1. Using membranes from stably or transiently transfected HEK293 cells cultured in 5-HT-free medium and expressing the recombinant human 5-HT(7) receptor splice variants (h5-HT(7(a)), h5-HT(7(b)) and h5-HT(7(d))), we compared their abilities to constitutively activate adenylyl cyclase (AC). 2. All h5-HT(7) splice variants elevated basal and forskolin-stimulated AC. The basal AC activity was reduced by the 5-HT(7) antagonist methiothepin and this effect was blocked by mesulergine (neutral 5-HT(7) antagonist) indicating that the inhibitory effect of methiothepin is inverse agonism at the 5-HT(7) receptor. 3. Receptor density correlated poorly with constitutive AC activity in stable clonal cell lines and transiently transfected cells. Mean constitutive AC activity as a percentage of forskolin-stimulated AC was significantly higher for the h5-HT(7(b)) splice variant compared to the h5-HT(7(a)) and h5-HT(7(d)) splice variants but only in stable cell lines. 4. All eight 5-HT antagonists tested inhibited constitutive AC activity of all splice variants in a concentration-dependent manner. No differences in inverse agonist potencies (pIC(50)) were observed between the splice variants. The rank order of potencies was in agreement and highly correlated with antagonist potencies (pK(b)) determined by antagonism of 5-HT-stimulated AC activity (methiothepin >metergoline> mesulergine > or = clozapine > or = spiperone > or = ritanserin > methysergide > ketanserin). 5. The efficacy of inverse agonism was not receptor level dependent and varied for several 5-HT antagonists between membrane preparations of transiently and stably transfected cells. 6. It is concluded that the h5-HT(7) splice variants display similar constitutive activity and inverse agonist properties.