The opiate agonists [3H]dihydromorphine (DHM, mu-selective ligand), [3H]bremazocine (potent kappa ligand), and [3H]etorphine bound stereospecifically, with high affinity, and reversibly to partially purified 3-(3-cholamidopropyl)dimethylammonio-1-propanesulfonate (CHAPS)-solubilized extract from rat brain membranes. Recoveries of the three binding activities were as follows: [3H]DHM, 47%; [3H]bremazocine, 55%; and [3H]etorphine, 17%. Each ligand exhibited (by Scatchard analysis) binding to a class of high-affinity sites (Kd = 0.8-2 nM). Hill analyses revealed Hill coefficients of n = 1.1-1.3. Many of the properties of solubilized brain opiate receptors are similar to those of membrane-associated opiate receptors. Opiate binding in soluble fractions was inhibited by a variety of protein-modifying agents, including trypsin, proteinase K, and N-ethylmaleimide as well as by heat treatment (60 degrees, 15 min). The relative potencies of a series of opiate narcotic agonists and antagonists in displacing 2 nM [3H]etorphine binding to the CHAPS-solubilized extract was similar to that determined for rat brain homogenates. In contrast, D-Ala2, D-Leu5-enkephalin (DADLE, putative delta-selective ligand) exhibited a much lower affinity for solubilized brain opiate receptors than for the membrane-bound receptors unless assayed in the presence of manganese chloride, sodium chloride, and GTP. Mu agonist binding to solubilized receptors was inhibited relatively selectively by sodium and guanyl nucleotides. These findings lend support to the pharmacological relevance of the solubilized opiate-binding component(s). The pI of the solubilized brain opiate receptor(s) was estimated by liquid isoelectrofocusing to be pH 4. The sizes of the solubilized, prelabeled [3H]etorphine-receptor complex of the solubilized mu and kappa receptor subtypes, as assayed by stereospecific binding of [3H]DHM and [3H]bremazocine binding, respectively, were estimated by molecular exclusion chromatography. The [3H]etorphine-receptor complex migrated as a broad radioactive peak at a position corresponding to a protein of Stoke radius 63 A. A secondary peak of radioactivity was observed at the salt peak. Mu receptor activity chromatographed as two major peaks. The first of these eluted just behind, but significantly separated from, the protein void peak and corresponded to a Stokes radius of 70 A; the second eluted just ahead of the salt peak and corresponded to a radius of less than 20 A. Kappa receptor activity eluted at positions corresponding to macromolecules of 50 A and less than or equal to 20 A. Together, these findings indicate that selective mu and kappa ligands interact with high molecular weight species of somewhat different sizes as well as a lower molecular weight species, which may represent a common subunit that can bind both ligands.