A Way for Finding Ligands for New Binding Sites

Abstract

Analysis of protein structures shows that most of them have potential binding sites that may be considered as applicable for new ligand design. The lack of known ligands interacting with such binding sites seriously complicated potential ligands selection. We have developed an approach that can increase the effectiveness of virtual screening for such ligands. It integrates methods of de novo ligand design, pharmacophore modeling, molecular docking, molecular dynamics, calculation of binding energies by MM- GBSA. This approach starts by the de novo design of virtual library of potential compounds followed by selection of favourable substructures and their correct positioning in a new ligand binding site. This generated library has been used for a development of pharmacophore models that have been used for a virtual screening of molecular databases. The selected compounds were docked to the putative binding site to check their ability to accommodate into it and their ability to locate the identified favorable fragments in the same region of the binding site as de novo generated molecules. The further evaluation of the selected ligands can be carried out by standard CADD methods.