Membrane preparations were made from Chinese Hamster Ovary (CHO) cells expressing 1.6 and 4.2 pmol/mg of recombinant human 5-HT1A receptors, as determined by saturation binding with the selective antagonist, [3H]-S 15535 ([3H]-4-(benzodioxan-5-yl)]-(indan-2-yl)piperazine). There was no change in the number of G-proteins activated by the full agonist, serotonin (5-HT; approximately 1.1 pmol/mg in each preparation, measured by [35S]-GTP gamma S saturation binding), therefore increasing the receptorG-protein ratio from approximately 1.41 (RGlow) to approximately 41 (RGhigh). Agonist efficacy was measured by stimulation of [35S]-GTP gamma S binding. The serotonergic agonist, eltoprazine, behaved as a partial agonist (Emax = 52.7%) at RGlow membranes but virtually as a full agonist (Emax = 93.2%) at RGhigh membranes, relative to 5-HT (= 100%). The latter exhibited a two-fold shift to the left in its concentration-response curve in RGhigh compared to RGlow membranes (P < 0.01). WAY 100,635 (N-¿2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl¿-N-(2-pyridinyl) -cyclo-hexane-carboxamide), did not alter [35S]-GTP gamma S binding from basal levels in either membrane preparation. In contrast, spiperone displayed inverse agonist activity, decreasing [35S]-GTP gamma S binding from basal levels by 17% in RGlow membranes but by 28% in RGhigh membranes. These data indicate that an increased receptorG-protein ratio (i) augments the potency of full agonists, (ii) increases the efficacy of partial agonists and (iii) increases the negative efficacy of inverse agonists at recombinant human 5-HT1A receptors. Furthermore, these data suggest that spiperone induces, or stabilises, a G-protein-coupled, but inactive conformation of the receptor.