Computational fragment screening of dimeric SARS-CoV-2 main protease.

Abstract

During the coronavirus pandemic the lack of antiviral therapeutics drove intense interest in rapid drug discovery by the scientific, and medical communities. The Voelz lab took part in the COVID Moonshot initiative, which focused on virtually screening small drug-like fragments as leads for inhibitors of the main protease (Mpro) of SARS-CoV-2, identified using high throughput crystallographic techniques. The Voelz lab performed absolute binding free energy calculations on the [email protected] distributed computing platform to massively parallelize the molecular dynamic (MD) simulations. Crystallographic fragment-based screening techniques can be time consuming, costly, and dependent on the availability of protein crystals. Could fragment based drug discovery be performed completely in silico? To address this question, we aim to perform simulations to virtually screen the original fragment library used in the COVID Moonshot, and assess the accuracy of the predictions. The GROMACS simulation package was used to construct a simulation topology of the apo form of dimeric Mpro from its crystal structure. The correct protonation states of titratable residues were taken from published experimental results. As a first step, we prepared and performed a set of simulations of Mpro in solution, each with one of 11 drug fragments selected from the crystallographic fragment screen, all of which are known to co-crystallize with Mpro experimentally. The simulations are analyzed to determine where the drug fragments like to bind to the Mpro receptor. These results are compared results to the experimental co-crystal structures to determine how accurately MD simulations can predict the known binding pose of small drug like fragments.