1. Chinese hamster ovary (CHO) cells have been reported to be devoid of 5-HT receptors and have frequently been used as hosts for the expression of cloned 5-HT receptors. Unexpectedly, 5-HT was found to induce profound inhibition of forskolin-stimulated cyclic AMP production in these cells and the aim of this study was to classify the 5-HT receptor involved. 2. In CHO(dhfr-) cells 5-HT was a potent agonist and caused 80-100% inhibition of forskolin stimulated cyclic AMP production. A study using several 5-HT1 receptor agonists revealed the following potencies (p[A50]): RU24969 (9.09 +/- 0.17) > 5-carboxamidotryptamine (8.86 +/- 0.20) > 5-HT (8.07 +/- 0.05) > CP-93,129 (7.74 +/- 0.10) > sumatriptan (5.93 +/- 0.04). All five agonists achieved a similar maximum effect. Irreversible receptor alkylation studies yielded a pKA estimate of 7.04 +/- 0.34 for 5-HT. 3. The 5-HT1A/1B antagonist, (+/-)-cyanopindolol (4-100 nM), caused parallel rightward shifts of the 5-HT concentration-effect curve with no change in asymptote. Schild analysis yielded a pKB estimate of 8.69 +/- 0.09 (Schild slope 1.13 +/- 0.10). (+/-)-Cyanopindolol actually behaved as a partial agonist with an intrinsic activity of 0.2-0.5 and a p[A50] of 8.55. 4. 5-HT (0.01-10 microM) also elicited a concentration-dependent increase in intracellular [Ca2+] in CHO(dhfr-) cells thus demonstrating that dual coupling is not a phenomenon restricted to systems in which there is overexpression of transfected receptors. 5. This agonist and antagonist profile is consistent with the presence of a 5-HT1B receptor. 8-OH-DPAT (1 microM) and renzapride (3 microM) were without effect on forskolin-stimulated cyclic AMP production and ketanserin (0.3 microM) did not antagonize the inhibition produced by 5-HT, thus excluding the involvement of 5-HT1A, 5-HT4, and 5-HT2 receptors. 6. The possibility that expression of a 5-HT1B receptor was associated with the dhfr- mutation was excluded since RU24969, 5-HT and CP-93,129 were also potent agonists in unmutated, CHO-K1 cells: p[A50] 9.03 +/- 0.03, 8.34 +/- 0.05, 7.69 +/- 0.07 respectively, and (+/-)-cyanopindolol (0.1 microM) shifted the 5-HT curve to the right and yielded a pA2 estimate of 8.70 +/- 0.06. 7. Little or no specific binding of [3H]-5-HT (0.1-200 nM) or of the high affinity ligand [125I]-iodocyanopindolol (0.01-3 nM) to CHO(dhfr-) cell membranes could be detected. 5-HT also failed to elicit any increase in the binding of [35S]-GTP gamma S to CHO membranes. 8. In conclusion, cultured CHO cells express 5-HT1B receptors which are negatively coupled to adenylyl cyclase and positively coupled to increases in intracellular calcium. The absence of radioligand binding was unexpected in view of the high potency of 5-HT and the partial agonist activity of the normally 'silent' competitive antagonist, (+/-)-cyanopindolol. This implies very efficient receptor-effector coupling of a low density of 5-HT1B receptors. Clearly, the absence of detectable radioligand binding cannot be assumed to mean the absence of receptors capable of eliciting a significant functional response.