Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Extracellular nucleotides have been found to modulate a variety of physiological processes. ATP evokes fast excitatory responses from neuronal cells, affects vascular tone, and may act as an autocrine and/or paracrine agent in the neuroendocrine system (Ralevic and Burnstock, 1998). The effects of ATP are mediated through two classes of cell surface receptors, a G protein-coupled, heptahelical family called P2Y receptors, and a family of ligand-gated ion channels called P2X (Abbracchio and Burnstock, 1994; Fredholm et al., 1997).

There are seven known members of the P2X family of ATP-gated ion channels cloned from rat tissues, designated P2X_1-7. These receptor subtypes differ with respect to tissue distribution, as well as to pharmacologic and kinetic profiles (Brake et al., 1994; Valera et al., 1994; Bo et al., 1995; Chen CC et al., 1995; Collo et al., 1996; Surprenant et al., 1996). The P2X receptor proteins share ~30 to 50% sequence identity and are 379 to 595 amino acids in length. Structurally, the receptors are characterized by two transmembrane (TM) domains with both amino and carboxyl ends being located cytoplasmically (Brake et al., 1994; Valera et al., 1994). The extracellular domains of these receptors are glycosylated and contain 10 conserved cysteine residues that participate in intrachain disulfide linkages (Newbolt et al., 1998). When expressed in heterologous systems, the P2X receptors form nonselective cation channels with differing sensitivities to ATP and other P2 receptor agonists, as well as to antagonists such as pyridoxal-5-phosphate-6-azophenyl-2',4'-disulfonic acid (PPADS) and suramin (reviewed in Soto et al., 1997). The receptors fall into two categories with respect to channel kinetics, with P2X₁ and P2X₃ forming relatively fast-desensitizing channels, and P2X₂, P2X₄, and P2X₇ forming channels having a slower rate of desensitization (Soto et al., 1997). Although expression of a single P2X subtype in heterologous systems is sufficient to create functional ion channels, it has been demonstrated that subunits from different P2X subtypes can associate to form heteromeric channels with unique characteristics (Lewis et al., 1995; Lê et al., 1998, 1999; Torres et al., 1998). Evidence has been most convincing for the association of P2X₂ and P2X₃ subunits, with the properties of the resulting heteromeric channel closely resembling those of native channels found in small-diameter nociceptive neurons of sensory ganglia (Lewis et al., 1995; Thomas et al., 1998).

The P2X₂ receptor subtype originally was isolated from PC12 rat pheochromocytoma cells (Brake et al., 1994). It has been found to be expressed in a variety of organs of the vascular, nervous, neuroendocrine, and sensory systems. Alternatively spliced messages for the P2X₂ receptor have been isolated from rat and guinea pig tissue, and functional studies have suggested that alternative splicing of the receptor serves to modulate the response of the receptor to prolonged agonist exposure (Housley et al., 1995; Brandle et al., 1997; Simon et al., 1997; Koshimizu et al., 1998b; Parker et al., 1998).

This report describes the isolation of novel human P2X receptor cDNAs. Based on structure as well as on pharmacologic profiles, it is concluded that they represent human homologs of the P2X₂ receptor. The human P2X₂ gene, as well as several splice variants of the receptor message, has been isolated, and the tissue expression and functional activity of these variants have been characterized.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. ATP, 2-methylthio-ATP tetrasodium (2-meSATP), -methylene ATP dilithium (,-meATP), suramin hexasodium, and PPADS were obtained from Research Biochemicals International (Natick, MA). 2' and 3'-O-(4-benzoylbenzoyl)-ATP tetraethylammonium salt (mixed isomers, BzATP), ADP5'-O-(3-thiotriphosphate) tetralithium salt (ATPS), ADP, and UTP were obtained from Sigma Chemical Co. (St. Louis, MO). ADP5'-O-(2-thiodiphosphate) trilithium salt (ADPS) and G418 sulfate were obtained from Calbiochem-Novabiochem Corp. (La Jolla, CA). Dulbecco's modified Eagle's medium (with 4.5 mg/ml glucose and 4 mM L-glutamine) and fetal bovine serum were obtained from Hyclone Laboratories, Inc. (Logan, UT). Dulbecco's PBS (with 1 mg/ml glucose and 3.6 mg/l Na pyruvate, without phenol red), hygromycin, Lipofectamine, restriction enzymes, and buffers were obtained from Gibco BRL Life Technologies (Gaithersburg, MD). Fluo-3-acetoxymethyl ester (fluo-3-AM) was obtained from Molecular Probes (Eugene, OR). -[³²P]UTP was obtained from Amersham, Inc. (Arlington Heights, IL).

Identification of Human P2X₂ Expressed Sequence Tag and Generation of Partial cDNA Clones. The TBLASTN database search tool (National Center for Biotechnology Information, Washington, DC) (Altschul, 1993) was used with the predicted amino acid sequence of the rat P2X₂ receptor (GenBank accession no. 1352688) to search the Lifeseq database (Incyte Pharmaceuticals, Inc., Palo Alto, CA) for human DNA sequences that would code for similar polypeptides. A sequence was identified from a cDNA clone derived from human fetal colon tissue that encoded a polypeptide with homology to a region of the rat P2X₂ receptor.

Isolation of the Human P2X₂ Gene. A polymerase chain reaction (PCR) primer pair was designed and synthesized based on the sequence compiled from the Incyte (Incyte Pharmaceuticals, Inc.) clone and partial 5'-RACE product, which yielded a product of 339 bases when amplifying human genomic DNA (sense, 5'-TCCTTCCTGTGCGACTGGATCTTG-3'; antisense, 5'-CAAACCTTTGGGGTCTGTGGGTG-3'). Using these primers, a P1 bacteriophage library screen was performed (Genome Systems, Inc., St. Louis, MO). Two clones containing the human P2X₂ gene were obtained and sequenced as above. The P1 phagemid DNAs were prepared for sequencing with Qiagen (Chatsworth, CA) Maxiprep plasmid purification system using a protocol supplied by the manufacturer. To facilitate sequencing, one of the clones was digested with the restriction enzyme BamHI and subcloned into the pBluescript II SK+ vector using the pCRscript cloning system (Stratagene, La Jolla, CA).

Isolation of Human P2X₂ cDNAs. Using the genomic sequence information surrounding the predicted initiation and termination codons of the human P2X₂ message, oligonucleotide primers were designed and synthesized to enable reverse transcription (RT)-PCR of the intact open reading frame of the mRNA. A consensus translation initiation signal was designed into the sense primer to optimize expression (Kozak, 1984). First-strand cDNA was synthesized from polyA+ RNA derived from pituitary gland tissue (1 µg; Clontech), using 10 pmol random hexamer primer, and Superscript II reverse transcriptase (Gibco BRL Life Technologies). A proofreading thermostable polymerase (cloned Pfu DNA polymerase; Stratagene) was used in the amplification to ensure high-fidelity amplification. The reaction mixture consisted of: 2 µl of cDNA, 10 pmol of each primer; hP2X₂ sense (5'-CCACCATGGCCGCCGCCCAGCCCAAGTA-3'); hP2X₂ antisense (5'-GGAAAGGAGCTCAGAGTTGAGCCAAACC-3'); and 1× Pfu reaction buffer and 200 µM 2'-deoxynuceleoside 5'-triphosphates. The reaction was preheated to 80°C before addition of the polymerase, after which the reaction was cycled 35 times under these conditions: 94°C for 15 s, 60°C for 20 s, and 72°C for 5 min. After cycling, the reaction was incubated for 10 min at 70°C. The reaction products were separated on a 0.8% agarose gel and products of ~1.5 kilobases (kb) were excised and purified via the Qiaquick (Qiagen) gel purification system. The products were subcloned into the pCRscript II+ vector and sequenced. To create the hP2X_2a message, an internal KpnI site was used to splice nucleotides 1 to 666 of hP2X_2d with nucleotides 595 to 1349 of hP2X_2c. The coding sequences of the clones were found to be identical with the predicted exons of the human P2X₂ genomic sequence isolated from the P1 library.

Northern Analysis. An -[³²P]UTP radiolabeled antisense RNA probe was synthesized from the hP2X_2a message using the T3 promoter of pCRscript vector and reagents from the Maxiscript in vitro transcription kit (Ambion, Inc., Austin, TX). A commercially prepared membrane Multiple Tissue Northern blot (Clontech) was probed overnight at 72°C in hybridization buffer (NorthernMax; Ambion). The blot was washed at 72°C in high-stringency buffer (NorthernMax). The hybridized membranes were analyzed on a Storm phosphoimager (Molecular Devices, Sunnyvale, CA).

RT-PCR Analysis of Tissues. PolyA+ RNAs from human tissues were obtained from Clontech. RNA (0.5 µg) was used in random hexamer-primed reverse transcriptase reactions using Thermoscript RT-PCR reagents (Gibco BRL Life Technologies). PCR primers were designed corresponding to nucleotides 2246 to 2269 (sense) and 3733 to 3710 (antisense) of the hP2X₂ gene. One-tenth of each RT reaction was used in PCR amplification reactions that also included 10 pmol of each primer, 1.5 mM MgSO₄, 200 µM 2'-deoxynucleoside 5'-triphosphates, 1× PCR× amplification buffer (Gibco BRL Life Technologies), 0.5 × PCR× enhancer (Gibco BRL Life Technologies), and 2.5 U platinum Taq polymerase (Gibco BRL Life Technologies). After initial incubation at 95°C for 5 min, the reactions were cycled 35 times under the following conditions: 94°C for 15 s, 55°C for 20 s, and 68°C for 1 min. The products were analyzed by agarose gel electrophoresis. To identify the products, they were isolated, purified (Qiaquick; Qiagen), and directly sequenced.

Cloning of the Rat P2X₂ and Human P2X₃ Receptors. The rat P2X₂ and P2X_2b receptor cDNAs were isolated through RT-PCR reactions from oligo(dT)-primed cDNA synthesized from total brain polyA+ RNA (Clontech). Primers were designed based on the published sequence for the rat P2X₂ receptors (GenBank accession nos.: rat P2X₂, U14114; rat P2X_2b, Y10473). The primers used were rP2X₂ sense, 5'-CACCATGGTCCGGCGCTTGGCCCGGGGC-3'; rP2X₂ antisense, 5'-TCAAAGTTGGGCCAAACCTTTGGGGTCCG-3'. The PCR amplification was set up using Pfu polymerase essentially as described above, and the reactions were cycled 35 times under the following conditions: 94°C for 20 s, 65°C for 20 s, and 72°C for 4 min. Reaction products were separated by agarose gel electrophoresis, and the major products of approximately 1.5 and 1.3 kb were isolated and cloned into the pCRscript vector. Inserts were sequenced and found to be identical with the published sequences for the rat P2X₂ (Brake et al., 1994) and P2X_2b receptors (Simon et al., 1997).

Electrophysiology of hP2X₂ Receptors. Adult female frogs (Xenopus laevis) were anesthetized with 0.2% tricaine, and sections of one ovary were removed. Oocytes were denuded of overlying follicle cells by agitation for 1 to 2 h in low-Ca²⁺ Barth's solution containing 88 mM NaCl, 2.5 mM KCl, 1.0 mM MgCl₂, and 10 mM Na-HEPES (pH 7.55) plus 2 mg/ml collagenase (Sigma Chemical Co.; type IA) and antibiotics. Selection of stage V and VI oocytes was begun after ~50% of the cells were denuded. Oocytes were maintained at 16 to 19°C in normal Barth's solution containing 90 mM NaCl, 1.0 mM KCl, 0.66 mM NaNO₃, 0.74 mM CaCl₂, 0.82 mM MgCl₂, 2.4 mM NaHCO₃, 2.5 mM Na-pyruvate, 10 mM Na-HEPES (pH 7.55) plus antibiotics. Cytoplasmic injections of 50 nl of 1 ng/nl cRNA or intranuclear injections of 12 nl of 1 ng/nl cDNA encoding hP2X₂ receptors were given within 24 h of preparation. For hP2X₂ and hP2X₃ coinjections (hP2X_2/3), equal concentrations of both cRNAs were premixed before a single 50-nl injection. The standard recording solution contained 96 mM NaCl, 2.0 mM KCl, 1.8 mM BaCl₂, 1.0 mM MgCl₂, 5.0 mM Na-pyruvate, and 5.0 mM Na-HEPES (pH 7.4). For long agonist applications, the recording solution contained 115 mM NaCl, 2.5 mM KCl, 1.8 mM MgCl₂, and 10 mM Na-HEPES (pH 7.3) (Brandle et al., 1997). Ca²⁺-free saline was used to minimize the contribution of endogenous calcium-dependent chloride channels to the ATP response.

Receptor Expression in Mammalian Cells. Human P2X₂ cDNAs were transferred into the mammalian expression vector pIRES(hyg) (Clontech) using a NotI and partial BamHI restriction digest of the cDNA in pCRscript and the corresponding sites of the pIRES vector. The resulting plasmids were sequenced and introduced into the 1321N1 human astrocytoma cells via transfection with Lipofectamine (Gibco BRL Life Technologies). Stable transformants were selected by resistance to hygromycin. Surviving individual colonies were isolated and screened for P2 receptor activity. The clones exhibiting the largest ATP-induced response were selected for additional characterization. Cells expressing recombinant P2X receptors were maintained at 37°C in Dulbecco's modified Eagle's medium with 4.5 mg/ml¹ glucose and 4 mM L-glutamine), 10% fetal bovine serum, and 100 µg/ml¹ hygromycin in a humidified 5% CO₂ atmosphere.

Ca²⁺ Flux Measurements. The pharmacologic characterization of the P2X₂ receptors was performed on the basis of agonist-mediated increases in cytosolic calcium concentration. A fluorescent calcium-chelating dye (fluo-3) was used as an indicator of the relative levels of intracellular calcium in a 96-well format using the Fluoresence Imaging Plate Reader (Molecular Devices, Sunnyvale, CA). Cells were grown to confluence and loaded with the fluo-3-AM (2 µM) in Dulbecco's PBS for 1 to 2 h at room temperature. Before the assay, each 96-well tissue culture plate was washed three times with Dulbecco's PBS to remove extracellular fluo-3-AM. Agonists were added 10 s after the start of the experimental run. Antagonists were added 10 s after the start of the experimental run, followed by agonists after a 180-s interval (t = 190 s). Fluorescence data were collected at 1- to 5-s intervals throughout each experimental run. Dose-response data were analyzed using a four-parameter logistic Hill equation in GraphPad Prism (GraphPad Software Inc., San Diego, CA).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

In an attempt to identify novel human P2X receptors, a search was performed on the Lifeseq (Incyte Pharmaceuticals, Inc.) database using the rat P2X₂ amino acid sequence as a query. The search identified one entry that had significant homology to the query and no identity to other known human or rat P2X receptors. Because the position of this sequence with respect to that of the rat P2X₂ sequence predicted that this was a partial cDNA, a series of antisense primers were designed to enable identification of the balance of the predicted open reading frame. The sequence was extended using 5'-RACE reactions. However, this approach was unsuccessful at generating products with open reading frames encoding a full-length P2X receptor. Hybridization screens of cDNA libraries also did not identify clones containing sequences encoding the N terminus of the receptor.

The human P2X₂ receptor gene was isolated by a PCR screen of a P1 phagemid library in an attempt to characterize the remaining receptor coding sequence. As shown in Fig. 1, the human P2X₂ gene contains 11 exons, retaining the structural rules for intron/exon boundaries (Mount, 1982). The gene structure is highly conserved between human and rat (Brandle et al., 1997).

View larger version (56K): [in this window] [in a new window]   Fig. 1.   Genomic organization of the human P2X2 gene. Eleven exons are blocked to phase, whereas the amino acid sequence is shown below the corresponding nucleotide sequence. The positions of the primers used in RT-PCR analysis of tissues are indicated by arrows. The sequence has been deposited into the GenBank database under accession no. AF190826.

Based on the human P2X₂ genomic sequence and on the size and structure of the rat P2X₂ open reading frame, primers were designed that encompassed the initiation and termination codons of the putative human P2X₂ receptor message. RT-PCR reactions were performed using polyA+-enriched RNA from human pituitary tissue, generating a series of products from 1.2 to 1.4 kb in length. These products were cloned and screened by sequencing. Three predominant messages were isolated initially. These sequences were compared with the P2X₂ genomic structure to deduce the origin of the structure and to verify the sequence. One message (called P2X_2b) is formed through the use of an internal splice site within exon 11, deleting a 201-bp sequence from the coding sequence (Fig. 2A). This deletion is in the identical position with that of the functional splice variant shown to be produced from the rat P2X₂ gene (Brandle et al., 1997). The second variant (called hP2X_2c) is formed by the in-frame deletion of exon 3 from the coding sequence (Fig. 2B). The third variant (called hP2X_2d) is formed from the inclusion of intron 10 in the message (Fig. 2C). Because the intron continues the open reading frame, this results in a 26-amino acid insertion within the predicted second TM region of the receptor. This variant also has been detected in the rat; however, the rat intron contains a termination codon, thereby encoding a truncated receptor (Housley et al., 1995). Several other P2X₂ cDNAs were isolated that contained deletions of several exons, resulting in drastically truncated open reading frames (data not shown). Based on the lack of function of the hP2X_2c form (see below), it was deemed unlikely that these cDNAs would encode functional receptors and they were not characterized further.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Splice variants of the human P2X2 message. A, the hP2X2b variant formed by splice donor and acceptor sites lying within exon 11. The resulting message removes 201 bp (67amino acids) from the C-terminal domain of the receptor. B, removal of exon 3 from the P2X2 message forms the hP2X2c variant. This deletes 24 amino acids from the putative extracellular domain of the receptor. C, the hP2X2d form is created by inclusion of intron 10 in the mRNA (lower case nucleotide sequence). This creates an insertion of 26 amino acids to the predicted polypeptide sequence (bold) in the predicted second TM region of the receptor. The sequences have been deposited into the GenBank database under accession numbers AF190822 hP2X2a, AF190823 hP2X2b, AF190824 hP2X2c, and AF190825 hP2X2d.

Initial cloning efforts failed to identify a message encoding a receptor homologous to the prototypic rat P2X₂ receptor, e.g., that having an intact C-terminal tail and no other deletions or insertions in exon sequences as predicted from the genomic sequence (Brake et al., 1994). To determine whether such a polypeptide would create a functional ATP-gated ion channel, we constructed a cDNA using portions of two of the hP2X₂ variants. This message, called hP2X_2a, was generated by splicing the 5' half of the hP2X_2d message to the 3' end of the hP2X_2c message, using a common KpnI restriction enzyme site. The resulting message encodes a 471-amino acid polypeptide with 68% identity with the prototype rat P2X₂ receptor (Fig. 3). The sequence identity of the predicted extracellular domains of the human and rat P2X₂ receptors is especially high. Within this region of the human polypeptide are three consensus sites for N-linked glycosylation at residues 133, 194, and 310. There also are three sites in the rat sequence, and two of these (at positions 194 and 310) are conserved positionally between the species. The N-terminal cytoplasmic domain of the human P2X_2a receptor is 12 amino acids longer than the rat homolog, but it is similar in length and sequence to the guinea pig P2X₂ receptor (human, MAAAQPKYPAGATAR... and guinea pig, MAATHPKAPTAQRLR... ) (Parker et al., 1998). Conversely, the C- terminal tail of the human P2X_2a receptor lacks 12 amino acids when compared with the rat and guinea pig receptors.

View larger version (53K): [in this window] [in a new window]   Fig. 3.   Alignment of the human and rat prototypic P2X2 receptor polypeptides. Analysis performed using the ClustalW algorithm and the BLOSUM 30 scoring matrix in MacVector 6.0 (Oxford Molecular Group, Oxford, UK). Identical residues are blocked and darkly shaded, and similar residues are blocked and lightly shaded. Predicted TM domains are illustrated by solid bars. The boundaries of the region deleted in the hP2X2b form of each species are illustrated by vertical bars ().

Northern analysis using a radiolabeled P2X₂ riboprobe detected a major mRNA species of ~2.4 kb in the pancreas (Fig. 4A). Extended exposures of the membrane detected faint bands of the same size in heart and brain (data not shown). Another faint band of ~5.0 kb was detected in all tissues except lung. A third, larger species was detected in the brain. Several tissues were analyzed by RT-PCR to examine distribution of the various P2X₂ splice variants. A set of primers was designed to flank exons 10 and 11 (see Fig. 1). This allowed identification of the hP2X_2b, hP2X_2d, and hP2X_2a/c products (products arising from the hP2X_2a and hP2X_2c form are identical). Results shown in Fig. 4B indicate that four expected primary products were obtained in all tissues except spinal cord, from which only two products were amplified. Direct sequencing of the products determined that the following were present (in descending order of size): the hP2X_2d form (879 bp), the hP2X_2a/c form (801 bp), and the hP2X_2b form (600 bp). The primary products detected in the spinal cord were derived from the hP2X_2d and hP2X_2b messages.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Tissue distribution of human P2X2 RNA. A, Northern analysis using the full-length hP2X2a probe. Size markers shown are shown on the left and expressed in kilobases (Kb.). B, RT-PCR of various human tissues with P2X2-specific primers. Reactions were performed with (+) or without () reverse transcriptase. The products arising from the hP2X2d, hP2X2b, and hP2X2a/c are labeled. A 100-bp ladder (Gibco BRL Life Technologies) is indicated by the high-intensity 600-bp band.

Four human and two rat P2X₂ messages were injected into Xenopus oocytes for functional characterization. Short (10 s) applications of 10 to 30 µM ATP were used for current-voltage relationship and antagonist studies. These applications produced nondesensitizing currents in oocytes expressing either the hP2X_2a or the human hP2X_2b receptor splice variant. Current-voltage relationships in oocytes expressing either hP2X_2a (n = 5) or hP2X_2b (n = 3) receptors were inwardly rectifying and reversed near 0 mV (Fig. 5C), indicating activation of a nonselective cation conductance. The P2X receptor antagonist PPADS was used to determine antagonist sensitivity of the hP2X_2a and hP2X_2b receptors. PPADS (10 µM) completely blocked ATP-evoked responses in oocytes expressing either hP2X_2a (n = 3) or hP2X_2b (n = 2) receptors, but had poor reversibility during wash. A high concentration of the ATP analog ,me-ATP (100 µM) produced no inward currents in oocytes injected with either receptor subtype, consistent with previous reports for the rat P2X₂ receptor (Brake et al., 1994). In contrast to hP2X_2a and hP2X_2b receptors, ATP (30 µM) failed to evoke a current response in oocytes injected with hP2X_2c or hP2X_2d cRNA.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Functional expression of P2X2 receptors in oocytes. A, representative inward currents evoked by 60-s applications of 300 µM ATP in oocytes expressing either hP2X2a (top) or hP2X2b (bottom) homomeric receptors. Desensitization time constants were similar between hP2X2a and hP2X2b. B, inward currents evoked as in A in oocytes expressing either rP2X2a (top) or rP2X2b (bottom) homomeric receptors. Significantly faster desensitization was seen in oocytes expressing rP2X2b receptors. C, current-voltage relationships for currents evoked by 30 µM ATP in oocytes expressing either hP2X2a  or hP2X2b  receptors. D, representative inward currents evoked by 10 µM , me-ATP in oocytes expressing either hP2X2a/3 (top) or hP2X2b/3 (bottom) heteromeric receptors. For A, B, and D, cells were voltage-clamped at 60 mV. Agonist application is denoted by the bar. For A and B, calibration bars are 1000 nA and 20 s, except hP2X2a, which is 400 nA and 20 s. For D, calibration bars are 400 nA and 10 s.

Longer applications (60 s) of a high concentration of ATP (300 µM) elicited slowly desensitizing inward currents in oocytes expressing either hP2X_2a (desensitization  = 106 ± 27 s; n = 10) or hP2X_2b ( = 102 ± 11 s; n = 9) (Fig. 5A). At the end of the 60-s ATP application, current had decayed to 84 ± 2% of peak (n = 11) for hP2X_2a, and to 84 ± 1% of peak (n = 11) for hP2X_2b receptors. Similarly, applications of 300 µM ATP for 60 s to oocytes expressing prototype rat P2X₂ receptors produced slowly desensitizing responses ( = 115 ± 21 s; n = 5), which decayed to 65 ± 3% of peak (n = 5) by the end of the application. However, as described previously (Simon et al., 1997), rat P2X_2b receptors exhibited more rapid ( = 17 ± 1 s; n = 6) and complete (14 ± 1% of peak; n = 6) desensitization (Fig. 5B). Although the human P2X_2b form is the structural homolog of the rat P2X_2b receptor (see Fig. 3), obvious differences exist with respect to the rate and extent of desensitization.

In addition to characterization of homomeric hP2X₂ receptor channels, the ability of hP2X_2a-d to form heteromeric receptors with hP2X₃ was investigated. Expression of the heteromeric receptor was confirmed by the appearance of a new phenotypic receptor having characteristics of both hP2X₂ and hP2X₃ receptors. ,-meATP was used as an agonist at these receptors to eliminate activation of any hP2X₂ homomeric receptors that also may have been expressed. In oocytes expressing either hP2X_2a/3 (n = 3) or hP2X_2b/3 (n = 3) receptors, ,-meATP (10 µM) evoked inward currents that did not desensitize in the presence of agonist (Fig. 5D), consistent with activation of the heteromeric receptor. For hP2X_2c/3- and hP2X_2/3-coinjected oocytes, ,-meATP evoked rapidly desensitizing inward currents that were indistinguishable from hP2X₃-mediated currents (data not shown), indicating that neither a functionally active hP2X₂ nor a functionally active heteromeric hP2X_2/3 receptor was formed in these cells.

Pharmacologic characterization of the functional hP2X₂ receptors was performed using Ca²⁺ influx measurements of 1321N1 human astrocytoma cells stably expressing either the hP2X_2a or the hP2X_2b subtype. Native 1321N1 cells have been shown previously to be unresponsive to extracellular nucleotides (Parr et al., 1994). Figure 6A shows representative fluorimetric traces of untransfected 1321N1 cells, as well as these cells stably expressing either the hP2X_2a or the hP2X_2b receptor in response to 10 µM ATP. Untransfected 1321N1 cells show no endogenous response to ATP, indicating a lack of functional P2X or P2Y response in this cell type. The pharmacologic profiles of cells expressing the hP2X_2a or hP2X_2b subtype were found to be virtually identical, as shown in Fig. 6 and Table 1. The agonists ATP, BzATP, ATPS, and 2-meSATP all had approximately the same potency toward either receptor form, with apparent EC₅₀ values of ~1 µM (Table 1). ADPS was weakly active for both receptors, with an EC₅₀ of 30 to 40 µM. The other ATP analogs, as well as UTP, were inactive. The P2 antagonist PPADS was active, with an IC₅₀ value of 5 µM. Suramin was a poor antagonist for either human P2X₂ receptor variant in 1321N1 cells, as measured by either fluorimetric means (IC₅₀ > 100 µM) or by electrophysiologic recordings (IC₅₀ > 50 µM). However, in electrophysiologic recordings from oocytes, suramin was found to be a functional antagonist for both hP2X_2a and hP2X_2b receptors (IC₅₀ = 6 µM). The reason for this apparent cell-specific effect on suramin potency at this receptor is unclear. In all other respects, the human P2X₂ receptors exhibited a similar rank order potency of prototypic P2 agonists and activity of PPADS to that observed previously for the rat P2X₂ subtype (Brake et al., 1994; Bianchi et al., 1999).

View larger version (12K): [in this window] [in a new window]   Fig. 6.   A, fluorimetric measurement of calcium levels in 1321N1 cells (solid line) and 1321N1 cells expressing the hP2X2a (dashed line) or hP2X2b (dotted line) receptors, as indicated in Experimental Procedures. RFU, relative fluorescence units. ATP (10 µM) was added 10 s into run. Dose-response curves of agonists determined by fluorimetric analysis of 1321N1 cells expressing the hP2X2a (B) or hP2X2b (C) receptors. Agonists are indicated as ATP (), 2-meSATP (), ,-meATP (), ATPS (), BzATP (), ADP (), ADPS (), and UTP ().

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

This report describes the cloning and characterization of a novel human P2X receptor gene and cDNAs. Pharmacologic and structural properties indicate that the gene is the human homolog of the previously identified rat P2X₂ subtype (Brake et al., 1994). Comparisons of the human P2X_2a and prototype rat P2X₂ receptors demonstrate a highly conserved structure, with the most variability occurring in the N- and C-cytoplasmic domains. The design of the sense primer for RT-PCR cloning of the full-length human P2X₂ receptor was based on the proximity of the predicted initiation methionine in the human genomic sequence to the initiation methionine of the aligned rat sequence. Although this creates an N-terminal region for the human P2X₂ receptors that is 12 residues longer than that of the rat receptor, the structure is very similar to that of the guinea pig P2X₂ receptor (which has a 9-amino acid extension relative to the rat receptor) (Parker et al., 1998).

The intron/exon structure of the P2X₂ gene was found to be highly conserved between rat and human, following the general structure found with other P2X receptor genes (Brandle et al., 1997; Souslova et al., 1997). Several recent reports have described the identification of splice variants arising from the rat P2X₂ gene (Housley et al., 1995; Brandle et al., 1997; Simon et al., 1997). In addition to the prototype receptor identified initially (Brake et al., 1994), only one of the splice variants, the rat P2X_2b receptor (alternatively called P2X_2-2) (Brandle et al., 1997; Simon et al., 1997), has been shown to form functional channels in heterologous expression systems. This message has a 207-bp deletion in the intracellular C terminus compared with the prototype receptor. Similar to the rat P2X₂ receptor, the human P2X₂ gene also gives rise to additional message heterogeneity, as demonstrated in this work. One of the human splice variants initially isolated was found to be a counterpart of the rat P2X_2b variant, with a deletion of the identical region of this receptor, and also produced a functional ATP-gated ion channel.

Because a cDNA encoding a form of the human P2X₂ receptor analogous to the prototype rat receptor was not found initially, this variant was constructed from the other forms using recombinant techniques and was called hP2X_2a. Subsequently, the naturally produced P2X_2a subtype was isolated from pancreas RNA. Although Northern analysis under high-stringency conditions indicates a high level of P2X₂ message in the pancreas, results of RT-PCR studies of different tissues demonstrated similar levels of splice variants in pancreas compared with other tissues examined. Additional Northern analysis of the pancreatic mRNA was performed using a riboprobe consisting of the 201-bp region of the P2X₂ gene exon 11, which is deleted in the P2X_2b message. This probe also recognized the 2.4-kb message that was detected by the full-length probe (data not shown), indicating that the major P2X₂ message in pancreas is not the C-terminal deletion (P2X_2b) form. This does not identify definitively the pancreas message as the P2X_2a form, however, because both the P2X_2c and the P2X_2d variants also contain the 201-bp sequence. Experiments using Northern analysis are underway to identify conclusively the message detected in the pancreas RNA. One explanation for the apparent discrepancies in P2X₂ message levels as determined by Northern and RT-PCR is that the structure of the P2X₂ RNA interferes with reverse transcription of the message. This is suggested by the failure in isolating the complete message in early RACE procedures and hybridization screens.

Two of the human P2X₂ clones isolated in the present work, the hP2X_2c and hP2X_2d, forms, were unable to form either functional homomeric channels or functional heteromeric channels with the human P2X₃ subunit. The P2X_2c form splices out exon 3 at predicted intron/exon boundaries, and the P2X_2d form may represent an incompletely processed stable intermediate. Interestingly, a splice variant of the human P2X₅ receptor (GenBank no. U49396) also has the corresponding region of the putative extracellular domain missing. A rat cDNA analogous to the human P2X_2d form, which also retains intron 10, was described by Housley et al. (1995). The role of the various P2X₂ receptor mRNAs is unclear. It is possible that the nonfunctional receptor messages may represent splicing artifacts that were detected because of the exquisite sensitivity of the RT-PCR technique. If the nonfunctional messages are artifacts, they appear to make up a high percentage of the total cell P2X₂ mRNA in most tissues, as judged by relative levels of products in RT-PCR analysis (Fig. 4B). The role of these transcripts, which do not form functional ion channels, is unknown at this time.

The rat P2X_2b receptor (alternatively called P2X_2-2) exhibits a more rapid desensitization on prolonged application of agonist compared with the prototypic P2X₂ receptor (Brandle et al., 1997; Simon et al., 1997). Expression of different levels of the respective rat P2X₂ forms has been proposed as a mechanism for modulating the cellular response to ATP in rat tissues such as the pituitary gland (Koshimizu et al., 1998b). Mutagenesis studies of the cDNAs have implicated certain regions of the rat P2X₂ C terminus that are deleted in the P2X_2b form as being important to imparting the slow-desensitizing phenotype of the full-length receptor (Koshimizu et al., 1998a). In contrast to studies with the rat P2X₂ variants, the desensitization rates of the two analogous human forms were found to be nearly identical. The similar desensitization characteristics exhibited by the human P2X_2a and P2X_2b receptors indicate that regions other than those exclusive to the P2X_2a form must play a role in conferring the slow-desensitizing phenotype to the human receptors. Studies using chimeras of the desensitizing rat P2X₁ receptor and nondesensitizing rat P2X₂ receptor have demonstrated that the nondesensitizing P2X phenotype appears to be dominant; both TM regions of rat P2X₁ receptor were required to confer desensitization to a rat P2X₂ receptor, whereas only one TM domain of rat P2X₂ creates a nondesensitizing rat P2X₁ receptor (Werner et al., 1996). This is exemplified in the behavior of heteromeric P2X_2/3 channels (Lewis, 1995; this study), in which the nondesensitizing characteristic of P2X₂ predominates. Structures within the rat P2X_2b variant that are not present in the human P2X_2b receptor may therefore promote a desensitizing phenotype in the absence of the extended C-terminal tail. The putative TM domains of the rat and human receptors are primarily conserved; however, the C-terminal domains of human and rat receptors do differ in that the rat receptor has an additional 10 residues present in both the P2X₂ and P2X_2b forms. Conversely, the N terminus of the human receptor is longer than that of the rat receptor by 12 residues. Differences in these or other structures of the rat and of the human receptors may account for the apparent differences in the desensitization kinetics of the splice variants of the two species.

Because there is no apparent difference in the kinetics of desensitization among the functional forms of the human P2X₂ receptor, there may be an additional role for these two forms of the same receptor. A candidate proline-rich site similar to a consensus SH3 binding domain exists within the P2X_2a C terminus (Yu et al., 1994). This may serve a functionally unique role for the P2X_2a form, such as directing subcellular localization, as has been shown for a similar region of the epithelial Na⁺ channel (Rotin et al., 1994). Association of P2X receptors with cellular proteins has been suggested in a recent study that described an effect of the state of the actin cytoskeleton on P2X₁ receptor function (Parker, 1998). Phosphorylation of the rat P2X₂ receptor C terminus also has been implicated in controlling its function (Chow and Wang, 1998). Interestingly, the protein kinase A phosphorylation site at Ser⁴³¹, which is found to negatively regulate function of the rat P2X₂ receptor, is not conserved in the human receptor (Pro⁴⁴⁰). Other potential phosphorylation sites do exist in the human C terminus, including Ser³⁷⁴ and Thr³⁸³ (Kennelly and Krebs, 1991). It is possible that the heterologous systems used in this study do not provide the cellular component(s) necessary to enable functional differentiation of the human P2X₂ variants. These factor(s) may be present in cells naturally expressing these receptors. Additional studies comparing the characteristics of the human and rat P2X₂ splice variants are needed to examine the physiologic function of this alternative splicing event.

To our knowledge, this is the first demonstration that a splice variant of the P2X₂ subtype can form heteromeric receptors with P2X₃ subunits. These human heteromeric receptors behave in a manner similar to that observed for the rat P2X_2/3 receptors (Lewis et al., 1995). Thus, the human P2X_2/3 receptors exhibit the kinetics of P2X₂ homomeric receptors, but the pharmacologic profile of P2X₃ receptors. The ATP-mediated responses of small-diameter nociceptive neurons of rat sensory ganglia resemble those produced by P2X_2/3 heteromeric receptors in recombinant systems (Thomas et al., 1998). Whether heteromeric receptors exist in human sensory neurons and which form(s) of P2X₂ receptor participate in heteromer formation is the subject of future investigation.

The presence of P2 receptors in the pancreas is not unexpected, because extracellular ATP has been implicated in the function of the neuroendocrine system. ATP induces secretion of hormones, including prolactin and leuteinizing hormone from cells of the pituitary gland (Carew et al., 1994; Chen ZP et al., 1995; Nunez et al., 1997; Villalobos et al., 1997). In addition, because ATP is coreleased with hormones such as insulin, prolactin, and leuteinizing hormone, as well as with catecholamines from adrenal chromaffin cells, it may act as a paracrine regulator of hormone release in these tissues (Leitner et al., 1975; Chen ZP et al., 1995; Tomic et al., 1996; Hollins and Ikeda, 1997; Nunez et al., 1997). Human P2X₂ cDNA was isolated from pituitary RNA (this study), and the P2X₂ receptor RNA and protein have been detected in rat pituitary tissue (Brake et al., 1994; Housley et al., 1995; Tomic et al., 1996; Vulchanova et al., 1996). There is evidence for the action of purinergic receptors on insulin-secreting  cells of the pancreas (reviewed in Hillaire-Buys et al., 1994). The existence of high levels of P2X₂ message in this human organ suggests a role for purinergic modulation of pancreatic function.

This work characterizes a novel human P2X receptor gene with structural and functional properties identifying it as the human homolog of the previously characterized rat P2X₂ receptor. Like the rat gene, the human P2X₂ gene produces multiple transcripts in a variety of tissues. Unlike the functional rat P2X₂ receptors, the two functional human receptors were found to be virtually identical in pharmacologic and electrophysiologic characteristics. The different characteristics exhibited by receptor splice variants from the two species underscore the importance of future studies to further examine the role of the various P2X₂ forms in purinergic signaling.