Organic synthesis combined with molecular modeling: A powerful approach to map the functional topography of dopamine and serotonin receptors
Previous molecular modeling studies of dopamine and serotonin receptor agonists docked into their receptors have produced mixed results, due mainly to the use of homology models based on the published crystal structure of inactive bovine rhodopsin. With the recently published structure of the more closely related β2 adrenergic receptor, whose endogenous ligands are also biogenic monoamine neurotransmitters, we have been able to develop an in silico-activated model of this G protein-coupled receptor (GPCR) based on Molecular Dynamics (MD) simulations of the membrane-bound and solvated protein. From this system, improved homology models of the serotonin 5-HT2A and dopamine D1 receptors have been created that appear well suited for the qualitative and quantitative study of agonist properties. ^ For the dopamine D1 receptor, the models were employed to: (1) elucidate the catechol hydrogen bonding network using the selective agonist A-77,636 and two analogs; (2) probe the steric space at the bottom of the binding site with analogs of dihydrexidine (DHX) methylated at the 7 and 8 positions; and (3) investigate the binding of analogs of DHX with substitutions at the pendant aromatic ring.^ In the case of the 5-HT2A receptor, the binding of different classes of ligands was studied, including phenylalkylamines, tryptamines, and ergolines. For the latter, the observations were used to determine the region of the binding pocket occupied by an 8-diethylamide group. Additionally, free energy of binding calculations using the Linear Interaction Energy (LIE) method were performed, obtaining a model that showed a squared correlation coefficient of 0.7018, with an associated error of 4.56 kJ/mol–1 . Validation of this model was performed with an independent test set, which showed a squared correlation coefficient of 0.7080. Notably, electrostatic interaction energies were not significant for this model.^ Finally, the models were used to study the binding profiles of newly synthesized ligands. During the course of this investigation, two sets of ligands were synthesized. The first group, targeted toward the D1 receptor, involved replacing the catechol group of DHX with a pyrazole or an aminopyrimidine ring, based on the structures of quinpirole and quinelorane. Only the pyrazole compound could be synthesized, and it showed poor binding affinity (18.3 μM), probably due to the incorrect orientation of the polar pyrazole proton. The second group of compounds, designed as 5-HT2A agonists, consisted of piperidine-based rigid analogs of a superpotent N-benzylphenethylamine. The (S,S)-2,6-disubstituted piperidine analog retained most of the affinity (2.5 nM) of the original ligand (0.19 nM), and showed dramatically increased selectivity for the 5-HT 2A over the 5-HT2C receptor of up to 128-fold, making it the most selective 5-HT2A agonist synthesized to date. Molecular modeling revealed that the rigidified N-benzyl group occupied the same general location in the 5-HT2A binding site as the original flexible compound.^ A better understanding of the subtle details of the activation process, as well as the interactions of various agonists with these receptors, will lead to new insights into the basis of their action, laying the groundwork for the design of more efficacious ligands with potential therapeutic and research applications.^
David E. Nichols, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our