Melanin-concentrating hormone (MCH) is a cyclic nona- Melanin-concentrating hormone (MCH)1 has been initially described in fish as a heptadecapeptide (Asp-Thr-Met-Arg-CysMet-Val-Gly-Arg-Val-Tyr-Arg-Pro-Cys-Trp-Glu-Val (1)). Its structure was relatively conserved throughout evolution, although in mammals the sequence of MCH is a nonadecapeptide with differences mainly in the N terminus (Asp-Phe-Asp-MetLeu-Arg-Cys-Met-Leu-Gly-Arg-Val-Tyr-Arg-Pro-Cys-Trp-GlnVal (2)). In rodents, there are now several lines of evidence for the involvement of MCH in the central regulation of feeding behavior as reviewed by Tritos and Maratos-Flier (3). The MCH peptide and its receptor are expressed in the hypothalamus, a region involved in energy balance and food intake (4 -7). In this particular brain area, MCH mRNA is overexpressed and up-regulated during fasting in ob/ob mice as well as in rats (8, 9). Intra-cerebroventricular injections of MCH promote feeding in mice and rats (9 -12). Finally, transgenic mice lacking the MCH gene are lean and hypophagic (13). Interestingly, in peripheral tissues, MCH also stimulates the release of leptin from isolated rat adipocytes (14). The lack of suitable binding conditions, mainly due to the hydrophobic and sticky nature of MCH itself or derivatives (15, 16), was probably a limitation for expression cloning of the receptor. The MCH receptor was nevertheless recently identified by several groups using reverse pharmacology (17-21). The MCH function was assigned to the previously described orphan receptor SLC-1 (22, 23), using inhibition of forskolin-stimulated cAMP production and induction of calcium rise.Receptor cloning and association of functional tests open the way to the search for pharmacological tools, especially receptor antagonists that are needed to study receptor functions. One of the possible strategies to this goal is the chemical modification of the natural peptide including peptide shortening, amino acid substitution, and conformation restriction with the help of structure-activity relationships and modeling studies toward optimized nonpeptide ligands. Such a strategy has been successful for the design of subtype-specific antagonists of neuropeptide Y receptors (24 -27).In the case of MCH, only two sets of data have been published on the pharmacological action and binding affinity of MCH analogues in vitro. A first series of experiments with fish MCH on fish, frog, or other batrachian skin assays were reported (28 -32), showing that fish MCH could be shortened at * This work was supported by a a Convention CIFRE between the Association Nationale de la Recherche Technique, the Institut de Recherches SERVIER and the Centre National de la Recherche Scientifique (to T. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.*

1 Shortened, more stable and weakly hydrophobic analogues of melanin-concentrating hormone (MCH) were searched as candidates for radioiodination. Starting from the dodecapeptide MCH 6 ± 17 , we found that: (1) substitution of Tyr 13 by a Phe residue; (2) addition of a 3-iodo-Tyr residue at the Nterminus; and (3) addition of a hydrophilic spacer 8-amino-3,6-dioxyoctanoyl between the 3-iodo-Tyr and MCH 6 ± 17 (compound S36057), led to an agonist more potent than MCH itself in stimulating

The screening of combinatorial libraries requires a deconvolution procedure to obtain, in fine, the most active compound of the starting library. The standard screening assays used in regular molecular pharmacology, have been poorly assessed when transposed to combinatorial chemistry-related experiments, particularly those involving large numbers of chemicals in a single assay. One key issue is the effect of the inactive analogs on the identification of the active ligand in mixtures. We chose melatonin receptors to measure the apparent affinity of a single ligand when tested alone or in mixtures of non-peptide low molecular weight compounds. Using ligands with IC50 from the micro- to the picomolar range, mixed with increasingly complex mixtures of 5 to 20 or 25 inactive compounds, we analyzed the displacements from the mt1 and MT2 melatonin receptor subtypes of the radioligand 2-iodomelatonin (KD= 25 pmol/l and 200 pmol/l, respectively) . The behavior of equimolar mixtures in displacement curves led to the conclusion that the observed binding affinity reflects the dilution effect of mixing the active component with inactive compounds but does not reveal noticeable interactions which would interfere with the binding process. From the practical point of view, the concentrations of the active species in the binding assay should be large enough to displace significantly the radioligand, a requirement which may be limited by the solubility of the ligand mixtures. In contrast, previous observations with peptide libraries report that the dilution effect is often compensated by additive or synergic action of structurally related analogs, thus making possible the deconvolution of very large (typically up to 10(7) compounds) peptide libraries.

No abstract