Using tiered computational screening to discover small molecule inhibitors of the SARS-CoV-2 NSP3 protein Mac1 domain

As new medications are used to treat COVID-19, many studies have reported that proteins such as spike, polymerase and proteases are prone to high levels of mutation that can create resistance to therapy over time. Thus, it becomes necessary to, not only target other viral proteins such as the non-structural proteins (NSP’s), but to also target the most conserved residues of these proteins. A synergistic combination of bioinformatics, computer-aided drug-design and in-vitro studies can feed into better understanding of SARS-CoV-2 (SC-2) and therefore help in the development of small molecule inhibitors against the NSP’s. As part of our initial anti-viral work, a pharmacophore study on NSP15 found a hit molecule (INS316) that made interactions with Ser293, Lys344 and Leu345 residues which are highly conserved across SC-2.

Our group was selected to enter an international challenge organized by CACHE to find inhibitors for the Mac1 domain of SC-2 NSP3. Our MSA alignment results of ~1 million NSP3 sequences indicated that the Mac1 domain is a highly conserved pocket that can be targeted for developing promising SC-2 inhibitors. We used a tiered screening workflow which included the use of volume/shape information of the binding pockets (fastROCS), use of in-house pharmacophore generation software (MoPBS/MOE) and performed docking in the binding pocket (FRED) to rank compounds for subsequent clustering and to identify hits that bind to these conserved pockets. The primary experimental validation results provided by CACHE found that two of our predicted hits show activity in HTRF and SPR assays.