Introduction

Summary Introduction Materials & Methods Results Discussion References

The binding of ATP to P2-purinergic receptors exerts widespread biological responses in different tissues, where it regulates central and peripheral nervous system signaling, blood vessel contractility, and endocrine secretion among other functions (for reviews, see Refs. 1-3). In 1985, based on the order of potency of different nucleotide analogs, these receptors were initially classified by Burnstock and Kennedy into P2X and P2Y subgroups (4). According to the signal transduction mechanism the P2X receptors constitute a subclass of ligand-gated channels while the P2Y subgroup represent metabotropic receptors.

The activation of P2X receptors leads to the opening, in the millisecond range, of nonselective cation channels permeable to Na⁺ and K⁺. Certain P2X receptors are also permeable to Ca²⁺ (5, 6). Pharmacologically, the agonist potency of ,-meATP and the antagonist effects of suramin and PPADS have been useful tools with which to define native P2X-mediated responses. In different cell types, the P2X receptors seem to have similar but not identical ligand binding selectivities (7), clearly suggesting the existence of multiple subclasses.

Recently, molecular biological studies have confirmed P2X diversity at the protein level. Seven isoforms have been cloned from different rat tissues, rP2X_1-7 (Ref. 8 and references therein; 9, 10), and one isoform has been cloned from human bladder, hP2X₁ (11). The rP2X subtypes share an overall amino acid identity of only 35-50%, but the conservation of the main structural features (two putative transmembrane segments flanking a putatively extracellular loop containing 10 cysteine residues) suggest a common three-dimensional structure (8). The different isoforms, except rP2X₃ (12), display a widespread mRNA distribution (9, 10, 13, 14). It is remarkable that rP2X₄ transcripts are detected in all the tissues analyzed, including central nervous system neurons (15).

In heterologous expression systems, the cloned P2X subunits assemble into functional homomeric receptors with unique pharmacological and kinetic properties. Despite some variation in ATP sensitivity, the main pharmacological distinction between the P2X isoforms is their relative sensitivities to the agonists ,-meATP and ,-meATP and the antagonists suramin and PPADS (16). Currents evoked by ATP in cells expressing the rP2X₄ (15) and rP2X₆ (16) receptors are weakly affected by suramin and PPADS. On the other hand, the same antagonists completely block the rP2X₁ (12), hP2X₁ (17)_, rP2X₂ (17), rP2X₃ (12), and rP2X₅ (9) receptors (IC₅₀ = 1-5 µM). Structural studies have demonstrated that the irreversible block of PPADS requires a lysine residue at position 249 (of rP2X₄) (16, 18), but no structural data have been described for the suramin binding pocket.

We report the cloning and functional characterization of the P2X₄ isoform from human brain (hP2X₄). In addition, we describe significant pharmacological differences concerning the antagonist potencies between the human and rat isoforms and delimit the structural requirements conferring the different phenotypes.

	Materials and Methods

Summary Introduction Materials & Methods Results Discussion References

Drugs. ATP (disodium salt), CTP (sodium salt), UTP (sodium salt), ,-meATP, (lithium salt), D-,-meATP (sodium salt), cibacron blue 3GA, bromphenol blue, basilen blue, (+)-tubocurarine (chloride), dATP (sodium salt), and AMP (sodium salt) were obtained from Sigma Chemical (St. Louis, MO). GTP (disodium salt) was purchased from Fluka Chemical (Ronkonkoma, NY). ADP (free acid) was obtained from Boehringer-Mannheim Biochemicals (Indianapolis, IN). PPADS (tetrasodium salt) and 2MeSATP (tetrasodium salt) were obtained from Research Biochemicals (Natick, MA). Suramin was purchased from Calbiochem (San Diego, CA). Fura-2 was obtained from Molecular Probes (Eugene, OR).

cDNA synthesis and PCR amplification. Poly(A)⁺ RNA was purified from human brain total RNA (75 µg) (Clontech, Palo Alto, CA) using oligo(dT)₂₅ magnetic beads (Dynal, Lake Success, NY) according to the manufacturer's recommendations. For the first-strand cDNA synthesis (19), the resulting mRNA was divided in two aliquots, primed separately with oligo(dT)_12-18 or random hexamers, and extended using the SuperScript Plus Reverse Transcriptase (GIBCO, Grand Island, NY). After enzyme inactivation, the differently primed cDNAs were diluted with 1 volume of H₂O and mixed for PCR analysis.

Library screening. A human brain cDNA library (Clontech) was screened by lifting 1 × 10⁶ phages to nylon membranes (Duralon-UV; Stratagene, La Jolla, CA) as described previously (21). The hP2X₄ PCR fragment isolated from human brain cDNA (~750 bp) was labeled with [-³²P]dCTP (specific activity, 3000 Ci/mmol) by random priming with the Redyprime kit (Amersham, Arlington Heights, IL). Hybridization of probes (5 × 10⁵ cpm/ml) with the filters and washing was performed under the conditions previously described (15). The  phages found to hybridize to the probe were subsequently plaque purified, subcloned into pBluescript SK(II) (Stratagene), and sequenced (see below). Other standard nucleic acid manipulations were performed using conventional protocols (21).

RT-PCR. First-strand cDNA synthesis reactions were performed using total RNA from various adult human tissues (Clontech) as indicated above. The PCR reactions were performed in 50-µl final volumes containing 1 µl of cDNA, 0.5 µM each of pairs of hP2X₄-specific primers, 200 µM dNTPs, 2.5 units of Taq DNA polymerase, and 1× Taq DNA polymerase reaction buffer (Promega, Madison, WI). The PCR thermal profile was 5 min at 94° and 30 cycles of 40 sec at 94°, 40 sec at 58°, and 60 sec at 72°. The hP2X₄ primers were (forward, located within the protein coding sequence) 5-CACCCACAGCAACGGAGTCT-3 and (reverse, located within the 3 untranslated region) 5-TTTGATGGGGCTGTGGAGAG-3. The PCR products were separated in an agarose gel and visualized with ethidium bromide staining. The specificity of the amplification bands was confirmed by determination of the nucleotide sequence. As negative controls, 0.5 µg of human genomic DNA and a sample with no target cDNA (H₂O) were subjected to the same amplification protocol.

FISH and detection. Cell preparation and FISH were performed as previously described (23). Briefly, lymphocytes isolated from human blood were cultured in -minimal essential medium supplemented with 10% fetal calf serum and PHA at 37° for 68-72 hr. The cells were washed three times with serum-free medium to release the block and recultured at 37° for 6 hr in -minimum essential medium with thymidine (2.5 µg/ml; Sigma). The cells were then harvested, and slides were made by using standard procedures involving hypotonic treatment, fixation, and air drying.

DNA mutagenesis and RNA synthesis. Chimeric DNA constructs were engineered by using common endonuclease restriction enzymes for both species. Single-point mutations were made using PCR approaches as described previously (25). The nucleotide sequence of the amplified DNA fragments was determined to ensure the absence of random mutations.

Electrophysiological characterization. Oocyte isolation and handling were performed using standard techniques (27). Two electrode voltage-clamp recordings were performed 1-10 days after cRNA injection. Unless otherwise indicated, the standard Mg²⁺ solution used to superfuse the oocytes contained 115 mM NaCl, 2.8 mM KCl, 1.8 mM MgCl₂, and 10 mM HEPES, pH 7.2. This Ca²⁺-free solution was used to avoid activation of Ca²⁺-dependent Cl channels. The drugs were prepared as concentrated stocks in 100 mM HEPES, pH 7.2, and stored at 20° until use. Solutions at the desired concentration were freshly made from the frozen stocks and used for 2 hr. The small volume of the bath in the recording chamber (<100 µl) and the high rate of perfusion (7-10 ml/min) allowed a rapid exchange of solutions. The recovery of the current was complete after <2 min of wash-out, even at the highest concentrations of ATP used. Nevertheless, we allowed a 3-min wash-out period between two successive recordings. Unless otherwise indicated, antagonists and Zn²⁺ were coapplied with ATP during perfusion with the standard Mg²⁺ solution. PPADS was preincubated for 4-8 min before ATP stimulation.

Fura-2-based measurements of Ca²⁺ influx. Permanent transfection of HEK 293 cells with a full-length hP2X₄ cDNA cloned in the mammalian expression vector pcDNA3 (InVitrogen, San Diego, CA) was carried out using the calcium phosphate precipitation method (28). Independent foci were selected and expanded in the continuous presence of geneticin (500 µg/ml; Sigma).

	Results

Summary Introduction Materials & Methods Results Discussion References

Structure and tissue distribution of the human P2X₄ receptor. Using degenerate oligonucleotides for conservative P2X sequences and human brain cDNA, we isolated an amplification fragment (~750 bp) that showed a high sequence homology with previously identified P2X receptors. Using this PCR fragment, we screened under low-stringency conditions a human cDNA library derived from total brain mRNA. Eleven independent cDNA clones were isolated that strongly hybridized to the ³²P-labeled probe. Results from restriction endonuclease mapping, Southern blotting, transcription unit search, and DNA sequence analysis revealed that overlapping cDNAs had been identified rather than a full-length clone. A common ScaI site was used to ligate two overlapping clones, thereby completing a full-length cDNA. The open reading frame of the hP2X₄ is 1164 bp long, encoding a 388-amino acid polypeptide. A computer search of GeneBank using BLAST (GCG, Wisconsin Sequence Analysis Software, Genetics Computer Group, Madison, WI) revealed that the encoded protein shares significant sequence identity (87%) (Fig. 1) with that of the recently cloned rP2X₄ (15), suggesting that the cDNA we isolated is the correspondent human isoform (hP2X₄). The hP2X₄ protein has six consensus sites for N-glycosylation, N-X-(S/T), which are conserved in the rat receptor. Differences between the amino acid sequences of both species are mostly conservative substitutions, located within the putative extracellular loop. In addition, three hydrophobic changes are present in the second putative transmembrane segment (Fig. 1).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Amino acid sequence alignment for the hP2X4 and rP2X4 receptors. Open boxes, nonconserved residues. Circles, N-glycosylation sites [consensus site N-X-(S/T)]. Underlined, proposed transmembrane segments. The approximate positions in the correspondent protein sequence for endonuclease restriction enzymes used to create chimeric human/rat constructs are indicated. *, Single-point mutations in the rat P2X4 receptor, Q78K and N127K.

View larger version (52K): [in this window] [in a new window]   Fig. 2.   Tissue distribution of hP2X4 mRNA determined by RT-PCR in a color-inverted video image of the DNA bands stained with ethidium bromide. cDNAs from various human tissues were analyzed by PCR using specific oligonucleotides for the hP2X4 gene, and the amplification products were separated in a 1% agarose gel. A single band of the predicted size was detected (792 bp) in every tissue. Samples using human genomic DNA (0.5 µg) and no DNA (H2O) were performed as negative controls. Size markers, 1-kb ladder (GIBCO).

Chromosomal localization on the hP2X₄ gene on chromosome 12. A 1.7-kb cDNA encoding the hP2X₄ receptor was labeled with biotinylated-dATP by nick translation and hybridized to normal metaphase chromosomes derived from PHA-stimulated peripheral blood lymphocytes using FISH. Under the conditions used, the hybridization efficiency was ~87% for this probe (among 100 checked mitotic figures, 87 of them showed signals on one pair of the chromosomes) (Fig. 3a). The DAPI banding was used to identify the specific chromosome; the assignment between signal from probe and the long arm of chromosome 12 could be obtained (Fig. 3b). There was no additional locus detected by FISH under these experimental conditions. Therefore, the hP2X₄ gene is located at human chromosome 12, region q24.32 (Fig. 3c).

View larger version (40K): [in this window] [in a new window]   Fig. 3.   FISH analysis of human metaphase chromosomes using a biotinylated-human P2X4 probe. a, Representative chromosome spread of PHA-stimulated human lymphocytes showing the localization of FISH signals (arrow). b, Same mitotic figure as in a stained with DAPI to identify the labeling of chromosome 12. c, Chromosome 12 denoting the localization of the hP2X4 gene at 12q24.32.

Functional expression of hP2X₄ yields ATP-activated homomeric channels. To characterize the functional and pharmacological properties of hP2X₄, we expressed the receptor in X. laevis oocytes. In cRNA-injected oocytes, ATP evoked a rapid inwardly rectifying current that slowly desensitized in the continuous presence of the agonist (Fig. 4, a and b). Noninjected oocytes routinely failed to respond to ATP. The rank order of agonist efficacy for different nucleotide analogs (100 µM) follows a pattern similar to that of the rat receptor: ATP  2MeSATP  CTP > ,-meATP > dATP. For rP2X₄ (15), we found the agonist 2MeSATP to be much less effective than ATP. Representative currents elicited by the various agonists are shown in Fig. 4a. No significant responses (<1% of ATP current) were detected with 100 µM of ADP, AMP, ,-meATP, GTP, and adenosine or with the neurotransmitter receptor agonists acetylcholine (with 100 µM atropine), nicotine, glutamate, -aminobutyric acid, glycine, and 5-hydroxytryptamine₃ (all at 100 µM). Construction of a dose-response curve to ATP (Fig. 4c) revealed that the EC₅₀ value was 7.4 ± 0.5 µM ATP and the Hill coefficient, n_H, was 1.4 ± 0.1 (five determinations). These values are virtually identical to the value previously reported for rP2X₄ (15). Partial agonist dose-response curves for CTP and 2MeSATP are also shown in Fig. 4c.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Functional expression of recombinant hP2X4 receptors in X. laevis oocytes. a, Representative inward currents elicited by application (5 sec) of various P2X agonists (100 µM). Holding potential is 70 mV. b, Maximal currents elicited by 50 µM ATP at different holding voltages (normalized to the response at 120 mV, three determinations). c, Dose-response curves of hP2X4 for ATP, 2MeSATP, and CTP (four determinations). Continuous line for ATP, fit to the data using the equation I = Imax/[1 + (EC50/L)nH], where I is the actual current for a ligand concentration (L), nH is the Hill coefficient, and Imax is the maximal current (EC50 = 7.44 ± 0.53, nH = 1.38 ± 0.12). The amount of current obtained at each ATP concentration was normalized to the current obtained with 100 µM ATP. a and b, Mg2+ in the standard solution was replaced by Ca2+. c, Standard Mg2+ solution was used.

Modulation by [Zn²⁺]_o of the hP2X₄. It has been reported that [Zn²⁺]_o increases the amplitude of ATP-activated currents through both native (31-33) and cloned (P2X₂ and P2X₄) P2X receptors. Coapplication of 5 µM ATP and 10 µM Zn²⁺ markedly increases the amplitude of the hP2X₄-evoked current (Fig. 5a). This modulation exhibits a bell-shaped dose-response dependence on the Zn²⁺ concentration (Fig. 5b). Maximal potentiation occurs at relatively low [Zn²⁺]_o in the range of 5-10 µM. A further rise of [Zn²⁺]_o (>100 µM) diminishes the relative potentiation, and at even higher concentrations (1 mM), a dramatic inhibition of the ATP-evoked response is observed. Both the potentiation and inhibition by [Zn²⁺]_o were fully reversed within <3 min of washout of the Zn²⁺ ions. To examine in more detail the [Zn²⁺]_o modulation of the hP2X₄ currents, we analyzed the agonist dose-response curve in the presence of 10 µM [Zn²⁺]_o. In the presence of Zn²⁺, the apparent affinity for ATP is increased (a ~3-fold reduction for EC₅₀) and the dose-response curve shifts to the left. However, the maximal response is not altered (Fig. 5c).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Modulation by [Zn2+]o of hP2X4 ATP-elicited currents. a, Recordings of current activated by 5 µM ATP in the presence or absence of 10 µM Zn2+. Recordings were obtained sequentially from the same oocyte, with a 3-min wash-out between applications. b, Relative enhancement or impairment of maximal ATP elicited currents (compared with control with 5 µM ATP alone) plotted as a function of [Zn2+]o (four determinations). Note the dual effect of Zn2+, potentiation at low concentrations (<10 µM), and block at higher concentration (>100 µM). c, Dose-response curve for ATP in the presence of 10 µM Zn2+. Data were fitted to the equation described in the legend to Fig. 4c (continuous line), giving an EC50 value of 2.39 ± 0.26 with an nH value of 0.99 ± 0.09 (three determinations). For comparison, the fit to the dose-response curve for ATP alone (data from Fig. 4c) is shown (dashed line). The amount of current obtained for each data point was normalized to the current obtained in the presence of 5 µM ATP, in the presence (continuous line) or absence (dashed line) of 10 µM Zn2+, respectively. d, Current ratios obtained under voltage ramps (100 to +70 mV in 150 msec) in the presence or absence of [Zn2+]o, I(presence)/I(absence). All recordings were obtained in standard Mg2+ solution at a holding potential of 70 mV.

The homomeric hP2X₄ channel has a high Ca²⁺ permeability. The hP2X₄ receptor, like the other characterized recombinant P2X receptors, is equally permeable to the monovalent ions Na⁺ and K⁺, whereas it is not permeable to the anion Cl (data not shown). It has been demonstrated that some native and recombinant P2X receptors, including the rP2X₄, conduct Ca²⁺ with high permeability for this divalent cation (5, 14, 15, 34). To determine the relative contribution of Ca²⁺ to the total current through recombinant hP2X₄ channels, we performed simultaneous measurements of membrane current using whole-cell patch-clamp techniques and intracellular Ca²⁺ using the fluorescence indicator Fura-2. The calibration of the receptor-dependent Ca²⁺ influx in physiological solutions was related to pure Ca²⁺ signals through the same receptors when Ca²⁺ was the only charge carrier (calibration solution) (34). After the application of 50 µM ATP to HEK 293 cells expressing recombinant hP2X₄ receptors, we calculated the ratio (F/Q) of the fluorescence decrement (F_390nm) to the total charge (Q_hP2X4) when both calibration and Ringer's solutions were used. A typical recording of F_390nm and voltage-clamped ATP-induced currents in test solutions is shown in Fig. 6. By dividing the F/Q ratio obtained in the Ringer's solution by the F/Q ratio in calibration solution, in which the entire ATP-activated current is expected to be carried exclusively by Ca²⁺, the fractional Ca²⁺ current (30) was calculated to be 8. 24 ± 0.36% (five determinations).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Ca2+ influx through hP2X4 receptors. Simultaneous (bottom) recordings of hP2X4 receptor currents and (F390nm, top) Fura-2 fluorescence at 390 nm were made from hP2X4-transfected HEK 293 cells. The cell was continuously perfused with the test solution and voltage-clamped at 70 mV. Solid bar above the current trace, application of the agonist (ATP, 50 µM). The decrement of F390nm (F390nm) caused by the rise of intracellular Ca2+ is indicated.

The hP2X₄ receptor shows a much higher sensitivity to P2X antagonists than the rP2X₄ receptor. The most unexpected pharmacological properties of the rat P2X₄ receptor are the low sensitivities to the antagonists suramin and PPADS (15). Indeed, sensitivity to suramin has been used to define P2X-mediated currents in native systems. Only in rat submandibular gland has it been reported that suramin is a weak antagonist for ATP-induced currents (18). The human P2X₄ receptor had a low sensitivity for suramin compared with other P2X receptors (Fig. 7a) (IC₅₀ = 178.1 ± 46.9; four determinations), but the antagonist effect was clearly more potent than that for the rat counterpart (IC₅₀ > 500 µM, 5 µM ATP) (15). Prolonged exposure to 100 µM suramin (for 15 min before ATP application) did not reveal any variation in the antagonist efficacy (data not shown). The suramin block was reversed completely after a 3-min perfusion with extracellular solution.

View larger version (14K): [in this window] [in a new window]   Fig. 7.   Inhibition of hP2X4 receptor currents by different antagonists. a, Relative maximal currents (5 µM ATP) obtained in the presence of increasing concentrations of suramin or PPADS. Continuous lines, fits to the data as described in the legend to Fig. 4c. Suramin (IC50 = 178 ± 46.9 µM; four determinations) was coapplied with the agonist. Prolonged exposure (15 min) did not affect the blocking efficiency. PPADS (IC50 = 27.5 ± 3.4 µM; five determinations) was preincubated for 8 min, and subsequently the agonist was applied. For PPADS concentrations of <50 µM, the agonist was coapplied with the appropriate PPADS concentration studied. The recordings were carried out in the standard Mg2+ solution at a holding potential of 70 mV. b, hP2X4 receptor block by using cibacron blue, basilen blue, and bromphenol blue. The maximal relative currents (5 µM ATP) obtained on coapplication of the antagonist and ATP are plotted with the fit to the data (continuous line) as described in the legend to Fig. 4c. The calculated IC50 values are basilen blue, 38 ± 2.1 µM; cibacron blue, 39 ± 10.9 µM; and bromphenol blue, 78 ± 4.7 µM (more than three determinations). Data were obtained in standard Mg2+ solution at a holding potential of 70 mV.

Specific protein domains determine the different pharmacology for antagonists. The important pharmacological differences between these very homologous proteins prompted us to investigate the structural determinants that confer the distinct antagonist phenotypes. Several chimeric receptors between the rP2X₄ and hP2X₄ receptors were made and then expressed in X. laevis oocytes. All of the constructions resulted in functional ATP-activated channels, with wild-type kinetics of activation and desensitization (not shown). We characterized the antagonist sensitivities of these chimeric proteins by using concentrations of suramin (200 µM) or PPADS (100 µM) that permitted us to differentiate between the rat and human phenotypes. Every chimeric receptor containing hP2X₄ sequences in the domain bracketed by the restriction enzymes Kpn2I and PstI (Fig. 1) demonstrated a human wild-type PPADS sensitivity. Furthermore, all of the constructs compassing a rat Kpn2I/PstI sequence were inefficiently blocked by the same drug, thereby resembling the rat receptor (Fig. 8). Therefore, molecular determinants located on the Kpn2I/PstI domain determine a higher PPADS sensitivity for the hP2X₄ receptor.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Specific domains determine the PPADS and suramin blocking differences between the hP2X4 and rP2X4 receptors. Left, analyzed constructions[black (rP2X4) or white (hP2X4) sequences, respectively]. The approximate positions for the endonuclease restriction enzymes used are indicated, as well as the single-point mutations on the rP2X4 receptor. Right, bars, percentage current obtained on coapplication of 200 µM suramin and the agonist (solid bars) or after 4-min preincubation of 100 µM PPADS (open bars) after agonist stimulation (5 µM ATP). The currents were normalized to the control response (100%) obtained on the application of the agonist in the absence of the blocker (six or more determinations). Data were obtained in standard Mg2+ solution at a holding potential of 70 mV.

	Discussion

Summary Introduction Materials & Methods Results Discussion References

We cloned the first P2X purinergic receptor from human brain, and it shows a high degree of homology with the described rP2X₄ (87% identity). The two receptors are equally long (388 amino acids) and differ mostly in the putative extracellular loop. Interestingly, three conservative changes are located in the second hydrophobic segment (M2 in Fig. 1), but complete identity is found in the first putative transmembrane segment (M1). Other important structural features, like N-glycosylation sites and proline or cysteine residues, are conserved.

Using hybridization of a hP2X₄ probe for FISH to normal metaphase chromosomes, the hP2X₄ has been located in chromosome 12, locus q24.32. The hP2X₁ receptor has been previously mapped to chromosome 17 p13.3 (11), and the hP2X₃ gene is located in the long arm of chromosome 11, locus q12 (35). These results reveal that every hP2X receptor gene that has been cloned maps to different chromosomes. A data base search for genetically linked human diseases did not reveal any known pathology that could be related with the hP2X₄ gene.

The pattern of hP2X₄ mRNA expression in human tissues observed by RT-PCR parallels the distribution previously found for the rat homologue receptor (15). We found expression in every tissue studied, including skeletal muscle. Although extracellular ATP activates contraction in vascular and visceral smooth muscle cells as well as in cardiomyocytes (1), fast ATP-activated currents have not been reported in adult skeletal muscle cells. However, ATP-gated ion channels have been extensively characterized in chick embryonic myoblast and myotubes (1). Unfortunately, the functional presence of P2X receptors in human skeletal muscle has not been studied, and the high sensitivity of the RT-PCR technique used in this report does not allow us to discriminate between a certain expression of transcripts in skeletal muscle cells and expression in the vascular system supplying this organ.

The hP2X₄ protein conforms functional homomeric ATP-activated channels when expressed in heterologous cellular systems. The agonist-induced currents recorded from X. laevis oocytes injected with hP2X₄ cRNA show a very similar agonist profile to that reported for the rat isoform (15). Thus, the hP2X₄ receptor exhibits a low apparent ATP affinity (~8 µM) and no sensitivity to ,-meATP (EC₅₀  100 µM). Nevertheless, ATP is a more powerful agonist than 2MeSATP for the hP2X₄, which in agreement with our published data for the rat isoform expressed in X. laevis oocytes (15) and in contrast with another report that described the same potency for both agonists when the rat receptor was expressed in HEK 293 cells (18).

Chelatable zinc is present in presynaptic vesicles of central excitatory neurons, most notably in the hippocampal hilus, CA1, and the mossy fibers (36). P2X₄ transcripts occur in many regions of the central nervous system (15), including all areas of the hippocampus where high density of synaptically released Zn²⁺ has been detected (36). We report here that the hP2X₄ receptor is modulated by [Zn²⁺]_o in a non-voltage-dependent manner. The allosteric modulation of the P2X₄ channel by [Zn²⁺]_o nicely mirrors the behavior of the ATP-activated currents in dissociated neurons from both rat nodose ganglia (31) and rat superior cervical ganglia (32, 33). The expression of the rP2X₄ receptor in these specific neurons (16) suggests that this protein could participate as a subunit of the native P2X receptor, thereby conferring [Zn²⁺]_o sensitivity. Because of the low concentration of Zn²⁺ required to modulate P2X₄ gating, it is reasonable to propose that vesicular release of Zn²⁺ into the synaptic cleft might be a physiological mechanism of regulating P2X₄ channel activity. Interestingly, it was recently reported that after transient forebrain ischemia in rats, [Zn²⁺]_o enters the cells through undetermined channels and accumulates specifically in degenerating neurons in the hippocampal hilus and CA1 (37). The fact that many cells in hypoxia release ATP into the extracellular media raises the intriguing possibility that P2X₄ might be involved in Zn²⁺-induced neuronal death. Nevertheless, the putative Zn²⁺ permeation through P2X receptors has not been addressed.

Although the relative Ca²⁺ permeability to monovalent cations of some recombinant P2X receptors, including rP2X₄, has been previously demonstrated from reversal potential measurements (15), the actual amount of Ca²⁺ influx is still not known. Using a combination of whole-cell patch-clamp and Fura-2 fluorescence measurements in HEK cells transfected with hP2X₄, we determined that the recombinant hP2X₄ channel permeates a substantial amount of Ca²⁺ ions under physiological ionic conditions. The percentage of current carried by Ca²⁺ is ~8%, a value very close to that previously reported for native ATP currents in sympathetic neurons from rat superior cervical ganglia (34). Among the ligand-gated channels, a higher Ca²⁺ fractional current (8-14%) has been reported only for some subunit combinations of N-methyl-D-aspartate-type glutamate receptors (see Ref. 30 and references therein). The high Ca²⁺ permeation through hP2X₄ receptors suggests that activation of these proteins may directly underlie activity-dependent Ca²⁺ signals while contributing to synaptic transmission (38).

The most remarkable difference between the rP2X₄ and hP2X₄ receptors concerns the antagonist sensitivities. The compounds suramin and PPADS block the hP2X₄ with much higher efficiency than rP2X₄ (Table 1), indicating the existence of specific structural determinants controlling the binding for these purinergic antagonists. We identified a domain of ~100 amino acids (81-183) located in the large extracellular loop of the receptor that account for the higher PPADS sensitivity in the human isoform. In the primary sequence, this region is upstream of amino acid 249 (for P2X₄), a position that requires a lysine residue to confer a high and irreversible PPADS sensitivity in the rP2X₄ receptor (18). Therefore, like other ligand-gated channels, a multivalency binding site that involves domains distant in the primary structure is probably essential for PPADS/receptor interactions. Among the 22 amino acids that differ within the PPADS-sensitive domain between the rat and human receptors, only one lysine residue is found exclusively in the human sequence (K127) but not in the rat homologue (N127). Mutation of the rP2X₄ receptor N127K did not generate the human phenotype, suggesting that this residue does not contribute to the chemical binding of PPADS.

A detailed analysis of the different apparent affinities for suramin of the rP2X₄ and hP2X₄ receptors allowed us to identify a single residue that critically controls the blocking efficiency of this compound. Thus, the single-point mutation rP2X₄ (Q78K) enhances the suramin affinity by >50-fold. At physiological pH, suramin is negatively charged; consequently, we propose that electrostatic interactions could stabilize the chemical interaction between suramin and P2X₄ receptors. Protein sequence alignments of the P2X_1-6 receptors (10) show that this residue (amino acid 78 in P2X₄) is different in all cloned channels and that only the hP2X₄ receptor possesses a lysine in this position. Furthermore, the suramin-sensitive receptors P2X₂ and P2X₃ exhibit a protein deletion in this region, suggesting that different channels use different chemical mechanisms for suramin binding and block. A more detailed study is necessary to define the precise structural domains that delineate the binding sites of suramin and PPADS in the purinergic ligand-gated channels.