A number of alpha-melanotropin (alpha-MSH) analogues have been designed de novo, synthesized, and bioassayed at different melanocortin receptors from frog skin (fMC1R) and mouse/rat (mMC1R, rMC3R, mMC4R, and mMC5R). These ligands were designed from somatostatin by a hybrid approach, which utilizes a modified cyclic structure (H-d-Phe-c[Cys---Cys]-Thr-NH(2)) related to somatostatin analogues (e.g. sandostatin) acting at somatostatin receptors, CTAP which binds specifically to micro opioid receptors, and the core pharmacophore of alpha-MSH (His-Phe-Arg-Trp). Ligands designed were H-d-Phe-c[XXX-YYY-ZZZ-Arg-Trp-AAA]-Thr-NH(2) [XXX and AAA = Cys, d-Cys, Hcy, Pen, d-Pen; YYY = His, His(1'-Me), His(3'-Me); ZZZ = Phe and side chain halogen substituted Phe, d-Phe, d-Nal(1'), and d-Nal(2')]. The compounds showed a wide range of bioactivities at the frog skin MC1R; e.g. H-d-Phe-c[Hcy-His-d-Phe-Arg-Trp-Cys]-Thr-NH(2) (6, EC(50) = 0.30 nM) and H-d-Phe-c[Cys-His-d-Phe-Arg-Trp-d-Cys]-Thr-NH(2) (8, EC(50) = 0.10 nM). In addition, when a lactam bridge was used as in H-d-Phe-c[Asp-His-d-Phe-Arg-Trp-Lys]-Thr-NH(2) (7, EC(50) = 0.10 nM), the analogue obtained is as potent as alpha-MSH in the frog skin MC1R assay. Interestingly, switching the bridge of 6 to give H-d-Phe-c[Cys-His-d-Phe-Arg-Trp-Hcy]-Thr-NH(2) (5, EC(50) = 1000 nM) led to a 3000-fold decrease in agonist activity. An increase in steric size in the side chain of d-Phe(7) reduced the bioactivity significantly. For example, H-d-Phe-c[Cys-His-d-Nal(1')-Arg-Trp-d-Cys]-Thr-NH(2) (24) is 2000-fold less active than 9. On the other hand, H-d-Phe-c[Cys-His-d-Phe(p-I)-Arg-Trp-d-Cys]-Thr-NH(2) (23) lost all agonist activity and became a weak antagonist (IC(50) = 1 x 10(-5) M). Furthermore, the modified CTAP analogues with a d-Trp at position 7 all showed weak antagonist activities (EC(50) = 10(-6) to 10(-7) M). Compounds bioassayed at mouse/rat MCRs displayed intriguing results. Most of them are potent at all four receptors tested (mMC1R, rMC3R, mMC4R, and mMC5R) with poor selectivities. However, two of the ligands, H-d-Phe-c[Cys-His-d-Phe-Arg-Trp-Pen]-Thr-NH(2) (9, EC(50) = 6.9 x 10(-9) M, 6.4 x 10(-8) M, 2.0 x 10(-8) M, and 1.4 x 10(-10) M at mMC1R, rMC3R, mMC4R, and mMC5R, respectively) and H-d-Phe-c[Cys-His(3'-Me)-d-Phe-Arg-Trp-Cys]-Thr-NH(2) (16, EC(50) = 3.5 x 10(-8) M, 3.1 x 10(-8) M, 8.8 x 10(-9) M, and 5.5 x 10(-10) M at mMC1R, rMC3R, mMC4R, and mMC5R, respectively) showed significant selectivities for the mMC5R. Worthy of mention is that neither of these two ligands is potent in the frog skin MC1R assay (EC(50) = 10(-7) M for 9 and EC(50) = 10(-5) M for 16). These results clearly demonstrated that binding behaviors in rodent MCRs are quite different from those in the classical frog skin (R pipiens) assay.