Computational guided identification of a citrus flavonoid as potential inhibitor of SARS-CoV-2 main protease

Abstract

Although vaccine development is being undertaken at a breakneck speed, there is currently no effective antiviral drug for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causing COVID-19. Therefore, the present study aims to explore the possibilities offered by naturally available and abundant flavonoid compounds, as a prospective antiviral drug to combat the virus. A library of 44 citrus flavonoids was screened against the highly conserved Main Protease (M^pro) of SARS-CoV-2 using molecular docking. The compounds which showed better CDocker energy than the co-crystal inhibitor of M^pro were further revalidated by flexible docking within the active site; followed by assessment of drug likeness and toxicity parameters. The non-toxic compounds were further subjected to molecular dynamics simulation and predicted activity (IC₅₀) using 3D-QSAR analysis. Subsequently, hydrogen bonds and dehydration analysis of the best compound were performed to assess the binding affinity to M^pro. It was observed that out of the 44 citrus flavonoids, five compounds showed lower binding energy with M^pro than the co-crystal ligand. Moreover, these compounds also formed H-bonds with two important catalytic residues His41 and Cys145 of the active sites of M^pro. Three compounds which passed the drug likeness filter showed stable conformation during MD simulations. Among these, the lowest predicted IC₅₀ value was observed for Taxifolin. Therefore, this study suggests that Taxifolin, could be a potential inhibitor against SARS-CoV-2 main protease and can be further analysed by in vitro and in vivo experiments for management of the ongoing pandemic.

Keywords: COVID-19; Flavonoids; Main protease; Molecular docking; Molecular dynamic; SARS-CoV-2.

Fig. 1

Docking interaction of a 3WL, b Taxifolin (CF3), c Eriodictyol (CF5), d Isoscutellarein (CF7), e Luteolin (CF8) and F Quercetin (CF10) with M^pro. The green dashed line indicates the H-bonds between the ligands and the interacting residues of M^pro

Fig. 2

a Root mean square deviation of the receptor-ligand complexes; b Root mean square fluctuations of the residues of receptor-ligand complexes and c Radius of gyrations of the receptor-ligand complexes within the 30 ns simulation. In all the cases the complex M^pro-3WL is represented by blue line, M^pro-CF3 is represented by green line, M^pro-CF5 is represented by yellow line and M^pro-CF8 is represented by red line

Fig. 3

Interaction of a 3WL, b Taxifolin, c Eriodictyol and d Luteolin with M^pro after MD simulation for 30 ns. The green dashed line indicates the H-bonds between the ligands and the interacting residues of M^pro

Fig. 4

Superposition of the protein ligand complex of a M^pro-3WL, b M^pro-CF3, c M^pro-CF5 and d M^pro-CF8. The green colour complexes represent the protein–ligand complexes before MD simulation and red colour complexes represent the protein–ligand complexes after 30 ns MD simulation

Fig. 5

Distance of different hydrogen bonds formed within the simulation period for a 3WL; b Taxifolin, c Eriodictyol and d Luteolin

Fig. 6

Flucutation of binding free energies (ΔG) of protein–ligand complexes during the MD simulation period. The blue line represents the complex M^pro-3WL, brown line represents M^pro-CF3, grey line represents M^pro-CF5 and yellow line represents M^pro-CF8 complexes

Fig. 7:

3D-QSAR plot of a training set and b test set

Fig. 8

Visualization of binding of M^pro with a 3WL and b Taxifolin before MD simulation and c 3WL and d Taxifolin after MD simulation in SeeSAR with quantification of HYDE of the important non-hydrogen atoms which contribute in binding affinities