We have examined the role of transmembrane domain amino acids in conferring subtype-selective ligand affinity to the human cholecystokinin-B (CCK-B)/gastrin receptor. Fifty-eight residues were sequentially replaced by the corresponding amino acids from the pharmacologically distinct CCK-A receptor subtype. 125I-CCK-8 competition binding experiments were performed to compare all mutant CCK-B/gastrin receptor constructs with the wild type control. Affinities for the nonselective agonist, CCK-8, as well as the subtype-selective peptide (gastrin), peptide-derived (PD135,158), and nonpeptide (L365,260 and L364,718) ligands were assessed. All of the mutants retained relatively high affinity for CCK-8, suggesting that the tertiary structure of these receptors was well maintained. Only eight of the amino acid substitutions had a significant effect on subtype selective binding. When compared with the wild type, single point mutations in the CCK-B/gastrin receptor decreased affinity for gastrin, L365,260, and PD135,158 up to 17-, 23-, and 61-fold, respectively. In contrast, the affinity for L364,718 increased up to 63-fold. None of the single amino acid substitutions, however, was sufficient to fully account for the subtype selectivity of any tested compound. Rather, CCK-B/gastrin receptor affinity appears to be influenced by multiple residues acting in concert. The 8 pharmacologically important amino acids cluster in the portion of the transmembrane domains adjacent to the cell surface. The spatial orientation of these residues was analyzed with a rhodopsin-based three-dimensional model of G-protein coupled receptor structure (Baldwin, J. M.(1993) EMBO J. 12, 1693-1703). This model [Abstract] predicts that the 8 crucial residues project into a putative ligand pocket, similar to the one which is well established for biogenic amine receptors (Caron, M. G., and Lefkowitz, R. J.(1993) Recent Prog. Horm. Res. 48, 277-290; Strader, C. D., Sigal, I. S., and Dixon, R. A.(1989) Trends Pharmacol. Sci. 10, Dec. Suppl., 26-30).
