Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

During the prolonged agonist occupancy, ATP-gated purinergic receptor-channels (P2XRs) become refractory to the stimulus and cellular responses decline. This process, called desensitization, is common for ligand-gated channels and occurs because liganded receptors enter stable conformations through which ion permeation is blocked or attenuated. Based on the observed differences in their desensitization kinetics, homomeric P2XRs are generally divided into three groups: P2X₁R and P2X₃R desensitize very rapidly and P2X₄R and P2X₆R desensitize with a moderate rate, whereas P2X₂R, P2X₅R, and P2X₇R show little or no desensitization (North and Barnard, 1997; Ralevic and Burnstock, 1998). Heteromultimerization results in channels that desensitize with different kinetics from those seen in cells expressing homomeric channels; the influence of participating subunits on channel desensitization pattern is well documented for P2X₂R+P2X₃R (Lewis et al., 1995; Radford et al., 1997). The differences in desensitization rates of P2XRs are reminiscent of those seen among subtypes of other ligand-gated receptor-channels (McBain and Mayer, 1994; Lerma et al., 2001).

The underlying molecular mechanisms of P2XR desensitization have been incompletely characterized. Calcium and other divalent cations influence the rate of desensitization and the rate of recovery from desensitization in native and cloned channels (Cook et al., 1998; Ding and Sachs, 2000). A highly conserved protein kinase C site located in the N terminus of P2XRs may control the rate of desensitization of P2X₁R, P2X₂R, and P2X₃R (Boue-Grabot et al., 2000; Paukert et al., 2001; Ennion and Evans, 2002). Phosphorylation of a protein kinase A site in the C terminus of P2X_2aR may also participate in receptor desensitization (Chow and Wang, 1998). Experiments with chimeras composed of P2X₂R and P2X₁R or P2X₃R subunits suggested that the rapid desensitization requires interactions between two transmembrane domains of receptor subunits (Werner et al., 1996). Several groups have also reported that the C-terminal splice variant of P2X₂R, called P2X_2bR or P2X_2-2R, lacks a stretch of 69 residues and desensitizes faster than the full-length channel, called P2X_2aR (Brandle et al., 1997; Simon et al., 1997; Koshimizu et al., 1998b). The site-directed mutagenesis experiments suggested important roles of different residues in the C-terminal tail in P2X₂R desensitization (Koshimizu et al., 1998a; Zhou et al., 1998; Smith et al., 1999). The variable C-terminal structures may also influence the desensitization rates of other members of P2XRs, including P2X₃R and P2X₄R (Koshimizu et al., 1999). A large C terminus of P2X₇R also accounts for the nondesensitizing pattern of these channels during repetitive stimulation (Surprenant et al., 1996).

Here, we examined the interactions between the ectodomain and C-terminal domain in controlling the pattern of P2XR desensitization. Specifically, we studied the effects of altering the agonist and substituting the ectodomain on the pattern of calcium signaling by P2XRs. For this purpose, we used P2X_2aR and P2X_2bR because of their identical ectodomains and distinct desensitization patterns in response to ATP. The P2X₂ receptor-subtype specific desensitization pattern was observed not only in current measurements, but also in single-cell calcium measurements (Koshimizu et al., 1998, 2000), indicating that such recordings are sufficient for studies with this particular receptor. These experiments revealed that the structure-dependent desensitization pattern of P2X_2aR and P2X_2bR reflects the efficacy of agonists for these receptors.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

DNA Constructs. The coding sequences of the rat P2X_2a, P2X_2b, P2X₃, and P2X₇ subunits were isolated by reverse transcription-PCR (Koshimizu et al., 1999), and subcloned into the bicistronic enhanced fluorescent protein expression vector pIRES2-EGFP (BD Clontech, Palo Alto, CA) at the restriction enzyme sits of XhoI/PstI for P2X_2aR and P2X_2bR, and XhoI/EcoRI for P2X₃R and P2X₇R. Chimeric subunits, termed P2X_2a+X₃EC and P2X_2b+X₃EC, contain extracellular domain from Val⁶⁰ to Phe³⁰¹ of P2X₃R instead of the native Ile⁶⁶-Tyr³¹⁰ sequence of P2X_2aR and P2X_2bR (Fig. 1). To exchange the corresponding extracellular regions, two restriction endonuclease sites for SacI and EcoRI were introduced into both P2X₂ and P2X₃ subunits using PCR-based overlap extension method (Horton et al., 1989). Primer sequences carrying silent nucleotide substitutions (underlined) are as follows: X2EcoL, 5'-AACATCGATTCGAATTCCATAGGCTTTGAT-3'; X3SacU, 5'-ATTGAGAGCTCAGTAGTTACAAAGGTG-3'; X3SacL, 5'-GCTCTCAATGGCGGTGTCCCTCACTTG-3'; X3EcoU, 5'-GGAATTCGCTTTGATGTGCTGGTA-3'; and X3EcoL, 5'-GCGAATTCCAAAAGCCTTCAGGAGTGT-3'. By two rounds of PCR, the entire protein coding regions for the SacI/EcoRI-carrying P2X_2a, P2X_2b, and P2X₃ subunits, termed P2X_2aSE, P2X_2bSE, and P2X₃SE, were amplified as described previously (Koshimizu et al., 1999). Before subcloning these PCR products, pBluescript vector (Stratagene, La Jolla, CA) was digested with SacI, treated with Klenow enzyme for end-filling, and self-ligated to remove this intrinsic SacI site. The PCR products were then subcloned into HincII/SmaI site of the modified pBluescript vector and sequenced. Correctly subcloned inserts were digested with SacI and EcoRI and the fragment corresponding to the putative extracellular loop of P2X₃R was transferred to P2X_2aSE and P2X_2bSE, generating chimeric receptors, P2X_2a+X₃EC and P2X_2b+X₃EC, respectively. For mammalian expression, the NheI/XhoI digested fragments of P2X_2a+X₃EC and P2X_2b+X₃EC were also transferred to pIRES2-EGFP.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the wild-type and chimeric constructs used in this study. White horizontal rectangles indicate P2X2aR and its spliced form lacking the Val370-Gln438 C-terminal sequence (shown in gray), called P2X2bR, black rectangles indicate P2X3R, and dotted rectangles indicate P2X7R. Vertical dashed rectangles indicate the positions of putative transmembrane domains TM1 and TM2. The constructed P2X2a+X3EC and P2X2b+X3EC chimeric receptors contain Val60-Phe301 extracellular domain sequence of P2X3R instead the native Ile66-Tyr310 sequence of P2X2aR and P2X2bR. The constructed P2X2a+X7EC and P2X2b+X7EC chimeric subunits contain Val61-Phe313 extracellular domain of P2X7R instead the native sequence Ile66-Tyr310.

Cell Culture and Transient Transfection. Mouse immortalized gonadotropin-releasing hormone-secreting GT1-7 cells (GT1 cells) were used to examine the patterns of Ca²⁺ signaling evoked by P2XRs as described previously (Koshimizu et al., 1999). GT1 cells were routinely maintained in Dulbecco's modified Eagle's medium/Ham's F12 medium (1:1), containing 10% (v/v) fetal bovine serum and 100 µg/ml gentamicin (Invitrogen Corp., Carlsbad, CA) in a water-saturated atmosphere of 5% CO₂ and 95% air at 37°C. Before the day of transfection, cells were plated on 25-mm coverslips coated with poly(L-lysine) (0.01% w/v; Sigma, St. Louis, MO) at a density of 0.75-1.0 × 10⁵ cells per 35-mm dish. For each dish of cells, transient transfection of expression constructs was conducted using 1 µg of DNA and 7 µl of LipofectAMINE 2000 Reagent (Invitrogen) in 3 ml of serum-free Opti-MEM. After 6 h of incubation, transfection mixture was replaced with normal culture medium. Cells were subjected to experiments 24 to 48 h after transfection.

[Ca²⁺]_i Measurements. Transfected GT1 cells were preloaded with 1 µM Fura-2 acetoxymethyl ester (Molecular Probes, Eugene, OR) for 60 min at room temperature in modified Kreb's Ringer buffer (120 mM NaCl, 5 mM KCl, 1.2 mM CaCl₂, 1.2 mM KH₂PO₄, 0.7 mM MgSO₄, 1.8 g/l Glucose, and 15 mM HEPES, pH 7.4). After dye loading, cells were incubated in modified Kreb's Ringer buffer and kept in the dark for at least 30 min before single-cell [Ca²⁺]_i measurement. Coverslips with cells were mounted on the stage of an Axiovert 135 microscope (Carl Zeiss, Oberkochen, Germany) attached to the Attofluor Digital Fluorescence Microscopy System (Atto Instruments, Rockville, MD). Cells were stimulated with various doses of agonists (added by pipette at room temperature) and the dynamic changes of [Ca²⁺]_i were examined under a 40× oil-immersion objective during exposure to alternating 340 and 380 nm light beams, and the intensity of light emission at 520 nm was measured. The ratio of light intensities, F₃₄₀/F₃₈₀, that reflects changes in [Ca²⁺]_i was simultaneously followed in several single cells. Apyrase (Grade I; Sigma, St. Louis, MO) was used at 0.2 U/ml throughout the incubation process, loading with Fura-2 acetoxymethyl ester, and [Ca²⁺]_i recording in cells expressing P2X₃R, P2X_2a+X₃EC, and P2X_2b+X₃EC receptors. Experiments with P2X_2aR, P2X_2bR, P2X₇R, and their chimeras were done without apyrase. GFP was used as a marker for cells with P2XR expression as described previously (Koshimizu et al., 1999, 2000). Cells expressing GFP were optically detected by an emission signal at 520 nm when excited by 488-nm ultraviolet light and were not detectable by 340- and 380-nm excitations in the absence of Fura-2.

Calculations. To minimize the impact of receptor saturation kinetics on the [Ca²⁺]_i profiles, agonists were added rapidly and were continuously present during the recording. Thus, the rise in [Ca²⁺]_i predominantly reflects the bound-open equilibrium, whereas the decay represents the equilibration into desensitization state (Auerbach and Akk, 1998). The time course of the [Ca²⁺]_i was fitted to a single exponential function using Prism software (GraphPad Software, San Diego, CA). All values in the text are reported as mean ± S.E.M. Significant differences, with P < 0.05, were determined by one-way analysis of variance with Newman-Keuls multiple comparison test. Concentration-response relationships were fitted to a four-parameter logistic equation using a nonlinear curve-fitting program (Kaleidagraph; Synergy Software, Reading, PA) that derives the EC₅₀ and Hill values. Calcium recordings were done in 15 to 50 cells simultaneously, and each experiment was repeated three or more times to ensure the reproducibility of the findings.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

P2X_2aR and P2X_2bR Exhibit Similar EC₅₀ Values for Agonists. The native and chimeric P2XRs were subcloned into GFP-expression vector pIRES2-EGFP, and the relative transfection efficiency of P2XRs constructs was estimated in single cells by analyzing the intensity of fluorescence signals, as described previously (Koshimizu et al., 2000). In the presence of fixed amount of expression constructs and comparable post-transfection times, the percentage of GFP+ATP-positive cells varied between 45 and 60% and was independent on the channel type expressed. When the average GFP fluorescence was similar for each set of cells (about 60 arbitrary units), the mean amplitude of peak [Ca²⁺]_i to 100 µM ATP were highly reproducible for the same channel types. No repetitive stimulation was done to avoid the possible impact of desensitization on the amplitude and pattern of [Ca²⁺]_i signals. Also, in all experiments agonists were added rapidly to the coverslip dish to minimize the impact of agonist diffusion on the profile of [Ca²⁺]_i signals.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Comparison of the effects of ATP and two analogs, BzATP and ,-meATP, on the peak calcium response and rates of signal desensitization in cells expressing P2X2aR and P2X2bR. In this and following figures, experimental records are shown by open circles (mean values from at least 15 traces in representative experiments) and fitted curves by full lines. A single exponential function was sufficient to describe the desensitization rates. The fitted function is extrapolated for clarity. Agonists were added in concentrations indicated below traces and were continuously present during the recording.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Concentration dependence of agonist-induced peak calcium response in GT1 neurons expressing homomeric P2X2aR and P2X2bR. A and B, comparison of the effects of ATP and its analogs, BzATP and ,-meATP, on peak calcium response in cells expressing P2X2aR (A) and P2X2bR (B). Data shown are means derived from three to eleven experiments per dose, each done in at least 15 single cells. S.E.M. were within 10%. Numbers above dotted vertical lines indicate the calculated EC50 values for three agonists.

The C-Terminal-Dependent Desensitization Pattern of P2X₂R Is Ligand-Specific. The desensitization rates of [Ca²⁺]_i signals generated by two receptors were also dependent on agonist concentrations. Figure 2, left, shows typical desensitization profiles in P2X_2aR- and P2X_2bR-expressing cells stimulated with increasing ATP concentrations. Consistent with the relevance of C-terminal domain structure of P2X₂Rs in control of receptor desensitization (Brandle et al., 1997; Simon et al., 1997; Koshimizu et al., 1998b), P2X_2bR desensitized more rapidly than P2X_2aR (Fig. 2, left). Stimulation with increasing BzATP and -meATP concentrations also produced a progressive increase in the rates of signal desensitization (Fig. 2, middle and right). Furthermore, P2X_2bR-expressing cells desensitized more rapidly than P2X_2aR-expressing cells during the prolonged stimulation with high concentrations of BzATP, whereas signals generated by two receptors desensitized with comparable rates in response to -meATP.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Ligand-specific receptor desensitization pattern of P2X2Rs. A-C, concentration-dependent effects of agonists on rates of signal desensitization in GT1 neurons expressing homomeric P2X2aR and P2X2bR. Circles represent means ± S.E.M. from four to five independent experiments per dose, each performed on 15 to 45 cells. Numbers indicate the calculated IC50 values for ATP, BzATP, and -meATP. The dotted horizontal line indicates comparable levels of P2X2aR desensitization rates at high agonist concentrations. To compare IC50 values with EC50 values, see Fig. 2. D-F, comparison of the effects of 2-MeS-ATP, ATP, ATPS, BzATP, and -meATP on the rate of P2X2Rs desensitization. D and E, differences in the rates of calcium signal desensitization in cells expressing P2X2bR. Traces shown by open circles are means derived from 15 to 45 cells in a representative experiment. All agonists were added in 500 µM concentrations. F, the relationship between EC50 values for agonists and desensitization rates calculated at 500 µM concentrations in P2X2aR-and P2X2bR-expressing cells.

Increase in P2X₂R Sensitivity for Agonists Facilitates Desensitization. We further examined the agonist-specific desensitization pattern of P2X₂Rs by producing the chimeric receptors with increased and decreased sensitivity for agonists. To make P2X₂Rs with high sensitivity for agonists, we constructed chimeric receptors containing the Val⁶⁰-Phe³⁰¹ extracellular domain sequence of P2X₃R instead the native Ile⁶⁶-Tyr³¹⁰ sequence, and called them P2X_2a+X₃EC and P2X_2b+X₃EC chimeras. In our experimental conditions, the native P2X₃R and P2X_2a+X₃EC and P2X_2b+X₃EC chimeras did not respond to ATP, BzATP, and -meATP or responded with irregular [Ca²⁺]_i patterns when cells were incubated without apyrase, whereas the pattern of responses was not affected in P2X_2aR-and P2X_2bR-expressing cells. These results confirm our earlier findings (Koshimizu et al., 1999) that endogenous ATP secretion is sufficient to desensitize receptors with high potency for ATP. When experiments were done in the presence of apyrase, all three receptors responded to agonist stimulation in a concentration-dependent manner and with the EC₅₀ values for ATP, -meATP, and BzATP in a submicromolar concentration range. Figure 5A, left, illustrates the dose-response to ATP in P2X_2aR+X₃EC-expressing cells. There was a ~30-fold decrease in EC₅₀ for ATP, a ~25-fold decrease for BzATP, and a ~150-fold decrease in EC₅₀ for -meATP in P2X_2aR+X₃EC-expressing cells compared with the native P2X_2aR (Fig. 5A). The P2X_2bR+X₃EC chimera exhibited the same shift in EC₅₀ values for two agonists (not shown). On the other hand, the peak amplitude of [Ca²⁺]_i responses in cells expressing chimeric receptors was 2- to 3-fold higher than that observed in P2X₃R-expressing cells (Fig. 5B).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Influence of the substitution of ectodomain at P2X2Rs on agonistic potency of ATP (left), BzATP (middle), and -meATP (right). A, change in the EC50 values for agonists in cells expressing P2X2a+X3EC receptor. The results shown are means ± S.E.M. B, comparison of the peak amplitude of [Ca2+]i signals in P2X2a+X3EC-expressing cells. P2X2b+X3EC receptors showed comparable leftward shifts in EC50 values for three agonists and a decrease in peak amplitude of [Ca2+]i responses. P2X2R+X3EC chimeras were constructed as described under Materials and Methods.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   The pattern of desensitization in cells expressing wild-type P2X3R and P2X2aR and P2X2a+X3EC chimera. A and B, comparison of the effects of 100 µM -meATP (A) and 100 µM ATP (B) on the rates of calcium signal desensitization in GT1 neurons expressing P2X3R (left traces) P2X2aR+X3EC chimera (central traces), and P2X2aR (right traces). Traces shown are representative from three to eight independent experiments, each done in at least 15 cells. Numbers above traces represent mean ± S.E.M. values of desensitization rates. Asterisks indicate significant differences compared with desensitization rates for P2X3R-and P2X2aR-expressing cells.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Patterns of BzATP-induced calcium signaling in GT1 cells expressing wild-type P2X3R and P2X2aR and chimeric P2X2a+X3EC receptors. A-C, concentration-dependent effects of BzATP on the rates of calcium signal desensitization in GT1 neurons expressing wild-types and chimeric receptors. Traces shown are representative from three to five independent experiments. Numbers above traces represent mean ± S.E.M. values of desensitization rates. Asterisks indicate significant differences compared with the desensitization rates for P2X3R-and P2X2aR-expressing cells.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Agonist-induced calcium signaling in GT1 cells expressing P2X2bR and P2X2b+X3EC receptors. A to C, typical patterns of P2X2bR+X3EC desensitization in response to 1 µM ATP (A), 1 µM BzATP (B), and 1 µM -meATP (C). Traces shown are representative from five to six independent experiments. Mean ± S.E.M. values for receptor desensitization kinetics are shown in Table 1.

Decrease in P2X₂R Sensitivity for ATP Blocks C Terminus-Dependent Desensitization. To further test the hypothesis about the relevance of ectodomain for C-terminal structure-dependent desensitization, we made chimeric P2X_2aR and P2X_2bR with lower sensitivity to ATP and higher sensitivity to BzATP, compared with the wild-type channels. This was achieved by constructing P2X_2a+X₇EC and P2X_2b+X₇EC chimeric receptors containing the Val⁶¹-Phe³¹³ extracellular domain sequence of P2X₇R instead of the native Ile⁶⁶-Tyr³¹⁰ sequence. In accordance with the literature (Surprenant et al., 1996), native P2X₇R expressed in GT1 neurons responded to BzATP stimulation with a rapid and nondesensitizing rise in [Ca²⁺]_i (Fig. 9A), with a calculated EC₅₀ of 8 µM (Fig. 9D). In a majority of cells, ATP also induced similar patterns of [Ca²⁺]_i signaling, albeit of smaller amplitude, whereas a fraction of cells (about 30% in response to 500 and 1000 µM ATP) responded with atypical [Ca²⁺]_i profiles (Fig. 9B). In all concentrations studied, ATP was less effective compared with 65 µM BzATP (Fig. 9C), and the estimated EC₅₀ was 485 µM (Fig. 9G).

View larger version (48K): [in this window] [in a new window]   Fig. 9.   Characterization of calcium signaling pattern by chimeric P2X2Rs containing the P2X7R ectodomain. A to C, characterization of agonist-induced calcium response in P2X7R-expressing cells. Concentration-dependent effects of BzATP (A) and ATP (B) on [Ca2+]i response. Numbers indicate concentrations of agonists that were continuously present during the recording. C, comparison of the agonistic potency of ATP and BzATP. D to F, comparison of the BzATP effects on peak [Ca2+]i response (D) and rates of receptor desensitization (E and F) in cells expressing P2X2R, P2X7R, and P2X2+X7EC receptors. G to I, comparison of the ATP effects on peak [Ca2+]i response (G) and rates of receptor desensitization (H and I) in cells expressing P2X2R, P2X7R, and P2X2+X7EC receptors. No difference in peak [Ca2+]i responses were observed between P2X2a+X7EC and P2X2b+X7EC and these results are shown combined. Arrows indicate differences in the EC50 values (D and G) and half-times for receptor desensitization (E and H). The results shown are means ± S.E.M. with three to eleven experiments per dose. In all, horizontal bars above and below traces indicate the duration of agonist stimulation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Two main hypotheses emerged from previous work on desensitization of P2XRs, one based on the structure of channels, and the other based on the actions of intracellular messengers. The dual control of P2XR may well be expected from studies on such allosteric proteins, and is reminiscent of those seen with other ligand-gated and voltage-gated channels. For example, desensitization of glutamate receptors depends on N-terminal domain (Krupp et al., 1998), the flip-flop cassette (Sommer et al., 1990), and M3-M4 domain (Partin et al., 1995), as well as on intracellular messengers in the postsynaptic cells, including Ca²⁺ (Krupp et al., 1996). The desensitization properties of AMPA receptors can be modified by alternative splicing and mRNA editing, and by heteromeric assembly of channels (Sommer et al., 1990; Robert et al., 2001). In cyclic nucleotide-gated channels, the agonist-binding domain is in the C terminus, and the N-terminal domain alters the efficacy of agonists through interactions with the ligand-binding site by a Ca²⁺-calmodulin-sensitive mechanism (Tibbs et al., 1997; Varnum and Zagotta, 1997). The structure of intracellular domains of voltage-gated channels are also critical for their voltage-dependent inactivation, whereas the functional control of these channels is mediated by various intracellular messengers (Hille, 1991).

Here we focused on the mechanism of C-terminal structure-dependent P2X₂R desensitization. Two sister receptors, P2X_2aR and P2X_2bR, exhibit comparable activation profiles for ATP and peak current/[Ca²⁺]_i responses but desensitize with different rates. P2X_2aRs desensitize slowly and partially, whereas P2X_2bRs desensitize rapidly and to the steady levels significantly lower than that of P2X_2aR (Brandle et al., 1997; Simon et al., 1997; Koshimizu et al., 1998b). Similar EC₅₀ values for ATP are consistent with identical structure of extracellular domains for these receptors. Different rates of receptor desensitization, on the other hand, indicate the relevance of Val³⁷⁰-Gln⁴³⁸ C-terminal sequence, deleted in P2X_2bR, for receptor-desensitization. In our experiments, potency of several agonists for P2X_2aR and P2X_2bR were in an order (2-MeS-ATP  ATP  ATPS < BzATP -meATP) comparable with results obtained by others (reviewed in Ralevic and Burnstock, 1998).

We also show that these two receptors desensitized in a concentration-dependent manner and with the same order of agonists. However, the IC₅₀ values for desensitization were right-shifted compared with the EC₅₀ values for activation of channels. This is a novel finding for P2XRs but has been shown for other ligand-gated channels. For example, the extent of AMPA receptor desensitization increases with agonist concentrations (Vyklicky et al., 1991). The half-maximal activation of kainate channels occurs at glutamate concentrations of 330 µM, whereas the half-maximal steady state desensitization occurs at ligand concentrations 20 times lower. A similar ratio was also observed for GluR6 homomers when kainate was used as an agonist (Lerma et al., 2001). Also, concentrations required to desensitize Torpedo californica receptors are nearly 1000-fold lower than those required for activation (Corringer et al., 1998). A dual aspect of agonist pharmacology may contribute to the shaping of synaptic currents and modulating the fraction of activatable channels (Jones and Westbrook, 1996). However, the slow and incomplete inactivation of P2X_2aR and the right-shifted IC₅₀ for ATP argue against such a role of receptor desensitization in neurons expressing these channels.

The receptor subtype-specific desensitization pattern was observed in response to ATP, the native agonist for these channels but also in response to stimulation with two analog agonists, 2-MeS-ATP and ATPS. However, the receptor-specificity of desensitization was less obvious when stimulated with BzATP and was lost when receptors were stimulated with -meATP. Furthermore, the ligand-specific desensitization patterns were recorded at maximal agonist concentrations, where peak amplitudes in [Ca²⁺]_i were comparable in P2X_2aR and P2X_2bR. Consistent with a role of agonist-binding domains in desensitization of other ligand-gated channels, both AMPA and glutamate maximally activate AMPA receptors, whereas kainate and domoate act as partial agonists, and produce much less desensitization than glutamate (Patneau and Mayer, 1990; Patneau et al., 1993; Swanson et al., 1997; Armstrong and Gouaux, 2000). The novel aspect in ligand-specific receptor desensitization emerging from this study is in coupling between ectodomain and C-terminal domain. In general, there was a parallelism between the rates of P2X_2aR and P2X_2bR desensitization and the EC₅₀ values for agonists. This suggests that C-terminal-dependent desensitization pattern is not an "all-or-none" phenomenon but a graded process that probably depends on ligand binding affinity and/or activation efficacy.

The relevance of ectodomain structure on C-terminal-dependent desensitization pattern was further confirmed in experiments with chimeric channels. The agonist-specific desensitization pattern of P2X₂Rs was lost by changing the native binding site of these channels with the P2X₃R extracellular domain. Both chimeras, P2X_2a+X₃EC and P2X_2b+X₃EC, exhibited about 30-, 25-, and 150-fold increase in the EC₅₀ values for ATP, BzATP and -meATP, respectively. The rates of desensitization for both chimeric receptors also increased for 2-3-fold. However, the C-terminal structure-dependent desensitization pattern was preserved; like native receptors, P2X_2b+X₃EC receptor desensitized more rapidly than P2X_2a+X₃EC receptor. Finally, the ligand-specific and receptor subtype-specific desensitization patterns reversed in cells expressing P2X_2a+X₇EC and P2X_2a+X₇EC receptors. Such chimeras showed lower sensitivity for ATP, compared with native P2X₂Rs and desensitized with comparable rates and higher sensitivity to BzATP accompanied with the C terminus-specific desensitization pattern.

At the present time, it is difficult to discuss the possible molecular mechanism of interactions between the ectodomain and C-terminal domain in development of desensitization. Calcium measurements used in our study provide several advantages. P2XR-generated calcium signals mediate the action of these receptors on cellular functions, including neurotransmission, hormone secretion, transcriptional regulation, and protein synthesis (Berridge, 1993). Thus, calcium rather than current profiles reflect the importance of a particular pattern of signaling on cellular functions. P2X₂Rs conduct calcium and the addition of nifedipine blocks the indirect (through voltage-gated L-type calcium channels) action of activated receptors in our expression system (Koshimizu et al., 2000), reflecting Ca²⁺ influx function of these channels. Single-cell calcium measurements can be done simultaneously in many cells, leading to better statistics, which are critical for interpretation of EC₅₀ and rates of desensitization. Measurements of GFP intensities also provide an effective mechanism for selection of cells with comparable expression of P2XRs and more reliable data on peak response and EC₅₀ values derived from these experiments (Koshimizu et al., 2000). However, [Ca²⁺]_i measurements also limit the interpretation of activation and desensitization properties of channels, because of calcium handling mechanism of the cells used in experiments.

Other limitation comes from the fact that the ligand-binding domain structure and the crystal structure of P2XRs have not been identified, in contrast to glutamate channels (Sun et al., 2002). In our chimeric receptors, the extracellular loop is derived almost entirely from P2X₃R and P2X₇R. Consistent with this, the ATP potency of the chimeric P2X₂+X₃EC receptors matches the ATP potency at native P2X₃R rather than native P2X₂R. However, the ATP potency at the P2X₂+X₇EC receptors is closer to the ATP potency at the parental P2X₂R rather than P2X₇R. These contradictory results suggest the relevance of flanking P2X₂R sequences on agonistic potency of ATP. We may speculate that these sequences act as "dominant-positive" domains to offset atypical low sensitivity of P2X₇R for native agonist. In accordance with this view, it has been reported recently that point mutations in the first transmembrane domain, specifically Phe⁴⁴, affect the ligand-selectivity of rat P2X₂R (Jiang et al., 2001).

In conclusion, our results show that homomeric P2X_2aR and P2X_2bR exhibit highly comparable EC₅₀ values for receptor activation by various agonists, but desensitize in a receptor- and agonist-specific manner. Pharmacological manipulations with activation of these receptors and molecular manipulations with their ectodomains indicate that the efficacy of agonists reflects the ligand-specificity of receptor desensitization; highly potent agonists trigger P2X₂R-subtype specific C terminus-mediated desensitization, whereas agonists with lower potency are less effective or ineffective. Thus, it seems that conformational changes needed for activation of P2X₂Rs are not always sufficient to trigger C terminus-controlled desensitization. These findings provide a solid base for further biophysical investigations on hypothesis that the affinity of agonists for receptors determines the strength of molecular conformational changes needed for development of C-terminal-controlled channel desensitization.