Broad-spectrum coronavirus inhibitors discovered by modeling viral fusion dynamics

Abstract

Development of oral, broad-spectrum therapeutics targeting SARS-CoV-2, its variants, and related coronaviruses could curb the spread of COVID-19 and avert future pandemics. We created a novel computational discovery pipeline that employed molecular dynamics simulation (MDS), artificial intelligence (AI)-based docking predictions, and medicinal chemistry to design viral entry inhibitors that target a conserved region in the SARS-CoV-2 spike (S) protein that mediates membrane fusion. DrugBank library screening identified the orally available, FDA-approved AXL kinase inhibitor bemcentinib as binding this site and we demonstrated that it inhibits viral entry in a kinase-independent manner. Novel analogs predicted to bind to the same region and disrupt S protein conformational changes were designed using MDS and medicinal chemistry. These compounds significantly suppressed SARS-CoV-2 infection and blocked the entry of S protein-bearing pseudotyped α,β,γ,δ,ο variants as well as SARS CoV and MERS-CoV in human ACE2-expressing or DPP4-expressing cells more effectively than bemcentinib. When administered orally, the optimized lead compound also significantly inhibited SARS-CoV2 infection in mice. This computational design strategy may accelerate drug discovery for a broad range of applications.