Melatonin receptors were studied in isolated mouse hepatocytes using the 2[125I]iodomelatonin binding assay. The binding of 2[125I]iodomelatonin to hepatocytes isolated from the mouse using collagenase was stable, saturable, reversible and of high affinity. The equilibrium dissociation constant (Kd) obtained from saturation studies was 10.0 ± 0.4 pmol/l (n = 16), which was comparable to the Kd obtained from kinetics studies (6.9 ± 1.2 pmol/l, n = 3), and the maximum number of binding sites (Bmax) was 2.9 ± 0.4 fmol/mg protein (n = 16). The relative order of potency of indoles in competing for 2[125I]iodomelatonin binding was 2-iodomelatonin > 2-phenylmelatonin > 6-chloromelatonin > melatonin > 6-hydroxymelatonin > N-acetylserotonin, indicating that the binding was mediated by the ML1 receptor subtype. The linear Rosenthal plots, the close proximity of the Hill coefficient to unity and the monophasic competition curves suggest that a single class of 2[125I]iodomelatonin binding sites is present in the mouse hepatocytes. Guanosine 5′-O-(3-thiotriphosphate) dose-dependently inhibited 2[125I]iodomelatonin by lowering the affinity of binding, while no inhibitory effects of adenosine nucleotides were observed, suggesting that the binding sites are G-protein linked. Western immunoblotting was used to identify the melatonin receptor subtype in mouse hepatocytes using anti-Mel1a and anti-Mel1b. Hepatocyte membrane extract reacted with anti-Mel1b but not anti-Mel1a giving a peptide-blockable band of 36 kD, supporting the hypothesis that the melatonin receptors in mouse hepatocytes are of the Mel1b subtype. Melatonin injection and a high plasma glucose level affected 2[125I]iodomelatonin binding in the whole mouse liver homogenates. Plasma glucose was elevated by mid-light intraperitoneal injection of melatonin (4 and 40 mg/kg body weight) in a dose-dependent manner with maximum elevation achieved 1 h after injection. 2[125I]Iodomelatonin binding at this time showed increased Kd with no changes in Bmax. When the plasma glucose returned to normal within 2 h, the binding remained lowered with increased Kd but no changes in Bmax. Elevation of plasma glucose by 2-deoxyglucose injection (500 mg/kg), on the other hand, decreased the binding by decreasing the Bmax without affecting the Kd. Suppression of plasma glucose by insulin injection (3 IU/kg) did not change the binding. Thus, melatonin may act directly on the liver to elevate the plasma glucose level, and changes in plasma glucose level itself may in turn affect hepatic melatonin binding.