Identifying natural product inhibitors against CDK9 enzyme via combined multicomplex-based pharmacophore modeling, interaction studies and molecular dynamics simulations

In the current work, multicomplex-based pharmacophore modeling was performed on the CDK9 enzyme. The generated models possess five, four, and six features, which were subjected to the validation process. Among them, six feature models were selected as representative models to conduct the virtual screening process. The screened drug-like candidates were chosen to perform molecular docking to study their interaction patterns within the binding cavity of the CDK9 protein. Based on the docking score and presence of crucial interactions, out of 780 filtered candidates, only 205 were docked. These docked candidates were further accessed via HYDE assessment. Based on ligand efficiency and Hyde score, only nine candidates passed the criteria. The stability of these nine complexes, along with the reference, was studied by molecular dynamics simulations. Out of nine, only seven displayed stable behaviour during the simulations, and their stability was further assessed by molecular mechanics‐Poisson‐Boltzmann surface area (MM-PBSA)‐based free binding energy calculations and per residue contribution. From the present contribution, we obtained seven unique scaffolds that can be utilized as the starting lead for the development of CDK9 anticancer compounds.