Our study was undertaken to characterize the functional properties of 5-hydroxytryptamine (5-HT)1D receptors in the rat midbrain raphe nuclei. In a first series of experiments, designed to assess whether 5-HT1D receptors are coupled to Gi/o proteins, the intracerebral injection of pertussis toxin into the dorsal raphe as well as incubation of midbrain raphe slices with the alkylating agent N-ethyl-maleimide (NEM) reduced the efficacy of the 5-HT1B/1D agonist sumatriptan to inhibit the electrically evoked overflow of [3H]5-HT from preloaded slices. Furthermore, preincubation with NEM also reduced the efficacy with which the 5-HT1B/1D antagonist GR 127935 enhanced evoked overflow of [3H]5-HT. These results indicate that, in rat midbrain raphe nuclei, 5-HT1D receptors are linked to Gi/o proteins. In an attempt to determine whether 5-HT1D receptors are located on 5-HT neurons, the inhibitory effect of sumatriptan and of the nonselective 5-HT agonist 5-carboxyamidotryptamine on K(+)-evoked overflow of [3H]5-HT was assessed in the presence of the Na+ channel blocker tetrodotoxin. Neither the inhibitory effect of sumatriptan nor that of 5-carboxyamidotryptamine were reduced by the addition of tetrodotoxin to the superfusion medium, suggesting that these 5-HT1D receptors are located on 5-HT neurons and may be considered autoreceptors. In a third series of experiments, rats were treated for 21 days either with the selective 5-HT reuptake inhibitor paroxetine (10 mg/kg/day, s.c.) or the reversible type A monoamine oxidase inhibitor befloxatone (0.75 mg/kg/day, s.c.) and superfusion experiments were performed after a 48-hr washout period. 5-HT1D receptors, similarly to 5-HT1A autoreceptors, desensitize after long-term treatment with a selective 5-HT reuptake inhibitor or a reversible type A monoamine oxidase inhibitor because the efficacy of sumatriptan and of 8-OH-DPAT to inhibit the electrically evoked overflow of [3H]5-HT was reduced after the administration of either drug.