A substance having benzodiazepine-binding inhibitory activity has been extracted from 18 kg of gray matter of bovine cerebral cortex and purified to homogeneity. This substance Inhibits competitively PHlflunitrazepam and ethyl The demonstration that benzodiazepines (BZDs) bind specifically and with high affinity to receptors in the central nervous system (1, 2) led to the search for a possible endogenous ligand that could recognize such receptor sites and modulate y-aminobutyric acid (GABA) neurotransmission. A number of substances extracted from brain were found to displace BZD from the receptor; however, in most cases the binding had low affinity-for example, purines (3, 4) and nicotinamide (5) had a Ki in the millimolar range. A number ofpeptides and proteins ofMr 1500-70,000 and other nonidentified substances were also isolated from brain (6-9). The best-characterized polypeptide is the diazepam-binding inhibitor, which is anxiogenic, has a Mr of 11,000, and has a relatively weak affinity (10). This substance has been submitted to controlled hydrolysis and an 18-amino acid fragment was isolated, which keeps the anxiogenic properties while partially losing affinity for the BZD receptor (11). So far, the most active compound isolated from urine and brain is ethyl ,-carboline-3-carboxylate (P-CCE), which has a Ki in the nanomolar range, similar to some BZDs (12). However, 3-CCE was found to be formed during the extraction in a hot ethanol HCO step. While recognizing that it was an artifact, Braestrup et al. (12) suggested that the true endogenous ligand could be a derivative of f-carboline-3-carboxylic acid. The complexities encountered in trying to demonstrate the existence of an endogenous ligand for the BZD receptor (see ref. 13) have led some investigators to disregard the presence of an endogenous ligand and to attribute the effect of BZDs only to their modulatory action on the GABA receptor (14).In the preliminary experiments, we used an aqueous extraction of rat or bovine cerebral cortices and found a heat-stable substance of Mr <1000, which specifically andto central BZD receptors (15). This substance was later extracted from larger quantities of bovine cerebral cortex by using a mild acetic acid and methanol extraction, followed by Sephadex G-25 filtration, countercurrent distributions, and several rounds of preparative and analytical high-performance liquid chromatography (HPLC) (16). We could detect several fractions having BZD binding inhibitory activity, one of which was prominent. This particular inhibitor is specific for the central BZD receptor and displaces [3H]FNZ binding with a Ki in the nanomolar range. Furthermore, it has no effect on other receptors such as the a1-, a2-, and 3-adrenoceptors; the cholinergic muscarinic and the GABAA receptors; as well as on the "peripheral" BZD binding site.In the present work, we describe the experiments done to identify the chemical nature of this major and most-purified peak of BZD inhibitory activity that we have isolated from bovine brain (16...