Cell Physiology

See original article:


Pannexin 1 (Panx1), an ortholog to invertebrate innexin gap junctions, has recently been proposed to be the pore induced by P2X7 receptor (P2X7R) activation. We explored the pharmacological action of compounds known to block gap junctions on Panx1 channels activated by the P2X7R and the mechanisms involved in the interaction between these two proteins. Whole cell recordings revealed distinct P2X7R and Panx1 currents in response to agonists. Activation of Panx1 currents following P2X7R stimulation or by membrane depolarization was blocked by Panx1 small-interfering RNA (siRNA) and with mefloquine > carbenoxolone > flufenamic acid. Incubation of cells with KN-62, a P2X7R antagonist, prevented current activation by 2′(3′)-O-(4-benzoylbenzoyl)adenosine 5′-triphosphate (BzATP). Membrane permeabilization to dye induced by BzATP was also prevented by Panx1 siRNA and by carbenoxolone and mefloquine. Membrane permeant (TAT-P2X7) peptides, provided evidence that the Src homology 3 death domain of the COOH-terminus of the P2X7R is involved in the initial steps of the signal transduction events leading to Panx1 activation and that a Src tyrosine kinase is likely involved in this process. Competition assays indicated that 20 μM TAT-P2X7 peptide caused 50% reduction in Src binding to the P2X7R complex. Src tyrosine phosphorylation following BzATP stimulation was reduced by KN-62, TAT-P2X7 peptide, and by the Src tyrosine inhibitor PP2 and these compounds prevented both large-conductance Panx1 currents and membrane permeabilization. These results together with the lack Panx1 tyrosine phosphorylation in response to P2X7R stimulation indicate the involvement of an additional molecule in the tyrosine kinase signal transduction pathway mediating Panx1 activation through the P2X7R.

  • gap junction blockers
  • permeabilization
  • P2Z
  • src-tyrosine kinase

the ionotropic purinergic P2X7 receptors (P2X7R) initially respond to agonists with the opening of a nonselective cation channel (24, 32). Prolonged exposure, however, leads to opening of a large pore that is permeable to molecules as large as 831 Da (44). Although this dual mode of function has been observed in many cell types, the pore has been reported to be absent in Xenopus oocytes in which P2X7 cRNA is expressed (40), leading to the hypothesis that the small-conductance cation pathway may correspond to the P2X7R (whose other family members display such permeation upon ATP binding: 32), whereas the pore might be formed by another molecule that is recruited in the response. A candidate molecule recently proposed for the highly permeable pore is pannexin1 (Panx1) (30, 38). This molecule, found in chordates, displays a low but significant degree of homology to innexins, the proteins that form gap junction channels in invertebrates (5, 52). Although Panx1 overexpression in Xenopus oocytes has been reported to result in a low degree of functional gap junction formation (10), formation of such intercellular channels by Panx1 in mammalian cells is unlikely (16, 21) given the presence of a glycosylation site in the extracellular docking domain of this protein (7, 8, 37). Nevertheless, Panx1 readily forms nonjunctional channels in both Xeonopus oocytes and in mammalian cells; these channels can be activated by stretch (3), elevations in intracellular Ca2+ (28), and depolarization above about +20 mV (3, 9). Moreover, Panx1 channels can be activated by prolonged stimulation of P2X7R (30, 38). In mammalian cells, activation of the P2X7R-Panx1 complex has been implicated in the processing of the proinflammatory cytokine interleukin-1β and the activation of caspase-1 in immune cells (38, 39) and in the release of ATP from several cell types, including astrocytes (41, 46), erythrocytes (29), and taste bud cells (22).

It has been proposed that nonjunctional (hemi-) channels can be formed by the gap junction proteins, the connexins, and that these may perform the function of large-conductance, highly permeable pores, allowing release of ATP and/or glutamate from astrocytes (45, 51; reviewed in Ref. 43). The stimulus used to activate connexin hemichannels is removal of divalent cations, a treatment long known to enhance P2X7R sensitivity to agonists (49). Moreover, evidence for such a role of connexin hemichanels is heavily dependent on use of compounds known to block intercellular gap junction channels, which have recently been shown to also block the P2X7R-mediated uptake of dye molecules (46). It therefore remains possible that Panx1 through its association with P2X7R may mediate the dye uptake and molecule release attributed to connexin hemichannels.

A previous study reported inhibition of flux of dye molecules only by one of several gap junction blockers tested and that dye uptake was not strictly correlated with large-conductance channel activation (38). Therefore, the present study was undertaken to further explore the pharmacological action of compounds known to block gap junctions on the cation channel and Panx1 pore associated with P2X7R and to gain insight into mechanisms involved in the interaction between these two groups of proteins. Our results provide further support that Panx1 is the permeabilization pore activated in association with the P2X7R and demonstrate that the antimalarial drug mefloquine is a highly potent inhibitor of Panx1 channels. We also provide, for the first time, evidence for the participation of a Src tyrosine kinase mediating the initial steps of the signal transduction pathway involved in the activation of Panx1 channels through the P2X7R.


Cell culture.

The mouse macrophage cell line J774 (obtained from Dr. A. Casadevall, Albert Einstein College of Medicine) was used in this study. Cells were cultured in DMEM, supplemented with 10% FCS (GIBCO), and 1% of penicillin/streptomycin and maintained in an humidified chamber with 5% CO2 at 37°C.


J774 cells were plated on cover slips 24–48 h before recordings. The whole cell patch-clamp configuration was performed at room temperature on cells bathed in external solution containing (in mM): 147 NaCl, 10 HEPES, 13 glucose, 2 CaCl2, 1 MgCl2, and 5 KCl, pH 7.4. The pipette solution contained (in mM): 130 CsCl, 10 EGTA, 10 HEPES, and 0.5 CaCl2. Activation of Panx1 channel by voltage was performed using a ramp protocol by applying voltages from a holding potential of −60 mV to final values of +100 mV. To analyze the participation of Panx1 channels in agonist-induced P2X7R activation, cell membrane potential was held at −60 mV after forming high-resistance (1–10 GΩ) seals and the whole cell recording configuration. 2′(3′)-O-(4-benzoylbenzoyl)adenosine 5′-triphosphate (BzATP, 50 μM) and ATP (1–5 mM) were superfused for 5 s. After the first response to the agonist, the gap junction channel blockers carbenoxolone (50 μM; CBX), mefloquine (10–100 nM; MFQ), and flufenamic acid (100–300 μM; FFA) and the P2X7R antagonist KN-62 (1 μM) were superfused for 5 min before the addition of 50 μM BZATP in the presence of the blockers. Cells were then washed with external solution for several minutes and further stimulated with 50 μM BzATP to evaluate reversibility of the blockers and the P2X7R antagonist. All data were recorded through an Axopatch 200B amplifier and digitized using an Axon Instruments Digitizer (Molecular Devices). Clampex 6.0 was used for recording and the Clampfit 9.0 program for analysis.


Preparation of oocytes and electrophysiological recording were performed using the two-microelectrode voltage-clamp technique, as previously described (30). Human Panx1 (MRS1) was kindly provided by Dr. Graeme Bolger, University of Alabama. Panx1, in Bluescript, was linearized with Kpn I and human P2X7 (kindly provided by Dr. Annmarie Suprenant, University of Manchester, Manchester, United Kingdom), in pcDNA3, with Avr II. In vitro transcription was performed with the polymerases T3 (Panx1) and T7 (P2X7R) using the mMessage mMachine kit (Ambion, Austin, TX). mRNAs were quantified by absorbance (260 nm), and the proportion of full-length transcripts was checked by agarose gel electrophoresis. In vitro transcribed mRNAs (∼20 nl) were injected in Xenopus oocytes. The oocytes were incubated at 18°C for 18–24 h in oocyte Ringer solution (OR2) (in mM: 82.5 NaCl, 2.5 KCl, 1 MgCl2, 1 CaCl2, 1 Na2HPO4, and 5 HEPES, pH 7.5). Whole cell membrane current of single oocytes was measured using a two-microelectrode voltage clamp and recorded with a chart recorder. Both voltage-measuring and current-passing microelectrodes (∼1.0 MΩ resistance) were pulled with a vertical Puller (Kopf) and filled with 3 M KCl. The recording chamber was perfused continuously with solution. The TAT-P2X7 peptides (see below) were applied for 30 min, and oocytes were washed by perfusion with OR2 before testing with 500 μM ATP. The holding potential was −30 mV.

Small-interfering RNA.

J774 cells were treated with 2 μg/1.5 ml small interference RNA (siRNA) corresponding to the mouse Panx1 sequence using 6 μl/1.5 ml oligofectamine reagent (Invitrogen), as previously described (30). After overnight exposure, transfection reagents were removed, and cells were incubated for 30–48 h in DMEM-FBS medium.

Immunoprecipitation and Western blots.

Confluent cultures of J774 cells untreated and treated with Panx1 siRNA were lysed in Tris-buffered solution containing 1% Triton X-100 (Sigma) and protease inhibitor cocktail (Sigma). Precleared lysates were incubated for 3 h with anti-P2X7 (Alomone Laboratories), with anti-phosphotyrosine (Upstate, Millipore) or with anti-Panx1 antibodies (see Ref. 30) at 4°C before the addition of agarose-conjugated IgG (UltraLink Immobilized Protein G; Pierce) or IgY (when using anti-Panx1 antibodies; ProSci) beads. After overnight incubation at 4°C, immunocomplexes were washed five to six times with lysis buffer solution, and bound proteins were eluted with 2× Laemmli buffer containing 100 mM dithiothreitol (DTT). Samples of immunoprecipitated proteins were subjected to electrophoresis using SDS-PAGE minigels (4–20%; Bio-Rad). Proteins were transferred to nitrocellulose membranes (Schleider & Schuell), and, after overnight blocking (2% nonfat dry milk in 1× PBS with 0.05% Tween 20), membrane were blotted using either anti-P2X7R (1:500) or anti-Panx1 (1:2,000) for 2 h at room temperature. After several washes in 1× PBS-Tween 20 solution, membranes were incubated for 2 h with goat-anti-rabbit or rabbit-anti-chicken horseradish peroxidase (HRP)-conjugated secondary antibodies (1:2,000; Santa Cruz Biotechnology). Parallel experiments were performed on whole cell lysates of untreated and Panx1 siRNA-treated J774 cells blotted with anti-P2X7R antibodies (1:500) and goat anti-rabbit HRP-conjugated secondary antibodies. Bands were visualized on X-ray films (Kodak) after incubation with the enhanced chemiluminescence reagent (Amersham). Quantification of Panx1 and P2X7R expression levels was performed using NIH image-J software.

Src phosphorylation.

Confluent cultures of J774 cells plated on 100-mm dishes were exposed for 15 min to BzATP (300 μM) in the absence or presence of the murine P2X7R antagonist KN-62 (1 μM), the TAT-P2X7 peptide (10 μM), and the src tyrosine kinase inhibitor PP2 (10 μM). After treatment, cells were lysed in Tris·HCl (50 mM; pH 7.2), NaCl (150 mM), NaF (40 mM), EDTA (5 mM), EGTA (5 mM), Nonidet P-40 (NP-40, 1%), Triton X-100 (1%), sodium deoxycolate (0.1%), phenylmethylsulfonyl fluoride (PMSF, 1 mM), and protease inhibitor cocktail (Complete-EDTA-free tablets; Roche). Supernatants (10,000 revolutions/min for 8 s) of whole cell lysates were diluted in 2× Laemmli buffer containing 100 mM DTT and electrophoresed in 4–20% minigels. Total protein (50 μg) was loaded in each lane. Nitrocellulose membranes were blotted with rabbit anti-Src[pY418] phosphospecific antibodies (1:500 dilution; Biosource) for 2 h at room temperature, washed, and then incubated with goat anti-rabbit HRP-conjugated antibody (1:2,000; Santa Cruz Biotechnology). After detection of bands, membranes were reprobed using the rabbit anti-pan-src (1:500; Biosource). Quantification of activated Src was performed by measuring the ratio of src[Y418] to pan-Src bands after densitometric analysis using NIH J-image software.

Competition assays.

Confluent cultures of J774 cells grown in 100-mm dishes were lysed in a buffer solution containing 50 mM Tris·Cl (pH 7.2), 150 mM NaCl, 40 mM NaF, 5 mM EGTA, 1% NP-40, 0.1% sodium deoxycholate, 1 mM Na3VO4, 1% Triton X-100, 1 mM PMSF, and complete protease inhibitor (Roche). Lysates were precleared by incubation with IgG agarose-conjugated beads (UltraLink Immobilized Protein G; Pierce) for 120 min at 4°C and brief centrifugation. Precleared lysates were then used for competition assays. Typically, 100–200 μl of precleared lysates were incubated with different concentrations (0–300 μM) of the TAT-P2X7-peptide (SLHDSPPTPGQGGGYKKRRQRRR: 451P) for ∼15 min at 4°C, before immunoprecipitation assays using anti-P2X7 antibodies (see above).

Dye uptake.

J774 cells plated on glass-bottomed dishes (MatTek) were bathed for 5 min in a Ca2+-containing or Ca2+-free PBS (pH 7.4) containing the cell-impermeant (629 Da) dye YoPro-1 (5 μM; Invitrogen-Molecular Probes), and basal fluorescence intensity was measured from regions of interest located on cells, as previously described (43). Cells were then exposed to a solution containing 150–300 μM BzATP and 5 μM YoPro, and fluorescence intensity was measured following 5–10 min stimulation. (Higher concentrations of agonist than those in the electrophysiological studies are necessary for the optimal detection of YoPro fluorescence.) After background subtraction, YoPro fluorescence intensity was normalized to values obtained from cells bathed in solution containing 5 μM YoPro. The effect of gap junction channel blockers on BzATP-induced dye uptake was tested on cells preincubated with CBX (50 μM) and MFQ (10 nM), and results were compared with those obtained from untreated and Panx1 siRNA-treated macrophages. YoPro fluorescence was captured using an Orca-ER charge-coupled device camera attached to an inverted Nikon microscope equipped with a ×20 dry objective and 488-nm filter set using Metafluor software, as previously described (30, 46).

TAT-P2X7 peptides.

Membrane permeant peptides (SLHDSP[P/L]TPGQGGGYKKRRQRRR; 78–90% purity) corresponding to the Src homology 3 domain (SH3) of the BalbC and C57Bl6 mouse P2X7R COOH terminus were synthesized by GeneScript. For dye uptake assays, cells were treated for 15 min with 10 μM of the TAT-peptides at 37°C and then used in optical measurements of large dye molecules as described above. For competition assays, 10–300 μM TAT-P2X7 were added to whole cell lysates for 3 h at 4°C before immunoprecipitation (see above). The effects of TAT-P2X7 peptides on ATP-induced inward currents were also evaluated. For that, whole cell recording was performed on untreated and peptide-treated J774 cells using the patch-clamp technique and on oocytes using two-microelectrode voltage clamp as described above.


Electrophysiology and pharmacology of the P2X7R-Panx1 complex.

To elucidate discrepancies in the literature (30, 38, 46) regarding the effects of compounds known to block gap junction channels on Panx1 channels and to verify the molecular identity of the P2X7R permeabilization pore, we used the J774 macrophage cell line to evaluate the pharmacological and electrophysiological properties of the P2X7R-Panx1 complex.

Although there is no specific P2X7R agonist, a pharmacological characteristic of these receptors is their higher sensitivity to BzATP than to ATP (reviewed in Ref. 12). Accordingly, we found that BzATP at lower concentrations induced larger currents in these cells than higher concentrations of ATP when the membrane potential was held at −60 mV. After 5 s BzATP (50 μM) stimulation, a biphasic inward current was recorded with an initial small inward current followed by a larger one (Fig. 1A, top). After ATP (1 mM) application, similar currents were recorded (Fig. 1A, bottom); however, the amplitudes of currents induced by 1 mM ATP were smaller than those induced by a 20-fold lower concentration of BzATP (50 μM), as expected for the P2X7R.

Fig. 1.

Blockade of P2X7 receptor (P2X7R)-induced currents by gap junction channel blockers and a P2X7R antagonist. A: representative current traces induced by 5-s application of 2′(3′)-O-(4-benzoylbenzoyl)adenosine 5′-triphosphate (BzATP, 50 μM; traces on top) and ATP (1 mM; traces on bottom) are shown to be greatly reduced by carbenoxolone (CBX, 50 μM); after washout of CBX, both agonists induced currents of similar amplitudes as those recorded after the first application. For all electrophysiological experiments, membrane potential was held at −60 mV. Note that after 50 μM BzATP application two distinct currents were recorded (an early and smaller current followed by a larger one). Bar histograms (right) show the mean ± SE values obtained from 6 experiments. B: representative current traces induced by 5-s application of BzATP (50 μM; top and bottom) are shown to be greatly reduced by 0.1 mM (top) and 0.3 mM (bottom) flufenamic acid (FFA); after washout of FFA, currents induced by BzATP were of slightly lower amplitudes than those recorded after the first application. Bar histograms (right) show the mean ± SE values obtained from 4 experiments. C and D: representative current traces induced by 5-s application of BzATP (50 μM) are shown to be greatly reduced by mefloquine (MFQ; 100 nM; C) and totally prevented by KN-62 (1 μM; D); after washout of MFQ and KN-62, BzATP induced currents with similar amplitudes as those recorded after the first application. E: representative current traces induced by 5-s application of BzATP (50 μM) in cells untreated (control) and treated for 48 h with pannexin1 (Panx1) small-interfering RNA (siRNA). Bar histograms (right) show the mean ± SE values obtained from 5 experiments; *P < 0.05, **P < 0.01, and ***P < 0.001. P values were obtained by ANOVA followed by Newman Keuls' test.

To investigate the possibility that the slowly developing, large-amplitude current recorded following P2X7R stimulation was mediated by Panx1 channels, we applied three compounds that have been shown to affect Panx1 and some gap junction channels (CBX, FFA, and MFQ) and measured current amplitudes evoked by 50 μM BzATP while holding membrane potential at −60 mV. All three compounds greatly reduced agonist-induced currents (Fig. 1, A–C). Incubation (5 min) with 50 μM CBX reduced the amplitudes of BzATP-evoked currents by 87.5% (P < 0.01; n = 6 cells; Fig. 1A, top). The blockade by CBX was reversible; after 2 min washout, the amplitudes of currents induced by BzATP were not significantly different from those measured in the absence of CBX (Fig. 1A). Similarly, CBX (50 μM) reversibly reduced by 83.4% the amplitudes of ATP-induced currents (P < 0.01; n = 6 cells; Fig. 1A, bottom).

FFA at 100 μM reduced current amplitude by 50% (Fig. 1B; P < 0.05; n = 4 cells) and, at 300 μM FFA, BzATP-induced inward current was reduced by 80% (Fig. 1B, bottom; P < 0.001; n = 4 cells). The effect of FFA was also largely reversible following a 2-min wash (Fig. 1B).

The anti-malarial drug MFQ has been previously reported to block gap junction channels in a connexin-specific manner (14) and to block the uptake of fluorescent molecules in cells expressing the P2X7R (46). Incubation of J774 cells for 2 min with 1 μM MFQ reduced BzATP-induced current by 78% (Fig. 1C; P < 0.01; n = 5 cells). After 5–6 min MFQ washout, current amplitudes were restored to their original values (Fig. 1C).

The amplitudes of the initial small currents induced by 50 μM BzATP (51.5 ± 3.7 pA; n = 18 cells) were not significantly different (P > 0.5; ANOVA) from those recorded in the presence of CBX (44.3 ± 5.6 pA; n = 5 cells), FFA (53.5 ± 7.0 pA; n = 5 cells), and MFQ (48.2 ± 3.7 pA; n = 6 cells). Thus these results showing that CBX, FFA, and MFQ blocked the slowly developing large-amplitude current but not the fast-developing and smaller-amplitude current suggest that these two currents may be mediated by distinct molecular pathways: the small cation conductance due to opening of the P2X7R and the large conductance provided by Panx1 channels. To verify that the initial conductance insensitive to CBX, FFA, and MFQ was related to ion influx through the ionotropic receptor, we used KN-62, a known P2X7R antagonist (4). Two minute preincubation of J774 with 1 μM KN-62 completely blocked both components of the BzATP-induced currents (Fig. 1D; n = 4 cells). To further test for the involvement of Panx1 in the late response, we transfected cells with Panx1 siRNA overnight and tested responses 30 h later. Knockdown of Panx1 with siRNA reduced by 40% the second component of BzATP-induced current, without affecting the small cation conductance (control=51.4 ± 3.7, n = 18 cells; siRNA=48.5 ± 6.3 pA, n = 7 cells; P > 0.5, t-test) (Fig. 1E).

Together, these results strongly indicate that the initial conductance insensitive to CBX, FFA, and MFQ corresponds to the opening of P2X7R channels and that the later large-conductance current sensitive to these three compounds corresponds to opening of Panx1 channels. Moreover, we found that activation of P2X7R is necessary for the opening of Panx1 at negative (−60 mV) membrane potentials, given that blockade of P2X7R with KN-62 prevented Panx1 currents (Fig. 1D).

Besides being activated through P2 receptors (30, 31; see Fig. 1), Panx1 channels are voltage gated, opening at positive membrane potentials (9). Further support for the functional presence of Panx1 channels in J774 cells is shown in Fig. 2. As previously reported in J774 macrophages (38), we also found an increase in whole cell currents at voltages above +20 mV, consistent with the presence of Panx1 in this cell line (Fig. 2A). These outward currents shared a similar pharmacological profile to Panx1 currents when exogenously expressed in Xenopus oocytes (9); specifically, these outward currents were greatly attenuated by 50 μM CBX and by higher concentrations of FFA (0.3 mM). Moreover, we found that the anti-malarial drug MFQ (50–100 nM) but not the P2X7R antagonist KN-62 (1 μM) greatly attenuated voltage-induced Panx1 outward currents (Fig. 2A). [The remaining currents, not sensitive to CBX, FFA, and MFQ, are unlikely to be tetraethylammonium-sensitive potassium channels given that reversal potential, measured by applying voltage steps after a ramp protocol, was found to be very close to 0 mV (Erev = −3.0 mV, N = 3; data not shown)]. Further support that Panx1 is responsible for the outward currents sensitive to CBX, FFA, and MFQ was obtained from experiments using Panx1 siRNA. Treatment (48 h) of J774 cells with Panx1 siRNA reduced current amplitude by 35% (Fig. 2A).

Fig. 2.

Pharmacology of outward (Panx1) currents and of the permeabilization pore in J774 cells. A: top, examples of outward currents recorded from J774 cells in the absence (control) and presence of MFQ (50 nM), and after washout. Bottom, bar histograms showing the mean ± SE values of the fraction of current amplitudes at the end of the current ramp (+100 mV) relative to control recorded from J774 macrophages bathed in solution containing CBX (10 and 50 μM), FFA (0.1 and 0.3 mM), and MFQ (10, 50, and 100 nM) and the P2X7R antagonist KN-62 (1 μM). Note that all three compounds known to block gap junction channels but not the mouse P2X7R antagonist greatly reduced voltage-activated outward (Panx1) currents. The last two bars on the right show the mean ± SE values of the fraction of current amplitudes recorded from untreated and Panx1 siRNA-treated cells. Data are from 6 independent experiments. B: top, representative curves of the mean value of relative YoPro fluorescence intensity (F/F0) changes recorded from J774 cells (n = 100) untreated (control) and treated with Panx1 siRNA following 300 μM BzATP application. Bottom, bar histogram showing that CBX (50 μM), MFQ (10 nM), and Panx1 siRNA greatly reduce YoPro uptake in BzATP (300 μM)-stimulated J774 cells (n = 300–400 cells/group from 3 independent experiments) bathed in a Ca2+-containing solution. C: representative Western blots of immunoprecipitates using P2X7 antibodies showing that the expression levels of Panx1 present in the P2X7R complex in the J774 cell line are greatly reduced (62%) following 48 h transfection with Panx1 siRNA (left) and that the expression level of P2X7R present in the whole cell lysates used for immunoprecipitation (right) is not altered by Panx1 siRNA treatment compared with untreated cells. Three independent experiments were performed. D: representative fluorescence images of J774 cells showing a reduced number of cells loaded with YoPro following 30 h transfection with Panx1 siRNA compared with control cultures. Arrows illustrate YoPro-loaded cells. Bar: 185 μm. *P < 0.05, **P < 0.01, and ***P < 0.001. P values obtained using ANOVA followed by Newman-Keuls' test.

Evidence that Panx1 is involved in P2X7R-mediated membrane permeabilization was obtained by measuring YoPro uptake in J774 cells treated with Panx1 siRNA and with CBX and MFQ (Fig. 2B). Each of these treatments led to >90% decrease in YoPro uptake. That Panx1 siRNA effectively reduced (61.7 ± 0.003%; n = 3 independent experiments) Panx1 expression levels is shown in Fig. 2C. Because we performed immunoprecipitation assays using P2X7R antibodies to evaluate the levels of Panx1 knockdown, these results also indicate that Panx1 is present in the P2X7R complex, as previously shown (38). Figure 2D shows representative fluorescence images of YoPro uptake in untreated and Panx1 siRNA-treated J774 cells to illustrate that, besides reducing the amount of dye uptake, knockdown of Panx1 caused a 69.40 ± 0.08% (n = 3 experiments) decrease in the number of cells loaded with the dye in response to 300 μM BzATP.

P2X7R-Panx1 signaling pathway.

To gain insight into pathways that could mediate the opening of Panx1 channels following activation of P2X7R, we considered the possibility that the proline-rich region located at the COOH-terminal domain of the P2X7R, which has been previously reported to be an important site involved in cell permeabilization (schematic in Fig. 3A ; see Refs. 1 and 28), could be involved. To test for that possibility, we designed membrane-permeant TAT-fusion peptides spanning that domain (see materials and methods) and evaluated effects on membrane permeabilization and on BzATP-induced currents. As shown in Fig. 3B, J774 macrophages treated with 10 μM TAT-P2X7 peptide for 15 min, but not cells treated with heat-inactivated TAT-P2X7 peptide, showed a substantial decrease in the amount of YoPro uptake following 150 μM BzATP.

Fig. 3.

The proline-rich region of P2X7R COOH terminus contributes to Panx1 activation. A: membrane topology of the P2X7R showing the region of the SH3 domain that was used to generate the TAT-P2X7 peptide. B: bar histograms showing the mean ± SE values of relative YoPro fluorescence intensity in J774 cells untreated and treated for 15 min with 10 μM active 451P TAT-P2X7 and heat-inactivated peptides following 10 min stimulation with 300 μM BzATP. Dye uptake was measured from cells bathed in a Ca2+-free solution. Note that the active TAT-P2X7 peptide greatly attenuated BzATP-induced dye uptake and that heat-inactivated peptide had only a minor effect. Data are from 3–4 independent experiments. C: bar histograms of the mean ± SE values obtained for current amplitudes induced by ATP (5 mM) in untreated (control n = 9 cells) and 10 μM TAT-P2X7 peptide (n = 14 cells)-treated J774 macrophages. The two TAT-P2X7 peptides (451P and 451L) used have the same amino acid sequence, except at position 451 where in one case a proline (P) was substituted for a lysine (L). About 10 cells/group were used; membrane potential was held at −60 mV, and channels were activated using the ramp protocol. D: bar histograms of the mean ± SE values (n = 3–4 independent experiments) obtained for current amplitudes induced by ATP (500 μM) in untreated and 10 μM TAT-peptide-treated Xenopus oocytes coexpressing P2X7R and Panx1 (see text for details). Inset: example of currents recorded from oocytes. E: bar histograms of the mean ± SE values (n = 3–4 independent experiments) obtained for current amplitudes induced by ATP (500 μM) in untreated and 10 μM TAT-peptides-treated Xenopus oocytes expressing only the P2X7R. F: bar histograms showing the mean ± SE relative current values (test/control; n = 4 independent experiments) obtained from Xenopus oocytes expressing only Panx1. Current amplitudes induced by voltage steps (−30 to +60 mV) during and after addition of 10 μM TAT peptides were normalized to the values obtained before TAT peptide addition. Inset: representative recording of voltage-activated currents in oocytes expressing Panx1 showing current changes during and after exposure to TAT peptides. The same TAT-P2X7 peptides described in B were used. **P < 0.01 and ***P < 0.001. P values were obtained by ANOVA followed by Newman-Keuls' test.

Electrophysiological recordings performed on J774 cells indicated that the late currents induced by 5 mM ATP were attenuated when cells were exposed to the TAT-P2X7 peptides (Fig. 3C). However, due to the large variability in the amplitudes of these late currents induced in individual cells by high concentrations of ATP in TAT-P2X7 peptide untreated cells, no significant difference was obtained compared with treated cells. Intracellular localization of active and heat-inactivated TAT-P2X7 peptides was confirmed optically using fluorescein isothiocyanate-conjugated peptides [Supplemental Figs. S1 and S2; Supplemental material for this article is available online at the American Journal of Physiology: Cell Physiology website.].

To better control for such variability, which is likely related to differences in P2X7R and Panx1 expression levels in this macrophage cell line, we used Xenopus oocytes coexpressing P2X7R and Panx1 to evaluate the effects of TAT-P2X7 peptides. As shown in Fig. 3, D and E, at a −30-mV holding membrane potential, TAT-P2X7 peptides greatly reduced the amplitudes of ATP-induced inward currents in Xenopus oocytes coexpressing P2X7R and Panx1 (Fig. 3D; control=2.65 ± 0.26 μA; TAT-451P peptide=0.68 ± 0.13 μA; TAT-451L peptide=1.00 ± 0.18 μA; P < 0.01; n = 3 independent experiments) without affecting currents of oocytes expressing only P2X7R (Fig. 3E; control=0.085 ± 0.011 μA and TAT peptides=0.053 ± 0.005 μA; P > 0.05; n = 4 independent experiments). In oocytes expressing exclusively Panx1, TAT-P2X7 peptide did not decrease membrane conductance when voltage was stepped from −30 to +60 mV. Instead, during TAT-peptide application, oocyte membrane conductance increased 1.47 ± 0.13-fold from control levels and to 1.84 ± 0.20-fold after washout (Fig. 3F). At holding membrane potential (−30 mV), inward currents increased 20% from control after TAT-peptide washout (Fig. 3F, inset).

Together, the dye uptake and electrophysiological data suggest that the proline-rich segment, which contains an SH3, of the P2X7R COOH-terminus is likely to be involved in the signal transduction pathway that leads to Panx1 activation following receptor activation. Possible scenarios for such action could involve the interaction and/or the activation of an intermediate molecule, most likely a tyrosine kinase. To answer the question of whether Src-tyrosine kinase participates in this process, we performed competition assays in which TAT-P2X7 peptides were incubated at different concentrations with J774 whole cell lysates before immunoprecipitation with P2X7 antibodies followed by Western blots using anti-Src[pY418] antibodies. Our results revealed a dose-dependent reduction of Src[pY418] bound to the P2X7R complex with an increased concentration of TAT-P2X7 peptides (IC50 = 20 μM; Fig. 4A).

Fig. 4.

Src tyrosine kinase mediates activation of Panx1 through P2X7R in J774 macrophages. A: dose-dependence curve showing the inhibition of src[Y418] binding to the P2X7R by the TAT-P2X7 peptide (451P). The IC50 value (20 μM) was obtained after fitting a single exponential decay curve to the densitometric values obtained for src[Y418] bands obtained in Western blots following pull-downs using P2X7R antibodies of J774 whole cell lysates treated with different concentration of TAT-P2X7 peptide (n = 4 independent experiments). An example of a Western blot obtained from competition assays is shown at top. B: representative current traces induced by 5-s application of BzATP (50 μM) are shown to be reduced by the Src tyrosine kinase inhibitor PP2 (20 μM; 5 min preincubation); after washout of PP2, agonist induced currents of similar amplitudes as those recorded after the first application. Bar histograms show the mean ± SE values obtained from 5 experiments. C: bar histograms of the mean ± SE values of dye uptake induced by 150 μM BzATP, showing that the PP2 (10 μM; 5 min preincubation) significantly decreased agonist-induced YoPro uptake. Dye uptake was measured from cells bathed in a Ca2+-free solution. D: bar histograms of the mean ± SE ratio values of the active (src[Y418]) to total src obtained from Western blot analysis of J774 whole cell lysates (n = 4 independent experiments). Before cell lysis, cultures were treated for 15 min with BzATP (300 μM) in the absence and presence of 1 μM KN-62, 10 μM TAT-P2X7 peptide (451P), and 10 μM PP2; the ratio levels of active src to total src obtained from treated cells was compared with those obtained from BzATP-untreated, control (ctrl) whole cell lysates. Inset: example of Western blot obtained from J774 cells showing that increased amount of tyrosine-phosphorylated src (srcY) induced by agonist stimulation was prevented by KN-62, TAT-P2X7 peptide, and PP2. **P < 0.01. P values obtained using ANOVA followed by Newman-Keuls' test.

To test the hypothesis that a Src tyrosine kinase is an intermediate protein mediating the activation of Panx1 through the P2X7R, J774 cells were treated with the Src tyrosine kinase inhibitor PP2 (10 μM) for 5 min to evaluate BzATP-induced Panx1 currents and membrane permeabilization. As shown in Fig. 4, B and C, PP2 attenuated both BzATP-induced currents and YoPro uptake. Further evidence for the participation of Src in the signal transduction pathway was obtained by measuring the amount of tyrosine-phosphorylated Src (active) in J774 cells stimulated with 300 μM BzATP in the absence and presence of 1 μM KN-62, 10 μM TAT-P2X7 peptide and 10 μM PP2. As shown in Fig. 4D, blockade of P2X7R with KN-62 or interference with Src binding and activation significantly reduced BzATP-induced Src phosphorylation. Thus these results support the hypothesis that Src activation following P2X7R stimulation is likely involved in the process leading to the opening of Panx1 channels.


The membrane-permeabilizing action of ATP (formation of a transmembrane pathway for flux of moderately sized molecules) first described as mediated by the so-called purinergic P2Z receptor (20) was later identified as being due to P2X7R (33, 47). The physical relationship between the P2X7R cation channel and the nonselective pore was until very recently unclear. Several models for pore formation have been proposed to explain the somewhat variable degree of agonist-induced permeabilization among various cell types; these include the recruitment of monomeric channel subunits (12, 13, 48) and the recruitment of a lytic pore (34), which has been recently suggested to be Panx1 (30, 38).

Pannexins are a newly discovered (but phylogenetically old) group of proteins that share no sequence homology with the vertebrate gap junction proteins, connexins, and a low but significant homology with the nonchordate gap junction proteins, the innexins (5, 35, 36, 52). When expressed in Xenopus oocytes, Panx1 and Panx2 have been reported to form gap junction channels to some extent and also to function as gap junction hemichannels that are sensitive to compounds known to block gap junction channels, including CBX and FFA (9, 10). Gap junction channel formation by pannexins in mammalian cells remains to be conclusively demonstrated (16) and is unlikely due to the presence of a glycosylation site in the extracellular loop of this protein (7, 8, 37). However, Panx1 readily forms nonjunctional channels (pannexons) that are voltage sensitive and large conductance (400–500 pS) and have been reported to be modulated by intracellular calcium and by mechanical stretch (3, 31); these large-conductance channels have been proposed to mediate ATP release from erythrocytes and mouse taste buds (22, 29) and to underlie dye uptake and ATP release formerly attributed to connexin43 hemichannels (3; reviewed in Ref. 43).

Our previous studies performed on Xenopus oocytes indicated that Panx1 channels are likely to form the permeabilizing pore of the P2X7R death complex (30). With the use of several cell types, including the J774 macrophage cell line, a previous study (38) provided evidence that P2X7R and Panx1 were coimmunoprecipitated; they also showed that knocking down Panx1 expression by siRNA and blocking Panx1 channels using either Panx1 mimetic peptides or the gap junction channel blocker CBX prevented P2X7R-induced uptake of large dye molecules (38). However, in this report (38), none of these treatments affected ATP-induced currents. This is an unexpected finding given that Panx1 forms large-conductance channels (400–500 pS) that are blocked by CBX (3, 9, 30). Thus it would be expected that Panx1 would contribute significantly to the amplitudes of ATP-induced currents.

In agreement with previous work (38), we found that, in J774 macrophages, Panx1 is associated with the P2X7R (Fig. 2C) and that activation of Panx1 channels by means of P2X7R induces uptake of YoPro (Fig. 2B). Contrary to previous findings (38), however, but consistent with our electrophysiological recordings, we found that both CBX and MFQ, which are shown to reduce P2X7R inward currents, prevented YoPro uptake (Figs. 1 and 2). Thus these results indicate that uptake of large dye molecules is associated with the opening of large-conductance channels provided by Panx1.

The pharmacological results described in the present report using the J774 macrophage cell line are in agreement with our previous studies (46) showing that several compounds known to block gap junction channels (CBX, MFQ, FFA, heptanol, and octanol) prevent P2X7R-induced influx of dye molecules in astrocytoma cells. However, our studies are substantially different from those reported by Pelegrin and Surprenant (38); these authors described that, of the three compounds (CBX, MFQ, and heptanol) that they employed, only CBX prevented P2X7-induced dye uptake in macrophages and astrocytoma cells. Such contradiction is likely related to the fact that these compounds were used (38) at concentrations that are either not effective or nonselective in blocking P2X7R-induced dye uptake (46), as in the case of heptanol (30-fold lower) and of MFQ (1,000-fold higher), respectively.

Early studies (6) proposed that the gap junction protein connexin43, which is expressed in the J774 cell line (2, 6, 19), was the permeabilization pore induced by ATP. Such possibility, however, was later discarded with the cloning of the P2X7R (47) and the finding that peritoneal macrophages derived from connexin43-null mice still displayed ATP-induced permeabilization (2). The pharmacological discrimination between pannexin and connexin channels can be problematic due to the fact that both are blocked by similar compounds, although at slightly different concentration, as is the case of CBX and FFA (9). However, here we found that MFQ is likely to be a more reliable compound to distinguish between Panx1 and connexin43 channels; our results showed that 100 nM is sufficient to block Panx1 channels, whereas total blockade of connexin43 gap junction channels requires 30 μM MFQ (14).

In the present study, we have not used “pannexin mimetic peptides” as a diagnostic tool for channel identification. Although such peptides might be expected to affect channels in a sequence-specific way, it has been shown recently that their inhibitory action is nonspecific. Pannexin mimetic peptides, besides inhibiting Panx1 channels, also inhibited the completely unrelated connexin46 channels with similar efficacy (50).

Very little is known about the signaling mechanisms by which activation of P2X7R leads to opening of Panx1 channels. Although influx of calcium has been proposed as a necessary step leading to activation of the P2X7R permeabilization pore (18), it is unlikely that this cation is the sole messenger given that membrane permeabilization was not found to be prevented by chelating intracellular Ca2+ in macrophages (15) or by removing extracellular Ca2+ from oocytes expressing P2X7R and Panx1 (30).

The P2X7R is unique among members of the ionotropic P2XR family because it has a long COOH terminus (∼32 kDa) that provides sites for protein-protein interactions (25, 26). The diversity of molecules found to interact with P2X7R indicates that these receptors are likely to mediate distinct functions. Binding of P2X7R to laminin, actin, actinin, integrins, phosphatidylinositol 4-kinase, and receptor protein phosphatase has been reported previously (25); however, there are no reports on the functional significance of such associations. Previous studies (27, 42, 47) have indicated that deletion of the COOH-terminal domain of the P2X7R prevented membrane permeabilization, suggesting that this domain is important for the signal transduction events leading to membrane permeabilization. Analysis of functional consequences of P2X7R polymorphisms has identified a single amino acid mutation (P451L) in the COOH-terminal domain of the rodent P2X7Rs that affected the ability of these receptors to induce membrane permeabilization in immune cells (1, 28).

Based on these reports, we hypothesized that the P2X7R COOH-terminal tail spanning this point mutation, which is also part of the death signal domain (17), is involved in the interaction between P2X7R and Panx1. Our results using membrane-permeant peptides spanning the proline-rich region of the COOH terminus of the P2X7R indicate that this SH3 domain containing the 451-point mutation is necessary to induce influx of large molecules through Panx1 channels. Moreover, these peptides (a decoy for an SH3 binding protein) are here shown to prevent the association between P2X7R and src tyrosine kinase, suggesting that this kinase is likely involved in the initial steps of the signal transduction pathway that leads to membrane permeabilization. This is in agreement with previous studies indicating that tyrosine kinases play a pivotal role in P2X7R-mediated cellular responses (23, 252; but see 15).

Our results thus support a mechanistic hypothesis whereby an Src tyrosine kinase is involved in the initial steps leading to the activation of Panx1 following P2X7R stimulation. Indeed, results presented here showing that an Src tyrosine kinase inhibitor, PP2, greatly attenuated BzATP-induced current and dye uptake in J774 cells provide support for phosphorylation-mediated activation of Panx1 channels. Such a tyrosine kinase-mediated transduction mechanism could either involve the direct tyrosine phosphorylation of Panx1 or the activation of another molecule. Database search (http://www.cbs.dtu.dk/services/NetPhos/) for a tyrosine residue on the cytoplasmic domain of Panx1 indicated that Y324 is a likely candidate. Evaluation of tyrosine phosphorylation of Panx1 performed on Western blots derived from immunoprecipitates (with anti-phosphotyrosine antibodies) of whole cell lysates of J774 cells that were treated with BzATP in the absence and presence of KN-62, TAT-P2X7 peptide, and PP2 did not reveal that Panx1 is tyrosine phosphorylated (unpublished observations). Although these results strongly suggest that activation of Panx1 channels through the P2X7R involves the phosphorylation of another molecule, the identity of such molecule remains to be determined.

In conclusion, our studies favor the hypothesis that Panx1 is most likely the permeation pore induced by the P2X7R in the J774 macrophage cell line and that a Src tyrosine kinase is one of the molecules that is involved in the transduction of ATP signals into membrane permeabilization.


This work was supported by National Institutes of Health Grants NS-41023 to E. Scemes, NS-041282 to D. C. Spray, and GM-48610 to G. Dahl, an American Heart Association graduate fellowship (to S. Locovei), and the Fogarty International Training Program (Fellowship to A. P. Alberto).


We appreciate the technical assistance of Marcia Urban-Maldonado and Melissa Aleksey.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract