Methoxy and methylthio-substituted trans-stilbene derivatives as CYP1B1 inhibitors – QSAR study with detailed interpretation of molecular descriptors

A quantitative structure–activity relationship (QSAR) model was built using multiple linear regression (MLR) to predict the ability of series methyl and/or methylthio trans-stilbene derivatives to inhibit CYP1B1. Twenty-four compounds with their activity expressed as the negative log of the IC50 value (pIC50 [M]) were split into a training (20 compounds) and a test set (four compounds) using Kennard and Stone algorithm. Molecular descriptors were calculated using alvaDesc software after compound optimization in the Gaussian 09 package in PL-Grid. The model characterized by the best validation parameters (R2TRAIN = 0.954, Q2LOO = 0.898, R2TEST = 0.880) was chosen based on the chemometric method – cluster analysis. The applicability domain has been determined, indicating that the regression model can give reliable prediction. The study shows that the inhibitory activity against CYP1B1 of the methyl and/or methylthio trans-stilbene derivatives can be predicted by RDF035m, Mor10m, Eig04_AEA(bo), RDF070s, MaxDD descriptors. Finally, the paper attempts to interpret three-dimensional descriptors by assessing the impact of interatomic interactions, following the partition of molecules into fragments, on the final value of descriptors.