1. Antagonists at 5-HT3 receptors have shown activity in animal models of mental illness, however, few radiolabeled 5-HT3 ligands are available for preclinical studies. MIZAC, an analogue of the selective 5-HT3 antagonist, zacopride, binds with high affinity (1.3-1.5 nM) to CNS 5-HT3 sites. The authors report here the selectivity of MIZAC for these sites in rat brain homogenates. 2. Ninety-seven percent of total specific binding of [125I]MIZAC (0.1 nM) of was displaced by bemesetron (3 microM), a selective 5-HT3 antagonist. Competition studies using ligands with known affinities for 5-HT3 sites give a high correlation with reported pKi values (r2 0.98). Bemesetron displaceable binding has a regional distribution consistent with that of the 5-HT3 receptor, i.e. highest in cortex and hippocampus, and lowest in striatum and cerebellum. 3. Potent antagonists present at concentrations sufficient to occupy 95% of other 5-HT receptor populations (1A, 1B, 1D, 2A, 2B, 2C, 5A, 5B, 6, and 7) showed minimal ability to displace [125I]MIZAC binding (3 nM). Specificity studies using radioligand binding assays selective for 5-HT4, 5-HT6, and 5-HT7 receptors, and for binding sites of other neurotransmitters indicate a high degree of selectivity of [125I]MIZAC for the 5-HT3 receptor. 4. [125I]MIZAC binds to an apparent low affinity (benzac) site having a unique pharmacology. Low affinity binding was displaceable by benztropine, but not by other muscarinic agents nor inhibitors of dopamine uptake. The regional distribution of the low affinity site differed markedly from that of the high affinity site. The apparent affinity of [125I]MIZAC for the benzac site is two orders of magnitude lower than for the 5-HT3 receptor. Given its high selectivity for 5-HT3 binding sites, [125I]MIZAC appears to be a promising ligand for labeling 5-HT3 receptors in vitro and in vivo.