Background and purpose:Low efficacy partial agonists at the D2 dopamine receptor may be useful for treating schizophrenia. In this report we describe a method for assessing the efficacy of these compounds based on stimulation of [35S]GTPγS binding.Experimental approach:Agonist efficacy was assessed from [35S]GTPγS binding to membranes of CHO cells expressing D2 dopamine receptors in buffers with and without Na+. Effects of Na+ on receptor/G protein coupling were assessed using agonist/[3H]spiperone competition binding assays.Key results:When [35S]GTPγS binding assays were performed in buffers containing Na+, some agonists (aripiprazole, AJ‐76, UH‐232) exhibited very low efficacy whereas other agonists exhibited measurable efficacy. When Na+ was substituted by N‐methyl D‐glucamine, the efficacy of all agonists increased (relative to that of dopamine) but particularly for aripiprazole, aplindore, AJ‐76, (−)‐3‐PPP and UH‐232. In ligand binding assays, substitution of Na+ by N‐methyl D‐glucamine increased receptor/G protein coupling for some agonists ‐. aplindore, dopamine and (−)‐3‐PPP – but for aripiprazole, AJ‐76 and UH‐232 there was little effect on receptor/G protein coupling.Conclusions and implications:Substitution of Na+ by NMDG increases sensitivity in [35S]GTPγS binding assays so that very low efficacy agonists were detected clearly. For some agonists the effect seems to be mediated via enhanced receptor/G protein coupling whereas for others the effect is mediated at another point in the G protein activation cycle. AJ‐76, aripiprazole and UH‐232 seem particularly sensitive to this change in assay conditions. This work provides a new method to discover these very low efficacy agonists.British Journal of Pharmacology (2006) 149, 291–299. doi:10.1038/sj.bjp.0706866