Artificial intelligence based virtual screening study for competitive and allosteric inhibitors of the SARS-CoV-2 main protease

Abstract

SARS-CoV-2 is a highly contagious and dangerous coronavirus that first appeared in late 2019 causing COVID-19, a pandemic of acute respiratory illnesses that is still a threat to health and the general public safety. We performed deep docking studies of 800 M unique compounds in both the active and allosteric sites of the SARS-COV-2 Main Protease (M^pro) and the 15 M and 13 M virtual hits obtained were further taken for conventional docking and molecular dynamic (MD) studies. The best XP Glide docking scores obtained were −14.242 and −12.059 kcal/mol by CHEMBL591669 and the highest binding affinities were −10.5 kcal/mol (from 444215) and −11.2 kcal/mol (from NPC95421) for active and allosteric sites, respectively. Some hits can bind both sites making them a great area of concern. Re-docking of 8 random allosteric complexes in the active site shows a significant reduction in docking scores with a t-test P value of 2.532 × 10⁻¹¹ at 95% confidence. Some specific interactions have higher elevations in docking scores. MD studies on 15 complexes show that single-ligand systems are stable as compared to double-ligand systems, and the allosteric binders identified are shown to modulate the active site binding as evidenced by the changes in the interaction patterns and stability of ligands and active site residues. When an allosteric complex was docked to the second monomer to check for homodimer formation, the validated homodimer could not be re-established, further supporting the potential of the identified allosteric binders. These findings could be important in developing novel therapeutics against SARS-CoV-2.

Communicated by Ramaswamy H. Sarma

Keywords:

• Artificial intelligence
• COVID-19
• deep docking
• molecular docking
• molecular dynamics simulation
• neural networks
• SARS-CoV-2 main protease

Acknowledgements

The author would like to acknowledge Bioinformatics Laboratory Facility of School of Biotechnology, KIIT Deemed to be University during the course of the study.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The project work is supported by the SERB-MATRICS project (MTR/2021/000191) from Department of Science and Technology, Govt. of India awarded to Dr. Rajani Kanta Mahapatra.