QSAR modeling of anthocyanins, anthocyanidins and catechins as inhibitors of lipid peroxidation using three-dimensional descriptors

Abstract

This work describes the quantiative structure–activity relationship (QSAR) study of lipid peroxidation inhibitory effect of catechins, anthocyanidins and anthocyanins using molecular descriptors and physicochemical parameters derived from optimised three-dimensional (3D) structure, since this set of studied compounds contains stereoisomers with different activities. Six groups of 3D descriptors have been used to generate QSAR models: geometrical, 3D molecule representation of structures based on electron diffraction (3D-MoRSE); Randic molecular profiles; geometry, topology and atom weights assembly (GETAWAY); radial distribution function (RDF); and weigthed covariance matrices (WHIM) descriptors. The 3D molecular descriptors and physicochemical parameters have been calculated applying the online software Parameter Client and HyperChem 8.0. The primary selection of 3D molecular descriptors and physicochemical parameters was based on their ability to discriminate stereoisomers. Further selection of predictor variables for multiple regression was performed by the best-subset and forward stepwise method. The best-developed QSAR models consisted of geometrical, RDF and Randic molecular profiles descriptors. Those descriptors could be used for the prediction of the biological activity of catechin stereoisomers and their derivates. The obtained models suggest that the inhibitory effect of studied compounds is related to the shape of the molecule and the three-dimensional distribution of atomic mass in the molecule.