Hydropathicity analysis of 39 G-protein-coupled receptors (GPCR) reveals seven hydrophobic stretches corresponding to membrane spanning alpha-helices. The alignment of the primary sequences shows a high degree of homology in the GPCR transmembrane regions. 3D models of 39 GPCRs were generated using the refined model of bacteriorhodopsin as a template. Five cationic neurotransmitter receptors (serotonergic 5-HT2, dopaminergic D2, muscarinic m2, adrenergic alpha 2 and beta 2 receptors) were taken as prototypes and studied in detail. The 3D models of the cationic neurotransmitter receptors, together with their primary structure comparison, indicate that the agonist binding site is located near the extracellular face of the receptor and involves residues of the membrane-spanning helices 3, 4, 5, 6, and 7. The binding site consists of a negatively-charged Asp located at the middle of transmembrane helix 3 and a hydrophobic pocket containing conserved aromatic residues on helices 4, 5, 6, and 7. To define the precise receptor-ligand interactions, the natural neurotransmitters were docked into the binding sites. Residues responsible for the affinity, selectivity, and eventually stereospecificity of dopamine, adrenaline, noradrenaline, serotonin, and acetylcholine for their receptors were identified. The ligands are involved in electrostatic interactions as well as hydrogen bonds and specific hydrophobic aromatic interactions. All the GPCRs possess invariant hinge residues, which might be responsible for a conformational change during agonist binding and therefore influence dissociation and association of G-proteins to the receptors. The role of hydrophobic interactions and hydrogen bonds in the conformational change of the receptors, modulating the coupling to the G-protein, is discussed with regard to these residues. The models are in agreement with published data obtained from mutagenesis and labeling studies and represent important working hypotheses to direct future mutagenesis studies. They also enable structure-activity relationship studies and more rational drug design. The 3D models of other G-protein-coupled receptors have been generated in a similar way.