The P2X7 receptor as a route for non-exocytotic glutamate release: dependence on the carboxyl tail

Authors


Address correspondence and reprint requests to Manuela Marcoli, Department of Pharmacy, University of Genova, Viale Cembrano 4, 16148 Genova, Italy. E-mail: marcoli@pharmatox.unige.it

Abstract

P2X7 receptors trigger Ca2+-dependent exocytotic glutamate release, but also function as a route for non-exocytotic glutamate release from neurons or astrocytes. To gain an insight into the mechanisms involving the P2X7 receptor as a direct pathway for glutamate release, we compared the behavior of a full-length rat P2X7 receptor, a truncated rat P2X7 receptor in which the carboxyl tail had been deleted, a rat P2X7 receptor with the 18-amino acid cysteine-rich motif of the carboxyl tail deleted, and a rat P2X2 receptor, all of which are expressed in HEK293 cells. We found that the P2X7 receptor function as a route for glutamate release was antagonized in a non-competitive way by extracellular Mg2+, did not require the recruitment of pore-forming molecules, and was dependent on the carboxyl tail. Indeed, the truncated P2X7 receptor and the P2X7 receptor with the deleted cysteine-rich motif both lost their function as a pathway for glutamate release, while still evoking intracellular Ca2+ elevation. No glutamate efflux was observed through the P2X2 receptor. Notably, HEK293 cells (lacking the machinery for Ca2+-dependent exocytosis), when transfected with P2X7 receptors, appear to be a suitable model for investigating the P2X7 receptor as a route for non-exocytotic glutamate efflux.

Abbreviations used
[Ca2+]i

intracellular Ca2+ concentration

A-438079

3-(5-(2,3-dichlorophenyl)-1H-tetrazol-1-yl)-methyl pyridine

BAPTA-AM

1,2-bis-(o-aminophenoxy)-ethane-N,N,-N',N'-tetraacetic acid tetraacetoxy methyl ester

BzATP

2'-3'-O-(benzoylbenzoyl) ATP

CBX

carbenoxolone

fura-2 AM

fura-2 acetoxymethylester

HEK293

human embryonic kidney 293

HEK293-rP2X2

HEK293 cells stably transfected with rat P2X2 receptor

HEK293-rP2X7

HEK293 cells stably transfected with full-length rat P2X7 receptor

HEK293-rP2X7tr

HEK293 cells stably transfected with truncated (1-375) rat P2X7 receptor

HEK293-rP2X7ΔCys-rich

HEK293 cells transfected with rat P2X7 receptor with the juxtamembrane cysteine-rich motif of the C-tail deleted

hP2X7

human P2X7

NMDG+

N-methyl-d-glucamine

panx1

pannexin-1

PBS

phosphate-buffered saline

rP2X7

rat P2X7

rP2X7tr

truncated (1-375) rat P2X7

rP2X7ΔCys-rich

rat P2X7 receptor with the juxtamembrane cysteine-rich motif of the C-tail deleted

The P2X7 receptor is a ligand-gated ion channel that, upon stimulation by extracellular ATP, can switch from a rapid-gating channel selective for small cations to more slowly developing ‘dilated pore’ conformations permeable to molecules up to 900 Da (North 2002). Widely expressed on cells of hematopoietic lineage, the receptor has also been described on astrocytes, microglia, oligodendrocytes, and Schwann cells, as well as on neurons in the central and peripheral nervous system (Ballerini et al. 1996; Atkinson et al. 2004; Duan and Neary 2006; Sperlagh et al. 2006; Burnstock 2007; Hamilton et al. 2008; Verkhrasky et al. 2009; Grygorowicz et al. 2010; Able et al. 2011).

The presence of P2X7 receptors on glutamatergic nerve terminals has repeatedly been reported in various brain regions (Deuchars et al. 2001; Lundy et al. 2002; Sperlagh et al. 2002; Miras-Portugal et al. 2003; Alloisio et al. 2008; Marcoli et al. 2008; Leon et al. 2008; see also Bennett et al. 2009). Entry of Ca2+ through the P2X7 receptors situated on the midbrain or cerebrocortical glutamatergic terminals (Lundy et al. 2002; Miras-Portugal et al. 2003; Alloisio et al. 2008; Leon et al. 2008) triggers Ca2+-dependent glutamate release (Leon et al. 2008; Marcoli et al. 2008), indicating that pre-synaptic P2X7 receptors can function as Ca2+ channels coupled with vesicular exocytotic neurotransmitter release. In addition, Ca2+-independent non-exocytotic glutamate release from cerebrocortical nerve terminals appears to occur through activated P2X7 receptors (Marcoli et al. 2008). Direct evidence that P2X7 receptors can function as a route for glutamate release has been obtained in cultured astrocytes prepared from the cortices of 1-day-old mice (Duan et al. 2003) or by measuring the efflux of [3H]d-aspartate in HEK293 cells (which lack the machinery for Ca2+-dependent vesicle exocytosis) stably transfected with full-length rat P2X7 receptors (Marcoli et al. 2008).

The P2X7 receptor is distinguishable from other P2X receptors by its long intracellular carboxyl tail (C-tail), which has multiple protein and lipid interaction motifs and a cysteine-rich 18-amino acid segment. The C-tail is thought to be involved in regulating receptor functions, including surface expression, signaling pathway activation, and permeability to large molecules. Indeed, deletion of the C-tail prevents the receptors from undergoing cytolytic pore formation and changes in their permeability to large organic cations or fluorescent dyes (Surprenant et al. 1996; Rassendren et al. 1997; Smart et al. 2003; Alloisio et al. 2010), while retaining the rapid-gated channel properties for small cations. The cysteine-rich 18-amino acid segment in the juxtamembrane region of the P2X7 receptor C-tail has been reported to contribute to specific P2X7 receptor properties and appears to be critical for the regulation of pore dilatation (Jiang et al. 2005; Roger et al. 2010). Removal of this cysteine-rich motif prevents the receptor from undergoing an increase in permeability to the large organic cation N-methyl-d-glucamine (NMDG+), while it does not affect the channel properties in normal extracellular sodium or the permeability of the P2X7 receptor to fluorescent dyes (Jiang et al. 2005). Conversely, the distal P2X7 C-tail is responsible for downstream pore formation, which allows the passage of fluorescent dyes such as the propidium dye, YO-PRO1, or ethidium bromide (Surprenant et al. 1996; Rassendren et al. 1997; Smart et al. 2003).

In this article, the role of the C-tail in the function of the P2X7 receptor as a route for non-exocytotic glutamate release was investigated on HEK293 cells expressing the full-length rat P2X7 (rP2X7) receptor, the truncated rat P2X7 (1-375) receptor – that is, lacking the C-tail (rP2X7tr) – and the rat P2X7 receptor with the cysteine-rich 18-amino acid motif of the C-tail deleted (rP2X7ΔCys-rich). Our study is the first to indicate that the function of the rP2X7 as a route for non-exocytotic glutamate release is dependent on the receptor C-tail and involves the juxtamembrane cysteine-rich motif of the tail.

Materials and methods

Cell cultures and transfection of HEK293 cells with P2X7 receptors

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% fetal 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. One day before stable transfection of HEK293 cells with full-length rP2X7 receptors, HEK293 cells were re-plated on plastic dishes (35 mm diameter) in an antibiotic-free growth medium. The DNA plasmid of P2X7 receptors was transfected by using cationic liposomes (Lipofectamine 2000) according to the manufacturer's instructions. At ~16 h post-transfection, the medium was replaced with one supplemented with 1.5 g/L of G418 sulfate. Procedures for stable transfection were carried out as previously described (Alloisio et al. 2006). The plasmid pcDNA3 containing the full-length rat P2X7-GFP cDNA and HEK293 cells stably transfected with 1-375 truncated rat P2X7 receptors (HEK293-rP2X7tr) were kindly provided by Prof. Francesco Di Virgilio (University of Ferrara, Italy). For transfection with rP2X7ΔCys-rich, HEK293 cells were maintained at 37°C, 5% CO2 in Dulbecco's Modified Eagle Medium/Ham's F12 (1 : 1 mixture) medium supplemented with 10% fetal bovine serum, penicillin (100 units/mL), and streptomycin (100 μg/mL). HEK293 cells were grown in 25 cm2 flasks. At the moment of the experiments, they were trypsinized and transfected with 20 μg of cDNA by electroporation (Equibio Easyject Plus) as described previously (Guida et al. 2008). The mutant rat P2X7 construct, which had undergone deletion of the 18-amino acid cysteine-rich segment (CCRSRVYPSCKCCEPCAV) immediately after the second membrane-spanning domain (rP2X7ΔCys-rich), was a generous gift from Prof. Alan North (University of Manchester, UK) and was generated as previously detailed (Jiang et al. 2005). HEK293 cells transfected with rP2X7ΔCys-rich receptors (HEK293-rP2X7ΔCys-rich) were examined for receptor expression after 48 h and functional experiments ([3H]d-aspartate efflux and intracellular calcium microfluorimetry) were performed in parallel. HEK293 stably transfected with rat P2X2 receptor (HEK293-rP2X2) were kindly provided by Prof. Annmarie Surprenant (University of Manchester, UK).

[3H]d-aspartate efflux from HEK293 cells

The presence of EAAT3 glutamate transporters on native HEK293 cells (Toki et al. 1998) was exploited in order to load native HEK293 cells and HEK293 cells expressing the full-length rat P2X7 (HEK293-rP2X7), HEK293-rP2X7tr, HEK293-rP2X7ΔCys-rich or HEK293-rP2X2 cells with [3H]d-aspartate. Briefly, native HEK293 cells and HEK293-rP2X7, HEK293-rP2X7tr, HEK293-rP2X7ΔCys-rich or HEK293-rP2X2 cells were incubated (30 min at 37°C) in the presence of [3H]d-aspartate (0.06 μM) in 5 mL of standard medium (mM: NaCl 135, KCl 2.4, MgSO4 1.2, KH2PO4 1.2, CaCl2 1.2, NaHCO3 5, HEPES 10 with glucose 10; pH 7.4). Cells were then transferred to parallel superfusion chambers at 37°C, stratified on Millipore filters, and superfused (0.5 mL/min) with standard medium (see Marcoli et al. 2004). After 31 min of superfusion, superfusate fractions were collected in 5 min samples (from B1 to B5) until the end of the experiment. After 40 min of superfusion, cells were exposed (5 min) to 2′-3′-O-(benzoylbenzoyl) ATP (BzATP) or ATP, and then reperfused with standard medium. The effect of 3-(5-(2,3-dichlorophenyl)-1H-tetrazol-1-yl)-methyl pyridine (A-438079), probenecid, or colchicine was evaluated by adding the drug 10 min before the agonist. The effect of lowering extracellular Mg2+ concentration, or of extracellular Ca2+ deprivation, was assessed in HEK293 cells superfused with a medium containing 0.01 mM MgSO4, or with Ca2+-free medium supplemented with EGTA 0.5 mM, starting 20 min before addition of the agonist. When indicated, cells were pre-incubated (30 min) in the presence of the membrane-permeable Ca2+ chelator 1,2-bis-(o-aminophenoxy)-ethane-N,N,-N',N'-tetraacetic acid tetraacetoxy methyl ester (BAPTA-AM; 50 μM). At the end of superfusion, the radioactivity of the filters on which cells were layered and of superfusate samples was determined by means of liquid scintillation counting.

The tritium fractional release and the drug-evoked tritium efflux were measured as reported in Marcoli et al. (2008); details are given in the Supporting Information.

Single-cell intracellular calcium microfluorimetry

Intracellular calcium was measured by using the fluorescent Ca2+ indicator fura-2 acetoxymethylester (fura-2 AM). Cells were loaded with 5 μM fura-2 AM dissolved in extracellular solution for 45 min at 37°C. To enhance dye solubility and uptake, 0.1 % pluronic acid T-127 was added. The extracellular bath solution was (mM): NaCl 135, KCl 5.4, CaCl2 1, MgCl2 1, HEPES 5, glucose 10, adjusted to pH 7.3 with NaOH and to osmolarity ~310 mOsm with mannitol. The microperfusion chamber containing the cell coverslip was placed on the stage of a Nikon TE200 (Nikon, Tokyo, Japan) inverted fluorescence microscope 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 fluorescence 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 Hz. Fura-2 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 changes in intracellular calcium concentration ([Ca2+]i). In some experiments, [Ca2+]i was calculated according to Grynkiewicz et al. (1985), using a KD of 140 nM for the Ca2+/fura-2 complex.

Immunocytochemistry

HEK293 cells grown for 2 days in vitro on poly-l-lysine-coated glass coverslips were fixed in 4% paraformaldehyde (10 min) and placed in blocking solution [phosphate-buffered saline (PBS) containing 3% bovine serum albumin and 2% fetal bovine serum; 30 min]. The cells were then incubated with a rabbit anti-extracellular epitope (aminoacids 136-152) of P2X7 receptor antibody (1 : 500; Alomone Labs, Jerusalem, Israel) in blocking solution overnight at 4°C, and subsequently with goat anti-rabbit Alexa Fluor 488 or 633 secondary antibody conjugates (1 : 1000; Invitrogen, Life Technologies Italia, Monza, Italy) in blocking solution for 1 h at 20 to 25°C. HEK293-rP2X7ΔCys-rich cells were examined for expression of the receptors on the cell membrane 48 h after transfection, in parallel with functional experiments. Addressing of rP2X7 on the plasma membrane was assessed by merging with the membrane protein Na+/K+-ATPase (details are given in the Supporting Information).

Cell surface biotinylation

HEK cells (about 106 cells) were surface biotinylated by incubating with Sulpho-NHS-SS-Biotin (Pierce, Thermo Fisher Scientific Inc., Rockford, IL, USA) at 2 mg/mL in ice-cold PBS. After 30 min, cells were washed and quenched with PBS containing 100 mM glycine. Cells were then lysed by sonication in 20 mM TRIS/HCl pH 7.4, 0.14 M NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP40. Protein concentration was determined using a bicinchoninic acid reaction kit (Pierce). Equal amounts of protein were incubated with NeutrAvidinTM Protein immobilized onto 6% cross-linked beaded agarose (Thermo Fisher Scientific Inc). After washing in NP-40 buffer, the beads were boiled in sample-buffer, fractionated in sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblotted according to Massone et al. (2007) with the rabbit anti-P2X7 extracellular antibody (1 : 400). Analysis software KODAK 1D 3.6 (Eastman Kodak Co., Rochester, NY, USA) was used for standardization and band quantification; surface expression of rP2X7tr was normalized for full-length rP2X7 expression.

Calculation and statistics

Log concentration-response relationships for the agonist and EC50 values (half-maximum effective concentrations; 95% confidence limits in parentheses) were obtained by using GraphPad Prism (GraphPad Software, San Diego, CA, USA). Means ± SEM of the numbers of experiments (n) are indicated throughout. Significance of the difference was analyzed by the non-parametric Mann–Whitney test, with statistical significance being taken at < 0.05.

Drugs

[3H]d-aspartate was from Amersham Radiochemical Centre (Buckinghamshire, UK); A-438079 from Tocris Cookson (Bristol, UK); ATP, BAPTA-AM, BzATP, colchicine, probenecid and trypan blue from Sigma-Aldrich (Milan, Italy); Lipofectamine 2000 from Gibco (Life Technologies). Drugs were dissolved in distilled water or physiological medium. All the salts and chemicals used for the preparation of cultures and the fluorimetric determinations of the [Ca2+]i were obtained from Sigma-Aldrich.

Results

BzATP-evoked [3H]d-aspartate release from HEK293 cells transfected with the rP2X7 receptor

The [3H]d-aspartate efflux in the absence of exogenously added agonists (basal [3H]d-aspartate efflux) in native and transfected HEK293 cells is reported in the supporting information; notably, transfecting HEK293 cells with rP2X7 receptor did not affect basal [3H]d-aspartate efflux (Fig. 1c).

Figure 1.

2′-3′-O-(benzoylbenzoyl) ATP (BzATP)-evoked [3H]d-aspartate release from HEK293 cells. (a) BzATP increased [3H]d-aspartate efflux from HEK293 cells stably expressing rP2X7 receptor, but not from native HEK293 cells (wt). Neither BzATP nor ATP increased [3H]d-aspartate efflux from HEK293 cells stably expressing rP2X2 receptor. (b) Time course of a representative experiment showing BzATP-evoked [3H]d-aspartate efflux from HEK293 cells stably transfected with rP2X7. [3H]d-aspartate efflux in the 5-min fractions collected from a control chamber (□) superfused with standard medium (no drug was added) and a chamber supplemented with BzATP 30 μM (●) is shown. The horizontal solid line indicates addition of the agonist to the superfusion medium. (c) No effect of rP2X7 transfection or rP2X7 antagonist on basal [3H]d-aspartate efflux. Mean [3H]d-aspartate efflux/min in the first two fractions collected (basal fractional [3H]d-aspartate release/min) is shown in native HEK293 cells (wt) and in HEK293 cells stably expressing rP2X7 receptor in the absence or in the presence of A-438079 (10 μM). Bars represent the mean ± SEM of 3–12 independent experiments in duplicate. (d) Antagonism by A-438079 of BzATP-evoked tritium efflux in HEK293 cells expressing rP2X7 receptor. Bars represent percentage increase in tritium efflux in the presence of the drugs at the concentrations indicated (mean ± SEM of three independent experiments in duplicate). (e) Log concentration-response curves for BzATP-evoked [3H]d-aspartate efflux from HEK293 cells expressing rP2X7 receptor: dependence on external Mg2+. BzATP effect in standard medium containing 1.2 mM Mg2+ (○) or in medium with 0.01 mM Mg2+ (●) is shown. Data are mean ± SEM of 3–9 independent experiments in duplicate. *p < 0.05 when compared with the effect of BzATP alone.

The rP2X7 agonist BzATP (30 μM) evoked [3H]d-aspartate efflux from HEK293-rP2X7 cells, but not from control native HEK293 cells or HEK293-rP2X2 (Fig. 1a). The time course of BzATP-induced [3H]d-aspartate efflux from HEK293-rP2X7 is shown in the representative experiment in Fig. 1b; returning of the efflux to basal level after agonist removal and trypan blue exclusion (Figure S1) made unlikely a BzATP-evoked cytolytic [3H]d-aspartate release. The rP2X7 antagonist A-438079 (10 μM), ineffective on basal [3H]d-aspartate efflux (Fig. 1c), prevented the response to BzATP (30 μM) (Fig. 1d).

The maximal [3H]d-aspartate-releasing effect of BzATP was greatly increased by lowering the extracellular Mg2+ concentration to 0.01 mM, while the BzATP half-maximal effective concentrations were unaffected (EC50 values: 16.2 (9.9–26.2) μM in 1.2 mM Mg2+ and 10.3 (9.7–11.1) μM in 0.01 mM Mg2+; Fig. 1e); this is consistent with a non-competitive block of the BzATP-activated channel by extracellular Mg2+. The BzATP (30 μM)-evoked [3H]d-aspartate efflux was increased by extracellular Ca2+ deprivation, and unmodified by the intracellular Ca2+ chelator BAPTA-AM (Fig. 2a).

Figure 2.

2′-3′-O-(benzoylbenzoyl) ATP (BzATP)-evoked [3H]d-aspartate release from HEK293 cells expressing rP2X7 receptor: (a) no requirement for Ca2+, and (b) ineffectiveness of probenecid or colchicine. Bars represent percentage increase in tritium efflux in the presence of the drugs at the concentrations indicated. BzATP was added for 5 min; Ca2+-free EGTA (0.5 mM)-containing medium was added 20 min before the agonist; probenecid or colchicine was added 10 min before the agonist; 1,2-bis-(o-aminophenoxy)-ethane-N,N,-N ′,N ′-tetraacetic acid tetraacetoxy methyl ester (BAPTA-AM) was pre-incubated for 30 min before superfusion. Data are mean ± SEM of 3–6 independent experiments in duplicate. *p < 0.05 when compared with the effect of BzATP in standard medium.

Recruitment of associated pore-forming molecules was not required for BzATP-evoked [3H]d-aspartate release from HEK293 cells transfected with the rP2X7 receptor

Pannexin-1 (panx1), a mammalian protein that functions as a hemichannel, has been reported to mediate the large-pore formation associated with P2X7 receptor activation (Pelegrin and Surprenant 2006; Di Virgilio 2007; Locovei et al. 2007). We previously reported that the BzATP-evoked [3H]d-aspartate efflux from HEK293-rP2X7 was unaffected by the panx1 and connexin inhibitor carbenoxolone (CBX, 10–50 μM; Marcoli et al. 2008); here, we report that the BzATP (30 μM)-evoked [3H]d-aspartate efflux was also unaffected by the panx1 inhibitor probenecid (1000 μM; Fig. 2b).

The existence of a separate permeation pathway for dyes, which is shared between P2X7 and P2X2 receptors and is sensitive to colchicine, has recently been hypothesized (Marques-da-Silva et al. 2011). Herein, we show that the BzATP (30 μM)-evoked [3H]d-aspartate efflux from HEK293-rP2X7 was unaffected by colchicine (25–50 μM; Fig. 2b).

Probenecid or colchicine at the concentrations used, or BAPTA-AM pre-incubation, did not per se significantly affect the basal [3H]d-aspartate outflow from HEK293-rP2X7 (data not shown).

BzATP did not evoke [3H]d-aspartate release from HEK293 cells transfected with either rP2X7tr or rP2X7ΔCys-rich receptors

BzATP (up to 100 μM) did not affect [3H]d-aspartate efflux in HEK293-rP2X7tr; in addition, lowering the extracellular Mg2+ concentration from 1.2 to 0.01 mM did not unmask any [3H]d-aspartate-releasing response to BzATP (Fig. 3). To determine whether the ineffectiveness of BzATP in evoking [3H]d-aspartate efflux from HEK293-rP2X7tr depended on loss of localization of the receptors on the cell membrane, parallel immunocytochemical staining experiments were performed. In Fig. 4a, the localization of stably transfected rP2X7 and rP2X7tr receptors on the HEK293 cell membrane is shown by using an antibody against an extracellular receptor epitope. Merging with the membrane protein Na+/K+-ATPase (Figure S2), and surface biotinylation (Fig. 4b; normalized surface expression for rP2X7tr: 54.7 ± 10.4%; n = 3) confirmed addressing of rP2X7tr at the cell membrane.

Figure 3.

Effect of 2′-3′-O-(benzoylbenzoyl) ATP (BzATP) on [3H]d-aspartate efflux from HEK293 cells transfected with full-length rP2X7 receptor (rP2X7), and ineffectiveness in HEK293 cells transfected with truncated rP2X7 receptor lacking the C-terminal tail (rP2X7tr), or rP2X7 receptor with the 18-amino acid cysteine-rich motif of the C-tail deleted (rP2X7ΔCys-rich), in 1.2 mM (empty bar) and 0.01 mM (gray bar) Mg2+-containing medium. Agonist was added for 5 min during superfusion; low Mg2+-containing medium was added starting 20 min before the agonist. Data are mean ± SEM values of = 3–13 experiments in duplicate.

Figure 4.

Surface expression on HEK293 cell membrane of full-length rP2X7 receptor, truncated rP2X7 receptor lacking the C-tail (rP2X7tr), or rP2X7 receptor with deletion of the 18-amino acid cysteine-rich motif of the C-tail (rP2X7ΔCys-rich). (a) Negative staining control in native wt HEK293 cells. Representative rP2X7 immunocytochemical staining in HEK293 cells expressing rP2X7, rP2X7tr, or rP2X7ΔCys-rich. Horizontal bar: 10 μm. (b) Surface biotinylation of native HEK293, HEK293-rP2X7 and HEK293-rP2X7tr cells; western blot analysis of one representative experiment out of three experiments.

The proximal cysteine-rich 18-amino acid segment (Cys-rich motif) of the P2X7 C-tail is responsible for the channel second large-state opening conformation (NMDG+ permeability), while the distal region of the C-tail is responsible for the downstream formation of pores, which allow the passage of molecules up to 900 Da (Smart et al. 2003; Jiang et al. 2005). We therefore looked for motifs in the rP2X7 sequence that might be involved in the dependence of [3H]d-aspartate efflux on the receptor C-tail. To this end, we investigated whether the rP2X7 receptor with this Cys-rich motif deleted retained the property of allowing glutamate release. We found that lowering the extracellular Mg2+ concentration to 0.01 mM did not unmask any [3H]d-aspartate-releasing response to BzATP (up to 100 μM) in HEK293-rP2X7ΔCys-rich cells (Fig. 3). In parallel experiments, we assessed the localization of the rP2X7ΔCys-rich receptor on the HEK293 cell membrane by using an antibody against an extracellular receptor epitope and by merging with the membrane protein Na+/K+-ATPase (Fig. 4a; Figure S2).

Calcium imaging data showed that HEK293-rP2X7, HEK293-rP2X7tr, and HEK293-rP2X7ΔCys-rich elicited a [Ca2+]i variation when exposed to BzATP. In HEK293-rP2X7tr cells, BzATP elicited a [Ca2+]i signal that was lower than that in HEK293-rP2X7 cells. The concentration dependency of the BzATP-induced [Ca2+]i increase is shown in Fig. 5a, from which we obtained the EC50 values 5.0 (4.82–5.04) and 35.6 (29.2–43.4) μM and a Hill coefficient of 4.5 and 1.6 in HEK293-rP2X7 and HEK293-rP2X7tr, respectively. In HEK293-rP2X7ΔCys-rich cells, BzATP caused a concentration-dependent [Ca2+]i increase, with an EC50 value of 5.8 (4.82–7.13) μM and a Hill coefficient value of 2 (Fig. 5a). Representative traces of a single-cell fluorescence ratio signal (F340/F380) from a HEK293-rP2X7 cell, a HEK293-rP2X7tr cell, and a HEK293-rP2X7ΔCys-rich cell are shown in Fig. 5b–d. As expected, in the absence of extracellular Mg2+, the Ca2+ response to BzATP was increased in HEK293-rP2X7, HEK293-rP2X7tr, and HEK293-rP2X7ΔCys-rich cells, thus suggesting that not only the rP2X7 receptor but also the mutant rP2X7 constructs are expressed and induce a [Ca2+]i increase when targeted by BzATP.

Figure 5.

The P2X7 agonist 2'-3'-O-(benzoylbenzoyl) ATP (BzATP) activates Ca2+ influx in full-length rP2X7, truncated rP2X7 receptor lacking the C-tail (rP2X7tr), and in rP2X7 receptor with deleted cysteine-rich motif in the cytosolic C terminus (rP2X7ΔCys-rich). (a) Quantitative analysis of [Ca2+]i elevations above basal levels elicited by BzATP (0.1–400 μM) at rP2X7, rP2X7ΔCys-rich and rP2X7tr receptors. Means ± SEM of at least 42 cells. (b, c, d) Mg2+ dependence of BzATP-evoked [Ca2+]i increase in cells transfected with full-length rP2X7, rP2X7tr, or rP2X7ΔCys-rich, respectively. Representative traces of single-cell fluorescence ratio depicting the [Ca2+]i increase in 1 mM Mg2+ and Mg2+-free extracellular saline. The amplitude of the [Ca2+]i response elicited by BzATP (3 or 10 μM, in the different HEK293-rP2X7 types) was strongly dependent on the presence of extracellular Mg2+.

Discussion

The P2X7 receptor as a route for glutamate release

We investigated the P2X7 receptor function as a pathway for non-exocytotic glutamate release by measuring the efflux of [3H]d-aspartate. This glutamate analog is taken up by the plasma membrane glutamate transporters, which are natively expressed on HEK293 cells (see Toki et al. 1998; Marcoli et al. 2008), and permeates P2X7 receptors to a similar degree to glutamate (Duan et al. 2003). The preferential P2X7 agonist BzATP was ineffective in eliciting [3H]d-aspartate efflux from native HEK293 cells (not bearing P2X7 receptors; see Humphreys et al. 1998; Alloisio et al. 2006). However, at low micromolar concentrations, it evoked [3H]d-aspartate efflux from cells transfected with full-length rP2X7. P2X7 receptors were reported to detect low basal levels of ATP and to undergo tonic, low-level activation depending on an autocrine/paracrine stimulation by secreted ATP (Adinolfi et al. 2005; Thompson et al. 2012). In our conditions – superfusion preventing accumulation of endogenously secreted substances in the receptor biophase – no evidence was found for rP2X7 receptor activation in the absence of exogenously added agonist. Pharmacological analysis has indicated that activation of rP2X7 is responsible for the [3H]d-aspartate-releasing response of HEK293-rP2X7 to BzATP: the response is eliminated by oxidized ATP or Brilliant Blue G (Marcoli et al. 2008). Effectiveness of the selective antagonist A-438079 confirmed the involvement of rP2X7. The BzATP response did not require Ca2+ permeation, being increased in the absence of extracellular Ca2+ (consistent with a rP2X7 response; North 2002), and unaffected by BAPTA-AM. The BzATP-evoked [3H]d-aspartate efflux was greatly amplified in Mg2+-free solution, which is consistent with a rP2X7-mediated response. The shift in the concentration–response curve indicates non-competitive Mg2+ antagonism of the BzATP-activated [3H]d-aspartate release. Interestingly, we previously observed differences in the modulatory action of extracellular Mg2+ on the gating modes through which rP2X7 conduct different ions: while the BzATP-elicited Ca2+ influx in HEK293-rP2X7 was inhibited by Mg2+ in a concentration-dependent manner that suggested a competitive block, Mg2+ behaved as a non-competitive antagonist of monovalent cation (Na+/NMDG+) currents through the receptor (Alloisio et al. 2010).

In HEK293-rP2X7 cells, no evidence has been found for BzATP-induced [3H]d-aspartate release by routes other than the P2X7 channel. Indeed, the inhibitor of glutamate transporters DL-TBOA (Shimamoto et al. 1998), the anion channel inhibitor niflumic acid (Nilius and Droogmans 2003) and the connexin hemichannel inhibitor CBX (Bruzzone et al. 2005; Pelegrin and Surprenant 2006) have proved ineffective (Marcoli et al. 2008). It should also be considered that CBX at the same concentration (50 μM) has been shown to effectively block glutamate release from astrocytes or HEK293 cells by swelling-activated anion channels (Benfenati et al. 2009; Ye et al. 2009). The ineffectiveness of CBX therefore excludes the involvement of either connexin hemichannels or swelling-activated anion channels in the BzATP-evoked efflux of [3H]d-aspartate. Thus, the BzATP-induced [3H]d-aspartate release from HEK293-rP2X7 cells most likely occurs through the rP2X7 receptor, as already shown for glutamate release from rodent cultured astrocytes (Duan et al. 2003; see also Duan and Neary 2006). The possibility that glutamate could be released through a non-conventional P2X7-activated way, for example, through the shedding of microvesicles (MacKenzie et al. 2001) or the budding of exosomes (Qu and Dubyak 2009; Guescini et al. 2012) that could be under the dependence of a C-tail signaling pathway, should be considered. However, no requirement for Ca2+ of rP2X7-evoked efflux makes unlikely the involvement of such alternative, mostly Ca2+-sensitive pathways (MacKenzie et al. 2001; Qu and Dubyak 2009).

Although rP2X2 can also undergo pore dilatation on prolonged exposure to an agonist (Virginio et al. 1999; Marques-da-Silva et al. 2011), no agonist-evoked [3H]d-aspartate efflux has been observed in HEK293-rP2X2 cells, indicating that the receptor does not function as a pathway for glutamate release. Indeed, functioning as a route for glutamate release appears to be a specific feature of rP2X7; the different behavior of rP2X2 and rP2X7 receptors as a pathway for glutamate release is unlikely to rest on different levels of Ca2+ loading.

P2X7 receptor requirements to function as a route for glutamate release

Notably, the concentrations of BzATP that are effective on HEK293-rP2X7 cells (≤ 10 μM) suggest that cytolytic pore-forming activity is not required for the receptor to function as a route for glutamate release. It is recognized that rP2X7 can activate multiple permeation pathways with different degrees of selectivity and different activation properties, which are dependent on the involvement of downstream signaling interactions (Jiang et al. 2005; Cankurtaran-Sayar et al. 2009; Alloisio et al. 2010; see also Pelegrin 2011): cytolytic pore-formation, which is related to dye uptake, has been reported to require the receptor C-tail (Surprenant et al. 1996; Rassendren et al. 1997; Smart et al. 2003) and to be associated with the opening of panx1 hemichannels (Pelegrin and Surprenant 2006; Di Virgilio 2007; Locovei et al. 2007). Panx1 can be blocked by low concentrations of the gap-junction blocker CBX (an effective inhibitor of panx1 currents at 1–20 μM and of gap-junction hemichannels at 10–500 μM; Bruzzone et al. 2005; Pelegrin and Surprenant 2006). Among gap-junction inhibitors, probenecid, an inhibitor of organic anion transporters, is specific for channels composed of pannexins, whereas connexins are insensitive to the drug (Silverman et al. 2008). We found that glutamate efflux through the recombinant rP2X7 receptor did not require panx1 activation, as indicated by the ineffectiveness of CBX or probenecid. Likewise, native rP2X7 expressed on cerebrocortical glutamatergic nerve terminals did not appear to require panx1 to function as a route for glutamate release (Marcoli et al. 2008). In addition, the ineffectiveness of colchicine indicates that glutamate efflux through the rP2X7 does not require the formation of the pore for big dyes, which has recently been found to be shared between P2X7 and P2X2 receptors (Marques-da-Silva et al. 2011).

The P2X7 receptor is distinguishable from other P2X receptors by its long intracellular C-tail, which is essential for receptor-mediated cytolytic pore-formation, and which also contains multiple interaction motifs that are important for the activation of P2X7 signal transduction pathways (Denlinger et al. 2001; Wilson et al. 2002; Barbieri et al. 2008; Costa-Junior et al. 2011). In this study, the truncated receptor lacking the intracellular C-tail lost the ability to function as a route for glutamate release. The inability to function as a route for glutamate efflux did not depend on loss of localization of the truncated receptor at the cell surface. Indeed, parallel experiments using antibodies against an extracellular receptor epitope, merging with plasma membrane protein Na+/K+-ATPase and biotinylation showed that the rP2X7tr was expressed at the plasma membrane of HEK293 cells; by measuring intracellular Ca2+, we showed that the receptor was functional. These findings fit in with previous evidence that the rP2X7 receptor truncated at 380, which does not retain pore-forming ability, is expressed at the cell surface, although at lower level than the wild-type receptor, and functions as an ion channel allowing Ca2+ influx (Smart et al. 2003).

We then searched for motifs in the rP2X7 C-tail that could be involved in ability of the receptor to function as a route for glutamate release. P2X7 receptors contain an 18-amino acid motif that is rich in six cysteine residues, and which is located immediately after the second transmembrane segment. This sequence is not found in other P2X receptors, is highly conserved in P2X7 receptor orthologs, and appears to contribute to properties specific to the P2X7 receptor (Jiang et al. 2005; Roger et al. 2010). The Cys-rich motif has been shown to be involved in receptor facilitation and permeability to NMDG+ but not in dye uptake (Jiang et al. 2005); moreover, it has been hypothesized to represent a hinge region for receptor conformational changes occurring during facilitation (Roger et al. 2010). Indeed, the proximal Cys-rich motif of the P2X7 C-tail is responsible for the channel second large-state opening conformation (NMDG+-permeability; see also Yan et al. 2010 and Pelegrin 2011), while the distal P2X7 C-tail region is responsible for the downstream pore formation that allow the passage of dyes (Smart et al. 2003; Jiang et al. 2005). We found that the receptor lacking this Cys-rich motif did not retain the ability to allow glutamate release. Conversely, and confirming previous evidence (Jiang et al. 2005; Roger et al. 2010), the receptor was expressed at the HEK293 cell membrane and functioned as a channel for Ca2+. The receptor pathway permeable to glutamate (which does not require accessory panx1 or colchicine-sensitive proteins, is lost in rP2X7tr and rP2X7ΔCys-rich and is non-competitively antagonized by Mg2+) bears a resemblance to the second large-state opening conformation permeable to NMDG+, which is lost in rP2X7tr (Alloisio et al. 2010) and in rP2X7ΔCys-rich (Jiang et al. 2005) and antagonized in a non-competitive way by Mg2+ (Alloisio et al. 2010).

Conclusions and functional implications

P2X7 receptors and glutamate release

We recently addressed P2X7 receptor gating behavior by comparing full-length and truncated rP2X7 receptors expressed in HEK293 cells. The model predicted two distinct conductive pathways: one for Ca2+, which still functioned in the truncated receptor lacking the C-tail; the other for NMDG+, which was lost in the truncated receptor (Alloisio et al. 2010). Taken together, our previous (Marcoli et al. 2008) and present data indicate that these two rP2X7 pathways may be linked to glutamate release: the pathway conductive for Ca2+ being associated with the receptor function as a Ca2+ channel coupled to glutamate exocytosis; the pathway permeable to NMDG+ being associated with the receptor function as a route for non-exocytotic glutamate release (see the model schematically presented in Fig. 6). When the C-tail, or the cysteine-rich motif of the C-tail, is absent, the receptor's function as a route for glutamate release is lost, while its function as a Ca2+ channel (linked to exocytotic glutamate release) is still present. The finding that non-exocytotic glutamate release is dependent on the C-tail may have significant consequences. Indeed, the two P2X7-activated modes of glutamate release could play different roles in glutamatergic synapse physiology and in the pathogenesis of disorders involving dysregulation of glutamatergic transmission. In such disorders, non-exocytotic glutamate release can become predominant (see Rossi et al. 2007 for the glutamate efflux modes during brain ischemia), and P2X7 receptor activation is likely to be involved (Sperlagh et al. 2006; Burnstock 2007; Verkhrasky et al. 2009; Arbeloa et al. 2012). In these conditions, the release of ATP through the P2X7 receptor (Ballerini et al. 1996; Pellegatti et al. 2005; Suadicani et al. 2006) might also contribute to spreading excitotoxicity by co-operating with glutamate in a self-sustaining cascade.

Figure 6.

Modes of glutamate release dependent on P2X7 receptor activation. (a) Schematic representation of the P2X7 receptor (only one subunit shown; the receptor is believed to be multimeric or trimeric) function as a route for glutamate release in HEK293 cells stably transfected with rP2X7 receptors. Two distinct pathways are functioning: the pathway permeable to Ca2+ and allowing Ca2+ influx (green) and the pathway permeable to glutamate and allowing glutamate release (red). The function as a route for glutamate release is lost in truncated P2X7 receptors lacking the C-terminal tail and in P2X7 receptors with the 18-amino acid cysteine-rich motif of the C-tail deleted; the truncated receptor and the receptor with the cysteine-rich motif deleted retain permeability to Ca2+. (b) Hypothetical model showing how activation of native pre-synaptic rP2X7 receptors can evoke glutamate release from glutamatergic nerve terminals (see Marcoli et al. 2008). Activation of glutamate vesicular exocytosis is linked to entry of Ca2+ (activation of the P2X7 conductive pathway for Ca2+, green); the mode does not require the receptor C-tail. Activation of non-exocytotic glutamate efflux (red); the mode is dependent on the receptor C-tail.

As a result of species differences in the P2X7 receptor properties, including functional responses such as facilitation and permeation properties and the pharmacological prophile (Humphreys et al. 1998; Roger et al. 2010; Bradley et al. 2011 and references therein), speculation on human P2X7 (hP2X7) receptor cannot be directly inferred from the findings on rP2X7. In the human brain, two variants of the P2X7 receptor – distinguishable by the presence (P2X7A) or absence (P2X7B) of the C-tail – are highly expressed (Cheewatrakoolpong et al. 2005; Adinolfi et al. 2010), and a multiplicity of polymorphic forms of the P2X7 receptor associated with the C-terminal region have been identified (see Ferrari et al. 2006; Fuller et al. 2009; Sun et al. 2010 and references therein). These naturally occurring hP2X7 variants – including disease-associated variants – would deserve to be evaluated for the ability to allow non-exocytotic glutamate release.

HEK293 cells as a suitable model for studying the P2X7 receptor as a route for non-exocytotic glutamate release

HEK293 cells, which do not natively express P2X receptors and which can be stably transfected with P2X7 receptors (including P2X7 receptor variants from different species), appear to be a suitable model for investigating the receptor function as a route for glutamate release. Indeed, native HEK293 cells are equipped with specific glutamate transporters that allow the cells to be loaded with the glutamate analog [3H]d-aspartate and endogenously express panx1 (Pelegrin and Surprenant 2006; Yan et al. 2008), the hemichannel that can be recruited in the cytolytic pore conformation of the receptor. It is noteworthy that, owing to the absence of the machinery for Ca2+-dependent vesicle exocytosis, HEK293 cells allow the investigation of P2X7 receptor function as a route for non-exocytotic glutamate efflux.

Acknowledgements

This study was supported by the Mariani Foundation of Milan (Grant no.R-07-66) to M.N.; the Ca.Ri.Ge of Genoa to M.N. and G.M.; and the University of Genoa to M.M. and to C.C. We thank Prof. Alan North (University of Manchester) for providing us with rP2X7ΔCys-rich cDNA and Prof. Francesco Di Virgilio (University of Ferrara) for providing full-length rP2X7-GFP cDNA and HEK293 cells stably transfected with 1-375 rP2X7 receptor. We thank Prof. Annmarie Surprenant (University of Manchester) for HEK293 stably transfected with rP2X2 receptor.

Conflict of interest

The authors declare that they have no conflicts of interest.

Ancillary