The binding characteristics of the kappa opioid ligands [3H]U69,593 and [3H]bremazocine, the mu opioid ligand [3H][D-ala2,N-Me-Phe4,glycol5]enkephalin and the delta opioid ligand [3H]p-Cl-[D-pen2,5]enkephalin were studied in rhesus monkey brain membranes in saturation binding experiments and were followed by competition binding experiments with a variety of peptidic and nonpeptidic opioid ligands. The [3H]U69,593 sites appeared to be a subset of kappa opioid receptors (kappa-1 receptors: Kd, 1.2 nM; Bmax, 66 fmol/mg). [3H]Bremazocine (in the presence of mu and delta receptor-masking agents), bound to a larger population of kappa receptors (kappa-all: Kd, 0.39 nM; Bmax, 227 fmol/mg), which presumably included the aforementioned kappa-1 sites. Competition binding experiments revealed that the presently defined kappa-1 sites were similar to previously reported sites in other mammalian species, particularly in terms of the higher kappa-1 selectivity observed with arylacetamide (e.g., U50,488) vs. benzomorphan kappa agonists (e.g., ethylketocyclazocine). The kappa-selective antagonist norbinaltorphimine (nor-BNI) displayed a very small (2.3-fold) selectivity for kappa-1 vs. kappa-all sites. This led to the prediction that in rhesus monkeys (n = 3), systemically administered nor-BNI (10 mg/kg s.c.) should have a very moderate degree of antagonist selectivity for the antinociceptive effects of a putative kappa-1-agonist, the arylacetamide U50,488 (0.1-3.2 mg/kg s.c.), vs. those of the benzomorphan kappa agonist ethylketocyclazocine (0.01-056 mg/kg s.c.). This prediction was confirmed in vivo because nor-BNI (10 mg/kg) caused a robust and long lasting (up to 21 days) antagonism of the antinociceptive effects of U50,488 and a small but significant antagonism of ethylketocyclazocine. The arylacetamide congener Cl-977 (enadoline), which displayed an 11-fold kappa-1 vs. kappa-all binding selectivity, was not sensitive to nor-BNI pretreatment. This indicates that the kappa subtype-binding profile of an agonist is not necessarily predictive of its sensitivity to nor-BNI in vivo. Overall, the present results suggest that at least two functional kappa receptor populations may be present in rhesus monkey brain.