Abstract

Dipeptidyl peptidase-IV is well thought out as one of the key targets for treatment of type 2 diabetes mellitus and diverse scaffolds have proven effective in designing novel dipeptidyl peptidase-IV inhibitors. To this end, three dimensional quantitative structure activity relationship analysis and molecular docking studies were performed on a set of 3-amino-N-substituted- 4-(substituted phenyl) butanamides to explore the structural requirements for dipeptidyl peptidase-IV inhibitory activity using k-nearest neighbour molecular field analysis and AutoDock Vina, respectively. The compounds were arbitrarily assigned to active, moderately active and less active classes according to their biological activities. The most significant k-nearest neighbour model exhibited internal cross validation coefficient (q²) and external cross validation coefficient (pred_r²) as 0.67 and 0.82, respectively while the best partial least squares regression model showed 70% internal and 77% external predictability. Subsequent rigorous external validation through q²_(F2), q²_(F3), concordance correlation coefficient and mean absolute error provided further confidence in the predictability of both the models. Apropos the docking results, the compounds from active, moderately active and less active classes exhibited different binding patterns. Whereas the active compounds such as 12x and 12v occupied a similar space in the active site cavity as the crystallographic ligand (sitagliptin) (PDBID:1X70); the binding modes of compounds from moderately active and inactive classes were distinct from the latter. The findings of this study can provide an impetus for design of prospective potent dipeptidyl peptidase-IV inhibitors.