The Metastasis Suppressor Gene KiSS-1 Encodes Kisspeptins, the Natural Ligands of the Orphan G Protein-coupled Receptor GPR54*

Natural peptides displaying agonist activity on the orphan G protein-coupled receptor GPR54 were isolated from human placenta. These 54-, 14,- and 13-amino acid peptides, with a common RF-amide C terminus, derive from the product of KiSS-1, a metastasis suppressor gene for melanoma cells, and were therefore designated kisspeptins. They bound with low nanomolar affinities to rat and human GPR54 expressed in Chinese hamster ovary K1 cells and stimulated PIP 2 hydrolysis, Ca 2 (cid:1) mobilization, arachidonic acid release, ERK1/2

G protein-coupled receptors constitute the largest family of membrane receptors (1). Over the last decade, a growing number of orphan G protein-coupled receptors have been made available by various cloning procedures such as PCR 1 amplification using degenerate oligonucleotides and systematic sequencing of cDNA libraries and genomes. In human, in addition to about 160 characterized receptors, an equal number of genes encode proteins obviously belonging to this family of receptors, but their ligands and functions remain to be determined. These orphan receptors potentially constitute elements of fundamental communication pathways in various systems. Some of these orphan receptors have been used as tools to uncover these signaling pathways and have led to the validation of novel drug targets (2). Nociceptin was the first of these novel ligands, purified from tissue extracts with the help of bioassays based on the expression of orphan receptors (3,4). Other recent examples of new molecules identified as ligands of orphan receptors include orexins, involved in the control of feeding and sleep (5,6), prolactin-releasing peptide (7), ghrelin (8), and apelin (9).
GPR54 is an orphan G protein-coupled receptor, originally cloned from rat brain and sharing 45% identity with galanine receptors (10). It is widely expressed in the rat central nervous system, including the hypothalamus, midbrain, pons, medulla, hippocampus, and amygdala (10). No natural or surrogate agonists of GPR54 have so far been identified, and its physiological functions are therefore unknown. The gene encoding human GPR54 is present in a bacterial artificial chromosome mapped to chromosome 2 (GenBank TM accession number AC023583).
In this work, we have tested tissue extracts on the basis of the distribution of GPR54 in human tissues. A biological activity was found in a placenta extract, and active peptides were purified and identified by mass spectrometry. They were found to derive from the previously described metastasis suppressor gene KiSS-1. The pharmacology of the peptides, signaling properties of the receptor, and in vivo function of this new system in vivo were investigated.

Tissue Distribution of GPR54 Expression in Human Tissues-RT-
PCR experiments were carried out as described previously (11). Seventy-five ng of poly(A) ϩ RNA from various human tissues (CLONTECH) were tested using sense (5Ј-TGTACAACCTGCTGGCGCTG-3Ј) and antisense (5Ј-CCACTGCTCCCTGGCTTCTG-3Ј) primers for human GPR54. Amplification of aldolase transcripts was performed in parallel as control.
Cloning and Expression of Rat and Human GPR54 -The coding sequence of rat (GenBank TM accession number AF115516) and human (GenBank TM accession number AY029541) GPR54 was amplified by PCR from rat or human hypothalamus cDNA using rat sense (5Ј-ATC-GGAATTCACCATGGCCGCAGAGGCGACGTTG-3Ј) and antisense (5Ј-ATCATCTAGATCAGAGTGGGGCAGTGTGTTCATC-3Ј) and human sense (5Ј-ATCGGAATTCACCATGCACACCGTGGCTACGTC-3Ј) and antisense (5Ј-ATCATCTAGAACAGAAGAATAGCCGCTGTTCCAGCA-G-3Ј) primers. Amplified DNA fragments were subcloned into the pEFIN3 bicistronic vector (12) and sequenced. These plasmids were transfected into CHO-K1 cells expressing G␣ 16  functional responses were analyzed using a bioassay based on the luminescence of aequorin as a result of intracellular calcium release, as described previously (12). Briefly, cells (25,000 cells in 50 l) were added to 50-l samples in 96-well plates, and the luminescence was recorded for 30 s (PerkinElmer Life Sciences luminometer).
Intracellular Calcium Release Assay-For intracellular Ca 2ϩ assays, cells pretreated or not with 100 ng/ml pertussis toxin for 15 h were incubated in Hanks' balanced salt medium containing 0.1% bovine serum albumin and 2.8 g/ml Fura-2 (Molecular Probes) at 37°C for 45 min as described (11). Cells were used at a density of 10 6 /ml, and intracellular Ca 2ϩ was measured using a luminescence spectrophotometer LS50B (PerkinElmer Life Sciences).
Actin Filament Reorganization-CHO-K1 cells were cultured on cover glasses with or without 5 g/ml C3 exoenzyme (Sigma) in 10% FCS medium for 24 h and further in FCS-free medium for 24 h. Cells were then incubated with 1 M kisspeptin-10 or vehicle for 2 h, fixed with 4% paraformaldehyde in phosphate-buffered saline, permeabilized with 0.1% Triton X-100 in phosphate-buffered saline, and stained with phycoerythrin-labeled phalloidin (Molecular Probes). Staining was observed by a confocal fluorescence microscopy (Axiovert 100, Zeiss) connected with MRC1024 imaging system (Bio-Rad).
Oxytocin Assay-Female Harlan Sprague-Dawley rats weighting 230 -250 g were anesthetized with ether and received 25 g of kisspeptin-10 in 100 l of physiological saline or vehicle alone into the left tail vein. After decapitation between 8 and 9 a.m., blood was collected into plastic syringes containing 500 units/ml aprotinin and 1 mg/ml Na 2 -EDTA. Plasma was rapidly separated by centrifugation at 4°C. Plasma oxytocin concentrations were measured by Oxytocin Correlate-EIA kit (Assay Designs).

RESULTS
Distribution of Human GPR54 -To identify potential sources of endogenous ligands, we tested the expression of GPR54 by RT-PCR in a set of human tissues. As shown in Fig. 1, GPR54 transcripts were particularly abundant in placenta, pituitary, spinal cord, and pancreas, whereas it was found at lower levels in other tissues, including various brain regions, stomach, small intestine, thymus, spleen, lung, testis, kidney, and fetal liver.
Purification of Natural Ligands of GPR54 -CHO-K1 cell lines expressing rat GPR54 (Fig. 2) together with G␣ 16 and apo-aequorin were used in a functional assay based on the luminescence of aequorin to identify natural ligands of the receptor. Extracts from various organs, including brain, placenta, intestine, stomach, lung, testis, and kidney were fractionated by HPLC, and individual fractions were tested in the aequorin-based assay using untransfected CHO-K1 cells and cell lines expressing other receptors as negative controls. Two peaks of biological activity specific for rat GRP54 were identified in fractions of a human placenta extract. These activities were purified through a number of HPLC columns (Fig. 3), and the active fractions resulting from the last step were analyzed by mass spectrometry. Sequencing and comparison of the corresponding masses with the protein and nucleotide data bases resulted in the identification in one peak of two short overlapping peptides of 13 and 14 amino acids, terminating by a common RF-amide motif (Fig. 4). In the other peak, the same peptides were associated with a longer form of 54 amino acids with the same C terminus. All three peptides were predicted to derive from a previously identified protein named KiSS-1 (16). The three peptides were therefore designated as kisspeptin-54, kisspeptin-14, and kisspeptin-13.

FIG. 3. Purification of kisspeptins from human placenta extracts.
A, elution profile of a placenta extract using a DeltaPak C18 column as the first fractionation step. The biological activity of the fractions on rat GPR54 using the aequorin-based functional assay is shown. Two peaks of biological activity, eluting at 32 min (fraction 64) and 33 min (fraction 66) were observed and were further purified separately. B, second step of purification for fraction 64 using a Symmetry 300 C18 column. A single peak of activity (90 min) was observed. C, third step of purification, using a Vydac C18 column. The active fractions were analyzed by mass spectrometry. RU, relative units. KiSS-1-derived GPR54 Ligands stimulated Ca 2ϩ mobilization and phosphatidylinositol turnover in these cells (Fig. 6, A and B). This effect was not affected by pertussis toxin pretreatment. No modification of cAMP accumulation was observed after stimulation of the rat or human GPR54 (data not shown). These results suggest that GPR54 is coupled with proteins of the G q class. Kisspeptins also stimulated arachidonic acid release from CHO cells expressing rat or human GPR54 in a concentration-dependent manner (Fig. 6, C and D).
Given the previous description of KiSS-1 as a tumor suppressor gene, we evaluated the effects of kisspeptins on cell proliferation and cytoskeleton organization. In CHO cells expressing human or rat GPR54, we observed a strong and sustained stimulation of phosphorylation of the MAP kinases ERK1 and ERK2 (Fig. 6, E and F). We also observed a weak stimulation of p38 MAP kinase phosphorylation (Fig. 6G) but not of stressactivated protein kinase/c-Jun NH 2 -terminal kinase (SAPK/ JNK) (Fig. 6H). However, despite the involvement of the MAP kinase pathways in the control of cell cycle and proliferation, culturing recombinant CHO cells in the presence of kisspeptin-10 resulted in a strong reduction of cell proliferation (Fig.  7). The peptide had no effect on untransfected CHO-K1 cells. No evidence of cell death by apoptosis was found either by flow cytometry using phycoerythrin-conjugated annexin V or a genomic DNA fragmentation assay (data not shown).
Actin microfilament reorganization in recombinant CHO-K1 cells was monitored using phycoerythrin-labeled phalloidin and confocal microscopy. Kisspeptin-10 stimulated stress fiber forma-tion in CHO cells expressing rat GPR54 (Fig. 6I). This effect was abolished in cells pretreated with C3 exoenzyme, suggesting that kisspeptin-10 stimulated stress fiber formation through activation of the Rho subfamily of G proteins. Identical results were obtained with cells expressing human GPR54, whereas kisspeptins had no effects on untransfected CHO-K1 cells (not shown).
In Vivo Activity of Kisspeptins-We have started to investigate the function of GPR54 and its ligands in vivo. Given the high expression of GPR54 in hypothalamus and pituitary, we postulated that it could be involved in the control of hormone release. We tested in a first approach the effect of kisspeptin-10 onto oxytocin plasma level in mature female rats. Administration of 25 g of kisspeptin-10 through the tail vein promoted a significant increase of oxytocin with a maximum 30 min after the injection (Fig. 8). DISCUSSION GPR54 is an orphan G protein-coupled receptor structurally related to neuropeptide receptors and, particularly, to galanine receptors. The rat and human cDNAs encoding GPR54 were cloned on the basis of sequences available in the data bases. As for most G protein-coupled receptors, predicted transmembrane segments displayed high interspecies conservation. Low similarity was, however, found in the third extracellular loop and the C-terminal intracellular domain (Fig. 2), suggesting that these regions do not play essential roles in the receptor's function. We determined the distribution of the receptor in FIG. 5. Pharmacology of rat and human GPR54. A, saturation binding assay using a CHO-K1 cell line expressing rat GPR54 and 125 I-kisspeptin-10 as tracer. Nonspecific binding was determined by using 500 nM unlabeled kisspeptin-10 as competitor. B, competition binding assay on rat GPR54 using different peptides as competitors. C, competition binding assay on human GPR54. D, aequorin-based functional assay using CHO-K1 cells coexpressing rat GPR54, apo-aequorin and G␣ 16 . E, aequorin-based assay for human GPR54. Graphs (mean and S.E. of triplicate data points) are representative of at least three independent experiments. human tissues by RT-PCR. GPR54 was found to be expressed in most tissues tested, including in the central nervous system, gut, and endocrine organs. The highest expression was found in pituitary and placenta. This distribution pattern was used to select the tissues in which the potential ligands of GPR54 could be present. No convincing biological activities were detected in extracts from brain, pituitary, and gut segments, whereas a strong biological activity was detected in a human placenta extract, allowing purification of the active compounds from this source. This led to the identification of kisspeptins, peptides encoded by the previously described gene KiSS-1.
KiSS-1 was isolated as a tumor metastasis suppressor gene for the human malignant melanoma cell line C8161 by a subtraction cloning method (16 -18). The predicted product of KiSS-1 is a 145-amino acid protein with a putative signal sequence. The secretory nature of the protein was, however, not demonstrated so far. KiSS-1 does not share similarities with any other protein in the data bases. The metastasis potential of melanoma and breast cancer cell lines in vivo was suppressed after KiSS-1 transfection without affecting the proliferation and migration properties of the cells (16,19). KiSS-1 was also reported to inhibit MMP-9 production when transfected in HT1080 cells (20). The molecular mechanisms underlying these observations were, however, not identified.
Three peptides of 54, 14, and 13 amino acids were identified within the active fractions, and their biological activity was confirmed for the synthetic peptides. The long peptide (54 amino acids) is predicted to result from the proteolytic processing of KiSS-1 by furin or prohormone convertases, as its sequence is surrounded by pairs of basic residues in the full-size protein (Fig. 4). No obvious cleavage sites were found that would result in the generation of the two shorter peptides. Their presence in the same fraction as the longer peptide suggested that the 54-amino acid peptide might be unstable and degraded into the shorter peptides during the extraction procedure, or even during mass spectrometry analysis. In binding and functional assays, it was found that kisspeptin-54, -14, and -13 as well as a shorter peptide designated kisspeptin-10 had the same affinity and efficacy on the rat receptor. This indicates that the C-terminal part of the peptides is responsible both for the high affinity binding and the activation of GPR54. The major role of the RF-amide moiety was demonstrated by the lack of binding and functional response for non-amidated synthetic kisspeptins. Other determinants within the C-terminal 10 amino acids are likely to contribute to the binding and functional properties of the peptides, as other RF-amide peptides had no detectable activity. The observations were similar for the human receptor, although in several functional assays, the EC 50 of the peptides was higher than for the rat receptor as a consequence of the lower expression level of the human receptor in the recombinant cell lines. The high affinity of kisspeptins for GPR54, their isolation from a tissue that expresses GPR54, and the predicted processing of KiSS-1 into kisspeptin-54 strongly suggest that these peptides indeed constitute natural agonists of the orphan receptor.
Analysis of the transduction pathways activated by GPR54 in CHO-K1 cells identified coupling to phospholipase C and intracellular calcium release through pertussis toxin-insensitive (presumably G q ) G proteins. GPR54 also stimulated arachidonic acid release, ERK1/2 and p38 activation, three pathways often activated by G protein-coupled receptors. Interestingly, despite activation of the ERK pathway, kisspeptins strongly inhibited the proliferation rate of the cells expressing GPR54. This proliferation arrest induced by GPR54 activation is not a property shared by all G q -coupled receptors, as the angiotensin AT 1A and vasopressin V 1 receptors have been reported previously to stimulate proliferation in CHO cells (21,22). Additional work will be necessary to determine which pathway is responsible for this proliferation arrest. Kisspeptins also induced stress fiber formation in GPR54-expressing cells through Rho activation. This observation might be linked to an earlier report describing faster spreading to culture plates of breast cancer cells transfected with KiSS-1 cDNA (19), although expression of GPR54 was not investigated in these cells. After completion of this study, the identification of kisspeptins as ligands of GPR54 was reported by Ohtaki et al. (23) and Muir et al. (24).
In conclusion, we have isolated KiSS-1-derived peptides from human placenta as high affinity natural agonists of a previously orphan G protein-coupled receptor designated GPR54. Tissue distribution suggested that GPR54 might be implicated in various hormonal function, a hypothesis supported by the demonstration that KiSS-1-derived peptides stimulate oxytocin release in rats. It is, however, likely that many other biological actions of kisspeptins in vivo will be uncovered in the future. The correlation of the present observations with the previous description of KiSS-1 as a tumor metastasis suppressor gene will require further investigations.