SARS-CoV-2 5’-UTR stem-loops activate the antiviral protein oligoadenylate synthetase 1 (OAS1)

Figures

• 

  Figure 1. Sequence and secondary structure of the SARS-CoV-2 genomic RNA 5’-end. Sequence and secondary structure of the overlapping SARS-CoV-2 5′-UTR and 5′-SE regions. Individual stem-loops are color-coded and the full structure shown as a cartoon representation (inset) as used in later figures.

• 

  Figure 2. SL1-8 of the SARS-CoV-2 5′-SE activates OAS proteins in cells and in vitro. Bioanalyzer analyses of rRNA integrity in A, wild-type, B, OAS3 KO, and C, RNase L KO A549 cells following transfection with SL1-8 RNA. Cleavage products of 28S and 18S rRNA are indicated by the solid and open arrows, respectively. Mock and poly (I:C) serve as negative and positive controls. Band intensities for cleaved and uncleaved 28S were quantified for three independent transfections and were used to calculate the fraction cleaved. D, immunoblots confirming expression of OAS1 (top) after IFN-β1α treatment in wild-type, OAS3 KO, and RNase L KO cells. OAS3 KO and RNase L KO are confirmed in middle and bottom blots, respectively. E, in vitro chromogenic OAS1 activity demonstrating that SL1-8 potently activates OAS1, with poly(I:C) and no RNA reactions serving as positive and negative controls, respectively.

• 

  Figure 3. Isolated SL1 and SL1-2 RNA regions from the SARS-CoV-2 5′-SE do not activate OAS1.A, sequence and secondary structure of SL1 (gray shading) and SL2 of the SARS-CoV-2 5′-SE and their approximate double-stranded helical lengths in base pairs (bp). Also indicated are the 5′-GG dinucleotides added to the 5′-end single stranded region (light blue; GG in outline font) and at the base of SL1 (black background) for IVT of SL1-2 and SL1-2(Δ5′), respectively. Boxed (top left), cartoon depicting potential SL1/SL2 coaxially stacking to form a continuous dsRNA helix of sufficient length to bind and activate OAS1 (∼17 bp). B, in vitro chromogenic assay showing that OAS1 is not activated by SL1, SL1-2, or the SL1-2 region lacking the single-stranded 5′-end sequence (SL1-2(Δ5′)).

• 

  Figure 4. Systematic truncation of the 5′-SE reveals SL1-4b as the shortest section that potently activates OAS1.A, cartoon of first 8 stem-loops (5′-SE) of the SARS-CoV-2 genomic RNA. B, in vitro OAS1 activity assay showing that truncation of 3′-end stem-loops does not decrease activation of OAS1 until SL4b is removed, making SL1-4b the shortest section that fully activates OAS1. SL1-4 exhibits intermediate capacity to activate OAS1, while SL1-3 and SL1-2 fail to activate. C, Bioanalyzer analysis of 28S and 18S rRNA integrity in wild-type, OAS3 KO, and RNase L KO A549 cells following transfection with SL1-4b or SL1-4. Cleavage products of 28S and 18S are shown by the black and white arrows, respectively. SL1-4b transfection results in more 28S cleavage product than SL1-4 mirroring in vitro OAS1 activation. D, fraction of cleaved 28S was quantified for three independent transfections of each RNA in wild-type and OAS3 KO cells, confirming that SL1-4b more potently activates the OAS/RNase L pathway than SL1-4.

• 

  Figure 5. SL4b is unstructured but alters the structure and OAS1 activation capacity of the SL1-4 region.A, plot (top) and mapping onto the SL1-4 RNA secondary structure (bottom) of average normalized SHAPE reactivity for three independent 2A3 modification experiments. Nucleotide reactivity is consistent with the formation of each SL structure. Grayshaded nucleotides indicate primer sites for which values could not be calculated. B, as panel A but for SL1-4b RNA. Nucleotide reactivity is again consistent with the formation of SL structures in the SL1-4 region but does not support formation of a structured SL4b under the conditions used. The low reactivity of two nucleotides on the 3′-side of SL4 which prompted testing of an additional RNA construct, SL1-4(UA), is indicated (shaded red box). C, in vitro OAS1 activity assay showing that SL1-4(UA) RNA promotes intermediate OAS1 activation, similar to SL1-4. D, in vitro OAS1 activity assay showing that the isolated SL4 does not activate OAS1; addition of SL4b in this RNA context results in a very weak activator. E, RNA UV thermal melting analysis of SL1-4b, SL1-4(UA), and SL1-4 in 100 mM NaCl, reveals an additional structural feature in SL1-4b…

• 

  Figure 6. Internal SL structures (SL2 and SL3) more subtly modulate OAS1 activation by SL1-4b.In vitro OAS1 activation assay with A, 5′-end truncations of the SL1-4b region: SL2-4b and SL3-4b, and B, internal deletion variants SL1-4b(ΔSL2), SL1-4b(ΔSL3), SL1-4b(ΔSL2-3) and SL1-4b(NL). Bioanalyzer analysis of 28S and 18S rRNA integrity in C, wild-type, D, OAS3 KO, and E, RNase L KO A549 cells. Cleavage products of 28S and 18S are shown by the black and white arrows, with cleavage of 28S being quantified for three independent transfections in wild-type and OAS3 KO cells. Quantification shows that transfection with SL1-4b(ΔSL2) leads to more cleavage of 28S by the OAS/RNase L pathway than SL1-4b(ΔSL3).

• 

  Figure 7. OAS1 interaction promotes changes on nucleotide dynamics in the regions identified as necessary for full OAS1 activation by SL1-4b. Average difference SHAPE reactivity (values without OAS1 subtracted from values with OAS1) from three independent modification experiments is mapped onto the RNA secondary structure for SL1-4b with A, 100 nM OAS1 or B, 1 μM OAS1, and for SL1-4 and C, 100 nM OAS1 or D, 1 μM OAS1. Changes in nucleotide dynamics are observed (redshaded regions) in SL1-4b in the regions required for full OAS1 activation, i.e. SL1, SL3, SL4 and SL4b. In contrast, SL2 exhibits similar minor differences in reactivity in both RNAs. Gray shaded nucleotides indicate primer sites for which values could not be calculated.

• 

  Figure 8. OAS1 activation by the SARS-CoV-2 5′-SE requires multiple features of the SL1-4b region.A, AlphaFold3 model of OAS1 (1–346) binding SL1-4b. The left model is color-coded by stem-loops. The right model is showing difference SHAPE results from Fig. 7 mapped onto RNA. OAS1 binds to SL4, which is coaxially stacking with SL3. SL1 and “SL4b” are interacting with the C-terminal region of OAS1 near the regions showing a decrease in SHAPE reactivity. B, the schematic model depicts SL4 (>17 bp dsRNA) as the main OAS1 interaction site, with major (large green arrows) and supporting contributions (small green arrows) from the other RNA elements, SL4b and SL1/SL3, respectively (see main text for details).