Mechanisms of agonist action at the G protein-coupled D2(short) dopamine receptor expressed in Chinese hamster ovary cells have been investigated. Agonist binding was assayed in the presence and absence of GTP (100 microM). Data in the absence of GTP were fitted best by a two-site model (apomorphine, dopamine, 10,11-dihydroxy-N-n-propylnorapomorphine hydrochloride, and quinpirole) or a one-site model [bromocriptine, dihydroergocristine, and (-)-3-(3-hydroxyphenyl)-N-propylpiperidine hydrochloride], whereas in the presence of GTP a one-site model was the best fit for all compounds. Agonist binding parameters were used to provide a measure of the ability of the agonist to stabilise the ternary complex of agonist/receptor/G protein. Agonist stimulation of [35S]guanosine 5'-O-(3-thiotriphosphate) ([35S]GTPgammaS) binding for a range of agonist concentrations was measured and the EC50 and maximal effects determined. The initial rates of [35S]GTPgammaS binding induced by maximally stimulating agonist concentrations were also recorded. Simultaneous inhibition of agonist-stimulated [35S]GTPgammaS binding and receptor occupancy by spiperone was determined. Agonist inhibition of forskolin-stimulated cyclic AMP accumulation was determined for a range of agonist concentrations and the EC50 and maximal inhibition recorded. The data on the maximal agonist responses showed that it was possible to detect a spectrum of agonist efficacy (partial and full agonism) in both functional assays. The data on the apparent potencies of agonists to elicit the functional responses showed that different extents of amplification of response were seen for different agonists in both assays. The maximal activity data have been compared with the stabilisation of the agonist/receptor/G protein ternary complex as measured in binding assays.