Computer-aided identification of natural lead compounds as cyclooxygenase-2 inhibitors using virtual screening and molecular dynamic simulation

Abstract

In order to find more natural lead-compounds as inhibitors for Cyclooxygenase-2, the essential structural features for human cyclooxygenase-2 inhibitors and 3D-Quantitative structure activity relationship (3D-QSAR) model were carried out based on dataset from three confirmatory bioassays using Phase program. Six point pharmacophore (AAHRRR) of COX-2 selective inhibitors was generated from training set of 52 compounds. The 3D-QSAR model was selected as having favourable statistic measures (R2 = 0.93, Q2ext = 0.81) for the training set and test set respectively. This model was developed using the best pharmacophore hypothesis that helped to reveal the essential features responsible for the anti-inflammatory activity. As a result, this pharmacophore hypothesis has aided in the identification of new four natural lead compounds from UNPD database (UNPD100208, UNPD168234, UNPD91145, UNPD57376) that can be used as potential anti-inflammatory agents or as a core structure to develop other more selective molecules. This result was confirmed by molecular docking, which showed that these four natural lead-compounds adopt the same orientation as Rofecoxib in the COX-2 active site. On the other hand, a molecular dynamic simulation (MDS) was applied and repeated on the top ranking complex 5KIR-UNPD100208 and compared with the results of MDS of 5KIR-Rofecoxib. The results from MDS revealed a good stability of the compound UNPD100208 in the active site based on several parameters such as RMSD, RMSF and potential energy, which can nominate it as a natural anti-inflammatory lead-compound candidate.