Fusion proteins were constructed between the delta opioid receptor and forms of the alpha subunit of G(i1) in which cysteine(351) was mutated to a range of amino acids. GDP reduced the binding of the agonist [(3)H]DADLE but not the antagonist [(3)H]naltrindole to both the receptor alone and all the delta opioid receptor-Cys(351)XaaG(i1)alpha fusion proteins. For the fusion proteins the pEC(50) for GDP was strongly correlated with the n-octanol/H(2)O partition co-efficient of G protein residue(351). Fusion proteins in which this residue was either isoleucine or glycine had similar observed binding kinetics for [(3)H]DADLE. However, the rate of dissociation of [(3)H]DADLE was substantially greater for the glycine-containing fusion protein than that containing isoleucine, indicating that more hydrophobic residues imbued greater stability to the agonist-receptor-G protein ternary complex. This resulted in a higher affinity of binding of [(3)H]DADLE to the fusion protein containing isoleucine(351). In expectation with the binding data, maximal DADLE-stimulated GTP hydrolysis by the isoleucine(351)-containing fusion protein was two-fold greater and the potency of DADLE seven-fold higher than for the version containing glycine. These results demonstrate that the stability of the ternary complex between delta opioid receptor, G(i1)alpha and an agonist (but not antagonist) ligand is dependent upon the nature of residue(351) of the G protein and that this determines the effectiveness of information flow from the receptor to the G protein.