The differential effect of SARS-CoV-2 NSP1 on mRNA translation and stability reveals new insights linking ribosome recruitment, codon usage, and virus evolution

Figure 1.. Differential sensitivity of reporter gene expression to NSP1. (A) Effect of increasing amount of NSP1 WT-expressing plasmid on expression of the indicated reporter genes in HEK293T cells. Data are the average of two independent experiments and are expressed as a percentage relative to the cells transfected with 0.6 μg of pNSP1 M1 (null mutant, control). Arrow indicates the amount of pNSP1 plasmid selected for further experiments. (B) Comparative analysis of the effect of NSP1 WT, oligo4VIC, and silvestrol on expression of FFL, RL, and EGFP reporters. The binding sites of these inhibitors on 48S-PIC model are indicated (some eIFs were omitted for simplicity). Data are the average of two independent experiments. (C) Comparative analysis of the effect of NSP1 WT and M2 (R124A/K125A, mRNA degradation null) on the expression of the indicated reporter genes. A representative western blot monitoring the expression of different NSP1 variants is also shown. (D) Comparative analysis of the effect of NSP1 WT and M2 on accumulation of the indicated reporter mRNAs 24 h after transfection. mRNA levels were quantified by RT-qPCR as described in the ‘Materials and methods’ section. Significance…

Figure 2.. Influence of mRNA codon usage bias on sensitivity to NSP1. (A) Correlation between tAIg of reporter mRNAs and their corresponding sensitivities to NSP1 WT or M2. Data were scored as described in Fig. 1. The R2 and P-values are indicated. (B) Effect of RL CDS recoding on sensitivity to NSP1 WT and M2. The CAI, tAIg, and CSCg values for parental (RL) and RL-1 to three variants are indicated. Right panel shows a relative quantification of RL and RL-3 mRNA levels in HEK293T cells expressing NSP1 WT or M2. (C) Correlation between TE, tAIg, and sensitivity to NSP1 WT and M2. Upper left panel shows TE calculated for the indicated mRNAs by polysome profiling followed of RT-qPCR. The resulting heat map of the relative quantification is shown. Correlation between TE and tAIg (upper right panel) or TE and sensitivity to mRNA-induced degradation by NSP1 WT or M2 (bottom panel). Data were scored as described above and the resulting R2 is indicated. Bottom right panel shows the correlation between TE of human mRNAs and sensitivity to NSP1-induced mRNA degradation in Calu3 cells infected with SARS-CoV-2 WT compared with that infected by a mutant virus defective in NSP1 (ΔNSP1). Data were taken…

Figure 3.. Influence of 5′ UTR length and secondary structure on sensitivity to NSP1. (A) RL plasmids with 5′ UTRs ranging from 8 to 650 nt were co-transfected with 82-FFL plasmid in HEK293T cells and the resulting normalized activities were expressed as a percentage relative to the 82-RL construct (-NSP1). The sensitivity to NSP1 WT expression (+NSP1) for every construct is expressed as described in previous figures. (B) Effect of NSP1 WT and M2 on 82-RL and TISU-RL expression (left panel) or 94-EGFP and TISU-EGFP expression assayed by western blot (middle panel). The effect of increasing amount of pNSP1 WT or M2 plasmids on TISU-EGFP expression is also shown. ND, not detected. Right panel shows mRNA level quantification of 94-EGFP and TISU-EGFP in cells expressing NSP1 WT or M2. Significance was scored as described before. (C) Influence of stable stem-loops (SLs) placed immediately before the AUG of FFL on the sensitivity to NSP1s. The placement and stability (ΔG°) of SLs are shown (left panel). Effect of SLs on FFL expression relative to 82-FFL (lower left panel) and sensitivity to NSP1 WT and M2 (top middle panel). SL50-FFL expression relief by NSP1 WT was also confirmed by western blot…

Figure 4.. The presence of Gs near the 5′ cap critically determines the sensitivity to NSP1. (A) Schematic diagram of G-less 5′ UTR and its variants including G, SL, or GA motifs at the indicated positions. The composition (% G) and stability (ΔG° in kcal/mol) of the resulting sequences are indicated. The Gs introduced in the sequences are in bold. The first 22-nt sequence of 5′ L N of SARS-CoV-2 is also shown. (B) Left panel shows a comparative analysis of RL expression with G-less or 82 5′ UTRs in combination with optimal or suboptimal codon usage (RL versus RL-3) in HEK293T cells. Effect of NSP1 WT or M2 on the expression of G-less 5′ UTR in combination with optimal or suboptimal codon usage (middle and right panels). Significance was scored as described in previous figures. (C) Sensitivity of G-less and its variants to increasing amounts of NSP1 WT. Data are the average of two independent experiments. (D) Reduced sensitivity of G-less-RL expression to eIF4A inhibition by silvestrol (left panel) or RocA (right panel) compared to RL, FFL, or M1 and M4 variants of G-less-RL. Inhibitors were added 1 h after transfection and measurements were taken 16 h later. Data are the mean ± standard…

Figure 5.. NSP1 promotes slotting of PIC to mRNA. (A) Effect of progressive shortening of 5′ UTR G-less on sensitivity to NSP1 WT in HEK293T cells. Co-transfections also included FFL plasmid to normalize for FFL activity. Data are expressed as log2 fold-change expression. Data points correspond to the mean from at least three independent experiments that were adjusted to a sigmoid function (R2= 0.92); shading represents the 95% confidence interval. The resulting estimated blind spot of 12–15 nt is shown. A model of 48S-PIC showing the position of mRNA residues along the mRNA channel of 40S is also represented (right panel). (B) Comparative analysis of the expression of G-less-RL and Slot-RL constructs (left panel). Comparative analysis of the sensitivity of 82-RL and Slot-RL to NSP1s (right panel). Significance was scored as described in previous figures. (C) Effect of upstream out-of-frame AUGs (uAUG) on expression of G-less-RL and sensitivity to NSP1. The uAUG was inserted where indicated and the resulting uORFs partially overlap the 5′ of the main RL ORF (line). The presence of eIF4A-binding motif (GA)6 is indicated. HEK293T cells were cotransfected with the indicated constructs together…

Figure 6.. A combination of 5′ UTR and codon usage bias promotes resistance of N mRNA SARS-CoV-2 to NSP1. (A) Plot of tAIg and GC3 (%) distributions among human and coronavirus (hCoV) mRNAs. Values corresponding to human ACTB and SARS-CoV-2 mRNA N and its variant N1 are indicated, along with the Pearson correlation. (B) A representative western blot analysis of N and EGFP accumulated in HEK293T cells co-expressing NSP1 M1, WT or M2 (upper panel) and the corresponding quantification (bottom left panel). Relative quantification of mRNA levels of N variants in cells co-expressing NSP1 WT (bottom right panel). Significance was scored as described in previous figures. The combination of 5′ UTR (95 or 5′ L) with CDS N is indicated at the top of each panel. The tAIg values for N and N1 are 0.21 and 0.248, respectively. (C) TE of 5′ L-N mRNA in HEK293T cells compared to some other reporters included in the co-transfection experiment (upper panel). Comparative analysis of the expression level and sensitivity to NSP1 among N variants with different combinations of 5′ UTR and codon usage. (D) Analysis of the expression of 5′ L-RL, G-less-RL and G-less_67-RL compared to 82-RL (left panel). The…

Figure 7.. Genome-wide analysis of features in human mRNAs associated with sensitivity to NSP1. (A) Violin plots of the differential sensitivity of human mRNAs to NSP1 in transfected cells or in the context of infection. For the analysis, we used the datasets from Yuan et al. (2020; top) and from Fisher et al. (2022; GSE200422, bottom). mRNAs showing a log2FC ←0.5 or > 0.5 were classified as downregulated or upregulated by NSP1, respectively. The cumulative fraction of log2FC values for mRNAs showing fast or slow ribosome transit for every dataset is shown. Mann–Whitney test for fast and slow groups is shown (B) Prediction of down and up classes by Random Forest classifier. The mean accuracy obtained was 0.81. The resulting three most important features captured by the model are shown. (C) Model of the differential effect of NSP1 on mRNA translation and stability. NSP1 binding prevents the penetration of mRNA through the entry channel of 40S (threading). However, lateral attachment of mRNA by slotting is still possible in the presence of NSP1. The resulting complex can proceed efficiently for mRNAs with G-low 5′ UTR. However, for mRNAs with eIF4A binding motifs (Gs or GAs) near the cap,…