Anti-HIV Potential of Origanum vulgare Compounds Targeting Viral Reverse Transcriptase with High Binding and Stability Validated by Machine Learning
TL;DR
Computational analyses have ranked four compounds as HIV-RT inhibitors that need further experimental verification and Machine learning-based quantitative structure–activity relationship (QSAR) prediction of experimentally validated HIV-RT inhibitors was applied to predict inhibitory potency.
Abstract
Human immunodeficiency virus (HIV) is one of the viruses that has co-evolved within human populations for a considerable time. With the evolution of new drug-resistant HIV strains, the necessity to discover novel drugs has become an issue of great concern, especially those that have increased binding affinities and inhibitory activity towards RT enzymes. This study used an in silico approach to identify potential HIV-RT inhibitory candidates from Origanum vulgare (oregano). An initial in silico screening of approximately 820 compounds was conducted, and based on molecular docking-derived binding energy scores, four compounds (IMPHY000687, IMPHY007084, IMPHY004619, and IMPHY012021), with docking scores of −9.42, −9.32, −9.24, and −9.23 kcal/mol, respectively, were selected as the top-ranked phytocompounds and were subsequently validated using multiple computational approaches. These compounds were geometrically optimized using quantum-chemical calculations, and detailed interaction analyses were performed using a redocking procedure. Reproducibility of the dynamic behavior was evaluated by carrying out independent replica molecular dynamics simulations for 300 ns each for all complexes. The ligand-dependent conformational dynamics were identified using RMSD and RMSF analyses, along with variations in positional changes during simulation times. PCA and FEL analyses helped in identifying the conformations sampled by the system under study. In addition, QM/MM calculations provided complementary information on the electronic characteristics of the individual protein–ligand systems. Machine learning-based quantitative structure–activity relationship (QSAR) prediction of experimentally validated HIV-RT inhibitors was applied to predict inhibitory potency, yielding predicted pIC50 values for the selected phytochemicals compared with the reference molecule. All in all, comprehensive computational analyses have ranked these phytochemicals as HIV-RT inhibitors that need further experimental verification.