Today’s post focuses on another protein encoded in open reading frame 1a (ORF1a) of many coronaviruses, the non-structural protein 9 (nsp9).
Nsp9 is thought to mediate viral replication. In SARS-CoV-1, nsp9’s role has been highlighted as a single-stranded RNA-binding subunit. (Egloff et al., 2004)
The structure of SARS-CoV-2 nsp9 was solved in both apo and peptide-bound forms by Litter et al. in July 2020.
The peptide identified seems to originate from a rhinoviral 3C protease sequence (LEVL). The structure of nsp9 is a homodimer, and the exogenous peptide is bound at the interface of the two monomers.
Previous research showed that disruption of nsp9 dimer formation by mutating the residues at the protein-protein interaction motif (GxxxG) debilitated the virus. (Miknis et al., 2009)
We assessed the druggability of nsp9 pockets and their amino acid variability across coronaviruses.
Using IcmPocketFinder (Molsoft), we identified two pairs of binding pockets on the SARS-CoV-2 nsp9 dimer (PDB code 6w9q). The first pair corresponds to the putative peptide-binding site identified by Littler et al. (figure 1 &2, Zenodo report). The second pair is within each monomer’s beta-barrel and distant from the protein interface (figure 3, Zenodo report).
By applying SiteMap (Schrodinger, NY), we assessed these pockets’ druggability using the druggability score (Dscore). Both pockets are druggable; nsp9 peptide-bound pocket has a Dscore of 1.152, and the beta-barrel has a Dscore of 0.956. (Halgren. 2009)
We determined the sidechains that surround the nsp9 pockets of interest within their 2.8 Å vicinity. The peptide-bound pocket is lined by 19 sidechains and the beta-barrel pocket by 14 sidechains. (Figure 4 & 5, Zenodo report)
We conduct a broad survey of viral proteins from the Alpha- and Beta-coronavirus genera to identify the most conserved druggable sites on viral proteins. We integrate the variability and druggability data to infer the most promising strategies for developing broad-spectrum viral inhibitors. Our analysis can be valuable in the context of the emergence of future coronaviruses from circulating viral strains in bat reservoir species.
We looked at twenty-seven reviewed nsp9 sequences; six entries from Alphacoronavirus genus and twenty-one entries from Betacoronavirus genus. Using these protein sequences, we assessed the variability of amino acids lining the nsp9 pockets by performing multiple sequence alignment. (Figure 6 & 7, Zenodo report) Based on this approach, the % identical residues at the nsp9 peptide-bound pocket is at 26% (53% sequence conservation), and none are identical at the beta-barrel pocket of nsp9.
Nicola De Maio, our collaborator from Nick Goldman’s lab at the European Bioinformatics Institute, looked at more than 15000 SARS-CoV-2 genome samples from COVID-19 patients and identified all mutations at the druggable nsp9 pockets.
He identified 11 non-synonymous variants at the peptide-bound pocket at nine unique amino acid positions of the nsp9 peptide-bound pocket (table 1, Zenodo report).
At the nsp9 beta-barrel pocket, he identified 4 non-synonymous variants at four unique amino acid positions (table 2, Zenodo report).
Our analysis shows that the nsp9 peptide-bound pocket, which is at the two nsp9 monomers’ interface, is more druggable than the beta-barrel pocket of nsp9.
The peptide-bound pocket has a relatively low amino acid identity at 26%, while the beta-barrel pocket has none.
Similarly, SARS-CoV-2 non-synonymous mutations are frequent at both sites (9 out of 19 residues for the peptide-bound pocket, 4 out of 14 residues for the beta-barrel pocket).
In summary, our data discourages the design of broad-spectrum inhibitors against nsp9 peptide-bound pockets and beta-barrel pockets.