Putative 5-ht5 receptor binding sites were visualized by in vitro autoradiography using [125I]LSD (in the presence of clozapine and spiperone) or [3H]5-carboxamidotryptamine (in the presence 8-OH-DPAT, GR127935 and spiperone). Under these conditions, no [3H]5-carboxamidotryptamine labeling was detected in the brain of mice lacking the gene encoding the putative 5-ht5a receptor (knockout mice), whereas intermediate densities of binding sites were seen in the olfactory bulb and neocortex of wild-type mice. [125I]LSD labeled the same areas as [3H]5-carboxamidotryptamine in wild-type mice. High densities of [125I]LSD binding sites were observed in the medial habenula of wild type and knockout mice. 5-CT competed for [125I]LSD binding sites with an affinity of 2 nM in the olfactory bulb and neocortex of wild-type mice and an affinity of 30 nM in the habenula of knockout mice, suggesting that habenular labeling might be accounted for by putative 5-ht5b receptors. In the presence of 5'-guanylylimidodiphosphate, 5-CT displaced [125I]LSD from putative 5-ht5a and 5-ht5b sites with a 6-times and 3-times lower affinity, respectively, suggesting that both receptor subtypes are coupled to G proteins in brain. We also studied the inhibitory effect of 5-CT on dural neurogenic inflammation in knockout mice. In wild type mice, 3 ng/kg 5-CT inhibited dural protein extravasation by 60%. A similar effect was observed in knockout mice, even in the presence of the 5-HT1B receptor antagonist GR127935. These results suggest that the inhibitory effects of 5-CT are not mediated by a site with the characteristics of the putative 5-ht5 receptor.