Digging for the discovery of SARS-CoV-2 nsp12 inhibitors: a pharmacophore-based and molecular dynamics simulation study

Abstract

Aim: COVID-19 is a global health threat. Therapeutics are urgently needed to cure patients severely infected with COVID-19. Objective: to investigate potential candidates of nsp12 inhibitors by searching for druggable cavity pockets within the viral protein and drug discovery. Methods: A virtual screening of ZINC natural products on SARS-CoV-2 nsp12's druggable cavity was performed. A lead compound with the highest affinity to nsp12 was simulated dynamically for 10 ns. Results: ZINC03977803 was nominated as the lead compound. The results showed stable interaction between ZINC03977803 and nsp12 during 10 ns. Discussion: ZINC03977803 showed stable interaction with the catalytic subunit of SARS-CoV-2, nsp12. It could inhibit the SARS-CoV-2 life cycle by direct interaction with nsp12 and inhibit RdRp complex formation.

Keywords: RNA-dependent RNA polymerase; SARS-CoV-2; molecular docking; molecular dynamics; natural product; pharmacophore-based drug discovery.

Figure 1.. Cavity no. 2 and its surface within the SARS-CoV-2 protein.

The nsp12 fingers are presented in Green color. Palm and thumb domains are also illustrated in cyan and magenta colors, respectively. As it is illustrated, cavity no. 2 (colored in black) is located adjustment to the left-handed finger (residues 366–581), palm (residues 582–620 and 680–815) and thumb (residues 816–920) domains.

Figure 2.. Schematic representation of adjunct pharmacophore features distributed around the cavity surface no.2.

The figure demonstrates the distribution of the pharmacophore features around the surface of cavity no. 2. H-bond acceptor (HBA) centers are shown in red, H-bond donor (HBD) centers are in blue, H-bond root is in cyan, and hydrophobic centers (HC) are in green, negative electrostatic center is in yellow, and positive electrostatic center is in gray. Pharmacophore features are shown in spears. The coordinates of HBA centers are overlapping with HBDs; therefore, they are not visible.

Figure 3.. Schematic of highlighted hits from different pharmacophores close or within the finger domain of the nsp12 protein.

The exact site of binding of the selected hits is presented with different colors on the surface of the nsp12 at the center of the figure. Further information, including pharmacophore number and their relative hit's affinity, is also presented. In this regard, the hit identified for pharmacophore no. 1 had the lowest affinity to its nsp12, and the hit highlighted for pharmacophore no. 4 had the highest affinity to nsp12 protein.

Figure 4.. Molecular docking validation.

The lead compound showed high affinity to the different crystallographic structures of SARS-CoV-2 nsp12. The lead compound had an affinity of -10.3 kcal.mol^-1 to 7BTF and it was slightly lower than -11 kcal.mol^-1 due to flexibility and non-redundancy of autodock Vina. PDB: Protein Data Bank.

Figure 5.. The binding site of ZINC03977803 within the nsp12 cavity pocket and interaction between nsp7/8 with nsp12.

(A) Shows the contacts of atom residues O and C with hydrogen donor and acceptor sites of the given amino acid residues within nsp12. Further 2D view and rotation of ZINC03977803 around cavity no. 2 are presented on the bottom-left side of the figure. (B) Right-top of the figure represents the nsp7 (green) and nsp8 (blue) interactive residues with nsp12 (red) in the crystallographic data of the SARS-CoV-2 proteins (PDB ID: 7BTF). Data showed higher numbers of nsp7 residues in close contact with nsp12. Further information on the interactive nsp12 residues is presented at the bottom-right of the figure.

Figure 6.. Root mean square of deviation and radius of gyration at nm scale within 10 ns of molecular dynamic production.

It has been shown that nsp12 RMSD is relative to the crystal structure. RMSD fluctuation differences between the equilibrated protein and the crystal protein are slight indications that the structure is stable favorably. Rg of nsp12 protein showed slight fluctuation along with simulation time, suggesting proper folding of the protein. Rg: Radius of gyration; RMSD: Root mean square deviation.

Figure 7.. The graphs of RMS, energy, hbond and mindist modules of GROMACS for ZINC03977803 and nsp12 backbone.

(A) RMSD fluctuation of the lead compound ZINC03977803 and its complex with nsp12 protein. The difference in the complex structure was substantially higher than that observed in the nsp12-CA and nsp12-Backbone. (B) Plot of the short-range potentials of LJ and Coul shows lower electrostatic and higher inter intermolecular interactions at the first 2ns of molecular dynamic simulation (MDS), which is converged after 2ns of MDS. (C) H-bonding between nsp12 amino acid residues and the lead compounds showed the N-terminal residues, Ala52, was the hydrogen bond donor. (D) Demonstrates the distance between nsp12-backbone interactive residues and the lead compound. (E) Shows the number of interactive residues of nsp12-backbone in close contact (<0.4 nm) with the lead compound during 10 ns of MDS. Coul-SR: Short-range Coulomb; LJ-SR: Short-range Lennard-Jones; RMSD: Root mean square deviation.

Figure 8.. Solvent accessible surface areas analysis of nsp12, nsp12–ligand complex and the lead compound.

(A) Demonstrates stable solvent-accessible surface area fluctuations during 8ns of the molecular dynamic simulated system. (B) Demonstrates stable solvent-accessible surface area of the lead compound during the molecular dynamic simulation. Only a slight reduction of surface area was observed in the complex of nsp12 with the lead compound, suggesting the location of the small molecule inside the cavity pocket.

Figure 9.. Covariance matrix (215 × 215) and PCA analysis of ZINC03977803 and SARS-CoV-2 nsp12.

(A) Eigenvector indexes indicate 100% eigenvalues of the covariance matrix in the 1st five eigenvector indexes. (B & C) Shows root mean square fluctuation and eigenvector components of the first five vectors. The result indicated atom motions at numbers 12 to 22 and 48 to 55, representing; (C) Aromatic ring rotations around the C-O-C bonds within the nsp12 protein. (D) Illustrates the 2D projection of nsp12 Cα atoms motion on eigenvectors 1 (-1.7 nm to 1.5 nm) and 5 (-0.9 nm to 1.5 nm).