An intracellular motif of P2X3 receptors is ... - Wiley Online Library

Journal of Neurochemistry, 2007, 102, 1357–1368

doi:10.1111/j.1471-4159.2007.04640.x

An intracellular motif of P2X3 receptors is required for functional cross-talk with GABAA receptors in nociceptive DRG neurons Estelle Toulme´,* Dominique Blais, Claire Le´ger,à Marc Landry,à Maurice Garret,* Philippe Se´gue´la and Eric Boue´-Grabot* *Universite´ Bordeaux 2, CNRS UMR 5227, Bordeaux, France Montreal Neurological Institute, McGill University, Montreal, QC, Canada àUniversite´ Bordeaux 2, INSERM U862, Bordeaux, France

Abstract Functional cross-talk between structurally unrelated P2X ATP receptors and members of the ‘cys-loop’ receptor-channel superfamily represents a recently-discovered mechanism for rapid modulation of information processing. The extent and the mechanism of the inhibitory cross-talks between these two classes of ionotropic receptors remain poorly understood, however. Both ionic and molecular coupling were proposed to explain cross-inhibition between P2X subtypes and GABAA receptors, suggesting a P2X subunit-dependent mechanism. We show here that cross-inhibition between neuronal P2X3 or P2X2+3 and GABAA receptors does not depend on chloride and calcium ions. We identified an intracellular QST386–388 motif in P2X3 subunits which is required for the functional

coupling with GABAA receptors. Moreover the cross-inhibition between native P2X3 and GABA receptors in cultured rat dorsal root ganglia (DRG) neurons is abolished by infusion of a peptide containing the QST motif as well as by viral expression of the main intracellular loop of GABAA b3 subunits. We provide evidence that P2X3 and GABAA receptors are colocalized in the soma and central processes of nociceptive DRG neurons, suggesting that specific intracellular P2X3-GABAA subunit interactions underlie a pre-synaptic cross-talk that might contribute to the regulation of sensory synaptic transmission in the spinal cord. Keywords: ATP, GABA receptor, pain, purinoceptor, sensory neurons, transmitter-gated channels. J. Neurochem. (2007) 102, 1357–1368.

Cross-talk between two structurally and functionally divergent neurotransmitter receptors represents a recently discovered mechanism for rapid modulation or integration of receptor activity (Brandon and Moss 2000; Salter 2003; Roberts et al. 2006). Interaction between G-protein-coupled receptors and receptor channels such as dopamine D5 and GABA receptors, or D1 and NMDA receptors, was shown to modulate function or plasma membrane surface expression of these receptors (Liu et al. 2000; Lee et al. 2002; Pei et al. 2004). Non-independent receptor function was also demonstrated between different ligand-gated channels for example between ATP-gated channels and several members of the cys loop receptor family including nicotinic, serotonin, and GABA receptor-channels (Barajas-Lopez et al. 1998, 2002; Searl et al. 1998; Zhou and Galligan 1998; Khakh et al. 2000; Sokolova et al. 2001; Boue-Grabot et al. 2003, 2004a; Karanjia et al. 2006), between glycine and GABAA receptors (Li and Xu 2002; Li et al. 2003), as well as between AMPA and NMDA receptors (Bai et al. 2002). Co-activation of P2X2 receptors and either nicotinic, serotonin

5-hydroxytryptamine type 3 (5-HT3), or GABAA/C receptors, leads systematically to a cross-inhibitory interaction that translates into non-additivity of the recorded currents (Roberts et al. 2006). Because fast neurotransmitters such as GABA/glycine (Jonas et al. 1998), ATP/acetylcholine (Galligan and Bertrand 1994; Redman and Silinsky 1994) or ATP/GABA are co-released in the nervous system (Jo and Schlichter 1999; Jo and Role 2002), interactions between their respective receptor channels may play a critical role in shaping synaptic currents. The cross-modulation of P2X2 Received February 9, 2007; revised manuscript received March 22, 2007; accepted March 22, 2007. Address correspondence and reprint requests to Eric Boue´-Grabot, CNRS UMR 5227, Laboratoire MAC, Universite´ Bordeaux 2, 146 rue Leó Saignat, 33076 Bordeaux cedex, France. E-mail: [email protected] Abbreviations used: 5-HT3, 5-hydroxytryptamine type 3; BSA, bovine serum albumin; DRG, dorsal root ganglia; HRP, horseradish peroxide; IL2, intracellular loop; SFV, Semliki Forest Virus; TBST, Trisbuffered saline Tween-20.

2007 The Authors Journal Compilation 2007 International Society for Neurochemistry, J. Neurochem. (2007) 102, 1357–1368

1357

1358 E. Toulme´ et al.

receptor with other channels has been extensively investigated (Roberts et al. 2006); however, mechanism and the extent of interaction between P2X receptor subtypes and ligand-gated channels of the ‘cys-loop’ family remain unclear. We previously provided evidence that P2X2 and serotonin- or GABA-gated channels interact physically (BoueGrabot et al. 2003, 2004a). In agreement with these data, fluorescence resonance energy transfer studies confirmed the ˚ ) of P2X2 and a4b2 close spatial arrangement (within 80 A nicotinic channels in the plasma membrane, indicating that a conformational spread might mediate cross-inhibition between P2X2 and ‘cys-loop’ receptors (Khakh et al. 2005). Nevertheless, P2X2 and nicotinic channels are not closely localized in midbrain neurons (Khakh et al. 2005) and the neurotransmitters ACh/ATP were shown to operate through independent receptors with additive responses in visceral ganglion neurons (Reyes et al. 2006). Moreover, a chloride- and calcium-dependent cross-inhibition between fast ATP and GABA currents mediated by P2X3 receptors and GABAA receptors was proposed in nociceptive dorsal root ganglia (DRG) neurons (Sokolova et al. 2001). Thus, cross-regulation between neuronal receptor channels appears to be a complex process, the exact mechanism of which is poorly understood. We show here that cross-inhibition between P2X3 or P2X2+3 and GABAA receptors, in heterologous system as well as nociceptive DRG neurons, is mediated by chloride- and calcium-independent intersubunit interactions, and identified QST motif within the P2X3 C-terminus which is required for interaction with GABAA receptors. Disruption of P2X3/GABAA coupling by competition abolished the current occlusion in nociceptive DRG neurons, suggesting that regulation of interaction between these receptor channels may modulate the pain signaling.

Materials and methods Molecular biology Wild-type rat P2X2, P2X3, GABA a2, ß3, c2, or c3 subunits cDNAs subcloned into pcDNA3 expression vector (Stratagene, La Jolla, CA, USA) were used in this study. All these constructs were available from previous work (Boue-Grabot et al. 2000, 2004a). Deletions (D) of the C-terminus of P2X3 subunits were generated by PCR using 5¢ primers (MWG-Biotech, Ebersberg, Germany) with a non-coding HindIII site and 3¢ primers derived from the amino acid sequence of P2X3 followed by an artificial stop codon and a terminal XhoI site. These PCR products were subcloned between HindIII and XhoI into pcDNA3 vector. Point mutations were constructed using the Quikchange site-directed mutagenesis system (Stratagene, La Jolla, CA, USA). All constructs were obtained using PfuTurbo DNA polymerase (Stratagene) to minimize artifactual mutations then verified by automatic sequencing. For viral expression, cDNAs encoding the intracellular loop (IL2) of b3 or c3 subunits (BoueGrabot et al. 2004a) were subcloned into the pSFV-IRES-GFP

vector kindly provided by Dr R. Dunn (McGill University, Montreal, QC, Canada). Electrophysiological recordings from oocytes Oocytes were removed from Xenopus laevis and prepared as previously described (Boue-Grabot et al. 2004b). Stage V and VI oocytes were microinjected in the nucleus with 10 nL water containing cDNAs mixture. The oocytes were then incubated in Barth’s solution containing 1.8 mmol/L CaCl2 and gentamycin (10 lg/mL, Sigma, Lyon, France) at 19C for 1–4 days prior to electrophysiological recordings. Two-electrode voltage-clamp recordings were carried out at 20C using glass pipets (1–2 MW) filled with 3 mol/L KCl solution to ensure a reliable holding potential. Oocytes were voltage clamped at )60 mVand the membrane currents were recorded with an OC-725B amplifier (Warner Instruments, Hamden, CT, USA) and digitized at 1 KHz on a Power PC Macintosh G4 using AXOGRAPH 4.9 software (Axon Instruments, Molecular Devices; Sunnyvale, CA, USA). Oocytes were perfused at a flow rate of 10–12 mL/min with Ringer solution, pH 7.4 containing (in mmol/ L) 115 NaCl, 5 NaOH, 2.5 KCl, 1.8 CaCl2, and 10 HEPES. Some experiments were performed in extracellular Ringer solution without CaCl2 or replaced with 1.8 mmol/L BaCl2. Agonists and drugs (Sigma) were prepared at their final concentrations in the perfusion solution and applied using a computer-driven valve system (Ala Scientific, Westbury, NY, USA). Applications (3–5 s) were separated by 3–4 min to avoid desensitization of the ATP receptors and only reproducible ATP, GABA, and ATP + GABA currents with full recovery after washout were taken in account. Dorsal root ganglia cell culture Adult male Sprague–Dawley rats (100–275 g) were killed by CO2 asphyxiation. Thoraco–lumbar dorsal root ganglia were rapidly dissected out and placed in Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA). After removal of the connective tissue, the ganglia were incubated in Dulbecco’s modified Eagle’s medium containing collagenase type D (0.28 U/mL; Sigma) for 80 min at 37C and then transferred to phosphate-buffered saline containing trypsin (25 000 U/mL; Sigma) for 10 min at 37C. Cells were subsequently dissociated by trituration with fire-polished pipettes and plated onto Petri dishes (Sarstedt, Marnay, France) pre-treated with 10 lg/mL laminin. DRG cells were cultured at 37C in 5% CO2 for 1–2 days in Ham’s F-12 medium (Invitrogen) supplemented with 1% penicillin–streptomycin, 10% fetal bovine serum, 1% L-glutamine, and 40 mmol/L glucose. Viral infection of DRG neurons in primary culture pSFV-IRES-EGFP-b3, pSFV-IRES-EGFP-c3 contructs were individually co-transfected with pSCA1 helper DNA into BHK21 cells using Polyfect (Qiagen, Courtaboeuf, France) for in vivo packaging of recombinant SFV (Semliki Forest Virus) particles. Virus stocks were collected 72 h later and infection tests on BHK21 cells were carried out with several dilutions. Prior to infection, recombinant SFV particles were activated with 10 mg/mL a-chymotrypsin for 45 min at 20C and the proteolysis was terminated by addition of 10 mg/mL aprotinin. 24-h-old DRG neuron cultures were infected with activated SFV-b3 or SFV-c3 virus stocks and electrophysiological recordings of infected neurons, visualized by fluorescence of EGFP, were performed 12–24 h later.


Interaction between P2X3 and GABAA receptors 1359

Peptides Peptides corresponding to the wild-type, mutated or scrambled sequence of the C-terminal domain (residues 385–395) of rat P2X3 subunit were synthesized by Tufts University Core Facility (Boston, MA, USA). Purified peptides were diluted at a final concentration of 200 lg/mL (160 lmol/L) in the patch pipette internal solution. Electrophysiological recordings from DRG neurons Whole-cell voltage clamp recordings from small to medium sized (20–35 lm) DRG neurons were performed 12–48 h after plating. Cells were superfused at room temperature (20–23C) with an external solution containing (in mmol/L): 152 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, 10 Glucose, pH 7.4. All drugs were rapidly delivered to neurons through a gravity-driven fast fluid changer (SF-77B, Warner Instruments) to ensure fast exchange rate. The duration of application was usually 3–5 s and the interval between two successive applications was 3–4 min to ensure complete recovery of currents. Currents were recorded, low-pass filtered at 2 kHz using an Axopatch 200B amplifier (Axon Instruments, Molecular Devices) and digitized at 500 Hz through a Digidata 1320A. Fire-polished glass pipettes (resistance between 3 and 5 MX) were filled with internal solution containing (in mmol/ L): 135 CsCl, 4.1 MgCl2, 2 EGTA, 10 HEPES and 2.5 ATPNa2, pH 7.2 (adjusted with CsOH). In competition experiments, the pipette tips were initially filled with internal solution and then backfilled with the same solution containing the peptides. Neurons were recorded 20 min after the infusion of the peptide. Data analysis To allow for differences in GABA- and ATP-current kinetics, we compared the peak of actual responses (IATP/GABA) to the peak of predicted additive responses (Ipredicted) obtained using AXOGRAPH 4.9 software (Axon Instruments) and not with the sum of the peaks of individual responses. Current amplitudes were normalized to I predicted and represented as the percentage of either the prediction or inhibition that is (Ipredicted ) IATP/GABA). Numerical values are presented as mean ± SEM from n determinations with statistical significance assessed using Student’s t-test or Kruskal–Wallis test with Dunn’s post-test for multiple comparisons with the same control group. The differences were considered significant with a p < 0.05 (PRISM 4.0; Graphpad, San Diego, CA, USA). Immunohistochemistry Rats were anesthetized with urethane (1.2–1.5 g/kg intraperitoneally) and perfusion fixed via the ascending aorta with 100 mL Kreb’s solution (in mmol/L) (124 NaCl, 2.4 KCl, 2.4 CaCl2, 1.3 MgSO4, 1.2 KH2PO4, 26 NaHCO3, 1.25 HEPES, and 10 glucose; Sigma, St Louis, MO, USA) at 30C (pH 7.4) followed by 300 mL of an ice-cold fixative containing 4% paraformaldehyde and 0.2% picric acid in 0.1 mol/L phosphate buffer (pH 7.4). The lumbar spinal cord, and L4–L5 dorsal root ganglia were rapidly dissected out, immersed in the same fixative for 2 h and rinsed for at least 24 h in 0.1 mol/L phosphate buffer (pH 7.4) containing 15% sucrose and 0.02% sodium azide (Sigma). The spinal cord was frozen, cut at 14 lm in a cryostat (Microm, Heidelberg, Germany), and collected onto slides. Sections were incubated in Tris-buffered saline containing 0.05% Tween 20 (TBST) and 1% bovine serum albumin (BSA). The slides were then immersed in sodium citrate

(a)

(b)

(c)

Fig. 1 Cross-inhibition between ionotropic ATP and GABA responses in rat DRG neurons. (a) whole cell voltage clamp recordings obtained from small diameter DRG neurons in culture, Vh = )60 mV. Co-application of ATP (30 lmol/L) and GABA (100 lmol/L) evoked an inward current (Iactual) significantly smaller than the predicted current (Ipredicted) corresponding to the sum of the individual currents. (b) Inhibition of P2X3 current amplitude is also observed when ATP is applied during GABA application (Iactual) compared with the algebraic sum of individual traces (Ipredicted), n = 7. Traces shown in (a) and (b) were recorded from individual neurons. (c) Pooled results of current inhibition recorded after co-application of ATP (30 lmol/L) plus GABA (100 lmol/L), ab-meATP (30 lmol/L) plus GABA (100 lmol/L), ATP (30 lmol/L) plus muscimol (100 lmol/L), and sequential applications of GABA then ATP. Successive applications were separated by a 3– 4 min washout period to ensure complete recovery of currents. Data were statistically compared with the sum of corresponding individual currents (Ipredicted). *p < 0.05; **p < 0.005; ***p < 0.0005.

buffer (pH 6) and heated three times in a microwave oven for 5 min at 700 W. They were then incubated with guinea-pig anti-P2X3 (1/ 1000, Chemicon, Temecula, CA, USA) and rabbit anti-GABA a2 subunits (1/1000; Alomone, Israel) antisera in 0.05% TBST and 1% BSA overnight at 4C. They were rinsed three times for 10 min in TBST, incubated with a blocking reagent (Perkin-Elmer Inc., Boston, MA, USA) and then with a swine horseradish peroxidase (HRP)-conjugated anti-rabbit antibody (1/100; Dako SA, Trappes, France) together with biotinylated goat anti-guinea-pig antibody (Jackson Immunoresearch, West Groce, PA, USA) for 30 min at



20C. After further rinsing, the peroxidase activity corresponding to GABAA immunoreactivity was developed with Cyanin 5-conjugated tyramide (Perkin-Elmer) according to the manufacturer’s instructions, as previously described (Landry et al. 2004). Sections were subsequently rinsed and P2X3 immunoreactivity was visualized with AlexaFluor 488-conjugated streptavidin (1 : 500 in Tris–BSA; Molecular Probes, Leiden, The Netherlands). Image processing Images from DRG and spinal cord were obtained from a Zeiss Axiophot 2 fluorescence microscope equipped with the appropriate filter sets. High-magnification double-immunostaining in the spinal cord was analyzed with a Leica DMR PCS SP2 AOBS confocal microscope (Leica, Heidelberg, Germany) using a 40· or 63· oil-immersion lens. In all cases, scans were carried out sequentially with the 488 and 647 nm lines of the laser. Only single optical sections have been used for illustrations. Digital images were finally optimized for image resolution (final resolution 300 dpi), brightness, and contrast using Photoshop 6.0 (Adobe, San Jose, CA, USA). Images were not altered in any way, e.g. by removing or adding image details.

Results

Cross-inhibition between P2X3 and GABAA receptors in rat DRG neurons Whole-cell voltage-clamp recordings were obtained from cultured DRG neurons. We examined responses mediated by application of ATP (30 lmol/L) and GABA (100 lmol/L) on small or medium diameter neurons (20–35 lm) held at )60 mV. The application of ATP evoked a fast desensitizing inward current (IATP = )493 ± 276 pA, n = 13) whereas GABA induced a slowly desensitizing inward current (IGABA = )479 ± 84 pA, n = 13; Fig. 1a). On average, 83% of cells displayed ATP and GABA responses, 6% were ATP only-responding cells and 11% of the cells responded to GABA only (224 cells). Identical responses (not shown) were observed in response to abme-ATP (30 lmol/L) or muscimol (100 lmol/L) supporting previously established data that ATP and GABA currents are mainly mediated by P2X3 and GABAA receptor channels in small diameter DRG neurons (Nakatsuka and Gu 2006). Complete recovery of ATP and GABA current was observed after a delay of 3 min between successive applications of agonists (see Fig. S1). To measure with accuracy peak current amplitude without variations attributable to current run down, a 3–4 min interpulse delay was kept in all following experiments (see Material and methods). The co-application of ATP and GABA evoked an inward current (Iactual = )543 ± 240 pA, n = 6) that was significantly smaller than the predicted current corresponding to the sum of the individual currents (Ipredicted = )774.8 ± 272.7 pA, n = 6; Fig. 1a and c). Iactual represented 66.90 ± 8.50% of Ipredicted. Similar non-additive responses were observed either during the co-application of abme-ATP + GABA, ATP + muscimol (respectively, Iactual were 74.95 ± 7.3%, n = 5 and 66.7 ±

4.1%, n = 6 of the prediction, Fig. 1a and c). The distinct kinetics of P2X3 and GABAA responses may have led to an underestimation of the peak responses during the co-application of both agonists. To eliminate this bias, ATP was applied during a sustained application of GABA and a similar peak current inhibition was observed (Iactual = 75.60 ± 3.80%, n = 7, Fig. 1b). These results showed that 25–35% current inhibition occurred between ATP- and GABA-gated channels expressed in DRG neurons in accordance with previous work (Sokolova et al. 2001). Double immunofluorescence labeling with antibodies directed against P2X3 and GABAA a2 subunits revealed that a2 subunit was expressed in P2X3-positive DRG neurons (Fig. 2c and i). In agreement with previous reports (Vulchanova et al. 1998; Labrakakis et al. 2003), the wide distribution of GABAA subunit in DRG neurons and the spinal cord (Fig. 2b and e) contrasted with the P2X3immunoreactivity restricted to small-size sensory neurons in rat DRG that terminate in inner lamina II of the spinal cord (Fig. 2a and d). Confocal imaging at higher magnification demonstrated the overlapping distribution of GABAA a2 and P2X3 subunits in primary afferent terminals in the superficial dorsal horn (Fig. 2g–i). These results support the idea that the cross-modulation between ATP- and GABA-gated channels observed in primary culture of DRG neurons may represent a physiological and native regulation. Properties of current occlusion between recombinant P2X3 and GABAA receptors We first co-expressed P2X3 with a2 and b3 GABA subunit cDNAs, the main a and b isoforms expressed with c2/3 subunits in DRG neurons (Bohlhalter et al. 1996), in Xenopus oocytes to investigate whether a current occlusion also occurs in heterologous expression system. As depicted in Fig. 3, ATP (30 lmol/L) and GABA (100 lmol/L) evoked typical inward current (IATP = )2.35 ± 0.57 lA and IGABA = )4.04 ± 1.64 lA, n = 11) when each agonist was applied alone. When ATP and GABA were co-applied, the measured current (Iactual = )3.78 ± 1.44 lA, n = 11) represented 75.60 ± 2.28% of the predicted current (Ipredicted = )4.98 ± 1.87 lA, n = 11). On oocytes expressing lower density of both channels (IATP = )0.97 ± 0.2 lA and IGABA = )1.19 ± 0.32 lA, n = 16), sequential application of agonists gave similar current inhibition: when ATP was applied during a sustained GABA application, the GABA/ ATP current (Iactual = )1.55 ± 0.36 lA, n = 16) was consistently inhibited, representing 75.74 ± 2.13% of the predicted current (Ipredicted = )2.16 ± 0.52 lA, n = 16). As illustrated in Fig. 3c, the fast desensitizing ATP current was sometimes followed by a rapid reduction of the GABA current, showing an instantaneous and reciprocal current occlusion. Coexpression of heteromeric P2X2/3 with a2b3 receptors as well as P2X3 with GABA receptors containing c2 or c3



GABA α2

P2X3

overlay

DRG

Fig. 2 Colocalization of P2X3 and GABA a2 subunits in DRG and spinal cord. Double immunohistochemical labeling for GABA a2 subunit (red) and P2X3 (green) in rat L5 DRG (a–c) and spinal cord (c–i) sections. (a–c) P2X3 is expressed in a subpopulation of GABA a2-expressing sensory neurons (arrows in a–c). A high number of cells remain single labeled for GABA a2 (arrowheads in b and c). (d–f) In the dorsal horn of the spinal cord, both GABA a2 and P2X3 are expressed in the superficial laminae, and in particular in lamina II (arrows in d–f). (g–i) At higher magnification, nerve endings show colocalization between GABA a2 and P2X3 in lamina II (arrows in g–i). GABA a2 single-labeled terminals are also found in the dorsal horn (arrowheads in g–i). Bar: 50 lm (a–f); 15 lm (g–i).

(a)

(b)

(c)

(e)

(f)

(h)

(i)

L5

(d) L5

(g)

subunits resulted in a similar inhibition of the current evoked by both agonists (abme-ATP or ATP in combination with GABA) in comparison with the prediction. Iactual represented respectively, 74.08 ± 2.8%, n = 16 and 78.41 ± 4.24%, n = 7 of the prediction (Fig. 3d and Figs S2 and S3). The fast desensitization of P2X3 receptors did not allow us to perform GABA application during ATP current and consequently prevented us from evaluating the reciprocity of the current occlusion between P2X3 and ab or abc GABA receptors. However, the observation of identical GABAevoked currents in absence or in presence of ATP after complete desensitization of ATP currents showed that occlusion is a receptor-activity dependent mechanism. Cross-inhibition between ATP- and GABA-gated channels expressed in DRG was proposed to be mainly due to the inhibitory action of Cl) efflux on P2X3 channels via open GABA channels and to the inhibitory action of Ca2+ ions on GABA-gated channels via open P2X3 channels (Sokolova et al. 2001). We first verified that current occlusion is independent of agonist cross-reactivity. In oocytes expressing P2X3 alone, GABA did not activate or modulate ATP-induced P2X3 currents (see Fig. S1). Previous results have shown that, conversely, ATP has no effect on GABA receptors (Boue-Grabot et al. 2004a,b). To investigate whether extracellular Cl) ions contribute to the current inhibition of P2X3 via a2b3 channels (results Fig. 3), oocytes were held at a holding potential of )20 mV.

As shown in Fig. 4a, GABA elicited outward current (i.e. Cl) influx; IGABA = 0.62 ± 0.23 lA, n = 5), whereas ATP elicited inward current via activation of P2X3 channels (IATP = )0.21 ± 0.04 lA, n = 5). Co-application of ATP and GABA induced a biphasic current with a fast inward followed by an outward component: the initial fast inward current is almost completely depressed and the maximal amplitude of the outward current (Iactual = 0.46 ± 0.21 lA, n = 5) is also significantly reduced in comparison with the amplitude of the sum of the individual responses (Ipredicted = 0.61 ± 0.23 lA, n = 5 ; Fig. 4d). At a potential of +20 mV, both GABA and ATP induced outward currents (IGABA = 0.71 ± 0.17 lA, n = 12; IATP = 0.24 ± 0.06 lA, n = 12). Again, co-application of both agonists elicited outward currents (Iactual = 0.72 ± 0.18 lA, n = 12) significantly smaller than the predicted current (Ipredicted ¼ 0.96 ± 0.2 lA, n = 12; Fig. 4b and d). The average current occlusion was 29.75 ± 4.54%, n = 5 and 23.08 ± 4.58%, n = 12, respectively at )20 and +20 mV revealing that current inhibition was not modified by the direction of Cl) ions or cations flow. Similarly, we investigated whether Ca2+ entry via open P2X3 channels participated in the observed cross-inhibition. P2X3 and GABAA receptors generated nonadditive responses when extracellular Ca2+ was removed from extracellular solution or substituted with Ba2+ (Fig. 4c and d). The current inhibition in absence of Ca2+ influx (CI = 28.31 ± 5.38%, n = 7 in 0 Ca2+; CI = 24.77 ± 2.50%,



(a)

(b)

(c)

(d)

Fig. 3 Current occlusion between recombinant GABAA and P2X3containing receptors co-expressed in Xenopus oocytes. (a and c) In oocytes co-expressing P2X3 and a2b3 GABAA subunits, simultaneous (a) or sequential (c) application of ATP (30 lmol/L) and GABA (100 lmol/L) evoked an inward current (Iactual) significantly smaller than the sum of individual IATP and IGABA (Ipredicted), Vh = )60 mV. Two examples of inward currents evoked by sequential application are shown in (c): the right panel shows a GABA current occlusion observed sometimes during ATP application. (b) Peak current data

normalized to the prediction of receptor independence when ATP was applied concomitantly with GABA (ATP + GABA) or during a sustained application of GABA (GABA then ATP). (d) Co-application of abmeATP and GABA evoked peak currents smaller than the sum of individual currents for oocytes co-expressing heteromeric P2X2/3 receptors and a2b3 GABAA receptors. Current inhibition was also observed when P2X3 receptors were co-expressed with GABAA receptors containing either c2 or c3 subunits.

n = 7 in 1.8 mmol/L BaCl2) was not significantly different from those observed in control conditions (24.30 ± 2.10%, n = 16, in 1.8 mmol/L CaCl2). These results showed that the cross-inhibition between P2X3 and GABAA receptors was independent from the direction of ionic flow, in contrast with a report on DRG neurons (Sokolova et al. 2001), raising the issue of the mechanism of the functional coupling between P2X3 and GABAA receptors expressed in heterologous cells or in sensory neurons.

inhibition with nicotinic, serotoninergic, and GABAergic receptors (Boue-Grabot et al. 2003, 2004a,b). C-terminal domains vary in length (ranging from 27 to 239 amino acids) and sequences among all P2X subunits (Vial et al. 2004). Besides, the P2X3 C-terminus (50 amino acids) is shorter than the one of P2X2 (120 amino acids) and both C-termini do not share conserved sequences except a YxxxK motif present in all P2X subtypes and involved in receptor trafficking (Chaumont et al. 2004). Thus, to determine whether the C-terminal domain of P2X3 subunit is involved in cross-inhibition, we first truncated this domain at different positions (Fig. 5a). All truncated forms of P2X3 subunit, except deletion at lysine 357 (P2X3DK357), resulted in the expression of functional P2X3-like receptors (IATP was ranging from )0.28 ± 0.07 to )1.48 ± 0.42 lA, n between

Current inhibition required specific P2X3 C-terminal residues We previously showed that the C-terminal tail of P2X2 subunits (residues 353–472) was involved in the cross-



Fig. 4 Functional coupling between recombinant P2X3 and GABAA receptors is independent from the direction of ionic flow and from extracellular calcium. (a) Currents evoked by successive applications of ATP (30 lmol/L), GABA (100 lmol/L), and ATP + GABA, on oocytes voltage-clamped at )20 mV co-expressing P2X3 and GABA (a2b3) receptors. (b) Application of ATP during sustained application of GABA induced an outward current that is smaller than the predicted current. Outward currents were recorded at a holding potential (Vh) of +20 mV. (c) Representative trace of ATP induced current inhibition during GABA application in Ca2+-free Ringer. (d) Summary bar graph showing current inhibition in all conditions tested, demonstrating the independence of P2X3-GABAA receptors interaction from direction of ionic flow and voltage.

(a)

(b)

(c)

(d)

(a) Fig. 5 Critical role of P2X3 C-terminal domain in current cross-inhibition between P2X3 and GABAA receptors. (a) Schematic transmembrane topology of wild-type rat P2X3 subunit and sequence of the C-terminal tail. Dots indicate last amino acids in truncated mutants that were expressed in Xenopus oocytes. (b) Summary bar graph representing current inhibition for wild-type and truncated mutant P2X3 receptors co-expressed with a2b3 GABA subunits, recorded when ATP (30 lmol/L) is applied during GABA application (100 lmol/L). *p < 0.05; **p < 0.005; ***p < 0.0005. NF, not functional. (c) Representative traces of current additivity recorded from one oocyte co-expressing P2X3DD379 and GABAA (a2b3) receptors. Vh = )60 mV. Iactual corresponding to the application of ATP during the application of GABA, is not different from the sum of individual currents (Ipredicted).

(b)

6 and 13). Mutant P2X3 subunits DK367, DD379, DK385, DQ386, DS387, DT388, DS390, DG391, DS394 were coexpressed with a2b3 receptors in oocytes. The amplitude of GABA and ATP responses was compared with the predicted

(c)

sum of the individual responses induced by ATP alone or GABA alone (Fig. 5b and c). Co-activation of P2X3 DK367, DD379, DK385, DQ386, DS387, and a2b3 receptors evoked currents not different from the sum of the individual



(a)

(b)

Fig. 6 Identification of the intracellular motif P2X3 QST386)388 required for the P2X3-GABAA cross-talk. (a) Current inhibition (%) of ATP-induced currents during GABA application in oocytes co-expressing a2b3 GABA subunits with either wild type P2X3 (WT) or mutant P2X3 receptors with single amino acid substitutions in the C-terminus. All residues from positions 385 to 391 were mutated to alanine. *p < 0.05; **p < 0.005; ***p < 0.0005. (b) Representative current additivity from one oocyte co-expressing P2X3S387A and a2b3 GABA receptors. Vh = )60 mV.

responses. Iactual represented respectively 101.3 ± 5.80%, 100.10 ± 2.66%, 97.06 ± 0.47%, 95.13 ± 2.52%, 96.24 ± 2.38% of the prediction. In other words, the current inhibition between P2X3 wt and a2b3 receptors was significantly abolished (CI < 4%) by the deletion of the last ten to thirty amino acids of P2X3 (Fig. 5b). Co-activation of P2X3DT388, DS390, DG391, DS394, and a2b3 GABAA receptors evoked current was significantly smaller than the prediction. Thus, cross-inhibition occurred when the last nine amino acids of P2X3 were suppressed although the extent of the current occlusion varied in function of the position of the deletion. As summarized in Fig. 5b, current inhibition observed between a2b3 and DS390 or DS394 (respectively, CI = 21.76 ± 5.11% and 23.10 ± 3.6%) was similar to current inhibition between

GABAA and wild-type P2X3. Current inhibition with DT388 or DG391 was reduced (CI = 14.11 ± 3.34% and CI = 7.71 ± 5.45% respectively), but without statistical differences (p > 0.05) with wild-type P2X3. These results showed that the distal part of the P2X3 C-terminus plays a prominent role in the functional cross-talk with GABAA receptors. To further identify P2X3 C-terminal residues important for this functional cross-talk, the residues K385 to G391 were alanine-substituted individually and the mutants were compared with wild-type P2X3 with respect to current occlusion when co-expressed with a2b3 receptors in oocytes (Fig. 6). Alanine substitution of K385 or D389 did not modify current inhibition between P2X3 and a2b3 receptors (respectively, CI was 18.46 ± 1.64% and 20.95 ± 2.32%, n = 11 for each) as compared with those measured with the wild-type P2X3 receptors (CI = 24.30 ± 2.10%, n = 16). A small decrease of the current inhibition was observed with S390A or G391A mutants (respectively, CI = 14.17 ± 5.20% n = 8 and 13.05 ± 4.90%, n = 10), although not significantly different from wild-type P2X3. In contrast, alanine substitution of residues Q386, S387, and T388 gave rise to significant reduction of the current occlusion compared with the wild-type. The percentage of inhibition was respectively, 11.91 ± 2.156%, n = 12; 7.83 ± 1.75%, n = 16; and 10.99 ± 3.56%, n = 10 (Fig. 6a and b). The replacement of the residue S387 by either a glutamine or a leucine induced a similar reduction of the current inhibition (respectively, CI was 7.02 ± 2.82%, n = 7 and 10.14 ± 2.40%, n = 12) indicating that this effect was not specifically due to the alanine residue. These results demonstrated the importance of the QST386–388 motif within the C-terminal domain of P2X3 subunits for the cross-inhibition with a2b3 GABA receptors and indicate that ion-independent molecular interactions underlie this coupling. Molecular coupling between P2X3 and GABAA channels in DRG neurons To test whether molecular interactions between native P2X3 and GABAA receptors are responsible for cross-talk in DRG neurons, we reasoned that peptides corresponding to critical P2X3 C-terminal sequence motifs would act as competitive inhibitors. As depicted in Fig. 7(a, b, and e), current inhibition (33.30 ± 4.10%, n = 6) observed between ATP and GABA responses in DRG neurons was abolished by intracellular infusion of the peptide corresponding to residues 385–395 of the C-terminus of P2X3 subunits (pepX3-CT 160 lmol/L). The remaining current inhibition, representing 3.90 ± 1.70% (p < 0.005, n = 7) was negligible in presence of pepX3-CT (Fig. 7d and e). Notably, the replacement of residue serine387 by an alanine within the peptide sequence (pepX3-S387A), caused significantly less disruption of the cross-talk. Indeed, current inhibition represented 18.22 ± 3.60% (n = 9) of



(a)

Fig. 7 Impact of the infusion of DRG neurons with P2X3 C-terminal peptides on the P2X3-GABAA cross-talk. (a) Schematic representation of the P2X3 subunit showing the 11 amino acids sequence of the peptides infused in DRG neurons corresponding to the sequence between residues 385–395 of: wild-type P2X3 (pepX3-CT), with an alanine substitution on serine 387 and or scrambled (pepX3-sc). (b) Representative traces of inward currents evoked by ATP (30 lmol/L), GABA (100 lmol/L) and ATP plus GABA (Iactual) showing peak current additivity after pepX3-CT infusion through the patch pipet. (c and d) ATP plus GABA co-application induced an inward peak current (Iactual) that is smaller than the sum of the individual ATP and GABA responses (Ipredicted) either in presence of the pepX3-S387A (c) or pepX3-sc (d). Whole-cell currents were measured at Vh = )60 mV. (e) Current inhibition normalized to predicted responses in control conditions or in the presence of peptides in the pipet (pepX3-CT, pepX3-sc, and pepX3-S387A). *p < 0.05; **p < 0.005. NS, not significant.

(b)

(c)

(d)

(e)

the prediction (Fig. 7c and e) in neurons infused with pepX3-S387A, which was not significantly different from the current inhibition observed in presence of the controlscrambled peptide pepX3-sc (current inhibition = 33.12 ± 8.5%, n = 6). We have shown previously that the main intracellular loop of GABAA b subunits was involved in the functional coupling between P2X2 and GABAA receptors expressed in oocytes (Boue-Grabot et al. 2004a), so we decided to test the role of the IL2 of GABAA subunits in their interaction with P2X3 in DRG neurons. For efficient high-level expression of recombinant IL2 domains in cultured DRG neurons, we used the SFV expression system and IRES-GFP-SFV constructs (see Material and methods and Fig. 8a). After infection with SFV-IL2b3-GFP or SFV-IL2c3-GFP of primary cultures of DRG neurons, the responses to ATP, GABA, and ATP + GABA were recorded from infected neurons visualized by green fluoresent protein fluorescence. Iactual was not signifi-

cantly different from Ipredicted in infected neurons expressing IL2b3 showing that expression of this domain drastically and significantly reduced the current inhibition between P2X3 and GABA receptors (CI = 4.94 ± 2.43%, n = 9, Fig. 8b and d). In contrast, current inhibition was maintained in IL2c3 expressing neurons. The current inhibition represented 19.94 ± 3.5%, n = 5 which was not significantly different from non-infected neurons (Fig. 8c and d). Thus, expression of the intracellular loop of b3 subunit, but not of c3 subunit, led to the disruption of the interaction by competition. Altogether, these experiments showed that the native crosstalk between P2X3 and GABAA in DRG neurons is based on molecular interactions which involved both the QST386–388 motif within the P2X3 C-terminal domain and most likely the intracellular loop of GABAA b subunits. However, we cannot exclude that intracellular domains of other GABAA subunits or other proteins may participate in this receptor– receptor coupling.



(a)

(c)

(b)

(d)

Fig. 8 Disruption of the P2X3-GABAA coupling by viral expression of intracellular domains of GABAA subunits in DRG neurons. (a) Schematic representation of GABAA subunits transmembrane topology. (b) Co-application of ATP (30 lmol/L) and GABA (100 lmol/L) on DRG neurons infected with SFV-IL2b3 induced an inward current (Iactual) not different from the predicted response (Ipredicted). (c) SFV-IL2c3 infection did not disturb the inhibitory cross-talk observed between P2X3 and GABAA, as co-application of ATP plus GABA (Iactual) on infected neurons evoked a current that is significantly smaller than the prediction (Ipredicted). (d) Means of current inhibition recorded on noninfected neurons (control), or SFV-infected neurons expressing the intracellular loop of b3 subunit (IL2b3) or c3 subunit (IL2-b3). **p < 0.005; ***p < 0.0005. NS, not significant.

Discussion

We investigated here functional interactions between P2X3 receptors and GABA receptor channels and observed that concomitant activation of fast ATP and GABA currents in small- or medium-size DRG neurons triggered instantaneous current occlusions, confirming the existence of a negative cross-talk between ATP- and GABA-gated channels in DRG (Sokolova et al. 2001). Sokolova et al. proposed that the current occlusion was mainly mediated by the inhibitory effect of GABA receptors-mediated chloride ions efflux on P2X receptors and conversely, but to a lesser extent, by an

inhibitory action of intracellular Ca2+ influx, via open P2X receptors, on GABA receptors. This contrasted with several studies showing that reciprocal current inhibition between P2X2 and either nicotinic, 5-HT3 or GABAA/C receptors in heterologous expression system or in native myenteric neurons was independent from ion flow direction, Ca2+ or voltage (Barajas-Lopez et al. 1998, 2002; Searl et al. 1998; Zhou and Galligan 1998; Khakh et al. 2000; Sokolova et al. 2001; Boue-Grabot et al. 2003, 2004a,b; Karanjia et al. 2006). We decided to investigate whether the mechanisms of current inhibition could depend on the P2X subunits involved. Compelling data from electrophysiology, immunohistochemistry, and genetic manipulation demonstrated that fast desensitizing ATP currents are mediated by homomeric P2X3 receptors whereas slowly desensitizing ATP currents are mediated by heteromeric P2X2/3 receptors in nociceptive DRG neurons (see for review Nakatsuka and Gu 2006). Therefore we probed the molecular aspects of the cross-talk in Xenopus oocytes co-expressing P2X3 or P2X2/3 with a2b3 or a2b3c GABA subunits and in native DRG neurons. The current inhibition (25–30% of the predicted current) observed in oocytes was similar to the one observed between ATP and GABA currents on DRG neurons. In both cell types, the level of occlusion was identical regardless of the agonist application sequence (co-application of both agonists or application of ATP during sustained GABA responses). Thus, current inhibition observed during coapplication did not result from differences between ATP and GABA current kinetics which may have led to an underestimation of the ATP peak current. Moreover, we did not observe any difference in the level of occlusion in oocytes co-expressing slowly desensitizing heteromeric P2X2/3 receptors with GABAA receptors. We previously showed by sequential application that reciprocity of current occlusion between P2X2 and GABAA receptors was c subunitdependent: ab or abc GABAA receptors inhibit P2X2 mediated current, whereas, P2X2 channels inhibit only ab GABAA (Boue-Grabot et al. 2004a). It would be of interest to test the reciprocity of the cross-inhibition between P2X3 and ab or abc GABA receptors, however, the fast desensitization kinetics of P2X3 channels did not allow us to perform application of GABA during ATP current and consequently to analyze whether activation of P2X3 inhibits GABA receptors containing a c subunit. We observed that mutual current occlusion was unaffected by changing the direction of cations and chloride ions flow or by removal/ substitution of extracellular calcium ions, indicating that inhibitory ionic effects did not contribute significantly to the cross-talk between P2X3 and GABA receptors. Conformational spread (Bray and Duke 2004) is the current model to explain the ability of P2X2 and nicotinic, 5-HT3 or GABAA/C receptors to display an activity-dependent cross-inhibition: the motion triggered by the gating of one channel type is communicated to the other channels and



induce their closure (Boue-Grabot et al. 2003, 2004a; Khakh et al. 2005). A prerequisite for such a mechanism is the close proximity of receptors which to date was shown only between P2X2 and cys loop receptors. Physical association between P2X2 and 5-HT3 or GABAC was demonstrated by co-purification (Boue-Grabot et al. 2003, 2004b) as well as by the alteration of the subcellular targeting of q1/GABAC subunits in transfected hippocampal neurons (Boue-Grabot et al. 2004b). Close spatial arrangement between P2X2 and a4b2 nicotinic receptors in the plasma membrane was demonstrated by fluorescence resonance energy transfer analysis (Khakh et al. 2005). We also provided evidence of the molecular basis of the current occlusion between recombinant P2X2 and cys loop receptors by showing that the C-terminal domain of the P2X2 receptors and the intracellular loop located between the third and fourth transmembrane domains of b subunit of GABAA receptors or q1 subunits of GABAC subunits are involved in the current occlusion (Boue-Grabot et al. 2004a,b). The C-terminal domains of P2X subunits display different length and low primary sequence homology, however additive responses were observed between GABAA receptors and several mutant P2X3 receptors with C-terminal deletions. It is important to notice that deletion of the last distal aminoacids prevented the cross-talk but did not modify the P2X3 receptor phenotype. Furthermore, the cross-inhibition between P2X3 and GABA receptors in DRG neurons was specifically abolished in presence of either a peptide corresponding to residues 385–395 of P2X3 or the intracellular loop of b3 subunits in nociceptive neurons. These results reinforced our conclusion that native cross-inhibition is an ion-independent process and argue in favor of molecular interactions between the C-terminus of P2X3 and intracellular loop of most likely b GABA subunits. Higher fluorescence resonance energy transfer efficiency measured between P2X2 and nicotinic b2 subunits rather than a4 subunits suggests that P2X2 interacts strongly and directly with b2 nicotinic subunit (Khakh et al. 2005). Thus, functional interactions between members of the ATP-gated channels and cys loop receptors may be explained by a general mechanism of state-dependent conformational spread from one receptor to the other. However, from our experiments we cannot exclude the possibility that protein complexes promote association between intracellular domains of P2X3 and GABA receptor channels. Alanine-substitution of residues K385 to G391 within the C-terminal domain of P2X3 allowed us to identify an intracellular motif of three consecutive residues Q386, S387, T388 important for the cross-inhibition with GABAA receptors in Xenopus oocytes. Co-expression of GABAA receptors with each mutant P2X3 Q386A, S387A, or T388A resulted in a significant reduction of the current inhibition while mutation of the neighboring residues K385 or D389 receptor did not have any effect. Furthermore, the absence of

modification of the ATP/GABA current occlusion in DRG neurons infused with the peptide pX3-S387A demonstrated that the QST motif of P2X3 subunit is essential for native interaction with GABAA receptors. This intracellular QST motif is conserved among all mammalian P2X3 subunits but is absent from other P2X subtypes including GABA receptor-interacting P2X2, which suggest that P2X subunitspecific molecular determinants are involved in functional interactions with other ligand-gated channels. To our knowledge, the QST motif was not previously described as a putative protein interaction domain. Searching for the QST motif in other proteins in the Swiss protein database using SCANPROSITE revealed 413 entries (http://expasy.org/ tools/scanprosite). By focusing on the intracellular location of this motif, we particularly noticed its presence in the intracellular C-terminal domain of adenosine A2B receptor (position 325–327), Kappa-type opioid receptor (position 355–357), and cannabinoid receptor CB2 (position 406– 408). It is tempting to speculate that these G protein-coupled receptors may interact with GABAA receptors or other receptor channels, as previously demonstrated for dopamine D5 or D1 receptors (Liu et al. 2000; Lee et al. 2002). In summary, P2X3 receptors interact via a specific intracellular motif in its C-terminus with GABAA receptors, leading to an inhibitory cross-talk of these receptor channels in Xenopus oocytes and in primary sensory DRG neurons. Double immunolabeling showing the co-expression of a2 GABA subunit in P2X3-positive neurons that terminate in superficial dorsal horn of spinal cord indicates that the molecular interaction between these channels may regulate nociceptive signal transmission. Cross-modulation between P2X3 and GABAA receptors may regulate their opposite pre-synaptic actions, which respectively facilitate (Gu and MacDermott 1997) and inhibit (Gmelin and Zimmermann 1983; MacDermott et al. 1999) the release of neurotransmitters and consequently, modulate the balance between inhibition versus excitation of sensory transmission in the spinal cord. Acknowledgements This work was supported by CNRS, Universite´ Bordeaux 2 and Bordeaux 1, by the ANR grant #JC-05–44799 (EB-G), by the CIHR grant #MOP-14718 (PS) and by exchange programs from CNRS/ CIHR and INSERM/FRSQ. ET is a postgraduate fellow from the Ministe`re de l’Education Nationale et de la Recherche, PS is a Killam scholar.

Supplementary material The following supplementary material is available for this article online: Fig. S1 GABA did not activate ATP P2X3 currents. Fig. S2 Cross-inhibition between P2X2+3 and a2b3 GABAA receptors.



Fig. S3 Cross-inhibition between P2X2+3 and a2b3c2 GABAA receptors. This material is available as part of the online article from http:// www.blackwell-synergy.com.

References Bai D., Muller R. U. and Roder J. C. (2002) Non-ionotropic cross-talk between AMPA and NMDA receptors in rodent hippocampal neurones. J. Physiol. 543, 23–33. Barajas-Lopez C., Espinosa-Luna R. and Zhu Y. (1998) Functional interactions between nicotinic and P2X channels in short-term cultures of guinea-pig submucosal neurons. J. Physiol. 513, 671–683. Barajas-Lopez C., Montano L. M. and Espinosa-Luna R. (2002) Inhibitory interactions between 5-HT3 and P2X channels in submucosal neurons. Am. J. Physiol. Gastrointest. Liver Physiol. 283, G1238–G1248. Bohlhalter S., Weinmann O., Mohler H. and Fritschy J. M. (1996) Laminar compartmentalization of GABAA-receptor subtypes in the spinal cord: an immunohistochemical study. J. Neurosci. 16, 283–297. Boue-Grabot E., Akimenko M. A. and Seguela P. (2000) Unique functional properties of a sensory neuronal P2X ATP-gated channel from zebrafish. J. Neurochem. 75, 1600–1607. Boue-Grabot E., Barajas-Lopez C., Chakfe Y., Blais D., Belanger D., Emerit M. B. and Seguela P. (2003) Intracellular cross talk and physical interaction between two classes of neurotransmitter-gated channels. J. Neurosci. 23, 1246–1253. Boue-Grabot E., Toulme E., Emerit M. B. and Garret M. (2004a) Subunit-specific coupling between gamma-aminobutyric acid type A and P2X2 receptor channels. J. Biol. Chem. 279, 52517–52525. Boue-Grabot E., Emerit M. B., Toulme E., Seguela P. and Garret M. (2004b) Cross-talk and co-trafficking between rho1/GABA receptors and ATP-gated channels. J. Biol. Chem. 279, 6967–6975. Brandon N. J. and Moss S. J. (2000) Receptor cross talk: ligand-gated ion channels start to communicate. Sci. STKE 2000, PE1. Bray D. and Duke T. (2004) Conformational spread: the propagation of allosteric states in large multiprotein complexes. Annu. Rev. Biophys. Biomol. Struct. 33, 53–73. Chaumont S., Jiang L. H., Penna A., North R. A. and Rassendren F. (2004) Identification of a trafficking motif involved in the stabilization and polarization of P2X receptors. J. Biol. Chem. 279, 29628–29638. Galligan J. J. and Bertrand P. P. (1994) ATP mediates fast synaptic potentials in enteric neurons. J. Neurosci. 14, 7563–7571. Gmelin G. and Zimmermann M. (1983) Effects of gamma-aminobutyrate and bicuculline on primary afferent depolarization of cutaneous fibres in the cat spinal cord. Neuroscience 10, 869–874. Gu J. G. and MacDermott A. B. (1997) Activation of ATP P2X receptors elicits glutamate release from sensory neuron synapses. Nature 389, 749–753. Jo Y. H. and Role L. W. (2002) Coordinate release of ATP and GABA at in vitro synapses of lateral hypothalamic neurons. J. Neurosci. 22, 4794–4804. Jo Y. H. and Schlichter R. (1999) Synaptic corelease of ATP and GABA in cultured spinal neurons. Nat. Neurosci. 2, 241–245. Jonas P., Bischofberger J. and Sandkuhler J. (1998) Corelease of two fast neurotransmitters at a central synapse. Science 281, 419–424. Karanjia R., Garcia-Hernandez L. M., Miranda-Morales M., Somani N., Espinosa-Luna R., Montano L. M. and Barajas-Lopez C. (2006) Cross-inhibitory interactions between GABAA and P2X channels in myenteric neurones. Eur. J. Neurosci. 23, 3259–3268. Khakh B. S., Zhou X., Sydes J., Galligan J. J. and Lester H. A. (2000) State-dependent cross-inhibition between transmitter-gated cation channels. Nature 406, 405–410.

Khakh B. S., Fisher J. A., Nashmi R., Bowser D. N. and Lester H. A. (2005) An angstrom scale interaction between plasma membrane ATP-gated P2X2 and alpha4beta2 nicotinic channels measured with fluorescence resonance energy transfer and total internal reflection fluorescence microscopy. J. Neurosci. 25, 6911–6920. Labrakakis C., Tong C. K., Weissman T., Torsney C. and MacDermott A. B. (2003) Localization and function of ATP and GABAA receptors expressed by nociceptors and other postnatal sensory neurons in rat. J. Physiol. 549, 131–142. Landry M., Bouali-Benazzouz R., El Mestikawy S., Ravassard P. and Nagy F. (2004) Expression of vesicular glutamate transporters in rat lumbar spinal cord, with a note on dorsal root ganglia. J. Comp. Neurol. 468, 380–394. Lee F. J., Xue S., Pei L., Vukusic B., Chery N., Wang Y., Wang Y. T., Niznik H. B., Yu X. M. and Liu F. (2002) Dual regulation of NMDA receptor functions by direct protein-protein interactions with the dopamine D1 receptor. Cell 111, 219–230. Li Y. and Xu T. L. (2002) State-dependent cross-inhibition between anionic GABA(A) and glycine ionotropic receptors in rat hippocampal CA1 neurons. Neuroreport 13, 223–226. Li Y., Wu L. J., Legendre P. and Xu T. L. (2003) Asymmetric crossinhibition between GABAA and glycine receptors in rat spinal dorsal horn neurons. J. Biol. Chem. 278, 38637–38645. Liu F., Wan Q., Pristupa Z. B., Yu X. M., Wang Y. T. and Niznik H. B. (2000) Direct protein-protein coupling enables cross-talk between dopamine D5 and gamma-aminobutyric acid A receptors. Nature 403, 274–280. MacDermott A. B., Role L. W. and Siegelbaum S. A. (1999) Presynaptic ionotropic receptors and the control of transmitter release. Annu. Rev. Neurosci. 22, 443–485. Nakatsuka T. and Gu J. G. (2006) P2X purinoceptors and sensory transmission. Pflugers Arch. 452, 598–607. Pei L., Lee F. J., Moszczynska A., Vukusic B. and Liu F. (2004) Regulation of dopamine D1 receptor function by physical interaction with the NMDA receptors. J. Neurosci. 24, 1149–1158. Redman R. S. and Silinsky E. M. (1994) ATP released together with acetylcholine as the mediator of neuromuscular depression at frog motor nerve endings. J. Physiol. 477, 117–127. Reyes E. P., Alcayaga J. and Zapata P. (2006) Are there interactions between acetylcholine- and ATP-induced responses at the level of a visceral sensory ganglion? Brain Res. 1107, 97–103. Roberts J. A., Vial C., Digby H. R., Agboh K. C., Wen H., AtterburyThomas A. and Evans R. J. (2006) Molecular properties of P2X receptors. Pflugers Arch. 452, 486–500. Salter M. W. (2003) D1 and NMDA receptors hook up: expanding on an emerging theme. Trends Neurosci. 26, 235–237. Searl T. J., Redman R. S. and Silinsky E. M. (1998) Mutual occlusion of P2X ATP receptors and nicotinic receptors on sympathetic neurons of the guinea-pig. J. Physiol. 510, 783–791. Sokolova E., Nistri A. and Giniatullin R. (2001) Negative cross talk between anionic GABAA and cationic P2X ionotropic receptors of rat dorsal root ganglion neurons. J. Neurosci. 21, 4958–4968. Vial C., Roberts J. A. and Evans R. J. (2004) Molecular properties of ATP-gated P2X receptor ion channels. Trends Pharmacol. Sci. 25, 487–493. Vulchanova L., Riedl M. S., Shuster S. J., Stone L. S., Hargreaves K. M., Buell G., Surprenant A., North R. A. and Elde R. (1998) P2X3 is expressed by DRG neurons that terminate in inner lamina II. Eur. J. Neurosci. 10, 3470–3478. Zhou X. and Galligan J. J. (1998) Non-additive interaction between nicotinic cholinergic and P2X purine receptors in guinea-pig enteric neurons in culture. J. Physiol. 513, 685–697.
