Polyclonal antibodies were raised against a synthetic hexadecapeptide corresponding to the portion of the second intracytoplasmic loop of the short form of the mouse 5-hydroxytryptamine-3A receptor subunit (5-HT3A-S), which differs from the long form (5-HT3A-L) by the removal of six amino acids. Antibodies were detected by enzyme-linked immunosorbent assay as soon as two months after the first injection to rabbits of the peptide coupled to keyhole limpet hemocyanin. Immunoblot detection of fusion proteins comprising glutathione-S-transferase and the second intracellular loop of 5-HT3A-S or 5-HT3A-L, and immunoprecipitation of cloned receptors showed that antibodies exhibited some selectivity for the short variant. Affinity chromatography allowed the purification of selective anti-5-HT3A-S antibodies which yielded a strong positive labeling of plasma membrane, reticulum and Golgi apparatus of COS-7 cells expressing murine 5-HT3A-S. In contrast, COS-7 cells expressing similar levels of 5-HT3A-L exhibited only a very weak labeling. Selectivity was also observed on immunoblots of cloned receptors transiently expressed in COS-7 cells, or stably expressed in CHO cells, both systems showing an immunolabeled component at 53,000-54,000 mol. wt. Immunoautoradiographic labeling of central nervous system sections showed that 5-HT3A-S-like immunoreactivity was found mostly within the nucleus of the solitary tract, the nucleus of the spinal tract of the trigeminal nerve, and the dorsal horn of the the spinal cord in the rat. After unilateral ablation of the nodose ganglion, 5-HT3A-S-like immunoreactivity decreased markedly in the ipsilateral part of the nucleus of the solitary tract, as expected of the presynaptic localization of 5-HT3 receptors. Finally, immunohistochemistry at the light and electron microscope levels revealed that 5-HT3A-S-like immunoreactivity was associated essentially with terminals and axonal profiles. All these results demonstrate that the immunolabeling exhibited by these antibodies is consistent with a specific and partially selective recognition of the short isoform of the 5-HT3A subunit. Because the pattern of immunoautoradiographic labeling matches the distribution previously established with selective radioligands, it can be inferred that these antibodies probably recognized the same fully assembled form of the 5-HT3A-S receptor subunit.