key: cord-0841744-9sfbra47
authors: Dolliver, Stacia M.; Kleer, Mariel; Bui-Marinos, Maxwell P.; Ying, Shan; Corcoran, Jennifer A.; Khaperskyy, Denys A.
title: Nsp1 proteins of human coronaviruses HCoV-OC43 and SARS-CoV2 inhibit stress granule formation
date: 2022-05-02
journal: bioRxiv
DOI: 10.1101/2022.05.02.490272
sha: c0ff9f91a2221a8bd7d574255b0cef3d20673992
doc_id: 841744
cord_uid: 9sfbra47

Stress granules (SGs) are cytoplasmic condensates that often form as part of the cellular antiviral response. Despite the growing interest in understanding the interplay between SGs and other biological condensates and viral replication, the role of SG formation during coronavirus infection remains poorly understood. Several proteins from different coronaviruses have been shown to suppress SG formation upon overexpression, but there are only a handful of studies analyzing SG formation in coronavirus- infected cells. To better understand SG inhibition by coronaviruses, we analyzed SG formation during infection with the human common cold coronavirus OC43 (HCoV-OC43) and the highly pathogenic SARS-CoV2. We did not observe SG induction in infected cells and both viruses inhibited eukaryotic translation initiation factor 2α (eIF2α) phosphorylation and SG formation induced by exogenous stress (e.g. sodium arsenite treatment). Furthermore, in SARS-CoV2 infected cells we observed a sharp decrease in the levels of SG-nucleating protein G3BP1. Ectopic overexpression of nucleocapsid (N) and non-structural protein 1 (Nsp1) from both HCoV-OC43 and SARS-CoV-2 inhibited SG formation. The Nsp1 proteins of both viruses inhibited arsenite-induced eIF2α phosphorylation, and the Nsp1 of SARS- CoV2 alone was sufficient to cause decrease in G3BP1 levels. This phenotype was dependent on the depletion of cytoplasmic mRNA mediated by Nsp1 and associated with nuclear retention of the SG- nucleating protein TIAR. To test the role of G3BP1 in coronavirus replication, we infected cells overexpressing EGFP-tagged G3BP1 with HCoV-OC43 and observed a significant decrease in infection compared to control cells expressing EGFP. The antiviral role of G3BP1 and the existence of multiple SG suppression mechanisms that are conserved between HCoV-OC43 and SARS-CoV2 suggest that SG formation may represent an important antiviral host defense that coronaviruses target to ensure efficient replication. Author Summary Host cells possess many mechanisms that can detect viral infections and trigger defense programs to suppress viral replication and spread. One of such antiviral mechanisms is the formation of stress granules – large aggregates of RNA and proteins that sequester viral components and cellular factors needed by the virus to replicate. Because of this threat, viruses evolved specific mechanisms that prevent stress granule formation. Understanding these mechanisms can reveal potential targets for therapies that would disable viral inhibition of stress granules and render cells resistant to infection. In this study we analyzed inhibition of stress granules by two human coronaviruses: the common cold coronavirus OC43 and the pandemic SARS-CoV2. We have demonstrated that these viruses employ at least two proteins – nucleocapsid protein (N) and the non-structural protein 1 (Nsp1) to suppress stress granules. These proteins act through distinct complementary mechanisms to ensure successful virus replication. Because both OC43 and SARS-CoV2 each dedicate more than one gene product to inhibit stress granule formation, our work suggests that viral disarming of stress granule responses is central for a productive infection.

is also readily infected with many RNA viruses, including OC43, which rapidly replicates in 293A cells 

As expected, in mock-infected cells, SGs were induced in nearly 100% of cells. By contrast, less than half 170 of the OC43-infected cells formed SGs following As treatment (Fig.1B,D ). Next, we tested levels of 171 eIF2α phosphorylation in infected cells and discovered that OC43 infection inhibited As-induced eIF2α 172 phosphorylation, with increasing efficiency from 12 to 48 hpi (Fig.1C) Fig. 2A,C) . As expected, Silvestrol treatment did not induce eIF2α phosphorylation (Fig.2B ).

Interestingly, in many infected cells we noticed brighter nuclear TIAR staining compared to uninfected 212 cells, possibly indicating disruption of normal nucleocytoplasmic shuttling of TIAR (Fig. 1A ,B, Fig. 2A ).

To confirm that OC43 effects were not limited to TIAR-containing SGs, we completed a series of 214 experiments using multiple SG marker proteins to analyze SG formation in infected cells treated with 215 either As or Silvestrol. These analyses confirmed that regardless of the markers used, including SG-11 216 nucleating proteins G3BP1, G3BP2, and TIA-1, as well as translation initiation factors eIF4G and eIF3B, 217 formation of SG foci was inhibited by OC43 (Fig.2D ). In addition, these analyses revealed that in a 218 fraction of infected cells that did form SGs, OC43 N protein was not accumulating in these foci (Fig. 2D ). 3A ). In addition, CoV2 was able to effectively suppress SG formation 242 induced by As (Fig. 3B ,C). Compared to OC43, CoV2 infection resulted in much greater SG inhibition,

with only about 8% of infected cells forming SGs following As treatment (Fig. 3C ). Similar to OC43,

CoV2 inhibited As-induced eIF2α phosphorylation (Fig. 3D) . Interestingly, we also observed an increase 245 in nuclear TIAR signal in CoV2 infected cells, which was even more prominent than that was observed in 246 OC43 infected cells (Fig.3A,B ). When we compared total levels of TIAR in mock and CoV2 infected 247 cells using western blot, we observed similar levels, indicating that the virus causes changes in subcellular 248 distribution of TIAR without drastically affecting its expression (Fig. 3D,E) . By contrast, we consistently 249 observed substantial reduction in the levels of G3BP1 protein in CoV2 infected cells compared to mock 250 (Fig. 3D,E) . Unlike OC43 infection, which resulted in detectable decrease in G3BP1 levels only at 48 hpi 251 (Fig. 1C) , in CoV2 infected cells G3BP1 decrease was observed much earlier ( Fig. 3D ,E). To examine if 252 these changes in G3BP1 expression were due to a decrease in its transcript levels, we isolated total RNA 253 from mock and CoV2-infected cells at 24 hpi and analysed G3BP1 and TIAR mRNA levels using RT-13 254 QPCR. Consistent with a known feature of CoV2 host shutoff causing cytoplasmic mRNA degradation,

we detected dramatic depletion of both G3BP1 and TIAR transcripts in infected cells (Fig. 3F ). The 256 magnitude of mRNA depletion was slightly higher for G3BP1 (on average 80% depleted for G3BP1 vs.

257 60% for TIAR), but alone it would not account for the observed differences in protein levels in infected 258 cells. To examine relative stability of G3BP1 and TIAR proteins in 293A cells, we treated uninfected 259 cells with two translation inhibitors, cycloheximide (CHX) and Silvestrol (Sil.), as well as the 260 transcription inhibitor Actinomycin D (ActD) for 12 hours and measured protein levels using western 261 blot. As expected, within 1 hour of treatment, CHX and Sil. potently blocked protein synthesis in 293A 262 cells, while ActD had no effect (Fig. 3G ). However, after 12 h only ActD treatment resulted in a small 263 (~25%) but statistically significant decrease in G3BP1 and, to a lesser extent, TIAR protein levels, while 264 translation inhibitors did not decrease either of these proteins ( Fig. 3H -J). This suggests that neither 265 G3BP1 nor TIAR are intrinsically unstable and rapidly degraded proteins. Instead, it points to a distinct 266 turnover mechanism for these RNA-binding proteins that is activated in response to either decrease in 267 transcription or total mRNA levels. As a control, we examined total levels of the translation initiation 268 factor eIF4G and detected no changes in its expression after ActD treatment (Fig. 3K ). These results

suggest that the dramatic decrease in G3BP1 levels in CoV2 infected cells is primarily due to mRNA 270 depletion rather than translation arrest, which are both features of CoV2 host shutoff. However, a direct 271 targeting of G3BP1 for degradation by a viral protein cannot be ruled out. In either case, given the central proteins were expressed at similar levels, and that neither of the N constructs affected G3BP1 expression 305 levels or As-induced eIF2α phosphorylation (Fig. 4A ). As expected, immunofluorescence microscopy 306 analyses showed that nearly 100% of the EGFP expressing cells formed SGs upon As treatment, while 307 CoV2 N expression efficiently inhibited SG formation (Fig. 4B,C) . In OC43 N expressing cells, SG 308 formation was also inhibited, but to a lesser extent (Fig. 4C) . Given that N proteins did not affect eIF2α

309 phosphorylation, we next tested if OC43 and/or CoV2 EGFP-N constructs could inhibit Silvestrol-16 310 induced SG formation. Indeed, both N constructs were able to inhibit Silvestrol-induced SG formation 311 (Fig 4D,E) , and the OC43 N protein was even better at inhibiting Silvestrol-induced SGs than the As-

induced SGs (Fig. 4C 4F) . Furthermore, OC43 Nsp15 did not significantly affect As-induced SG 319 formation (Fig. 4G,H) . Thus, we conclude that Nsp15 is not contributing to inhibition of As-induced SGs 320 by OC43 in our experimental system. (Fig. 5B ). In addition, western blot analysis revealed 358 that both OC43 and CoV2 Nsp1 significantly attenuated As-induced eIF2α phosphorylation (Fig. 5B,C) . As. In contrast, in CoV2 EGFP-Nsp1 expressing cells, G3BP1 signal was much lower than in bystander 367 untransfected cells, EGFP expressing control cells, or OC43 EGFP-Nsp1 cells (Fig. 5D ). Both TIAR and 368 G3BP1 are important SG nucleating proteins, therefore the CoV2 Nsp1 mediated redistribution of TIAR 369 into the nucleus and depletion of G3BP1 levels could potentially disrupt cytoplasmic SG condensation.

Alternatively, using TIAR or G3BP1 as SG markers in these cells could compromise visualization of SGs 371 that may lack these proteins. To distinguish between these possibilities, we stained for G3BP2, another 372 well established SG marker. Unlike with G3BP1 staining, we did not see significant decrease in G3BP2 373 signal in either OC43 or CoV2 Nsp1 expressing cells, instead we observed that some CoV2 Nsp1 374 expressing cells were forming small G3BP2-positive SGs upon As treatment (Fig. 5E) . We analysed the 375 number and size of As-induced G3BP2-positive SGs that form in CoV2 Nsp1 expressing cells and 376 compared them to SGs formed in control EGFP expressing cells. This analysis revealed that SGs that do 377 form in many CoV2 Nsp1-expressing cells are significantly smaller than SGs that form in control cells,

while their average number remains the same (Fig. 5F ). We observed depletion of G3BP1 protein and mRNA in CoV2 infected cells. Next, we tested if G3BP1 380 depletion by CoV2 Nsp1 but not by OC43 Nsp1 is linked to mRNA degradation induced by the former.

We generated two CoV2 Nsp1 amino acid substitution mutants that are defective for mRNA degradation (Fig. 5H ). In addition, only the wild type CoV2 Nsp1 overexpression caused depletion of 393 G3BP1 protein (Fig. 5H, lane 3) . This alteration in G3BP1 levels was not observed in OC43 Nsp1- type CoV2 Nsp1 caused decrease in G3BP1, G3BP2, and β-actin (ACTB) mRNAs (Fig. 5J ). This 399 indicated that unlike CoV2 Nsp1, the OC43 Nsp1 does not cause mRNA degradation, and that the amino 400 acid substitution mutants we generated behave as expected. Interestingly, none of the Nsp1 constructs 401 decreased TIAR transcript levels, with the wild type CoV2 Nsp1 causing modest but statistically 402 significant increase in TIAR mRNA compared to EGFP control (Fig. 5J ).

To measure and compare SG inhibition by our panel of Nsp1 constructs, we quantified As-induced SG 404 formation using different markers. We saw that when SGs were stained using TIAR as a marker, like in 405 Fig. 5A and 5G, the wild type CoV2 Nsp1 appeared significantly better at preventing SG formation 406 compared to 99A or 125A mutants that did not cause nuclear accumulation of TIAR (Fig. 5I) . By 407 contrast, when we quantified G3BP2-positive SGs, wild type and mutant Nsp1 constructs inhibited SG 408 formation to the same degree (Fig. 5F ). While all Nsp1 constructs inhibited SG formation visualized 409 using G3BP2 as a marker, more than 50% of cells still formed SGs. Therefore, it is apparent that in many 410 wild type CoV2 Nsp1-expressing cells SGs still form, but they contain very little TIAR and G3BP1.

To examine if inhibition of eIF2α phosphorylation is the main mechanism of SG suppression by OC43 

Error bars = standard deviation. Two-tailed Students t-Test was done to determine statistical significance. 

G3BP1 is one of the most important SG nucleating proteins. Apart from its function in nucleating SG

formation it is also involved in antiviral signaling. Since our work revealed that coronaviruses are (Fig. 6B,C) . Importantly, this was not due to lentiviral integration or non-485 specific effect of EGFP overexpression as infection rates were the same between 293A[EGFP] and 486 parental untransduced cells (Fig. 6B,C) . Western blot analysis confirmed that the EGFP-G3BP1 fusion 487 protein was expressed at higher levels than the endogenous G3BP1 (Fig. 6D) . The ectopic overexpression 488 of EGFP-G3BP1 but not the EGFP control caused noticeable decrease in endogenous G3BP1 and G3BP2 the tuberous sclerosis complex (TSC) to lysosomes and suppressing activation of the mechanistic target 585 of rapamycin complex 1 (mTORC1) (72). These functions of G3BP1 are independent from SG formation.

Thus, in addition to interfering with SG nucleation, depletion of G3BP1 in CoV2 infected cells could 587 benefit viral replication by both blunting the cellular innate immune responses and by upregulating 588 biosynthetic pathways through relieving mTORC1 suppression. In our study we showed that 589 overexpression of G3BP1 inhibits OC43 infection without increasing SG formation, suggesting that 590 G3BP1 is antiviral towards OC43 and that some of the SG nucleation-independent functions of G3BP1 591 could be responsible.

Existence of multiple mechanisms that interfere with translation arrest and SG formation in cells infected 593 with both the seasonal common cold OC43 and the pandemic CoV2 viruses described in this study 594 highlights the importance of overcoming these antiviral mechanisms by diverse coronaviruses. In 595 addition, our work discovers a novel feature of Nsp1-mediated host shutoff that simultaneously blocks 596 host translation initiation and promotes continuous regeneration of GTP-bound translation initiation-597 competent eIF2 by inhibiting eIF2α phosphorylation. In the follow up work focusing on CoV2 and OC43 598 Nsp1, we aim at characterizing the mechanism by which these host shutoff factors inhibit eIF2α 599 phosphorylation and the contribution of this function to viral replication fitness and suppression of host 600 antiviral responses. 

Whole-cell lysates were prepared by direct lysis of PBS-washed cell monolayers with 1× Laemmli 702 sample buffer (50 mM Tris-HCl pH 6.8, 10% glycerol, 2% SDS, 100 mM DTT

704 stored at −20°C. Aliquots of lysates thawed on ice were incubated at 95°C for 3 min, cooled on ice, 705 separated using denaturing PAGE, transferred onto PVDF membranes using Trans Blot Turbo Transfer 706 System with RTA Transfer Packs

CoV2 N (1:1,000; rabbit, Novus Biologicals, NBP3-05730); eIF2α (1:1000; rabbit

#5324); eIF4G (1:1000; rabbit

G3BP1 (1:4000; mouse

G3BP2 (1:2500; rabbit

GFP (1:1000; rabbit

HA tag (1:1,000; mouse

MAB9012); phospho-S51-eIF2α (1:1000; rabbit

TIAR (1:1000; rabbit

For band visualization, HRP-conjugated anti-rabbit IgG (Goat, Cell Signaling, #7074) 715 or anti-mouse IgG (Horse, Cell Signaling, #7076) were used with Clarity Western ECL Substrate on the 716

Where indicated, total protein was visualised 717 post-transfer to PVDF membranes on ChemiDoc using Stain-free fluorescent dye

For analyses of protein band intensities, western blot signals were quantified using Bio-Rad Image Lab 719 5.2.1 software and values normalized to the Stain-free signal for each lane

The puromycin incorporation assay was performed as described in (77) with the following modifications

Puromycin was added to the medium at the final concentration of 10 μg/ml for 10 min

The molecular virology of 759 coronaviruses

Coronavirus, the King Who Wanted More Than a Crown: From Common to the Highly 761

Is the Key in the Accessory Genes? Front Microbiol

Imbalanced host 765 response to SARS-CoV-2 drives development of

A Novel Coronavirus from Patients with 767 Pneumonia in China

Human Coronavirus: Host-Pathogen Interaction

The Coronavirus Replicase

Mechanisms and enzymes 775 involved in SARS coronavirus genome expression

Coronavirus genome structure and replication

Virus-encoded proteinases and proteolytic processing in the 779 Nidovirales

Structure, Function, and Evolution of Coronavirus Spike Proteins

Structural and functional analysis of the surface protein of human coronavirus 783 OC43

Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease 786 Inhibitor

Angiotensin-converting enzyme 2 788 is a functional receptor for the SARS coronavirus

All hands on deck: SARS-CoV-2 proteins that block early anti-viral 790 interferon responses

Translation inhibition and stress granules in the antiviral immune 792 response

Bronchitis Coronavirus Limits Interferon Production by Inducing a Host Shutoff That Requires 795 Accessory Protein 5b

Inhibition of anti-viral stress granule 797 formation by coronavirus endoribonuclease nsp15 ensures efficient virus replication. Thiel V, 798 editor

Pattern Recognition and Signaling Mechanisms of RIG-I and 800 MDA5

Type I interferons in host defense

The dsRNA protein kinase PKR: virus and cell control

The mechanism of eukaryotic translation initiation and 805 principles of its regulation

Evidence That Ternary Complex (eIF2-807 GTP-tRNAiMet)-Deficient Preinitiation Complexes Are Core Constituents of Mammalian Stress 808

Stress granules and cell signaling: more than just a passing 810 phase?

Principles and Properties of Stress Granules

RNA-binding proteins TIA-1 and TIAR link 814 the phosphorylation of eIF-2 alpha to the assembly of mammalian stress granules

G3BP-Caprin1-USP10 817 complexes mediate stress granule condensation and associate with 40S subunits

Critical role of an antiviral 820 stress granule containing RIG-I and PKR in viral detection and innate immunity

DHX36 enhances RIG-I 823 signaling by facilitating PKR-mediated antiviral stress granule formation

The Stress Granule Protein G3BP1 Recruits Protein Kinase R To Promote 826

Multiple Innate Immune Antiviral Responses. Sandri-Goldin RM, editor

Influenza A virus inhibits cytoplasmic stress granule 829 formation

Virus Host Shutoff Disables Antiviral Stress-Induced Translation Arrest. Pekosz A, editor. PLoS 832 Pathog

Who 834 Regulates Whom? An Overview of RNA Granules and Viral Infections

SARS-CoV-2 838 nucleocapsid protein binds host mRNAs and attenuates stress granules to impair host stress 839 response. iScience

SARS-CoV-2 NSP5 and N protein 842 counteract the RIG-I signaling pathway by suppressing the formation of stress granules. Signal 843 Transduct Target Ther

SARS-CoV-2 nucleocapsid protein impairs stress 845 granule formation to promote viral replication

Characterization of 847 SARS2 Nsp15 nuclease activity reveals it's mad about U

Coronavirus nonstructural protein 1: 850 Common and distinct functions in the regulation of host and viral gene expression

Severe acute respiratory syndrome 853 coronavirus protein nsp1 is a novel eukaryotic translation inhibitor that represses multiple steps of 854 translation initiation

Structural 856 basis for translational shutdown and immune evasion by the Nsp1 protein of SARS-CoV-2

SARS coronavirus 859 nsp1 protein induces template-dependent endonucleolytic cleavage of mRNAs: viral mRNAs are 860 resistant to nsp1-induced RNA cleavage

Severe acute respiratory syndrome 862 coronavirus nsp1 facilitates efficient propagation in cells through a specific translational shutoff of 863 host mRNA

The N-terminal 865 domain of SARS-CoV-2 nsp1 plays key roles in suppression of cellular gene expression and 866 preservation of viral gene expression

Proximity-dependent biotinylation detects associations between SARS coronavirus nonstructural 869 protein 1 and stress granule-associated proteins

Thiopurines inhibit coronavirus Spike protein processing and incorporation into progeny virions 872 [Internet]. bioRxiv; 2022

Evaluation of four cell lines for assay of infectious adenoviruses 875 in water samples

Group V and X secretory phospholipase A2 prevents 877 adenoviral infection in mammalian cells

The Scattered Twelve Tribes of HEK293

Ribosome Recruitment Induces Stress Granule Formation Independently of Eukaryotic Initiation 882 Factor 2α Phosphorylation

Eukaryotic Translation Initiation Factor 4A Inhibitors Block Influenza A Virus Replication. 885 Viruses

Harnessing stress granule formation by small molecules to 887 inhibit the cellular replication of SARS-CoV-2

Proteomic 889 analysis identifies the RNA helicase DDX3X as a host target against SARS-CoV-2 infection

Comparative host-892 coronavirus protein interaction networks reveal pan-viral disease mechanisms

Large scale discovery of 895 coronavirus-host factor protein interaction motifs reveals SARS-CoV-2 specific mechanisms and 896 vulnerabilities

Nsp1 binds the ribosomal mRNA channel to inhibit translation

The RasGAP-901 associated endoribonuclease G3BP assembles stress granules

UV damage induces G3BP1-dependent stress granule formation that is not 904 driven by translation arrest via mTOR inhibition

Coronavirus in human diseases: Mechanisms and advances 906 in clinical treatment

Direct Observation of Repeated Infections With Endemic Coronaviruses. J 908 Infect Dis

Pharmacologic profiling reveals lapatinib as a novel antiviral against SARS-CoV-2 in vitro

Activation of STING Signaling 913

Pathway Effectively Blocks Human Coronavirus Infection

High-Throughput Screening and 915 Identification of Potent Broad-Spectrum Inhibitors of Coronaviruses

Photodynamic 918 Inactivation of Human Coronaviruses. Viruses

Improving human coronavirus OC43 (HCoV-OC43) research 920 comparability in studies using HCoV-OC43 as a surrogate for SARS-CoV-2

Multi-color super-resolution imaging to 923 study human coronavirus RNA during cellular infection

Kinase-mediated Phosphorylation of Eukaryotic Translation Initiation Factor 2 Inhibits Translation

Induces Stress Granule Formation, and Mediates Survival upon Arsenite Exposure

UBAP2L Forms Distinct 930 Cores that Act in Nucleating Stress Granules Upstream of G3BP1

Arginine Demethylation of 933 G3BP1 Promotes Stress Granule Assembly

The stress granule protein G3BP1 promotes pre-condensation of cGAS to allow rapid 935 responses to DNA

The stress granule protein G3BP1 binds viral dsRNA and RIG-I 937 to enhance interferon-β response

TDRD3 is an antiviral restriction factor that promotes IFN 939 signaling with G3BP1

RNase L Amplifies Interferon Signaling by Inducing 941

G3BPs tether the TSC complex to lysosomes and suppress mTORC1 signaling

Replication Kinetics of Severe Acute Respiratory Syndrome Coronavirus 2. Emerg 948 Infect Dis

Human 950 coronaviruses disassemble processing bodies

Selective Degradation 954 of Host RNA Polymerase II Transcripts by Influenza A Virus PA-X Host Shutoff Protein. Palese P, 955 editor

KSHV activates unfolded protein response sensors but 957 suppresses downstream transcriptional responses to support lytic replication

Methods for the characterization of stress granules in virus 960 infected cells

We would like to thank Dr. Arinjay Banerjee (Vaccine and Infectious Disease Organization (VIDO), Dr.

Karen Mossman (University of McMaster) and Dr. Samira Mubareka (University of Toronto) for the