This study examined the activity of chemically diverse alpha2 adrenoceptor ligands at recombinant human (h) and native rat (r) alpha2A adrenoceptors compared with 5-HT1A receptors. First, in competition binding experiments at h alpha2A and h5-HT1A receptors expressed in CHO cells, several compounds, including the antagonists 1-(2-pyrimidinyl)piperazine (1-PP), (+/-)-idazoxan, benalfocin (SKF 86466), yohimbine and RX 821,002, displayed preference for h alpha2A versus h5-HT1A receptors of only 1.4-, 3.6-, 4-, 10- and 11-fold, respectively (based on differences in pKi values). Clonidine, brimonidine (UK 14304), the benzopyrrolidine fluparoxan and the guanidines guanfacine and guanabenz exhibited intermediate selectivity (22- to 31-fold) for h alpha2A receptors. Only the antagonist atipamezole and the agonist dexmedetomidine (DMT) displayed high preference for alpha2 adrenoceptors (1290- and 91-fold, respectively). Second, the compounds were tested for their ability to induce h5-HT1A receptor-mediated G-protein activation, as indicated by the stimulation of [35S]GTPgammaS binding. All except atipamezole and RX 821,002 exhibited agonist activity, with potencies which correlated with their affinity for h5-HT1A receptors. Relative efficacies (Emax values) were 25-35% for guanabenz, guanfacine, WB 4101 and benalfocin, 50-65% for 1-PP, (+/-)-idazoxan and clonidine, and over 70% for fluparoxan, oxymetazoline and yohimbine (relative to 5-HT = 100%). Yohimbine-induced [35S]GTPgammaS binding was inhibited by the selective 5-HT1A receptor antagonist WAY 100,635. In contrast, RX 821,002 was the only ligand which exhibited antagonist activity at h5-HT1A receptors, inhibiting 5-HT-stimulated [35S]GTPgammaS binding. Atipamezole, which exhibited negligeable affinity for 5-HT1A receptors, was inactive. Third, the affinities for r alpha2A differed considerably from the affinities for h alpha2A receptors whereas the affinities for r5-HT1A differed much less from the affinities for h5-HT1A receptors. This affected markedly the affinity ratios of certain compounds. For example, (+/-)-idazoxan was only 3.6-fold selective for h alpha2A versus h5-HT1A but 51-fold selective for r alpha2A versus r5-HT1A receptors. Conversely, yohimbine was tenfold selective for h alpha2A versus h5-HT1A adrenoceptors but 4.2-fold selective for r alpha2A versus r5-HT1A receptors. Nevertheless, both atipamezole and DMT were highly selective for both rat and human alpha2A versus rat or human 5-HT1A receptors. In conclusion, these data indicate that (1) the agonist DMT and the antagonist atipamezole are the ligands of choice to distinguish alpha2-mediated from 5-HT1A-mediated actions, whilst several of the other compounds show only low or modest selectivity for alpha2A over 5-HT1A receptors; (2) caution should be exercised in experimental and clinical interpretation of the actions of traditionally employed alpha2 ligands, such as clonidine, yohimbine and (+/-)-idazoxan, which exhibit marked agonist activity at 5-HT1A receptors.