Calcium/calmodulin-dependent serine protein kinase (CASK) is a new intracellular modulator of P2X3 receptors


Address correspondence and reprint requests to Elsa Fabbretti, Center for biomedical sciences and engineering, University of Nova Gorica, SI-5000 Nova Gorica, Slovenia. E-mail:


ATP-gated P2X3 receptors of sensory ganglion neurons are important transducers of painful stimuli and are modulated by extracellular algogenic substances, via changes in the receptor phosphorylation state. The present study investigated the role of calcium/calmodulin-dependent serine protein kinase (CASK) in interacting and controlling P2X3 receptor expression and function in mouse trigeminal ganglia. Most ganglion neurons in situ or in culture co-expressed P2X3 and CASK. CASK was immunoprecipitated with P2X3 receptors from trigeminal ganglia and from P2X3/CASK-cotransfected human embryonic kidney (HEK) cells. Recombinant P2X3/CASK expression in HEK cells increased serine phosphorylation of P2X3 receptors, typically associated with receptor upregulation. CASK deletion mutants also enhanced P2X3 subunit expression. After silencing CASK, cell surface P2X3 receptor expression was decreased, which is consistent with depressed P2X3 currents. The reduction in P2X3 expression levels was reversed by the proteasomal inhibitor MG-132. Moreover, neuronal CASK/P2X3 interaction was up-regulated by nerve growth factor (NGF) signaling and down-regulated by P2X3 agonist-induced desensitization. These data suggest a novel interaction between CASK and P2X3 receptors with positive outcome for receptor stability and function. As CASK-mediated control of P2X3 receptors was dependent on the receptor activation state, CASK represents an intracellular gateway to regulate purinergic nociceptive signaling.

Abbreviations used



time constant of fast current decay


time constant of slow decay of the current


calcium/calmodulin-dependent kinase


calcium/calmodulin-dependent serine kinase


trigeminal ganglia

P2X3 receptors are ATP-gated membrane proteins, expressed predominantly by sensory neurons, and are important transducers of peripheral nociceptive signals (Wirkner et al. 2005; Burnstock 2006). Previous studies have indicated that P2X3 receptors undergo rapid activity-dependent turnover (Xu and Huang 2004; Vacca et al. 2009; Pryazhnikov et al. 2011), a process that requires a complex series of events, including oligomerization of protein subunits as well as post-translational folding and modification (Royle and Murrell-Lagnado 2003). Several intracellular kinases modulate the expression and function of P2X3 receptors of sensory neurons by discrete phosphorylation of intracellular domains of this receptor (Fabbretti and Nistri 2012). We have previously shown that calcium/calmodulin-dependent kinase CaMKII and the Ser/Thr kinase Cdk5 are highly involved in this process (Nair et al. 2010a, b). Because adaptor and scaffold molecules are likely to participate in the compartmentalized signal transduction essential for receptor stabilization and function, we investigated the potential role of calcium/calmodulin-dependent serine protein kinase (CASK), a membrane protein belonging to the membrane-associated guanylate kinase (MAGUK) family, in regulation of P2X3 expression and function. CASK is mainly known for its function as a neuronal scaffolding protein (Cohen et al. 1998; Hsueh 2009). CASK deficiency impairs synaptic transmission and dendritic spine formation (Lu et al. 2003; Zordan et al. 2005; Hsueh 2006; Chao et al. 2008; Sun et al. 2009; Slawson et al. 2011) as CASK expression is essential for receptor trafficking to membrane level (Butz et al. 1998; Jeyifous et al. 2009) and the link between plasma membrane to cytoskeletons (Cohen et al. 1998; Chao et al. 2008).

We wondered if CASK might have a role also in the regulation of the expression and function of P2X3 receptors in sensory ganglion neurons. For this purpose, using silencing experiments we check role of CASK in P2X3 stability and with immunoprecipitation experiments we investigated the association of CASK and P2X3 in trigeminal sensory ganglia in situ and in culture. Furthermore, we explored the dynamic nature of the interaction of CASK and P2X3 and the impact of CASK on P2X3 receptor function.

Materials and methods

Tissue and culture preparation

Postnatal day 12–15 C57-Black/6J mice were fully anesthetized by slowly raising levels of CO2 and decapitated, a procedure in accordance with the Italian Animal Welfare Act and approved by the Ethical Committee of SISSA. The animals were from both sexes and were obtained from in-house animal facility at SISSA. Trigeminal ganglia were excised and processed for molecular and cell biology experiments or dissociated (10–20 min at 37°C) in a solution containing 0.25 mg/mL trypsin, 1 mg/mL collagenase, and 0.2 mg/mL DNase (Sigma, Milan, Italy) in F-12 medium (Invitrogen, San Diego, CA, USA) to prepare primary cultures. Cultured cells were used 24 h after plating (Simonetti et al. 2006) unless otherwise indicated. In a set of experiments, cultures were incubated with a neutralizing anti-NGF antibody (12 ng/mL, 14 h; Sigma) to suppress endogenous nerve growth factor (NGF)-mediated signaling (D'Arco et al. 2007, 2009), or with NGF (100 ng/mL, 5 min; Alomone, Jerusalem, Israel) applied immediately before cell lysis.

Cell culture and transfection

Human embryonic kidney (HEK)-293T cells were obtained from our in-house cell bank and used as previously described. For transfection experiments, the following DNA plasmids were used: pEGFP (Clontech, Mountain View, CA, USA), pcDNA3-P2X3 (rat sequence, NCBI accession number: CAA62594), pcDNA3-CASK (kindly provided by Dr. Li-Huei Tsai, Cambridge, MA, USA) and the CASK deletion mutants (Gene Bank Accession number: 447110; Chao et al. 2008) ΔCAMK (Δ3–323), ΔPDZ (Δ488–568), ΔSH3 (Δ604–665), ΔGK (Δ721–895). In selected experiments, Myc-tag CASK plasmid (Hsueh et al. 1998) was used. As in preliminary experiments ΔPDΖ and ΔSH3 CASK showed limited expression in HEK cells, in this instance we used 10% larger concentration of plasmid transfection with respect to the standard protocol (Sundukova et al. 2012).

Immunofluorescence and microscopy

For immunofluorescence staining, paraformaldehyde-fixed ganglia or cultures were immunostained with antibodies against the P2X3 receptors (dilution 1 : 200; Alomone), anti-CASK (MAB5230; dilution 1 : 100; Millipore, Milan, Italy), and neuron-specific β-tubulin III (dilution 1 : 300; Sigma). 4′,6-Diamidino-2-phenylindole (DAPI) was used for nuclear staining. Immunofluorescence reactions were visualized using as secondary antibodies AlexaFluor 488, AlexaFluor 546, or Streptavidin 647 (dilution 1 : 500; Invitrogen). Images were analyzed and quantified using a Zeiss fluorescence microscope or Leica confocal microscope (Leica Microsystems, Heidelberg, Germany) and dedicated software. The perimembrane region was defined as the one within 5 μm from the cell surface. Unless otherwise stated, an average of 500 cells was analyzed in each test, and data are the mean of three independent experiments.

siCASK silencing

For siRNA experiments, cultured trigeminal neurons (from two mice) were transfected with mouse CASK siRNA SmartPools (100 nM; Dharmacon RNAi Technology, Lafayette, CO, USA) using the DharmaFECTTM-1 transfection reagent (Dharmacon). For transfection efficiency control, cells were transfected with scramble RNA and siGLO RISC-Free siRNA (Dharmacon). Efficiency of CASK silencing was tested with western immunoblotting as exemplified in Figure S1. Forty eight hours after silencing, cells were used for protein expression and patch clamping experiments.

Protein lysates, immunoprecipitation and immunoblotting

For western blotting (WB) and immunoprecipitation (IP) experiments, proteins from neuronal cultures were extracted in TNE 1.5X buffer (10 mM Tris-HCl at pH 7.5, 150 mM NaCl, 2 mM EDTA, 100 mM NaF, 20 mM Na orthovanadate, and 1% Triton-X 100) plus protease inhibitors (Sigma) (D'Arco et al. 2009). The following antibodies were used: anti-P2X3 (WB, 1 : 300; Alomone), anti-P2X3 (H-60, IP, 1 : 500; Santa Cruz Biotechnology, Santa Cruz, CA, USA); anti-β-tubulin III (1 : 2000; Sigma); anti-β-actin (1 : 3000; Sigma); anti-CASK (C-19 or H-107, 1 : 500; Santa Cruz); anti-phospho-serine antibody (1 : 600; Millipore) and anti-Myc (9B11; IP, 1 : 1000; WB, 1 : 4000; Cell Signaling, Danvers, MA, USA). Western blot signals were detected with the enhanced chemiluminescence light system (GE Healthcare, Milano, Italy). For quantification of intensities of the immunoreactive protein bands (expressed in optical density absolute units, AU), we used Scion Image software (NIH, Bethesda, MD, USA) or the digital imaging system UVTEC (Cambridge, UK).

Membrane biotinylation

Membrane protein biotinylation and streptavidin pull down experiments were performed as described previously (D'Arco et al. 2007). Briefly, cultures were incubated with 2 mg/mL EZ-Link Sulfo-NHS-LC-biotin (Thermo Fisher Scientific corp., Erembodegem, Belgium) for 30 min at 4°C in phosphate-buffered saline containing 1 mM MgCl2 and CaCl2 (pH 8.0). Cells were quenched in 100 mM ice-cold glycine for 30 min, lysed in 10 mM Tris-HCl (pH 7.5), 150 mM NaCl, 20 mM EDTA, and 1% Triton-X-100 plus protease inhibitors (Roche Products, Welwyn Garden City, UK). Pulldown of biotinylated proteins was obtained with ImmunoPure Immobilized Streptavidin beads (Thermo Fisher Scientific) for 2 h at 4°C in a rotator. Streptavidin bead complexes washed twice with phosphate-buffered saline with 0.1 M CaCl2/MgCl2 and 0.1% Triton-X. Supernatant was eluted by boiling in presence of 0.1 M dithiothreitol (5 min) and loaded on sodium dodecyl sulfate gel. Biotinylation experiments resulted free of intracellular protein contaminants when tested with β-actin.

Patch clamp recordings

As previously described (D'Arco et al. 2009; Sundukova et al. 2012), currents were recorded from small-/medium-size mouse trigeminal neurons in culture under whole cell voltage clamp mode at holding potential of −80 mV. Such cells were continuously superfused at 20°C with control solution containing (in mM): 152 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 glucose, 10 HEPES; pH 7.4 adjusted with NaOH. Patch pipettes had a resistance of 3–4 MΩ when filled with (in mM): 140 KCl, 2 MgCl2, 0.5 CaCl2, 2 ATP-Mg, 2 GTP-Li, 20 HEPES, 5 EGTA; pH 7.2 adjusted with KOH. After establishing whole cell configuration, access resistance was never > 10 MΩ and was routinely compensated by 70%. Data were acquired and analyzed with the pCLAMP software Clampex 9.2 (Molecular Devices, Palo Alto, CA, USA). The receptor agonist α,β-methyleneATP (α,β-meATP; Sigma) was applied by rapid solution changer system (Perfusion Fast-Step System SF-77B; Warmer Instruments, Hamden, CT, USA). Membrane currents were analyzed in terms of current density (calculated as amplitude divided by cell slow capacitance) to take into account cell size. Current decay as a result of receptor desensitization during agonist application was fitted with either a monoexponential (for 1 μM and 3 μM doses of α,β-meATP) or biexponential (for 10 μM) function using pCLAMP Clampfit 9.2. Recovery from desensitization was assessed by a paired pulse protocol over 30 s intervals in accordance with previous reports (Sokolova et al. 2006; Nair et al. 2010a, b).

Data analysis

All data are presented as mean ± standard error of the mean (SE); n is the number of cells. Statistical significance was evaluated with paired Student's t-test (for parametric data) or Mann-Whitney rank sum test (for nonparametric data) using OriginPro 7.5 (OriginLab, Northampton, MA, USA), p < 0.05 was considered statistically significant.


CASK expression in mouse trigeminal ganglion neurons

Figures 1a and b show immunofluorescence experiments to examine CASK and P2X3 receptor expression in mouse trigeminal ganglion neurons, which were labeled by neuron-specific β-tubulin. While CASK/P2X3 expression was observed in the majority of trigeminal neurons, cell size analysis indicates that most P2X3 and CASK co-localization was found in medium-size neurons (15–20 μm somatic size; Simonetti et al. 2006), although large-diameter neurons, typically lacking P2X3 receptors, also expressed CASK (Fig. 1b, right). This expression pattern was maintained in trigeminal ganglion culture in which CASK signal was found in both neurons and non-neuronal cells (Fig. 1c and d). Under our conditions, no CASK nuclear signal was found in trigeminal ganglia neurons.

Figure 1.

P2X3/calcium/calmodulin-dependent serine protein kinase (CASK) co-expression in trigeminal ganglion neurons in vivo and in vitro. (a) Trigeminal ganglion neurons immunostained with antibodies against P2X3 receptors (green), CASK (red). (b) Histograms quantify the percentage of P2X3 and CASK expressing neurons (left, 100% is the total number of β-tubulin III positive neurons) and the somatic size distribution of neurons expressing CASK and P2X3 receptors (right), where open bars represent the percentage of P2X3 expressing neurons and filled bars the percentage of CASK expressing neurons (β-tubulin III positive). The database for analysis comprises an average of 1000 cells. (c) Example of trigeminal ganglion cultures (24 h) immunostained as in (a). (d) Histograms quantify the percentage of P2X3/CASK co-expressing neurons (left) and their somatic size distribution (right) in culture. Data are the mean of three independent experiments. Calibration bars in (a) and (c): 50 μm.

CASK modulates the P2X3 protein stability

Lack of selective pharmacological inhibitors of CASK prompted us to use RNA interference to knock down endogenous CASK (siCASK) in trigeminal ganglion cultures and to look for its effects on P2X3 receptor expression. Figure 2a shows that, after siCASK, while expression of total P2X3 receptor in whole cell lysates was reduced by about 40% (see open bar in right histograms), protein biotinylation experiments showed that an even stronger decrease (by nearly 2/3rd; shaded bar in Fig. 2a, right) was observed for the surface expression of P2X3 receptors. To check if siCASK-induced P2X3 receptor loss was because of proteasomal degradation, trigeminal cultures were treated with the proteasome inhibitor MG-132 (30 μM; 4 h; Vacca et al. 2011), that per se increased P2X3 receptor protein expression already in control basal conditions (Figure S2a). Figure 2b indicates that application of MG-132 to trigeminal ganglion cultures was sufficient to significantly (p < 0.04) reverse the effect of siCASK, therefore preventing siCASK-mediated decrease in P2X3 receptor expression.

Figure 2.

Calcium/calmodulin-dependent serine protein kinase (CASK) silencing reduces P2X3 expression by trigeminal neurons. (a) Example of total lysates and membrane biotinylation western immunoblots of trigeminal ganglion culture lysates from control or siCASK samples probed with the anti-P2X3 antibody. Gel loading is shown with β-Tubulin signal. In the right panels, the total or surface P2X3 protein levels in siCASK cultures were compared to those of cells transfected with scramble siRNA. n = 5. Note significantly lower expression of P2X3 receptors in total lysates (*p = 0.0028) and membrane samples (p = 0.038) from siCASK cultures. (b) Examples of P2X3 or CASK immunoblots from trigeminal neuron lysates in control, siCASK or siCASK treated with MG-132 (30 μM; 4 h) conditions. β-Tubulin shown for gel loading. Histograms compare P2X3 receptor expression of control, siCASK or siCASK and MG-132 treated samples. In the latter case, P2X3 receptor expression is significantly larger versus siCASK (n = 3 experiments, *p = 0.035). (c) Left, confocal microscopy images of single cultured trigeminal neurons in control, or after siCASK. Localization of nuclei is indicated with N. Bar = 10 μm. Right, cell line-scan immunostaining profile for P2X3 or CASK indicates sparse and decreased P2X3 immunostaining after siCASK. Similar results were obtained in at least three independent experiments (n = 15 cells).

We performed confocal line-profile analysis of trigeminal neurons in culture with or without CASK silencing and immunostained for P2X3 and CASK (Fig. 2c). While in control conditions the P2X3 receptor signal was uniformly distributed across the cell section (Fig. 2c), resembling the staining of neurons in intact ganglia (as exemplified in Figure S2b), after CASK knock down, P2X3 receptor staining was highly reduced especially at the perimembrane region, in accordance with the decrease in cell surface expression demonstrated in Fig. 2a and b. P2X3 receptor trafficking was explored in siCASK-treated trigeminal neuronal cultures either in control or in the presence of brefeldin A (BFA), which alters the transport from the endoplasmatic reticulum to Golgi complex (Valenzuela et al. 2011). These experiments confirmed larger expression of P2X3 subunits after BFA treatment (5 μg/mL; 30 min; Fabbretti et al. 2006), while P2X3 expression remained sensitive to siCASK (Figure S3).

Characteristics of P2X3/CASK interaction

As these data suggest that CASK had a role in the stability of P2X3 receptors in trigeminal neurons, we explored whether CASK and P2X3 could be part of a unique macromolecular complex. To address this issue, P2X3/CASK co-immunoprecipitation experiments were performed using lysates from intact murine trigeminal ganglia or from cultures. Figure 3a shows that CASK signal was clearly detected in immunopurified P2X3 receptors from ganglion tissue or cultures (control signals for P2X3, CASK and β-tubulin are also shown in Fig. 3a). After siCASK in trigeminal cultures, the CASK/P2X3 complex was clearly reduced (Fig. 3b). The P2X3/CASK interaction was also observed in a neuronal free environment, as similar results were obtained with HEK cells co-transfected with CASK and P2X3 plasmids (Fig. 3c, right lane). Even endogenous CASK of HEK cells could co-precipitate with the P2X3 protein as shown in Fig. 3c (middle lane indicated as P2X3/GFP). Similar data were observed with pull down of either P2X3 or CASK-myc, providing further validation of CASK pull down experiments (Fig. 3d and e). Interestingly, the expression of recombinant P2X3 receptors was also affected by silencing endogenous CASK in HEK cells (Figure S1, left), thus supporting a strong role of CASK in the stability of P2X3 receptors (Fig. 2b). After P2X3/CASK co-expression in HEK cells, immunoprecipitation with anti-P2X3 antibodies and western blot with an anti-phosphoserine antibody showed significant increment in serine phosphorylation of P2X3 receptors (Fig. 3f), which is known to be associated with receptor function (Nair et al. 2010a, b).

Figure 3.

Characteristics of calcium/calmodulin-dependent serine protein kinase (CASK)/P2X3 interactions. (a) Example of P2X3/CASK co-immunoprecipitation experiments from either trigeminal ganglion tissue or culture. β-Tubulin shows gel loading in different lanes. (b) Similar approach is shown for P2X3/CASK immunoprecipitation from cultures in control or after siCASK. (c) Panels show total expression of human embryonic kidney (HEK) cell lysates transfected with P2X3 and CASK or GFP. β-actin is used for gel loading control. NT shows no signal in non-transfected HEK cell lysates. Histograms quantify increased P2X3/CASK co-immunoprecipitation after CASK over-expression in HEK cells (n = 3 experiments, *p = 0.03). (d) HEK cells transfected with P2X3 and GFP or P2X3 and CASK are used for P2X3 immunoprecipitation experiments as above. Different negative control lanes include (from left to right): Ab, antibody alone showing no crossreactivity with IgG heavy chains; NT shows no signal in P2X3 immunoprecipitates from non-transfected HEK cell lysates; no primary, indicating lack of unspecific signal of secondary antibody cross-reactivity. (e) Example of P2X3 pull down with anti-Myc antibody immunoprecipitation of P2X3/CASK-myc co-transfected lysates. Note stronger expression of P2X3 protein in the complex when co-transfected with CASK. (f) Example of enhanced P2X3 receptors serine phospshorylation in P2X3/CASK co-transfected HEK cells versus P2X3/GFP alone. To quantitatively evaluate P2X3 receptor phosphorylated residues, we loaded equal amounts of P2X3 receptor in different lanes (D'Arco et al. 2009). Histograms quantify serine phosphorylation of P2X3 receptors with or without CASK over-expression in HEK cells (n = 3 experiments, *p = 0.03). Immunoprecipitates of cell extracts with an unrelated IgG antibody (Ab) were also shown as control.

To further explore the role of CASK in the stability of P2X3 receptors, we used a series of CASK deletion mutants (Chao et al. 2008) for P2X3 receptor co-transfection in HEK cells. Figure 4 (left) shows representative immunoblots of CASK and P2X3 expression in cell lysates after transfection with wild type CASK or the CASK mutants ΔCaMKII, ΔPDΖ, ΔSH3 and ΔGK (characterized by different molecular weight). We observed that total P2X3 receptor expression was significantly enhanced by over-expressing either CASK (see also Fig. 3c) or anyone of these mutants (Fig. 4).

Figure 4.

Calcium/calmodulin-dependent serine protein kinase (CASK) domains necessary for P2X3 receptor interaction. Left: western blot of human embryonic kidney (HEK) cells lysates co-transfected P2X3 with full length CASK and domain deletion mutants ΔCAMK, ΔPDΖ, ΔSH3 and ΔGK, as indicated. Transfection with P2X3 only was used as control (see Methods), while β-Actin shows loading control. Note upregulation of total P2X3 protein expression in co-transfected cells. Right: histogram quantifies the relative P2X3 expression in different groups, as indicated. n = 4 experiments; *p < 0.05.

Receptor state-dependent interaction of P2X3 and CASK

We next inquired if the CASK/P2X3 interaction was a dynamic process related to the activity state of P2X3 receptors at membrane level. We, therefore, tested if enhanced P2X3 receptor activity by NGF or by P2X3 desensitization via by sustained application of the selective agonist α,β-meATP was accompanied by changes in the P2X3/CASK interaction.

To determine whether NGF signaling could modulate P2X3/CASK interaction, trigeminal ganglion cultures were grown in basal conditions or under NGF deprivation (D'Arco et al. 2007, 2009). The potent algogenic action of NGF-TrkA signaling results in enhanced P2X3 receptor function via modulation of PKC (D'Arco et al. 2007) and the Csk/Src pathways (D'Arco et al. 2009). Acute application of NGF (100 ng/mL; 5 min) to anti-NGF pre-treated cultures (Fig. 5a) significantly (p = 0.003; n = 5) increased CASK/P2X3 co-precipitation, while the anti-NGF antibody per se induced a modest CASK/P2X3 signal decrease (p = 0.03, n = 3).

Figure 5.

Dynamic changes in P2X3/calcium/calmodulin-dependent serine protein kinase (CASK) association. (a) Left, CASK expression in purified P2X3 receptor samples from trigeminal ganglion cultures is enhanced by acute nerve growth factor (NGF) application (100 ng/mL, 5 min) following overnight neutralizing NGF antibody (αNGF; 14 h) application to block endogenous NGF. P2X3 or CASK input is also shown. To exclude artifact signals as a result of contamination of heavy chain of anti-NGF antibody used for the NGF deprivation treatment, an aliquot of an unrelated IgG was also loaded in first lane as control. Right, Histograms quantify increase in co-precipitation of CASK and P2X3 receptors by NGF application (n = 3 experiments, *p = 0.03); data of immunoprecipitated CASK are normalized with respect to total P2X3 expression. (b) Left, Example of effect of sustained application of α,β-meATP (100 μM, 30 s) on CASK expression by P2X3 receptors immunoprecipitation from transfected human embryonic kidney (HEK) cells. The action of α,β-meATP is blocked by co-applying the P2X3 antagonist A-317491 (10 μM, 30 s). Right, Histograms quantify the effects of α,β-meATP exemplified on the left (n = 5 experiments, *p = 0.027; data expressed as in (a). (c), Immunoblots run with protocol similar to the one in (b), show the P2X3/CASK complex in trigeminal primary cultures after a short α,β-meATP pulse (100 μM, 30 s). Histograms quantify the effect (n = 3, p > 0.05).

Full desensitization of P2X3 receptors evoked by sustained application of the selective agonist α,β-meATP (100 μM; 30 s; Sokolova et al. 2006) induced a significant (p = 0.0027; n = 5) decrease in P2X3/CASK interaction, an effect prevented by co-application of α,β-meATP together with the selective P2X3 antagonist A-317491 (10 μM; 20 s; Jarvis et al. 2002, 2004) (Fig. 5b). Similar results were obtained also in trigeminal ganglion primary cultures (Fig. 5c).

We subsequently studied the action of CASK on the properties of P2X3 receptor activation on trigeminal neurons in culture. Patch clamping was used to measure membrane currents evoked by short pulses of α,β-meATP (1–10 μM; 2 s) to minimize receptor desensitization. Figure 6a compares sample responses elicited by α,β-meATP in control conditions or after siCASK: in the latter condition, there was significant depression of P2X3 receptor currents as quantified in Fig. 6b. No significant change in fast desensitization properties was, however, observed as indicated by similar values for current decay (Fig. 6c) nor recovery from desensitization 30 s later (Fig. 6d). CASK silencing did not change cell capacitance (18 ± 1 vs. 18 ± 1 pF control; n = 31 cells), or input resistance (278 ± 24 vs. 280 ± 30 MΩ control), indicating that this treatment did not modify basic neuronal properties under resting condition (Fig. 6c and d). Patch clamp experiments were also performed to study the functional impact of co-expression of P2X3 receptors with CASK mutants in HEK cells as summarized in Fig. 6e. Thus, significantly larger amplitude of P2X3 receptor currents induced by 2 s pulses of 10 μM α,β-meATP was observed after co-transfection with WT CASK (~ 40%), ΔCAMK (~ 50%), ΔPDΖ (~ 40%) whereas no significant change versus control currents was detected after co-expressing P2X3 with ΔSH3 or ΔGK.

Figure 6.

Electrophysiology of P2X3 receptors after siCASK. (a) Examples of neuronal currents elicited by short (2 s) application of α,β-meATP at the doses indicated above traces. Note smaller responses after siCASK. (b) Histograms quantify responses (expressed as current density) evoked by α,β-meATP that were significantly depressed after siCASK. Current density: for 1 μM, Control n = 24 versus siCASK n = 20, (*p = 0.02). For 3 μM, control n = 19 versus n = 24 for siRNA (*p = 0.04). For 10 μM, control n = 11 versus n = 7 for siRNA (*p = 0.02). After siCASK, no change in the time constant of current decay (c), or current recovery at 30 s (d), expressed as% of first response following a double pulse application of α,β-meATP) is observed. (e) Histograms quantify effects of wildtype CASK or its mutants ΔCAMK, ΔPDΖ, ΔSH3 and ΔGK on currents evoked by short (2 s) pulses of α,β-meATP applied to P2X3/CASK co-transfected human embryonic kidney (HEK) cells or P2X3/GFP co-transfected controls cells. CASK over-expression or CASK mutants ΔCAMK and ΔPDΖ significantly enhanced P2X3 receptor responses (n = 19 for P2X3; n = 21, *p = 0.03 for CASK WT; n = 15, *p = 0.03 for ΔCAMK; n = 13, p = 0.73 for ΔSH3; n = 13, *p = 0.02 for ΔPDΖ; n = 9, p = 0.07 for ΔGK; data were compared vs. control).


The principal finding of this report is the demonstration that the multidomain scaffold protein CASK is a novel interactor of P2X3 receptors to control their membrane dwelling and responsiveness. CASK is, therefore, proposed to be a new member of the family of intracellular signaling molecules that determine the efficiency of P2X3 receptors on sensory neurons (Fabbretti and Nistri 2012).

CASK is important for P2X3 protein expression

CASK is molecular scaffold playing a major role in the assembly of multiprotein complexes at specialized regions of neuronal plasma membranes (Butz et al. 1998; Borg et al. 1999) and is associated with pre- and post-synaptic proteins and cytoskelethon (Hata et al. 1996; Butz et al. 1998; Hsueh et al. 1998; Biederer and Südhof 2000; Jeyifous et al. 2009). Hence, CASK is reported to be involved in different neuronal functions, including synaptic strength, vesicle release, and receptor expression and trafficking as well as neuronal development (Olsen et al. 2005): indeed, its genetic ablation is lethal to mice (Atasoy et al. 2007). Our present data revealed for the first time that, in trigeminal ganglion neurons, CASK was widely expressed by P2X3-positive neurons, suggesting its potential role in sensory neuron physiology and nociceptive activity. The association of CASK with P2X3 receptors, confirmed in a recombinant neuron-free cell system, appeared important for the stability of P2X3 receptors on cell surface, as indicated by the biotinylation experiments. These results outline a new role of CASK in modulating the properties of P2X3 receptors and, perhaps, pain transduction.

Characteristics of CASK/P2X3 interaction

After knocking down of CASK, P2X3 receptor expression was decreased, an effect prevented by the proteasome inhibitor MG-132, suggesting that CASK could stabilize surface P2X3 receptors normally subjected to a relatively rapid turnover via internalization and proteasomal degradation (Vacca et al. 2011). In addition, the cell distribution of P2X3 immunoreactivity was altered by siCASK, indicating a change in receptor compartmentalization. We further tested the Golgi export inhibitor BFA (effective in controlling P2X3 receptor sorting; Fabbretti et al. 2006) that did not change the consequences of siCASK on P2X3 receptor expression. A likely possibility is that CASK modulates the stability and clustering of P2X3 receptors at membrane levels, where surface redistribution is a prerequisite of appropriate receptor function and plasticity.

Furthermore, we report that CASK/P2X3 receptor co-expression in HEK cells led to increased P2X3 serine phosphorylation, which is known to be a signature for enhanced receptor function and trafficking (Fabbretti et al. 2006; Nair et al. 2010b). These observations accord with the notion that CASK functions as a molecular scaffold to recruit signaling molecules to the membrane domain for modulation of signal transduction (Vaccaro and Dente 2002). Because CASK can act as an adaptor for various proteins including other kinases (Kaech et al. 1998; Lu et al. 2003), it is likely that manipulating CASK expression modulated P2X3 receptors through influencing other kinase anchors or adaptor proteins.

The question then arises how CASK can modulate P2X3 receptor expression. To address this issue, we explored four CASK mutants missing the CaMK, PDZ, SH3 or GK domains and observed that P2X3 receptor expression was still enhanced when co-expressed with anyone of such mutants. In this study, we have not tested the role of the L27 domain, because the epitope of the CASK antibody is located at this site. Future experiments are necessary to clarify the nature of the P2X3/CASK complex as it is not ruled out that CASK uses an undefined protein motif to associate with P2X3.

We noted that expression of these mutants differentially changed P2X3 receptor-mediated currents. In fact, ΔCaMK or ΔPDZ expression largely enhanced the current amplitude, while this was not the case of ΔSH3. Thus, against a background of comparable increment in P2X3 subunit expression by these mutants, the functional data suggest that expression of CASK or the ΔCaMK or ΔPDZ mutants might have recruited other interactors to modulate the P2X3 receptor function. This observation strengthens the notion that, in addition to stabilizing P2X3 protein levels, CASK also modified P2X3 activity. A body of evidence indicates that CaMKII has an important role in P2X3 expression and function in sensory neurons via intracellular calcium increase (Xu and Huang 2004; Simonetti et al. 2008; Nair et al. 2010a, b). This study examined if the function of CaMKII included regulation of the CASK/P2X3 complex: our data suggest that P2X3 receptors expression and function could be controlled by CaMKII (Simonetti et al. 2008; Nair et al. 2010a, b) even with mechanisms independent from its binding to CASK. Because of the multiple partners of CASK (including CaMKII and PDZ) and its heterogeneous roles in neurons (Hsueh 2011), it is difficult to precisely explain how CASK modulates P2X3 receptor currents. It is possible that the differential affinity of CASK for multiple intracellular adaptors can modulate various components of receptor responses comprising clustering, gating or internalization after agonist binding. Future studies are, therefore, necessary to identify key components in the P2X3/CASK complex and multiple mechanisms modulating their interaction.

Functional impact of P2X3/CASK interaction

Acute application of NGF, known to up-regulate P2X3 receptor currents (D'Arco et al. 2007) enhanced P2X3/CASK interaction, while NGF deprivation had the opposite effect. Conversely, desensitization of P2X3 receptors with a sustained application of large concentrations of α,β-meATP, led to strong decrease in P2X3/CASK interaction, indicating that prolonged receptor conformational modifications elicited by the agonist (Sokolova et al. 2006) were sufficient to perturb the P2X3/CASK domain sites. These results propose that the strength of association between P2X3 and CASK was dynamically regulated by the receptor state. Patch clamp experiments on trigeminal neurons after siCASK demonstrated a small, yet significant depression of P2X3 receptor mediated responses evoked by short pulses of α,β-meATP. This phenomenon was, however, not associated with a change in receptor desensitization or recovery from it. Thus, under basal conditions, block of CASK expression likely decreased the amount of P2X3 receptors (or of unidentified interactors) at membrane level and, therefore, impaired neuronal functional responses without altering their kinetic properties.


P2X3 receptors are highly regulated by soluble mediators (including ATP itself) and complex intracellular signaling pathways (Burnstock et al. 2011; Giniatullin et al. 2008; Illes et al. 2011). This property offers an array of molecular processes underlying the dysregulation of P2X3-mediated nociception especially in chronic pain syndromes (Abbracchio et al. 2009; Burnstock 2012; Wirkner et al. 2005). As this report shows that CASK has a key role in ensuring P2X3 receptor stability and function, it will be interesting to study whether sensory neurons of chronic pain models possess enhanced CASK/P2X3 interaction as a mechanism to support this syndrome.


This work was financially supported by grants from the Telethon Foundation – Italy (GGP10082), the Cariplo Foundation (2011-0505), the EU Crossborder Cooperation Initiative (MINA project) managed by the Friuli Venezia Giulia Region Government, the AHA-MOMENT grant from Slovenian Ministry for Education and Science. We acknowledge the help by Dr. Manuela Simonetti for preliminary experiments.

Conflict of interest

The authors state that the content of this article does not create any conflict of interest.