| A variety of Quantitative Structure-Activity Relationship (QSAR) methods were applied to various data sets of D1 dopamine agonists and antagonists. These QSAR methods included the three-dimensional (3D) approaches of conventional Comparative Molecular Field Analysis (CoMFA) and the more recently developed Cross-Validated R2 Guided Region Selection (q2-GRS) CoMFA procedure. Our 3D QSAR experiments were complemented with the concurrent application of the two-dimensional (2D) QSAR methods of Genetic Algorithm-Partial Least Squares (GA-PLS) and the K Nearest Neighbor (KNN), recently developed analysis methods that utilize topological descriptors of chemical structures. As a result, we have produced numerous highly predictive models of D1 agonist and antagonist affinity and intrinsic activity. Our antagonist data set included seven novel 1-phenyltetrahydroisoquinoline analogs, synthesized as probes of the purported hydrophobic region of the D1 receptor, and was comprised of a total of 29 diverse compounds. Likewise, our agonist affinity data set was large and made up of compounds from several chemical classes, with a total of 32 compounds. Finally, our model of intrinsic activity is the first to be reported for the D1 receptor, and only the second reported for any receptor class. The data set for this model of intrinsic activity contained 30 dopamine full agonists, partial agonists, and antagonists.; In addition to these ligand-based methods of structural determinants of ligand-receptor interactions, we have developed docking models for the D1 receptor. These studies have served to direct our own mutagenesis studies, the preliminary results of which have been incorporated into an iterative refinement of our docking model. Our resulting ligand-receptor interaction model depicts the D1 agonist dihydrexidine forming intermolecular hydrogen bonding and ionic interactions with Ser198, Ser202, and Asp103 of the dopamine receptor. |
