Covalent and Non-Covalent Binding Free Energy Calculations for Peptidomimetic Inhibitors of SARS-CoV-2 Main Protease

COVID-19, the disease caused by the newly discovered coronavirus — SARS-CoV-2, has created global health, social, and economic crisis. At the time of writing (November 12, 2020), there are over 50 million confirmed cases and more than 1 million reported deaths due to COVID-19. Currently, there are no approved vaccines, and recently Veklury (remdesivir) was approved for the treatment of COVID-19 requiring hospitalization. The main protease (M^pro) of the virus is an attractive target for the development of effective antiviral therapeutics because it is required for proteolytic cleavage of viral polyproteins. Furthermore, the M^pro has no human homologues, so drugs designed to bind to this target directly have less risk for off-target reactivity. Recently, several high-resolution crystallographic structures of the M^pro in complex with inhibitors have been determined — to guide drug development and to spur efforts in structure-based drug design. One of the primary objectives of modern structure-based drug design is the accurate prediction of receptor­-ligand binding affinities for rational drug design and discovery. Here, we perform rigorous alchemical absolute binding free energy calculations and QM/MM calculations to give insight into the total binding energy of two recently crystallized inhibitors of SARS-CoV-2 M^pro, namely, N3 and α-ketoamide 13b. The total binding energy consists of both covalent and non-covalent binding components since both compounds are covalent inhibitors of the M^pro. Our results indicate that the covalent and non-covalent binding free energy contributions of both inhibitors to the M^pro target differ significantly. The N3 inhibitor has more favourable non-covalent interactions, particularly hydrogen bonding, in the binding site of the M^pro than the α-ketoamide inhibitor. But the Gibbs energy of reaction for the M^pro–α-ketoamide covalent adduct is greater than the Gibbs reaction energy for the M^pro–N3 covalent adduct. These differences in the covalent and non-covalent binding free energy contributions for both inhibitors could be a plausible explanation for their in vitro differences in antiviral activity. Our findings highlight the importance of both covalent and non-covalent binding free energy contributions to the absolute binding affinity of a covalent inhibitor towards its target.