Role of domain calcium in purinergic P2X2 receptor channel desensitization

Abstract

Activation of P2X2 receptor channels (P2X2Rs) is characterized by a rapid current growth accompanied by a decay of current during sustained ATP application, a phenomenon known as receptor desensitization. Using rat, mouse, and human receptors, we show here that two processes contribute to receptor desensitization: bath calcium-independent desensitization and calcium-dependent desensitization. Calcium-independent desensitization is minor and comparable during repetitive agonist application in cells expressing the full size of the receptor but is pronounced in cells expressing shorter versions of receptors, indicating a role of the COOH terminus in control of receptor desensitization. Calcium-dependent desensitization is substantial during initial agonist application and progressively increases during repetitive agonist application in bath ATP and calcium concentration-dependent manners. Experiments with substitution of bath Na⁺ with N-methyl-d-glucamine (NMDG⁺), a large organic cation, indicate that receptor pore dilation is a calcium-independent process in contrast to receptor desensitization. A decrease in the driving force for calcium by changing the holding potential from −60 to +120 mV further indicates that calcium influx through the channel pores at least partially accounts for receptor desensitization. Experiments with various receptor chimeras also indicate that the transmembrane and/or intracellular domains of P2X2R are required for development of calcium-dependent desensitization and that a decrease in the amplitude of current slows receptor desensitization. Simultaneous calcium and current recording shows development of calcium-dependent desensitization without an increase in global intracellular calcium concentrations. Combined with experiments with clamping intrapipette concentrations of calcium at various levels, these experiments indicate that domain calcium is sufficient to establish calcium-dependent receptor desensitization in experiments with whole-cell recordings.

purinergic p2x receptors (P2XRs) are ATP-gated nonselective cation channels expressed in various tissues and have multiple physiological roles (4, 30). Some notable examples of their roles in the central and peripheral nervous system include presynaptic P2X2Rs that contribute to an increase of glutamate release in mice interneurons (24), postsynaptic P2X4Rs that contribute to long-term potentiation in CA1 neurons from mice (34), involvement of P2X4Rs and P2X7Rs in neuropathic pain (10, 36, 37), and P2X3Rs and P2X2R acting through a heteromeric receptor for pain transmission in mammalian sensory neurons (28). On the other hand, the specific properties of each P2XR, such as ATP affinity (7), pore dilation (32), and kinetics of receptor desensitization (30), confer unique modes of action on these receptors that define their physiological functions. Furthermore, several endogenous and exogenous compounds can modulate ATP gating of these receptors, including essential trace metals, ivermectin, reactive oxygen species, phospholipids, and alcohols. The effects of these modulators are wide, including shifts in the EC₅₀ values, peak amplitude of current responses, and changes in pharmacological profiles (6).

Divalent cations have an important role on P2XR functioning, acting mainly as allosteric modulators. For example, Zn²⁺, a trace metal, inhibits the currents mediated by P2X1R, P2X3R, P2X5R, and P2X7R, whereas it potentiates the currents mediated by P2X2R and P2X4R (7). Copper, another trace element, potentiates the P2X2R but inhibits the activity of all other P2XRs (6, 7). Specific residues that mediate cation effects, most probably participating in metal coordination, have been identified, and several cation-binding sites have been proposed (7). The macroelements Ca²⁺ and Mg²⁺ also have a modulatory role on P2XRs; their main effect in a millimolar concentration range is inhibition of agonist-induced currents (30), but there is also evidence suggesting a positive effect of Ca²⁺ in recovery from desensitization of P2X3Rs (8). In cells expressing P2X2R, an inhibitory effect of Ca²⁺ and Mg²⁺ has also been detected with single-channel recording as a reduction of unitary currents (13). The same group also observed Ca²⁺-dependent receptor desensitization although the mechanism of its action was not determined (12).

In a recent work, our group showed that rat P2X2R can dilate and desensitize simultaneously and that the desensitization process includes Ca²⁺-dependent and -independent mechanisms. During repetitive agonist application, with washout periods of 4–5 min, there is a progressive increase in the rates of calcium-dependent desensitization (CDD), a phenomenon termed use-dependent desensitization (23). Here, we characterized CDD of P2X2R. We found that CDD is extended to human, rat, and mouse P2X2Rs and that its action relies on the transmembrane/intracellular domains of this receptor subtype but not other P2XRs. Moreover, experiments with the wild-type and chimeric receptors clamped at different potentials to facilitate or decrease Ca²⁺ influx indicate that Ca²⁺ influx through the channel pore is critical for CDD. Finally, we found that Ca²⁺ exerts its effects at very low intracellular concentrations in the whole-cell configuration, indicating the role of a domain rather than global calcium signaling in receptor desensitization.

MATERIALS AND METHODS

Receptor transfection and cell culture.

Experiments were done with the rat, mouse, and human P2X2Rs, rat P2X3R and P2X4R, as well as the rat receptor chimeras P2X4/X2a, P2X4/X2b (18, 19), P2X3/X2a (42), and P2X2a/V49-V61X4 (20), all inserted in the pIRES2-enhanced green fluorescent protein (GFP) plasmid and expressed in HEK293 cells or GT1-7 cells. HEK293 cells were routinely maintained in DMEM containing 10% (vol/vol) fetal bovine serum (Biofluids) and 1% (vol/vol) penicillin-streptomycin (Invitrogen). GT1-7 cells were maintained in DMEM/Ham's F-12 medium (1:1) supplemented with 10% FBS and 1% penicillin-streptomycin. For electrophysiological experiments, cells were grown on 35-mm culture dishes at a density of 500,000 cells/dish. The transient transfection was conducted 24 h after we plated the cells using 2 μg of DNA and 5 μl of Lipofectamine 2000 reagent (Invitrogen) in 2 ml of serum-free Opti-MEM. After 4.5 h of incubation, the transfection mixture was replaced with normal culture medium. The experiments were performed 24–48 h after transfection, and P2XRs expressing cells were identified by GFP fluorescence.

Current measurements.

Electrophysiological experiments were performed on cells at room temperature using whole-cell patch-clamp recording techniques. The currents were recorded using an Axopatch 200B patch-clamp amplifier (Molecular Devices) and were filtered at 2 kHz using a low-pass Bessel filter. Patch electrodes, fabricated from borosilicate glass (type 1B150F-3; World Precision Instruments), using a Flaming Brown horizontal puller (P-87; Sutter Instruments), were heat polished to a final tip resistance of 2–4 MΩ. All current records were captured and stored using the pClamp 9 software packages in conjunction with the Digidata 1322A analog-to-digital converter (Molecular Devices). The current responses were recorded from single cells clamped at −60 mV. Concentration-response data were collected from recordings of a range of ATP concentrations applied to single cells, with a washout interval of 4 min between each application, and normalized to the highest current amplitude. I–V relations were used to evaluate changes in reversal potential during ATP application and were obtained by voltage ramps from −80 to +80 mV, delivered twice per second for 45 s.

Bath and intrapipette media.

The Ca²⁺-containing bath solution (Krebs-Ringer-like) contained the following (in mM): 142 NaCl, 3 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES. The Ca²⁺-deficient solution contained the following (in mM): 145 NaCl, 3 KCl, 10 glucose, and 10 HEPES, with trace Ca²⁺ concentration estimated by atomic absorption spectroscopy at about 90 μM. The N-methyl-d-glucamine (NMDG⁺) bath solution contained 155 NMDG⁺, 10 glucose, and 10 HEPES. If not otherwise specified, patch electrodes were filled with a solution containing the following (in mM): 142 NaCl, 10 EGTA, and 10 HEPES, pH adjusted to 7.35 with 10 M NaOH; estimated free Ca²⁺ concentration was 0.8 nM. In some experiments, calcium was added to elevate free Ca²⁺ concentrations to 25 nM and 200 nM. Calculations of free intrapipette [Ca²⁺] were done using the WEBMAXC Standard software (http://web.stanford.edu/∼cpatton/webmaxcS.htm). The osmolarity of the internal solutions was 305 mosmol/kgH₂O. ATP was prepared daily in bath buffer and applied using a rapid solution changer system (RSC-200; Biologic Science Instruments).

Simultaneous current and calcium recordings.

For these experiments, 5 μM Indo-1 was added to the pipette solution; the composition of this solution was the same as that used in the other electrophysiological experiments but did not include EGTA, except for control experiments for global calcium increases in which EGTA was also included. Cells were cultured in 25-mm coverslips, and single cells were selected and clamped at −60 mV. For determination of intracellular Ca²⁺ levels, the same cell was excited at 340 nm, and fluorescence was detected with a photomultiplier at 405 and 480 nm (Nikon).

Data analysis.

Concentration-response curves were performed by applying ATP for 5 s in the 0.1–1,000.0 μM concentration range. Curves were normalized against the concentration of ATP that evoked the maximal response, in the presence of 0.09 mM and 2.00 mM Ca²⁺. Likewise, the maximal ATP current (I_max) was obtained from each ATP concentration-response curve. Each experiment was repeated in at least four separate cells. Curve fitting, EC₅₀, IC₅₀, I_max values, and statistical analyses, including nonparametric Kruskal-Wallis and Mann-Whitney tests, were obtained with GraphPad software. The desensitization constant (τ_des) was calculated from each recording by curve fitting of the desensitization phase with a predefined monoexponential function [f(t) = B exp(−t/τ)] using the Clampfit 10.0 software (Molecular Devices).

RESULTS

Bath Ca²⁺ is required for P2X2R desensitization.

In a previous study, we reported the effects of bath Ca²⁺ on P2X2R desensitization during repetitive ATP applications (23). Figure 1A shows two examples, using two splice forms of rat P2X2Rs (rP2X2Rs), the full size termed rP2X2aR and the splice form missing 69 amino acid residues in COOH terminus termed rP2X2bR (26). At a holding potential of −60 mV, desensitization was more pronounced in cells expressing rP2X2bR in the absence and presence of bath Ca²⁺, but both receptors desensitize more rapidly in the presence of 2 mM bath calcium (Ca²⁺-containing medium; black traces) than in the presence of trace 0.09 mM Ca²⁺ (Ca²⁺-deficient medium; gray traces) during the first agonist application. Furthermore, the rP2X2aR exhibited progressive CDD during repetitive ATP application, accompanied with a small decrease in current peak amplitudes (Table 1). Under identical experimental conditions, the human receptors behaved similarly as the rat models (data not shown). Mouse P2X2Rs have three functional splice forms: mP2X2aR, mP2X2bR, and mP2X2eR, and the last splice variant lacks 90 amino acid residues in its COOH-terminal domain (27). CDD was preserved in mP2X2aR and mP2X2bR, whereas desensitization of the P2X2eR was very fast (Fig. 1B), indicating that missing 21 amino acids contributes to the control of receptor gating independently of Ca²⁺.

Fig. 1.

Bath calcium increases P2X2R desensitization. A and B: representative recordings of ATP-induced inward currents in HEK293 cells expressing rat (r) P2X2aR and P2X2bR (A) and mouse (m) P2X2aR, P2X2bR, and P2X2eR (B). Black and gray traces correspond to recordings in cells bathed in Ca²⁺-containing (2 mM Ca²⁺) and Ca²⁺-deficient medium (0.09 mM Ca²⁺), respectively. Traces shown in Ca²-containing medium were obtained during repetitive application of 100 μM ATP to the same cell, with washout periods of 4 min between each pulse. Numbers correspond with each ATP application. Traces in Ca²⁺-deficient medium are from different cells during the first ATP application. In this and following figures, horizontal bars above traces illustrate duration of ATP application. In all experiments, the whole-cell recording was used with the pipette containing 142 mM NaCl, 10 mM HEPES, and 10 mM EGTA, at a holding potential of −60 mV. Recordings are representative of at least 4 different experiments.

Table 1.

Changes in the peak current amplitudes and rates in receptor desensitization during repetitive agonist application

	ATP Applications, 100 μM
Parameters	1st	2nd	3rd	4th
τdes, s	13.4 ± 3.4 (7)	7.3 ± 1.7 (7)	5.2 ± 0.4 (6)	3.3 ± 0.6 (4)
Imax, pA/pF	274 ± 25 (7)	215 ± 32 (7)	204 ± 37 (6)	193 ± 42 (4)

Values are means ± SE, with n in parentheses. Applications were for 40 s followed by 4-min washout periods in HEK293 cells expressing rat P2X2a receptor.

τ_des, desensitization constant; I_max, maximal ATP current.

In further experiments, we characterized in more detail the effect of Ca²⁺ on rP2X2aR desensitization. Receptor desensitization depended on both extracellular Ca²⁺ and ATP concentration (Fig. 2). In Ca²⁺-containing medium, ATP application induced a concentration-dependent desensitization of the currents (Fig. 2A). In Ca²⁺-deficient medium, ATP also induced a concentration-dependent desensitization, but the rates of receptor desensitization were significantly smaller (Fig. 2B). Removal of bath Ca²⁺ slightly enhanced the potency of ATP [Ca²⁺-containing medium EC₅₀ = 14.6 ± 1.5 μM (n = 7) vs. Ca²⁺-deficient medium EC₅₀ = 8.7 ± 1.3 μM (n = 5), P < 0.05; Fig. 2C]. In contrast, bath Ca²⁺ did not affect the efficacy of ATP expressed as maximal current density; values were 274 ± 25 and 293 ± 17 pA/pF in the presence and absence of Ca²⁺, respectively (n = 12). Using a fixed ATP concentration (100 μM), we also compared effects of extracellular Ca²⁺ concentration on receptor desensitization. Figure 2D shows that an increase in bath Ca²⁺ concentration resulted in faster receptor desensitization. Together, these results indicate that rP2X2aR desensitizes in a Ca²⁺-dependent and -independent manner and that Ca²⁺-dependent receptor desensitization is facilitated by an increase in bath ATP and Ca²⁺ concentrations.

Fig. 2.

Dependence of rP2X2aR desensitization rate on bath ATP and calcium concentrations. A and B: concentration-dependent effects of ATP on the rate of receptor desensitization in cells bathed in Ca²⁺-containing (+Ca²⁺; A) and -deficient (-Ca²⁺; B) medium. Normalized representative currents of receptor-expressing HEK293 cells activated by 10, 30, or 100 μM ATP are shown. C: concentration-dependent effects of ATP on the peak current response in cells bathed with Ca²⁺-containing (●) and Ca²⁺-deficient (○) medium; means ± SE, n = 5. D: concentration-dependent effects of bath Ca²⁺ on the rate of receptor desensitization. Data shown are normalized representative recordings of currents. The tracings shown and means ± SE are from naïve receptors during initial ATP application.

Bath Ca²⁺ facilitates rP2X2R desensitization independently of pore dilation.

The rP2X2R dilates and desensitizes simultaneously (23), and here we focused on the role of Ca²⁺ in both processes. In rP2X2aR-expressing HEK293 cells bathed in Ca²⁺-deficient/NMDG⁺-containing medium and clamped at −60 mV, we observed a rapid outward Na⁺ current followed by a slowly developed and stable inward NMDG⁺ current (Fig. 3A, left), indicative of pore dilation. Consistent with this, in −80/+80-mV ramp protocols (delivered twice per second), we also observed a shift in the reversal potential from −60 to −24 mV (Fig. 3, B and C, left). When cells were bathed in Ca²⁺/NMDG⁺-containing medium, however, the inward NMDG⁺ current dramatically declined during the sustained agonist application (Fig. 3A, right). In contrast, bath Ca²⁺ had no effect on the peak amplitude of current when expressed as pA/pF [NMDG⁺ = 119.3 ± 20.2 (n = 4) vs. NMDG + Ca²⁺ = 149.8 ± 42.7 (n = 4); nonsignificant]. It also did not affect the shift of the reversal potential and only affected the current amplitude, deduced by the decrease in ramp slope (Fig. 3, B and C, right). We also tested the shift in reversal potential after three ATP applications to a single cell in the absence and presence of extracellular Ca²⁺, with no significant differences (Fig. 3D). These results indicate that receptor desensitization is facilitated in the presence of extracellular Ca²⁺ independently of the presence of other ions in bath solution and pore dilation.

Fig. 3.

Bath calcium does not affect rP2X2aR pore dilation. A–C: representative recordings from HEK293 cells. A: cells clamped at −60 mV were bathed in N-methyl-d-glucamine (NMDG⁺) medium (155 mM NMDG⁺, 10 mM HEPES, and 10 mM glucose) in the absence (left) or presence (right) of 2 mM Ca²⁺. B: voltage-ramp experiments; 0.485-s voltage ramps from −80 to +80 mV were delivered twice per second. In A and B, notice the difference in the profile of current in cells bathed in Ca²⁺-containing and -deficient medium. C: shifts in the reversal potential; 10 out of 50 ramps are shown (corresponding to the first 20 s of ATP application), and the reversal potential value is shown for the first and last ramp. All experiments in A–C were done with naive cells. D: summary of the changes in reversal potential after 3 ATP applications to the same cell bathed in Ca²⁺-containing and -deficient medium; n = 4.

Dependence of receptor desensitization on Ca²⁺ influx through the channel pores.

To clarify whether Ca²⁺ influx is critical for development of CDD, we stimulated cells clamped at −60 mV but included BAPTA instead of EGTA as the Ca²⁺ buffer; under these conditions, CDD was still preserved (Fig. 4A). We then tested the pattern of current response at different positive holding potentials, +40 mV, +60 mV, and +120 mV. Positive holding potentials decrease the driving force for Ca²⁺ influx, which should affect the rate of receptor desensitization if it is dependent on Ca²⁺ influx through the channel pores and/or accumulation of Ca²⁺ in cytosol. In cells clamped at positive potentials, ATP application generated outward currents. CDD was still observed at +40 and +60 mV (Fig. 4, B and C). However, in cells clamped at +120 mV, a membrane potential that almost completely reduces Ca²⁺ influx, CDD was greatly reduced (Fig. 4D). These results indicate that Ca²⁺ influx through the channel pores at least partially accounts for CDD and/or that membrane potential influences the conformation of the receptor affecting its gating properties.

Fig. 4.

Influence of holding potentials on Ca²⁺-dependent desensitization (CDD). A–D: development of CDD in cells clamped at different potentials: −60 mV (A), +40 mV (B), +60 mV (C), and +120 mV (D). Recordings were done with pipette solution containing 10 mM BAPTA (A) or 10 mM EGTA (B–D) in Ca²⁺-deficient (gray traces) and Ca²⁺-containing (black traces) medium. Numbers indicate consecutive ATP applications, with 4 min of washout between pulses. Recordings are representative of at least 4 different experiments.

The transmembrane/intracellular domains of the rP2X2R are required for CDD.

We next assessed which domains of the rP2X2R are relevant for CDD. To test that, we used three chimeric receptors, rP2X4/X2aR, rP2X4/X2bR, and rP2X3/X2aR, and compared their gating with parental receptors. Calcium dependence of rP2X2R and rP2X2bR desensitization is shown in Fig. 1A. The rP2X4R exhibits a small CDD (Fig. 5A), and the rP2X3R desensitizes rapidly and independently of bath Ca²⁺ concentration (Fig. 5B). Chimeric receptors have the Ile⁶⁶-Tyr³¹⁵ and Val⁶⁰-Phe³⁰¹ ectodomains from the rP2X4R and rP2X3R, respectively, instead of the native Ile⁶⁶-Tyr³¹⁰ sequence of rP2X2Rs (18, 19); three chimeras showed CDD similar to that observed in rP2X2Rs. These results indicate that transferring the vestibule, transmembrane, and intracellular regions of the rP2X2aR to rP2X4R and rP2X3R ectodomains is sufficient to confer CDD to the respective chimeras, as reflected in the desensitization rates in the absence and presence of bath Ca²⁺ (Fig. 5F).

Fig. 5.

CDD is a P2X2R-specific phenomenon and requires transmembrane and/or intracellular domains. A–E: representative whole-cell recordings from HEK293 cells expressing rP2X4R (A), rP2X3R (B), rP2X4/2aR (C), rP2X4/2bR (D), or rP2X3/2aR chimeras (E) and bathed in Ca²⁺-deficient (gray traces) or Ca²⁺-containing (black traces) medium. In all cases, cells were clamped at −60 mV, the pipette solution contained 10 mM EGTA, currents were gated with 100 μM ATP, and washout periods between pulses were 4 min. F: summary of Ca²⁺ dependency of P2XRs and chimeras, reflected in the decrease of desensitization constant (τ_des) from cells bathed Ca²⁺-deficient (gray circles) Ca²⁺-containing medium (black circles, representing the third ATP application except for P2X3R, which corresponds to the first application); means ± SE values, n = 4.

CDD depends on the amplitude of rP2X2R current.

We next used the rP2X2a/V49-V61X4R chimera in which the Ile⁵⁰-Glu⁶³ fragment of P2X2aR was replaced with the Val⁴⁹-Val⁶¹ fragment of rP2X4R (20). This region is part of the lateral portals that have been suggested as the entry pathway for cations to access to the channel pore (33). This chimera was functional, but the peak amplitude of current recorded at −60 mV in response to 100 μM ATP application was dramatically reduced (0.49 ± 0.06 nA vs. 5.72 ± 0.26 nA observed with the wild-type rP2X2aR). Here, we used this chimera as a receptor model to test the dependence of CDD on current amplitude.

This chimera showed only a small CDD, with no further increases after repetitive applications, when bathed in Ca²⁺-containing medium (Fig. 6, A and B). The lack of development of progressive CDD in cells expressing rP2X2a/V49-V61X4R when recording was done using 10 mM EGTA-containing pipette solution could indicate that small amplitude currents generated by this chimera are not sufficient to elevate intracellular Ca²⁺. To test this hypothesis, we performed two experiments. First, we increased holding potential to −120 mV to increase the driving force for Ca²⁺ influx. Under these experimental conditions, CDD started to appear (Fig. 6C). Second, when the intrapipette EGTA was lowered from 10.00 mM to 0.05 mM, we observed the development of CDD (Fig. 6D). When 2 mM Ca²⁺ was added to the intracellular solution, cells immediately desensitized, and CDD was maximal at the third ATP application (Fig. 6, E and F). These results indicate that the balance between buffering capacity of EGTA and the amplitude of Ca²⁺ influx determines development of CDD of P2X2R.

Fig. 6.

Dependence of CDD on current amplitudes, studies with rP2X2a/V49-V61X4 chimera. A and B: lack of repetitive ATP application to generate CDD in cells expressing chimeric receptor using standard whole-cell recording. Representative recordings from a single cell bathed in Ca²⁺-containing medium (A) and means (white curves) ± SE (black and gray areas above and below white curves) from 5 recordings (B) obtained in Ca²⁺-containing (black) and Ca²⁺-deficient (gray) medium. Notice small amplitude of currents generated by this chimera. ECS, extracellular solution; ICS, intracellular solution. C: recordings from a chimera-expressing cell, clamped at −120 mV to facilitate Ca²⁺ influx. D: recordings from a single chimera-expressing cell with low intrapipette EGTA to facilitate intracellular accumulation of Ca²⁺. E: rapid receptor desensitization in a chimera-expressing cell clamped with 2 mM Ca²⁺ through pipette. F: summary of τ_des of the P2X2a/V49-V61X4 chimera at the third ATP application with different pipette solution compositions, n = 5.

Domain calcium is sufficient to facilitate receptor desensitization.

To further study dependence of rP2X2aR desensitization on intracellular calcium concentrations, two types of experiments were performed. In the first series of experiments, the intrapipette calcium concentration was fixed to 0.8 nM, 25.0 nM, and 200.0 nM, and Ca²⁺-deficient bath medium was used without calcium chelators, containing about 90 μM free calcium. At 0.8 nM free intrapipette [Ca²⁺], ATP induced slow desensitizing currents, and CDD was not developed (Fig. 7A). However, at 25 nM free intrapipette [Ca²⁺], the development of a progressive CDD was observed (Fig. 7B). When free intrapipette [Ca²⁺] was increased to 200 nM, a strong CDD was developed instantaneously, and additional ATP applications were unable to elicit P2X2aR-mediated currents, probably attributable to receptor rundown (Fig. 7C).

Fig. 7.

Domain Ca²⁺ is sufficient for development of CDD. A–C: recordings of rP2X2aR-expressing HEK293 cells in the absence of extracellularly added Ca²⁺ and with estimated free intrapipette [Ca²⁺] of 0.8 nM (A), 25.0 nM (B) or 200.0 nM (C). In all experiments, currents were measured using the whole-cell configuration at a holding potential of −60 mV. Recordings are representative of at least 4 different experiments. D and E: simultaneous current and [Ca²⁺]_i recordings in GT-1 cells expressing the rP2X2aR and stimulated with 100 μM ATP in the absence (D) and in the presence (E) of 10 mM EGTA in the intracellular solution. Inset: comparison of the rates of current growth during first 5 s of ATP application shown in D and E.

In the second series of experiments, simultaneous measurements of current and [Ca²⁺]_i were done. Generation of global calcium signals was observed only when GT1-7 cells, which do not express endogenous P2YRs, were bathed in 2 mM Ca²⁺-containing medium and intrapipette medium was without EGTA (Fig. 7D). In regular 10 mM EGTA-containing intrapipette solution, no rise in [Ca²⁺]_i was observed although the typical profile of current was still observed (Fig. 7E). These results indicate that the domain Ca²⁺ is sufficient for development of CDD.

DISCUSSION

P2XRs conduct Ca²⁺ and Mg²⁺ (14, 21, 30), but these metals can also inhibit those receptors at millimolar concentrations (5, 13, 30, 38). This raised the possibility that these cations, acting as extracellular or intracellular messengers, may control the receptor gating. In the present work, we focused on the role of Ca²⁺ in receptor desensitization by changing its bath and intrapipette concentrations and keeping the Mg²⁺ concentration gradient constant. Our results indicate that P2X2R desensitization has two components, bath Ca²⁺-independent desensitization and CDD. Calcium-independent receptor desensitization is minor in cells expressing P2X2aR but more pronounced in cells expressing shorter splice versions of this receptor, P2X2bR and P2X2eR. These findings are consistent with the literature (3, 26, 27, 35), indicating a role of the COOH terminus in control of receptor desensitization.

The main focus in the present work is on the bath Ca²⁺-dependent receptor desensitization. In whole-cell recording, CDD is pronounced in P2X2aR-expressing cells and has a form of use-dependent desensitization, a finding consistent with our previous report on this topic (23). CDD was observed with P2X2aR from three species, mouse, rat, and human, further indicating that this is not a species-specific phenomenon. CDD is also visible in P2X2bR-expressing cells but not in P2X2eR-expressing cells. The lack of development of CDD in P2X2eR does not necessarily indicate that the missing segment in the COOH terminus accounts for CDD but suggests that rapidly desensitizing receptors do not permit development of CDD. Consistent with this hypothesis, CDD was not observed in P2X3R-expressing cells but was visible in slower desensitizing P2X4R. Finally, we show that receptor desensitization is facilitated in the presence of bath Ca²⁺ in a concentration-dependent manner independently of the composition of other ions in medium and the accompanied pore dilation.

In general, bath Ca²⁺ could influence P2X2R gating extracellularly, acting on the pore segment, or intracellularly. Experiments with positive holding potential and the use of receptor chimeras further revealed a minor contribution of extracellular actions of Ca²⁺ on receptor desensitization, with an extracellular Ca²⁺ sensor(s) located presumably in the lateral portals (22, 33) in a sequence that does not contain the putative phosphorylation sites but contains negatively charged residues that could account for allosteric action of bath Ca²⁺. Furthermore, these experiments could suggest that Ca²⁺ influx through the channel pore is responsible for the development of CDD, in addition to the putative alterations of gating properties attributable to changes in membrane potential.

Several studies have addressed the importance of ectodomain, transmembrane, and intracellular domains in P2XR desensitization. For example, a number of extracellular residues has been identified to contribute to P2X3R desensitization (15). The involvement of the Cys-rich domain of the P2X1R in desensitization has also been proposed (29). When transmembrane domains of P2X2R are transferred to P2X1R or P2X3R, the resulting chimeras exhibit slower desensitization profiles (40). The P2X2R COOH-terminal domain also plays a role in receptor desensitization, as can be deduced from the different properties of its COOH-terminal splice variants (3, 25–27). The NH₂-terminal residues also participate in P2XR desensitization; the threonine-18 residue has been reported to be important for P2X2R desensitization (2), and the hydrophobic serine-15 residues have been reported for P2X3R desensitization (17). In the P2X1R, the contribution of both NH₂ and COOH termini to receptor desensitization domains has been found with the use of chimeric receptors and voltage-clamp fluorometry (1, 16). Thus desensitization of P2XRs is a complex process that depends on multiple domains and probably involves the interaction of those domains and conformational changes that occur during the gating process. At the present time, we could not dissociate whether the transmembrane and/or intracellular domains are critical for development of CDD.

However, we progressed in understanding the nature of Ca²⁺ signals needed for the development of CDD. Experiments with the P2X2a/V49-V61X4R chimera revealed that CDD was not observed in normal electrophysiological recording conditions but was evident when EGTA was lowered from the pipette solution. Under the same experimental conditions, CDD was developed in cells expressing the wild-type P2X2aR; the main difference between these receptors was the amplitude of current response, the former representing only ∼9% of that observed with the wild-type receptor. At the present time, we do not know whether the small amplitudes of current reflect the low plasma membrane expression of this chimera or changes in the gating properties. Independently of that, we believe that these data indicate that amplitude of current response, i.e., the amount of Ca²⁺ influx and the buffering capacity of intrapipette solutions, determines development of CDD.

Our experiments with simultaneous [Ca²⁺]_i and current measurements clearly indicate that, in experimental conditions where we observe CDD, there is no formation of global calcium signals. Furthermore, the twofold increase in global [Ca²⁺]_i concentrations above resting concentration (around 100 nM) causes rapid receptor desensitization. We observe regular development of CDD in cells with intrapipette [Ca²⁺] of only 25 nM, indicating that the process of receptor desensitization could be triggered at global [Ca²⁺]_i below resting. This indicates that domain rather than global Ca²⁺ signals are sufficient for development of CDD. In accordance with this, a recent analysis using a mathematical model indicates that Ca²⁺ chelators are not sufficient to abolish domain [Ca²⁺]_i signals; in fact, they can even increase them (39). We suggest that Ca²⁺ concentration in the P2X2R pore or the intracellular domain located immediately after the pore domain signals for receptor desensitization.

In general, the effect of elevated basal [Ca²⁺]_i on CDD could be direct, i.e., Ca²⁺ acting as an allosteric regulator, as it has been recently shown for P2X7R (41), or indirect, i.e., Ca²⁺ activating some signaling molecule. For example, it has been suggested that P2X7R-mediated currents are facilitated by calmodulin in a Ca²⁺-dependent manner (31). For the P2X2R, an interaction has been suggested in neurons with the Ca²⁺ sensor VILIP1 that could regulate some of the receptor properties, such as ATP sensitivity and maximal response (9). It has been also suggested that PKC and PKA may contribute to the control of P2XR gating (2, 11), and calcium directly contributes to activation of PKC and indirectly to activation of PKA by stimulating Ca²⁺-sensitive adenylyl cyclases. Our ongoing pharmacologically based experiments are oriented to address these questions to establish the mechanism(s) by which Ca²⁺ increases P2X2R desensitization.

In conclusion, we studied in detail the role of Ca²⁺ in P2X2R gating. Our results indicate that receptor desensitization is drastically affected by lowering bath and intrapipette calcium concentrations. The main targets for Ca²⁺ effects are the P2X2R transmembrane and/or intracellular domains, suggesting the existence of a high-affinity Ca²⁺ microdomain in this receptor, although we do not know whether these effects are directly or indirectly mediated by Ca²⁺. Further studies are also needed to identify the residue(s) responsible for sensing Ca²⁺ and the mechanism of its action.

GRANTS

This work was funded by FONDECYT Initiation Grant no. 11121302 (C. Coddou) and Intramural Research Program of the Eunice Kennedy Shiver National Institute of Child Health and Human Development, National Institutes of Health (S. Stojilkovic, Z. Yan, and C. Coddou).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: C.C., Z.Y., and S.S.S. conception and design of research; C.C. and Z.Y. performed experiments; C.C. and S.S.S. analyzed data; interpreted results of experiments; prepared figures; drafted manuscript; and edited and revised manuscript; C.C., Z.Y., and S.S.S. approved final version of manuscript.