Identification of potential anti-COVID-19 drug leads from Medicinal Plants through Virtual High-Throughput Screening

Abstract

Natural compounds are widely used as attractive and valuable starting points for drug lead discovery. The present study aims to identify phytochemical compounds found in medicinal plants as potential COVID-19 inhibitors, using ensemble docking simulations. To this end, a phytochemical library from the PHCD database – a database of natural chemical compositions of Persian medicinal herbs (https://persianherb.com) – have been virtually screened against four key protein targets in the SARS-CoV-2 life cycle – the Mpro and PLpro proteases and the Spike and human ACE2 proteins. Several potential antiviral lead candidates have been identified based on the “Computational Multitarget Screening” approach, in which favourite candidates interact simultaneously with all four targets. Four of the bioactive phytochemicals identified – Chelidimerine, Gallagyldilacton, Hinokiflavone, and Physalin Z – show the highest binding affinities to all the targets and are suggested to be the best choices for drug design research. Also, several important medicinal plants, including Chelidonium majus L., Punica granatum, Rhus coriaria, Capparis spinose, Cichorium intybus, and Cynara scolymus, with the most phytochemicals interacting with all the host and viral proteins, have been identified that can be considered as the most important herbal resources for drug development with the medicinal plant formulations against COVID-19.

Virtual Screening and binding mode analysis of selected FDA approved drugs against PLpro target: An effort to identify therapeutics to combat COVID-19

“…Based on our recent researches, the top-ranked poses (with the lowest-energy poses) from the first and the most populated clusters are chosen as the representative poses for each docked ligand [20,37,47]. Therefore, we focus on the two representative docking poses for each ligand and use their docking scores, instead of considering the mean over docking score of the top-ranked poses of all representative PL pro conformations or taking the best-scoring binding poses from an aggregation of all predicted docking poses (the lowest scoring poses of ensemble docking), as routinely utilized in an ensemble docking protocol [18,[43][44][49][50].…”

“…In total, 1600 (= 8 × 200) poses by AutoDock and 160 (= 8 × 20) poses by Vina were calculated for each ligand. All the predicted docking poses for each ligand were collected separately for each docking program and re-clustered based on the symmetry corrected heavy-atom RMSD algorithm implemented in AutoDock4 with an RMSD cutoff of 2.0 Å [20,[47][48][49][50][51]. As a result, all predicted poses of a given ligand into multiple different conformations of the PL pro binding site are simultaneously organized and clustered for identifying representative poses and subsequent analyses.…”