Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

P2X receptors represent a family of ATP-gated ion channels that play a significant role in fast excitation and synaptic transmission in many excitable tissues (Ralevic and Burnstock, 1998). As far as sensory neurons are concerned, P2X receptors have been implicated in the direct excitation of primary afferent nerve fibers (Bland-Ward and Humphrey, 1997; Cook et al., 1997; Dowd et al., 1998; Burgard et al., 1999; Hamilton and McMahon, 2000), in ascending excitation of second order sensory neurons in the dorsal horn of the spinal cord and dorsomediolateral nuclei of the brainstem (Li and Perl, 1995; Bardoni et al., 1997; Gu and MacDermott, 1997; Scislo et al., 1997; Li et al., 1998; Tsuda et al., 1999) and in the excitation of higher order neurons in the central nervous system (Dreissen et al., 1998; Ralevic et al., 1999).

Thus far, cDNA sequences for seven P2X subunit proteins (P2X_1-7) have been cloned from mammalian tissues. These P2X subunit proteins can combine to form ion channels, as homomeric and heteromeric assemblies of three, possibly four, P2X subunits (Kim et al., 1997; Nicke et al., 1998; Torres et al., 1999). Transcripts for six subunits (P2X_1-6) have been found in sensory neurons (Chen et al., 1995; Collo et al., 1996; Cook et al., 1997). However, immunohistochemical studies reveal a predominance of P2X₃-like protein in the cell bodies of sensory ganglia and associated sensory nerve endings in skin and lamina II of the spinal cord. P2X₃-like immunoreactivity frequently colocalizes with P2X₂-like immunoreactivity in sensory neurons (Cook et al., 1997; Vulchanova et al., 1997; Bradbury et al., 1998; Xiang et al., 1998). Small sensory DRG neurons containing P2X₃-immunopositive material also contain specific biochemical markers for nociceptor cells (Chen et al., 1995; Bradbury et al., 1998).

It is well known that sensory neurons show a marked variability in the time course of responses to P2X receptor agonists. Agonist-evoked ion currents in many rat DRG and trigeminal ganglion neurons are known to activate and inactivate rapidly, in a manner similar to agonist responses of homomeric P2X₃ receptors (Chen et al., 1995; Cook et al., 1997; Rae et al., 1998). Such fast agonist responses are absent in sensory neurons of P2X₃-null (/) animals (Cockayne et al., 2000). Rat nodose ganglia and bullfrog DRG neurons often respond to P2X agonists with slowly inactivating ionic currents that are similar to the responses of heteromeric P2X_2/3 receptors (Bean, 1990; Khakh et al., 1995; Lewis et al., 1995). Furthermore, some DRG neurons produce composite responses with both rapidly and slowly decaying ion currents (Burgard et al., 1999; Grubb and Evans, 1999).

These differences in the operational profile of native P2X receptors may be due to different levels of expression of P2X₂, P2X₃, and P2X_2/3 receptors in individual sensory neurons (Burgard et al., 1999). Here, we have investigated the operational profiles of the recombinant P2X receptors generated by coexpressing varying amounts of P2X₂ and P2X₃ subunits in Xenopus oocytes. The resultant spectrum of operational profiles was attributed to a heterogeneous mixture of homomeric and heteromeric P2X receptors, and this mixture might conveniently explain the range of operational profiles of sensory neurons that also express P2X₂ and P2X₃ receptor subunits in varying amounts.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of Oocytes and Expression of Recombinant P2X Receptors. Xenopus laevis were anesthetized with Tricaine (0.2%, w/v) and killed by decapitation. The preparation of defolliculated Xenopus oocytes has been described previously (King et al., 1997). Defolliculated oocytes (stages V and VI) were injected (40 nl) cytosolically with capped ribonucleic acid (cRNA) encoding either rat P2X₂ or rat P2X₃ receptor subunits. In coexpression experiments, oocytes were injected with a mixture of cRNAs prepared at four concentration ratios (1:500, 1:100, 1:50, and 1:20), by mixing P2X₂ cRNA (2, 10, 20, and 50 µg ml¹) with an equal volume of P2X₃ cRNA (1 mg ml¹). Injected oocytes were incubated at 18°C in Barth's solution (pH 7.5) containing (mM): NaCl 110, KCl 1, NaHCO₃ 2.4, Tris-HCl 7.5, Ca(NO₃)₂ 0.33, CaCl₂ 0.41, MgSO₄ 0.82, supplemented with gentamycin sulfate 50 µg l¹ for 48 h to allow full receptor expression and then stored at 4°C for up to 12 days.

Solutions and Electrical Recording of cRNA-Injected Oocytes. Nucleotide-evoked membrane currents were recorded from cRNA-injected oocytes studied under voltage-clamp conditions using a twin-electrode amplifier (Axoclamp 2B, Foster City, CA). Intracellular microelectrodes had a resistance of 1 to 2 M when filled with 3 M KCl. Oocytes were perfused constantly (at 5 ml min¹) with Ringer's solution containing (mM): NaCl 110, KCl 2.5, HEPES 5, BaCl₂ 1.8, pH 7.4 to 7.5. All recordings were made at room temperature (18°C) at a holding potential of 50 mV (unless stated otherwise). Electrophysiological data were recorded on a chart recorder (Gould 2200S, Ilford, UK).

Compounds and Salts Used. All common salts were AnalaR grade (Aldrich Chemicals, Gillingham, UK). Adenosine 5'-triphosphate disodium salt (ATP) was purchased from Boehringer (Mannheim, Germany). Adenosine-5'-O-(3-trio)triphosphate (ATPS), P¹,P⁵-diadenosine pentaphosphate (Ap₅A), ,-methylene ATP (,-meATP), ,-methylene ATP (,-meATP), uridine 5'-triphosphate (UTP), and Tricaine (3-aminobenzoic acid ethyl ester) were purchased from Sigma Chemical Co. (Poole, UK), pyridoxal-5-phosphate-6-azophenyl-2',4'-disulfonic acid (PPADS) from RBI-Sigma (Natick, NJ) and 2',3'-O-trinitrophenyl-ATP (TNP-ATP) from Molecular Probes (Cambridge, UK). Suramin was a gift from Bayer plc (Newbury, UK) and Ip₅I a gift from Dr. Jesus Pintor (University of Complutense, Madrid, Spain). Solutions of agonists and antagonists were prepared daily from a stock solution (10 or 100 mM, stored frozen) made up in extracellular bathing solution.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Three groups of cRNA-injected defolliculated oocytes were tested. The first and second groups comprised oocytes injected with P2X₂ or P2X₃ cRNAs, respectively. The third group of oocytes was injected with a fixed amount of P2X₃ cRNA and varying amounts of P2X₂ cRNA. The first and second groups behaved as previously observed for homomeric P2X₂ and P2X₃ receptors (King, 1998). The third group of oocytes showed a spectrum of operational profiles that ranged from predominantly P2X₃-like to predominantly P2X₂-like.

Use of ,-meATP to Distinguish Types of P2X Receptors. ,-meATP reliably distinguishes homomeric P2X₂ receptors from homomeric P2X₃ receptors (Fig. 1A). ,-meATP (100 µM) did not activate any detectable current at P2X₂ receptors, whereas a rapidly activating and rapidly inactivating current was evoked at P2X₃ receptors. Sham-injected control oocytes did not produce any detectable responses to ,-meATP (100 µM).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   A, ,-meATP (100 µM) evoked whole-cell inward currents at homomeric P2X3 receptors but not homomeric P2X2 receptors. B, whole-cell currents evoked by ,-meATP (3 µM) in oocytes injected with P2X2 and P2X3 cRNAs (at concentration ratios of 1:500, 1:100, 1:50, and 1:20). Each recording consisted of fast (I1) and slow (I2) inward currents. C, consecutive applications of ,-meATP (1 µM, applied 40 s apart) evoked whole-cell inward currents where the initial fast current became desensitized after the first application of the agonist, whereas the amplitude of slow current was unchanged. A through C, Vh = 50 mV.

Potency and Efficacy of ,-meATP at P2X Receptors. Concentration/response (C/R) curves for the fast current (I₁) evoked by ,-meATP were reliably determined for those oocytes injected with mixed cRNAs ratios of 1:500 and 1:100 (Fig. 2A). Here, EC₅₀ values (and Hill coefficients) for agonism were 1.7 ± 0.3 µM (0.8 ± 0.1; n = 7) and 1.6 ± 0.4 µM (0.8 ± 0.2; n = 5), respectively. For oocytes injected with mixed cRNAs ratios of 1:50 and 1:20, the amplitude of the fast current to ,-meATP could not be easily resolved from the slow current (I₂) over the full range of agonist concentrations used and, accordingly, it was not possible to determine EC₅₀ values here. In contrast, ,-meATP yielded an EC₅₀ value of 1.9 ± 0.3 µM (0.8 ± 0.1; n = 6) at homomeric P2X₃ receptors (Fig. 2A).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   A, C/R curves are shown for fast currents (I1) activated by ,-meATP in oocytes injected with cRNA ratios of 1:500 and 1:100, and oocytes expressing homomeric P2X3 receptors. Here, the agonist was equipotent for fast responses (I1) and responses of P2X3 receptors. B, C/R curves are shown for slow currents (I2) activated by ,-meATP in oocytes injected with cRNA ratios of 1:500, 1:100, 1:50, and 1:20. Changing the amounts of cRNAs injected failed to influence agonist potency for slow responses (I2). C, the mean amplitudes of fast (I1) and slow (I2) currents are shown for oocytes injected with cRNAs at 1:20, 1:100, and 1:500 ratios. Here, agonist efficacy for I1 and I2 responses was highly dependent on the relative amounts of cRNA injected. Data are expressed as mean ± S.E. of five to seven cells. cRNA (P2X2:P2X3): , 1:20; , 1:100; , 1:500.

Slow Currents to ATP and ,-meATP Involve Different P2X Receptors. ATP and other nucleotides were tested on oocytes coinjected with mixed cRNAs (1:500 ratio), then tested again on homomeric P2X₂ receptors. EC₅₀ values for these nucleotides at different P2X receptors are given in Table 1. ATP, 2-methylthioATP (2-MeSATP), and ATPS evoked slow currents at homomeric P2X₂ receptors and produced slow ion currents in oocytes coexpressing P2X₂ and P2X₃ subunits. The potency of ATP and 2-MeSATP, but not ATPS, was lower at homomeric P2X₂ receptors than at the P2X receptor population formed by P2X₂ and P2X₃ subunit coexpression. Furthermore, four nucleotides (,-meATP, ,-meATP, Ap₅A, and UTP) were inactive at homomeric P2X₂ receptors but were agonists at presumptive P2X_2/3 receptors (Table 1).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   A, C/R curves are shown for slow currents (I2) activated by a series of nucleotides at P2X receptors formed by P2X2 and P2X3 subunit coexpression. The maximum response (or efficacy) to ATP, 2-MeSATP, and ATPS was markedly greater than that of ,-meATP. B, C/R curves for slow currents activated by a second series of nucleotides at P2X receptors formed by P2X2 and P2X3 subunit coexpression. The efficacy of Ap5A, ,-meATP, and UTP was similar to, or less than, that of ,-meATP. C, the successive application of ,-meATP and ATP activated two independent slow currents. The sum of these two slow currents (see arrows) equaled the amplitude of an earlier response to ATP, where ,-meATP desensitization had not yet occurred.

Antagonism of the Heteromeric P2X_2/3 Receptor. A way was devised to study P2X_2/3 receptors in isolation, using ,-meATP to activate P2X_2/3 receptors and employing a desensitizing procedure of two agonist pulses (40 s apart) to inactivate all homomeric P2X₃ receptors present (as shown in Fig. 1C). The blocking activity of a series of P2 receptor antagonists was tested on the residual slow response evoked by the second pulse of ,-meATP. Suramin (0.01-100 µM), RB-2 (0.03-100 µM), PPADS (0.03-100 µM), and TNP-ATP (0.001-1 µM) caused a concentration-dependent inhibition of P2X_2/3 receptors, whereas Ip₅I (0.01-100 µM) was inactive. IC₅₀ values are given in Table 2. The heteromeric P2X_2/3 receptor was remarkably sensitive to TNP-ATP (IC₅₀, 11 nM), a feature shared with the homomeric P2X₃ receptor (IC₅₀, 0.3 nM). Except for suramin, all antagonists, including TNP-ATP appeared to work in noncompetitive manner, although their blocking actions were slowly reversible after washout. For suramin, a pA₂ value of 5.9 ± 0.4 was determined at the heteromeric P2X_2/3 receptor.

Modulation of the Heteromeric P2X_2/3 Receptor. The same procedure of ,-meATP agonism and twin agonist pulses (40 s apart) were used again to study P2X_2/3 receptors in isolation and to assess the modulatory effects of extracellular H⁺ and Zn²⁺ ions. Under these circumstances, alteration in the extracellular pH (pH 8.0-6.5) caused only a modest change in agonist potency at P2X_2/3 receptors (Fig. 4A). These findings contrasted with the strong effect of pH on agonist potency at P2X₂ receptors and weak inhibitory effect of H⁺ on agonism of P2X₃ receptors (see Table 3).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   A, C/R curves are shown for slow currents (I2) activated by ,-meATP at P2X2/3 receptors for the given extracellular pH levels. B, effect of extracellular Zn2+ ions on the amplitude of slow currents activated by ,-meATP (1 µM) at P2X2/3 receptors.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study was carried out for two reasons. First, we wished to characterize the pharmacological profile of the P2X_2/3 receptor, about which surprisingly little is known apart from ,-meATP agonism (Lewis et al., 1995) and TNP-ATP antagonism (Thomas et al., 1998). Second, we wished to shed further light on the variability of P2X responses in sensory neurons, in which P2X₂ and P2X₃ subunits commonly occur and where P2X_2/3 receptors may play a role in sensory transmission.

Sensory neurons show a marked variability in the time course of responses to P2X receptor agonists. Grubb and Evans (1999) reported that the majority (80%) of adult rat DRG neurons responded to nucleotides with a P2X₃-like phenotype, and the remainder possessed noninactivating P2X receptors of undetermined phenotype. Burgard and colleagues (1999) reported similar findings in adult rat DRG neurons, where 70% of IB4-labeled nociceptors showed fast-desensitizing P2X₃-like agonist responses and the remaining 30% showed either mixed (fast and slow responses) or slow responses. To exemplify these phenotypes, we carried out a brief survey of ATP responses in rat neonatal DRG neurons and found qualitatively and quantitatively similar results (Fig. 5). In this exploratory survey, 50% of small DRG neurons produced fast P2X₃-like ATP responses, 30% gave slowly desensitizing responses, and 20% gave mixed responses. Burgard and colleagues (1999) have attributed the last phenotype to mixed P2X receptors resulting from the coexpression of P2X₂ and P2X₃ subunits, and this observation motivated us to study the pharmacological properties of heteromeric P2X_2/3 receptors as fully as possible.

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Whole-cell inward currents activated by ATP (10 µM) in rat neonatal DRG neurons. Three phenotypes were observed in 209 cells studied under patch-clamp conditions (Vh = 60 mV) [for methods, see Dunn et al. (2000)].

Coexpression of P2X₂ and P2X₃ subunits in approximately equal amounts in HEK 293 cells was reported, in the first study of its kind, to generate a hybrid P2X_2/3 receptor that showed the agonist profile of P2X₃ receptors and inactivation properties of P2X₂ receptors (Lewis et al., 1995). However, we have found that the outcome of P2X subunit coexpression is not straightforward. Where P2X₃ expression was kept constant and the level of P2X₂ expression gradually increased, the resultant P2X receptors changed their phenotype from predominantly P2X₃-like to a mixed phenotype of P2X₂, P2X₃ and P2X_2/3 receptors. Thus, coexpression of two P2X subunits, where each can independently form functional homomeric receptors, seems to generate a heterogeneous population of P2X receptorsin this case P2X₂, P2X₃, and P2X_2/3 receptors. This finding suggests that the process of P2X receptor assembly occurs in a stochastic manner, the probability of forming either homomeric or heteromeric receptors being dependent on the relative numbers of the two types of P2X subunit available for receptor assembly.

For each of the four levels of P2X₂/P2X₃ subunit coexpression studied, the resultant population of recombinant P2X receptors gave rise to complex agonist responses. These responses involved an initial fast current that, we believe, was mediated by a small population of homomeric P2X₃ receptors. This conclusion was based in part on the finding that ,-meATP was equipotent for the P2X₃-like fast current and homomeric P2X₃ receptor itself, and ,-meATP caused prolonged desensitization in both cases. Also, other P2X₃ receptor agonists (e.g., Ap₅A) evoked fast currents in those oocytes coexpressing P2X₂ and P2X₃ subunits. Additionally, we have already shown that Ip₅I blocks the P2X₃-like fast current, blocks P2X₃ receptors, and also blocks the fast current of DRG sensory neurons with similar IC₅₀ values (King et al., 1999; Dunn et al., 2000). Lastly, the amplitude of the P2X₃-like fast current was reduced as more P2X₂ subunits were expresseda consequence, we believe, of the reduction in the numbers of P2X₃ receptors as their constituent subunits were incorporated instead into P2X_2/3 receptors.

The initial P2X₃-like fast current was followed by a slow current that, in all probability, involved P2X_2/3 receptors when ,-meATP was the agonist, and both P2X₂ and P2X_2/3 receptors when ATP was the agonist. Significantly, other studies of P2X₂ and P2X₃ coexpression have also alluded to the dual formation of P2X₂ and P2X_2/3 receptors (Thomas et al., 1998; Ueno et al., 1998). A slow current to ,-meATP is a recognized signature of P2X_2/3 receptors (Lewis et al., 1995; Ueno et al., 1999), but surprisingly little else can be said with certainty about its agonist profile. Agonism of P2X_2/3 receptors was reassessed recently (Bianchi et al., 1999), although these investigators may not have taken into account the dual formation of P2X₂ and P2X_2/3 receptors in their expression system.

The P2X_2/3 receptor was further characterized in our experiments by an agonist potency order of Ap₅A > ,-meATP > ,-meATP > UTP, which took into consideration that none of these nucleotides can activate P2X₂ receptors. In all probability, ATP, 2-MeSATP, and ATPS are agonists at P2X_2/3 receptors also, but a true assessment of their potency will require the discovery of selective P2X₂ receptor antagonists to strip away the complication of coactivating P2X₂ and P2X_2/3 receptors. The agonist potency order for homomeric P2X₃ receptors was Ap₅A > ,-meATP > ,-meATP > UTP (setting aside data for ATP, 2-MeSATP, and ATPS), which matched the rank order for heteromeric P2X_2/3 receptors, although the EC₅₀ values for these two P2X subtypes are not identical (see Table 1).

The antagonist potency order at P2X_2/3 receptors was TNP-ATP > PPADS = suramin > RB-2, with Ip₅I inactive. These antagonists were tested only on slow currents to ,-meATP and, therefore, any interaction of these blocking agents with P2X₂ and P2X₃ receptors can be discounted. TNP-ATP was potent in the near nanomolar concentration range, in agreement with an earlier study (Thomas et al., 1998). However, our analysis of the inhibition curve (n_H ~ 1) suggested only a single population of heteromeric P2X_2/3 receptors. PPADS, suramin, and RB-2 also blocked the P2X_2/3 receptor in micromolar concentrations, where the potency order (PPADS = suramin > RB-2) differed from that for homomeric P2X₂ receptors (RB-2 > PPADS > suramin) and P2X₃ receptors (PPADS > suramin > RB-2). Regarding our negative results with Ip₅I, it is of note that slow currents to ,-meATP in sensory neurons (nodose and DRG) are similarly unaffected by this dinucleotide (Dunn et al., 2000) and provide indirect proof that such slow responses probably involve native P2X_2/3 receptors.

The potentiating effect of extracellular H⁺ and Zn²⁺ was modest at recombinant P2X_2/3 receptors compared with their reported facilitatory actions on P2X₂ receptors (Wildman et al., 1998). Regarding the effects of H⁺ ions, our results revealed a 4-fold increase in agonist potency at P2X_2/3 receptors in changing from pH 8.0 to 6.5, and a 21-fold increase for P2X₂ receptors. Regarding the effects of Zn²⁺ ions, agonist activity was enhanced about 2-fold at P2X_2/3 receptors, but 14-fold at P2X₂ receptors. Both H⁺ and Zn²⁺ ions potentiate ATP-activated slow currents in nodose ganglia, which express P2X₂ and P2X₃ subunits (Li et al., 1997). From our present data, it is likely that any H⁺- and Zn²⁺-based potentiation of ATP-activated responses in sensory ganglia will involve both P2X₂ and P2X_2/3 receptors.

In summary, we have carried out a sequential pharmacological survey of the possible homomeric and heteromeric P2X receptors that could be generated by coexpression of P2X₂ and P2X₃ subunits in the same cell. Our experiments indicate that, where cell types naturally express more than one P2X subunit, it is likely that a mixture of homomeric and heteromeric receptors will be present in these cells. Where the P2X_2/3 receptor was studied in isolation, we found that the pharmacological profile is comparatively unique and does not seem to be governed by either of the constituent P2X₂ and P2X₃ subunits. Thus, ,-meATP (but not the structurally related ,-meATP) is a potent agonist at P2X_2/3 receptorsa profile that does not fit either homomeric P2X₂ or P2X₃ receptors. Also, the potency of the antagonists tested here on P2X_2/3 receptors, particularly TNP-ATP, RB-2, and the inactive Ip₅I, did not mirror our findings with homomeric P2X₂ and P2X₃ receptors. These pharmacological findings indicate that the therapeutic value of P2X subunit-selective agonists and antagonists (a popular objective at this point in time) might not be entirely appropriate and that selective agonists and antagonists for those P2X heteromultimers found in tissues need also be sought.