impaired MAVS aggregation induced by viral RNA as well as its recruitment of downstream components such as TRAF3, TBK1, and IRF3. These results suggest that the M protein evades the innate antiviral response by impairing MAVS activation.
Previously, SARS-CoV and MERS-CoV were shown to have developed multiple mechanisms to evade the innate immune response. For example, SARS-CoV M, N, 9b, and NSP1 as well as MERS-CoV M, ORF4a, and ORF4b have been demonstrated to target different signaling components in the innate immune response pathways. 28,29,35,36 Some recent screenings have identified SARS-CoV-2 M, N, ORF3a, ORF6, NSP1,NSP3, NSP12, NSP13, NSP14 and NSP15 as potential candidates for inhibiting the induction of IFNβ . 37,38 In particular, it has been shown that the SARS-CoV M protein impairs the innate immune response by inhibiting the TRAF3/TANK/TBK1 complex, 28,35 whereas the MERS-CoV M protein impairs the TBK1-mediated phosphorylation of IRF3. 39 Considering these results and those of our studies, it is possible that the SARS-CoV family of viruses utilizes their M proteins as a general strategy to evade the host innate immune system. Our current findings facilitate the understanding of the complicated interaction between SARS-CoV-2 and the host and provide a potential target for the development of interventions for SARS-CoV-2.

Plasmid construction
The IFNβ promoter, ISRE and NF-κB luciferase reporter plasmids, and mammalian expression plasmids for Flag-and HA-tagged RIG-I, MDA5, MAVS, TRAF3, and TBK1 were previously described 14,43 Flag-, MyC-, and HA-tagged M and their mutants were constructed by standard molecular biology techniques.
Transfection and reporter assays Luciferase assays were performed using a dual-specific luciferase assay kit (Promega). Cells (1 × 10 5 ) were seeded in 48-well plates and transfected the following day by standard calcium phosphate  3   TM1  TM2 TM3   TM2  TM3   TM1  TM3   TM1  TM2   TM3   TM1   TM1  TM2 TM3   TM1  precipitation. In the same experiment, empty control plasmid was added to ensure that the same amount of total DNA was used for each transfection. To normalize for transfection efficiency, 0.01 μg of the pRL-TK (Renilla luciferase) reporter plasmid was added to each transfection. 31 Real-time PCR Total RNA was isolated for qPCR analysis to measure the mRNA levels of the indicated genes. The data shown are the relative abundances of the indicated mRNAs normalized to that of GAPDH. Primer sequences for IFNB1, ISG56, CXCL10, TNF, and GAPDH were previously described. 31,40 The SARS-CoV-2 S primers CTTCC CTCAGTCAGCACCTC (forward) and AACCAGTGTGTGCCATTTGA (reverse) as well as the SARS-CoV-2 E primers TCGTTTCGGAAGA GACAGGT (forward) and CACGAGAGTAAACGTAAAAAGAAGG (reverse) were utilized. ELISA HEK293-M or control cells were infected with SeV or poly (I:C) for 8 h. The culture media were collected for measurement of IFN-β and TNF-α by ELISA.
GST pull-down assay GST-M was bound to glutathione agarose beads and incubated with purified His-MAVS or lysate of HEK293T cells transiently expressing Flag-TBK1, Flag-IRF3, or Flag-TRAF3 along with an empty or HA-M plasmid for 3 h. The beads were collected and washed three times with lysis buffer (20 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 10 μg/mL aprotinin, 10 μg/mL leupeptin, 1 mM phenylmethylsulfonyl fluoride). The beads were then mixed with SDS loading buffer and boiled for 10 min. The input/eluates were resolved by SDS-PAGE and analyzed by Coomassie staining and/or immunoblot.

MTT assay
The cells were plated in 96-well culture plates (2000 cells/200 µL/ well for all cell lines) and incubated in a 5% CO2 incubator. Further, at the respective time points, 10 µL of the stock MTT solution (5 mg/mL) was added, and the cells were incubated in a 5% CO2 incubator for 4 h. The medium was removed, and formalin crystals formed by the cells were dissolved in 100 µL of DMSO. The absorbance was read at a 570 nm wavelength on a multiwell plate reader.

Statistics
Unpaired Student's t test was used for statistical analysis with GraphPad Prism software; *P < 0.05 and **P < 0.01 were considered significant, and ns indicates no significance.