The C-terminal Src Inhibitory Kinase (Csk)-mediated Tyrosine Phosphorylation Is a Novel Molecular Mechanism to Limit P2X3 Receptor Function in Mouse Sensory Neurons

The C-terminal Src Inhibitory Kinase (Csk)-mediated TyrosinePhosphorylation Is a Novel Molecular Mechanism to LimitP2X3 Receptor Function in Mouse Sensory Neurons*□S

Received for publication, February 20, 2009, and in revised form, May 19, 2009 Published, JBC Papers in Press, June 9, 2009, DOI 10.1074/jbc.M109.023051

Marianna D’Arco‡, Rashid Giniatullin§, Vanessa Leone‡¶, Paolo Carloni‡¶, Nicol Birsa‡, Asha Nair‡, Andrea Nistri‡,and Elsa Fabbretti‡�1

From the ‡Neurobiology Sector and Italian Institute of Technology Unit, International School for Advanced Studies (SISSA),34014 Trieste, Italy, the ¶INFM-DEMOCRITOS Modelling Centre for Research in Atomistic Simulation, via Beirut 4, 34014 Trieste, Italy, the§A. I. Virtanen Institute, University of Kuopio, 70211 Kuopio, Finland, and the �University of Nova Gorica, SI-5000 Nova Gorica, Slovenia

On sensory neurons, sensitization of P2X3 receptors gated byextracellular ATP contributes to chronic pain.We explored thepossibility that receptor sensitizationmay arise from down-reg-ulation of an intracellular signal negatively controlling receptorfunction. In view of the structural modeling between the Srcregion phosphorylated by the C-terminal Src inhibitory kinase(Csk) and the intracellular C terminus domain of the P2X3receptor, we investigated how Csk might regulate receptoractivity. Using HEK cells and the in vitro kinase assay, weobserved that Csk directly phosphorylated the tyrosine 393 res-idue of the P2X3 receptor and strongly inhibited receptor cur-rents. Onmouse trigeminal sensory neurons, the role of Cskwastightly controlled by the extracellular level of nerve growth fac-tor, a known algogen. Furthermore, silencing endogenous CskinHEKor trigeminal cells potentiated P2X3 receptor responses,confirming constitutive Csk-mediated inhibition. The presentstudy provides the first demonstration of an original molecularmechanism responsible for negative control over P2X3 receptorfunction and outlines a potential new target for trigeminal painsuppression.

ATP-activated P2X3 receptors are expressed almost exclu-sively bymammalian sensory neurons to play an important rolein the transduction of painful stimuli to the central nervoussystem (1). Activation of P2X3 receptors by ATP released dur-ing acute and chronic pain is thought to send nociceptive sig-nals to central pain-related networks (2). In view of the multi-tude of environmental stimuli normally reaching sensoryterminals, the question then arises how inappropriate activa-tion of P2X3 receptors is normally prevented. This process maycontribute to suppression of continuous pain sensation in con-junction with central synaptic inhibition.The molecular pathways triggered by algogenic substances

and responsible for modulating P2X3 receptor structure andfunction remain incompletely understood. This topic is of par-

ticular interest because it can provide original clues for novelapproaches related to treat pain. The nerve growth factor,NGF,2 is one of the most powerful endogenous substanceswhich elicit pain and inflammation via the tyrosine kinasereceptor TrkA (3). This neurotrophin stimulates an intracel-lular cascade that elicits PKC-dependent P2X3 receptorphosphorylation with ensuing facilitation of receptor cur-rents. Conversely, suppression of NGF signaling powerfullydown-regulates P2X3 receptor function (4). These observationsare consistent with the raised NGF levels in acute or inflamma-tory pain conditions (3). The molecular mechanisms underly-ing these effects remain, however, unclear.A dynamic balance between tyrosine phosphorylation and

dephosphorylation is a major factor controlling the activity ofmany neurotransmitter receptors (5). TrkA stimulation acti-vates intracellular signaling including Src tyrosine kinases (6)that, in neurons, are important modulators of ligand-gatedreceptors like nicotinic (7), NMDA receptors (8), and TRPV1receptors (9). All these receptors are involved inmediating var-ious types of pain in the spinal cord and sensory ganglia. Thereis, however, no available data on the role of tyrosine phospho-rylation on P2X3 receptor function.

The fundamental regulator of Src signaling is the C-terminalSrc kinase (Csk) that blocks it via tyrosine phosphorylation(Tyr-527, Refs. 10, 11). We explored whether tyrosine phos-phorylation might regulate P2X3 receptors of sensory neuronsby focusing on the P2X3 C-terminal domain Tyr-393 residue,which is included in a regionwith significant similarity with theCsk-phosphorylating region of Src. Our data demonstrate thatCsk activation induced an increased tyrosine (Tyr-393 residue)P2X3 receptor phosphorylation with decreased receptor func-tion, observed both in mouse trigeminal sensory neurons aswell as a cell expression system. We, thus, propose that Csk-mediated P2X3 receptor inhibition is a novel mechanism tolimit overactivation of P2X3 receptors.

EXPERIMENTAL PROCEDURES

Plasmids and Constructs—pCDNA3-P2X3 (rat sequence,NCBI accession number: CAA62594) was provided by Dr. A.

* This work was supported by grants from the Telethon Foundation(GGP07032), the Italian Institute of Technology (IIT), Fondi per gli Investi-menti della Ricerca di Base project (to A. Nistri), and the Slovenian ResearchAgency ARRS project (to E. F.).

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental Figs. S1–S3 and information.

1 To whom correspondence should be addressed. Fax: 386-5-331-5224;E-mail: [email protected] or [email protected].

2 The abbreviations used are: NGF, nerve growth factor; �,�-meATP, �,�-methylene-ATP; AU, arbitrary units; Csk, C-terminal Src inhibitory kinase;FSK, forskolin; ODG, n-octyl �-D-glucopyranoside; PKC, protein kinase C;TrkA, tyrosine kinase receptor; wt, wild type; PKA, cAMP-dependent kinase;siRNA, small interfering RNA; HEK, human embryonic kidney cells.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 284, NO. 32, pp. 21393–21401, August 7, 2009© 2009 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

AUGUST 7, 2009 • VOLUME 284 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 21393

North (University of Manchester, UK). pCDNA3-Csk (12) waskindly provided by Dr. X. Y. Huang (Cornell University).pGEX-rat P2X3C-terminal domain (13) was gently provided byDr. P. Seguela (McGill University). pCDNA3-P2X3 or pGEX-P2X3 mutants were obtained using the QuikChange mutagen-esis kit (Stratagene, La Jolla, CA) and the following primers:Y393A 5�-GACTCAGGGGCCGCTTCTATTGGTCACTAG-3�; Y393F 5�-GACTCAGGGGCCTTTTCTATTGGTCAC-TAG-3�; E384A 5�-TTCACCAGCGACGCGGCCACAGCG-GAG-3� and Q380A 5�-CAGGCCACAGCGGCGAAGCAGT-CCACCGAT-3�. Correct mutagenesis was confirmed byautomatedDNAsequencing, while correct expressionwas con-firmed by immunofluorescence microscopy experiments andWestern blotting.Cell Cultures and Transfection—Primary cultures of trigem-

inal ganglion neurons from C57-Black mice (12–14 days old)were cultured as described and used 24 h after plating (4). Thefollowing substances were added to the culture medium asrequired: anti-NGF neutralizing antibody (6 �g/ml, 1:5000;Sigma; 4); NGF (50 ng/ml; Alomone, Jerusalem, Israel), H89 (10�M, Sigma), forskolin (10 �M, Sigma). Inhibitors were pre-ap-plied for 30 min to serum-starved cultures. To block tyrosinephosphatases, living cells were incubatedwith sodiumpervana-date (50 �M Na3VO4, Sigma) for 10 min at 37 °C (9, 14).Transient transfection of pCDNA3-P2X3 receptors in

HEK293T cells was carried out with calcium/phosphatemethod. Mock transfection was obtained with co-transfectedwith pEGFP plasmid (Clontech, Mountain View, CA).In Vitro Kinase Assay—C-terminal P2X3 peptides (amino

acid residues 346–397) were used for in vitro kinase assay.Reactionswere performed using 1�g of substrate peptide and 1�g of recombinant kinase GST-Csk (Cell Signaling Technol-ogy, Danvers, MA) according to the manufacturer’s protocol.pGEX-P2X3 peptides were expressed in Escherichia coli BL21cells (GEHealthcare, Uppsala Sweden) and purified with aGSTSpinTrap Purification module (GE Healthcare). The qualityand quantity of input target peptides were checked with SDS-PAGE andCoomassie Blue staining. Activity of the commercialrecombinant Csk was determined by the Supplier (Cell Signal-ing) using a radiometric method in a kinase dose-dependentassay based on Csk HTScanTM kit (Cell Signaling).Protein Analysis, Immunoprecipitation, and Immunoblotting—

Trigeminal cultures were lysed in ODG buffer (2% n-octyl �-D-glucopyranoside, 1%Nonidet P-40, 10mMTris, pH 7.5, 150mM

NaCl, 100 mMNaF, 20 mM orthovanadate) plus protease inhib-itors mixture (Complete, Roche Applied Science). For phos-phorylation studies, proteins were extracted in ODG buffer,immunopurified in TNE buffer (10 mM Tris, 150 mM NaCl, 2mM EDTA plus 100 mMNaF, 20 mM orthovanadate, and prote-ase inhibitors) with rabbit anti-P2X3 or anti-Csk antibodies (0.5�g/ml, Santa Cruz Biotechnology, Santa Cruz, CA) and pulldownwith protein A/G PLUS-agarose (Santa Cruz Biotechnol-ogy) for 4 h at 4 °C. For co-immunoprecipitation, proteins wereextracted in the presence of 1% Triton X-100. Membrane pro-tein biotinylation experiments were performed as previouslydescribed (4). Total cellular membranes were purified by ultra-centrifugation at 100,000 � g for 1 h at 4 °C (14). Samples wereseparated on 10% polyacrylamide gel and processed for West-

ern immunoblotting using the following antibodies: anti-P2X3(1:300; Alomone), anti-Csk (1:500, Santa Cruz Biotechnology),anti-Src-p527 (1:500; Cell Signaling), anti-phospho-tyrosinehorseradish peroxidase (HRP)-conjugated (1:3500; clone Y20;Invitrogen, San GiulianoMilanese, Italy). To avoid detection ofimmunoglobulin heavy chains in Western blot, a mouse anti-rabbit IgG-HRP-conjugated (Jackson ImmunoResearch, Suf-folk, UK) was used as secondary antibody.Quality and correct reactivity of anti-Csk, anti-Src-p527 and

anti-Src-p416 antibodies were tested analyzing with Westernimmunoblotting HEK protein lysates after Csk overexpressioncells (not shown, 15).Loading controls were performed processing total lysates

withWestern immunoblotting or stripping (RestoreTM, Pierce)and successive probing. Western blot signals were detectedwith enhanced chemiluminescence light ECL (GE Healthcare).For quantification, band density was measured using ScionImage software.Co-immunofluorescence and Microscopy—Paraformalde-

hyde fixed trigeminal neurons were processed with a guineapig anti-P2X3 antibody (1:400, Neuromics, Edina,MN) and a rab-bit anti-Cskantibody (1:200,SantaCruzBiotechnology)ora rabbitanti-phospho-Src527 antibody (1:50, Cell Signaling). Immunoflu-orescence reactions were visualized using suitable Alexa-Fluor-conjugated secondary antibodies, or, for triple immu-nofluorescence reactions, with biotinylated antibody andstreptavidin-AlexaFluor647-conjugated (1:500, Invitrogen).Cells stainedwith secondary antibodies only, showedno immu-nostaining.Membrane in vivo labelingwas obtainedwithwheatgerm agglutinin (WGA) AlexaFluor488-conjugated (1:200,Invitrogen). Specimens were observed with a confocal LeicaTCS SP5 microscope (Ar-He, Ne laser). Quantitative analysiswas obtained with MetaMorph software (Molecular Devices,Downingtown, PA). Data are themean of at least five independ-ent experiments where an average of 260 cells was analyzed.RNA Silencing—For siRNA experiments, trigeminal neurons

(from 2 mouse) or HEK293 cells (5 � 104 cells/well in 12-wellplate) were transfected, respectively, withmouse or humanCsksiRNA SmartPools (100 nM, Dharmacon RNAi Technology,Lafayette, CO) using the DharmaFECTTM-1 transfection rea-gent (Dharmacon). For transfection efficiency control, cellswere transfected with scramble RNA and siGLO RISC-FreesiRNA (Dharmacon). Efficiency of Csk silencing was testedwith Western immunoblotting and immunofluorescence. 24 hafter silencing HEK cells were transfected with pEGFP orpCDNA3-P2X3 plasmids and analyzed 48 h later.Patch Clamp Recording—Trigeminal neurons or HEK cells

were recorded in whole-cell configuration as described (4). Intra-cellular solution of recording pipette contained (in mM): 140 KCl,0.5 CaCl2, 2 MgCl2, 2 Mg2ATP, 2 GTP, 10 HEPES, and 10 EGTA(pH 7.2). Cells were continuously superfused (2ml/min rate) withphysiological solution containing (in mM): 152 NaCl, 5 KCl, 1MgCl2, 2 CaCl2, 10 glucose, and 10 HEPES (pH 7.4).

Responses to selective P2X3 receptor agonist�,�-methylene-ATP (�,�-meATP, resistant to ectoATPase hydrolysis, Sigma)were measured in terms of peak amplitude. To express agonistpotency in terms of concentration producing 50% of the maxi-mum response (EC50 values), dose-response curves for �,�-

Csk Regulates P2X3 Receptors

21394 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 284 • NUMBER 32 • AUGUST 7, 2009

meATPwere constructed by applying different agonist doses tothe same cells and fitting them with a logistic equation (Origin6.0,Microcal, Northampton,MA). The onset of desensitizationwas estimated by calculating the first time constant of currentdecay (�fast) in the presence of agonist in accordance with ourprevious reports. Recovery from desensitization was assessedby paired-pulse experiments (4).Computational Modeling and Src Family Sequence

Alignment—To perform the sequence alignment, 346–397amino acid residues of C-terminal portion of the rat P2X3receptor (CAA62594) were aligned with the C-terminal aminoacid residues 478–533 of Src (Q9JJ10) involved in Csk/Srcinteraction. To this aim, we used ClustalW and T-COFFEEalgorithms (16, 17). After this step, computational homologymodeling, in silico P2X3-Csk docking and alanine scanningwere performed to identify candidate residues suitable formutagenesis within the putative Csk-docking site of the P2X3domain. Details are provided as supplemental information.Accordingly to computational predictions, we mutated poten-tial residues (Gln-380 and Glu-384) in P2X3 sequence thatmight destabilize Csk-P2X3 C-terminal complex and weexplored the consequences using Csk in vitro kinase assays.Statistics—Data are expressed asmeans� S.E., where n indi-

cates the number of experiments in molecular biology/immu-nocytochemistry or the number of investigated cells in electro-physiology. Statistical analysis was performed using theStudent’s t test, the Mann-Whitney rank sum test or theANOVA test, as appropriate. A p value of �0.05 was acceptedas indicative of a statistically significant difference.

RESULTS

Csk-mediated P2X3 Receptor Phosphorylation Leads toDown-regulation of Receptor Function—The tyrosine kinaseSrc is negatively controlled by the C-terminal Src kinase, Csk,via its docking to the Src C-terminal residues 504–518 (Fig. 1;11, 18). By exploring the C-terminal region of the P2X3 recep-tor, we could identify a sequence homologous to theC-terminalportion of Src specifically binding Csk (40% homology, 33.3%identical residues and 6.7% strongly similar residues, Fig. 1 andsupplemental Fig. S1).In particular, the P2X3 C-terminal Tyr-393 residue, a residue

specific for the P2X3 subtype and is not conserved in the otherP2X family members, appeared fully aligned with the target of

Csk, Src Y527. Furthermore computational modeling (19) sug-gested also similar accessibility of these residues to the solvent,rendering them suitable targets to a similar kinase.These observations prompted us to examine if P2X3 Tyr-393

was a target for Csk activity. Hence, we performed a cell-free invitro kinase assay plus aWestern blottingwith phosphotyrosineantibodies. Using recombinant Csk and GST-P2X3 C-terminaldomain purified from E. coli, we observed P2X3 tyrosine phos-phorylation (n � 5; Fig. 2A, lane 1). This signal was not detect-able when the same assay was run in the presence of a P2X3antibody recognizing the P2X3 peptide 383–397 (n� 3, Fig. 2B,lane 2) or when GST-P2X3Y393F mutant was tested as sub-strate (n � 5, Fig. 2B, lane 3), demonstrating thus that P2X3Tyr-393 was a target for Csk phosphorylation in vitro. Further-more, following our computational modeling data, we per-formed mutagenesis of the specific P2X3 residues Q380A andE384A within the P2X3 region compatible with putative Cskdocking (residues 371–386, see Fig. 1). This approach shouldhelp to explore Csk docking to the P2X3 subunits as suggestedby the in vitro kinase assay (Fig. 2C). These experiments dem-onstrated a significantly reduced phosphorylation of mutantsQ380A and E384A (50.2 � 6.6%; and 51 � 8.1% respectively,taken 100% wt signal, n � 3, Fig. 2C, lanes 2 and 3) in compar-ison with wt (lane 1), thus, supporting the existence of a Cskdocking site on the P2X3 receptor C-terminal domain.

Co-expression of full-length P2X3 receptors together withCsk in HEK cells was therefore carried out to investigate func-tional consequences of receptor tyrosine phosphorylation. Cskand P2X3 co-transfection led to robust co-expression of bothproteins (Fig. 2D, lane 2) with a significant increase in P2X3receptor tyrosine phosphorylation (Fig. 2D, lane 2; n � 5). Inaccordance with the notion that PKA-mediated phosphoryla-tion of Csk is essential to enable this kinase activity (20), weobserved that, in HEK cells, the PKA inhibitor H89 (10 �M; 30min) prevented P2X3 tyrosine phosphorylation (10 �M; 30min;Fig. 2D, lane 3, n � 4). It is noteworthy that basal adenylylcyclase in these cells was sufficient to trigger activation ofendogenous Csk, since H89 suppressed basal P2X3 tyrosinephosphorylation (supplemental Fig. S2A).In basal condition, a small portion of Csk and P2X3 receptor

interacted at cellular level, as observed by weak P2X3/Csk co-immunoprecipitation in HEK cells (Fig. 2E, lane 3). However,the co-immunoprecipitation signal was strongly detectedwhenusing the Y393F mutant (Fig. 2E, lane 2) despite the lack ofphosphorylation (see Fig. 2A, lane 2). These data suggested thatCsk and P2X3 receptors are part of a complex; however, highlydependent by phosphorylation state and conformation.Whole-cell patch clamp experiments on HEK cells co-trans-

fected with Csk and P2X3 indicated that P2X3 receptor-medi-ated currents evoked by the selective agonist �,�-meATP (�10�M)were significantly (p� 0.02 for 100�M, p� 0.05 for 10�M)smaller than those recorded from cells expressing the P2X3alone (Fig. 2G; n � 10), although this effect was not associatedwith a significant variation in membrane receptor expression(n� 3, p� 0.05; Fig. 2F). Such smaller receptor responses werenot accompanied by changes in the agonist EC50 value (indica-tive of drug receptor affinity, Table 1), in the Hill coefficient(indicative of the stoichiometry of drug receptor interaction,

FIGURE 1. Similarity of C-terminal domains of P2X3 and of c-Src. Aminoacid alignment of C-terminal domains of P2X3 and c-Src, natural substrate ofCsk, was examined. Note alignment of Src Tyr-527 and P2X3 Tyr-393 residues.The Src/Csk docking domain and residues involved in the docking are indi-cated. Identical (red), strongly similar (green), and weakly similar (blue) resi-dues are indicated. Within this domain, Src/P2X3 identical residues corre-spond to 33.3%, while 6.7% residues are strongly similar, and 26.7% residuesare weakly similar. The Q380A and E384A residues important for Csk dockingon P2X3 have been indicated.

Csk Regulates P2X3 Receptors

AUGUST 7, 2009 • VOLUME 284 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 21395

FIGURE 2. Csk interaction with P2X3 subunits in HEK cells yields increased receptor phosphorylation and decreased function. A, in vitro kinaseassay of GST-Csk and C-terminal P2X3 domain wt (GST-P2X3) or Y393F (GST-P2X3-Y393F). In vitro P2X3 tyrosine phosphorylation was visualized withWestern blot and anti-phosphotyrosine antibody. Input of recombinant P2X3 is shown (bottom lanes). B, neutralization of Csk activity was obtained withpolyclonal anti-C-terminal P2X3 antibodies (lower panel, lane 2 and histogram, *, p � 0.03). wt signal 0.52 � 0.15 AU, n � 5; wt plus peptide 0.17 � 0.015AU, n � 3; Y393F mutant 0.13 � 0.015 AU, n � 5. C, in vitro kinase assay of GST-Csk and C-terminal P2X3 domain wt (GST-P2X3) or mutants (GST-P2X3-Q380A and GST-P2X3-E384A) visualized with Western blot and anti-phosphotyrosine antibody. Input of recombinant P2X3 is shown (bottom lanes).D, Western immunoblots using an anti-phosphotyrosine antibody on immunopurified P2X3 receptors. P2X3 receptor phosphorylation increases withrespect to mock cells (lane 1) when P2X3 is co-expressed together with Csk (lane 2). No co-immunopurification of Csk and P2X3 receptor was observedin these conditions. H89 (10 �M; 30 min) blocks P2X3 receptor phosphorylation (lane 3). Middle and bottom lanes show total P2X3 and Csk expression.Histograms show P2X3 phosphotyrosine levels (n � 4, *, p � 0.05). Values are normalized with respect to the P2X3 input signals. E, examples of Csk/P2X3co-immunopurification show association between these two proteins in P2X3 and Csk co-expressing HEK cells (n � 3). Note higher co-purification levelswhen P2X3 receptor Y393F mutant is expressed (lane 2). P2X3 protein input used in immunoprecipitation is shown in the lower panel. F, membraneprotein biotinylation and Western immunoblotting reveal the amount of P2X3 receptor expressed at the cell membrane level in HEK cells mock-transfected or transfected with Csk. Histograms show values normalized versus total P2X3 receptors (0.55 � 0.21 AU for P2X3/mock and 0.75 � 0.34 AUfor P2X3/Csk expression, n � 3, p � 0.05). G, representative examples of currents induced by �,�-meATP (�,�, black bar) on HEK cells expressing P2X3 andmock plasmid or P2X3 together with Csk. Note smaller current amplitude after P2X3/Csk co-expression. The right panel shows dose-response curves for�,�-meATP-evoked currents in the two experimental conditions (P2X3 receptor plus mock plasmid, open squares; n � 10; co-expression of P2X3 plus Csk,filled squares; n � 10, *, p � 0.02 and #, p � 0.05). Note that co-expression of Csk depresses the P2X3- mediated current.

Csk Regulates P2X3 Receptors

21396 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 284 • NUMBER 32 • AUGUST 7, 2009

Table 1) or in the time constant of current decay (indicative ofdesensitization onset, 101 � 8 ms for control versus 126 � 10ms for P2X3/Csk co-expression; n � 17 and 13, respectively).

Likewise, no significant changes were observed in terms ofP2X3 current recovery from desensitization (tested with a dou-ble agonist pulse application, 4) in the presence (58 � 5% of theprevious response; n � 8) or in the absence of Csk (60 � 4% ofthe previous response, n � 18).

Globally, our results suggest that Csk phosphorylated P2X3receptors in cells and in vitro that was associated withdecreased receptor function, when tested with high agonistconcentrations, without concomitant alteration in receptorkinetic properties.P2X3 Tyr-393 Receptor Mutants Are Insensitive to the Action

of Csk in HEK Cells—Because Tyr-393 appeared to be the Csktarget in the in vitro assays, we mutated this residue to exploreits functional consequences with patch clamp recording fromtransfected HEK cells. Despite similar expression observed

with immunofluorescence micros-copy and membrane biotinylationassay (Fig. 3, A and B), the currentamplitude evoked by �,�-meATP(100 �M) was larger in the P2X3Y393A and Y393F mutants than inwt receptors (see examples andaverage data in Fig. 3C). The agonistdose response curves for P2X3mutants displayed enhanced maxi-mal effect, even though similar EC50values suggested analogous agonistaffinity (Table 1). Co-expression ofthe P2X3 mutants with Csk did notshowed significant changes eitherin receptor tyrosine phosphoryla-tion (n � 3, Fig. 3D) or currentamplitude (Fig. 3E, n � 13). Thesedata suggest that Tyr-393 wasimportant to exploit the Csk mod-ulation of P2X3 receptor func-tional response in an in vitroexpression system.NGF Neutralization Activated

Csk in Trigeminal Neurons—Tounderstand the physiological impli-cations of the interaction betweenCsk and P2X3 receptors, we studiedmouse trigeminal sensory neuronsthat constitutively express P2X3receptors potently modulated byNGF: indeed, NGF neutralizationdecreases trigeminal pain in vivoand P2X3 receptor currents in neu-ronal cultures (4). It seemed, there-fore, interesting to investigate thepotential contribution by Csk sig-naling to the action of NGF.On trigeminal sensory neurons

cultured in the presence of anti-NGF antibodies (4), membraneexpression of Csk was significantlyenhanced by NGF neutralization

FIGURE 3. C-terminal P2X3 receptor Tyr-393 residue is sufficient to control receptor properties and sensitivityto Csk. A, examples of HEK cells expressing the wt P2X3 receptors or the mutated forms Y393A or Y393F. Calibrationbar: 10 �m. B, membrane protein biotinylation and Western blot reveal wt or mutants P2X3 receptor expression atcell membrane level (see histograms, values normalized versus total P2X3 receptors; n � 6, 3 or 5). C, representativeexamples of currents induced by �,�-meATP (�,�, black bar, 100 �M) and dose-response curves obtained from HEKcellsexpressingwtP2X3 receptors (open squares; n�10)orY393AorY393Fmutants(filled triangles and filled squares,respectively; n �13 or 6; for 100 �M: p �0.03, #, p �0.05; for 10 �M: *, p �0.01; for 1 �M: *, p �0.008, #, p �0.01). Notelarger current amplitude for both mutated receptors. D, example of Western immunoblots of immunopurifiedY393A receptor expressed alone or together with Csk. Bottom lanes show input of P2X3 levels. Histograms quantifytyrosine phosphorylation levels in the Y393A mutant when expressed with or without Csk (0.21 � 0.04 and 0.16 �0.03 AU for tyrosine phosphorylation signals in the absence and in the presence of Csk co-expression, respectively,n � 3). Experiments performed with the Y393F mutant showed similar results (0.4 � 0.23 and 0.32 � 0.14 AU fortyrosine phosphorylation in the absence and in the presence of Csk, respectively, n � 4). E, dose-response curves for�,�-meATP evoked currents of Y393A receptors expressed a with mock plasmid (filled circles, n � 13) or togetherwith Csk (open circles, n � 12).

TABLE 1Characteristics of P2X3 receptors and its Tyr-393 mutantsThe table shows the kinetic properties of P2X3 receptors and its 393 mutants.

Receptor EC50 Hill coefficient Na

�M

P2X3 wt 2.05 � 0.47 1.06 � 0.11 10Wt/Csk 1.74 � 0.51 1.05 � 0.13 10Y393A 1 � 0.21 1.24 � 0.16 13Y393F 1.56 � 0.66 0.89 � 0.16 6Y393A/Csk 1.02 � 0.25 1.23 � 0.14 12Wt siRNA mock 2.44 � 0.92 1.10 � 0.42 6Wt siRNA Csk 1.67 � 0.34 1.21 � 0.42 10

a N indicates the number of neurons tested for these analyses.

Csk Regulates P2X3 Receptors

AUGUST 7, 2009 • VOLUME 284 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 21397

(Fig. 4A) together with a strongerCsk activity as demonstrated byconcomitant Src inhibition (seeanti-phospho-Src527 antibody andFig. 4A).To further validate Csk activation

we examined, as a reliable index ofthis process, Csk translocation tothe neuronal membrane (21). Con-focal microscopy showed, afterNGF neutralization, redistributionof Csk to membrane (Fig. 4B).Quantification of microscopy ex-periments showed a significantlylarger number of P2X3 receptorexpressing neurons with a ring-likedistribution of Csk (Fig. 4B, p �0.009). Conversely, rapid disappear-ance of membrane-bound Csk wasinduced when exogenous NGF(100 ng/ml for 5 or 15 min) wasapplied to anti-NGF antibody-treated mouse trigeminal neurons(p � 0.02 n � 3, Fig. 4C), indicatinga close relation between extracellu-lar NGF levels and Csk location atthe membrane level.Following neutralization of extra-

cellular endogenous NGF with anti-NGF antibodies (24 h, 4), mouse tri-geminal neurons showed a largeincrease in tyrosine phosphoryla-tion of their P2X3 receptors (p �0.03, Fig. 4D). The effects of NGFdeprivation on P2X3 phosphoryla-tion in culture were even more evi-dent following tyrosine phosphataseinhibition with pervanadate (250�M; 25 min; 0.79 � 0.1 AU in con-trol versus 2.24� 0.58 AU after per-vanadate;n� 3; p� 0.001), suggest-ing that tyrosine phosphorylationrather than dephosphorylation wasan important mechanism for theaction by Csk on P2X3 receptors.On mouse trigeminal neurons,application of H89 (10 �M; 30 min)induced a slight increase of the con-stitutive N-terminal threoninephosphorylation of P2X3 receptors(0.24 � 0.02 absolute gray levels forcontrol and 0.26 � 0.003 after H89treatment, n � 3; supplemental Fig.S2B) previously described as attrib-uted PKC activity (4). These resultscannot exclude that, on trigeminalneurons, PKC or PKA should exert,at least partially, distinct and con-

FIGURE 4. Neutralization of NGF activates Csk in mouse trigeminal neurons. A, purified membrane extractsfrom trigeminal neurons grown in control conditions or after application of anti-NGF antibody, tested withanti-Csk or anti-phospho-Src527 antibodies. The amount of total Src in membrane preparation is shown.Loading control was done with anti-�-actin antibody. Histograms show quantification of Western blot for Cskor Src527 optical density values (n � 5 experiments, *, p � 0.04 and *, p � 0.02, respectively). Signals of Src-p527are normalized with respect to total Src. B, confocal images show cellular immunoreactivity of Csk (red) incontrol conditions (upper panels) and in the absence of NGF (lower panels). Membrane staining was obtainedwith in vivo labeling with WGA (green). Note Csk re-localization at membrane level after NGF neutralization.P2X3 receptors immunoreactivity (blue) was also shown. Calibration bar: 5 �m. Histograms quantify neuronswith ring-like localization of Csk as percent of total of P2X3 immunoreactive neurons (n � 60 or 68 cells, *, p �0.02). A total of twenty optical stack sections for each cell were analyzed (at 1-�m interval). C, decrease in Cskmembrane localization after acute NGF (100 ng/ml, 5 or 15 min) applied to NGF-deprived (24 h) cultures.Amount of input P2X3 is also shown. Histograms show quantification of the effect in the different conditions(from 4.5 � 0.75 AU versus 2.28 � 0.30, *, p � 0.03, n � 4 after 5 min of NGF application or versus 1.52 � 0.32 AU,p � 0.02 n � 3 after 15 min of NGF application, n � 3 for 15 min NGF; *, p � 0.03). D, Western immunoblots ofimmunopurified P2X3 receptor tyrosine phosphorylation (pTyr) from trigeminal cultures grown (24 h) in con-trol conditions (lane 1) or after applying the anti-NGF antibody (lane 2). P2X3 inputs are shown (bottom lanes).Histograms show pTyr values expressed normalized with respect to the P2X3 input signal (0.79 � 0.1, AU incontrol, n � 10, versus 1.17 � 0.12 AU after anti-NGF, n � 9; *, p � 0.03).

Csk Regulates P2X3 Receptors

21398 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 284 • NUMBER 32 • AUGUST 7, 2009

trasting effects of the threonine and tyrosine phosphorylationstate of P2X3 receptors.Role of Endogenous Csk on P2X3-mediated Currents of HEK

Cells and Trigeminal Neurons—We tested the hypothesis thatendogenous Csk exerts a constitutive control over P2X3 recep-tor function. Because transgenic mice lacking Csk are not vital(22), we examined our hypothesis by silencing Csk in HEK cellsas well as in mouse trigeminal neurons (Fig. 5). Immunofluo-rescence andWestern blotting demonstrated efficient silencingof Csk in HEK and in trigeminal ganglion cultures (n � 3, Fig.5A and supplemental Fig. S2B).Western blotting with a phosphotyrosine antibody demon-

strated that, in P2X3-expressing-HEK cells, there was noincrease in P2X3 receptor tyrosine phosphorylation after

silencing endogenous Csk (Fig. 5B),together with a significant rise inthe maximal amplitude of the �,�-meATP-evoked current with re-spect to responses measured frommock silenced cells (p � 0.04 for 10�M and p � 0.02 for 100 �M �,�-meATP, Fig. 5C). To explore anypotential activity of residual Csk leftafter silencing in HEK cells, wedecided to maximize the cAMP/PKA-mediated stimulation of en-dogenous residual Csk with forsko-lin application (10 �M, 5 min;known to be a strong Csk activator;Ref. 23) to either mock or siRNA-transfected HEK cells. Under theseconditions, no significant change inP2X3 receptor phosphorylation orreceptor current was found (supple-mental Fig. S3), suggesting efficientsilencing of endogenous Csk.We next tested the consequences

of Csk silencing on the responses ofP2X3 receptors of trigeminal neu-rons (Fig. 5D). Under these condi-tions, controlled experiments run inparallel with neurons treated withsiRNACsk or control non-targetingsiRNA showed that the currentamplitude evoked by 10 �M �,�-meATP was significantly largerafter Csk silencing (n � 29 in con-trol and n � 24 cells for Csk siRNA;p � 0.04; Fig. 5D). These data indi-cate that endogenous Csk wasessential to curtail constitutive,functional overactivation of P2X3receptors in an expression systemand in trigeminal neurons.

DISCUSSION

The principal finding of the pres-ent study is the demonstration that

the kinase Csk exerted a powerful inhibitory control on thefunction of P2X3 receptors via phosphorylation of the P2X3Tyr-393 residue. These data shed light on a new mechanismthat controls the efficiency of signaling by pain-sensing P2X3receptors. It is tempting to speculate that Csk activity repre-sents a key endogenous brake to avoid inappropriate paintransduction.Multiple Phosphorylation Processes Control the Operation of

P2X3 Receptors in Trigeminal Neurons—In analogy with otherionotropic receptors (24), phosphorylation of P2X3 receptors isamechanism tomodulate how effectively this protein can senseits ligands (4). In keeping with this notion, we have recentlyshown that NGF enhances PKC-mediated threonine phospho-rylation of P2X3 receptors with resulting gain of function (4).

FIGURE 5. Endogenous Csk has a constitutive inhibitory role on P2X3 receptors. A, Western blots showingvery low expression of Csk after Csk siRNA in trigeminal neurons and HEK cells. Loading control is confirmed by�-actin signals (bottom lanes). B, example of Western blot showing tyrosine phosphorylation (pTyr) of P2X3receptors in relation to silencing. After Csk siRNA, there is no change in total P2X3 receptor signal, disappear-ance of Csk and minimal pTyr of the P2X3 subunits. C, representative examples of currents induced by �,�-meATP (�,�, black bar, 100 �M) and dose-response curves from HEK cells expressing wt P2X3 receptors incontrol conditions (open circles; n � 11) or after Csk siRNA (filled circles; n � 6) that enhances responses to�,�-meATP, revealing background depression by endogenous Csk (*, p � 0.02 at 100 �M; *, p � 0.04 at 10 �M).D, �,�-meATP (�,�, 10 �M) evoked currents recorded from trigeminal neurons after mock (n � 29) or Csksilencing (n � 24). Histograms show quantification of mean current amplitude values (333 � 57 pA for controln � 29 cells; 559 � 100 pA for Csk siRNA, n � 24 cells; *, p � 0.04).

Csk Regulates P2X3 Receptors

AUGUST 7, 2009 • VOLUME 284 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 21399

Less common is, however, to observe phosphorylation mecha-nisms to depress receptor activity. The present study has indi-cated, for the first time, how an important contribution to theNGF-mediated action on P2X3 receptor function was themod-ulation of Csk activity. Furthermore, native P2X3 receptors oftrigeminal neurons generated significantly larger membranecurrents after Csk silencing, confirming that these receptorswere kept under constitutive inhibition by this kinase activity.Because anti-NGF treatment produces mouse trigeminal anal-gesia in vivo (4), depression of P2X3 receptor activity by tyrosinephosphorylation could contribute to lessened pain, despiteunchanged receptor expression. In general, our data extend therole of Csk as a negative regulator of pain, because Csk wouldalso inhibit the Src-mediated facilitation of TRPV1 (9) andNMDA (8) receptors that are important contributors to painsensitivity.Interaction of Csk with P2X3 Receptors—Co-purification of

Csk and P2X3 subunits as well as in vitro kinase assay demon-strated that a certain fraction of receptors directly interactswithCsk. Nevertheless, because Csk recognizes a tertiary struc-ture of intracellular domains without a common consensussequence (25), it was difficult a priori to demonstrate a discretesite for Csk binding. Starting from the sequence alignmentbetween P2X3 and the Csk target domain of Src (11), wedetected a compatible Csk binding region in the P2X3 C-tail.When we mutated the key P2X3 residues Q380A and E384Awithin the region compatible with Csk docking, the in vitroCsk-mediated phosphorylationwas prevented. Theweak phos-phorylated signal for the wt P2X3 C-terminal domain in com-parison with total substrate available was likely to be due to theabsence of the complete receptor conformation obtainable onlyunder in vivo conditions. Furthermore, although very few non-Src targets for Csk have been reported (26), we cannot com-pletely exclude other indirect mechanisms coexisting in neu-rons to explain Csk-mediated P2X3 phosphorylation. Forinstance, certain signaling elements (or adaptor proteins) of theSrc pathways, as well as cytoskeleton elements, could contrib-ute to the action of Csk on P2X3 receptors. Nonetheless, it isunnecessary to assume the role of an unknown kinase in P2X3phosphorylation because P2X3 receptors exhibited poor tyro-sine phosphorylation after Csk silencing, a condition thatshould have per sehyper-activated other Src family kinases (27).Furthermore, P2X3 receptor tyrosine hyperphosphorylationand reduced functional activity were observed even after phar-macological inhibition of phosphatases, suggesting that P2X3receptors displayed regulatorymechanisms distinct from otherpurinergic receptors, like P2X7 (28).The Csk-mediated reduction of P2X3 receptor function was

observedwhen a substantial fraction of receptors was activated.In analogy with P2X2 receptors (29), one might speculate thatphosphorylation of trimeric P2X3 receptors could change theoverall electrical charge of this domain, leading to a receptorconformation associated with impaired channel function andefficacy.The Csk-mediated control over the P2X3 current amplitude

was unlikely to be caused by concomitant changes in receptorsensitivity because of the unchanged threshold and slope of theagonist dose response plots. Likewise, no alteration in desensi-

tization kinetics was detected, suggesting that the modulatoryrole of Csk was different from the one seen, for example, withthe peptide NGF (4). In view of the well known compartmen-talization of Csk and Src to membrane lipid rafts (30), it is fea-sible to presume that already described discrete redistributionof P2X3 receptors between rafts and non-raft domains (31)might determine the efficiency of the global cell response to theP2X3 agonist.ANewRole for theP2X3ReceptorC-terminalResidueTyr-393—

Electrophysiological studies indicate the high sensitivity of theP2X receptor family to sequence alterations because pointmutations of conserved or nonconserved residues generatedramatic differences in receptor function (32). In the presentstudy, the tyrosine residue Tyr-393, nonconserved among theP2X family members, was a gain setter for the P2X3 receptorfunction. Mutating this residue rendered the receptor insensi-tive to the depressant action of Csk and limited P2X3 tyrosinephosphorylation, suggesting that the Tyr-393 residue was a tar-get for Csk, necessary and sufficient for its activity on P2X3receptors (see model in Fig. 6, A and B). To the best of ourknowledge, this is the only available example of a powerful gainof function of P2X3 receptorswith a single amino acidmutationof the intracellular domains, without concomitantly alteringthe receptor kinetic properties.Under basal conditions, namely the absence of strong stim-

ulation of P2X3 receptors, it is possible to hypothesize that theconstitutive activity of PKA was an important factor to ensure

FIGURE 6. Proposed mechanism of action of Csk on P2X3 receptors oftrigeminal sensory neurons. A, during pain states, excess NGF binds TrkAreceptors to stimulate downstream signaling (including PKC), a process thattogether with enhanced release of ATP, strongly activates P2X3 receptors.PKC-mediated phosphorylation of threonine residues of the N-terminaldomain of P2X3 subunit, together with relief of Csk inhibition of P2X3 recep-tors amplifies functional receptor responses. The outcome of all these effectsis an increase in nociceptive neuronal firing and pain. B, under pain-free con-ditions when levels of ATP and NGF are low, PKA-mediated Csk activity isenvisioned to inhibit Src (via Tyr-527 phosphorylation) and to depress P2X3receptor responses (via Tyr-393 phosphorylation). This process intrinsicallylimits P2X3 receptor activity.

Csk Regulates P2X3 Receptors

21400 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 284 • NUMBER 32 • AUGUST 7, 2009

Csk-mediated inhibition of P2X3 receptor function (Fig. 6B).During pain states, onemight assume that elevated intracellularCa2� should facilitate the action of PKAonCsk: however, this isactually unlikely to take place because, under such conditions,strong activation of Src has been reported (8) to imply that Cskhad transiently lost its function.Under such circumstances, the prevailing phosphorylation

state of the P2X3 receptor is proposed to be PKC-mediated (Fig.6A) with facilitation of the receptor currents. Indeed, the pos-sibility of mutually exclusive effects of PKA and PKC on theP2X3 phosphorylation state and functional responses makesthese kinases as potentially attractive yin and yang regulators ofpain sensitivity (Fig. 6, A and B).AScheme for Regulation of P2X3Receptors byTyrosineKinase—

In chronic pain states, the strong activation of P2X3 receptorsby large concentrations of ATP (2) is potentiated by algogenicsubstances like NGF (34) via complex multifactorial pathwaysthat, as shown in the present study, include relief of Csk-medi-ated constitutive P2X3 inhibition (Fig. 6).Because ATP-activated P2X3 receptors transduce acute

(mechanical) and chronic (neuropathic) painful stimuli to thebrain (33), inappropriate activation of this system may evokepain hypersensitivity (hyperalgesia) even to non-painful stimuli(allodynia), a condition typical of severe pain disorders, includ-ing trigeminal neuralgia. If Csk is one endogenous inhibitor ofP2X3 receptor signaling,manipulatingCsk activity could prom-ise a novel approach to pain suppression. This is an attractivehypothesis since Csk can target strong responses mediated byP2X3 receptors. The outcome of the action of Csk would be tocreate a low-pass filter for less painful stimuli that could benormally relayed without altering the sensory system gain.

Acknowledgments—We thank P. Seguela for pGEX plasmid, P.Magerfor structural modeling data of P2X3 receptors, and A. Gnanaseka-ran, R. Abbate, and the staff of Centro Servizi Polivalenti di Ateneo(University of Trieste) for help.

REFERENCES1. Souslova, V., Cesare, P., Ding, Y., Akopian, A. N., Stanfa, L., Suzuki, R.,

Carpenter, K., Dickenson, A., Boyce, S., Hill, R., Nebenuis-Oosthuizen, D.,Smith, A. J., Kidd, E. J., and Wood, J. N. (2000) Nature 407, 1015–1017

2. Burnstock, G. (2008) Nat. Rev. Drug Discov. 7, 575–5903. Pezet, S., and McMahon, S. B. (2006) Annu. Rev. Neurosci. 29, 507–5384. D’Arco, M., Giniatullin, R., Simonetti, M., Fabbro, A., Nair, A., Nistri, A.,

and Fabbretti, E. (2007) J. Neurosci. 27, 8190–82015. Kalia, L. V., Pitcher, G. M., Pelkey, K. A., and Salter, M. W. (2006) EMBO

J. 25, 4971–49826. Abram, C. L., and Courtneidge, S. A. (2000) Exp. Cell Res. 254, 1–137. Charpantier, E., Wiesner, A., Huh, K. H., Ogier, R., Hoda, J. C., Allaman,

G., Raggenbass, M., Feuerbach, D., Bertrand, D., and Fuhrer, C. (2005)J. Neurosci. 25, 9836–9849

8. Liu, X. J., Gingrich, J. R., Vargas-Caballero, M., Dong, Y. N., Sengar, A.,Beggs, S., Wang, S. H., Ding, H. K., Frankland, P. W., and Salter, M. W.(2008) Nat. Med. 14, 1325–1332

9. Zhang, X., Huang, J., and McNaughton, P. A. (2005) EMBO J. 24,4211–4223

10. Nada, S., Okada,M.,MacAuley, A., Cooper, J. A., andNakagawa,H. (1991)Nature 351, 69–72

11. Levinson, N. M., Seeliger, M. A., Cole, P. A., and Kuriyan, J. (2008) Cell134, 124–134

12. Lowry, W. E., Huang, J., Ma, Y. C., Ali, S., Wang, D., Williams, D. M.,Okada, M., Cole, P. A., and Huang, X. Y. (2002) Dev. Cell 2, 733–744

13. Bernier, L. P., Ase, A. R., Chevallier, S., Blais, D., Zhao,Q., Boue-Grabot, E.,Logothetis, D., and Seguela, P. (2008) J. Neurosci. 28, 12938–12945

14. Khanna, S., Roy, S., Park, H. A., and Sen, C. K. (2007) J. Biol. Chem. 282,23482–23490

15. Zhang, S. Q., Yang,W., Kontaridis, M. I., Bivona, T. G.,Wen, G., Araki, T.,Luo, J., Thompson, J. A., Schraven, B. L., Philips, M. R., and Neel, B. G.(2004)Mol. Cell 13, 341–355

16. Thompson, J. D., Higgins, D.G., andGibson, T. J. (1994)Nucleic Acids Res.22, 4673–4680

17. Lesk, A. M. (2002) Introduction to Bioinformatics, pp. 160–215, Oxford,NY

18. Cowan-Jacob, S. W., Fendrich, G., Manley, P. W., Jahnke, W., Fabbro, D.,Liebetanz, J., and Meyer, T. (2005) Structure 13, 861–871

19. Mager, P. P., Weber, A., and Illes, P. (2004) Curr. Top. Med. Chem. 4,1657–1705

20. Abrahamsen, H., Vang, T., and Tasken, K. (2003) J. Biol. Chem. 278,17170–17177

21. Kawabuchi, M., Satomi, Y., Takao, T., Shimonishi, Y., Nada, S., Nagai, K.,Tarakhovsky, A., and Okada, M. (2000) Nature 404, 999–1003

22. Imamoto, A., and Soriano, P. (1993) Cell 73, 1117–112423. Vang, T., Torgersen, K.M., Sundvold, V., Saxena,M., Levy, F.O., Skålhegg,

B. S., Hansson, V., Mustelin, T., and Tasken, K. (2001) J. Exp. Med. 193,497–507

24. Smart, T. G. (1997) Curr. Opin. Neurobiol. 7, 358–36725. Wang, D., Huang, X. Y., and Cole, P. A. (2001) Biochemistry 40,

2004–201026. Baumeister, U., Funke, R., Ebnet, K., Vorschmitt, H., Koch, S., and Vest-

weber, D. (2005) EMBO J. 24, 1686–169527. Nada, S., Yagi, T., Takeda, H., Tokunaga, T., Nakagawa, H., Ikawa, Y.,

Okada, M., and Aizawa, S. (1993) Cell 73, 1125–113528. Kim,M., Jiang, L. H.,Wilson,H. L., North, R. A., and Surprenant, A. (2001)

EMBO J. 20, 6347–635829. Fujiwara, Y., and Kubo, Y. (2006) J. Physiol. 576, 135–14930. Oneyama, C., Hikita, T., Enya, K., Dobenecker, M. W., Saito, K., Nada, S.,

Tarakhovsky, A, and Okada, M. (2008)Mol. Cell 30, 426–43631. Vacca, F., Giustizieri, M., Ciotti, M. T., Mercuri, N. B., and Volonte, C.

(2009) J. Neurochem. 109, 1031–104132. Vial, C., Roberts, J. A., and Evans, R. J. (2004) Trends Pharmacol. Sci. 25,

487–49333. Donnelly-Roberts, D., McGaraughty, S., Shieh, C. C., Honore, P., and

Jarvis, M. F. (2008) J. Pharmacol. Exp. Ther. 324, 409–41534. Giniatullin, R., Nistri, A., and Fabbretti, E. (2008) Mol. Neurobiol. 37,

83–90

Csk Regulates P2X3 Receptors

AUGUST 7, 2009 • VOLUME 284 • NUMBER 32 JOURNAL OF BIOLOGICAL CHEMISTRY 21401