Receptor-Based Pharmacophore Modeling in the Search for Natural Products for COVID-19 M^pro

Considering the urgency of the COVID-19 pandemic, we developed a receptor-based pharmacophore model for identifying FDA-approved drugs and hits from natural products. The COVID-19 main protease (M^pro) was selected for the development of the pharmacophore model. The model consisted of a hydrogen bond acceptor, donor, and hydrophobic features. These features demonstrated good corroboration with a previously reported model that was used to validate the present model, showing an RMSD value of 0.32. The virtual screening was carried out using the ZINC database. A set of 208,000 hits was extracted and filtered using the ligand pharmacophore mapping, applying the lead-like properties. Lipinski's filter and the fit value filter were used to minimize hits to the top 2000. Simultaneous docking was carried out for 200 hits for natural drugs belonging to the FDA-approved drug database. The top 28 hits from these experiments, with promising predicted pharmacodynamic and pharmacokinetic properties, are reported here. To optimize these hits as M^pro inhibitors and potential treatment options for COVID-19, bench work investigations are needed.