QSAR and docking analysis of A(2B) adenosine receptor antagonists based on non-xanthine scaffold

Abstract

Partial least squares (PLS) method as ligand-based method was applied for building quantitative structure-activity relationships (QSAR) regression model to predict the inhibitory activity of some A(2B) antagonists. The accuracy and predictability of the developed model were evaluated by several validation methods using external and internal test sets and also the criteria recommended by Tropsha and Roy were met. The result of the PLS model had a high statistical quality (R (2) = 0.936 and Q (2) = 0.867) for predicting the activity of the compounds. Evaluation of a test set of seven compounds with the developed PLS model revealed that this model is reliable and has a good predictability. Because of high correlation between the predicted and experimental values of activity, PLS model proved to be a highly predictive QSAR approach. Also, the reliability of the model was assessed through docking for the selected antagonists as structure-based method. A potential binding site of A(2B)AR was verified according to the previous studies of site-directed mutagenesis. Phe173, Glu174, His251, Asn254, Lys269, Ile276, and His 280 were determined to be involved in ligand-receptor interactions.