Introduction

Top Summary Introduction Procedures Results Discussion References

There are seven P2X receptor subunits, which assemble into ATP-activated ion channels either as homomers or heteromers (reviewed by North, 1996; North and Barnard, 1997). At the molecular level, any pair of the subunits has 35-50% identical amino acids. At the functional level, several subgroups have been distinguished. For example, in one subgroup (P2X₁ and P2X₃ homomeric channels), meATP and ATP are equally effective agonists, and the currents desensitize during agonist applications of more than several hundred milliseconds. None of the other homomeric channels is activated by meATP, and the currents show much less desensitization. A distinct class of channel is formed by the coexpression of P2X₂ and P2X₃ subunits; this class is activated by meATP and ATP but it shows little desensitization. A further distinguishing feature is the ability of PPADS to block the currents evoked by ATP; P2X₄, P2X₆, and P2X₇ receptors are relatively insensitive. Finally, P2X₇ homomeric channels are fundamentally different from all the others because repeated or prolonged agonist application results in cell permeabilization as measured by the uptake of fluorescent dyes and, eventually, cell lysis (North, 1996; Surprenant et al., 1996; North and Barnard, 1997).

The assignment of functional roles for P2X receptors in intact tissues depends critically on the use of receptor antagonists. Indeed, the main evidence that ATP mediates synaptic transmission between neurons (Edwards et al., 1992; Evans et al., 1992) or from nerve to muscle (Sneddon and Westfall, 1984; Evans and Surprenant, 1992) has been the block of the postsynaptic responses by suramin and/or PPADS (Sneddon and Westfall, 1984; Dunn and Blakeley, 1988; Ziganshin et al., 1994). However, the low affinity and limited specificity of these compounds restricts their usefulness and, as mentioned above, some P2X receptors are not blocked (Buell et al., 1996). There is a clear need to identify more receptor antagonists.

Trinitrophenyl analogs of ATP have been widely used for the fluorescent labeling of ATP binding sites in proteins, including P2X receptors (Mockett et al., 1994). We first examined their effects on cloned and expressed P2X receptors with such an application in mind. In the course of those experiments, it became clear that, for some P2X receptors, the analogs were able to block responses to ATP at nanomolar concentrations. Here we report the characterization of this observation.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

HEK 293 cells that stably or transiently express the following P2X receptors were used in these studies: human P2X₁, rat P2X₂, rat P2X₃, rat or human P2X₄, rat P2X₂ together with rat P2X₃ (heteromer), and rat P2X₇. Generation of stable P2X receptor-expressing cell lines and methods of transient lipofectin transfection have been described in detail previously (Evans et al., 1995; Buell et al., 1996; Evans et al., 1996; Kawashima et al., 1997). HEK cells stably transfected with the human P2X₄ receptor were generously provided by Professor W. Stuhmer, Max-Planck Institute (Gottingen, Germany). Cells were plated onto 12-mm glass coverslips and maintained in Dulbecco's modified Eagle's medium, Nutrient Mix F-12 (GIBCO-BRL, Bethesda, MD) supplemented with 10% heat-inactivated fetal calf serum (FAKOLA, Bern, Switzerland) and 2 mM L-glutamine at 37° in a humidified 5% CO₂ incubator.

Whole-cell recordings were made 12-48 hr after transient transfection (rat P2X₁, P2X₃, P2X₄) and 6-72 hr after passage of stable cell lines (human P2X₁, P2X₃, P2X₄, and rat P2X₂, P2X_2/3, and P2X₇). Currents were recorded with an EPC9 patch-clamp amplifier (HEKA Elektronik, Lambrecht/Pfalz, Germany), acquired (1-2 kHz) and analyzed with Pulse and PulseFit 8.02 (HEKA). Patch pipettes (4-7 M[/omega]) contained 140 mM NaCl, 10 mM HEPES, and 11 mM EGTA. The external solution was 147 mM NaCl, 10 mM HEPES, 12 mM glucose, 2 mM KCl, 2 mM CaCl₂, and 1 mM MgCl₂. Osmolarity and pH values of both solutions were maintained at 300-315 mOsM/liter and 7.3, respectively. Unless otherwise stated, experiments were performed at a holding potential of 60 mV and at room temperature. Agonists were applied using a fast-flow U-tube delivery system (Fenwick et al., 1982). Antagonists were added to both the bath superfusate and the fast-flow solution. ATP was the agonist in all experiments on P2X₁, P2X₂, P2X₄, and P2X₇ receptors. Both ATP and meATP were used at the P2X₃ receptor and only meATP was used at the heteromeric P2X_2/3 receptor (Kawashima et al., 1997). Agonists were applied for 0.5-2-sec duration at 2 min intervals at all receptors except P2X₁ and P2X₃ where 4-5 min intervals were required to allow recovery from desensitization (Evans et al., 1995).

Agonist concentration-response curves for each cell were fit by the least-squares method to I = I_max[1 + (EC₅₀/[A])^n_H] where I is the peak current evoked by agonist concentration [A], I_max is the peak current evoked by a maximal agonist concentration, EC₅₀ is the concentration giving half the maximal current, and n_H is the Hill coefficient. Antagonist concentration-inhibition curves were obtained in individual cells by using a fixed agonist concentration close to the EC₅₀ (1 µM ATP at P2X₁, 10 µM ATP at P2X₂ and P2X₄, 300 µM ATP at P2X₇, 1 µM ATP or meATP at P2X₃, and 5 µM meATP at P2X_2/3), and progressively increasing the concentration of antagonist: IC₅₀ values were calculated by least squares fitting to I = I₀/[1 + (IC₅₀/[Ant])ⁿH], where I and I₀ are peak currents in the presence and absence of antagonist at concentration [Ant]. All EC₅₀ and IC₅₀ values given in text and tables are the mean ± standard error from individual cells; however, the graphs in the figures were drawn by averaging results from all experiments and fitting a single concentration-response curve to the pooled data.

TNP-AMP sodium salt, TNP-ADP disodium salt, TNP-ATP trisodium salt, and TNP-GTP trisodium salt were obtained from Molecular Probes (Eugene, OR). GTP, ATP, and meATP were from Sigma (St. Louis, MO) and 2,4,6-trinitrophenol (picric acid) was from Fluka (Buchs, Switzerland). TNP-A was prepared from adenosine and 2,4,6-trinitrobenzenesulfonate according to the procedure published by Azegami and Iwai (1975); the red precipitate of TNP-A that crystallized from the solution was purified by reprecipitation from acetone (1 ml) by adding 10 volumes of toluene.

	Results

Top Summary Introduction Procedures Results Discussion References

TNP-ATP strongly inhibited currents in cells expressing P2X₁, P2X₃, or P2X_2/3 receptors (IC₅₀ about 1 nM), but was very much less effective in cells expressing P2X₂, P2X₄, or P2X₇ receptors (IC₅₀ > 1 µM)(Figs. 1 and 2; Table 1). The inhibition was concentration-dependent and well fitted by the logistic function (see Experimental Procedures); the IC₅₀ values are shown in Table 1, and the coefficient ^n_H was not significantly different from unity. The inhibition reversed within 4-15 min of TNP-ATP washout, although reversal was sometimes incomplete for near maximal concentrations. The inhibition was the same at holding potentials of 60 mV and 40 mV (n = 4, 6, 5, and 3 for cells expressing P2X₃, P2X_2/3, P2X₂, and P2X₄ receptors, respectively). The P2X₇ receptor was particularly insensitive to blocking by TNP-ATP; at the highest concentration tested (30 µM), the inhibition was 39 ± 2% (n = 6). TNP-ATP (1 nM-30 µM) had no agonist action at any of the P2X receptors.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   TNP-ATP is a potent antagonist at P2X1, P2X3, and P2X2/3 receptors. Each set of records consists of superimposed currents recorded from individual HEK 293 cells expressing the indicated receptor before, during, and after application of TNP-ATP at 10 nM (P2X1, P2X3, and P2X2/3) or at 30 µM (P2X2, P2X4, and P2X7). Currents shown in the presence of TNP-ATP are after 4-min application; currents shown after TNP-ATP are at 4-min wash except for P2X1, in which case the washout was for 8 min.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   TNP-ATP concentration-inhibition curves generated from all experiments as illustrated in Fig. 1. Results are plotted as normalized current, where current in absence of antagonist is equal to 1. Points, mean ± standard error of 4-8 experiments. Lines, least-square fits to a logistic equation (see Experimental Procedures).

View this table: [in this window] [in a new window]   TABLE 1 Inhibition of currents evoked by ATP or meATP in cells expressing P2X receptor Values are mean ± standard error with numbers of experiments in parentheses.

The effect of TNP-ATP was mimicked by TNP-ADP and TNP-AMP, as well as TNP-GTP (Fig. 3), although TNP-A had no effect (n = 4). These compounds were also highly effective at P2X₁, P2X₃, and P2X_2/3 receptors but much less so at P2X₂, P2X₄, and P2X₇ receptors. Complete antagonist-inhibition curves were generated for P2X₂, P2X₃, and P2X_2/3 receptors (Fig. 3), and IC₅₀ values are provided in Table 1. The dose-inhibition curves at the heteromeric P2X_2/3 receptors were consistently to the right of those for the homomeric P2X₃ receptor (Fig. 3), although the difference in IC₅₀ estimates was significant only in the case of TNP-ATP (Table 1). As for TNP-ATP, the other TNP-nucleotides (up to 30 µM) had no agonist action. TNP-A (0.1-1 µM), GTP (0.1-10 µM), and picric acid (10 µM) had no agonist or antagonist action at P2X₃, P2X_2/3, or P2X₂ receptors.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   TNP-AMP, TNP-ADP, and TNP-GTP are also potent antagonists at P2X3 and P2X2/3 receptors. Sets of traces, currents recorded from a single cell expressing P2X2 (left), P2X3 (middle), or P2X2/3 (right) receptors in the absence and presence of increasing concentrations of TNP-GTP, as indicated. Graphs are concentration-inhibition curves obtained from all such experiments for TNP-AMP (left), TNP-ADP (middle), and TNP-GTP (right). Points, mean ± standard error of 4-8 experiments.

The nature of the inhibition was examined further in the case of the P2X₃ receptor by constructing full agonist concentration-response curves. With either ATP or meATP as the agonist, TNP-ATP (3 and 10 nM, respectively) caused both a rightward shift and a depression of the maximal current, indicating insurmountable antagonism. For the two antagonist concentrations ([B])(3 and 10 nM, respectively), the curves were fit by an expression appropriate to noncompetitive antagonism [I/I_max = [1 + (EC₅₀/[A])]¹ (1 + K_B/[B])¹], which provided estimates of K_B of about 2 nM. Similar results were obtained for TNP-ADP and TNP-GTP (data not shown).

	Discussion

Top Summary Introduction Procedures Results Discussion References

The results indicate that certain nucleotides with a ribose-substituted trinitrophenyl group are potent antagonists at those P2X receptors that can be activated by meATP (P2X₁, P2X₃, and P2X_2/3). However, several observations are not readily reconciled with the notion that the TNP-nucleotides are binding to the site occupied by the agonists when they act to open the P2X receptor channel. First, the antagonism is noncompetitive (Fig. 4). Second, both guanine and adenine nucleotides are equally effective; this is in marked contrast to the lack of any agonist activity by GTP itself. Third, removal of one or even two phosphate groups from TNP-ATP had no significant effect on the antagonism; yet in terms of agonist action, ADP is more than 100-fold less potent than ATP at the P2X₃ receptor (Lewis et al., 1995) and AMP (100 µM) has no effect at the P2X₁ (Evans et al., 1995), P2X₃ (Chen et al., 1995) or P2X_2/3 receptor (unpublished observations). Removal of the third phosphate, as in TNP-A, resulted in loss of antagonism. In brief, the antagonist binding site differs from the agonist binding site in that it does not discriminate between guanine or adenine bases, absolutely requires the 2',3'-trinitrophenol, and will accept one, two, or three (but not zero) 5'-phosphates.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   TNP-ATP is a noncompetitive antagonist at the P2X3 receptor. a, Currents recorded from a single cell expressing the P2X3 receptor in the absence (top) and presence (bottom) of 10 nM TNP-ATP. b and c, Agonist concentration-response curves for meATP (b) and ATP (c) in the absence and presence of 3 nM or 10 nM TNP-ATP. Points, mean ± standard error of 8-9 experiments for meATP and 3-6 experiments for ATP.

It is possible that the TNP-nucleotides directly block the conducting pathway of the channel, as found for the outwardly rectifying chloride channel (Paulmichi et al., 1992; Venglarik et al., 1993). This seems unlikely both because ATP is negatively charged and the channel is cation-selective, and because the inhibition by TNP was not different for inward and outward currents. In any event the concentrations of extracellular ATP and TNP-ATP that block the outwardly rectifying chloride current are still some hundred-fold higher than those effective at P2X₁, P2X₃, and P2X_2/3 receptors. The most likely mechanism, therefore, is the binding of TNP-nucleotides to an allosteric site on the large extracellular region of the receptor. In this case, the P2X₁ and P2X₃ subunits might provide a common domain that interacts with the strongly electronegative trinitrophenyl moiety. It is interesting that the most potent antagonists in a series tested on the rat urinary bladder (which expresses P2X₁ receptors) also had large aromatic 3' substitutions (Bo et al., 1994; Burnstock et al., 1994); these bound with affinities in the 10-100 nM range.

The weak antagonism of TNP-ATP at other receptors had previously been reported for cochlear hair cells isolated from guinea pig organ of Corti (Mockett et al., 1994), which are known to express P2X₂ receptors (Housley et al., 1995; Brandle et al., 1997). In that case, 75 µM TNP-ATP almost completely blocked the current evoked by 10 µM ATP. TNP-ATP is not an effective antagonist in the rat parotid gland (Soltoff et al., 1993), which contains P2X₄ (Buell et al., 1996) and P2X₇ receptors (Collo et al., 1997). It will clearly be important to test the TNP analogs on P2X responses to ATP in other tissues. On the basis of the present work, one might expect blockade in the nanomolar concentration range to indicate that the underlying receptor contains P2X₁ or P2X₃ subunits. The results with heteromeric receptors, such as are expressed by some primary afferent neurons (Lewis et al., 1995; Cook et al., 1997), might be less straightforward. In the present work, we used meATP as the agonist in experiments on the cells expressing the heteromeric P2X_2/3 receptor, on the assumption that it activates only heteromers and not any homomeric P2X₂ receptors that might also be present. The nerve-released transmitter would be ATP rather than meATP, and a combined action of P2X₂ and P2X₃ subunits might result in an intermediate sensitivity to TNP-ATP. For the interpretation of such experiments it would also be useful to know whether these TNP analogs have blocking action at other receptors types, including the P2Y receptors.