Pharmacophore and quantitative structure activity relationship modelling of UDP-glucuronosyltransferase 1A1 (UGT1A1)...

UDP-glucuronosyltransferase 1A1 (UGT1A1) is a polymorphic enzyme responsible for the glucuronidation of structurally diverse drugs, non-drug xenobiotics and endogenous compounds (e.g. bilirubin). Thus, definition of UGT1A1 substrate and inhibitor selectivities and binding affinities assumes importance for the identification of compounds whose elimination may be impaired in subjects with variant genotypes, and for the prediction of potentially inhibitory interactions involving xenobiotics and endogenous compounds metabolized by UGT1A1. We report the generation of two- and three-dimensional (2D and 3D) quantitative structure activity relationships (QSAR) and pharmacophore models for 23 known UGT1A1 substrates with diverse structure and binding affinity. Initially, a simple procedure was developed to determine apparent inhibition constants (Ki,app) for these compounds. Eighteen substrates were subsequently used to construct models and the remaining five to validate the predictive ability of the models. Three different models were constructed: (i) three feature pharmacophore model able to predict the Ki,app on the basis of the degree to which a substrate can fit to the arrangement of 3D features (r2 = 0.87, Ki,app for all five test substrates predicted within log unit); (ii) 3D-QSAR using a 'common features' pharmacophore to align the substrates (r2 = 0.71, Ki,app for four out of five test substrates predicted within one log unit); (iii) 2D-QSAR constructed with six chemical descriptors (r2 = 0.92, Ki,app of all five test substrates predicted within one log unit). The common features pharmacophore demonstrated the importance of two hydrophobic domains separated from the glucuronidation site by 4 A and 7 A, respectively. These models, which represent the first generalized predictive models for a UGT isoform, complement each other and are an important first step towards computer based (in silico) models of UGT1A1 for high throughput prediction of metabolism.