1. The interaction of melatonin (N-acetyl-5-methoxytryptamine) with 5-hydroxytryptamine4 (5-HT4) receptors and/or with melatonin receptors (ML1, ML2 sites) has been assessed in isolated strips of the guinea-pig proximal colon. In the same preparation, the pharmacological profile of a series of melatonin agonists (2-iodomelatonin, 6-chloromelatonin, N-acetyl-5-hydroxytryptamine (N-acetyl-5-HT), 5-methoxycarbonylamino-N-acetyltryptamine (5-MCA-NAT)) was investigated. 2. In the presence of 5-HT1/2/3 receptor blockade with methysergide (1 microM) and ondansetron (10 microM), melatonin (0.1 nM-10 microM), 5-HT (1 nM-1 microM) and the 5-HT4 receptor agonist, 5-methoxytryptamine (5-MeOT: 1 nM-1 microM) caused concentration-dependent contractile responses. 5-HT and 5-MeOT acted as full agonists with a potency (-log EC50) of 7.8 and 8.0, respectively. The potency value for melatonin was 8.7, but its maximum effect was only 58% of that elicited by 5-HT. 3. Melatonin responses were resistant to atropine (0.1 microM), tetrodotoxin (0.3 microM), and to blockade of 5-HT4 receptors by SDZ 205,557 (0.3 microM) and GR 125487 (3, 30 and 300 nM). The latter antagonist (3 nM) inhibited 5-HT-induced contractions with an apparent pA2 value of 9.6 GR 125487 antagonism was associated with 30% reduction of the 5-HT response maximum. Contractions elicited by 5-HT were not modified when melatonin (1 and 10 nM) was used as an antagonist. 4. Like melatonin, the four melatonin analogues concentration-dependently contracted colonic strips. The rank order of agonist potency was: 2-iodomelatonin (10.8) > 6-chloromelatonin (9.9) > or = N-acetyl-5-HT (9.8) > or = 5-MCA-NAT (9.6) > melatonin (8.7), an order typical for ML2 sites. In comparison with the other agonists, 5-MCA-NAT had the highest intrinsic activity. 5. The melatonin ML1B receptor antagonist luzindole (0.3, 1 and 3 microM) had no effect on the concentration-response curve to melatonin. Prazosin, an alpha-adrenoceptor antagonist possessing moderate/ high affinity for melatonin ML2 sites did not affect melatonin-induced contractions at 0.1 microM. Higher prazosin concentrations (0.3 and 1 microM) caused a non-concentration-dependent depression of the maximal response to melatonin without changing its potency. Prazosin (0.1 and 1 microM) showed a similar depressant behaviour towards the contractile responses to 5-MCA-NAT. 6. In the guinea-pig proximal colon, melatonin despite some structural similarity with the 5-HT4 receptor agonist 5-MeOT, does not interact with 5-HT4 receptors (or with 5-HT1/2/3 receptors). As indicated by the rank order of agonist potencies and by the inefficacy of luzindole, the most likely sites of action of melatonin are postjunctional ML2 receptors. However, this assumption could not be corroborated with the use of prazosin as this 'ML2 receptor antagonist' showed only a non-concentration-dependent depression of the maximal contractile response to both melatonin and 5-MCA-NAT. Further investigation with the use of truly selective antagonists at melatonin ML2 receptors is required to clarify this issue.