1The ATP-activated P2X2(a) and P2X2(b) receptor splice variants, which differ only in their C-terminal sequences, desensitize at different rates. We used mutational analysis to investigate the involvement of the C-terminal region in receptor desensitization. Rat wild-type and mutant P2X2 receptors were expressed in Xenopus oocytes and currents were measured using the two-electrode voltage-clamp technique.
2Truncating P2X2 at the Lys369 splice site increased the rate of desensitization by >100-fold. Recovery from desensitization was slowed by ∼5-fold.
3Addition of Val370 onto the C-terminus of the truncated receptor slowed desensitization by ∼70-fold. Point mutations that substituted either smaller or larger hydrophobic amino acids for Val370, within the P2X2(a) splice variant, had profound effects on the rate of desensitization. The rate decreased with increasing hydrophobicity but was not dependent upon the precise structure of the side group.
4A mutant receptor, with only nine amino acids, Val-Asp-Pro-Lys-Gly-Leu-Ala-Gln-Leu, beyond the Lys369 splice site, desensitized at a similar rate to P2X2(a). Injection of the peptide of this sequence into oocytes expressing P2X2(a) increased the rate of desensitization, whereas the eight-residue peptide lacking the valine had no effect.
5Neutralizing lysines in the vicinity of the splice site increased the rate of receptor desensitization. Substituting glutamine for Lys365 produced the greatest effect (∼30-fold increase), whereas mutating lysines that were further upstream or downstream of this position had progressively less of an effect.
6We conclude that the C-terminal splice site of the P2X2 receptor is located within a region that is critically involved in regulating the rate of receptor desensitization. The valine at position 370 interacts with an intracellular hydrophobic site to slow the rate of desensitization. Nearby lysines may facilitate this interaction.
Extracellular ATP is a signalling molecule which mediates its effects through activation of two distinct families of cell surface receptors: metabotropic P2Y purinoceptors and ionotropic P2X receptors (Abbracchio & Burnstock, 1994). P2X receptor channels are selective for cations and are appreciably permeable to Ca2+ (Benham & Tsien, 1987; Bean, 1992; Evans et al. 1995; Humphrey et al. 1998). They have been shown to be involved in mediating fast synaptic transmission both in the central and peripheral nervous systems.
All of the recombinant P2X receptor subunits form functional homomeric complexes when expressed in heterologous systems. These receptors differ in their pharmacological and biophysical properties. P2X1 and P2X3 receptors are α,β-methyleneATP-sensitive and desensitize rapidly in the continued presence of agonist. In contrast, P2X2 and P2X4-7 are insensitive to α,β-methyleneATP and desensitize slowly over a time period of tens of seconds. P2X2 and P2X3 have been shown to form functional heteromers with properties that differ from either of the parent homomers in that they are α,β-methyleneATP-sensitive but desensitize relatively slowly (Lewis et al. 1995). Recent studies suggest that heteromeric channels with novel phenotypes also form between P2X4 and P2X6 (Lêet al. 1998) and between P2X1 and P2X5 subunits (Torres et al. 1998). In vitro, P2X1, P2X2, P2X3, P2X5 and P2X6 can associate with each other to form stable complexes (Torres et al. 1999).
The mechanism(s) by which P2X receptors desensitize has not been elucidated. Werner et al. (1996) employed a chimeric approach to identify the structural differences between P2X1 and P2X2, underlying rapid desensitization in the former and slow desensitization in the latter. They showed that fast desensitization was introduced into P2X2 by transference of both transmembrane segments from either P2X1 or P2X3 and concluded that desensitization of P2X1 involves an interaction between these two regions. The C-terminal region of the P2X2 receptor has also been shown to be involved in regulating desensitization. Two P2X2 isoforms which differ only in their C-terminal region, desensitize at different rates (Brändle et al. 1997; Simon et al. 1997; Koshimizu et al. 1998a). The P2X2(b) isoform is missing a sixty-nine amino acid segment that is present in P2X2(a), and it desensitizes μ5-fold faster than the longer isoform. Koshimizu et al. (1998b) reported that four residues within the 69 amino acid segment, Pro373-Lys374-His375-Pro376, are involved in slowing receptor desensitization.
In this study we used site-directed mutagenesis to investigate the involvement of the P2X2 C-terminus in controlling the rate of receptor desensitization. We found that truncating the receptor at the Lys369 splice site dramatically increased the rate of desensitization to give a time constant of current decay of < 1 s, similar to P2X3 receptor currents. We identified two structural features which are important in producing the slowly desensitizing phenotype that is characteristic of the P2X2 subtype. One of these is the presence of a hydrophobic amino acid immediately adjacent to the splice site, at position 370, and the other is the presence of several lysines within the 20 amino acid region that lies proximal to TM2.
Deletion mutants were constructed by two stage PCR. In the first stage, either P2X2(a)-pcDNAIamp or P2X2(b)-pCMV BK was used as a template and the regions of DNA flanking the segment to be deleted were amplified in separate reactions. In the second stage, the two amplified fragments were annealed and used as a template for further amplification. The PCR product was then spliced into either P2X2(a) or P2X2(b) using the Rapid DNA Ligation Kit (Boehringer-Mannheim, Germany). Point mutations were generated using a similar approach. The primers contained one or several mismatch bases for the replacement of the targeted amino acids. All mutations were verified by DNA sequencing of the mutated region (Cambridge Bioscience, Cambridge, UK).
In vitro transcription
cRNAs were transcribed in vitro using the mmessage mmachine RNA transcription kit (Ambion, Austin, TX, USA) according to manufacturers instructions.
Preparation and injection of oocytes
Stage V/VI oocytes were obtained by ovariectomy from Xenopus laevis frogs anaesthetized with 0.3 % 3-aminobenzoic acid ethyl ester methane sulphonate (Sigma). Following the first ovariectomy the Xenopus was allowed to recover from anaesthesia. Prior to a second ovariectomy, the Xenopus was terminally anaesthetized in 0.6 % anaesthetic for 1 h. It was subsequently killed by opening the chest cavity and removing the heart. Oocytes were dissociated from connective tissue using 0.3 % collagenase IA (Sigma) in calcium-free OR2 solution containing 82.5 mm NaCl, 2.5 mm KCl, 1 mm MgCl2 and 5 mm Hepes (pH 7.6). Isolated oocytes were microinjected 24 h later with 50 nl cRNA (50-100 ng) dissolved in water. The cells were maintained in ND96 medium containing 96 mm NaCl, 5 mm Hepes, 2 mm KCl, 1.8 mm CaCl2, 1 mm MgCl2 (pH 7.6) and 0.1 mg ml−1 gentamicin at 18°C for up to 7 days.
Two-electrode voltage-clamp recordings were made 1–6 days after microinjection using an OC-725C amplifier (Warner Instruments Corp., Hamden, CT, USA), interfaced to a Power Macintosh 8100/100 computer using an ITC-16 A/D board (Instrutech Corp., Greatneck, NY, USA). Current traces were recorded with pulse acquisition software (version 8.11, HEKA electronics, Lambracht, Germany). Microelectrodes, filled with 3 M KCl, had resistances of 0.5–2 MΩ. In all experiments oocytes were voltage clamped at a membrane potential of −60 mV. The bath solution contained 96 mm NaCl, 2 mm Hepes, 1 mm MgCl2 and 0.1 mm BaCl2 (pH 7.4) and was continuously perfused at a flow rate of 2 ml min−1. The volume of the bath chamber was μ100 μl and all drugs were prepared in bath solution at their final concentration and applied by pipette after the inflow to the bath was stopped. Complete solution exchange was achieved in μ2 s. Experiments were carried out at room temperature (21-22°C). ATP (sodium salt) and cytochalasin D were purchased from Sigma (Poole, UK). Cytochalasin D was initially dissolved in dimethyl sulfoxide before being diluted at least 200-fold in bath solution to give the final working concentration. The peptides, VDPKGLAQL and DPKGLAQL were purchased from Research Genetics Inc. (Huntsville, AL, USA).
Electrophysiological records were analysed using Igor Pro software (version 3, Wavemetric, Inc.). Single exponentials were fitted to the decay phase of the macroscopic currents with the Levenberg-Marquadt least-squares algorithm. The time constant was taken as an estimate of the rate of receptor desensitization. For a few of the mutants the currents could not be well fitted with a single exponential and so only the initial phase of decay was fitted. Statistical analyses were performed with Students unpaired t test using InStat software (v2.01; GraphPad, San Diego, CA, USA). Data were assumed to be normally distributed and the difference between the means was significant when P was less than 0.05. Averaged data are presented as means ±s.e.m.; n is the number of times the experiment was repeated.
Previously, we have shown that the P2X2(b) splice variant, expressed in either Xenopus oocytes or HEK 293 cells, exhibits a μ5-fold faster rate of desensitization than the P2X2(a) splice variant in the presence of 100 μm ATP (Simon et al. 1997). These splice variants differ only in the C-terminal region that lies downstream of the Lys369 splice site (Fig. 1A). To investigate the role of the C-terminus in controlling the rate of receptor desensitization we generated a truncated P2X2 receptor with the C-terminus deleted at position 369. This leaves only 16 residues beyond the end of the putative TM2 region. The wild-type and mutant receptors were expressed in Xenopus oocytes and currents evoked by application of 100 μm ATP were measured at −60 mV. Unlike the wild-type receptors, the mutant receptor, P2X2(b)Δ34, desensitized extremely rapidly (Fig. 1B and C). Single exponential functions were fitted to the decay phase of the currents and values for the time constants were 110.8 ± 9.7 s (n= 28) for P2X2(a), 24.5 ± 1.0 s (n= 40) for P2X2(b) and 0.9 ± 0.1 s (n= 14) for P2X2(b)Δ34. The P2X2(b)Δ34 currents had a similar time course to the currents we measured from oocytes expressing the P2X3 receptor (τ= 0.9 ± 0.1 s, n= 4; Fig. 1B). For both receptors, the speed of the response was probably limited by the time taken for the solution to be exchanged around the oocyte.
The rate at which the P2X2(b) and P2X2(b)Δ34 receptors recovered from desensitization was measured by giving repeated 2 min applications of 100 μm ATP, which desensitized the initial response by > 90 %, separated by wash periods of increasing length. The P2X2(b) response recovered fully with a time constant of 4 min whereas the P2X2(b)Δ34 response recovered with a time constant of 18 min (Fig. 1D). We conclude from these experiments that the C-terminal tail of P2X2 receptors dramatically slows the rate of receptor desensitization and also increases the rate of recovery to an activatable state. Thus, it appears to play a crucial role in producing the slowly desensitizing phenotype that is characteristic of the P2X2 subtype.
To identify structural elements within the C-terminal tail which are involved in slowing the rate of desensitization, we made deletions of increasing length within the P2X2(b) C-terminus (Fig. 2). Deleting either nine or 17 amino acids beyond Val370 produced no significant change in the rate of desensitization whereas deleting 21 residues produced a small but significant slowing (τ= 35.2 ± 3.3 s, n= 5. P < 0.05). Surprisingly, deleting 25 amino acids slowed the rate of desensitization still further; the P2X2(b)Δ25 mutant has only nine amino acids downstream of the Lys369 splice site and yet its currents decayed with a time constant similar to that of the P2X2(a) splice variant (τ= 114.8 ± 4.4 s, n= 20). Finally, all amino acids beyond the splice site, except for Val370, were removed. Inward currents through this mutant receptor (P2X2(b)Δ33) decayed with a time constant of 70.4 ± 6.8 s (n= 10). Thus, addition of a single valine onto the C-terminal end of the P2X2(b)Δ34 mutant appears to be sufficient to slow desensitization by at least 70-fold.
The contrasting kinetics of the P2X2(b)Δ34 and P2X2(b)Δ33 mutants suggests that Val370 plays an important role in controlling the rate of P2X2 receptor desensitization. To test this further, we generated four point mutations at the 370 position of the P2X2(a) splice variant. Val370 was either mutated to alanine, leucine, isoleucine or phenylalanine. The V370A mutation, which reduced the size and hydrophobicity of the side group, produced a μ13-fold increase in the rate of desensitization (Fig. 3). The other mutations, which all increased the hydrophobicity of the side group, slowed the rate of desensitization (Fig. 3); the time for the current to decay to 70 % of its peak value was 321.6 ± 41.5 s (n= 4) for P2X2(a)V370L, 280.3 ± 36.3 s (n= 3) for P2X2(a)V370I, and 225.4 ± 40.2 s (n= 3) for P2X2(a)V370F. These were all significantly slower than P2X2(a) (P < 0.05) which desensitized to 70 % of the peak current in 62.7 ± 8.0 s (n= 4). Thus, the rate of desensitization appears to be correlated with the hydrophobicity of the amino acid at this position. However, the exact structure of the side group does not appear to be an important determinant of channel gating.
In the P2X2(b)Δ25 mutant, the eight amino acid region downstream of Val370 significantly slowed the rate of desensitization compared with the P2X2(b)Δ33 mutant (P < 0.0001). However, deleting the identical sequence at the C-terminus of the P2X2(b) splice variant did not change the decay time constant (τ= 26.0 ± 1.5 s, n= 6). This suggests that amino acids that lie in close proximity to Lys369-Val370 exert a greater influence on the rate of receptor desensitization than those further away. P2X2(a) and P2X2(b)Δ25 have in common a Pro-Lys sequence which is three to four residues downstream of Val370 and is not present at the equivalent position of P2X2(b) or the other deletion mutants. Mutating the proline to alanine in P2X2(b)Δ25 caused only a small, albeit significant, increase in the rate of decay (τ= 69.1 ± 5.3 s, n= 10, P < 0.0001). However, mutating the charged residues to alanines, P2X2(b)Δ25,D371A,K373A, caused a > 4-fold increase in the rate of decay (τ= 26.1 ± 1.2 s, n= 12). To test whether or not Lys374 contributes to the slowly desensitizing behaviour of P2X2(a), we substituted alanine. This caused a μ5-fold increase in the rate of desensitization, such that the time course of the currents resembled the P2X2(b) receptor currents (Fig. 4A). Thus, Lys374 appears to be at least partly responsible for the slower desensitization of the P2X2(a) splice variant compared with the P2X2(b) splice variant.
The 16 amino acid region from the end of TM2 to the Lys369 splice site contains several lysines (Fig. 4). To investigate their involvement in slowing desensitization, we substituted either alanine or glutamine. Substituting alanine for Lys369 produced an μ8-fold increase in the rate of desensitization (Fig. 4B). Slightly further upstream is Lys366 which is conserved throughout P2X1-6. Mutating Lys366 to glutamine inhibited the expression of a functional receptor. Mutating the adjacent Lys365 to glutamine dramatically increased the rate of desensitization. Although the entire decay phase was not well fitted with just a single exponential, there was a major fast component which had a time constant of 3.7 ± 0.5 s, (n= 8) (Fig. 4B). Similarly, the K365Q mutation in the P2X2(b) splice variant caused an μ13-fold increase in the rate of desensitization (τ= 1.9 ± 0.2 s, n= 7; Fig. 4C). Substituting the negatively charged glutamic acid for Lys365 did not produce a significant further increase in the rate of decay compared with a glutamine at this position (τ= 1.4 ± 0.2 s, n= 8, P > 0.05; Fig. 4C). However, the time course of the responses was probably limited by the rate of solution exchange around the oocytes. Finally we mutated the two lysine residues closest to the TM2 in P2X2(b). The K358Q,K360Q double mutation caused only a 3-fold increase in the rate of current decay (τ= 7.9 ± 0.8 s, n= 16; Fig. 4C). Interestingly, this double mutation prevented the full recovery of the receptor from the desensitized state. Following a 1 min application of 100 μm ATP, the currents recovered rapidly to μ50 % of the original peak value (τ= 1.6 min) but showed no further recovery, even with wash periods of greater than 20 min (Fig. 4D). This is in contrast to the wild-type P2X2(b) receptor currents which recovered fully with a time constant of μ4 min (Fig. 1D). The P2X2(b)K365Q receptor currents also recovered fully but with a time constant of 6.8 min (Fig. 4D). Thus, all of the lysines within the C-terminal region adjacent to TM2, appear to contribute to slowing the rate of P2X2 receptor desensitization. They also affect the stability of the desensitized states.
To investigate the mechanism by which the C-terminal region slows receptor desensitization, we took advantage of our finding that the nine amino acid region at the C-terminus of P2X2(b)Δ25 (VDPKGLAQL) was as effective at slowing desensitization as the 103 amino acids at the C-terminus of P2X2(a). A synthetic peptide of these nine amino acids was injected into oocytes expressing the P2X2(b)Δ34 mutant to test whether or not this region needs to be covalently attached to the rest of the receptor protein in order to slow the rate of decay. Recordings were made 2 and 6 h after injecting 100 nl of 100 mm peptide. There was no significant change in the time course of the ATP-evoked currents (P > 0.05; τ= 1.3 ± 0.3 s (n= 9) 6 h after injection of peptide and τ= 0.7 ± 0.1 s (n= 4) 6 h after injection of water). This suggests that this region does need to be covalently linked to the rest of the receptor in order to inhibit the desensitization process.
Having shown that the peptide could not directly slow desensitization, we next tested whether or not it could compete with the C-terminal tail of the P2X2(a) receptor and increase the rate with which this receptor desensitizes. Four hours after injection of 100 nl of 100 mm of the peptide into oocytes expressing wild-type P2X2(a), the time constant of desensitization was 74.8 ± 4.8 s (n= 5, Fig. 5), which was significantly faster (P < 0.01) than for oocytes injected with water (τ= 107.3 ± 5.9 s, n= 5). A similar injection of an eight-residue peptide that lacked the N-terminal valine (DPKGLAQL) did not significantly change the rate of P2X2(a) desensitization (τ= 104.5 ± 6.1 s, n= 4, P > 0.05). This suggests that the nine-residue peptide competes with the C-terminal tail of P2X2(a) for binding to an intracellular site which is involved in slowing desensitization, and that the valine is critical for binding.
The results from this study indicate that the C-terminal splice site of the P2X2 receptor is located within a region that is critically involved in regulating the rate of receptor desensitization. Truncating the receptor at the Lys369 splice site dramatically increased the rate of desensitization to give a time course resembling P2X3 receptor currents. Two structural features of this proximal C-terminal region were shown to play an important role in producing the slowly desensitizing phenotype that is characteristic of the P2X2 subtype. One of these was the presence of a hydrophobic amino acid at position 370 and the other was the presence of several lysines within this region.
Although splicing occurs at Lys369, both P2X2(a) and P2X2(b) have a valine at position 370. The first indication that this valine is important in regulating the rate of desensitization was the extreme difference in the time course of the currents carried by the two truncated receptors, P2X2(b)Δ34 and P2X2(b)Δ33. P2X2(b)Δ34 is truncated at Lys369 and P2X2(b)Δ33 at Val370, and the additional valine at the C-terminus caused a > 70-fold reduction in the rate of desensitization. Furthermore, point mutations which substituted either smaller or larger hydrophobic amino acids for Val370 within the P2X2(a) splice variant caused the rate of desensitization to vary byμ70-fold. The rate of desensitization decreased with increasing hydrophobicity but was not dependent upon the precise structure of the side group; the aromatic side group of phenylalanine was as effective as the aliphatic side groups of leucine and isoleucine in slowing desensitization. Finally, the ability of the nine amino acid peptide, VDPKGLAQL, to compete with the C-terminal tail of P2X2(a) was crucially dependent upon the presence of the N-terminal valine. These results suggest that Val370 slows the rate of channel closure by interacting with another hydrophobic region, either on the receptor protein itself or on a different protein that is associated with the receptor. In order for this interaction to be coupled to a decrease in the rate of closure, the valine needs to be covalently linked to the proximal C-terminal region of the receptor protein.
The possibility that another associated protein is involved in regulating desensitization is intriguing. The intracellular location of the proximal C-terminal segment makes it a suitable candidate region for interacting with either internal structural elements or regulatory factors. There is evidence to suggest that internal factors regulate the rate with which P2X2 receptors desensitize. The P2X2 splice variants appear to desensitize more rapidly when expressed in human kidney HEK 293 cells than in oocytes (Simon et al. 1997). Zhou et al. (1998) reported that the rate of desensitization of P2X2 receptors expressed in either Xenopus oocytes or HEK 293 cells varied widely from cell to cell, which suggests that there were variable levels of some intracellular regulatory factor. However, we did not observe a similar variation in the time course of the wild-type receptor currents between batches of oocytes. Zhou & Hume (1998) also reported that in outside-out membrane patches, P2X2 receptors desensitize considerably faster than in cell-attached patches. This could be because of the loss of a soluble factor, or because of disruption of the cytoskeleton. This implies that in the intact cell the action of a putative regulatory factor is to slow receptor desensitization. There is no indication of an endogenous soluble protein acting in a similar way to the nine amino acid peptide to speed receptor desensitization. For P2X1 receptors expressed in HEK 293 cells, an interaction with the actin cytoskeleton appears to slow the rate of desensitization (Parker, 1998). We observed no effect of pretreatment with cytochalasin D on P2X2(a) receptor currents (F. M. Smith, unpublished observations), although we have not demonstrated that this disrupts the oocyte cytoskeleton.
The lysines are the other feature of the proximal C-terminal region that contribute to the slowly desensitizing kinetics of the P2X2 receptor. The relative importance of the different lysines in determining the rate of channel closure was dependent upon their position within this region. Neutralizing Lys365 produced the greatest increase in the rate of current decay, whereas lysines located further upstream or downstream of this position had progressively less of an effect on the desensitization process. In the wildtype P2X2(a) receptor, substituting alanine for Lys374 produced a μ5-fold increase in the rate of desensitization. This lysine, which is not present in the P2X2(b) splice variant, appears to be responsible for the slower rate of desensitization of the longer splice variant. Our results are consistent with the findings of Koshimizu et al. (1998b). They compared the rates of desensitization of wild-type and mutant P2X2 receptors expressed in the rat GT1-7 cell line by monitoring changes in intracellular [Ca2+]. They concluded that four residues present in P2X2(a) but not in P2X2(b), Pro373-Lys374-His375-Pro376, were required for sustained Ca2+ influx through this receptor.
The mechanism by which these lysines slow receptor desensitization is unknown. The pore of P2X2 receptors is lined by the TM2 region (Rassendren et al. 1997; Egan et al. 1998) and Zhou et al. (1998) recently showed that neutralizing an aspartate residue towards the cytoplasmic end of the TM2 in P2X2(a) caused a pronounced increase in the rate of desensitization. The lysines within the proximal C-terminal region might be involved in long-range electrostatic interactions with the aspartate in TM2 to inhibit pore closure. Alternatively, if another protein is involved in regulating receptor desensitization, then long range electrostatic interactions might promote the association of this other protein with Val370.
The lysines also affect the stability of the desensitized state. Neutralizing the lysines nearest to the TM2, (K358, K360) revealed at least two desensitized states, one of which was extremely stable, such that μ50 % of the initial peak current could not be recovered even after a prolonged wash period. These lysines lie closest to the TM2 region and substitutions here might exert a more direct, steric effect on the desensitizing conformational change.
For ligand-gated ion channels, functional diversity has been shown to be increased by alternative splicing (Sommer & Seeburg, 1992; Wafford & Whiting, 1992; Hollmann et al. 1993; Hope et al. 1993; Werner et al. 1993; Downie et al. 1994; Zukin & Bennett, 1995). For the P2X2 receptor, the C-terminal splice site occurs within a region that appears to have an important functional role in controlling the rate at which the ATP-bound receptor closes. The differences in the rate of desensitization of P2X2(a) and P2X2(b) appear to be much less than the differences between P2X2 and P2X1 or P2X3 (Valera et al. 1994; Brake et al. 1994; Chen et al. 1995; Lewis et al. 1995). It seems unlikely that either splice variant would display significant desensitization in response to synaptically released ATP. Whether or not they are exposed to more prolonged periods of ATP, as a result of tissue injury for example, is not yet fully established. However, the rate at which P2X receptors desensitize and the rate at which they recover from desensitization does appear to be regulated by external and internal factors other than ATP (Khiroug et al. 1997; Parker, 1998; Cook et al. 1998). Therefore the physiological importance of P2X receptor desensitization remains an open question. In this study we have shown that an intracellular region of the P2X2 receptor that lies adjacent to the TM2 region and contains the splice site can determine whether the receptor desensitizes over a period of many tens of seconds or in less than a second. The possibility that this region interacts with regulatory factors that dramatically alter the rate of channel gating needs to be investigated further.