Following the cloning and sequencing of the A subunit of the 5-HT3 receptor, two alternatively spliced isoforms, 5-HT3-AS and 5-HT3-AL, have been identified. In order to analyse the distribution of the receptor, a polyclonal antibody has been produced against the short form which is the most abundant in the central nervous system [Doucet et al. (2000) Neuroscience 95, 881-892]. As expected from the recognition of functional 5-HT3 receptors, immunostaining by this anti-5-HT3-R-AS antibody matched the distribution of the high-affinity 5-HT3 binding sites in the rat brain and spinal cord. 5-HT3-AS-like immunoreactivity was detected at low levels in the limbic system, particularly in the amygdala and the hippocampus, and in the frontal, piriform and entorhinal cortices. High levels of immunoreactivity were found in the brainstem, mainly in the nucleus tractus solitarius and the nucleus of the spinal tract of the trigeminal nerve, and in the dorsal horn of the spinal cord. At the ultrastructural level, immunostaining was generally found associated with axons and nerve terminals (70-80%) except in the hippocampus, where labelled dendrites were more abundant (56%). This preferential localization on nerve endings is consistent with the well-documented physiological role of 5-HT3 receptors in the control of neurotransmitter release. However, the different distribution in the hippocampus raises the question of whether differential addressing mechanisms exist for preferentially targeting 5-HT3 receptors to postsynaptic dendritic sites as compared to presynaptic nerve endings, depending on the nature of the neurons bearing these receptors.