Genome-Wide Interrogation of SARS-CoV-2 RNA-Protein Interactions Uncovers Hidden Regulatory Sites

Fig. 1.. eCLIP maps SARS-CoV-2 viral RNA interactions with NSP8, NSP12 and N. a) Enhanced crosslinking and immunoprecipitation resolves RNA bound to SARS-CoV-2 proteins in virus infected cells. After SDS-PAGE and membrane transfer, the region corresponding to the Immunoprecipitated (IP) protein and 75kDa above is excised for RNA extraction. Size-matched input (IN) is the same region in the cell lysate to determine background crosslinked RNA. b) Read density plots show RNA enriched by N, NSP8 and NSP12, distributed across the SARS-CoV-2 genome on the positive strand. Two biological replicates were performed for each sample and condition. c) Bar plot indicating the number of significantly enriched 20 nucleotide (nt) windows of IP vs IN samples for each replicate sample and condition. Significance, log2foldchange>1.5, qvalue<0.01. d) Bar plot showing the breakdown of the number of significant windows in each region. 5UTR, 5′ untranslated region; 3UTR, 3′ untranslated region. e) Scatter plot showing the mean read density of each 20 nt window across the positive sense virus genome in N IP samples at 6h and 48h. f) Heatmap showing the read density of N IP samples (colorbar) as 200 nt sliding…

Fig 2.. eCLIP highlights interactions between replicase proteins NSP8 and NSP12 at the 5′ and 3′ untranslated regions. a) Secondary structure of the region from the 5′ untranslated region to the end of the conserved stem loop 5 (SL5). Genome coordinates use positive strand numbering. Highlighted in blue are regions bound by NSP8 and NSP12, corresponding to yellow highlighted regions in b. Conserved stem loop structures (SLs), start codons (green) and stop codon (maroon) are annotated for the upper open reading frame (uORF) and the beginning of ORF1ab. Structure is predicted using RNAfold. b) Read density plots showing RNA enriched by NSP8 and NSP12 eCLIP on the positive (left) and negative (right) strand near the 5′ end of the positive sense genome. c) Secondary structure of the region from the N protein stop codon to the 3′ end of the positive strand. Highlighted in red are regions bound by NSP8 and NSP12, corresponding with yellow highlighted regions in d. Structure predicted using RNAfold and pseudoknot annotation from ref41. d) Read density plots showing RNA enriched by NSP8 and NSP12 on the positive (left) and negative (right) strand near the 3′ end of the positive sense genome.

Fig 3.. NSP12 binds to an RNA sequence within the conserved NSP3 Y1 domain coding sequence. a) Read density plot highlighting the peak that shows maximum read density across the genome in the NSP12 IP samples. Black dotted lines indicate translation start of annotated open reading frames and the start of each non-structural protein, corresponding to the schematic of SARS-CoV-2 genome. Orange dashed line box zooms into the bound region. Genome annotation indicates the bound region is in NSP3. NSP3 domains are illustrated in a schematic with different grey boxes, the Y1 domain coding sequence is outlined in orange, to which NSP12 binds. Y1+Y2+Y3 form the Y1+CoV Y domain conserved among coronaviruses. b) Zoomed in read density plots indicating enrichment of NSP12 bound RNA at the region 7436 – 7526, marked by dotted lines. c) Conservation analysis determined from an alignment of 9 coronavirus species using MAFFT v7.453 (2019/Nov/8). The phylogenetic tree is constructed based on the alignment using a neighbor joining algorithm in JalView version 2.11.4.1. d) The aligned sequences in (c) were used to calculate the Shannon entropy of nucleotide identity at each genomic position. Top: Shannon…

Fig 4.. NSP12 differentially regulates RNA levels upstream of its binding site. a) Schematic showing a dual luciferase reporter assay where 5' Y1 + NSP12 footprint sequences are inserted in between Cypridinia and Gaussia luciferases, CLuc and Gluc respectively. As negative controls, the DNA sequences are scrambled while preserving the amino acid sequence. b-d) Luciferase activity of CLuc (b), GLuc (c), and ratio of CLuc vs. GLuc (d) when reporters in (a) are co-transfected with NSP12 and mCherry negative control. q value is computed from multiple two-tailed unpaired t-test using Benjamini-Hochberg correction. e) Ratio of CLuc to GLuc RNA levels determined by RT-qPCR. ∆∆CT, log2 fold difference of CLuc RNA compared to GLuc RNA. f) PolyA mRNA-seq data showing the abundance of virus RNA sequencing read density across the genome. The orange line indicates the NSP12 footprint region 7436 – 7526. g) Schematic representation of a dual luciferase reporter assay utilizing two plasmids encoding Renilla luciferase (RLuc) and Firefly luciferase (FLuc). Six MS2 RNA hairpins are integrated into the 3′ untranslated region (UTR) of RLuc, enabling binding by the MS2 phage coat protein (MCP). A regulatory…

Fig 5.. N interacts with the 5′ base of stem loop 5 in the 5′ untranslated region. a) Secondary structure from the 5′ untranslated region (UTR) to the end of the conserved stem loop 5 (SL5). Highlighted in yellow is a region that interacts with N corresponding to the highlighted regions in b. b) Read density plots showing RNA enriched by N eCLIP near the 5′ start of SARS-CoV-2 RNA. c) Shannon entropy conservation analysis focused on the 5′ region of SARS-CoV-2 RNA; the orange-highlighted region corresponds to the 139 – 181 nt region bound by N. d) Schematic showing the region targeted by a locked nucleic acid (LNA) complementary to the N interaction site. e) RT-qPCR results indicating log2 fold change of SARS-CoV-2 genomic RNA levels in virions released from cells treated with anti_SL5_base LNA, compared to a scrambled LNA control and siRNAs targeting ACE2. *p-values are from two-tailed unpaired t-tests.

Fig 6.. Interaction between N and 5′ untranslated region impacts translation a) Schematic showing a dual luciferase reporter assay where different SARS-CoV-2 5′ end regions are inserted upstream of a Gaussia luciferase (GLuc) to regulate its expression. Cypridina luciferase (CLuc) serves as a transfection control in all four reporter constructs. pLT1 is a negative control vector with no 5′ UTR inserted. b) Luciferase activity GLuc/CLuc normalized to pLT1 when co-transfected with N, with mCherry as a negative control. *q values from multiple unpaired two-tailed t-test with Benjamini-Hochberg correction. c) Ratio of CLuc to GLuc RNA levels determined by RT-qPCR. d) Fold change in the RLuc/FLuc luciferase activity ratio for MCP-tagged proteins. 3×FLAG serves as a non-regulatory control, CNOT7 as a downregulatory control, and BOLL as an upregulatory control. *p-values from two-tailed unpaired t-tests. e) AlphaFold3 structure prediction of N and 5′ UTR-SL5 RNA. f) Magnified regions indicating N amino acid and nucleobase interactions. g) Relative GLuc/CLuc luciferase activity ratio due to overexpressing N mutants (corresponding to d) compared to wildtype N and mCherry. *p-values from…

Fig 7.. N binds upstream of the frameshift element and regulates ribosomal frameshifting a) Schematic showing the frameshift element, attenuator loop, slippery site and pseudoknot, and highlighted in orange is a region that interacts with N, corresponding with the blue highlighted region in b. Anti-FSE LNA, locked nucleic acid targeting the N footprint. b) Read density plots showing RNA enriched by N eCLIP (N footprint). c) Shannon entropy zoomed into the region around the frameshift element; orange highlighted sequence indicates the 13,415 – 13,435 nt region bound by N. d) Schematic showing a dual luciferase reporter assay where the SARS-CoV-2 frameshift region is inserted upstream of a Gaussia luciferase (GLuc) to regulate its expression, with (FSE3), without (FSE6) the N footprint and with a scrambled region (FSE8). GLuc is −1 frameshifted from an upstream Cypridina luciferase (CLuc). GLuc is observed only if −1 frameshifting occurs. e) Luciferase activity GLuc/Cluc normalized to FSE6 when frameshift reporters are co-transfected with N. *q values from multiple Welch’s t-test with Benjamini-Hochberg correction. f) RT-qPCR results indicating log2 fold change of SARS-CoV-2 genomic RNA…