P2X receptors are ATP-gated nonselective cation channels highly permeable to calcium that contribute to nociception and inflammatory responses. The P2X4 subtype, upregulated in activated microglia, is thought to play a critical role in the development of tactile allodynia following peripheral nerve injury. Posttranslational regulation of P2X4 function is crucial to the cellular mechanisms of neuropathic pain, however it remains poorly understood. Here, we show that the phosphoinositides PI(4,5)P2 (PIP2) and PI(3,4,5)P3 (PIP3), products of phosphorylation by wortmannin-sensitive phosphatidylinositol 4-kinases and phosphatidylinositol 3-kinases, can modulate the function of native and recombinant P2X4 receptor channels. In BV-2 microglial cells, depleting the intracellular levels of PIP2 and PIP3 with wortmannin significantly decreased P2X4 current amplitude and P2X4-mediated calcium entry measured in patch clamp recordings and ratiometric ion imaging, respectively. Wortmannin-induced depletion of phosphoinositides in Xenopus oocytes decreased the current amplitude of P2X4 responses by converting ATP into a partial agonist. It also decreased their recovery from desensitization and affected their kinetics. Injection of phosphoinositides in wortmannin-treated oocytes reversed these effects and application of PIP2 on excised inside-out macropatches rescued P2X4 currents from rundown. Moreover, we report the direct interaction of phospholipids with the proximal C-terminal domain of P2X4 subunit (Cys360–Val375) using an in vitro binding assay. These results demonstrate novel regulatory roles of the major signaling phosphoinositides PIP2 and PIP3 on P2X4 function through direct channel–lipid interactions.
Mammalian P2X receptors are ATP-gated nonselective cation channels, generated from the trimeric assembly of homologous subunits (P2X1–7). Each subunit is composed of two transmembrane domains connected by a large ectodomain, with both N and C termini facing the intracellular side of the plasma membrane (North, 2002). In the nervous system, P2X receptor channels are expressed at the plasma membrane of neurons, astrocytes, and microglia, where they contribute to intercellular purinergic communication.
The P2X4 subunit forms a functional homomeric channel with high permeability to calcium ions (Egan and Khakh, 2004). It can also heteromerize with P2X1 (Nicke et al., 2005), P2X6 (Lê et al., 1998), or P2X7 subunits (Guo et al., 2007). It is expressed at high levels in central neurons and in microglial cells (Lê et al., 1998; Rubio and Soto, 2001; Tsuda et al., 2003). In recent years, microglial P2X4 receptor channels have been found to play a critical role in the development of neuropathic pain induced by peripheral nerve injury (Tsuda et al., 2003; Inoue et al., 2004; Trang et al., 2006). The upregulation of P2X4 expression in activated microglia following nerve injury is critical to the development of tactile allodynia (Tsuda et al., 2003) via the release of microglial BDNF and KCC2 cotransporter-mediated disinhibition of dorsal horn neurons (Coull et al., 2003, 2005). Despite the involvement of the P2X4 receptor subtype in the pathogenesis of chronic pain, very little is known about its posttranslational regulation.
A lot of attention has been brought to phosphoinositides as modulators of membrane proteins. These phospholipids are produced by the combinatorial action of selective lipid kinases, namely phosphoinositol 3-kinase (PI3K), phosphoinositol 4-kinase (PI4K) and phosphoinositol 5-kinase, which phosphorylate the inositol ring at the 3′, 4′, or 5′ positions. Synthesis of the major phosphoinositide PI(4,5)P2 (PIP2) is completed through successive phosphorylations by PI4K and phosphoinositol(4)5-kinase. Formation of PI(3,4,5)P3 (PIP3) from PIP2 necessitates one additional phosphorylation by PI3K. Phosphoinositides are regulators of several important cellular processes, such as vesicle trafficking, cytoskeleton remodeling, and ion channel function (Hilgemann et al., 2001; Suh and Hille, 2005). Interestingly, it has been shown that P2X1, P2X2, and P2X7 receptors are positively modulated by different phosphoinositides (Fujiwara and Kubo, 2006; Zhao et al., 2007; Bernier et al., 2008).
Phosphoinositides can interact with membrane proteins via electrostatic binding to positively charged residues (Rosenhouse-Dantsker and Logothetis, 2007). The P2X4 subunit contains a phosphoinositide-binding motif characterized by a cluster of lysine and arginine residues in the intracellular C-terminal domain in close proximity to its second transmembrane domain. In this study, we used electrophysiological techniques to elucidate the role major phosphoinositides play in the modulation of microglial and recombinant P2X4 receptor channel function. We also investigated the impact of phosphoinositides on P2X4-mediated calcium entry and the structural characteristics of their interactions with P2X4 using a lipid binding assay.
Materials and Methods
Patch-clamp recordings in BV-2 cells.
Mouse BV-2 microglial cells were a kind gift from Dr. Jean Labrecque (AstraZeneca R&D) and were routinely maintained at 37°C and 5% CO2 in DMEM supplemented with 5% heat-inactivated FBS, 10 mm pyruvate (Sigma), 5% l-glutamine (Invitrogen), penicillin, and streptomycin. Whole-cell patch-clamp recordings of BV-2 cells (Vhold = −60 mV) were performed using pipettes filled with internal solution, pH 7.2, containing (in mm) 120 K-gluconate, 1 MgCl2, 4 NaOH, and 10 HEPES. The normal recording solution, pH 7.4, comprised (in mm) 130 NaCl, 5 NaOH, 3 KCl, 1 MgCl2, 2 CaCl2, and 10 HEPES. Low divalent recording solution contained 0.3 mm CaCl2 without Mg2+. Membrane currents were recorded using an Axopatch 200B amplifier and digitized at 500 Hz with a Digidata 1330 interface (Axon Instruments). Drugs were dissolved in recording solution and applied using a SF-77B fast perfusion system (Warner Instruments) at a rate of 1 ml/min. All experiments were performed at room temperature.
Ratiometric measurement of [Ca2+]i in BV-2 cells.
BV-2 cells were loaded with Ca2+-sensitive fluorescent dye fura-2 AM (5 μm, Molecular Probes) in Ringer's solution (in mm: 130 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES; 37°C, pH 7.4) with 1% BSA for 30 min at 37°C in the dark. Cells were washed three times to remove extracellular fura-2 AM and left for at least 30 min for complete hydrolysation of acetoxymethyl ester before imaging. Fluorescence measurement of Ca2+ in individual cells were performed on an inverted TE2000-U microscope (Nikon) equipped with 40× oil-immersion objective [CFI super(S) fluor, Nikon]. Fura-2 AM was excited alternatively at 340 and 380 nm with a Lambda 10-2 filter wheel (Sutter Instrument) coupled to a xenon arc lamp (XPS-100, Nikon). The fluorescent emission was monitored at 510 nm and recorded by a high-resolution cooled CCD camera (Cool Snap-HQ, Roper Scientific/Photometrics) interfaced to a Pentium III PC. Pairs of 340 and 380 nm images were acquired every second and ratio images were calculated using Metafluor 7.0 software (Molecular Devices). The experiment was performed at room temperature. Cells were superfused with Ringer's solution and drugs at a flow rate of 1 ml/min. Results were expressed as the 340/380 nm emission ratio proportional to the intracellular calcium. The magnitude of responses to ATP was calculated as the difference between the peak of the response and the basal ratio (Δ ratio). To test the viability of the cells, they were challenged with 1 μm ionomycin at the end of each recording. In each experiment, all conditions were tested (control, ivermectin, and ivermectin plus wortmannin) and data were collected from 3 independent experiments.
Two-electrode voltage clamp recordings in Xenopus oocytes.
Oocytes were surgically removed from Tricaine-anesthetized female Xenopus laevis frogs and were incubated in OR2 solution containing 1–2 mg/ml type IA collagenase at room temperature for 2 h under vigorous agitation. Stage V and VI oocytes were then manually defolliculated before intranuclear microinjection of plasmid DNA coding for rP2X4 (3 ng). After injection, oocytes were incubated in Barth's solution containing 1.8 mm CaCl2 at 19°C for 24 to 72 h before electrophysiological recordings. Two-electrode voltage-clamp recordings (Vhold = −60 mV) were performed using glass pipettes (1–3 MΩ) filled with 3 m KCl solution. Oocytes were placed in a recording chamber and perfused at a flow rate of 10 to 12 ml/min with Ringer's solution, pH 7.4, containing (in mm) 115 NaCl, 5 NaOH, 2.5 KCl, 1.8 CaCl2, and 10 HEPES. Membrane currents (d.c., 1 kHz) were recorded using a Warner OC-725B amplifier (Warner Instrument) and digitized at 500 Hz. Agonists were dissolved in the perfusion solution and applied using a computer-driven valve system. All series of recordings consisted of three successive applications of ATP 100 μm, with a 4 min washout with Ringer's solution between each application.
Macropatch recordings in Xenopus oocytes.
Inside-out macropatch experiments were performed using an Axopatch 200B amplifier (Axon Instruments). PClamp 9 was used for data acquisition and analysis. The vitelline membrane of Xenopus oocytes was removed using fine forceps before recordings. Electrodes with resistance of 0.5–1.0 MΩ were used. The currents were evoked by ramps from +100 mV to −100 mV. The time course of the currents at −80 mV was analyzed. The pipette solutions contained 100 μm ATP and dioctanoyl (diC8)-PIP2 (5 μm) was dissolved in the bath solution. The solutions were applied through a gravity-driven perfusion system. For each experiment, a minimum of two batches of oocytes were tested.
Lipid binding assay.
Oligonucleotides coding for amino acid sequences proximal to the C terminus (C360–V375) or the N terminus (F12–V29) of P2X4 were inserted into the pGEX-2T vector (New England Biolabs) for the production of the corresponding recombinant GST fusion proteins. The lipid binding analysis of the GST-fusion proteins was conducted using lipid coated hydrophobic membranes (PIP Strips, Echelon Biosciences). The membranes were first blocked with TBS+T solution supplemented with 3% BSA for 1 h at room temperature. The membranes were then incubated overnight in TBS+T with 3% BSA and 1 μg/ml GST fusion protein. The membranes were then washed with TBS+T six times, 5 min per wash. The primary antibody (mouse anti-GST, 1:1000) was then added in TBS+T, 3% BSA solution for 1 h. Washes with TBS+T were repeated, followed by incubation with the secondary antibody (goat anti-mouse HRP, 1:5000) in TBS+T, 3% BSA. The membranes were washed again in TBS+T, then bound proteins were detected in ECL.
Peak currents, defined as the maximal amplitude recorded during agonist application, were measured for comparative analysis. For recovery experiments, the amplitude of the third response was compared with the first and expressed as a percentage. The results obtained after a 2 h drug incubation were always compared with those obtained after a 2 h control incubation in Barth's solution containing vehicle (100 nm or 35 μm DMSO). For current kinetics experiments, activation rate was measured as the rise time (in seconds) from 10% to 90% of the peak amplitude. For desensitization rate, the time constant was measured with a one-exponential fit of the inactivation phase of the current. Data are presented as mean ± SEM and analyzed using Student's t test or one-way ANOVA followed by a Bonferroni's multiple comparison test. Normalized data were analyzed using nonparametric Mann–Whitney test.
Depletion of phosphoinositides decreases P2X4 current density in BV-2 microglial cells
To assess the role phosphoinositides might play on native P2X4 currents, we tested the effect of blocking PIP2 and/or PIP3 synthesis with wortmannin in BV-2 microglial cells. The furanosteroid wortmannin has been extensively used to block key lipid kinases in the phosphoinositides synthesis pathways. It is a potent inhibitor of PI3K at nanomolar concentrations while it blocks also PI4K at micromolar concentrations (Nakanishi et al., 1995; Vanhaesebroeck et al., 2001). Murine BV-2 cells express endogenous P2X4 and P2X7 channels (Raouf et al., 2007). To isolate P2X4 currents from P2X7 currents, we took advantage of the differential agonist sensitivity of the mouse P2X4 and P2X7 subtypes. Mouse P2X7 receptors exhibit a markedly lower sensitivity to the agonist ATP than the mouse P2X4 subtype (ATP EC50 >500 μm and <10 μm respectively) (Townsend-Nicholson et al., 1999). Ivermectin is an allosteric modulator of P2X4 receptors that enhances currents by stabilizing the agonist-induced open state (Priel and Silberberg, 2004). Therefore, ivermectin provides an additional tool to isolate P2X4 currents from microglial cells. Application of 100 μm ATP evoked ivermectin-potentiated inward currents typical of the P2X4 subtype (Fig. 1A). Vehicle-treated cells showed an ATP-evoked current density of 1.7 ± 0.3 pA/pF, while current density following 3 μm ivermectin treatment was increased to 13.9 ± 2.2 pA/pF (p < 0.0001, n = 9–10) (Fig. 1A,B). Depletion of PIP3 levels with 100 nm wortmannin (PI3K blockade) for 2 h led to a significant decrease of ivermectin-potentiated P2X4 current density (6.3 ± 0.6 pA/pF; p = 0.0096, n = 7). A comparable decrease was observed by blocking both PI3K and PI4K with 35 μm wortmannin for 2 h when compared with ivermectin alone (4.2 ± 0.9 pA/pF; p = 0.0025, n = 7).
P2X4-mediated calcium entry is sensitive to depletion of phosphoinositides
Taking advantage of the fact that P2X4 receptor channels are highly calcium permeable (Egan and Khakh, 2004), we used ratiometric fura-2 fluorescence imaging to measure changes in intracellular calcium concentration ([Ca2+]i) induced by ATP in BV-2 cells. Application of 100 μm ATP for 30 s induced a calcium response that was evaluated as the difference between the fluorescence ratio at the peak of the response and the basal ratio (Δ ratio) (Fig. 2A). In 77 cells recorded in 3 independent experiments, the increase of [Ca2+]i corresponded to a mean Δ ratio of 0.24 ± 0.02 (Fig. 2A,D). Applications of 1 μm ionomycin evoked typical and reproducible rises of [Ca2+]i. Given that ivermectin selectively enhanced ATP-evoked P2X4-current densities, we then examined the sensitivity of ATP-induced Ca2+ ion entry to ivermectin. Two minutes pretreatment with 3 μm ivermectin strongly potentiated the calcium response to 100 μm ATP (Δ ratio = 0.55 ± 0.05, n = 71) (Fig. 2B,D). The rise of fluorescence ratio induced by ATP was increased by 229%. Depletion of PIP2 and PIP3 levels by preincubation with 35 μm wortmannin (PI3K and PI4K blockade) for 10 min significantly decreased the P2X4-mediated calcium entry (Δ ratio: 0.32 ± 0.04, n = 45) (Fig. 2C,D). The responses to ionomycin were similar in BV-2 cells exposed to wortmannin plus ivermectin compared with those exposed to ivermectin alone (Δ ratio = 0.24 ± 0.02, n = 45 and 0.28 ± 0.02, n = 71, respectively). Using published calibration methods (Grynkiewicz et al., 1985), we estimated the rise of P2X4-mediated [Ca2+]i at 240 ± 10 nm for vehicle-treated cells, 650 ± 70 nm for cells treated with 3 μm ivermectin, and 360 ± 30 nm for cells treated with both 3 μm ivermectin and 35 μm wortmannin. Altogether, these results show that disrupting PIP2 and PIP3 levels significantly decrease the capacity of P2X4 channels to mediate calcium entry into microglial cells.
P2X4 current amplitude is decreased by wortmannin treatment in Xenopus oocytes
To further investigate the modulation of P2X4 currents by phosphoinositides, we conducted experiments on Xenopus oocytes expressing recombinant P2X4 receptor channels. Oocytes were incubated for 2 h in wortmannin at 100 nm (PI3K blockade), or at 35 μm (PI3K and PI4K blockade). The amplitude of the first evoked P2X4 current was recorded (Fig. 3). The average current amplitude in control conditions was 4.2 ± 0.6 μA (n = 6). As shown in Figure 3, A and B, 100 nm wortmannin induced a significant decrease of ∼50% in the P2X4 current amplitude, down to 2.1 ± 0.6 μA (p = 0.0347, n = 6), whereas 35 μm wortmannin led to an even greater inhibition of P2X4 current amplitude (0.8 ± 0.4 μA; p = 0.0007, n = 7). Dose–response curves confirmed a fourfold decrease in maximal current amplitudes after treatment with wortmannin (Fig. 3C). They also revealed a small albeit significant increase in the EC50 of ATP (18.7 ± 1.8 μm in vehicle-treated oocytes, 28.7 ± 2.4 μm in wortmannin-treated oocytes, n = 4–9, *p < 0.05).
Impact of phosphoinositides on the speed of recovery of P2X4 currents
We then tested whether the depletion of PIP2 or PIP3 could affect the speed of recovery of P2X4-mediated currents. We activated the receptor with three successive applications of 100 μm ATP at 4 min intervals, and measured the amplitude of the third current relative to the first current response, defined as the third/first recovery ratio. Under control conditions, a 4 min washout between each agonist application was enough for the receptor to fully recover: the third/first ratio was of 100 ± 8% (n = 7) (Fig. 4A). Blocking PIP3 synthesis with 100 nm wortmannin preincubation led to a gradual decrease of the currents, the amplitude of the third currents being 66 ± 11% of that of the initial current (Fig. 4B). When both PIP2 and PIP3 levels were depleted with 35 μm wortmannin preincubation, the third current went down to 26 ± 5% of the first one (Fig. 4C). The third/first ratios under both 100 nm and 35 μm wortmannin were significantly smaller than under control conditions (p = 0.0251, n = 6 and p < 0.0001, n = 7, respectively).
Injection of phosphoinositides rescues P2X4 current amplitude in wortmannin-treated oocytes
To test the selective involvement of phosphoinositides in the modulation of P2X4 channel responses by wortmannin, we conducted experiments in which diC8-PIP2 or diC8-PIP3 were injected into the oocyte cytoplasm. Initial channel activation was recorded, followed by injection of vehicle (PBS) or phosphoinositides, to reach an intracellular concentration of 200 μm. The subsequent current activation was then recorded and compared with preinjection amplitude. Notably, under basal conditions, injection of either diC8-PIP2 or diC8-PIP3 did not lead to any change in the P2X4 current amplitude (data not shown), suggesting that PIP2 and PIP3 levels are saturating under basal conditions. However, injection of phosphoinositides rescued P2X4 current amplitudes in conditions of wortmannin-induced depletion (Fig. 5A,B). As shown in Figure 5C, diC8-PIP3 injection induced a 396 ± 64% increase in P2X4 (p = 0.012, n = 4), whereas diC8-PIP2 injection led to a 827 ± 244% increase (p = 0.006, n = 4).
P2X4 current kinetics is sensitive to PIP2 levels
To assess changes in P2X4 current kinetics in Xenopus oocytes, we examined the activation rate, calculated as the 10–90% rise time of the onset current, and the desensitization rate calculated as the mono-exponential fit of the inactivation current phase (Fig. 6A). Preincubation with 35 μm wortmannin (PI3K and PI4K blockade) induced an increase of the 10–90% rise time (1.15 ± 0.1 s, p = 0.0068, n = 7) compared with vehicle-treated oocytes (0.72 ± 0.08 s, n = 10) indicative of a slower activation rate of P2X4 channels (Fig. 6B,C). The 10–90% rise time was not affected by treatment with 100 nm wortmannin (PI3K blockade) (0.59 ± 0.06 s, n = 7). The same kinetics parameter was studied at 35 μm wortmannin followed by injection of diC8-PIP2 or vehicle. Under these conditions, cytoplasmic diC8-PIP2 injection led to a faster activation rate compared with vehicle-injected oocytes, with respective rise times of 0.90 ± 0.07 s and 1.61 ± 0.28 s respectively (p = 0.0263, n = 13–14).
The desensitization rate of P2X4-mediated currents was studied under similar experimental conditions. Vehicle-treated oocytes showed a desensitization rate of 6.73 ± 0.85 s, whereas preincubation with 35 μm wortmannin slowed it significantly to 15.76 ± 2.8 s (p = 0.0089, n = 5–6). Under these conditions, cytoplasmic diC8-PIP2 injection led to a faster desensitization rate compared with that of vehicle-injected oocytes, with inactivation times of 5.15 ± 0.34 s and 9.17 ± 0.59 s respectively (p < 0.0001, n = 13–17). A treatment with 100 nm wortmannin did not modify the desensitization rate of P2X4 currents (6.23 ± 1.96 s, n = 6).
PIP2 activates P2X4 current responses in excised inside-out macropatches
Excised inside-out patch recordings were performed to test the direct effects of phosphoinositides on heterologously expressed P2X4 channels. Upon patch excision, a fast rundown of the current was observed. Application of 100 μm polylysine, which sequesters endogenous anionic phospholipids, induced a further rundown of the residual current. Subsequent application of 5 μm diC8-PIP2 to the intracellular side of the membrane activated P2X4 currents (Fig. 7A,B). A quantitative analysis shows a significantly larger P2X4 current amplitude (∼1.7-fold increase, p = 0.0059) in the presence of diC8-PIP2 (32 ± 6 pA, n = 7) compared with rundown baseline levels (19 ± 4 pA, n = 7).
A proximal C-terminal sequence of P2X4 subunit is critical for phosphoinositide binding
Phosphoinositides are negatively charged lipids so we investigated potential interactions with basic residues on the intracellular domain of the P2X4 channel subunit, using a lipid strip assay. The proximal region of the C-terminal domain displays a cluster of basic amino acids potentially involved in the interaction with phosphoinositides. A P2X4 peptide containing the sequence C360–V375 (Fig. 8A) was expressed as a GST-fusion protein, whereas GST protein alone was used as negative control for the lipid binding assay. No specific binding was observed when the GST protein alone or GST fused with the N-terminal sequence F12–V29 was tested (Fig. 8B). However the GST–P2X4 (C360–V375) construct selectively bound to several phosphatidylinositides including PI(3)P, PI(4)P, PI(5)P, PI(3,4)P2, PI(3,5)P2, PIP2, and PIP3. It also showed strong binding to phosphatidylserine and phosphatidic acid. The neutral lipids phosphatidylethanolamine and phosphatidylcholine did not display any affinity for the GST–P2X4 (C360–V375) construct.
We demonstrate here a specific modulation of P2X4 receptor function by phosphoinositides. This modulation is physiologically relevant as both native P2X4 currents and P2X4-mediated calcium entry are sensitive to PIP2 and PIP3 levels in microglial cells. Microglia are immunocompetent cells that respond to adverse conditions in the CNS, such as neurodegeneration, trauma, ischemia, inflammation, and infection. Increasing evidence indicate that microglia also play a critical role in the pathogenesis of neuropathic pain, a debilitating chronic condition that can occur after peripheral nerve damage. After nerve injury, microglia located in the ipsilateral dorsal horn of spinal cord become activated and upregulate the expression of P2X4 subunit (Tsuda et al., 2003). Recent findings suggest that ATP-induced activation of microglia evokes the release of BDNF and downregulation of the neuronal cotransporter KCC2, leading to postsynaptic changes in the electrochemical gradient of chloride ions and disinhibition (Coull et al., 2003, 2005). BV-2 cells have been established as a model of murine microglia. Native microglial cultures have serious limitations including short survival time ex vivo, heterogeneity in activation states (Lawson et al., 1990; Ren et al., 1999) and the process of isolating primary microglia can result in phenotypical changes (Frank et al., 2006). P2X4 and P2X7 are the major P2X subtypes expressed in BV-2 cells (Raouf et al., 2007). We successfully isolated P2X4 currents from P2X7 currents by using a dual pharmacological strategy based on the higher sensitivity of mouse P2X4 to ATP compared with that of mouse P2X7 (Townsend-Nicholson et al., 1999; Raouf et al., 2007) and on the use of ivermectin, a selective allosteric potentiator of P2X4-containing receptors (Khakh et al., 1999; Toulmé et al., 2006). Our patch-clamp recordings data showed reduced P2X4 current density after blockade of PI3K and PI4K with wortmannin at 100 nm and 35 μm, respectively. The ATP-induced microglial release of the proalgesic cytokines TNF-α and interleukin-6 involve Ca2+-dependent pathways (Hide et al., 2000; Shigemoto-Mogami et al., 2001; Hanisch, 2002). P2X4 receptor channels are highly permeable to Ca2+ ions (Egan and Khakh, 2004) and we used this property to measure in ratiometric imaging a significant reduction of ATP-induced ivermectin-sensitive Ca2+ ion entry after treatment with wortmannin. Altogether, our results show that both PIP2 and PIP3 levels regulate the activity of native P2X4 receptor channels in BV-2 microglial cells. It is possible that a large Ca2+ influx through P2X4 receptor channels could activate Ca2+-dependent isoforms of phospholipase C, leading to decreased PIP2 levels and subsequent decreased channel responses. Interestingly, such a mechanism of Ca2+-induced inactivation of PIP2-sensitive channels through activation of phospholipase C has been described recently for TRP channels (Rohács et al., 2005; Thyagarajan et al., 2008).
The modulation of P2X4 responses by phosphoinositides does not depend on a cell-specific context since we reconstituted it in the Xenopus oocyte expression system. Wortmannin-induced phosphoinositides depletion in oocytes led to reduced maximal P2X4 currents, lower sensitivity to ATP, longer recovery times following repeated agonist applications, as well as slower current activation and inactivation kinetics. We conclude that ATP becomes a partial agonist for P2X4 when PIP2 and PIP3 levels are depleted, i.e., phosphoinositides are required cofactors for the full expression of P2X4 function. Treatment with wortmannin at micromolar concentrations could potentially affect many intracellular pathways. However we validated the selective blockade of phosphoinositides synthesis by normalizing P2X4 current amplitude as well as kinetics with cytoplasmic injections of PIP2 or PIP3. Interestingly, depleting PIP3 levels has an impact on the amplitude of P2X4 currents and on their recovery time, but not on their kinetics of activation or inactivation which are sensitive to PIP2 only. Therefore P2X4 channels are sensitive to both PIP2 and PIP3 albeit through different mechanisms of regulation.
Although earlier studies conducted on the modulation of ion channels by phosphoinositides showed a predominant role of PIP2 as a signaling molecule (Suh and Hille, 2008), there is increasing evidence that PIP3 also regulates various ion channels including K+ inward rectifier channels (Rohács et al., 2003), Ca2+-activated K+ channels (Srivastava et al., 2005), epithelial Na+ channels (Pochynyuk et al., 2006) as well as cyclic nucleotide-gated channels (Zhainazarov et al., 2004). Among the P2X receptors, the desensitization of the P2X2 subtype is modulated by PI(3)P and PI(3,5)P2 (Fujiwara and Kubo, 2006) and we have recently reported that the P2X1 and P2X7 subtypes are sensitive to PIP2, but not to PIP3 (Zhao et al., 2007; Bernier et al., 2008). Therefore the sensitivity of P2X4 receptor channels to both PIP2 and PIP3 indicates that phosphoinositides regulate P2X receptor channel function in a subunit-specific manner.
We also provide evidence that modulation of P2X4 channels results from the direct effects of protein-lipid interactions on channel function, and not from indirect effects on trafficking and endocytosis since PIP2 is able to rescue P2X4 currents from rundown in inside-out macropatches excised from oocytes. P2X4 channel activity was stimulated by direct application of PIP2 to the intracellular side of the macropatch, likely by promoting the exit of P2X4 channels from their desensitized state. This could explain the longer recovery time and the slower kinetics of P2X4 currents recorded in whole cell configuration when PIP2 levels are low after treatment with wortmannin. Another strong argument for the direct effect of phospholipids on P2X4 function is based on the fact that PIP2 and PIP3, as well as other phosphoinositides, are able to interact directly with the proximal C-terminal subdomain Cys360–Val375 of P2X4 subunits in vitro. This region contains a cluster of basic residues at the right location for interacting with membrane-bound anionic phospholipids, and it shows significant conservation among all known P2X subunits. By analogy with other channels (Suh and Hille, 2008), the highly negatively charged head group in phosphoinositides is thought to interact directly with a distinct positively charged binding site in this region of the P2X4 subunit. The exact contribution of each of the 7 candidate residues (6 lysines and 1 arginine, see sequence in Fig. 8) to the binding of PIP2 and/or PIP3 will remain to be investigated.
The fact that the N-terminal domain of P2X4 did not bind to any of the phosphoinositides tested confirms an important regulatory role for the C-terminal domain that also contains determinants of desensitization (Fountain and North, 2006) and internalization (Royle et al., 2005). Our results in macropatch recordings and in lipid binding assay strongly suggest that direct interactions between P2X4 channels and phosphoinositides mediate the modulation observed in microglia as well as in Xenopus oocytes.
The relevance of the modulation of microglial and neuronal P2X4 function by PIP2 and/or PIP3 in pathophysiological conditions like neuropathic pain will remain to be established. However, in a broader context, our results highlight a positive and subunit-specific modulation of P2X receptor channels by phosphoinositides through direct channel–lipid interactions. Many metabotropic receptors regulate phosphoinositide levels through the activation of PI3K or phospholipase C, therefore we will need to integrate their exact role on P2X function in our understanding of ionotropic ATP signaling in the nervous system and in peripheral tissues.
This work has been supported by operating grants from Canadian Institutes of Health Research (CIHR) (P.S.), Agence Nationale pour la Recherche and Projet International de Coopération Scientifique du Centre National de la Recherche Scientifique (E.B.G.), and National Institutes of Health (D.L.). L.P.B. holds a studentship from the CIHR training grant “Pain: Molecules to Community.”
- Correspondence should be addressed to Dr. Philippe Séguéla, Montreal Neurological Institute, 3801 University, Suite 778, Montréal, Québec, Canada H3A 2B4.