Discovery of SARS-CoV-2 papain-like protease inhibitors through machine learning and molecular simulation approaches

Abstract

The papain-like protease (PLpro), a cysteine protease found in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), plays a crucial role in viral replication by cleaving the viral polyproteins and interfering with the host's innate immune response through deubiquitination and deISGylation activities. Consequently, targeting PLpro has emerged as an appealing therapeutic strategy against SARS-CoV-2 infection. Despite considerable efforts in the development of PLpro inhibitors, there is currently no drug available on the market that specifically targets PLpro. Improving drug screening strategies and identifying additional candidate compounds could significantly contribute to the advancement of antiviral agents targeting PLpro. To address this pressing issue, our present study has developed a highly efficient compound screening strategy based on a supervised machine learning approach. Integrated with further molecular simulation approaches such as molecular docking, molecular dynamics simulations, and quantum chemical calculations, we have identified seven compounds with potent inhibitory activity against PLpro. Notably, two of these compounds exhibited superior activity compared to Jun12682, which is currently considered the best-performing inhibitor against PLpro. Furthermore, some crucial residues in SARS-CoV-2 PLpro were recognized as favorable contributors to the binding with inhibitor, which would provide valuable insights for the development of more potent and highly selective SARS-CoV-2 PLpro inhibitors. The compound screening strategy and potential PLpro inhibitor candidates revealed in the present study would hold promise for advancing the development of antiviral drugs targeting SARS-CoV-2 and its variants.

Keywords: SARS-CoV-2; machine learning; molecular simulation; papain-like protease inhibitor; virtual screening.