Combined 3D-QSAR modeling and molecular docking study on spiro-derivatives as inhibitors of acetyl-CoA carboxylase

Acetyl-CoA carboxylases (ACC) have long been recognized as an attractive target for metabolic syndrome because of its indispensable role in the regulation of fatty acid metabolism in most living organisms. Based on the lead compound CP-640186, a series of spiro-compounds of benzothiazole derivatives have been identified as ACC inhibitors, which displayed high ACC inhibitor activities and good metabolic stability. Nevertheless, no quantitative structure-activity relationship study has been developed for this kind of ACC inhibitors. In order to evaluate the correlation between changes on structures and variation on biological activity, three-dimensional quantitative structure-activity relationship analysis was employed. Satisfactory results were obtained after performing a leave-one-out cross-validation study with q2 and r2 values of 0.625 and 0.927 by the comparative molecular field analysis model, 0.779 and 0.937 by the comparative molecular similarity indices analysis model, respectively. The predictive ability of the models was validated using external validation test set. The contour map analysis of comparative molecular field analysis and comparative molecular similarity indices analysis models indicated that electropositive and bulky groups substituted at the R1 position of carbamido group, and hydrophilic groups substituted at the phenyl ring were favorable for the inhibitory activity. Furthermore, FlexX docking was employed to investigate the binding mode between human ACC1 and its inhibitors and to validate the three-dimensional quantitative structure-activity relationship models. As a result, based on the results of three-dimensional quantitative structure-activity relationship and FlexX docking, a series of ACC inhibitors with higher predictive activities have been designed. This work might provide valuable information in further optimization of human ACC inhibitors.