Cloning of a human receptor of the NPY receptor family with high affinity for pancreatic polypeptide and peptide YY.

Neuropeptide Y (NPY), peptide YY (PYY), and pancreatic polypeptide (PP) are structurally related peptides found in all higher vertebrates. NPY is expressed exclusively in neurons, whereas PYY and PP are produced primarily in gut endocrine cells. Several receptor subtypes have been identified pharmacologically, but only the NPY/PYY receptor of subtype Y1 has been cloned. This is a heptahelix receptor that couples to G proteins. We utilized Y1 sequence information from several species to clone a novel human receptor with 43% amino acid sequence identity to human Y1 and 53% identity in the transmembrane regions. The novel receptor displays a pharmacological profile that distinguishes it from all previously described NPY family receptors. It binds PP with an affinity (Ki) of 13.8 pM, PYY with 1.44 nM, and NPY with 9.9 nM. Because these data may identify the receptor as primarily a PP receptor, we have named it PP1. In stably transfected Chinese hamster ovary cells the PP1 receptor inhibits forskolin-stimulated cAMP synthesis. Northern hybridization detected mRNA in colon, small intestine, pancreas, and prostate. As all three peptides are present in the gut through either endocrine release or innervation, all three peptides may be physiological ligands to the novel NPY family receptor PP1.

Pancreatic polypeptide (PP) 1 forms a family of 36-amino acid peptides together with neuropeptide Y (NPY) and peptide YY (PYY) (Larhammar et al., 1993). PP was the first of these to be discovered (Kimmel et al., 1968), but in evolutionary terms it seems to be the most recent member and probably arose by duplication of the PYY gene in early tetrapods (Larhammar et al., 1993). PP is exclusively localized to subsets of endocrine cells in the pancreas and inhibits pancreatic secretion, gall bladder contraction, and gut motility (see Hazelwood (1993) for review). PYY is expressed in PP cells and in somatostatin cells (Böttcher et al., 1993;Upchurch et al., 1994) as well as in endocrine cells of the large intestine (El-Salhy et al., 1983;Lundberg et al., 1982), has similar actions to PP, and, in addition, redistributes blood flow in gut vessels (see Laburthe (1990) for review). Both peptides are released into the circulation in response to a meal (see Hazelwood (1993)). In contrast, NPY is present in the central nervous system but is involved in gastrointestinal function through potent induction of feeding in the hypothalamus.
The NPY family peptides exert their actions via heptahelix (seven-transmembrane region) receptors coupled to G-proteins. Several receptor subtypes have been defined by their ability to bind NPY, PYY, PP, and derivatives of these peptides (Gehlert, 1994). Both NPY and PYY bind to the Y1 and Y2 receptors, while the Y3 receptor binds only NPY. The hypothalamic feeding receptor seems to be distinct from all of these (see Gehlert (1994)). An additional receptor has been described that displays a preference for PYY over NPY and is found in the rat small intestine (Laburthe et al., 1986) and in dog adipocytes (Castan et al., 1992), where it mediates reduction of lipolysis. PP does not bind to any of these subtypes but seems to have a unique receptor in dog intestinal mucosa (Gilbert et al., 1988;Gilbert et al., 1986), rat phaeochromocytoma PC12 cells (Schwartz et al., 1987), and rat adrenal cortex and medulla (Whitcomb et al., 1992) as well as in rat vas deferens (Jørgensen et al., 1990) and rat brain area postrema (Whitcomb et al., 1990). Finally, there is a PP-fold-recognizing receptor located in the distal colon in rabbit (Ballantyne et al., 1993) that binds all three peptides. While the discovery of selective peptide agonists has allowed a preliminary receptor classification, the lack of specific receptor antagonists has made functional studies difficult. For instance, it is unclear which receptor mediates the feeding induction reported for human PP in rats (Clark et al., 1984) and dogs (Inui et al., 1991).
To date only the Y1 receptor has been cloned. Cloning of additional receptor subtypes would be helpful to determine their preferences for the three endogenous peptides and to distinguish their physiological roles. The object of the present investigation was to isolate DNA clones encoding additional members of the NPY receptor subfamily. For this purpose we designed degenerate PCR primers based upon the Y1 receptor sequences from human (Herzog et al., 1992;Larhammar et al., 1992), rat (Eva et al., 1990), mouse (Eva et al., 1992), and Xenopus laevis . This approach allowed the cloning of a human receptor that has a higher degree of amino acid sequence identity to the Y1 receptor than to other heptahelix receptors. We also describe functional expression of this receptor to identify it as a PP-preferring receptor, hence named PP1.

EXPERIMENTAL PROCEDURES
Generation of a Rat Y1-like Clone by PCR-Degenerate primers were used in different pairwise combinations for PCR on rat genomic DNA using the following conditions: 5 min at 99°C for one cycle and then 1 min at 94°C, 2 min at 42°C, and 3 min at 72°C for 25 cycles using Taq polymerase. The product of one primer combination was subcloned. The 5Ј primer was a 29-mer with the sequence CGG GAT CCT A(C/T)A

CI(C/T) T(G/A/T/C)A TGG A(C/T)C A(C/T)T GG corresponding to a
BamHI cloning site and positions 362-382 in TM2 of the rat Y1 sequence (GenBank accession code Z11504). The 3Ј primer had the sequence CGG GAT CCC C(A/G)T A(A/G)A A(G/A/T)A TIG G(G/A)T T(G/ A/T/C)A C(A/G)C A corresponding to a BamHI site and positions 1004 -1026 in TM7. The PCR product was separated on an agarose gel, and a band corresponding to 670 base pairs was cut out, added to 500 l of water, boiled for 5 min, and reamplified for 1 min at 94°C, 2 min at 42°C, and 2 min at 72°C for 25 cycles. An aliquot of the generated product was ligated to 25 ng of the plasmid vector pT7Blue (Novagen) with T4 DNA ligase (U.S. Biochemical Corp.) and transformed into E. coli DH5␣ cells. One clone called R4 -7, containing an insert of the expected size, was obtained and characterized further.
Sequencing of the Rat PCR Product-Sequence determinations were performed with dideoxy chain termination in an automated flourescent dye DNA sequencer (Applied Biosystems Inc.) or manually using [␣-35 S]dATP followed by autoradiography. Primers JS1 and JS2 of nucleotide sequence GAGCGGATAACAATTTCACACAGG and GCC-AGGGTTTTCCCAGTCACGACGA were used for the ABI sequencing. For manual sequencing either a T7 primer or a M13 forward primer were used.
Generation of a PCR Probe for Screening of Library-A PCR product was generated with the rat clone R4 -7 as a template and using the same degenerate primers as described previously using the following conditions: 1 min at 94°C, 1 min at 55°C and 2 min at 72°C for 25 cycles. The product was labeled with [␣-32 P]dCTP by the random priming method (Pharmacia Oligo Labeling kit).
Screening of a Human Genomic Library-A human genomic library made from lymphocytes in the vector lambda DASH (Stratagene) was plated out with Escherichia coli LE392 as bacterial host strain. Approximately 600,000 plaques were lifted with nylon membranes, and hybridization was done with the rat probe for 16 h with high stringency at 65°C in 25% formamide, 6 ϫ SSC, 10% Dextran sulfate, 5 ϫ Denhardt's solution, and 0.1% SDS. Filters were washed twice at room temperature in 2 ϫ SSC, 0.1% SDS and twice for 30 min at 65°C in 0.2 ϫ SSC, 0.1% SDS. Screenings were carried out in three consecutive steps to obtain single plaques. Six individual clones were selected.
Phage Clone Characterization-Phage DNA was digested with various restriction enzymes and run on agarose gel. The gel was denatured and blotted onto a nylon membrane that was hybridized as described above with the rat probe. The six phage clones were found to be nonidentical but contained the same hybridizing region as shown by identical restriction sites. Hybridizing fragments from different phages were identified and cloned into the plasmid vector Bluescript KS ϩ . A restriction map was constructed for overlapping hybridizing plasmid inserts.
Sequencing of the Human Y1-like Gene-Sequencing was carried out as described above. Several different plasmid clones from two different phage clones were sequenced. One plasmid clone, Hubert (from phage gH2), containing a 1.45-kb PvuII fragment cloned into the SmaI site of Bluescript KSϩ, was found to contain the entire coding region. From this clone a StyI-BamHI deletion subclone of 750 base pairs was generated for further sequencing.
Southern Hybridization-Genomic DNA was purified from human leucocytes and digested with restriction enzymes. The probe was a PCR-generated fragment corresponding to the entire coding region of the clone Hubert. Hybridization and washes were done at high stringency.
Northern Hybridization-Northern membrane containing mRNA from different human organs and brain regions was purchased from Clontech. The membranes were prehybridized for 4 h at 42°C in 5 ϫ SSPE, 10 ϫ Denhardt's solution, 100 g/ml scheared salmon sperm DNA, 50% formamide, and 2% SDS. This was followed by hybridization overnight at 42°C in the same buffer containing 2 ϫ 10 6 cpm of a nick-translated (Life Technologies, Inc.) 1.2-kb fragment containing the entire coding region of the clone Hubert. The membranes were washed three times for 10 min each at room temperature (2 ϫ SSC, 0.05% SDS). This was followed by a 40-minute wash at 50°C (0.1 ϫ SSC, 0.1% SDS) with one change of solution. The blots were then visualized using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Cloning into Expression Vector-As no suitable restriction sites were available flanking the receptor gene for cloning into the expression vector, two oligonucleotides were used as PCR primers to generate a fragment containing the entire coding region. The 5Ј primer contained a HindIII cloning site and had the sequence CCG GGA AGC TTC CCG CGT CAT CCC TCA AGT GTA TC, and the 3Ј primer had an EcoRI cloning site and the sequence CGG AAT TCC GGC AAG GGA CAT GGC AGG GAG. The PCR was run with Vent DNA polymerase (Biolabs) and the clone Hubert as a template under the following conditions: 1 min at 94°C, 1 min at 50°C, and 2 min at 72°C for 25 cycles. An aliquot of the PCR reaction was separated on an agarose gel and displayed the expected product of 1.25 kb. The remainder of the reaction was phenolextracted and cut with HindIII-EcoRI, and the fragment was purified on an agarose gel and ligated into the expression vector pTEJ-8 (Johansen et al., 1990) to give the clone Hubert-pTEJ. This clone was completely sequenced to ascertain identity to the genomic clone.
Transient Transfection Protocol-COS1 African green monkey kidney cells (COS-1) were seeded at a density of 1 ϫ 10 6 cells/150-mm dish and incubated for 48 h at 37°C. Each dish was transfected with 600 l of Lipofectace (Life Technologies, Inc.) containing 25 g of Hubert-pTEJ according to kit protocol. Cells plus DNA/Lipofectace mixture were incubated for 6 h. Cells were harvested in PBS 48 h after transfection and pelleted by centrifugation.
Binding Assays-The homogenate binding studies were conducted as described previously (Gehlert et al., 1992). Cell pellets were resuspended using a glass homogenizer in 25 mM HEPES (pH 7.4) buffer containing 2.5 mM CaCl 2 , 1 mM MgCl 2 , and 2 g/liter Bacitracin. Saturation experiments were performed in a final volume of 200 l containing various concentrations of 125 I-pPYY (SA 2200 Ci/mmol, DuPont NEN) and 5-10 g of protein for 2 h at room temperature. Nonspecific binding was defined as the amount of radioactivity remaining bound to the cell homogenate after incubation in the presence of 1 M unlabeled human PP (hPP). In competition studies, various concentrations of the peptides (hPP, hPYY, hNPY, porcine [Leu 31 -Pro 34 ]NPY, porcine NPY2-36, and porcine NPY13-36) (Peninsula, Belmont, CA, or Bachem, King of Prussia, PA) were included in the incubation mixture along with 125 I-pPYY. Incubations were terminated by rapid filtration through GF/C filters (Wallac, Gaithersburg, MD), which had been presoaked in 0.3% polyethyleneimine (Sigma), using a TOMTEC (Orange, CT) cell harvester. The filters were washed with 5 ml of 50 mM Tris (pH 7.4) at 4°C and rapidly dried at 60°C. The dried filters were treated with MeltiLex A (Wallac) and melt-on scintillator sheets, and the radioactivity retained on the filters was counted using the Wallac 1205 Betaplate counter. The results were analyzed using the Prism software package (Graphpad, San Diego, CA) or the Cheng-Prushoff equation. Protein concentrations were measured using Coomassie Protein Assay Reagent (Pierce) with bovine serum albumin for standards.
cAMP Assay-A cell line with stable PP1 expression was obtained by transfection of Chinese hamster ovary cells with Hubert-pTEJ. cAMP was assayed in whole cells treated for 20 min at 37°C with 100 M isobutylmethylxanthine. Cells were incubated with 15 M forskolin and various concentrations of hPP, hPYY, and hNPY for 15 min at 37°C. Reactions were terminated by the addition of EDTA to 0.4 M and heating in a boiling water bath for 4 min. Sample buffer containing cAMP was removed and lyophilized. cAMP was quantitated using radioimmunoassay (Amersham Corp.). Protein content of each well was measured using the Coomassie Protein Assay Reagent (Pierce) with bovine serum albumin as the standard.

RESULTS
Isolation of a Y1-related Rat PCR Product-To generate primers for PCR, we analyzed the sequences for the Y1 receptor from several species. Several degenerate primers were designed and used for PCR on rat genomic DNA. Two of the primers corresponding to TM2 and TM7 generated a product of the expected size. The fragment was cloned; one clone was sequenced and found to have higher sequence identity to the Y1 receptor than to all other receptor sequences.
Isolation of a Full-length Human Homologue-A human genomic library was screened under conditions of high stringency with the rat PCR product as a probe. Many clones hybridized, and six of the most strongly hybridizing ones were analyzed further. Five nonidentical clones contained the same hybridizing fragments in a Southern blot, while the sixth clone had a hybridization pattern indicating that it was truncated near the hybridizing segment but contained the same gene as the other five. Fragments of appropriate sizes were subcloned, and a restriction map was constructed. The subclone Hubert of 1.45 kb was found by sequencing to contain the entire coding region of a receptor with high identity to the rat PCR product. This clone encodes a heptahelix (7-TM) receptor of 375 amino acids (Fig. 1). It has greater amino acid sequence identity to the Y1 receptors (Fig. 2) than to any other receptor with 53% identity in the transmembrane regions and 43% overall identity. The closest non-Y1 receptor is the dog gastrin receptor (Kopin et al., 1992) with an overall identity of about 30% (Fig.  2). The novel receptor gene lacks the intron immediately after TM5 that is present in the Y1 receptor genes in all four species characterized to date. The nucleotide sequence identity to the human Y1 sequence is 58%.
The receptor protein deduced from the nucleotide sequence displays many of the characteristic features of heptahelix receptors ( Figs. 1 and 2). The amino terminus has three potential glycosylation sites, and a fourth is present in the second extracellular loop (as in the Y1 receptor). Four extracellular cysteines, one in the amino-terminal region and one in each of the three extracellular loops, presumably form two disulfide bridges (again like the Y1 receptor). A cysteine in the cytoplasmic tail probably serves as an attachment site for palmitate inserted into the cell membrane.
The sequence similarity to Y1 is most prominent in the transmembrane regions, but the loops also show blocks of resemblance. The sequenced portion of the gene extends 180 base pairs beyond the termination codon, but no polyadenylylation signal was found in agreement with the large size of the mRNA (see below).
Southern Hybridization-A single band corresponding to the isolated receptor gene was observed at high stringency (not shown), suggesting that the human genome contains a single PP1-receptor gene.
Northern Hybridization-The expression of receptor mRNA in various human organs and brain regions was investigated by Northern hybridization. Among the organs (Fig. 3), colon, small intestine, pancreas, and prostate showed a band in the range 6 -7 kb. All of the other peripheral organs gave no signal. In the nervous system, faint signals were observed in cerebellum, medulla, and spinal cord after long exposure (not shown).
Binding Properties of the Novel Human Receptor-The coding portion of the clone Hubert was cloned into the expression vector pTEJ-8 and transfected into COS1 cells. Membranes prepared from these cells exhibited concentration-dependent binding of 125 I-pPYY (Fig. 4). This radioligand identified a single class of high-affinity binding sites with an affinity constant (K d ) of 148 Ϯ 29 pM (n ϭ 3, ϮS.E.)for 125 I-pPYY and B max of 258 Ϯ 46 fmol/mg protein. Nontransfected COS1 cells exhibited no specific binding of 125 I-pPYY (data not shown). Competition experiments were performed using PYY, PP, NPY, and various peptide analogues (Fig. 5). Both hPP and hPYY were potent inhibitors of 125 I-pPYY binding with inhibition constants (K i ) of 13.8 Ϯ 0.4 pM (n ϭ 4, ϮS.E.) and 1.44 Ϯ 0.2 nM (n ϭ 4, ϮS.E.), respectively. NPY was less potent, with a K i of 9.88 FIG. 2. Amino acid sequence alignment. The human PP1 receptor serves as master sequence in alignment with the human Y1 receptor (Larhammar et al., 1992) and the dog gastrin receptor (Kopin et al., 1992). In the two latter sequences only positions that differ from the PP1 sequence are shown, while dots mean identities. Dashes represent gaps introduced to optimize alignment. The hydrophobic segments assumed to be embedded in the cell membrane are underlined. Four tripeptides in extracellular parts underlined with dotted lines conform to the consensus sequence for N-linked glycosylation. Diamonds show four extracellular cysteines and one intracellular cysteine. Ϯ 1.13 nM (n ϭ 4, ϮS.E.). The difference for PYY between K d (148 pM) and K i (1.44 nM) is probably because porcine PYY was used for the former and human PYY for the latter, and the two species differ at two amino acid positions (Larhammar et al., 1993). The Y1-selective analog h[Leu 31 -Pro 34 ]NPY was slightly less potent than NPY, with a K i of 21.2 Ϯ 2.0 nM (n ϭ 4, ϮS.E.) and substantially less potent than PYY and PP. Also porcine NPY2-36 was slightly less potent than intact NPY with a K i of 42.2 Ϯ 1.6 nM (n ϭ 4, ϮS.E.). The Y2-selective fragment, porcine NPY13-36, had very low potency, with a K i of 139 Ϯ 4 nM (n ϭ 4, ϮS.E.). Several unrelated peptides were also tested and did not significantly affect binding at 1 M concentrations.
cAMP Assay-A stably transfected Chinese hamster ovary cell line was assayed for cAMP after stimulation of adenylyl cyclase with forskolin in the presence of hPP, hPYY, or hNPY. Both PP and PYY produced a dose-dependent inhibition of adenylyl cyclase activity (Fig. 6). Under these conditions, maximal inhibition was approximately 50%, and IC 50 was 7 nM for hPP and 95 nM for hPYY. hNPY at concentrations of up to 10 uM did not appear to affect adenylyl cyclase activity under these conditions (not shown).

DISCUSSION
Binding studies to different tissue preparations and cell lines have demonstrated the existence of several distinct receptor subtypes that bind NPY family peptides and peptide analogues. The molecular and physiological characterization of these receptors requires access to molecular clones that can be used for functional expression in cell lines and design of specific DNA and RNA probes. So far only the Y1 receptor has been cloned. We have used molecular biology approaches to find clones for additional receptor subtypes related to Y1 and describe here one such clone that displays greater homology to the Y1 receptor than to any other G-protein-coupled receptor. Because the novel receptor preferentially binds PP among the NPY family peptides, we call the receptor PP1.
The human PP1 receptor consists of 375 amino acids with 53% identity to the human Y1 receptor in the TM regions. This degree of identity is similar to that between different subtypes of tachykinin or somatostatin receptors in the TM regions. The overall identity to hY1 is 43%. The PP1 receptor shares several features with Y1 such as three amino-terminal glycosylation sites and four extracellular cysteines. Intracellular loops 1 and 2 have multiple identical positions; loop 1 has seven out of ten identities, and in loop 2 the first nine amino acids are identical. This motif, ERHQLIINP, is also conserved among all four known Y1 sequences . It will be interesting to see whether other NPY family receptor subtypes have the same motif. The sequence similarity in the intracellular loops is consistent with the finding that the PP1's messenger response, namely inhibition of forskolin-stimulated cAMP synthesis (Fig. 6), is similar to that of Y1.
A recent study of the human Y1 receptor by site-directed mutagenesis suggested four acidic residues as points of interaction with basic side chains in NPY, namely Asp-104, Asp-194, Asp-200, and Asp-287 (Walker et al., 1994). We have previously shown that three of these positions are negatively charged also in the Xenopus laevis Y1 receptor , whereas the position corresponding to Asp-194 is a glycine. The PP1 receptor presented here has negatively charged side chains in all four corresponding positions, namely Asp-105, . The similar pattern in negatively charged residues might indicate that PP and PYY bind this receptor in a manner similar to NPY binding to Y1. However, because NPY binds less strongly to PP1 than to Y1, there clearly must be additional structural aspects that diminish NPY binding to the PP1 receptor.
While many heptahelix receptor genes lack introns in their coding regions, the Y1 gene was found to have a small intron immediately after the segment encoding TM5 (Eva et al., 1992;Herzog et al., 1992;Larhammar et al., 1992). The human PP1 receptor gene described here lacks this intron as does the rat genomic fragment generated with PCR that was used to isolate the human PP1 gene. Evolutionary studies may show whether the intron was present in the ancestral NPY family receptor gene. Southern hybridizations to human genomic DNA at high stringency suggest a single PP1 receptor gene.
The functional expression binding studies of the PP1 receptor revealed a high affinity for hPP with a K i of only 13.8 pM. The PP1 receptor also exhibits high affinity for hPYY (1.44 nM) and hNPY (9.9 nM). No previous reports in the literature have described a human PP receptor. When comparing the pharmacological profile of the human PP1 receptor with PP-preferring receptors described for other species in the literature, some important differences emerge. 125 I-bPP has been reported to bind to a receptor on rat PC12 cells that differs in pharmacology to the Y1 receptor also found on these cells (Schwartz et al., 1987). However, while the PC12 receptor has high affinity for bPP, it exhibits very low affinity for NPY (Ͼ1000 nM) whereas the human PP1 receptor binds NPY with a K i of 9.9 nM. We have recently cloned the rat ortholog of PP1, 2 and this receptor, too, binds NPY with higher affinity than the PC12 receptor. In the rat vas deferens, both PP and NPY mediate an inhibition of the electrically evoked twitch response with similar IC 50 values (Jørgensen et al., 1990); however, this effect is probably mediated by separate receptor populations. A previously reported PP receptor in the basolateral membranes of the canine intestine (duodenum, jejunum, ileum, and colon) (Gilbert et al., 1988;Gilbert et al., 1986) displayed high binding to bPP. However, because it had very low affinity for PYY and NPY and these were almost equal to one another, this receptor seems to be distinct from the PP1 receptor described here. The PP-fold receptor found in rabbit distal colon was reported to bind all three peptides with almost equal affinity (Ballantyne et al., 1993). Naturally, some of these differences in pharmacology may be due to species differences.
The selectivity (but not the affinity) of NPY for the Y1 receptor can be improved by replacing the amino acids found in positions 31 and 34 with those found in PP, Leu, and Pro, respectively. In the present study, we found that h[Leu 31 ,Pro 34 ]NPY has a 2-fold lower affinity with a K i of 21.2 nM, whereas NPY has 9.9 nM. Thus, in this respect our novel receptor is reminiscent of, but distinct from, Y1. However, the Y1 as well as the Y2 receptor has low affinity for PP . The Y3 receptor in bovine adrenal chromaffin cells has high affinity for NPY but relatively lower affinity for PYY and PP .
The presence of mRNA for the human PP1 receptor in colon, small intestine, and pancreas (Fig. 3) is consistent with the known effects of PP. The faint mRNA band in medulla (not shown) may indicate a relationship to the binding sites for 125 I-bPP that have been localized in the nucleus of the solitary tract in rat (Whitcomb et al., 1990). Our recently cloned rat PP1 receptor will allow investigation of this possibility.
Thus, the novel human receptor PP1 has pharmacological properties that are consistent with a PP receptor but distinguish it from all pharmacologically characterized receptor subtypes for PP, PYY, and NPY.