In Silico Identification of Inhibitory Compounds for SARS-Cov-2 Papain-Like Protease.

Abstract

Virtual screening with high-performance computers is a powerful and cost-effective technique in drug discovery. A chemical database is searched to find candidate compounds firmly bound to a target protein, judging from the binding poses and/or binding scores. The severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2) infectious disease has spread worldwide for the last three years, causing severe slumps in economic and social activities. SARS-Cov-2 has two viral proteases: 3-chymotrypsin-like (3CL) and papain-like (PL) protease. While approved drugs have already been released for the 3CL protease, no approved agent is available for PL protease. In this work, we carried out in silico screening for the PL protease inhibitors, combining docking simulation and molecular mechanics calculation. Docking simulations were applied to 8,820 molecules in a chemical database of approved and investigational compounds. Based on the binding poses generated by the docking simulations, molecular mechanics calculations were performed to optimize the binding structures and to obtain the binding scores. Based on the binding scores, 57 compounds were selected for in vitro assay of the inhibitory activity. Five inhibitory compounds were identified from the in vitro measurement. The predicted binding structures of the identified five compounds were examined, and the significant interaction between the individual compound and the protease catalytic site was clarified. This work demonstrates that computational virtual screening by combining docking simulation with molecular mechanics calculation is effective for searching candidate compounds in drug discovery.