Quantitative structure-activity relationships/comparative molecular field analysis (QSAR/CoMFA) for receptor-binding...

The 3-D quantitative structure-activity relationships/comparative molecular field analysis (QSAR/CoMFA) paradigm, which considers the primary importance of the molecular fields in biological recognition, is now widely used to analyze and predict receptor-binding properties of various ligands. CoMFA was applied to build 3-D QSAR models of substituted estradiol-receptor interactions, employing 3-D molecular databases of more than 40 molecules. Ligands included the 17 alpha-ethynyl- and isomeric 17 alpha (20E/Z)-(iodovinyl)estradiols and their 7 alpha-, 11 beta-, and 12 beta-methyl (-methoxy) and -ethyl (-ethoxy) derivatives as well as selected 2- and 4-halogenated analogs. The influence of different CoMFA descriptors was studied in order to achieve the highest possible cross-validated r2, as derived from partial least-squares calculations. Special emphasis was put on the analysis of the nature of H-bonding (donor/acceptor) interactions. The model with the best predictive performance (r2 = 0.895) was used to visualize steric and electrostatic features of the QSAR (standard deviation*coefficient contour maps) and to predict receptor-binding affinities (RBA) of substituted estradiols other than those included in the original database. Twenty-seven test molecules were selected, including five which had previously been reported by other investigators. For the latter, a very good correlation with literature RBA values was obtained, which together with the high cross-validated r2 provides evidence for the high predictive capacity of the model. Among the unknown structures, the model suggests several new substitutions to derive at reasonable affinity ligands for the estrogen receptor.