In Silico Drug-designing Studies on Sulforaphane Analogues: Pharmacophore Mapping, Molecular Docking and QSAR Modeling

Abstract

Aims: In the presented work we successfully discovered several novel NQO1 inducers using the computational approaches. Background: The phytochemical sulforaphane (SFN) is a potent inducer of carcinogen detoxication enzymes like NAD(P)H:quinone oxidoreductase 1 (NQO1) through the Kelch-like erythroid cellderived protein with CNC homology[ECH]-associated protein 1 (Keap1)–[NF-E2]-related factor 2 (Nrf2) signaling pathway. Objective: In this paper, we report the first QSAR and pharmacophore modeling study of sulforaphane analogues as NQO1 inducers. The pharmacophore model and understanding the relationships between the structures and activities of the known inducers will give useful information on the structural basis for NQO1 enzymatic activity and lead optimization for future rational design of new sulforaphane analogues as potent NQO1 inducers. Methods: In this study, a combination of QSAR modeling, pharmacophore generation, virtual screening and molecular docking was performed on a series of sulforaphane analogues as NQO1 inducers. Results: In deriving the QSAR model, the stepwise multiple linear regression established a reliable model with the training set (N: 43, R: 0.971, RMSE: 0.216) and test set (N: 14, R: 0.870, RMSE: 0.324, Q2: 0.80) molecules. The best ligand-based pharmacophore model comprised two hydrophobic (HY), one ring aromatic (RA) and three hydrogen bond acceptor (HBA) sites. The model was validated by a testing set and the decoys set, Güner–Henry (GH) scoring methods, etc. The enrichment of model was assessed by the sensitivity (0.92) and specificity (0.95). Moreover, the values of enrichment factor (EF) and the area under the receiver operating characteristics curve (AUC) were 12 and 0.94, respectively. This well-validated model was applied to screen two Asinex libraries for the novel NQO1 inducers. The hits were subsequently subjected to molecular docking after being filtering by Lipinski’s, MDDR-like, and Veber rules as well as evaluating their interaction with three major drugmetabolizing P450 enzymes, CYP2C9, CYP2D6 and CYP3A4. Ultimately, 12 hits filtered by molecular docking were subjected to validated QSAR model for calculating their inducer potencies and were introduced as potential NQO1 inducers for further investing action. Conclusions: Conclusively, the validated QSAR model was applied on the hits to calculate their inducer potencies and these 12 hits were introduced as potential NQO1 inducers for further investigations.