INTRODUCTION

Pancreatic polypeptide (PP)¹ is a hormone found in the general circulation, where it is released after meal ingestion (1, 2, 3). Little is known about the function of PP, which is produced and secreted primarily by certain endocrine cells of the pancreatic islets (1, 2, 3). PP-containing cells can be found also in gut and intestine (3, 4). The main known biological effects of PP are those on gastrointestinal tract and pancreatic secretion (1, 2, 3, 4, 5). Some activities of PP are mediated centrally via brainstem sites (6). In addition, PP may serve some unknown function in other tissues in which PP binding sites were described, including adrenal gland, brain, prostate, and liver (6, 7, 8, 9, 10).

PP, neuropeptide Y (NPY), and peptide YY (PYY) belong to a family of structurally related 36-amino-acid peptides, which produce their effects through the interaction at multiple receptors, i.e. Y1, Y2, Y3, Y1-like, as well as a PP receptor (4, 5, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21). Recently, we and others have reported cloning and characterization of a PP receptor from human, rat, and mouse, termed Y4 or PP1 (15, 16, 18, 19). The Y4/PP1 receptor has subnanomolar affinity to PP and nanomolar affinity to PYY and NPY, raising the possibility that it may be shared by all three pancreatic polypeptides. Tissue distribution studies in humans and mice suggest potential roles for Y4/PP1 receptor in the gastrointestinal tract, heart, and prostate, as well as in neural and endocrine signaling.

In this report we describe cloning of a novel member of the NPY/PYY/PP receptor family termed PP2. We show that the mouse version of this gene is coding for a functional PP receptor and describe pharmacological properties of recombinant mouse PP2 receptors expressed in COS-7 cells. We also present data that suggest that the human PP2 gene is mutated, having a single base deletion, and that human PP2 receptors expressed in COS-7 cells do not interact with pancreatic polypeptides.

EXPERIMENTAL PROCEDURES

Library screening, cloning, Northern and Southern blotting, PCR, and other manipulations were carried out by standard methods (22) as described previously (16). Reverse transcription coupled with PCR (RT-PCR) was carried out with human brain mRNA (Clontech Labs, Palo Alto, CA) as a template and using a RT-PCR kit (Perkin Elmer Cetus, Norwalk, CT). Degenerate oligonucleotide primers corresponding to cloned NPY receptors (11, 12, 13) were as follows: 5-ATG GAY CAY TGG RTI TTY GGI GA-3 and 5-ATG AAG CAI ARI GGI CCR AAR TAY TG-3 (where I is inosine, Y is C or T, R is A or G). PCR conditions were 94 °C, 1 min, 55 °C, 2 min, and 72 °C, 1 min, for 36 cycles. PCR product of 370 bp was purified on 1.5% agarose gel, digested with BclI to remove amplified Y1 receptor cDNAs and subcloned. Several sequenced clones, including pG8 (Table I), represented a novel sequence related to NPY receptors. Plasmid pY3.12 and additional longer cDNAs for this novel gene (Table I) were obtained by screening of a human heart cDNA library (Stratagene, La Jolla, CA) using the insert of pG8 as the probe. Inserts of plasmids pT7(a-e) (Table I) were obtained by PCR of human genomic DNA of five unrelated individuals (Bios Laboratories, New Haven, CT) with the following primers: 5-CAA TGA GAA CAA GAG GAT CAA CAC-3 (FNPYRB) and 5-ACC AAG TGG CAA ACT ACA AAT ACC-3 (RNPYRB). Mouse genomic clones were obtained by screening of 500,000 recombinant phage from a mouse 129SV genomic library in phage  Fix II (16). Two positive phage were plaque-purified and one of these was selected for detailed analysis. A 1.6-kb HindIII fragment was subcloned into pcDNA3 (Invitrogen, San Diego, CA) and found to contain the entire coding region of the mouse PP2 gene. Manual sequencing was carried out by Lark Sequencing Technologies (Houston, TX) with Sequenase kit (U. S. Biochemical Corp.); in some sequencing reactions dITP was used. Automated sequencing was carried out at Yale University W. M. Keck facility, using cycle sequencing with TAQ polymerase, fluorescent-dideoxynucleotide terminators (Perkin-Elmer Corp.) and an Applied Biosystems 373A Stretch DNA sequencing apparatus. The nucleotide sequences of human and mouse PP2 DNAs have been deposited with Genbank^TM Data Library under accession nos. U59431[GenBank] and U59430[GenBank], respectively.

Table I. Summary of sequenced human PP2 DNAs containing a single base deletion All listed plasmids were determined to have a single base deletion as shown in Fig. 1 (following nucleotide 864 of human PP2 sequence). Plasmids were sequenced on both strands as described in experimental procedures. Plasmid Insert Insert source Cloning method Sequencinga Specific bindingb kb pG18 1.5 Heart cDNA library Library screening M, MI, A Not detectable pY3.12 1.8 Heart cDNA library Library screening M, MI, A Not detectablec pY3.12dH 1.5 pY3.12 derivative Library screening A Not detectable pG8 0.37 Brain mRNA RT-PCR A Not tested pG27 1.2 P1 genomic DNA PCR A Not tested pG23 1.2 Heart mRNA RT-PCR A Not detectable pY3.8 1.4 Heart cDNA library Library screening A, M Not tested pY3.14 1.2 Heart cDNA library Library screening A, M Not tested pT7(a-e)d 0.2 Genomic DNA PCR A Not tested a  M, manual sequencing; MI, manual sequencing with dITP; A, automated sequencing. b  Specific binding was tested with 125I-rPP and 125I-PYY as described under ``Experimental Procedures.'' c  pY3.12 was not tested for 125I-rPP specific binding. d  pT7(a-e) represents human genomic DNAs from five unrelated individuals of three ethnic backgrounds.

A sequence tagged site (STS) for human PP2 was developed with the following two primers: 5-CAT ATG AGA CAG GCA GTC TTA GC-3 (reverse primer RHYB) and 5-AGA TTG GCT CGT ATA ACA ACA GG-3 (forward primer FHYB). PCR with these two primers gave a 174-bp product, which could be confirmed by digestion with XmnI. This STS marker was used to type a panel of somatic cell hybrids for chromosome mapping (Bios Laboratories). Southern blots of genomic DNA digested with EcoRI and BamHI were purchased from Bios Laboratories. A P1 human genomic clone E2182 for PP2 was obtained by custom screening of a P1 human genomic library with cDNA insert of pY3.12 at Bios Laboratories. The identity of this P1 clone was confirmed by Southern blotting and by using our STS marker. FISH was carried out at Bios Laboratories. DNA from clone E2182 was labeled with digoxigenin dYTP by nick translation. Labeled probe was combined with sheared human DNA and hybridized to normal metaphase chromosomes derived from PHA stimulated peripheral blood lymphocytes in a solution containing 50% formamide, 10% dextran sulfate, and 2 × SSC. Specific hybridization signals were detected by incubating the hybridized slides in fluorescein conjugated antidigoxigenin antibodies. The chromosomes were then counter stained with propidium iodide and analyzed.

Multiple tissue Northern blots were purchased from Clontech Labs. Membranes were hybridized and washed under high stringency conditions as described previously (16). A human -actin cDNA was used to probe all filters to control for unequal loading or transfer. For RT-PCR, human poly(A)-containing RNA from various tissues (Clontech Labs) was reverse transcribed and resulting cDNAs were used for PCR with the primers FHYB and RHYB as described above. The PCR product of 174 bp was resolved on 1.5% agarose gels and confirmed by Southern blot analysis.

COS-7 cells were grown and transfected using the LipofectAMINE method (Life Technologies, Inc.) as described previously (16). ¹²⁵I-Labeled rat pancreatic polypeptide (¹²⁵I-rPP), ¹²⁵I-labeled rat peptide YY (¹²⁵I-PYY), and rat ¹²⁵I-NPY were purchased from DuPont NEN. PYY, NPY, and (2-36)PYY were synthesized by Bayer Corp. Rat pancreatic polypeptide (rPP), human pancreatic polypeptide (hPP), and (Leu³¹Pro³⁴)NPY were purchased from Peninsula Laboratories Inc. (Belmont, CA). Expression vectors were constructed in pcDNA3 (Invitrogen). Samples for binding consisted of 2-150 µg of membrane protein, 100 pM ¹²⁵I-rPP, and peptide concentration in a final volume of 200 µl. Nonspecific binding was defined by 1 µM rPP. The binding assays were performed on GF/C Millipore 96-well plates as described previously (16). Data were analyzed using the nonlinear regression curve-fitting program RS/1 (BBN Software Products Corp., Cambridge, MA).

To remove the 5-untranslated region of clone Y3.12, we utilized a unique HindIII site within the 5 leader. Plasmid pY3.12 was cut with HindIII and religated. The resultant plasmid pY3.12dH had a 5-untranslated leader of ~100 bp. To insert a single T base following the CCCC sequence in TM6, we used PCR mutagenesis. PCR primer TFIXNPYB, 5-TCC ATC GTG GTG ACC TTT GGA GCC TGC TGG CTG CCC CTG AAT ATC TTC AAT G-3, was designed to take advantage of a unique BstEII site upstream of the mutated position. PCR was carried out using pY3.12dH as a template and primers TFIXNPYB and SP6. The PCR product was digested with BstEII and EcoRV and purified. pY3.12dH was digested with BstEII and EcoRV, purified, and ligated to PCR fragment BstEII-EcoRV. The resultant plasmid pG22 was sequenced from BstEII site until 3 end to confirm the sequence and the single base insertion.

RESULTS

The nucleotide and deduced amino acid sequences of the mouse PP2 receptor and its human homologue are shown in Fig. 1. The mouse sequence was obtained from a genomic clone, showed no evidence of introns, and revealed a 1113-bp open reading frame encoding 371 amino acid residues with a calculated molecular mass of 42,713 Da (Fig. 1). The nucleotide sequence identity between mouse and human PP2 is 83% within the coding region.

Sequence of the human PP2 was determined with clones pY3.12 and pG18, which represent cDNAs isolated from a human heart cDNA library. The sequences of pY3.12, pG18, and all other cDNAs for the human PP2 consistently showed a frameshift in the coding region. Sequence alignment with mouse PP2 as well as with other members of the NPY receptor family determined that the frameshift is due to a single base deletion in the region corresponding to TM6, following a proline residue (Fig. 1, nucleotide 865). A summary of all sequenced human PP2 DNAs is shown in Table I. Both strands of clone pY3.12 were sequenced completely by two independent methods, manual and automated sequencing. As indicated in Table I, several additional cDNAs were sequenced in the critical region with identical result. Furthermore, the same single base deletion was detected in a P1 human genomic clone and in human genomic DNA of five unrelated individuals of three ethnic backgrounds (African-American, Asian, and Caucasian).

Analysis of mouse PP2 protein for regional hydrophobicity revealed seven putative transmembrane domains (TMs), the typical hallmark of G protein coupled receptors. There are three potential N-linked glycosylation sites, two in the NH₂ terminus and one in the second extracellular loop connecting TMIV and TMV. The mouse PP2 protein contains residues conserved in many G protein receptors, including an acidic residue (Asp) in TMII and cysteine residues in the first and second extracellular loop that may form a disulfide bridge (23). The carboxyl-terminal region contains 4 serines and 3 threonines that may be substrates for protein kinases and cysteine residues, which may be sites for fatty acid modification (24, 25). Within human PP2 cDNA, a single base insertion following to sequence CCCC in TM6 would shift the reading frame and extend it throughout TM7 and the COOH terminus, similar to mouse PP2 (Fig. 1, reading frame shown in italics). Comparison of the mouse PP2 protein with human PP2 (up to TM6) reveals 78% identity. Additional comparisons of the mouse PP2 sequence reveal 42% identity with the cloned Y4/PP1, 53% with mouse Y1 receptor and lower, 31%, identity with the cloned Y2 receptor (Table II). A Drosophila NPY receptor (26) has only 28% identity with mouse PP2, and unrelated G protein receptors have approximately 25-29% identity. An alignment of the mouse PP2 with other cloned mammalian NPY/PYY/PP receptors is shown in Fig. 2.

Table II. Percent sequence identity between members of the NPY/PYY/PP receptor family (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21) Abbreviations are as in Fig. 2. Percent identity hY1 mY1 hY2 hY4 mY4 hPP2 mPP2 hY1   94 31 43 45 51 52 mY1   31 46 46 51 53 hY2   33 33 29 31 hY4/PP1   76 38 42 mY4/PP1   38 42 hPP2   78 mPP2

The coding region of the mouse PP2 was cloned into the expression vector pcDNA3 and the resultant plasmid pG29 was transfected into COS-7 cells. Binding of ¹²⁵I-rPP to COS-7 cell membranes was linear up to at least 50 µg of protein, and typically greater than 75% specific binding was observed with ¹²⁵I-rPP. Of the peptides tested, both rat and human PP had the highest affinity for PP2 (Fig. 3). There were only small differences observed when comparing the pharmacological profiles of mouse PP2 and Y4/PP1 receptors (Table III). Specific binding of ¹²⁵I-PYY to membranes was lower, with relatively high background of nonspecific binding, and was not further investigated.

Table III. IC50 values for various NPY family peptides in competition with 125I-rPP binding to transiently expressed mouse PP2 receptors, as well as mouse Y4/PP1 receptors (16) for comparison Values are the average ± standard error of the mean for at least three independent experiments. Peptide IC50 Mouse PP2 Mouse Y4/PP1 nM Human PP 0.057  ± 0.007 0.054 Rat PP 0.065  ± 0.013 0.097 (Leu31Pro34)NPY 4.5  ± 1.1 4 PYY 514  ± 140 500 NPY 638  ± 126 790 (2-36)NPY >1,000 >1,000 (3-36)NPY >1,000 >1,000 (13-36)NPY >1,000 >1,000

In contrast to mouse PP2, transient expression of human PP2 in COS-7 cells did not yield any specific binding to ¹²⁵I-rPP, ¹²⁵I-PYY, or ¹²⁵I-NPY. A number of different plasmids were tested with identical results (Table I). In addition, the plasmid construct pY3.12dH, which had its 5-untranslated leader sequence removed (see ``Experimental Procedures''), did not show any specific binding to the above radioligands. Furthermore, the plasmid construct pG22, which had a single base T insertion to restore the reading frame found in the mouse PP2 sequence, did not show any specific binding to the above radioligands. In control experiments, NPY Y1 and mouse PP2 constructs tested positive.

The human PP2 gene coding region could be amplified by PCR from our P1 genomic clone (plasmid pG27, Table I) and showed no evidence for introns, similar to the mouse gene. Southern blot analysis of mouse and human genomic DNA with mouse and human PP2 probes suggested the existence of a single gene in both species. Cross-hybridization with mouse PP2 probe suggested that its human species homologue is indeed the human PP2 gene (Fig. 4).

Typing of a somatic cell hybrid panel with our STS marker revealed that PP2 gene maps to human chromosome 5 (data not shown). To map the human PP2 gene more precisely, a human genomic P1 clone was isolated and used for FISH. This experiment resulted in the specific labeling of the long arm of a group B chromosome. A second experiment was conducted in which a genomic probe containing the nucleolar protein gene NPM, which is known to localize to 5q35 (27), was cohybridized with our P1 plasmid E2182 in order to confirm the identity of the specifically labeled chromosome as chromosome 5. Measurements of 10 specifically labeled chromosomes 5 determined that E2182 hybridization signal is located at a position that is 70% of the distance from the centromere to the telomere of the chromosome arm 5q, an area that corresponds to 5q31 (Fig. 5). A total of 80 metaphase cells were analyzed with 74 exhibiting specific labeling.

Northern analysis of mouse and human tissues was performed to examine transcript sizes and regional differences in mRNA abundance. The human PP2 message was detected as an abundant 3.4-kb transcript in the heart. At lower levels, the same transcript was detectable also in skeletal muscle, gastrointestinal tissues, adrenal glands and some other tissues (Fig. 6). Using RT-PCR, the message was also detectable in various parts of the human brain (data not shown). In the adult mouse tissues, PP2 message was not detectable by Northern blot analysis (data not shown). However, in the developing mouse, a 9.8-kb PP2 message was detected in 7-day-old embryo but not in 11-, 15-, and 17-day-old embryo (Fig. 7).

DISCUSSION

The results presented here show that the PP2 gene is a novel member of the NPY/PYY/PP receptor family. The mouse PP2 receptor has very high affinity to PP, indicating that PP may be one of its endogenous ligands. Both Y4/PP1 and PP2 receptors may also be able to interact with PYY and NPY in vivo, although we have not investigated ¹²⁵I-PYY binding to mouse PP2 receptors in detail. The affinities of PYY and NPY were low in the ¹²⁵I-rPP assay, as is the case for mouse Y4/PP1 receptors (16). The percent identity between mouse Y4/PP1 and mouse PP2 proteins is relatively low (42%) considering that both receptors have very similar affinities to rat and human PP and similar rank orders of potencies for other peptides tested (Table III). While we have not investigated coupling to second messenger systems, similarity of mouse PP2 receptors to Y4/PP1, Y1 and Y2 receptors within second and third intracellular loops (Fig. 2) may indicate potential coupling to adenylate cyclase and calcium signaling systems (13, 15, 17).

Surprisingly, the human PP2 homologue displays properties of an inactivated gene, i.e. pseudogene. The single base deletion in the human PP2 gene is not due to a cloning or sequencing artifact, since DNAs were cloned by three independent methods and sequenced by three different methods. Furthermore, the deletion was found in several different cDNAs and in genomic DNAs from five unrelated individuals of three ethnic backgrounds (Table I). The deletion predicts a truncated protein without the seventh transmembrane region and the cytoplasmic carboxyl terminus. Such deletion likely results in a nonfunctional receptor, since G protein-coupled receptors generally require TM7 for activity (e.g. see Unson et al. (28)). A number of different constructs containing the coding region of human PP2 were transfected in COS-7 cells, but none of these bound to ¹²⁵I-rPP, suggesting a loss of function. It is possible, however, that functional human PP2 polypeptide is not expressible in COS-7 cells. Although we have not formally shown protein production from our various human PP2 expression constructs, control experiments with mouse PP2 and rat Y1 DNAs (constructed in the same pcDNA3 expression vector) tested under identical conditions, validate the common COS-7 cell transient expression system that we employed. Insertion of a single base following to CCCC sequence in TM6 would correct the reading frame of human PP2, so that its COOH terminus would be similar to and co-linear with mouse PP2 (Fig. 1, reading frame in italics). We have constructed plasmid pG22 with such an insertion (insertion of a single T base predicts a Leu-277 following Pro-276) and determined that this construct also failed to direct production of functional PP/PYY/NPY receptors in COS-7 cells. This may indicate that the coding region of human PP2 has accumulated additional mutations that have inactivated the active site for pancreatic polypeptides. The relatively abundant expression of human PP2 transcripts in the heart and several other tissues is of unknown significance; it may represent an attempt at compensatory up-regulation of the PP2 gene expression. In conclusion, the human PP2 gene is mutated, does not seem to interact with pancreatic polypeptides and may represent a pseudogene. Less likely, the truncated human PP2 protein may serve some other function, which does not require interaction with pancreatic polypeptides. In some instances, truncated GPCRs have been re-activated by coexpression with other truncated GPCRs (29), and cytoplasmic domains of GPCRs might inhibit GPCR signaling (30).

Mouse PP2 message is detectable in 7-day-old mouse embryo, which is prior to the formation of pancreas by invagination of duodenal endoderm. It would be interesting to determine if PP2 message is co-localizing to the pancreatic anlage, which is already noticeable before 20-somite stage (about day 8.5-9) (31). The ligand for embryonic PP2 receptors at this early stage of development is probably not PP, since PP message can be detected in embryonic foregut or pancreas for the first time only later, at 30-somite stage (about day 10) (31); immunohistochemical data place onset of PP expression to a much later stage (32). These issues are of interest since they may shed light on a central question in organogenesis of the pancreas, i.e. identification of a common, multipotential progenitor cell from which endocrine pancreatic islet cells arise (31, 32, 33, 34). Available data indicate that the earliest expressed islet hormone gene may be PYY (34), which is a possible endogenous ligand for mouse PP2 receptor. Future studies will be also required to elucidate the reasons for dramatic evolutionary change in structure and function of mouse versus human PP2 receptors.