SARS-CoV-2 nucleocapsid protein inhibits the PKR-mediated integrated stress response through RNA-binding domain N2b

Abstract

The nucleocapsid protein N of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) enwraps and condenses the viral genome for packaging but is also an antagonist of the innate antiviral defense. It suppresses the integrated stress response (ISR), purportedly by interacting with stress granule (SG) assembly factors G3BP1 and 2, and inhibits type I interferon responses. To elucidate its mode of action, we systematically deleted and over-expressed distinct regions and domains. We show that N via domain N2b blocks PKR-mediated ISR activation, as measured by suppression of ISR-induced translational arrest and SG formation. N2b mutations that prevent dsRNA binding abrogate these activities also when introduced in the intact N protein. Substitutions reported to block post-translation modifications of N or its interaction with G3BP1/2 did not have a detectable additive effect. In an encephalomyocarditis virus-based infection model, N2b - but not a derivative defective in RNA binding-prevented PKR activation, inhibited β-interferon expression and promoted virus replication. Apparently, SARS-CoV-2 N inhibits innate immunity by sequestering dsRNA to prevent activation of PKR and RIG-I-like receptors. Similar observations were made for the N protein of human coronavirus 229E, suggesting that this may be a general trait conserved among members of other orthocoronavirus (sub)genera.

Fig 1. SARS-CoV-2 N inhibits PKR-induced ISR, preventing translation arrest and SG formation in HeLa and A549 cells.

(A) Schematic representation of the PKR-induced ISR activation upon the pEGFP-N3 transfection in eukaryotic cells. Created with BioRender.com. (B) HeLa and A549 cells were transfected with expression vector pEGFP-N3 (EGFP) or pEGFP-N3 derivatives encoding SARS-CoV-2 N, MERS-4a, IAV-NS1 or BwCoV-AcP10 genetically fused to EGFP. Induction of the ISR via dsRNA-mediated PKR activation or suppression thereof was assessed by comparing EGFP fluorescence intensity and SG formation as detected by immunofluorescence staining for G3BP2 (see also S1A Fig). (left panel) Transfected HeLa cells; (immune)fluorescence microscopy images, representative results. Scale bars: 50 μm. Settings of image acquisition (laser intensity, exposure time) and processing conditions used throughout were chosen to avoid over-exposure in high-expressing cells. Note that under these conditions, EGFP expression is too low to be detected in a sizeable population of stressed and non-stressed cells (see also S1B Fig). (right panel) Bar graphs for Hela and A549 cells showing the percentage of EGFP expressing cells containing G3BP2-positive SGs. The results are representative of three independent experiments counting >200 cells per sample. Standard deviation indicated by error bars; *** p = 0.001, **** p< 0.001, ns = not significant (one-way ANOVA with Dunnett post hoc test). (C) HeLa PKR^KO cells transfected as in B. Representative (immuno)fluorescence microscopy images are shown. (D) Western-blot analysis for PKR, phosphorylated PKR (p-PKR), EGFP-fused proteins and β-actin. HeLa wt and PKR^KO cells, mock-transfected or transfected with indicated plasmid were lysed at 24 hr. Of note, the larger yield of EGFP versus N-EGFP in either cell type is counter intuitive but can be explained from the fact that (i) this is an ensemble measurement (for the total transfected cell population) versus the analysis of individual cells by fluorescence microscopy and (ii) basal expression levels of the codon optimized EGFP prior to ISR activation will exceed those the N-EGFP fusion protein, which is non-codon optimized and three times larger in size (see also S4 Fig). (E) Inhibition of SG formation and translational arrest by pEGFP-N3-expressed N proteins of MERS-CoV and HCoV-229E. (Immuno)fluorescence analysis (left panel), quantitative representation of the results and statistical analysis as in B. (F) Western-blot analysis for eIF2α, phosphorylated eIF2α (p-eIF2α), EGFP-fused proteins and β-actin. HeLa cells were transfected to express EGFP or EGFP-tagged N proteins of SARS-CoV-2 and HCoV 229E. Mock-transfected cells were either left untreated (mock) or treated with sodium arsenite (+Arsenite) to induce eIF2α phosphorylation.

Fig 2. Nucleocapsid domain N2b suppresses translational arrest and SG formation.

(A) Schematic representation of truncated derivatives of SARS-CoV-2 N protein fused to EGFP (left panel). The proteins were overexpressed from pEGFP-N3-based vectors in HeLa cells and percentages of SG-positive cells were determined as in Fig 1A (right panel). The data are representative of three independent experiments with more than 200 cells counted for the presence of G3BP2-positive SGs per individual sample. Standard deviation indicated by error bars. For a statistical analysis of the results, see S1 Table. (B) Representative results from (immuno)fluorescence analysis of HeLa cells transfected to express the SARS-CoV-2 N protein derivatives. Scale bar: 50 μm.

Fig 3. N2b mutations that disrupt dsRNA binding abrogate suppression of translational arrest and SG formation.

(A) (left panel) Surface representation and (middle panel) cartoon representation of the SARS-CoV-2 N2b dimer (PDB ID code 7C22 [59] with monomers in light blue (chain A) and wheat (chain B). Side chains of mutated charged/polar residues labeled and shown in sticks (Chain A: Lys²⁵⁷ and Lys²⁶¹; Chain B: Q²⁷², Q²⁸⁹, R²⁷⁶ and R²⁹³). (right panel) N2b dimer surface representation colored according to calculated charge distributions, displaying the positively-charged central RNA-binding groove. Critical residues Lys²⁵⁷ (single asterisk) and Lys ²⁶¹ (double asterisk) marked for both monomers. Top view images at a forward 45° tilt; generated with UCSF ChimeraX version 1.6.1 [79]. (B) Electrophoretic mobility shift assays (EMSA). (left panel) EMSA with bacterially expressed N2b domain and single (ss) and double-stranded (ds) RNA oligonucleotides, designed after SARS-CoV-2 stem-loop II motif (s2m). Assays were performed with N2b in 10- or 20-fold molar excess. Non-bound RNA was included as a control (np, no protein). (middle panel) The effect of N2b amino acid substitutions on binding of s2m dsRNA or (right panel) a scrambled version thereof (right). EMSAs performed with N2b and mutant derivatives in 20-fold molar excess (middle and right panels). (C-D) Mutational analysis of SARS-CoV-2 N2b, full-length SARS-CoV-2 N and full-length HCoV-229E N. Select surface-exposed charged residues were substituted by Ala either individually or in combination as indicated and the effect on IRS-induced translational arrest (C) and SG formation (D) was analyzed in HeLa cells as in Fig 1. For a statistical analysis of the results, see S2 Table.

Fig 4. N-mediated suppression of SG formation and ISR-induced translational arrest is not affected by posttranslational modifications or G3BP1 interaction.

(A) The effect on N-mediated ISR inhibition by mutations introduced to prevent posttranslational modifications (K³⁷⁵Q, K³⁷⁵R, R⁹⁵K), to disrupt the G3BP1/2 binding site (1A-4A mutants) or to prevent phosphorylation of the SR element (13S>A) was tested in HeLa cells by (immune)fluorescence analysis as in Fig 1. Scale bar: 50 μm. (B). Quantification of the results in (A) representative of three independent experiments, with >200 cells counted per sample as in Fig 1. (C) Pull down assay to test SARS-CoV-2 N mutants for their association with endogenous G3BP1 in HeLa wt (left panel) and PKR^KO cells (right panel).

Fig 5. The effect of N-G3BP interaction on arsenite-induced SGs in HeLa PKR^KO cells.

HeLa PKR^KO cells were transfected to express EGFP, SARS-CoV-2 N wt, N mutants defective in G3BP binding (1A-4A mut), N-K²⁵⁷A+K²⁶¹A (N2b mut) or a mutant ‘N1a + N2b mut’ with substitutions in both the G3BP binding site (2A mut) and N2b (K²⁵⁷A+K²⁶¹A). At 24 hrs post transfection, cells were sodium arsenite-treated to induce HRI-mediated ISR with ensuing formation of SGs. (A) Representative (immune)fluorescence images. Scale bar: 50 μm. (B) Quantification of the results based on three independent experiments by counting all cells with detectable EGFP fluorescence or (C) highly expressing cells exclusively. For the statistical analysis, see S3 Table.

Fig 6. Quantitative assessment of N-mediated rescue of ISR-induced translational arrest.

HeLa cells were transfected to express EGFP, SARS-CoV-2 N2b, the full-length N proteins of SARS-CoV-2 and HCoV-229E, and mutants thereof from pEGFP-N3-based expression vectors to induce PKR-activated ISR. The capacity of these proteins to rescue translational repression of red fluorescent protein (RFP) in trans or lack thereof was measured by flow cytometry at 24 hr post transfection (left panels) and fluorescence microscopy (see S7 Fig). Representative flow cytometry histograms shown. The transfected cells were divided into non-RFP-expressing, low (L) RFP-expressing and high (H) RFP-expressing populations (see dashed lines in histograms). For each mutant protein, the [H/L] ratio was calculated and normalized, with those of the corresponding wildtype proteins set at 100%. The bar graphs show mean values with standard deviations based on three independent experiments (unpaired t-test; ****, P<0.001; ns, non-significant) (right).

Fig 7. The N2b domain inhibits PKR phosphorylation, prevents SG formation, and reduces the type I interferon response in EMCV infected cells.

(A) Structure of the polyprotein of recombinant EMCV-L^zn-N2aN2b. Schematic representation, mature cleavage products shown as boxes. Red asterisks indicate the locations of mutations (C¹⁹A/C²²A) in the Zn-finger domain of leader protein L, the red arrowhead that of a newly engineered EMCV 3C cleavage site and the yellow and grey boxes the native N-terminus of the EMCV polyprotein (see S8 Fig for details) and a Strep2 tag, respectively. (B) SG formation in EMCV-infected cells is prevented by SARS-CoV-2 N2a2b. HeLa cells were infected at MOI 10 with EMCV-L^zn-N2aN2b^wt or EMCV-L^zn-N2aN2b-K²⁵⁷A+K²⁶¹A (‘N2a2b-Mut’). Cells infected with wildtype EMCV or EMCV-L^zn served as controls as were cells infected with recombinant EMCV-L^zn derivatives encoding established coronavirus ISR antagonists MERS 4a and BW-Acp10. Cells fixed at 8 hpi were analyzed by IFA for dsRNA as a marker for infection and for SG scaffold protein G3BP2. (C) Percentages of infected cells with stress granules at 6 and 8 hpi. Bar graphs show the means of three independent experiments with >200 infected cells counted per sample. Standard deviations indicated by error bars; **** p< 0.0001; ns, non-significant (two-way ANOVA with Dunnett post hoc test). (D) Western-blot analysis for PKR, phosphorylated PKR (p-PKR), IRF3, phosphorylated IRF3 (p-IRF3), EMCV capsid proteins and β-actin. HeLa cells, (mock-)infected with recombinant EMCVs at MOI 10, were lysed at 6 or 8 hpi. Images are representative of three independent experiments. (E) Suppression of type I interferon response in EMCV-infected cells by N2a2b but not N2aN2b-Mut. Infected HeLa cells (MOI 10) were lysed at 6 or 8 hpi. Total RNA was analyzed by RT-qPCR for IFN-β and actin. IFN-β levels were calculated as fold induction compared to levels in mock-infected cells, after correction for actin mRNA levels, and normalized with EMCV-L^zn IFN-β levels set at 100%. Data represent means of three independent experiments. Standard deviations indicated by error bars; statistical significance compared to the results for EMCV-L^zn or EMCV-L^zn-N2aN2b infected cells calculated by two-way ANOVA with the Dunnett post hoc test; ** p<0.01; *** p<0.001; **** p<0.0001.