Previously, a 5-hydroxytryptamine receptor has been described in neuronal membrane preparations. This receptor corresponds to a recognition site for 5-HT, and is seemingly related to an effector which consists of an adenylate cyclase activated by 5-HT and having a high apparent affinity constant (KD close to 1 nM). Moreover, this neuronal receptor is distinct of the glial serotoninergic receptor (Fillion et al., 1980). The method used to measure the binding of 3H-5-HT has been previously reported (FILLION et al., 1978). Adenylate cyclase activity was determined by measurement of cAMP using a radioimmunoassay (FILLION et al., 1979 a, b). Membranes were pretreated as follows: (1) membranes were preexposed to 5-HT (i.e.: incubated in the presence of 5-HT, generally 10 nM) at 37 degrees C, washed by centrifugation and resuspended in a medium free of 5-HT (or diluted in the same medium) or (2) membranes were incubated with a single concentration of 5-HT for a prolonged duration. Results were as follows: 1. The binding of 3H-5-HT, incubated for 10 min at 37 degrees C, corresponds to a KD close to 20 nM whereas a prolongation of the incubation duration to 15-25 induces the binding of 5-HT with a KD close to 2 nM. 2. Preexposure of the membranes to 5-HT leads to the same increase in affinity with, apparently, a reduced number of sites. 3. N-ethylmaleimide pretreatment of the membranes prior to their preexposure to 5-HT inhibits the change in affinity. 4. GTP, but not GDP, reverses the affinity change. 5. The adenylate cyclase activated by 5-HT is desensitized by preexposure of the membranes to 5-HT. 6. GTP reverses the process of desensitization. These results support the hypothesis of the existence of a regulatory mechanism for the 5-HT neuronal receptor. The regulatory mechanism would involve structural conformation changes of the recognition site corresponding to different affinities of the adenylate cyclase. A nucleotide binding protein is presumably involved in the coupling of these two subunits of the receptor.