Molecular Cloning and Functional Characterization of MCH2, a Novel Human MCH Receptor*

Melanin-concentrating hormone (MCH) is involved in the regulation of feeding and energy homeostasis. Recently, a 353-amino acid splice variant form of the human orphan receptor SLC-1 (1) (hereafter referred to as MCH1) was identified as an MCH receptor. This report describes the cloning and functional characterization of a novel second human MCH receptor, which we designate MCH2, initially identified in a genomic survey sequence as being homologous to MCH1 receptors. Using this sequence, a full-length cDNA was generated with an open reading frame of 1023 base pairs, encoding a polypeptide of 340 amino acids, with 38% identity to MCH1 and with many of the structural features conserved in G protein-coupled receptors. This newly discovered receptor belongs to class 1 (rhodopsin-like) of the G protein-coupled receptor superfamily. HEK293 cells transfected with MCH2 receptors responded to nanomolar concentrations of MCH with an increase in intracellular Ca2+ levels and increased cellular extrusion of protons. In addition, fluorescently labeled MCH bound with nanomolar affinity to these cells. The tissue localization of MCH2 receptor mRNA, as determined by quantitative reverse transcription-polymerase chain reaction, was similar to that of MCH1 in that both receptors are expressed predominantly in the brain. The discovery of a novel MCH receptor represents a new potential drug target and will allow the further elucidation of MCH-mediated responses.

Melanin-concentrating hormone (MCH) 1 is a cyclic neuropeptide that was first discovered in teleost fish, in which it acts as a skin color-regulating hormone (2). In rodents its tissue distribution in the perikarya of the lateral hypothalamus and the zona incerta suggests that MCH may be involved in a variety of behavioral responses (3). Similar tissue distributions have been reported in both bird (4) and monkey (5). Reports implicating MCH in the regulation of feeding behavior show that increased food intake occurs after direct administration of MCH into the brain (6) and that MCH is up-regulated after fasting and in obese leptin-deficient mice (7). There are also reports that suggest MCH may be involved in aggressive behavior, anxiety, and reproductive function (8,9).
Recently, several groups independently identified a 353-amino acid splice variant of the orphan G protein-coupled receptor (GPCR) SLC-1 (1) as an MCH receptor (10 -14). In view of the findings of the current study, we propose that this form of SLC-1 be hereafter referred to as MCH 1 . Southern blot and related studies have indicated the absence of additional MCH receptor subtypes that closely resemble MCH 1 at the DNA level (3,10,15). However, because degeneracy in receptor-ligand pairings throughout the GPCR superfamily is common, we reasoned that other MCH receptors with low homology to the MCH 1 receptor may exist. This suggestion is supported by reports of pharmacological differences between the MCH 1 receptor and MCH binding sites in various cell lines and tissues (10).
Sequencing of the human genome resulted in the deposition of a vast amount of unannotated sequence in public data bases in recent years. In the present study, we first describe how we identified a sequence with low but significant homology to the MCH 1 receptor from one of these data bases and cloned a full-length cDNA from this. We then demonstrate that the predicted 340-amino acid polypeptide product of this sequence, which we term MCH 2 , exhibits many of the structural features of the GPCR superfamily, is a member of the class 1 (rhodopsin-like) subfamily, and after heterologous expression in HEK293 cells is selectively activated by nanomolar concentrations of MCH. Last, to investigate the biological significance of this finding we compare the tissue distribution of MCH 1 and MCH 2 receptors by RT-PCR analysis.

Receptor Cloning, Transient Expression, and Generation of Stable
Cell Lines-A 195-base pair genomic survey sequence (GenBank TM accession number AQ311725) was identified that when translated exhibited 42% identity at the amino acid level to the transmembrane-4 region of MCH 1 . The primers were designed to perform 5Ј (5Ј-CAGAGTACATCGTCAGGGGATGTCAAATCAAAA-3Ј) and 3Ј (5Ј-TACTTTGCCCTCGTCCAACCATTT-3Ј) rapid amplification of cDNA ends on a Marathon human fetal brain cDNA template (CLONTECH). Extension of the known sequence at both the 5Ј and 3Ј ends revealed a coding sequence of 1023 base pairs with an in-frame upstream stop codon at position Ϫ24 (Fig. 1). The full-length gene was amplified from the fetal brain cDNA template using forward (5Ј-ACAATGAATCCATT-TCATGCATCTTGT-3Ј) and reverse (5Ј-TGCTGCTAAGAGTCACAAG-TACACAAGAAG-3Ј) primers. The cDNA was cloned into pcDNA3.1/V5/His-TOPO (Invitrogen), and both strands were sequenced on an ABI sequencer. HEK293 cells were transfected with the recombinant plasmid using LipofectAMINE plus reagent (Life Technologies, Inc.) following the manufacturer's instructions. Stable cell lines were generated by selection in geneticin (16), and clones were screened by MCH-induced calcium mobilization on a fluorometric imaging plate * 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF347063.

reader (as described below).
Calcium Mobilization Assays-Intracellular calcium assays were carried out essentially as described previously (17). The maximum change in fluorescence above baseline, measured on a fluorometric imaging plate reader (FLIPR, Molecular Devices), was used to determine the agonist response. For cross-screening studies, HEK293 cells were screened against a large library of over 1500 known and putative GPCR agonists including all known mammalian neuropeptides as described previously (17). Peptides in this library were tested at a final concentration of Ͼ100 nM, and other potential agonists were tested at a final concentration of Ͼ1 M. For antagonist studies, test substances were added 30 min before the addition of an EC 50 concentration of agonist. The data were analyzed using GraFit (Erithacus Software). Peptides were purchased and synthesized as described previously (10).
Microphysiometry-Changes in the extracellular pH of HEK293 cells stably transfected with MCH 2 receptors were monitored using the Cytosensor microphysiometer (Molecular Devices) (18). Cells were seeded into poly-L-lysine-coated Cytosensor capsules (0.2 million cells/capsule) and cultured overnight. The capsules were placed on the Cytosensor and equilibrated for 2 h with modified RPMI 1640 medium (Life Technologies, Inc.), pH 7.4. MCH was applied to the cells for 20 s prior to the "get rate," and the cells were then washed with running medium to remove ligand. Extracellular acidification rates were determined as the rate of change of sensor output during the periodic interruption of media flow. The medium was allowed to flow for 80 s and stopped for 38 s. Rates were measured over 30 s, starting 8 s after the flow was stopped.
Laser Scanning Cytometry Binding Assays-HEK293 cells transiently expressing MCH 2 receptors were seeded into 16-well chambers  (Lab-Tek, Nalge Nunc International). The cells were grown in Eagle's minimum essential medium (including L-glutamine, 10% fetal calf serum, 1% nonessential amino acids, and 400 g/l G418) for 24 h and then incubated at 37°C for 30 min with indodicarbocyanine-labeled MCH (Amersham Pharmacia Biotech) at concentrations ranging from 0 to 40 nM in HEPES-buffered saline (including 2.5 mM MgCl 2 , 1.5 mM CaCl 2 , and 0.5% bovine serum albumin). The cells were then washed in HEPES-buffered saline minus the bovine serum albumin and fixed with 4% paraformaldehyde. Analysis was performed using a laser scanning cytometer (CompuCyte). Excitation of the indodicarbocyanine label by a 5-milliwatt helium/neon laser resulted in the emission of fluorescence that was collected through a 650-nm long path filter and measured by monitoring the red fluorescence maximal pixel intensity. Nonspecific binding was determined in the presence of 40 M unlabeled MCH, and specific bindingderived fluorescence was determined by subtraction of this from total binding. The data were analyzed by using Sigma Plot (SPSS, Inc.).

RESULTS
The MCH 2 receptor was cloned from human fetal brain tissue cDNA. The 1023-base pair open reading frame encodes a 340-amino acid protein that structurally resembles members of the GPCR superfamily (Fig. 1). The receptor contains a short N terminus, seven distinct hydrophobic membrane-spanning domains, and the highly conserved DRY motif located at the interface between the third transmembrane helix and the cytoplasm (20). The receptor has only one initiator methionine in the open reading frame, which contrasts with three such putative initiator methionines in MCH 1 . There are two putative N-linked glycosylation sites, and there is no characteristic signal peptide. BLAST analysis of public data bases revealed MCH 1 to be its most homologous relative. The two GPCRs are 57% identical at the nucleotide level and 59% similar and 38% identical at the amino acid level (Fig. 2). Somatostatin receptors were the most similar receptors to these two putative paralogs, with ϳ26% identity at the amino acid level.
We transiently transfected HEK293 cells with MCH 2 recep-tor cDNA and tested these cells for responsiveness to MCH. MCH (100 nM) induced a clear, robust, and transient increase in intracellular Ca 2ϩ in MCH 2 receptor-transfected cells but not in control cells transfected with the same vector containing a opiate receptor (Fig. 3A). The dose dependence of this response was investigated, and an EC 50 value (Ϯ S.E., n ϭ 3) of 8.57 Ϯ 0.62 nM was determined. To confirm the specificity of MCH 2 receptor activation by MCH, we also screened these cells against a large library of known and putative GPCR ligands including all known mammalian neuropeptides at final concentrations greater than 100 nM. MCH was the only substance in this library observed to elicit MCH 2 receptor-mediated Ca 2ϩ responses in these cells. A HEK293 cell line stably expressing the MCH 2 receptor was established and used in all further functional studies. The concentration dependence of MCH 2 receptor activation by MCH and related peptides was investigated in the intracellular Ca 2ϩ assay (Fig. 3B). MCH, salmon MCH, and [Phe 13 ,Tyr 19 ]MCH all behaved as agonists with similar potencies (EC 50 values (Ϯ S.E., n ϭ 3) of 5.65 Ϯ 1.78, 7.14 Ϯ 3.13, and 4.29 Ϯ 0.48 nM, respectively). Variant MCH, the putative product of a second, variant form of the MCH gene (21), was a weak agonist (EC 50 Ͼ 3000 nM). To determine the type of G protein mediating this response, we pretreated cells with pertussis toxin (100 ng/ml for 16 h). The toxin treatment had no effect on calcium mobilization by MCH in these cells (data not shown), suggesting that the MCH 2 receptor is coupled to G proteins of the G q/11 subfamily.
In a variety of native receptor studies, a number of peptides have been reported to either exhibit functional antagonism of MCH responses or inhibit binding to MCH binding sites (22)(23)(24). To investigate whether these effects are mediated via MCH 2 receptors, we tested the following peptides in agonist and antagonist modes over a range of concentrations up to 10 M in an intracellular Ca 2ϩ assay: rat atrial natriuretic peptide (1-28), rat atrial natriuretic peptide (3-28), human C-type natriuretic peptide-22, human brain natriuretic peptide-32, ␥-endorphin, and ␣-melanocyte-stimulating hormone. We also tested somatostatin-14, somatostatin-28, and cortistatin-14 because of the similarity of MCH 1 and MCH 2 receptors to somatostatin receptors. We also tested a number of putative products of the authentic MCH precursor and of a second, variant form of the MCH gene: neuropeptide EI, neuropeptide GE, MCH gene-overprinted peptide-14, and variant neuropeptide EI. None of these peptides were active as agonists or antagonists at concentrations up to 10 M.
To confirm further that the MCH 2 receptor responded specifically to MCH, we monitored extracellular acidification rates

FIG. 4. MCH increases extracellular acidification rates in HEK293 cells stably transfected with MCH 2 receptor (filled symbols) but not in nontransfected cells (open symbols)
. Cells were exposed to either 10 (squares or diamonds) or 1 nM (triangles) MCH over the time course indicated by the bar. Data are taken from a single experiment representative of a total of three such experiments. in cells stably expressing the receptor using the technique of microphysiometry. 1 and 10 nM MCH caused robust increases in the rate of proton extrusion in transfected cells but not in nontransfected HEK293 cells (Fig. 4). Control studies demonstrated that 100 M muscarine elicited similarly robust acidification responses in both MCH 2 receptor-transfected and nontransfected cells (data not shown), indicating that both cell lines are capable of responding in this assay and thus confirming the specificity of the response to MCH. The dose dependence of the acidification response was also investigated, and an EC 50 value (Ϯ S.E., n ϭ 4) of 1.43 Ϯ 0.44 nM was determined.
We used laser scanning cytometry to measure the binding parameters of fluorescently labeled MCH to MCH 2 receptors. Saturation analysis of indodicarbocyanine-labeled MCH binding to HEK293 cells transiently transfected with MCH2 receptors showed specific high affinity binding with a K d (Ϯ S.E., n ϭ 3) of 6.02 Ϯ 0.46 nM (Fig. 5), although expression levels were apparently low with the fluorescence signal (specific binding) arising from ϳ5% of the population. This value was comparable with the potencies observed for nonlabeled forms of MCH obtained in the two functional assays.
We compared the tissue distributions of MCH 1 and MCH 2 receptors using quantitative TaqMan RT-PCR. The mRNA profiles of the two receptors were similar, showing predominant expression in the brain (Fig. 6). However, a major difference was the much higher relative levels of MCH 1 compared with the MCH 2 mRNA in pituitary. The distribution of the two receptors within individual regions of the brain is similar, but there are subtle differences. For example, the hypothalamus, locus coeruleus, medulla oblongata, and cerebellum seem to express higher relative levels of MCH 1 compared with MCH 2 mRNA.

DISCUSSION
This report describes the cloning and functional characterization of a second, novel MCH receptor. Several lines of evidence suggest that MCH is involved in the regulation of food intake and energy balance. Mice have been generated that carry a targeted deletion of the MCH gene (25). These knockout mice show reduced food intake, have reduced body weight, and are leaner. The localization of MCH in the lateral hypothalamus and the zona incerta, areas that are involved in the regulation of ingestive behavior, is consistent with this role in feeding. However, perikarya from these areas project widely throughout the central nervous system, suggesting an involve-ment of MCH in a wide range of behaviors. Thus MCH has also been shown to interact within the hypothalamo-pituitary-adrenal axis or stress axis (26,27) and to be anxiogenic (28). Hyperactivity of the stress axis is known to occur in individuals suffering from depression. Thus MCH may also have a significant role to play in psychiatric disorders.
Despite the large amount of data concerning the physiological actions of MCH, a receptor had not been identified until recently. Within a short period of time several groups (10 -14) discovered that a 353-amino acid splice variant form of SLC-1 (MCH 1 ) is an MCH receptor. Within the GPCR superfamily, there are numerous examples of natural ligands that can activate more than one molecular species of GPCR. We therefore reasoned that additional MCH receptors may exist. However, low stringency Southern blot studies (10) and related studies (3,15) suggest that additional receptors with high sequence identity to MCH 1 are unlikely to exist. We therefore interrogated public data bases to identify sequences with low but significant levels of homology to MCH 1 . A short sequence was identified in a genomic survey sequence with homology to MCH 1 . Extension of this sequence revealed a full-length cDNA with many of the motifs characteristic of a GPCR, which we designated MCH 2 in view of its sequence similarity to MCH 1 and its subsequent characterization. MCH 2 , similar to MCH 1 , shares low but significant homology to somatostatin receptors. We tested three naturally occurring ligands for somatostatin receptors against MCH 2 receptors and did not observe any agonism or antagonism of MCH responses. Similar studies have demonstrated that somatostatin does not interact with MCH 1 receptors (10,11,15).
Recent studies have investigated the molecular mechanism by which MCH interacts with MCH 1 (29) and demonstrate that Asp 123 in transmembrane-3 plays a key role in the formation of a complex between receptor and MCH, possibly by direct interaction with the Arg 11 of MCH. Interestingly, this transmembrane-3 aspartate is conserved within MCH 2 (Asp 113 ) and may therefore play a similar role in both paralogs. An equivalent aspartate residue with functional significance for receptor activation has also been identified in the related families of opiate (30) and somatostatin (31) receptors, and the same residue has been recognized for many years to be conserved in all biogenic amine GPCRs, in which it acts as a counter-ion for the protonated amine moiety of the ligand (32).
FIG. 6. Tissue localization of MCH 1 and MCH 2 receptors using TaqMan RT-PCR methodology. The expression profile of the two receptors over a range of 20 different human tissues and 19 different brain regions is shown. The data are presented as mean (Ϯ S.E.) mRNA copies detected of four nondiseased individuals (two males and two females (except prostate)) for each tissue from 1 ng of Poly(A) ϩ RNA. Each of the four individual intestine samples represents a 50:50 mix of one person's small intestine and another person's colon (same sex). Each brain sample represents an equal part of the RNA pool of the 19 key brain regions used. The GAPDH distribution profiles for the same samples are shown for comparison.
Both functional and binding assays were used to confirm MCH as the cognate ligand for this receptor. Further characterization using an intracellular Ca 2ϩ assay demonstrated that salmon MCH and the synthetic analogue [Phe 13 ,Tyr 19 ]MCH were equipotent with MCH and were full agonists at MCH 2 . However, variant MCH, the putative product of a second, variant form of the MCH gene (21), was ϳ500-fold less potent than MCH. This profile of agonist activity is essentially similar to that observed at MCH 1 (10). These data also demonstrate that variant MCH is unlikely to act as a natural ligand for either MCH 1 or MCH 2 receptors.
To further characterize the MCH 2 receptor we tested, as both agonists and antagonists, a number of peptides that have been reported to bind with low affinity to MCH binding sites in various cell lines and tissues. Thus, MCH binding sites in mouse melanoma cells (22), human keratinocytes (23), and human brain (24) are weakly displaced by a number of natriuretic peptides. Because these peptides do not bind to MCH 1 receptors, it has been proposed that they may bind to novel MCH receptors (10). In the present study we demonstrate that the MCH 2 receptor is not activated or antagonized by these peptides. This may therefore imply the existence of still further subtypes of the MCH receptor. However, recent studies (33) have demonstrated that the specific binding to some of these cell lines is unlikely to be caused by the presence of a receptor involved in signal transduction, because binding is largely localized to microsomal and not plasma membranes, and the internalization kinetics are not typical of a receptor-mediated event.
We also tested a number of peptides that have been reported either to functionally antagonize MCH responses or are putative products of either the authentic or variant MCH gene. Thus the ␣-melanocyte-stimulating hormone and MCH have mutually antagonistic effects on a number of different physiological functions including feeding behavior (34,35). The lack of activity of ␣-melanocyte-stimulating hormone at MCH 2 receptors parallels similar findings with the MCH 1 receptor (10 -12) and confirms that both peptides exert their effects via separate receptor families. Likewise, we observed that none of the additional putative products of the authentic and variant MCH precursor (neuropeptide GE, neuropeptide EI, variant neuropeptide EI, MCH gene-overprinted peptide-14) seem to interact with MCH 2 receptors. Previous studies have also demonstrated that these peptides do not interact with MCH 1 receptors (10,11). Taken together, these data indicate that any biological effects of these peptides are most likely to be mediated by receptors other than the two known MCH receptors.
Both MCH 1 and MCH 2 receptors are expressed predominantly in the brain. Within the brain, the pattern of expression of the two paralogs is similar. Both paralogs are widely distributed, but the mRNA profiles show noticeably higher contributions from limbic areas such as amygdala, hippocampus, and parahippocampal gyrus and in a number of cortical regions (cingulate-, medial frontal-, and superior frontal-gyri). The widespread distribution of MCH 1 is in agreement with that observed previously in the rat brain (11,36). Because the patterns of expression of these paralogs are so similar, these data suggest that neither paralog can yet be selectively implicated in mediating any specific effect of MCH. The lower relative levels of MCH 2 mRNA in pituitary and hypothalamus compared with MCH 1 suggests that this newly discovered MCH receptor may be involved in physiological processes other than feeding or neuroendocrine modulation.
The identification of a second MCH receptor will help to further elucidate the role of MCH in energy homeostasis and feeding behavior as well as in disorders such as obesity and social anxiety disorder.