A Truncated P2X7 Receptor Variant (P2X7-j) Endogenously Expressed in Cervical Cancer Cells Antagonizes the Full-length P2X7 Receptor through Hetero-oligomerization*

A truncated naturally occurring variant of the human receptor P2X7 was identified in cancer cervical cells. The novel protein (P2X7-j), a polypeptide of 258 amino acids, lacks the entire intracellular carboxyl terminus, the second transmembrane domain, and the distal third of the extracellular loop of the full-length P2X7 receptor. The P2X7-j was expressed in the plasma membrane; it showed diminished ligand-binding and channel function capacities and failed to form pores and mediate apoptosis in response to treatment with the P2X7 receptor agonist benzoyl-ATP. The P2X7-j interacted with the full-length P2X7 in a manner suggesting heterooligomerization and blocked the P2X7-mediated actions. Interestingly, P2X7-j immunoreactivity and mRNA expression were similar in lysates of human cancer and normal cervical tissues, but fulllength P2X7 immunoreactivity and mRNA expression were higher in normal than in cancer tissues, and cancer tissues lacked 205-kDa P2X7 immunoreactivity suggesting lack of P2X7 homo(tri)-oligomerization. These results identify a novel P2X7 variant with apoptosis-inhibitory actions, and demonstrate a distinct regulatory property for a truncated variant to antagonize its full-length counterpart through hetero-oligomerization. This may represent a general paradigm for regulation of a protein function by its variant.

The receptor P2X 7 belongs to the P2X subfamily of P2 nucleotide receptors (1,2), which are membrane-bound, ligand-operated channels (3)(4)(5). ATP is the naturally occurring ligand for the P2X 7 and activation of the receptor by brief exposure to extracellular ATP opens cation channels that allow Ca 2ϩ , Na ϩ , and K ϩ influx (6). Longer exposure to ATP allows passage of cations with progressively larger diameters, up to 900 Da, through formation of pores (7). The mechanism of pore formation is unclear, and opinions vary between decreased filter selectivity of existing channels (8) to rearrangement of receptor molecules (9). P2X 7 receptors function in a cell-specific manner and effects of receptor activation are determined by receptor expression (10), trafficking and plasma membrane localization (11)(12)(13), oligomerization (5), and postactivation internalization, recycling, and degradation (14). Expression of P2X 7 can be regulated hormonally; in human cervical epithelial cells epinephrine down-regulates expression of the glycosylated form of the P2X 7 and increases receptor degradation, and the effects can be potentiated by epidermal growth factor (15). Evidence for the physiological role of the P2X 7 comes from studies of P2X 7 -deficient mice, indicating its role in inflammatory (16) and immune processes (17).
Epithelial cells of the female lower reproductive tract express the P2X 7 (18), and in human cervical epithelial cells ligand binding induces apoptosis by a mechanism that involves pore formation, augmented calcium influx, and calcium-dependent activation of the apoptotic mitochondrial pathway (19,20). Because human cervical epithelial cells secrete ATP into the extracellular milieu at concentrations that suffice to induce P2X 7 pores (19), it was proposed that growth of cervical cells in vivo is controlled by autocrine-paracrine P2X 7 -mediated apoptosis (14,15,19,20).
Cervical neoplasia is a common disease in women. Although most cases are detected and managed at an early stage of development, an estimated 13,000 women progress to invasive cancer and about 4000 women die annually of the disease in the United States (21). Until recently little was known about the role of the P2X 7 system in human cancer cervical cells. ATP and the P2X 7 -specific agonist 2Ј,3Ј-O-(4benzoylbenzoyl)-adenosine 5Ј-triphosphate (BzATP) 3 can induce apoptosis in both normal and cancer cervical cells. However, the effects are greater in normal than in cancer cells (19), suggesting that cancer cervical cells have evolved mechanisms that protect them from P2X 7 -mediated apoptosis. Understanding these phenomena is important because mechanisms that block apoptosis could provide growth advantage to cells and enhance the growth of cancers.
The present paper reports our discovery of a novel variant of the human P2X 7 (P2X 7-j ) that lacks the entire intracellular carboxyl terminus, the second transmembrane domain, and the distal third of the extracellular loop of the P2X 7 . This variant was deficient in ligand binding, but interacted with the full-length P2X 7 and blocked P2X 7 -mediated channel activity. Because pore formation depends on oligomerization of P2X 7 molecules (22), the present results suggest that the P2X 7-j variant hetero-oligomerizes with the full-length P2X 7 to form nonfunctional P2X 7 oligomers that do not mediate P2X 7 -dependent apoptosis. Because in cancer cervical cells the P2X 7-j is co-expressed with the wildtype P2X 7 (present results), it is hypothesized that co-expression of the P2X 7-j could lead to defective apoptosis and enhance the growth of the cancer cervical cells.

MATERIALS AND METHODS
Reagents, Cells, and Human Tissues-All chemicals, unless specified otherwise, were obtained from Sigma. The following types of cells were used (all from ATCC): human cervical epithelial cancer CaSki, HeLa, SiHa, and HT3 cells; MDCK (Madin-Darby canine kidney cells, strain II); and HEK293 (human embryonic kidney 293 cells). MDCK cells were cultured in minimal essential medium containing Earle's salts and supplemented with 5% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Cell culture conditions for the other types of cells were described (19).
Discarded human uterine cervical tissues from women undergoing hysterectomy for indications unrelated to the present study were obtained according to IRB protocols 12-03-50 and 03-90-300 from the Human Tissue Procurement Facility of University Hospitals of Cleveland and the Comprehensive Cancer Center Tissue Procurement Core Facility (CTPC), Case Western Reserve University, Cleveland, OH; and from the Cooperative Human Tissue Network (CHTN) (National Cancer Institute) through the Human Tissue Resource Network (HTRN), Department of Pathology at the Ohio State University, Columbus, OH. Upon removal, tissues were washed in cold saline, snap frozen in liquid N 2 , transferred in liquid N 2 , or shipped on dry ice to the laboratory and stored at Ϫ80°C until assayed. The present experiments utilized a total of three histologically normal human uterine cervical tissues and three tissues designated histologically as squamous cell carcinomas of the cervix. The histological diagnoses were assigned by the Departments of Pathology at University Hospitals of Cleveland or at Ohio State University. All six tissues were from premenopausal women ages 44 -49. For assays, cultured cells were lysed as described (14). Tissues (about 30 mg) were minced using mortar and pestle in liquid nitrogen, and lysed in ice-cold lysis buffer (phosphate-buffered saline plus 1% Triton X-100 and protease inhibitors (Halt Protease Inhibitor Mixture Kit, Pierce). Protein quantification was performed using the Bio-Rad Protein Assay (Bio-Rad). Samples were mixed with 2 ϫ SDS sample buffer, boiled at 100°C for 5 min, and stored at Ϫ80°C.
Molecular Biology Techniques and Transfections-Total RNA extracted from CaSki cells using the RNeasy Mini kit (Qiagen, Valencia, CA) was used for RT-PCR using the RT-PCR kit (Invitrogen Corp.) to amplify the human P2X 7 gene (Ref. 23, GenBank TM accession number Y09561). Oligo(dT) primer was used for the RT reaction. A pair of P2X 7 primers (BIOSOURCE, Camarillo, CA) of sense (TTTTTAAGCTTATGCCGGCCTGCTGC-AGCTG) and antisense (TTTTTGCGGCCGCTCAGTAAGGACTCT-TGAAGCC) were used for PCR amplification. The amplified P2X 7 genes were subcloned into pcDNA5/FRT vector (Invitrogen) with HindIII and NotI sites. For inducible expression of P2X 7 receptors in MDCK cells that lack endogenous expression of the receptors, P2X 7 genes were subcloned into pcDNA4/TO vector (Invitrogen) with HindIII and NotI sites. A c-Myc tag was also attached to the NH 2 -terminal of the P2X 7 genes with a new Myc-containing sense primer (TTTTAAGCTTATGGAACAAAAACT-TATTTCTGAAGAAGATCTGCCGGCCTGCTGCAGCTGA) and the same antisense primer. For simultaneous heterologous expression of both the P2X 7 and the P2X 7-j in HEK293 cells, P2X 7 and Myc-P2X 7-j or Myc-P2X 7 and P2X 7-j DNAs were subcloned into pBud 4.1 plasmid (Invitrogen) with HindIII and NotI sites or XbaI and BglII sites, respectively. All genes were fully sequenced for both strands (Cleveland Genomics, Cleveland, OH).
Transfections of P2X 7 plasmid DNAs were performed using Gene-PORTER Transfection Reagent (GST Inc. San Diego, CA). For generation of stable MDCK cells, zeocin at 300 g/ml was used. Stable clonal cells were maintained in minimal essential medium containing 5% tetracycline-free fetal bovine serum and 100 g/ml zeocin. Inducibility of expression of P2X 7 genes (Myc-P2X 7 and Myc-P2X 7-j ) in the stable MDCK cells was confirmed by RT-PCR with sense primer (CTGTTC-CTCTGACCGAGGTT) and antisense primer (TCCGTTTCTCAA-CATTGTTTTCC). The PCR products for the Myc-P2X 7 and Myc-P2X 7-j were 539 and 401 bp, respectively (Fig. 2). Expression of Myc-P2X 7 and Myc-P2X 7-j proteins in stable MDCK cells was induced by 100 ng/ml doxycycline. Stable HEK293 clonal cells were generated by transfection of P2X 7 genes in pBud4.1 vector or pcDNA5/FRT vector in Flp-In TM -293 cells (Invitrogen) and selection in the presence of 300 g/ml zeocin or 500 g/ml hygromycin, respectively. Transfections of P2X 7 genes in pcDNA5/FRT vector were carried out in the presence of pOG44 plasmid (Invitrogen) following the manufacturers instruction. The expression of P2X 7 genes in both stable MDCK and stable HEK293 cells was confirmed by immunocytochemistry and Western blotting. The hygromycin-inducible HEK293 cells expressing the full-length P2X 7 were kindly provided by Dr. George Dubyak, Case Western Reserve University. Stable HEK293 clones were maintained in the presence of 100 g/ml zeocin, 500 g/ml hygromycin, or both.
Real time PCR assays utilized SYBR Green PCR master mixture machine with ABI 7500 Real-time PCR SDS software (Applied Biosystems, Foster City, CA) and experiments were carried out according to the manufacturer's instructions. Primers were as follows: full-length P2X 7 : forward, ATA-CAGTTTCCGTCGCCTTG; reverse, AACGGATCCCGAAGACTTTT. The truncated P2X 7 variant P2X 7-j : forward, TTTCAGATGTGGCAAT-TCAGATA, reverse, AAGTAGGAGAGGGTTGAGCC; glyceraldehyde-3-phosphate dehydrogenase: forward, CAATGACCCCTTCATTGACC; reverse, GACAAGCTTCCCGTTCTCAG. The reaction mixture was composed of 0.5 l of primers (5 M), 9 l of diluted cDNA, and 10 l of SYBR Green PCR master mixture. PCR conditions were 50°C for 2 min, 95°C for 10 min, 40 cycles of 15 s at 95°C, and 60°C for 1 min. Results were calculated using the comparative threshold cycle (C t ) method of relative quantitation (RQ).
Protein Methods-Western blots were carried out on post-nuclear supernatant of cell lysates (14), and total proteins were quantified using the Bio-Rad Protein Assay (Bio-Rad). Samples were mixed with 2 ϫ SDS sample buffer and boiled at 100°C for 5 min. Fractionation of cultured cells was done using the Proteo-Extract Subcellular Proteome Extraction kit (EMD Biosciences, Inc., San Diego, CA) according to the manufacturer's instructions. Aliquots of cell lysates (normalized to 15 g of protein) were separated in SDS-polyacrylamide 6 -10% gels by gel electrophoresis (PAGE) and blotted by Western analysis. Receptor polypeptides were visualized using 1.5 g/ml of primary antibodies. The rabbit polyclonal anti-P2X 7 antibody (Alomone Labs, Jerusalem, Israel) recognizes the stretch of 12 amino acids (KKGWMDPSKGIQTGRC) from 136 to 152 of the mouse P2X 7 receptor. Rabbit monoclonal anti-Myc antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-tubulin antibody (hybridoma supernatant, clone E7, generated with the antigen of ␤-tubulin-galactosidase/ftz fusion protein) was from the Developmental Studies Hybridoma Bank, University of Iowa (Iowa City, IA), and was used at 1:500 dilution. Anti-rabbit peroxidase-conjugated secondary antibody was used for visualization (ECL kit, Santa Cruz Biotechnology, Santa Cruz CA). Co-immunoprecipitation, immunostaining, light microscopy, and confocal laser scanning microscopy were described (14).
Cell Number, Cell Cycle, Flow Cytometry, and Apoptosis Assays-Cell number was evaluated using the CyQUANY cell proliferation assay kit (Molecular Probes, Eugene, OR) according to the manufacturer's instructions.
For cell cycle analysis cells were synchronized by culturing in 10% fetal bovine serum, Dulbecco's modified Eagle's medium overnight and serum starvation in Dulbecco's modified Eagle's medium for 6 h. Cells were shifted to 10% fetal bovine serum, Dulbecco's modified Eagle's medium containing 1:500 defined keratinocyte growth factors from bovine pituitary extract (Invitrogen) for 18 h, and treated with 100 M BzATP (or the vehicle) for 8 h. Harvested cells were washed with phosphate-buffered saline and fixed with 100% methanol at Ϫ20°C overnight. Following treatment with 20 g/ml RNase, the cells were stained with 50 g/ml propidium iodide. Aliquots of 5 ϫ 10 4 cells were analyzed for DNA content by flow cytometry using Beckman Epics XL-MCL.
Apoptosis was quantified using Roche Cell Death Detection ELISA Kit (Roche Applied Science) according to the manufacturer's instructions. Briefly, cells plated in 96-multiwell plates at 10 4 cells/well and maintained overnight in 10% serum-enriched medium were serumstarved for 6 h in the absence or presence of added BzATP. At the completion of incubation the medium was removed, spun, and the pellet stored. The cells together with the stored pellet were lysed with the provided lysis buffer, and the mixture was spun at 200 ϫ g for 10 min at room temperature. An aliquot from the supernatant, containing the cytoplasmic fraction of oligonucleosomes was added to a streptavidincoated multiplate and mixed with pre-prepared reaction reagent containing the anti-histone-biotin and anti-DNA-POD antibodies. After a 2-h incubation on a shaker at room temperature the solution was collected and mixed with 2,2Ј-azino-bis(3-ethylbenziazolin-6-sulfonic acid) solution at room temperature for 15 min. Absorbance was measured at 405 nm against blank, and the degree of apoptosis was determined in reference to the control/standard provided in the kit.
Dynamic Confocal Laser Scanning Microscopy-We used a described method (24) with minor modifications. Briefly, cells were seeded at 2-3 ϫ 10 5 on 35-mm glass bottom Petri dishes (MatTek Corp., Ash-land, MA), and allowed to reach confluence. Cells were loaded with 5 M Fluo-4/AM (Molecular Probes) in 0.1% bovine serum albumin, Ringer's solution for 20 min at 37°C, rinsed twice with Ringer's solution, and incubated for additional 20 min at 37°C. Cells were imaged with a Zeiss LSM 510 inverted real time confocal microscope equipped with a ϫ20 water immersion objective at room temperature (Comprehensive Cancer Center Microscopy Facility of Case Western Reserve University). Images were collected at 488/505 nm (excitation/emission) before and after treatment with 100 M BzATP, added to both the luminal and subluminal perfusates, at intervals of 10 to 15 s afterward. For ethidium bromide influx experiments, cells cultured on glass-bottomed dishes were loaded onto the microscope. Images (collected at 488/505 nm (excitation/emission)) were taken before, and at intervals of 30 s after adding 5 M ethidium bromide to both the luminal and subluminal perfusates. Average fluorescence intensity was quantified from collated images using MetaVue software (Fryer Company Inc., Huntley, IL) by subtracting the basal intensity value.
Densitometry-Densitometry was done using a AGFA Arcus II scanner (AGFA, New York) and Un-Scan-It gel automated digital software (Silk Scientific, Orem, UT).
Statistical Analysis-Data are presented as mean Ϯ S.D. and significance of differences among means was estimated by Student's t test. Trends were analyzed by analysis of variance.

RESULTS
Cloning and Expression of the Truncated P2X 7 Variant P2X 7-j -A PCR product (ϳ1652 bp) of smaller size than the expected full-length P2X 7 (ϳ1789 bp) was identified in RT-PCR experiments trying to amplify the full-length P2X 7 gene (data not shown). DNA sequencing and gene analysis of the 1652-bp PCR product revealed an identical P2X 7 gene that lacks the putative exon 8 except the A 882 , with a shift of coding frame to a new variant (Fig. 1A). As a result, the gene product becomes a truncated variant of only 258 amino acids (compared with 595 of the full-length P2X 7 ) with 10 altered unique residues at the COOH-terminal (Fig. 1B). The predicted sequence of this variant lacks the entire carboxyl cytoplasmic domain, the second transmembrane domain, and the distal third of the extracellular domain of the fulllength P2X 7 (Fig. 1B, lower panel). We named this newly discovered variant as P2X 7-j because previous studies identified splice variants iso-forms designated P2X 7-b -P2X 7-h (Ref. 25, accession numbers AY847 (298 -304)), and a truncated P2X 7 variant 2 (149 residues) (Ref. 26, accession number NM177427). The sequence of our novel P2X 7-j variant has been deposited in the GenBank (accession number DQ399293). Blast search of available gene data base showed that this variant has not been reported before.
To determine the translational product of the P2X 7-j gene, the P2X 7-j cDNA was subcloned into various expression vectors as described under "Materials and Methods," and was introduced into MDCK and HEK293 cells by transfection. MDCK cells, in contrast to the HEK293 In C, lanes are total homogenates prior to the immunoprecipitation (b, "before"); the immunoprecipitate mixtures (IP); and post-immunoprecipitate mixtures (a, "after"). D, lysates were immunoblotted with the anti-P2X 7 -R antibody in the absence or presence of the P2X 7 -R antigen (used to generate the anti-P2X 7 -R antibody). E and F, heterologous expression of human P2X 7 and P2X 7-j proteins in HEK293 cells. HEK293 cells were transfected with P2X 7 , P2X 7-j , or N-Myc-P2X 7-j cDNAs. G and H, naturally occurring expression of the P2X 7 and P2X 7-j in human cancer cervical cell lines. All experiments were repeated three to six times with similar trends.

Truncated P2X 7 Receptor Variant (P2X 7-j )
cells form sheets of cells with typical epithelial characteristics. Stable lines of MDCK and HEK293 cells expressing the P2X 7-j were successfully generated, and the expression of P2X 7-j was confirmed at both the mRNA level by RT-PCR ( Fig. 2A) and the protein level by Western blots (Fig. 2, A-E). Specific bands at 75 kDa for the full-length P2X 7 proteins and 45-42 kDa for the P2X 7-j proteins expressed in MDCK cells were visualized with both anti-Myc antibodies and anti-P2X 7 antibodies (Fig.  2, B-D). Similar patterns of expression of the P2X 7 and the P2X 7-j proteins were also observed in HEK293 cells (Fig. 2, E and F). To determine whether the P2X 7-j gene is also naturally translated into protein, CaSki cells expressing P2X 7-j mRNA were examined by Western blot with anti-P2X 7 antibody. Fig. 2G shows two main specific forms at 75 and 42 kDa that can be blocked by the P2X 7 antigen. The 75-kDa form is most likely the full-length P2X 7 (14,15). The 42-kDa is likely the P2X 7-j because the sizes of heterologously expressed P2X 7-j in MDCK cells (Fig.  2, B-D) and HEK293 cells (Fig. 2, E and F) were also 42 kDa. The presence of the P2X 7-j variant (45-42 kDa) was confirmed also in the human cancer cervical cell lines HT3, SiHa, and HeLa (Fig. 2H). Collectively, these data suggest that the P2X 7-j is a naturally occurring P2X 7 variant.
Cellular Localization of the P2X 7-j -In MDCK cells and HEK293 cells, both the 75-kDa P2X 7 and the 45-42-kDa forms were present predominantly in the plasma-membrane fraction. Small amounts of the 75-kDa P2X 7 were detected also in the cytosol, nuclear, and cytoskeletal fractions (Fig. 3). Interestingly, a significant amount of the 45-42-kDa form was also detected in the nuclear fraction. These findings are supported by the results obtained using laser confocal microscopy (Fig. 3C), which also show that some of the P2X 7-j localizes in nuclear/perinuclear regions of the HEK293 cells.
Deficient Induction of Apoptosis and Activation of Pore Formation-In cervical cells an important function of the full-length P2X 7 is induction of apoptosis (19). To examine whether the P2X 7-j preserves this action, the cell number, cell cycle, and direct apoptosis assays were performed using both MDCK and HEK293 cells. In MDCK cells, Induc-ible expression of P2X 7 alone, but not P2X 7-j alone resulted in a significant decrease in cell numbers (ϳ25%, Fig. 4A). A similar effect was observed in HEK293 cells where induced expression of the P2X 7 alone, but not the P2X 7-j , decreased cell number by about 35% compared with mock HEK293 cells (Fig. 4A). In contrast, P2X 7-j expressing cells, in particular HEK293 cells, showed an increase in cell number (by about 35%, Fig. 4A).
Cell cycle assays in MDCK and HEK293 cells showed that expression of the full-length P2X 7 induced a 2-4-fold increase of cells in the sub-G 1 phase. Treatment with BzATP augmented the effect (Fig. 4B) in P2X 7expressing HEK293, suggesting increased apoptosis. In contrast, expression of the P2X 7-j altered only little the percent of cells in the sub-G 1 phase as compared with control cells (Fig. 4B). Direct apoptosis assays in MDCK and HEK293 cells showed that expression of the P2X 7 increased apoptosis by about 2-fold, and treatment with BzATP augmented the effect (Fig. 4C). In contrast, expression of the P2X 7-j resulted in no change in baseline or in BzATP-induced apoptosis (Fig. 4C). In HEK293 cells, co-expression of the P2X 7-j blocked P2X 7 -mediated apoptosis (Fig. 4C).
In cervical cells, P2X 7 -mediated apoptosis requires formation of pores (19,20). The failure of P2X 7-j -expressing cells to undergo apoptosis could be the result of the inability of the P2X 7-j to effectively activate pore formation. As expected, in both MDCK and HEK293 cells stimulation of P2X 7-j with BzATP induced only a negligible influx of ethidium bromide compared with effects in cells expressing the P2X 7 (Fig. 5A), suggesting deficient pore formation. In HEK293 cells, co-expression of the P2X 7-j inhibited by about 80% the P2X 7 -mediated influx of ethidium bromide in the presence of BzATP (Fig. 5A).
Deficient BzATP-induced Acute Ca 2ϩ Influx-An early event of P2X 7 receptor activation is the acute channel opening that precedes, and possibly mediates P2X 7 pore formation (22). MDCK and HEK293 cells expressing P2X 7 or P2X 7-j were loaded with the Ca 2ϩ -sensitive dye Fluo-4 and examined using dynamic confocal laser scanning microscopy in the absence and presence of BzATP. In P2X 7 -expressing cells BzATP induced an acute tran-

Truncated P2X 7 Receptor Variant (P2X 7-j )
sient increase in cytosolic calcium (Fig. 5B). The BzATP effects could be blocked by pre-treatment with 1.2 mM EGTA (to chelate extracellular calcium) (data not shown), indicating that the acute increase in cytosolic calcium is the result of calcium influx. In P2X 7-j -expressing cells BzATP induced a significantly smaller increase in cytosolic Ca 2ϩ (Fig. 5B). In P2X 7expressing HEK293 cells, co-expression of the P2X 7-j inhibited the BzATPinduced increase in cytosolic Ca 2ϩ to levels seen in cells expressing P2X 7-j alone (Fig. 5B).
Collectively, the data in Figs. 4C, and 5, A and B, indicate that co-expression of the P2X 7-j inhibits P2X 7 receptor activation, pore formation, and apoptosis. Fig. 5C shows that in P2X 7 -expressing HEK293 cells co-expression of the P2X 7-j had no significant effect on the expression of the full-length P2X 7 . Therefore the P2X 7-j -mediated inhibition of P2X 7 actions cannot be explained by down-regulation of P2X 7 receptor expression. P2X 7-j /P2X 7 Interaction-Previous studies suggested that the membrane-bound full-length P2X 7 functions as homo-oligomer, presumably a homotrimer (22). This speculation is supported by the experiment shown in Fig. 6A where differential co-immunoprecipitation and Western blot analysis in lysates of MDCK cells co-expressing Mycand HA-tagged P2X 7 resulted in co-immunoprecipitation of the Myc-and HA-tagged isoforms. Fig. 6B shows co-immunoprecipitation of Myc-and HA-tagged isoforms in lysates of HEK293 cells co-expressing Myc-and HA-tagged P2X 7-j . Fig. 6C shows co-immunoprecipitation of Myc-and HA-tagged isoforms in lysates of HEK293 cells co-expressing Myc-tagged P2X 7-j and HA-tagged P2X 7 . These data confirm the possible existence of homo-and hetero-oligomers in the forms of P2X 7 / P2X 7 , P2X 7-j /P2X 7-j , and P2X 7-j /P2X 7 .
The results in Fig. 7 further clarified these speculations by using 6 -8% gels to better detect and separate high M r forms. In HEK293 cells expressing the P2X 7 or co-expressing the P2X 7-j plus the P2X 7 we confirmed that the wild-type P2X 7 (usually referred to as 75 kDa) is expressed as a cluster of 85-(glycosylated) and 65-kDa (non-glycosylated) forms (Fig. 7) (15). In HEK293 cells expressing the P2X 7 alone, in addition to the 85-65-kDa forms two other specific P2X 7 immunoreactivities could be detected, a major form of 220 kDa and a minor form of 115 kDa. In HEK293 cells expressing the P2X 7-j alone, in addition to the 45-42-kDa forms at least two other specific P2X 7 immunoreactivities were observed, a major form of 135 kDa and a minor form of 90 kDa (Fig. 7). In HEK293 cells co-expressing the P2X 7 plus the P2X 7-j , in addition to the 85-65-kDa (P2X 7 monomers) and 45-42-kDa forms (P2X 7-j monomers) four other forms were observed: a major form of 135
Homologous Expression of the P2X 7 and P2X 7-j Forms-P2X 7 -R immunoreactivity was examined in lysates of three human squamous cell carcinoma tissues of the cervix (Fig. 8A, lanes a-c), and results were compared with those of three histologically normal cervical tissues (Fig.  8A, lanes d-f). The 45-42-kDa forms (P2X 7-j ) were equally expressed in cancer and normal tissues, but the expression of the full-length 85-65-kDa forms (P2X 7 ) was significantly greater in normal than in cancer tissues (Fig. 8A). Densitometry, relative to the constitutively expressed tubulin protein revealed a ratio of P2X 7 /tubulin in excess of about 8 -32 in the normal versus the cancer tissues (Fig. 8B). Interestingly, 205-kDa P2X 7 -R immunoreactivity was found in lysates of the normal tissues, but it was lacking in lysates of the cancer tissues (Fig. 8A).
Total RNA samples from the six tissues of Figs. 8A were also analyzed by real time PCR. The results revealed no significant differences in the mRNA ratios of P2X 7-j and the constitutively expressed glyceraldehyde-3-phosphate dehydrogenase among cancer and normal tissues. In contrast, the ratios of P2X 7 /glyceraldehyde-3-phosphate dehydrogenase were 3-200-fold greater in normal than in cancer tissues (Fig. 8C).

DISCUSSION
The present data describe a novel truncated form of the P2X 7 receptor, designated P2X 7-j , which is expressed naturally in normal and cancer cervical cells in humans. The P2X 7-j protein is composed of the  proximal 248 amino acids of the wild-type P2X 7 plus an altered stretch of 10 amino acids at its carboxyl terminal. It lacks the distal 337 amino acids of the P2X 7 , including the entire intracellular carboxyl terminus, the second transmembrane domain, and the distal third of the extracellular loop. When expressed heterologously, the P2X 7-j was localized to the plasma membrane, and treatment with the P2X 7 -specific ligand BzATP evoked some channel activity but failed to induce pore formation and apoptosis. The P2X 7-j interacted with the full-length P2X 7 , suggesting hetero-oligomerization between the P2X 7-j and P2X 7 . Coexpression in host cells of the P2X 7-j plus the full-length P2X 7 blocked apoptosis and pore formation and inhibited acute Ca 2ϩ influx in response to BzATP activation of the P2X 7 . These results suggest that co-expression of P2X 7-j results in formation of nonfunctional P2X 7-j / P2X 7 hetero-oligomers that fail to promote apoptosis. Therefore, in cells expressing both P2X 7 receptor forms, abundance of expression of the P2X 7-j or paucity of expression of the full-length P2X 7 receptor may cause resistance to apoptosis.
Eight variants of the human P2X 7 resulting from alternative splicing also were previously reported, designated by Cheewatrakoolpong et al. (25) as P2X 7b-h (GenBank AY847 (298 -304)), and by Georgiou et al. (26) as P2X 7 variant 2 (GenBank NM177427). For consistency we propose to designate the P2X 7 variant 2 as P2X 7-i and the present novel variant as P2X 7-j . Of the nine human P2X 7 variants five (P2X 7-b , P2X 7-e , P2X 7-g , P2X 7-I , and P2X 7-j ) are truncated, lacking the carboxyl terminus of the wild-type P2X 7 receptor, and have been shown (P2X 7-b and P2X 7-j ) or presumed to be ineffective as far as pore formation and apoptosis induction. The presently described P2X 7-j differs from the other truncated variants in structure and size. It is also the only P2X 7 form to be ineffective when expressed alone, and more importantly, when coexpressed with the full-length P2X 7 receptor. The present results show that it can interact and possibly oligomerize with the full-length receptor and thereby render it ineffective. P2X 7-j preserves the first four of the five putative N-glycosylation sites present in the full-length P2X 7 (Asn-187, -202, -213, -241, and -284). The calculated molecular masses of P2X 7 and P2X 7-j are 69 and 36 kDa, respectively. N-Glycosylation would increase the size of the P2X 7 to about 85 kDa, as was observed (Ref. 15, and present results). The predicted size of the P2X 7-j (258 amino acids) is compatible with the de novo induced protein cluster of 45-42 kDa in HEK293 or MDCK cells transfected with the P2X 7-j cDNA. In cells expressing the P2X 7-j , treatment with BzATP evoked some acute transient calcium influx but no influx of ethidium bromide (i.e. no pore formation). The failure of the P2X 7-j to form pores is probably due to lack of the intracellular carboxyl terminus. This is consistent with conclusions from mutagenesis studies (6,8,35), with the E496A mutation (27), and with the P2X 7-b (⌬-C) variant (25), that the distal two-thirds of the carboxyl terminus of the full-length receptor is required for pore formation but is dispensable for channel activity (36).
Plasma membrane localization of P2X 7 receptors determines their functionality (22). Surface expression depends on receptor glycosylation, trafficking, and sorting to the plasma membrane, and on receptor redistribution as determined by internalization, degradation, and recycling (14). Activation of wild-type P2X 7 induces GRK-3, ␤-arrestin-2, and dynamin-dependent internalization into clathrin domains, and receptor recycling into the plasma membrane (14). In CaSki cells pore formation is associated with post-activation redistribution of the P2X 7 into the plasma membrane, supporting the hypothesis that pore formation is not only the result of the conformational change of existing receptor molecules, but it involves recruitment of additional receptors to the plasma membrane (37)(38)(39).
In HEK293 and MDCK cells expressing the P2X 7-j form, the P2X 7-j localized also in nuclear/perinuclear regions. Nuclear/perinuclear localization of the wild-type P2X 7 was previously described in epithelial (40,41), neuronal (e.g. Ref. 42), and smooth muscle cells (43), but the biological significance of those findings is at present unclear. In the case of the P2X 7-j a possible explanation is a defect in sorting the truncated protein to the plasma membrane, resulting in localization of the receptor in the cytoplasm at perinuclear domains. Support for this hypothesis comes from studies suggesting that domains within the carboxyl-terminal tail of P2X receptors (43,44), including the P2X 7 (6,25,45), direct trafficking to the plasma membrane and stabilize expression of the receptors. However, the present data in MDCK and HEK293 cells show that most of the truncated isoform localized in the plasma membrane despite lacking the entire intracellular carboxyl terminus of the fulllength P2X 7 . Also, the highly preserved N-glycosylation sites in P2X 7-j could facilitate transport and expression of the receptor in the plasma membrane (22).
Another explanation is that the P2X 7-j is sorted to the plasma membrane and upon activation undergoes endocytosis into clathrin-coated endosomes (46), similar to the wild-type P2X 7 (14). However, in contrast to the P2X 7 , the truncated form is retained longer in submembranal domains and fails to undergo degradation. One of the conse-   JUNE 23, 2006 • VOLUME 281 • NUMBER 25 quences would be feedback inhibition of internalization and longer residence of the P2X 7-j in the plasma membrane. The end result in cells expressing both isoforms would be competition with the wild-type P2X 7 monomers for oligomerization. Support for this hypothesis comes from analysis of the presumed phosphorylation sites in the P2X 7-j . The P2X 7-j lacks 7 of 10 tyrosine and 12 of 15 serine phosphorylation sites, but it retains 6 of the 7 threonine phosphorylation sites. Activation of the full-length P2X 7 induces a transient increase in receptor phosphorylation on tyrosine, serine, and threonine residues, which are associated with redistribution of the P2X 7 (14). In CaSki cells, redistribution of the P2X 7 was associated with phosphorylation on tyrosine and serine residues, whereas phosphorylation on threonine residues increased ATP requirements for tyrosine and serine phosphorylation, and could therefore control the effect (14). The lack of most tyrosine and serine residues, but only one of the seven threonine residues predicts attenuated ligandinduced phosphorylation of the P2X 7-j on tyrosine and serine residues, with only minimal effect on threonine phosphorylation. The consequence of this effect would be attenuated internalization of the P2X 7-j and prolonged residence in the plasma membrane.
P2X channels function as oligotrimers (9,22,47,48). P2X 1 -P2X 6 receptors can hetero-oligomerize, but the P2X 7 receptor forms only homo-oligomers (49). Our data in HEK293 cells expressing the P2X 7 support the speculation of formation of homotrimers, because in addition to the 85-65-kDa monomeric form a specific P2X 7 immunoreactivity of 220-kDa form was detected, compatible with a P2X 7 homotrimer ([P2X 7 ] 3 ). An additional minor and smaller 115-kDa form in those cells was either a partially degraded form or a denatured homodimer. A similar mechanism possibly involves the P2X 7-j because in HEK293 cells expressing the P2X 7-j , in addition to the 45-42-kDa monomeric forms, a 135-kDa form was also detected, compatible with a P2X 7-j homotrimer ([P2X 7-j ] 3 ). The 90-kDa form in those cells was possibly a denatured homodimer as well.
The novel and important finding in this paper was that co-expression of the P2X 7-j in cells expressing the full-length P2X 7 receptor blocked BzATP activation of the P2X 7 receptor. The data in Figs. 6 and 7 may provide a mechanistic explanation. In HEK293 cells co-expressing the P2X 7 plus the P2X 7-j , four specific P2X 7 immunoreactivities were detected in addition to the 85-65-kDa (P2X 7 monomers) and the 45-42-kDa (P2X 7-j monomers) forms. In order of expression levels those were the 135-, 160-, 220-, and 200-kDa bands. The 135-kDa form could possibly be the oligotrimer [P2X 7-j 7-j ]. These data predict that co-expression of the P2X 7 plus the P2X 7-j would favor formation of inactive complexes of the P2X 7 receptor.
The above conclusion is also supported by the patterns of P2X 7 -R protein and mRNA expression in human cervix squamous cell carcinoma versus normal cervical tissues. Whereas the cancer and normal tissues showed similar expression of the truncated P2X 7-j form, the expression of full-length P2X 7 was significantly greater in normal than in cancer tissues. Furthermore, normal tissues expressed the 205-kDa  carcinoma (columns a-c) and normal tissues (columns  d-f). Tissues designated a-f were obtained from different women. Assays utilized discarded human tissues as described under "Materials and Methods." A, total homogenates were resolved on 8% PAGE, Western blotted using the anti-P2X 7 antibody, and reprobred with the anti-tubulin antibody. B, data of A were analyzed by densitometry and presented in terms of the ratios of the P2X 7 and P2X 7-j proteins relative to tubulin. AU, arbitrary units. C, total RNA obtained from lysates of specimens a-f were reverse transcribed and assayed by real time PCR. Shown are the ratios of P2X 7 and P2X 7-j mRNA relative to glyceraldehyde-3-phosphate dehydrogenase mRNA. Rq, relative quantification. The P2X 7-j /GPDH mRNA levels are shown as ϫ10 (to fit in the figure). P2X 7 -R immunoreactivity, which could represent the [P2X 7 ] 3 homotrimer; in contrast, cancer tissues lacked such an expression. These data therefore suggest a mechanistic explanation for the defective P2X 7 receptor-related apoptosis in human cancer cervical cells (19). Accordingly, decreased P2X 7 transcription and lesser expression of the P2X 7 in cancer cells would favor formation of P2X 7-j -dominated inactive [P2X 7-j ]-[P2X 7 ] complexes. As such, determinations of the cellular content of the full-length P2X 7 , alone or relative to the P2X 7-j could provide a novel marker of defective apoptosis and possibly cancer.
The molecular mechanism by which single P2X 7 molecules oligomerize is unclear. The present results support the speculation that ectodomains of the P2X receptor are essential to carry out the oligomerization (47). However, our data do not support the speculation that residues in or near the second membrane-spanning segment are critical for multimerization of P2X receptors (50) because the P2X 7-j could interact with the full-length P2X 7 despite lacking the second transmembrane stretch.
The P2X 7-j protein is a translational product of alternative splicing. Alternative splicing determines the binding properties, intracellular localization, enzymatic activity, protein stability, and post-translational modifications of a large number of proteins that descend from the relatively low number of human genes (51). Although expression of aberrant mRNA is usually controlled (52), inherited and acquired defects of pre-mRNA processing are increasingly recognized as causes of human disease including cancer (53,54). We identified the P2X 7-j mRNA in a number of different types of human cancer cells (data not shown). As discussed above hetero-oligomerization of the P2X 7-j with the P2X 7 could result in inactive receptor and may be an important mechanism for regulation of P2X 7 effects wherein the non-functional form P2X 7-j could compete with the full-length P2X 7 receptor for oligomerization and block its function. Abrogated P2X 7 -mediated apoptosis could play a role in tumor neogenesis.
Most proteins in mammals have splice variants, which greatly expand the proteome. Understanding the function of protein variants is critical to the advance of system biology and medicine. Unfortunately, the function and mode of action for most protein variants identified this far remain elusive. Our finding that a truncated variant antagonizes its full-length counterpart through hetero-oligomerization with the fulllength form may represent a general paradigm for regulation of protein function by its variant.