Common Requirements for Melanocortin-4 Receptor Selectivity of Structurally Unrelated Melanocortin Agonist and Endogenous Antagonist, Agouti Protein*

The activity of melanocortin receptors (MCR) is regulated by melanocortin peptide agonists and by the endogenous antagonists, Agouti protein and AgRP (Ag-outi-related protein). To understand how the selectivity for these structurally unrelated agonists and antagonist is achieved, chimeric and mutants MC3R and MC4R were expressed in cell lines and pharmacologically analyzed. A region containing the third extracellular loop, EC3, of MC4R was essential for selective Agouti protein antagonism. In addition, this part of MC4R, when intro-duced in MC3R, conferred Agouti protein antagonism. Further mutational analysis of this region of MC4R demonstrated that Tyr 268 was required for the selective interaction with Agouti protein, because a profound loss of the ability of Agouti protein to inhibit 125 I-labeled [Nle 4 , D -Phe 7 ] a -melanocyte-stimulating hormone (MSH) binding was observed by the single mutation of Tyr 268 to Ile. This same residue conferred selectivity for the MC4R selective agonist, [ D -Tyr 4 ]MT-II, whereas it inhib-ited action and antagonism. To this end, chimeric and mutant and were cell lines and were tested interaction with Agouti protein, In addition, the results show that selectivity for agonists and antagonist achieved in a similar manner.

The activity of melanocortin receptors (MCR) is regulated by melanocortin peptide agonists and by the endogenous antagonists, Agouti protein and AgRP (Agouti-related protein). To understand how the selectivity for these structurally unrelated agonists and antagonist is achieved, chimeric and mutants MC3R and MC4R were expressed in cell lines and pharmacologically analyzed. A region containing the third extracellular loop, EC3, of MC4R was essential for selective Agouti protein antagonism. In addition, this part of MC4R, when introduced in MC3R, conferred Agouti protein antagonism. Further mutational analysis of this region of MC4R demonstrated that Tyr 268 was required for the selective interaction with Agouti protein, because a profound loss of the ability of Agouti protein to inhibit 125  Melanocortin (MC) 1 receptors are activated by the POMC (pro-opiomelanocortin)-derived ACTH (adrenocorticotropic hormone) and MSH (melanocyte-stimulating hormone) peptides. MC3R and MC4R are the main MC receptors in the brain, and MC4R is thought to play a prominent role in the regulation of body weight in both rodents and human (1)(2)(3). The identification of the endogenous MC receptor antagonists, Agouti protein and Agouti related-protein (AgRP), gave rise to an additional level of regulation of MC receptors, in which ligands with opposite activities control MCR activity (4,5). It is unknown, however, whether the molecular interaction with the receptor of these oppositely acting ligands is achieved in a similar manner.
Mouse Agouti protein is a 131-amino acid protein normally expressed in hair follicles, which acts on MC1R-expressing melanocytes to regulate pigmentation (6,7). In mice that ectopically overexpress Agouti protein, chronic MC1R blockade by Agouti protein results in a yellow coat color. In addition, these mice display severe obesity, hyperphagia, and increased plasma levels of the adipose derived satiety factor, leptin (8,9). In these mice, obesity is thought to be the result of continuous blockade of the hypothalamic MC4R, since recombinant murine Agouti protein has been demonstrated to act as a high affinity antagonist for both mouse MC1R and MC4R in vitro but not for MC3R and MC5R (10 -12). This determination is in agreement with the finding that MC4RϪ/Ϫ mice recapitulate the obesity phenotype as observed in yellow obese mice (2).
In wild-type mice, Agouti protein is not expressed in the brain. But its homologue, AgRP, is expressed in the hypothalami of mice, rats, primates, and humans (13)(14)(15)(16). Recombinant human AgRP acts as a high affinity antagonist for the MC3R and MC4R, and to a lesser extent for MC5R, but not for MC1R and MC2R (17). Transgenic mice that overexpress AgRP display an obesity phenotype similar to that found in MC4RϪ/Ϫ mice (4). Thus, pharmacological, histochemical, and genetic studies suggest that hypothalamic AgRP is an important endogenous stimulator of feeding and exerts this function by inhibiting MCR signaling.
Even though an important role of MC4R in body weight homeostasis is evident, the role of MC3R in the control of body weight and other processes is ill-defined because of the absence of MC3R selective ligands. Knowledge of the interaction between MCR and its ligands at the level of molecular detail would contribute to the design of selective MC3R and MC4R ligands. This is important not only for the understanding of MCR subtype-specific functions, but in addition, ligands that selectively activate or block MC4R may be therapeutically useful in the treatment of obesity (18) and anorexia (19,20).
To investigate whether MCR agonists and the structurally unrelated antagonist, Agouti protein, are regulated in a similar manner, the aim of this study was to identify which part of the human MC4R is important for selective Agouti protein inter-* 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. ]Lys-␥ 2 -MSH were synthesized using solid phase Fmoc chemistry and purified as described previously (21). Fulllength human Agouti protein and AgRP were prepared as described previously (17,22).
Radioiodination of NDP-␣-MSH-Iodination was performed exactly as described previously (23). In short, 4 g of NDP-␣-MSH was mixed with 1.2 IU bovine lactoperoxidase (Calbiochem) and 1 mCi of Na 125 I (ICN) in a final volume of 100 l of 0.05 M phosphate buffer (pH 6.5). Then, 5 l of 0.003% H 2 O 2 was added every 60 s. After 4 min, 50 l of 1 mM dithiothreitol was added to stop the reaction. The sample was purified by high pressure liquid chromatography with a Bondapak C 18 column, 3.9 ϫ 300 mm (Waters, Div. of Millipore) by elution with a 22-52% acetonitrile gradient in 10 mM ammonium acetate (pH 5.5) in 40 min. The specific activity of 125 I-NDP-␣-MSH was 2.25 ϫ 10 6 Ci/mol.
Adenylate Cyclase Assay-B16G4F cells, stably expressing human MC4R, rat MC3R, and chimerae 3AB, 3B, 3C, 3D, or 4D were grown in 24-well plates (Corning Costar). Agonist-stimulated adenylate cyclase activity was measured as described previously (26). In short, after prelabeling with 500 l of [ 3 H]adenine (PerkinElmer Life Sciences) in a concentration of 2 Ci/ml, the cells were incubated for 30 min at 37°C in Dulbecco's modified Eagle's medium containing 0.1 mM isobutylmethylxanthine, ␣-MSH in a concentration ranging from 0.3 nM to 3 M, with and without 40 nM Agouti protein. The cells were harvested, and [ 3 H]cAMP formation was calculated as the percentage of [ 3 H]ATP converted into [ 3 H]cAMP. For each curve, 12 duplicate data points were collected. EC 50 values were determined by fitting the data to a sigmoidal curve with variable slope using GraphPad Prism 2.01 for Windows 95/NT (GraphPad Software Inc., San Diego, CA). Experiments were performed two times with the same results.
Receptor Binding Assay-MC receptor expressing B16G4F, 293 HEK cells, or BHK cells were grown in 24-well plates (Corning Costar). The cells were incubated with 100,000 cpm (Ϸ0.1 nM) of 125 I-NDP-␣-MSH and various concentrations of nonradioactive peptides diluted in binding buffer consisting of Ham's F-10 medium (Life Technologies, Inc.) (pH 7.4), 2.5 mM calcium chloride, 0.25% bovine serum albumin, 10 mM Hepes, and 50 g/ml aprotinin (Sigma). After a 30-min incubation at room temperature, the cells were washed twice with ice-cold Trisbuffered saline containing 2.5 mM calcium chloride and lysed in 1 M sodium hydroxide. Radioactivity of the lysates was counted in a Packard Cobra ␥-counter. Competition curves were fitted from 12 duplicate data points with GraphPad Prism 2.01 for Windows 95/NT, nonlinear regression, one-site competition. IC 50 values were calculated with 99% confidence interval. Experiments were repeated at least two times with the same results.

RESULTS
The affinities (IC 50 values, using 125 I-NDP-␣-MSH as radioligand) of full-length human Agouti protein, AgRP, and NDP-␣-MSH were determined for the wild-type MC4R and MC3R expressed in B16G4F cells. NDP-␣-MSH and AgRP possessed comparable IC 50 values for both MC3R and MC4R (Fig. 2). In contrast, Agouti protein displayed an almost 40-fold lower affinity for MC3R than AgRP and NDP-␣-MSH, whereas Agouti The gray circles indicate a stretch of residues with complete homology between MC3R and MC4R that were used as boundaries for construction of the chimeric MC3/MC4 receptors. The larger circles indicates residues belonging to the third extracellular loop (EC3, residues 267-282). The black circles with doubleresidue symbols (for example, Y/I at position 268) indicate that the first residue (Y) is present in MC4R and the second residue (I) is present in MC3R at the corresponding position. The white circles in EC3 are homologous in MC4R and MC3R. The numbering corresponds to the human MC4R amino acid sequence. The N-and C-terminal residues are not shown.
protein displayed high affinity for MC4R. The IC 50 of Agouti protein was 10-fold higher for MC3R than for MC4R. 2 Next, the antagonistic properties of Agouti protein were analyzed. Agouti protein at a concentration of 40 nM was a potent antagonist for MC4R (Fig. 3) and was able to increase the EC 50 of ␣-MSH almost 20-fold (Table I). In contrast, the same concentration of Agouti protein was not able to significantly alter the EC 50 of ␣-MSH for MC3R.
In addition, it was investigated as to which region of MC4R is required for the antagonistic properties of Agouti protein. Therefore, the ability of Agouti protein to antagonize ␣-MSHstimulated adenylate cyclase activity of chimeric MC3/MC4 receptors was analyzed. Agouti protein was able to effectively antagonize ␣-MSH-stimulated adenylate cyclase activity for all chimeric receptors except chimera 3D, which only has the Cterminal portion of MC3R starting at TM5 (Fig. 3 and Table I). Furthermore, the reverse chimera, 4D, which is an MC3R with only the C-terminal portion (starting at TM5) of MC4R, was able to confer Agouti protein antagonism to MC3R. Agouti protein was unable to significantly increase the EC 50 of ␣-MSH for chimera 3D as observed for MC3R.
To assess which region of MC4R determined the high affinity of Agouti protein for MC4R, the IC 50 values of Agouti protein for chimeric MC3/MC4 receptors was analyzed. Table II summarizes the IC 50 values of Agouti protein for all chimeric receptors. Chimerae 3AB, 3B, and 3C all displayed a high affinity for Agouti, which was not significantly different from the IC 50 for MC4R. However, chimera 3D (MC4R with the C-terminal portion of MC3R, starting at TM5), displayed a significant reduction in the affinity for Agouti protein. The affinities of Agouti protein for chimera 3D and MC3R were comparable (Fig. 4). Conversely, chimera 4D (the reverse of 3D, MC3R with the C-terminal portion of MC4R) displayed high affinity for

Melanocortin-4 Receptor Selectivity
Agouti protein, which was not different from wild-type MC4R.
Next, the role of EC3 of MC4R in interaction with Agouti protein was further specified. To this end, wild-type MC4R, mutant MC4(Tyr 268 3 Ile), and chimeric MC4R with small parts of EC3 of MC3R were expressed in 293 HEK cells. The chimera were: loop (MC4R with EC3 from MC3R), 1st half (MC4R with first half of EC3 from MC3R), and 2nd half (MC4R with second half of EC3 from MC3R). Fig. 5 and Table III show that the affinities of NDP-␣-MSH for the MC4R, chimeric, and mutant receptors were comparable. However, when EC3 of MC4R was replaced by MC3R sequence (loop), the IC 50 for Agouti protein increased more than 4-fold as compared with MC4R. This loss of affinity was also observed when only the first half of EC3 was replaced (1st half). Moreover, when only Tyr 268 in the first half of EC3 of MC4R was mutated to Ile, the corresponding residue of MC3R, a more than 10-fold drop in Agouti protein affinity was still observed. In contrast, replacement of the second half of EC3 (2nd half chimera) did not alter the affinity for Agouti protein as compared with wild-type MC4R. Next, MC3(Ile 265 3 Tyr), the reverse mutant of MC4 (Tyr 268 3 Ile), was analyzed by determining the IC 50 values of NDP-␣-MSH, Agouti protein, [Nle 4 ]Lys-␥ 2 -MSH, and [D-Tyr 4 ]MT-II affinity ( Fig. 6 and Table IV). The latter two peptides were shown previously to display increased and decreased affinity, respectively, for MC4 (Tyr 268 3 Ile) (21). The data show that the IC 50 of NDP-␣-MSH and Agouti protein for MC3R and MC3(Ile 265 3 Tyr) did not significantly differ. However, the affinity of [Nle 4 ]Lys-␥ 2 -MSH was ϳ40 times lower for the mutant as compared with MC3R. In contrast, the affinity of [D-Tyr 4 ]MT-II was significantly higher for MC3(Ile 265 3 Tyr).

DISCUSSION
The melanocortin system is unique in the sense that MCR activity is regulated by two structurally unrelated endogenous peptides with opposing activities: on the one hand, POMCderived melanocortin agonists, and on the other hand, Agouti protein and AgRP, which act as antagonists for MCR. Cur-rently, selective MCR ligands are necessary to obtain a better understanding of the role of brain MC3R and MC4R in processes like the regulation of body weight. Because Agouti protein is a high affinity antagonist for the MC4R, this study was aimed at characterizing the selective interaction of human Agouti protein with the human MC4R. The pharmacological analysis of chimeric and mutant MC3R and MC4R suggested that a single amino acid residue in the first half of the third extracellular loop of MC4R conferred selectivity for Agouti protein.
The results confirm that human Agouti protein is a high affinity antagonist for MC4R but not for MC3R. In contrast, AgRP does not discriminate between MC3R and MC4R. The selective interaction of Agouti protein with MC4R was studied using chimeric and mutant MC3R and MC4R. Apparently, the lower affinity of Agouti protein for chimera 3D (MC4R with EC3 and surrounding domains of MC3R) and MC3R wild type accounts for the inability of Agouti protein to antagonize ␣-MSH action on these receptors. Chimera 4D was tested for a gain of function for Agouti binding and antagonism to test whether this region (EC3 and surrounding domains of MC4R)   alone is sufficient to obtain an MC4R-specific pharmacological profile. Indeed, this EC3-containing region of MC4R fully conferred Agouti protein antagonism to MC3R. Thus, this region changes the MC3R pharmacological profile to that of MC4R.
A more detailed analysis of the role of EC3 in Agouti protein binding showed that the first half of EC3, in particular position Tyr 268 of MC4R, was critical for high affinity Agouti protein binding because, when it was replaced by Ile, a loss of affinity for Agouti protein occurred. Previously, it was shown that Tyr 268 was also required for the selective interaction of the agonist [D-Tyr 4 ]MT-II with MC4R, whereas it hindered interaction with the MC3R selective agonist [Nle 4 ]Lys-␥ 2 -MSH (21). Thus, Tyr 268 is required for selective agonist and antagonist binding of the MC4R. This is a striking observation because there exists no obvious amino acid homology between melanocortins and Agouti protein (Fig. 7).
To test whether Ile 265 of MC3R hindered interaction with Agouti protein, Ile 265 of MC3R was mutated into Tyr. MC3 (Ile 265 3 Tyr) did not show a gain of affinity for Agouti protein, indicating that Tyr 268 , in the context of the MC3R conformation, does not confer selectivity for antagonists. Surprisingly, MC3 (Ile 265 3 Tyr) displayed increased affinity for [D-Tyr 4 ]MT-II and decreased affinity for [Nle 4 ]Lys-␥ 2 -MSH. Thus, this residue is able to change the pharmacological profile of MC3R toward that of MC4R for agonists but not for antagonists. These data imply that it may not be feasible to design MC3Rselective antagonists based upon MC3R selective agonists (such as ␥-MSH) and MC4R-selective antagonists. This approach, however, may be applicable for MC4R because the structural requirements for selective agonists and antagonists interaction with MC4R is more alike.
In conclusion, through pharmacological analysis of chimeric and mutant MC3R and MC4R, it was demonstrated that Tyr 268 in EC3 of MC4R is critical for selective Agouti protein interaction. This study and previous data show that the mutation of MC4R Tyr 268 to Ile decreases affinity for MC4R-selective ligands but increases affinity for the MC3R-selective ligands. Moreover, Tyr 268 in MC4R determined both agonist and antagonist selectivity. This is the first report describing details of the molecular recognition of the endogenous MCR antagonist, Agouti protein, by MC4R. This report demonstrates that the selectivity of structurally unrelated peptide ligands with opposite activities (melanocortin peptide agonist versus Agouti protein antagonist) is achieved in a similar manner, strongly suggesting that they use the same binding pocket. An understanding of the molecular basis governing selective ligand interaction with MC receptors contributes to the rational design of new selective ligands.