Comparative modeling and molecular dynamics studies of the delta, kappa and mu opioid receptors.

Molecular models of the trans-membrane domains of delta, kappa and mu opioid receptors, members of the G-protein coupled receptor (GPCR) superfamily, were developed using techniques of homology modeling and molecular dynamics simulations. Structural elements were predicted from sequence alignments of opioid and related receptors based on (i) the consensus, periodicities and biophysical interpretations of alignment-derived properties, and (ii) tertiary structure homology to rhodopsin. Initial model structures of the three receptors were refined computationally with energy minimization and the result of the first 210 ps of a 2 ns molecular dynamics trajectory at 300K. Average structures from the trajectory obtained for each receptor subtype after release of the initial backbone constraints show small backbone deviations, indicating stability. During the molecular dynamics phase, subtype-differentiated residues of the receptors developed divergent structures within the models, including changes in regions common to the three subtypes and presumed to belong to ligand binding regions. The divergent features developed by the model structures appear to be consistent with the observed ligand binding selectivities of the opioid receptors. The results thus implicate identifiable receptor microenvironments as primary determinants of some of the observed subtype specificities in opiate ligand binding and in functional effects of mutagenesis. Networks of interacting residues observed in the models are common to the opiate receptors and other GPCRs, indicating core interfaces that are potentially responsible for structural integrity and signal transduction. Analysis of extended molecular dynamics trajectories reveals concerted motions of distant parts of ligand-binding regions, suggesting motion-sensitive components of ligand binding. The comparative modeling results from this study help clarify experimental observations of subtype differences and suggest both structural and dynamic rationales for differences in receptor properties.