key: cord-261688-njlxrxv6
authors: Yang, Ziwei; Zhang, Xiaolin; Wang, Fan; Wang, Peihui; Kuang, Ersheng; Li, Xiaojuan
title: Suppression of MDA5-mediated antiviral immune responses by NSP8 of SARS-CoV-2
date: 2020-08-12
journal: bioRxiv
DOI: 10.1101/2020.08.12.247767
sha: 
doc_id: 261688
cord_uid: njlxrxv6

Melanoma differentiation-associated gene-5 (MDA5) acts as a cytoplasmic RNA sensor to detect viral dsRNA and mediates type I interferon (IFN) signaling and antiviral innate immune responses to infection by RNA viruses. Upon recognition of viral dsRNA, MDA5 is activated with K63-linked polyubiquitination and then triggers the recruitment of MAVS and activation of TBK1 and IKK, subsequently leading to IRF3 and NF-κB phosphorylation. Great numbers of symptomatic and severe infections of SARS-CoV-2 are spreading worldwide, and the poor efficacy of treatment with type I interferon and antiviral agents indicates that SARS-CoV-2 escapes from antiviral immune responses via an unknown mechanism. Here, we report that SARS-CoV-2 nonstructural protein 8 (NSP8) acts as an innate immune suppressor and inhibits type I IFN signaling to promote infection of RNA viruses. It downregulates the expression of type I IFNs, IFN-stimulated genes and proinflammatory cytokines by binding to MDA5 and impairing its K63-linked polyubiquitination. Our findings reveal that NSP8 mediates innate immune evasion during SARS-CoV-2 infection and may serve as a potential target for future therapeutics for SARS-CoV-2 infectious diseases. Importance The large-scale spread of COVID-19 is causing mass casualties worldwide, and the failure of antiviral immune treatment suggests immune evasion. It has been reported that several nonstructural proteins of severe coronaviruses suppress antiviral immune responses; however, the immune suppression mechanism of SARS-CoV-2 remains unknown. Here, we revealed that NSP8 protein of SARS-CoV-2 directly blocks the activation of the cytosolic viral dsRNA sensor MDA5 and significantly downregulates antiviral immune responses. Our study contributes to our understanding of the direct immune evasion mechanism of SARS-CoV-2 by showing that NSP8 suppresses the most upstream sensor of innate immune responses involved in the recognition of viral dsRNA.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is an emerging severe coronavirus that is 54 currently causing a global outbreak of Coronavirus Disease 2019 . It has infected more than twenty 55 million patients and caused more than two-thirds of a million deaths. The number of patients and deaths are 56 still rapidly increasing; however, no effective therapy, vaccine or cure are available. Knowledge of SARS-CoV-57 6 their induction upon stimulation, and NSP8 once again had suppressive effects on the expression of these 134 cytokines (Fig.2c , red vs. blue columns). Taken together, these results suggest that NSP8 suppresses MAVS-135 dependent innate immune responses, probably by acting on either MAVS or upstream RNA sensors. 136 137

Since viral RNA of coronaviruses contains a methylated 5'-cap and 3'-polyA tail that is similar to cellular 139 mRNA, we assumed that NSP8 may preferentially regulate MDA5-mediated responses rather than RIG-I-140 mediated responses that recognize 5'-pppRNA. Hence, we performed coimmunoprecipitation (co-IP) analysis 141 and found that NSP8 interacts with MDA5 (Fig.3a) . Furthermore, we mapped the binding domain and 142 determined that its CARD domain is responsible for the NSP8 interaction (Fig.3b) . As a result, MDA5-mediated 143 ISRE-luc activity was inhibited by NSP8 in a dose-dependent manner (Fig.3c ). In addition, confocal 144 microscopy demonstrated that NSP8 tightly colocalized with MDA5 inside cells (Fig.3d) . These results suggest 145 that NSP8 interacts with MDA5 and directly suppresses MDA5-mediated immune responses. 146

To further understand the molecular mechanism by which NSP8 interacts and interferes with MDA5, we 147 predicted the MDA5 CARD, NSP8 and K63-Ub tertiary structures with SWISS-MODEL, and then the predicted 148 structures were input into ZDOCK SERVER for simulation. Predicted docking models were processed in 149

PyMOL for visualization. Surprisingly, we found that NSP8 possesses a long α-helix ( Supplementary Fig.2a) , 150 which is tightly packed in the ravines formed by the two α-helixes of MDA5 CARDs. The random coil and a 151 short α-helix in the N terminus of NSP8 occupy the area or space that interacts with K63-Ub ( Fig.3e and 152 Supplementary Fig.2b ). Further calculation of vacuum electrostatics for this binding model demonstrated that 153 the contact area in the chain of MDA5 CARDs is positively charged, while the corresponding area in the chain 154 of NSP8 is negatively charged (Fig.3f and Supplementary Fig.2c ), implying that there is a likely interaction of 155 these two structures. We further searched the polar contacts in the interface of the binding model with PyMOL 156 and found that 7 paired residues anchor with each other, one locates in the N-terminal coil of NSP8 while the 157 others are in the long α-helix (Fig.3g) .Thus, computer-based molecular structural prediction and modeling 158

implies that NSP8 interacts with the MDA5 CARD domain probably through ionic interactions and dipolar 159 surfaces between the NSP8 and MDA5 CARD binding pockets. 160 7 Next, we sought to determine how NSP8 inhibits MDA5 activation. It is well documented that upon virus 161 infection, the MDA5 CARD domain undergoes K63-linked polyubiquitination and recruits MAVS to form a 162 signalosome 20 . The structural prediction of the NSP8-MDA5 CARD interaction showed that NSP8 may 163 interrupt this process since it interacts with MDA5 at its CARD domain and shields the binding area or space 164 for K63-ubiquitin linkage (Fig.3e, Supplementary Fig.2 , and pdb file). The polyubiquitination of MDA5 was 165 analyzed in the presence or absence of NSP8 expression. The MDA5-expressing plasmid was cotransfected 166 into HEK293T cells with a WT-, K48-, or K63-linked ubiquitin-expressing plasmid, and an in vivo ubiquitination 167 assay showed that MDA5 WT-and K63-linked polyubiquitination were strongly inhibited (Fig.4a,c) , while K48-168 linked polyubiquitination was barely affected (Fig.4b) . Thus, these results reveal that NSP8 interferes with the 169 such as IL-16, IL-17A, IL-17F, and IL17C, was also downregulated. Furthermore, a decreasing transcription 178 tendency was also observed for the inflammatory receptors IL-1RI, IL-1RII, IL-2Rα, and IL18RII; NK cell-179 associated activation receptors, such as NKp44, NKp46, and NKG2B; and the trans-acting T-cell-specific 180 transcription factor GATA3 (Fig.5a) , indicating that the activation of T cells and NK cells was attenuated by 181 NSP8 through the suppression of these key factors. Although these decreased cytokines and receptors may 182 not be directly activated by IRF3 or NFκB, they could be regulated by downstream cytokines or other factors 183 derived from these two pathways. In contrast, the cytokine IL-2 and IFN-gamma suppression gene FOXP3 184 was significantly increased with NSP8 overexpression. 185

To further confirm that the downregulation of these immune and inflammatory cytokines and genes is 186 mediated by NSP8 under physiological conditions, A549 cells were transfected with an NSP8-expressing 187 8 plasmid or empty vector and then stimulated with poly(I:C) mimicking viral RNA for the indicated times. The 188

inhibition of the expression of key cytokines and related genes was verified, and NSP8 negatively regulated 189 the expression of these immune and inflammatory genes (Fig.5b) . Collectively, these results suggest that 190 NSP8 could strongly impair the expression of genes involved in antiviral immune and inflammatory responses. in MAVS KO HEK293T cells that also express NSP8 successfully restored the NSP8 inhibitory activity. 205

Consequently, a series of antiviral immune and inflammatory cytokines and related genes were further strongly 206 downregulated by NSP8 expression. 207

Herein, we speculated that NSP8 may act on RIG-I or MDA5, two upstream viral RNA sensors. Our results 208

showed that NSP8 directly interacted with MDA5 on its CARD domains, and MDA5-mediated type I IFN 209 signaling activities were strongly inhibited at the same time. Polyubiquitinated modification of MDA5 is crucial 210 for its antiviral responses, K48-linked polyubiquitination mediates MDA5 proteasomal degradation 21 , and K63-211 linked polyubiquitination mediates MDA5-induced type I IFN expression 20 . We speculated that NSP8 may 212 inhibit type I IFN signaling through the polyubiquitinated modification of MDA5. To test our hypothesis, we 213 determined the status of WT-, K48-and K63-linked polyubiquitination of MDA5 in the absence or presence of 214 9 NSP8 and confirmed that WT-and K63-linked polyubiquitination were impaired by NSP8, while K48-linked 215 polyubiquitination was barely changed. We therefore conclude that under certain circumstances, NSP8 216 jeopardizes antiviral responses by impairing MDA5 K63-linked polyubiquitination. 217

Based on our existing experimental data, we propose a simple working model to illustrate how NSP8 218 negatively regulates innate immune responses by inhibiting MDA5 K63-linked polyubiquitination (Fig.5c) . to the attenuation of RIG-I-mediated type I IFN antiviral responses 24 . In addition, the NS3 protein of ZIKA virus 241 10 interacts with scaffold proteins 14-3-3ϵ and η separately through its 14-3-3 binding motif, hence blocking the 242 translocation of RIG-I and MDA5 from the cytosol to mitochondria, impairing signalosome formation with MAVS, 243 and antagonizing innate immunity 25 . Our studies revealed that NSP8 of SARS-CoV-2 acts as a binding partner 244 of MDA5 to shield its K63-linked polyubiquitination and then impairs the formation or activation of the MDA5 245

In summary, our study provides insights into the potential mechanisms of SARS-CoV-2 NSP8 in the 247 inhibition of type I IFN signaling and antiviral responses. We provide compelling evidence that NSP8 plays a 248 critical negative role in MDA5-mediated antiviral responses and demonstrate specific orchestration of the viral 249 dsRNA-triggered signalosome and signal cascade by NSP8. Importantly, considering that MDA5 plays a key 250 pathological role in antiviral immunity towards severe coronaviruses, antagonists of NSP8 could serve as a 251 promising therapeutic target for COVID-19 therapies. Immunoprecipitation and immunoblot analysis 308

For immunoprecipitation, whole cell extracts were prepared after transfection or stimulation with appropriate 309 ligands, followed by incubation for 1 h at 4°C with anti-GFP agarose beads (AlpaLife). Beads were washed 4 310 times with low-salt lysis buffer, and immunoprecipitants were eluted with 2x SDS loading buffer and then 311 resolved by SDS-PAGE. Proteins were transferred to PVDF membranes (Millipore) and further incubated with 312 the appropriate primary and secondary antibodies. The images were visualized using Odyssey Sa (LI-COR). 313 314 Computer-based prediction and structural modeling 315 NSP8.pdb, MDA5-CARDs.pdb and K63-Ub.pdb were generated in SWISS-MODEL 27 . MDA5-CARDs.pdb 316 was input into ZDOCK-SERVER 28 as a receptor, while NSP8 or K63-Ub was input as a ligand for docking 317 computation. MDA5-CARDs with NSP8.pdb and MDA5-CARDs with K63-Ub.pdb were the best fit prediction 318 models chosen from the results. All the pdb files were processed and visualized with PyMOL (Schrödinger. 319 320

Middle East Respiratory Syndrome: Emergence of a Pathogenic 519

Annual review of medicine 68

Dysregulated Type I Interferon and Inflammatory Monocyte-Macrophage 522

Responses Cause Lethal Pneumonia in SARS-CoV-Infected Mice

Emerging coronaviruses: Genome structure, replication, and pathogenesis

Transcriptomic characteristics of bronchoalveolar lavage fluid and peripheral blood 527 mononuclear cells in COVID-19 patients. Emerging microbes & infections 9

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia 532

China: a descriptive study

Recognition of pathogen-associated molecular patterns by TLR family

Type I interferons (alpha/beta) in immunity and 537 autoimmunity

SARS coronavirus nsp1 protein induces template-dependent endonucleolytic cleavage of 540 mRNAs: viral mRNAs are resistant to nsp1-induced RNA cleavage

Severe acute respiratory syndrome 543 coronavirus nsp1 facilitates efficient propagation in cells through a specific translational shutoff of host

Severe acute respiratory syndrome 546 coronavirus papain-like protease ubiquitin-like domain and catalytic domain regulate antagonism of IRF3 547 and NF-kappaB signaling

MAVS-mediated apoptosis and its inhibition by viral proteins

Ribose 2'-O-methylation provides a molecular signature for the distinction of self and non-self 551 mRNA dependent on the RNA sensor Mda5

Biochemical and structural insights into the mechanisms of SARS coronavirus RNA ribose 2

O-methylation by nsp16/nsp10 protein complex

Crystal structure and functional analysis of the SARS-coronavirus RNA cap 2'-O-556

Phage display technique identifies the interaction of severe acute respiratory syndrome 559 coronavirus open reading frame 6 protein with nuclear pore complex interacting protein NPIPB3 in 560 modulating Type I interferon antagonism

SARS-coronavirus open reading frame-9b suppresses innate immunity by targeting 563 mitochondria and the MAVS/TRAF3/TRAF6 signalosome

MERS-CoV 4b protein interferes with the NF-κB-dependent innate immune response during 566 infection

Structural basis for translational shutdown and immune evasion by the Nsp1 protein of 568 SARS-CoV-2

Ubiquitin-induced oligomerization of the RNA sensors RIG-I and MDA5 activates antiviral 570 innate immune response

The E3 Ubiquitin Ligase TRIM40 Attenuates Antiviral Immune Responses by Targeting MDA5 572 and RIG-I

Pangolins Lack IFIH1/MDA5, a Cytoplasmic RNA Sensor That Initiates 574

Innate Immune Defense Upon Coronavirus Infection

Regulation of Retinoic Acid Inducible Gene-I (RIG-I) Activation by the Histone Deacetylase 6

Herpesvirus deconjugases inhibit the IFN response by promoting TRIM25 autoubiquitination 579 and functional inactivation of the RIG-I signalosome

Zika Virus NS3 Mimics a Cellular 14-3-3-Binding Motif to Antagonize RIG-I-and MDA5-582

NLRP4 negatively regulates type I interferon signaling by targeting the kinase TBK1 for 584 degradation via the ubiquitin ligase DTX4

SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary 586 information

ZDOCK server: interactive docking prediction of protein-protein complexes and symmetric 588

Author contributions 322 13 X.L. and E.K. initiated the concept. Z.Y., X.L. and E.K. designed the experiments and analyzed the data. P.W. 323 provided the reagents. Z.Y., X.Z., F.W. performed the experiments. Z.Y. and E.K. wrote the paper. 324 325 326

We thank all the members of our laboratory for their critical assistance and helpful discussions. This work is 328 supported by grants from the Natural Science Foundation of China (81671996 and 81871643) to E.K. and the 329 Natural Science Foundation of China (81971928) to X.Li. 330 331 332

The authors declare no competing financial interest. Twenty-four hours post transfection, cells were treated with poly(I:C) (5 μg/ml) for the indicated time points, 361 and then total RNA was extracted and subjected to RT-PCR analysis for TNF-α, IFN-β, IFIT1, IFIT2, IL-6 and 362 CCL20 expression. The data are shown as the mean values ± SD (n = 3). *, p < 0.0332; **, p < 0.0021; ***, p 363 < 0.0002; ****, p < 0.0001; by Sidak's multiple comparisons test. were collected, and total RNA was extracted and subjected to RT-PCR analysis for the expression of the 432 indicated genes. "▼" indicates genes that were significantly changed in two independent NSP8-expressing 433 samples. 434 b. A549 cells were transfected with an empty vector or Flag-NSP8. Twenty-four hours post transfection, cells 435were treated with poly(I:C) (5 μg/ml) for the indicated time points, and then total RNA was extracted and 436 subjected to RT-PCR analysis of the expression of selected genes. 437The data are shown as the mean values ± SD (n = 3). *, p < 0.0332; **, p < 0.0021; ***, p < 0.0002; ****, p < 438 The sequences of primer pairs used in the RT-PCR array (Fig.5a)