Three-dimensional quantitative structure-activity relationship analysis ofthe new potent sulfonylureas using comparative molecular similarity indices analysis

The present study describes the implementation of a new three-dimensional q uantitative structure-activity relationship (3D-QSAR) technique: comparativ e molecular similarity indices analysis (CoMSIA) to a set of novel herbicid al sulfonylureas targeted acetolactate synthase. Field expressions in terms of similarity indices in CoMSIA were applied instead of the usually used L ennard-Jones and Coulomb-type potentials in CoMFA. Two different kinds of a lignment techniques including field-fit alignment and atom-by-atom fits wer e used to produce the molecular aggregate. The results indicated that those two alignment rules generated comparative 3D-QSAR models with similar stat istical significance. However, from the predictive ability of the test set, the models from the alignment after maximal steric and electrostatic optim ization were slightly better than those from the simple atom-by-atom fits. Moreover, systematic variations of some parameters in CoMSIA were performed to search the best 3D-QSAR model. A significant cross-validated q(2) was o btained, indicating the predictive potential of the model for the untested compounds; meanwhile the predicted biological activities of the five compou nds in the test set were in good agreement with the experimental values. Th e CoMSIA coefficient contour plots identified several key features explaini ng the wide range of activities, which were very valuable for us in tracing the properties that really matter and getting insight into the potential m echanisms of the intermolecular interactions between inhibitor and receptor , especially with respect to the design of new compounds.