Evidence for two conductive pathways in P2X7 receptor: differences in modulation and selectivity

Authors


Address correspondence and reprint requests to Dr Mario Nobile, Institute of Biophysics, CNR, Via De Marini, 6 – 16149 Genoa, Italy. E-mail: nobile@ge.ibf.cnr.it

Abstract

J. Neurochem. (2010) 113, 796–806.

Abstract

The P2X7 receptor (P2X7R) is an ATP-gated cation channel whose biophysical properties remain to be unravelled unequivocally. Its activity is modulated by divalent cations and organic messengers such as arachidonic acid (AA). In this study, we analysed the differential modulation of magnesium (Mg2+) and AA on P2X7R by measuring whole-cell currents and intracellular Ca2+ ([Ca2+]i) and Na+ ([Na+]i) dynamics in HEK293 cells stably expressing full-length P2X7R and in cells endowed with the P2X7R variant lacking the entire C-terminus tail (trP2X7R), which is thought to control the pore activation. AA induced a robust potentiation of the P2X7R- and trP2X7R-mediated [Ca2+]i rise but did not affect the ionic currents in both conditions. Extracellular Mg2+ reduced the P2X7R- and trP2X7R-mediated [Ca2+]i rise in a dose-dependent manner through a competitive mechanism. The modulation of the magnitude of the P2X7R-mediated ionic current and [Na+]i rise were strongly dependent on Mg2+ concentration but occurred in a non-competitive manner. In contrast, in cells expressing the trP2X7R, the small ionic currents and [Na+]i signals were totally insensitive to Mg2+. Collectively, these results support the tenet of a functional structure of P2X7R possessing at least two distinct conductive pathways one for Ca2+ and another for monovalent ions, with the latter which depends on the presence of the receptor C-terminus.

Abbreviations used:
[Ca2+]i

free cytosolic Ca2+ concentration

AA

arachidonic acid

BBG

brilliant blue G

BzATP

3′-O-(4-benzoyl)benzoyl-ATP

NMDG+

N-methyl-d-glucamine

P2X7R

P2X7 receptor

The P2X7 receptor (P2X7R) is a ligand-gated ion channel and a member of the P2X family of purinergic receptors (P2X1–P2X7) (Surprenant et al. 1996). The gene for the P2X7R encodes a 595-amino acid subunit that is structurally divergent from the other P2X subtypes, because it has a long (∼200 amino acids) intracellular C-terminal tail. Chemical cross-linking and blue native polyacrylamide gel electrophoresis analysis of recombinant receptors have revealed a trimeric structure of P2X ion channels (Nicke et al. 1998), which has been confirmed by atomic force microscopy (Barrera et al. 2005) and by crystallography (Kawate et al. 2009). The P2X7R is expressed in mammalian immune cells and epithelia with immunomodulatory function (Surprenant et al. 1996; Di Virgilio et al. 2001; Ferrari et al. 2006; Di Virgilio 2007). In the past few years, the role of this receptor in the CNS has also received considerable attention (Duan and Neary 2006; Sperlágh et al. 2006). There is evidence of its expression in neuronal, astroglial cells and microglia, where it is involved in different functions (Franke et al. 2001; Duan et al. 2003; Nobile et al. 2003; Franke et al. 2007; Marcoli et al. 2008). Moreover, P2X7R has been proposed to be a potential therapeutic target site in CNS disorders, such as ischaemia–reperfusion injury, Alzheimer’s disease, spinal cord injury and neuropathic pain (Wang et al. 2004; McLarnon et al. 2006; Milius et al. 2008; Wang et al. 2009).

A unique property of the P2X7R is that it exhibits two agonist-activated conductance modes. Upon activation with ATP or the potent agonist 3′-O-(4-benzoyl)benzoyl-ATP (BzATP), the receptor functions as a non-selective cation channel, permeant to small cations, such as Na+, K+, and Ca2+ but this activation mode is dependent on extracellular divalent cations (Virginio et al. 1997; Ding and Sachs 2000; Ma et al. 2006; Sperlágh et al. 2006; Jiang 2009). Upon repeated or prolonged application of agonist, the presence of extracellular divalent cations and H+ protons, or depending on the receptor density (Virginio et al. 1997; Chessell et al. 1998; Michel et al. 1999; Virginio et al. 1999; Yan et al. 2008), the P2X7R becomes permeable to larger molecules like ethidium bromide, N-methyl-d-glucamine or neurotransmitters such as glutamate and ATP (Marcoli et al. 2008; Hamilton et al. 2008). Several mechanisms have been proposed to explain this phenomenon (Egan et al. 2006). However, pore formation can be observed in some cell types expressing P2X7R (Smart et al. 2003) but not in other cell types, including Xenopus oocytes, which exhibit only the cation permeable mode (Petrou et al. 1997). These observations support the notion that other molecular components are necessary for the transition in the pore mode. Recently, evidence were provided that Pannexin-1 (panx1), a hemichannel protein, is the large pore associated to P2X7R (Pelegrin and Surprenant 2006; Locovei et al. 2007; Pelegrin and Surprenant 2007). In contrast, other studies reported that the sustained rise in P2X7R current and permeability was caused by dilatation of the P2X7R itself (Chaumont and Khakh 2008; Yan et al. 2008). Other data indicated that the uptake pathways activated by P2X7R may be two: one for cationic and another for anionic dyes (Cankurtaran-Sayar et al. 2009; Schachter et al. 2008). Finally, in cultured astrocytes and in HEK293 cells stably transfected with the P2X7R, the polyunsaturated fatty acid arachidonic acid (AA) induced a strong potentiation, independent of AA metabolism, of the P2X7R-mediated [Ca2+]i rise but not of the P2X7R-mediated cation currents, thereby suggesting alternative pathways of permeability (Alloisio et al. 2006).

In the attempt to unravel a plausible molecular frame that can explain the manifold characteristics of ATP-gated P2X7R here, we have analysed the modulation of some of the functional properties of the P2X7R by extracellular AA and Mg2+ on [Ca2+]i and [Na+]i signalling and ionic currents comparing the behaviours of the recombinant P2X7R and the trP2X7R heterologously expressed in HEK293 cells. Our results indicate that the effect of AA and Mg2+ on P2X7R is mediated by one of the at least two different pathways that constitute the P2X7R.

Experimental procedures

Cell cultures

Cultures of the human embryonic kidney cell line HEK293 were maintained in Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F-12 Ham supplemented with 10% foetal bovine serum, and gentamicin/glutamine (5 mg/mL and 200 mM, respectively). Culture flasks were maintained in a humidified incubator at 37°C in a 5% CO2 enriched atmosphere. At confluence cells were enzymatically dispersed (trypsin-EDTA, 0.5–0.2 g) and were re-plated in 20-mm glass coverslips at a density of 5 × 103 per coverslip. The plasmid pcDNA3 containing the full length rat P2X7-GFP cDNA and HEK293 cells stably transfected with 1–375 truncated rat P2X7R were kindly provided by Prof. Di Virgilio F. (Department of Experimental and Diagnostic Medicine, University of Ferrara, Italy).

Stable transfection of HEK293 cells with full length P2X7R

One day before transfection, HEK293 cells were replated on plastic dishes (35-mm diameter) in antibiotics-free growth medium. DNA plasmid of P2X7R was transfected using cationic liposomes (Lipofectamine 2000) according to the manufacturer’s instruction. At ∼16 h post-transfection, the medium was changed with one supplemented with 1.5 g/L of G418 sulfate. Procedures for stable transfection were as previously described (Alloisio et al. 2006).

Electrophysiology

The electrophysiological studies were performed at room temperature (20–22°C) using the whole-cell configuration of the patch-clamp technique. Patch pipettes were pulled from borosilicate glass capillaries (Clark Electromedical Instruments, Reading, UK) and heat-polished to obtain a resistance of 2–4 MΩ when filled with the aforesaid solutions. The external standard solution was composed of (mM): 135 NaCl, 5.4 KCl, 1 CaCl2, 1 MgCl2, 5 HEPES, 10 glucose, adjusted to pH 7.3 with NaOH and to osmolarity ∼310 mOsm with mannitol. Ca2+-free standard solution was similar but Ca2+ was replaced with 5 mM EGTA. The standard pipette solution contained (mM): 135 NMDGCl, 1 MgCl2, 5 Hepes, 5 EGTA, 5 glucose, adjusted with NMDGOH to pH 7.3. When we used external solutions with different ionic compositions (Na+ replaced with NMDG+ as the principal flowing ion), salts were equimolary replaced. Osmolarity was measured with a vapour pressure osmometer (Wescor 5500, Delcon s.r.l., Milan, Italy). The different saline solutions containing the pharmacological agents were applied with a gravity-driven, local perfusion system at a flow rate of ∼300 μL/min positioned within ∼100 μm of the recorded cell.

Membrane currents were recorded using an AXOPATCH 200A integrating patch clamp amplifier (Axon Instruments, Foster City, CA, USA), and were low-pass filtered at 1 kHz before acquisition. Both voltage stimulation and data acquisition were obtained using a 12-bit interface (Axon Instruments) and a microcomputer equipped with pClamp 6 software (Axon Instruments). Experiments were performed at a holding potential of −50 mV or with a 2-s-long voltage ramp pulses protocol from −50 to 50 mV.

Single cell [Ca2+]i microfluorimetry

Intracellular calcium measurements were performed by using the fluorescent Ca2+ indicator fura-2AM or fura-FF AM. Cells were loaded with 5 μM fura-2AM or fura-FF AM dissolved in extracellular solution for 45 min at 37°C. Pluronic acid (0.1%) was added to improve the dyes uptake. The extracellular bath solutions were the same used for electrophysiological experiments. The microperfusion chamber containing the cell coverslip was placed on the stage of an inverted fluorescence microscope Nikon TE200 (Nikon, Tokyo, Japan) equipped with a dual excitation fluorometric calcium imaging system (Hamamatsu, Sunayama-Cho, Japan). Low-density seeded cells were continuously perfused at a rate of about 2.5 mL/min. Emission florescence of selected cells was passed through a narrow-band filter and acquired with a digital CCD camera (Hamamatsu C4742-95-12ER). Monochromator settings, chopper frequency and complete data acquisition were controlled by dedicated software (Aquacosmos/Ratio U7501-01, Hamamatsu). The sampling rate was 0.25 or 0.5 Hz. Fura-2AM or fura-FF AM loaded cells were excited at 340 and 380 nm and emission fluorescence measured at 510 nm. The fluorescence ratio F340/F380 was used to monitor [Ca2+]i changes. In some experiments, [Ca2+]i was calculated according to Grynkiewicz et al. (1985), using a KD of 140 nmol/L for the Ca2+/fura-2 complex.

For measurement of intracellular Na+ concentration ([Na+]i), cells were loaded for 75 min at 37°C with the fluorescent dye SBFI-AM (15 μM) dissolved in extracellular solution complemented with pluronic acid (0.1%). SBFI-AM loaded cells were excited at 345 and 385 nm and emission fluorescence measured at 510 nm. The fluorescence ratio F345/F385 was used to monitor [Na+]i changes.

Solutions and chemicals

All the salts and chemicals used for the preparation of cultures, the electrophysiological measurements and the fluorimetric determinations of the intracellular Ca2+ concentrations were obtained from Sigma-Aldrich S.r.L. (Milan, Italy).

Statistics

All data are given as mean ± SEM. The statistical significance of differences between mean values was assessed using Student’s t-test. Differences were regarded as statistically significant for < 0.05.

Results

AA effect on BzATP-elicited ions influx in HEK293 cells stably transfected with recombinant full-length and C-terminal truncated rat P2X7R s

We previously showed that cultured rat cortical type-1 astrocytes and HEK293 cells transfected with full-length rat P2X7R challenged with micromolar concentrations of the relatively selective agonist BzATP exhibited a [Ca2+]i signal elicited by Ca2+ influx that was potentiated by co-application of AA (Alloisio et al. 2006). We also reported that the whole-cell cationic currents mediated by P2X7R remained of the same magnitude when BzATP was co-applied with AA. This differential behaviour may be exploited to gain insights on the molecular mechanisms through which P2X7R can conduct different ions depending on its gating mode. To address this issue in this study, we compared the properties in terms of sensitivity to AA and extracellular Mg2+ of the recombinant rat P2X7R and the P2X7 (1-375) in which the C-terminus has been deleted (trP2X7R).

Initial experiments were devoted to define some properties of control HEK293 cells in P2X7R- and trP2X7R-expressing HEK293 cells. In agreement with studies on P2Y1,2,4 receptor subtypes expression (Fischer et al. 2005), untransfected HEK293 challenged with 10 μM ATP displayed a transient [Ca2+]i response whereas, exposure to 10 μM BzATP did not produce any [Ca2+]i signal (Fig. 1a; n = 32). In contrast, in HEK293 cells transfected with P2X7R, BzATP promoted a large, dose-dependent [Ca2+]i rise, that remained stable throughout BzATP application (Fig. 1b; n = 120). Similar results were also obtained in HEK293 cells transfected with trP2X7R. This P2X7R and trP2X7R [Ca2+]i response was dependent on extracellular Ca2+, indicating that the BzATP action depended on Ca2+influx (Fig. 1c and d; n = 83). Furthermore, after depletion of the intracellular Ca2+stores by extracellular application of the endoplasmic reticulum Ca2+-ATPase inhibitor cyclopiazonic acid (10 μM) in a Ca2+-free solution, a sustained [Ca2+]i increase mediated by store-dependent capacitative Ca2+entry pathway was observed upon re-addition of extracellular Ca2+in both P2X7R and trP2X7R expressing cells. Application of 5 μM BzATP in the presence of cyclopiazonic acid induced a further augment of the steady-state [Ca2+]i response, a result which supports the notion that intracellular stores are not involved in the P2X7R-mediated [Ca2+]i elevation (Fig. 1e and f; n = 62; < 0.01). Figure 2(a) shows that whereas application of the threshold BzATP concentration (3 μM) promoted a small but significant [Ca2+]i rise co-application of BzATP and AA (5 μM) elicited a robust, sustained [Ca2+]i elevation, which decayed to the AA-induced [Ca2+]i level upon BzATP removal. In P2X7R-HEK293 cells, clamped at a holding potential of −50 mV and using standard internal and external solutions, rapidly activating, non-inactivating whole-cell inward currents were elicited in response to 3 μM BzATP (Fig. 2b). Upon repeated (every 2 min, 20-s-long) applications of 3 μM BzATP, the evoked currents increased in magnitude and reached their maximal levels after about four applications (n = 24). Following maximal activation, co-application of BzATP (3 μM) and AA (5 μM) elicited currents that did not increase further. In trP2X7R-HEK293 cells, application of 3 μM BzATP also promoted a small [Ca2+]i rise, which was strongly enhanced by co-application of BzATP and AA (5 μM) (Fig. 2c). Likewise P2X7R sustained currents, the small ionic currents induced by 30 μM BzATP in trP2X7R-HEK293 cells were unaffected by co-application of BzATP and AA (Fig. 2d). The results, summarised in Fig. 2e, show that AA potentiated the BzATP-induced [Ca2+]i signal caused by Ca2+ entry but did not alter the Na+ influx measured as whole-cell currents in truncated and full-length P2X7R. All the BzATP-induced signals were inhibited by the specific P2X7R antagonist brilliant blue G (100 nM; BBG).

Figure 1.

 Purinoceptor agonists differently regulate [Ca2+]i responses in HEK293 cells. (a) Representative control HEK293 cell fluorescence ratio signal (F340/F380) depicting the transient [Ca2+]i rise evoked by 10 μM ATP and no response to 10 μM BzATP (n = 32). (b) In HEK293 cells stably transfected with rat P2X7 receptor (P2X7R), 1 μM BzATP did not elicit [Ca2+]i increase, but a subsequent application of 10 μM BzATP caused a robust [Ca2+]i rise that rapidly returned to basal level upon BzATP removal (n = 120). (c, d) In P2X7R and trP2X7R transfected cells, in the absence of extracellular Ca2+, only 10 μM ATP induced a [Ca2+]i response similar to that produced in non-transfected cells (n = 83). (e, f) Typical experiments showing the effect of BzATP (5 μM) on capacitative Ca2+entry (CCE) elicited by exposure to cyclopiazonic acid (CPA) in a Ca2+-free solution to deplete the intracellular stores of HEK293 cells expressing P2X7R and trP2X7R, respectively (n = 62). Horizontal bars indicate the duration of the agonist application.

Figure 2.

 Arachidonic acid (AA) effect on BzATP-induced [Ca2+]i signals and whole-cell currents in P2X7R and trP2X7R. (a, c) Representative experiments of co-application of AA (5 μM) and BzATP (3 μM) at P2X7R (a) and trP2X7R (c), showing a sustained [Ca2+]i signal larger than that evoked by individual applications. (b, d) Whole-cell current traces obtained during application of BzATP alone or BzATP plus AA (5 μM) on HEK293 cells expressing P2X7R or trP2X7R. Agonist concentration was 3 μM at P2X7R (b) and 30 μM at trP2X7R (d). No significant differences in current amplitude could be observed between individual (BzATP) or co-applications (BzATP plus AA) in both P2X7R and trP2X7R. (e) Histogram illustrating the effect evoked by individual applications of BzATP or AA and the synergistic action of BzATP plus AA on [Ca2+]i increase (left) and current amplitude (right). Each bar is the mean ± SEM of at least 33 cells. **< 0.01. The bars graph also shows the significant block effect of 100 nM BBG in all conditions. *< 0.05.

Mg2+ effect on BzATP-elicited ions influx in HEK293 cells stably transfected with recombinant full-length and C-terminal truncated rat P2X7Rs

There is clear indication that the P2X7R is potently inhibited by divalent cations such as Mg2+ (Acuña-Castillo et al. 2007). Thus, next we addressed the question whether extracellular Mg2+ also blocks BzATP-induced Ca2+ influx. The data showed that P2X7R-HEK293 cells exposed to micromolar concentrations of BzATP elicited a [Ca2+]i signal that increased markedly when the standard solution (1 mM Mg2+) was replaced with a nominally Mg2+-free saline (Fig. 3a and b). To rule out possible artefacts or the P2X7R switch from the channel to pore conformation, we used BzATP concentrations no higher than 10 μM and we monitored the parallel changes in F340 and F380 throughout the experiments (Fig. 3c and d). Curiously, in Mg2+-free saline BzATP applications longer that 120 s, induced a [Ca2+]i increase, which reached a steady maximal level that persisted even after agonist wash-out in extracellular Ca2+-free solution (Fig. 3e and f). A plausible explanation is that the [Ca2+]i homeostatic mechanisms of the cells are altered under these conditions as a result of prolonged agonist stimulation. The quantitative analysis illustrates that, in Mg2+-free, the [Ca2+]i elevation saturated at ∼3 μM BzATP whereas in 1 mM Mg2+the [Ca2+]i levelled off at ∼10 μM, with an EC50 at 1.2 μM and a Hill coefficient of 3.8 in Mg2+-free solution and an EC50 of 5.0 μM and a Hill coefficient of 4.5 in 1 mM Mg2+(Fig. 3g). Interestingly, the two EC50 values were significantly different whereas the maximal current amplitudes were similar, indicating a competitive mechanism for Mg2+action. Figure 3(h) shows the concentration–response curve for Mg2+inhibition on 3 μM BzATP-elicited [Ca2+]i increase depicting an IC50 of 0.2 mM and the Hill coefficient of 0.9. As it was suggested the inadequacy of the sensitive dye fura-2 to measure high [Ca2+]i elevation in P2X7R-expressing cells (Koshimizu et al. 2000), we repeated these experiments with the less sensitive dye fura-FF. The dose-response curves obtained by using fura-FF indicated a lower maximal Δ ratio but a similar competitive mechanism of the Mg2+ action with an EC50 at 4.0 μM and a Hill coefficient of 2.8 in Mg2+-free solution and an EC50 of 22.8 μM and a Hill coefficient of 2.5 in 1 mM Mg2+ (Fig. 3j). In HEK293 transfected with trP2X7 the maximal [Ca2+]i elevation was diminished by ∼50% compared with the P2X7R but was strongly increased in Mg2+-free solution. The BzATP-induced responses were dose dependent with a EC50 value of 7.9 μM and a Hill coefficient of 1.2 in Mg2+-free solution and a EC50 value of 30 μM and a Hill coefficient of 1.7 in 1 mM Mg2+ (Fig. 4a–c). The concentration-response curve for Mg2+ inhibition on BzATP-elicited [Ca2+]i increases displayed a IC50 value of 0.2 mM, which is similar to the value obtained for P2X7R, and a Hill coefficient of 0.8 (Fig. 4d). Interestingly, the dose-dependence curve for BzATP indicates that the Mg2+-induced reduction of [Ca2+]i rise was because of a competitive block as shown for the full-length P2X7R.

Figure 3.

 Mg2+ dependence of BzATP-evoked [Ca2+]i increase in cells transfected with rat P2X7R. (a, b) Representative traces of single cell fluorescence ratio depicting the [Ca2+]i increase in 1 mM Mg2+ and Mg2+-free extracellular saline. The amplitude of [Ca2+]i response elicited by 3 μM BzATP was strongly dependent on the presence of extracellular Mg2+. (c) Representative experiment performed in 1 mM Mg2+ and Mg2+-free extracellular solutions on the same cell (n = 22) and (d) relative fluorescence intensities [FI; λex = 340 nm (F340), λex = 380 nm (F380)]. (e) In Mg2+-free experiments, a BzATP-evoked [Ca2+]i saturation level persisted even after BzATP and extracellular Ca2+ removal; conversely, in extracellular solution with 1 mM Mg2+ the complete or partially recovery of the Ca2+ signalling was obtained (data not shown). (f) The fluorescence intensities under 340 and 380 nm excitation indicated the experimental quality. (g) Quantitative analysis of [Ca2+]i elevations above basal levels elicited by BzATP (0.1–10 μM) in Mg2+-free saline and in control conditions. The fit by the Hill equation yielded an EC50 of 1.2 and 5.0 μM in 0 Mg2+ and in 1 mM Mg2+, respectively. Each point is the mean ± SEM of at least 46 cells. Notably, the reduction of EC50 value in absence of extracellular Mg2+ and the dose–response curves shape were indicative of a competitive block. (h) Concentration–response curves for Mg2+ inhibition of BzATP-evoked [Ca2+]i rise in P2X7R-expressing cells (IC50 of 0.2 mM). Mean ± SEM of at least 34 cells. < 0.05. (j) Quantitative analysis of [Ca2+]i elevations by using the low affinity fluorescent Ca2+ indicator fura-FF AM. Each point is the mean ± SEM of at least 24 cells.

Figure 4.

 BzATP-evoked [Ca2+]i increase in single HEK293 cells expressing trP2X7R lacking the full C-terminal. (a, b) [Ca2+]i responses recorded in standard and Mg2+-free saline (10 μM BzATP). The maximal cell [Ca2+]i signal was about twofold lower than full-length P2X7R-mediated [Ca2+]i response but had a similarly dependence to extracellular Mg2+ concentration. (c) Quantitative analysis of [Ca2+]i elevations above basal levels elicited by BzATP (0.1–200 μM) in control and in Mg2+-free recording solution. The fit by the Hill equation yielded an EC50 of 7.9 and 30.0 μM in 0 Mg2+ and in 1 mM Mg2+, respectively. Mean ± SEM of at least 42 cells. Notably, the EC50 values and the dose–response curves shape were indicative of a competitive block as P2X7R. (d) Concentration–response curves for Mg2+ inhibition of BzATP-evoked [Ca2+]i rise in trP2X7R-expressing cells. Similarly to P2X7R the IC50 value was 0.2 mM. Mean ± SEM of at least 26 cells. < 0.05.

Next we measured whole-cell currents in 1 mM Mg2+ or Mg2+-free extracellular salines. The amplitude of the inward currents evoked by 10 μM BzATP in the presence of 1 mM Mg2+ was twofold lower than that in Mg2+-free saline (Fig. 5a and b). Under these ionic conditions, the mean value of the reversal potential of BzATP-induced currents determined by a ramp protocol was 15.4 ± 0.4 mV (n = 8), indicating the predominance of cation influx. In cells challenged with BzATP in extracellular saline in which Na+ was replaced by NMDG+ currents were smaller and reversed at 0 ± 0.5 mV (n = 9). In standard solutions, the amplitude of the currents depended on BzATP concentration with an EC50 at 6.1 μM and a Hill coefficient of 3.5 in Mg2+-free solution and an EC50 of 7.2 μM whereas in 1 mM Mg2+ the Hill coefficient was 4.9 (Fig. 5c). The results indicate that the values of the two EC50 were not significantly different but the maximal current amplitude in 1 mM Mg2+ was strongly reduced, denoting a non-competitive Mg2+ block of the BzATP-activated channel. The Mg2+ inhibition was dose dependent (Fig. 5d) with an IC50 of 0.7 mM and the Hill coefficient of 0.9.

Figure 5.

 Whole-cell currents analysis of P2X7R stably expressed in HEK293 cells. (a, b) 10 μM BzATP-evoked current, recorded at −50 mV holding potential, in 1 mM Mg2+ and in nominally Mg2+-free recording solutions. The absence of Mg2+ increased about twofold the whole-cell currents. (c) BzATP concentration–response curve for membrane currents obtained in experiments as illustrated in (a) and (b). The fit by the Hill equation yielded an EC50 of 6.1 and 7.2 μM in the absence and in the presence of extracellular Mg2+, respectively. Notably, the similarity of EC50 values and the potentiation of maximal current amplitudes are indicative of a non-competitive blocking mechanism. Data are the average of several current values and each point is the mean of at least 10 cells; error bars indicate SEM. (d) Concentration–response curves for Mg2+ inhibition of currents in P2X7R-expressing cells. A concentration dependence was observable with a IC50 of 0.7 mM. Each bar is the mean of at least nine cells.

Finally, we verified the role of the C-terminal domain of P2X7R on the properties described by performing a series of experiments in trP2X7R-HEK293. The finding that the trP2X7R, even upon challenge with a high concentration of BzATP (100 μM) showed smaller currents (∼100-fold reduction; Fig. 6a) compared with P2X7R was in line with previous report (Smart et al. 2003). The reversal potential of BzATP-induced currents under these ionic conditions was 14.5 ± 0.7 mV (n = 7). The currents through trP2X7R retained their sensitivity to BBG (80 ± 8% reduction; n = 9; data not shown) but, in contrast to the [Ca2+]i signals evoked in trP2X7R, they did not show any Mg2+ dependence (Fig. 6a and b). Moreover, at variance with P2X7R, in symmetrical NMDG+ saline no currents were elicited in response to 100 μM BzATP (Fig. 6c; > 0.05).

Figure 6.

 Characterisation of the BzATP-evoked currents recorded from HEK293 cells stably transfected with trP2X7R. (a, b) Membrane currents recorded in standard and in nominally Mg2+-free recording solution in presence of 100 μM BzATP. Notably, the cell current amplitude was about 20 times lower than that recorded in full-length P2X7R-transfected cells and was independent from extracellular Mg2+. (c) Summary of results obtained in extracellular standard saline and in nominally Mg2+-free solution. When the extracellular Na+ was replaced with NMDG+ as the principal flowing ion, no detectable currents were obtained both in 0 Mg2+ and in the presence of 1 mM Mg2+. Each bar is the mean ± SEM of at least 16 cells.

Because of the different results obtained using imaging and electrophysiological assays, we compared [Na+]i versus [Ca2+]i ratio-imaging measurements. P2X7R-HEK293 cells loaded with the Na+ indicator SBFI-AM (Harootunian et al. 1989) and exposed to micromolar concentrations of BzATP elicited a [Na+]i signal that increased markedly when the standard solution was replaced with a nominally Mg2+-free saline. The quantitative analysis illustrates that, in standard solution, the [Na+]i elevation depended on BzATP concentration with an EC50 at 2.4 μM and a Hill coefficient of 2.3 in Mg2+-free solution and an EC50 at 4.0 μM and a Hill coefficient of 2.6 in 1 mM Mg2+ saline (Fig. 7a and c). Differently, the trP2X7R even upon challenge with a high concentration of BzATP (100 μM) showed an insignificant [Na+]i rise not sensitive to extracellular Mg2+ (Fig. 7b; n = 54; > 0.05). Altogether, the [Na+]i ratio-imaging results were consistent with those obtained measuring whole-cell currents.

Figure 7.

 Mg2+ dependence of BzATP-evoked [Na+]i increase in cells transfected with rat P2X7R. (a) Representative traces of single cell fluorescence ratio depicting the [Na+]i increase in 1 mM Mg2+ and Mg2+-free extracellular saline. The amplitude of [Na+]i response elicited by 10 μM BzATP was significantly dependent on the presence of extracellular Mg2+. (b) BzATP-evoked [Na+]i responses from HEK293 cells stably transfected with trP2X7R. [Na+]i signals in standard and in nominally Mg2+-free recording solution in presence of 100 μM BzATP. Notably, the [Na+]i signal was very low and was independent from extracellular Mg2+. (c) Quantitative analysis of [Na+]i elevations above basal levels elicited by BzATP (0.1–100 μM) in control and in Mg2+-free recording solution. The fit by the Hill equation yielded an EC50 of 2.4 and 4.0 μM in 0 Mg2+and in 1 mM Mg2+, respectively. Mean ± SEM of at least 34 cells.

Discussion

The biophysical properties of the purinergic receptor P2X7R are still largely elusive because the channel displays unique architecture and unusual characteristics. In particular, there is no clear information on the mechanism whereby it conducts ions and large molecules across the membrane or how to explain these properties in terms of molecular structure. We previously demonstrated a synergism of BzATP and AA on [Ca2+]i signal of native and recombinant P2X7R which, however, was not detectable when measuring whole-cell currents (Alloisio et al. 2006). Here, we compared the behaviours of cultured HEK293 cells stably transfected with recombinant rat P2X7R and HEK293 cells expressing truncated rat P2X7R (1–375; trP2X7R). As previously shown by Smart et al. (2003), the trP2X7R showed low current amplitudes even upon challenge with a relatively high concentration of the potent synthetic P2X7R agonist BzATP, demonstrating that the C-terminal domain influences the channel dilatation or the pore mode formation. As for full-length P2X7R, the presence of extracellular AA did not produce any increase in current amplitude. The BzATP-activated [Ca2+]i rise in cells expressing trP2X7R was lower than that generated in P2X7R but it was synergistically increased by the presence of AA. The same differences were observed concerning the modulatory action of extracellular Mg2+.

The BzATP-evoked whole-cell currents recorded in HEK293 cells transfected with rat P2X7R confirmed previous observations about an increase in amplitude upon repeated agonist applications, inhibition by BBG (Chessell et al. 1998; Pelegrin and Surprenant 2006; Yan et al. 2008). Although the modulation by divalent cations of various P2X receptor subtypes has been already observed in previous studies (Virginio et al. 1997; Ding and Sachs 2000; Ma et al. 2006; Sperlágh et al. 2006; Jiang 2009), here we provide evidence of two distinct Mg2+ effects on ionic fluxes: (i) the ionic current through P2X7R (Na+ or NMDG+ as the principal flowing ion) and [Na+]i signals were reduced by Mg2+ by a mechanism indicative of a non-competitive blocking mechanism; (ii) the BzATP-elicited Ca2+ influx was potently inhibited by Mg2+, but with a dose dependency indicative of a competitive block. The Mg2+ modulation of P2X7R was also investigated using trP2X7R. BBG affected the trP2X7R currents as it did with the P2X7R. However, at variance with P2X7R, no currents were elicited in NMDG+ recording solution and the Mg2+-dependency was lost. Moreover, [Ca2+]i signals were still detected, albeit they were lower than that those mediated by P2X7R indicating a reduction in receptor function (Becker et al. 2008). Previous experiments with fura-2 revealing the inconsistency to report high [Ca2+]i elevation in P2X7R expressing cells (Koshimizu et al. 2000), prompted us to use the less sensitive fura-FF dye. However, the results were similar showing a competitive Mg2+ modulation even if the Hill and EC50 values were slightly different. In fura-2 experiments, the Hill coefficients and EC50 values obtained for P2X7R and trP2X7R were in good agreement with the results of Klapperstück et al. (2001) and Becker et al. (2008). They reported evidence of two distinct types of ATP activation sites in the human P2X7R and the functional loss of a ‘low-affinity’ site in truncated P2X7R (1–436). Both the N- and C-terminal domains appeared to be important determinants for P2X7R properties regulating the low-affinity ATP sensitivity.

Given the striking differences in behaviour between P2X7R- and trP2X7R-mediated current, [Na+]i and [Ca2+]i responses, we propose a new functional and structural model able to fit P2X7R activity. The P2X7R could be constituted of a complex forming two interacting conductive pathways: one with the ATP binding sites, the sensitivity to extracellular Mg2+ and controlling the permeability to Ca2+ (pathway A; P2X7R-A); a second with activation and divalent sensitivity depending on A and permeable to cations (pathway B; P2X7R-B). This model can explain our results: (i) We previously described a potentiating effect of AA on BzATP-induced P2X7R-mediated [Ca2+]i signal, which was not observable on membrane currents (Alloisio et al. 2006). We did not find any plausible explanation, but those results can be fit with the proposed model, in which AA interacts with the P2X7R-A. (ii) The P2X7R-mediated Na+ or NMDG+ fluxes, but not Ca2+ influx, were strongly diminished when the C-terminus was truncated. These data are consistent with the existence of a macromolecular complex, in which the C-terminal domain is not the physical link to keep the receptor assembled, but can modulate the interaction between the P2X7R-A and the P2X7R-B. (iii) The differential modulation by Mg2+, which is lost in BzATP-induced trP2X7R currents but evident in [Ca2+]i responses. (iv) The dose–response curves of BzATP-induced [Ca2+]i signals in the absence and in the presence of Mg2+ are typical of a competitive block as depicted by a right shift of the dose–response curve for the agonist without a change in the maximal ratio value. These results are compatible with the suggestion of Virginio et al. (1997) on divalent cations acting as allosteric modulators to alter the affinity of ATP binding to the P2X7R-A. In contrast, the Mg2+ inhibition of P2X7R currents is indicative of a different, indirect non-competitive mechanism of action on the P2X7R-B (Li et al. 1997).

The model which is represented schematically in Fig. 8 is the simplest model that can explain our results. It defines that at least two separate pathways, distinct from the pore mode, constitute the P2X7R. As we do not consider the apoptotic pore but only the physiological behaviour of the P2X7R, it is likely that this functional model shall be refined. However, our results illustrate the high degree of plasticity of this receptor and may explain some controversies on P2X7R functional properties, thereby providing useful information for the search of the complete molecular structure of P2X7R.

Figure 8.

 Schematic representation of the two distinct pathways, called P2X7R-A and P2X7R-B, proposed to be present in P2X7R. One of them is permeable to Ca2+and the other to monovalent and large cations. (a) The data on P2X7R are consistent with the existence of a macromolecular complex, in which the C-terminal domain is not the physical link to keep the receptor assembled, but can modulate the interaction between the P2X7R-A and the P2X7R-B. (b) When the P2X7R was investigated using trP2X7R, the characteristics of the P2X7R-A pathway were unchanged whereas the P2X7R-B was completely inefficient.

Acknowledgements

The financial support of Ca.Ri.Ge Foundation, Mariani Foundation (Grant n. R-07-66) and Linear S.r.l are gratefully acknowledged.

Ancillary