Material and Methods
The in silico studies were performed on Lenovo ThinkPad which has 12 GB RAM, 1TB hard disk, Intel i7 generation with 6 core. Molecular docking was performed on GLIDE (Grid-based Ligand Docking with Energetics) module of maestro 12.0 (Schrodinger LLC 2019, USA) between ligand/s molecules with a receptor macromolecule, mainly protein.
Some coumarin derivatives, both naturally derived and chemically derived were found to have good antiviral activity. Based on biological activity 2755 compounds of coumarin derivatives weredownloadedfromthewebsitePubChem (https://pubchem.ncbi.nlm.nih.gov/#query = coumarin), is a chemical database in SDF format. These ligands were prepared using the LigPrep module of Maestro. LigPrep performs 3D low energy structure conversion from 2D with accurate chiralities. Ionization were retained in original states with realistic bond lengths and bond angles, tautomers, ring conformation were generated, using the OPLS–2005 force field.
The protein structures, namely, 6w61 (methyltransferase-stimulatory factor complex of NSP16 and NSP10), 6vww (NSP15 Endoribonuclease), 6vxs (ADP ribose phosphatase (ADRP)of NSP3) and 6lu7 (protease) were retrieved from protein data bank (https://www.rcsb.org/) as anti-COVID–2019 targets. Retrieved structures were subjected to the protein preparation wizard of Maestro for preparation of the structures. The selected structures were processed for the incorporation of creating zero bond order for metal, assigning proper bond orders, creating disulphide bonds, and the addition of missing hydrogens. Optimization of hydrogen bonds was assigned using sample water orientations and Non-hydrogen atoms of protein structures were energy minimized until the RMSD (root mean square deviation) reaches the value of 0.3Å.
Generation of Receptor Grid
Cocrstallized structures downloaded from PDB had ligand was separated. The partial atomic charge cutoff was 0.25 and Van der Waals radii of receptor atoms were 1.0 Å by defaults. The centroid of the workspace ligand has an enclosing box to represent the activity of receptor/s.
Center of the bound ligand in receptor/s was selected to generate a grid box by using the above protocol in the receptor grid generation module of Maestro.
Virtual Screening and Molecular Docking
Virtual screening of prepared coumarin derivatives was performed using virtual screening workflow (VSW) of GLIDE (maestro, Schrödinger, LLC, New York, NY, USA). Running QuikProp, prefilters by Lipinski rule and remove ligands with reactive function groups were selected in VWS for filtering the ligands. VWS of Glide has three levels of Choose a docking precision, HTVS (High Throughput Virtual Screening) for rapid screening of large number ligands docking, standard-precision (SP) for screening ligands of unknown quality in large numbers, Extra-precision (XP) a more powerful and discriminating procedure for docking and scoring of ligands. Extra precision (XP) docking was performed on coumarin derivatives by keeping parameters of the scaling factor at 0.80 and partial charge cutoff at 0.15. The binding affinity of docked compounds for the receptors, methyltransferase-stimulatory factor complex of NSP16 and NSP10, NSP15 Endoribonuclease, ADP ribose phosphatase of NSP3 and protease, the active site was calculated from the docking binding energy.
Binding Free Energy calculation
Binding free energy of receptors and docked ligands were predicted using the Prime MM- GBSA (Molecular Mechanics-Generalized Born Surface Area) approach of Maestro 12.0 which includes the OPLS_2005 molecular mechanics energies, nonpolar term of solvation and VSGB solvent model . Prime MM-GBSA was performed using pose viewer file (generated after docking) of lead molecules to calculate binding free energy. The following descriptors were used to calculate changes in energy upon binding.
MM-GBSA DG bind = Ligand binding free energy, MM-GBSA E complex = complex free energy, MM-GBSA E protein = free energy of the receptor without the ligand, MM-GBSA E Ligand = unbound ligand-free energy.
Absorption, distribution, metabolism, excretion, and toxicity (ADME/T) properties studies reveal the medication properties of drugs. QuikProp tool in VSW (maestro 12.0) calculated the ADME/T (auxiliary, physicochemical, biochemical, pharmacokinetics, and harmfulness properties) information of ligands