Recently, a series of xanthone analogues has been identified as α-glucosidase inhibitors. To provide deeper insight into the three-dimensional (3D) structural requirements for the activities of these molecules, CoMFA and CoMSIA approaches were employed on 54 xanthones to construct 3D-QSAR models. Their bioactive conformations were first investigated by docking studies and optimized by subsequent molecular dynamics (MD) simulations using the homology modeled structure of the target protein. Based on the docking/MD-determined conformers, 3D-QSAR studies generated several significant models in terms of 47 molecules as the training set. The best model (CoMSIA-SHA) yielded q 2 of 0.713, r 2 of 0.967 and F of 140.250. The robustness of the model was further externally confirmed by a test set of the remaining molecules (q 2 = 0.793, r 2 = 0.902, and k = 0.905). Contour maps provided much information for future design and optimization of new compounds with high inhibitory activities towards α-glucosidase. Graphical Abstract CoMSIA/SHA contour map of xanthone α-glucosidase inhibitor.
Keywords: 3D-QSAR; Docking; Homology modeling; Xanthone derivatives; α-glucosidase.