Theoretical Studies On The Predictions Of Pregnane X Receptor Activators And Non-Activators

The pregnane X receptor（PXR） is an important transcriptional regulator,recognizing the metabolism and excretion of a variety of different structural endogenous and exogenous compounds. The activation of PXR can mediate the expression and metabolism of many enzymes, including cytochrome P450 enzymes,glutathione-S-transferases, and some important drug transporters, such as multidrug resistance proteins（MDRs）, multi-drug resistance proteins（MRPs） and the organic anion transporter polypeptides（OATPs）. 1-7The PXR-activating clinical drugs may enhance the metabolism and efflux of other co-consumed drugs and mediate drug-drug interaction. Therefore, identification of PXR activators will be helpful to analyze the interaction of ligands-reporter and detect the potential drug-drug interactions.In this study, based on 532 structurally diverse compounds, we presented a comprehensive analysis with the aim to build accurate classification models for distinguishing PXR activators from non-activators by using na?ve Bayesian classification technique. First, the distributions of eight important molecular physicochemical properties of PXR activators versus non-activators were compared,illustrating that the hydrophobicity-related molecular descriptors（AlogP and logD）showed slightly better capability to discriminate PXR activators from non-activators than the others. Then, based on molecular physicochemical properties, Vol Surf descriptors and molecular fingerprints, the na?ve Bayesian classifiers were developed to separate PXR activators from non-activators. The results demonstrated that the introduction of molecular fingerprints is quite essential to build the classifiers with satisfactory prediction accuracy. The best Bayesian classifier based on 21 physicochemical properties, VolSurf descriptors and the LCFC₁0 fingerprints descriptors yields the prediction accuracy of 92.7% for the training set and that of85.2% for the test set. Moreover, by exploring the important structural fragments derived from the best Bayesian classifier, we observed that the flexibility is an important structural pattern for PXR activation. Besides, chemical compounds containing more halogen atoms and unsaturated alkanes chains relevant to π-π stacking are tend to be PXR activators. We believe that the na?ve Bayesian classifier can be used as a reliable virtual screening tool to predict PXR activation in the drug design/discovery pipeline.