Predicting the inhibitory Potentials of 1-heteroaryl-2-alkoxyphenyl analogues against SARs-coV-2 virus via QSAR Models and Molecular Modelling

The COVID-19 pandemic, caused by the SARS-CoV-2 virus, has led to the hospitalization of over 287 million individuals worldwide and resulted in more than 7.1 million deaths globally. This pandemic has sparked a significant surge in medical research globally. Computer-aided drug design enables the study of the behavior of existing molecules. It makes predictions on the properties of new derivatives, which could show promise as safer and more effective alternatives. Therefore, this work considered machine learned quantitative structure-activity modeling in predicting the bioactivity of some 1-heteroaryl-2-alkoxyphenyl compounds. The grid-search support vector regression (GS-SVR) model was adopted as the representative machine learned QSAR model as it outperformed other models, R² = 0.924 (≥ 0.6) and Q²cv = 0.858 (≥ 0.5). Furthermore, structural activity relationship studies generated 102 new compounds which were evaluated using the GS-SVR model. Lead compounds were selected and docked at the active sites of SARS-CoV-2 main protease (M^pro), papain-like protease (PL^pro), and spike glycoprotein. These compounds showed visible interactions with the amino acid residues at the active sites of these receptors. However, the docking scores of the compounds with the spike glycoprotein were higher than those of the standard drug, CP-100356, a spike glycoprotein inhibitor, unlike in the Mpro and PL^pro. The stability and flexibility of the complexes were determined via molecular dynamics simulation. The average RMSD values of the unbound protein and the complexes were below 2.50 Å, confirming their stability. Given the promising results, these molecules are currently being considered for synthesis and further investigation against SARS-CoV-2 cell lines.