Abstract
Agonist properties of the P2X7 receptor (P2X7R) differ strikingly from other P2X receptors in two main ways: high concentrations of ATP (> 100 μM) are required to activate the receptor, and the ATP analog 2′,3′-O-(4-benzoyl-benzoyl)ATP (BzATP) is both more potent than ATP and evokes a higher maximum current. However, there are striking species differences in these properties. We sought to exploit the large differences in ATP and BzATP responses between rat and mouse P2X7R to delineate regions or specific residues that may be responsible for the unique actions of these agonists at the P2X7R. We measured membrane currents in response to ATP and BzATP at wild-type rat and mouse P2X7R, at chimeric P2X7Rs, and at mouse P2X7Rs bearing point mutations. Wild-type rat P2X7R was 10 times more sensitive to ATP and 100 times more sensitive to BzATP than wild-type mouse P2X7R. We found that agonist EC50 values were determined solely by the ectodomain of the P2X7R. Two segments (residues 115-136 and 282-288), when transposed together, converted mouse sensitivities to those of rat. Point mutations through these regions revealed a single residue, asparagine284, in the rat P2X7R that fully accounted for the 10-fold difference in ATP sensitivity, whereas the 100-fold difference in BzATP sensitivity required the transfer of both Lys127 and Asn284 from rat to mouse. Thus, single amino acid differences between species can account for large changes in agonist effectiveness and differentiate between the two widely used agonists at P2X7 receptors.
The P2X7 receptors belong to a family of cation-permeable membrane proteins gated by extracellular ATP. They can be distinguished from other family members (P2X1-P2X6)by several properties. First, receptor activation is followed in several seconds by the appearance of a permeation pathway that allows passage of molecules up to 900 Da (North, 2002). Second, activation by extracellular ATP rapidly engages a series of cytoskeletal and mitochondrial alterations, which include actin/α-tubulin rearrangements, phosphatidylserine translocation, mitochondrial swelling and loss of mitochondrial membrane potential, and membrane blebbing (MacKenzie et al., 2001, 2005; Le Feuvre et al., 2002; Morelli et al., 2003; Verhoef et al., 2003; Pfeiffer et al., 2004; Elliott et al., 2005; Ferrari et al., 2006). Third, the P2X7 receptor in immune cells of monocyte/macrophage lineage becomes upregulated and functionally active in response to inflammatory stimuli (Guerra et al., 2003; Ferrari et al., 2006); its activation there engages cascades that culminate in processing and release of interleukin-1β, release of tumor necrosis factor α, and activation of nuclear factor-κB (North, 2002; Ferrari et al., 2006). Finally, studies using mice in which the P2X7 receptor has been deleted further support a role in inflammatory processes (Solle et al., 2001; Labasi et al., 2002; Chessell et al., 2005).
P2X7 receptors can also be readily distinguished from other family members when membrane ionic current is measured directly. First, they are more potently inhibited by extracellular calcium and/or magnesium. Second, they are unusually insensitive to ATP; the EC50 value (>300 μM) is approximately 100-fold greater than for other P2X receptors (North, 2002). Third, the ATP analog 2′,3′-O-(4-benzoyl-benzoyl)ATP (BzATP) is considerably more potent that ATP itself at P2X7 receptors, whereas at other P2X receptors it is less potent (North, 2002; Baraldi et al., 2004). Fourth, certain antagonists are selective for the P2X7 receptor, although few have been studied at the primary effect of membrane current (Humphreys et al., 1998; Jiang et al., 2000).
There are marked species differences in agonist and antagonist pharmacology for P2X7 receptors. The effectiveness of BzATP relative to ATP, which often has been used as the primary distinguishing feature of the P2X7 receptor, is not equally reliable among species (Surprenant et al., 1996; Chessell et al., 1998a,b; Hibell et al., 2000; Young et al., 2006). For example, isoquinolone derivatives such as KN-62 and KN-04 block human P2X7 receptors with low nanomolar affinity but are without effect at rodent P2X7 receptors, even at high micromolar concentrations (Humphreys et al., 1998; Baraldi et al., 2004). Conversely, Brilliant Blue G is 20 times more potent at rat than human P2X7 receptors (Jiang et al., 2000).
The differences in properties between the P2X7 receptors of different species are important for several reasons. First, quite different polymorphisms have been described in the human and mouse P2X7 receptors (Adriouch et al., 2002; Gu et al., 2004; Cabrini et al., 2005; Shemon et al., 2005); these may affect their properties in different ways. Second, pharmacological characterizations and structure-function studies of P2X7 receptors have mainly been obtained from heterologous expression of the rat ortholog; such studies will be of limited value if properties of agonists and antagonists are very different from human. Third, increasing the use of P2X7 receptor knockout mice as animal models of disease (e.g., neuropathic and inflammatory pain, Chessell et al., 2005; osteopenia, Li et al., 2005; and joint inflammation, Labasi et al., 2002) drives the need for a more precise knowledge of the properties of this receptor. In general, information concerning residues involved in agonist action at P2X receptors is extensive for P2X1, P2X2, and P2X4 receptors but minimal for P2X7 receptors (North, 2002; Vial et al., 2004; Zemkova et al., 2004).
The purpose of the present study was to identify residues in P2X7 receptors that may be involved in agonist action. We recently noticed that the sensitivity to BzATP was different between rat and mouse P2X7 receptors (Young et al., 2006). Of the 595 amino acids of the P2X7 receptors, 88 are different between mouse and rat (Fig. 1A). In the present study, we have investigated which of these might be responsible for the differences in effectiveness of ATP and BzATP. We have identified two residues in the ectodomain of P2X7R that can account for the differential agonist sensitivities: residue 127, which primarily influences BzATP affinity; and residue 284, which can fully account for ATP sensitivity.
Materials and Methods
Cell Culture, Transfection, and Site-Directed Mutagenesis. Both rat and mouse P2X7 constructs (Surprenant et al., 1996; Chessell et al., 1998b) were subcloned in the same expression vector background (pcDNA3; Invitrogen, Paisley, UK) and bore C-terminal glutamic acid-glutamic acid epitope tags (EYMPME) to allow detection of protein expression by Western blotting. The 3′- and 5′-noncoding regions of the mouse P2X7 construct was engineered to be identical with that of the rat P2X7 construct to eliminate any differences in expression as a result of noncoding sequences. Point mutations were generated from the above constructs using the PCR overlap extension method and Accuzyme proof-reading DNA polymerase (Bioline, London, UK). Single chimeras (Fig. 1B) were produced using 21-nucleotide synthetic oligonucleotides designed with an inframe 9-nucleotide 5′-adapter tail to introduce overlapping sequences to fuse chimeras between rat and mouse P2X7R sequences. These oligonucleotides were used in combination with the T7 sense and BGH antisense oligonucleotides annealing in the pcDNA3 expression vector sequence. Overlapping amplification products were purified from a 1% agarose gel electrophoresis and used in combination for a second PCR amplification using the T7 sense and BGH antisense oligonucleotides. Three consecutive overlapping PCR amplifications were necessary to produce double chimeras (Fig. 1B). Final T7/BGH amplified products were double-digested with HindIII and XbaI and replaced back in the HindIII-XbaI positions of the original vector. A high concentration of template vector in combination with a proof-reading DNA polymerase and a low number of cycling steps were used in all amplification reactions to minimize random mutations. All subcloned products were confirmed by sequencing (CEQ 2000 Dye Terminator; Beckman Coulter, Fullerton, CA), and protein expression was verified by Western blotting.
Human embryonic kidney 293 cells were transiently transfected using Lipofectamine 2000 (Invitrogen, Paisley, UK). Cells were plated onto 13-mm glass coverslips and maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal calf serum and 2 mM l-glutamine at 37°C in a humidified 5% CO2 incubator. We were concerned that differences in expression levels between constructs (Young et al., 2006) might change the pharmacological properties of the currents. Reducing the rat P2X7R cDNA concentration by 10-fold (from the standard 1to0.1 μg/ml) resulted in approximately 50% reduction in maximum currents to ATP or BzATP without a significant change in agonist EC50 values (Fig. 2A). Therefore, in all subsequent experiments we used the standard (1 μg/ml cDNA) concentration for all transfections.
Protein Solubilization, Deglycosylation, and Western Blotting. Confluent cells were washed with phosphate-buffered saline and pelleted. Cell pellets were lysed in phosphate-buffered saline containing 1% Triton X-100 and antiproteases (Complete; Roche, Lewes, UK) for 1 h at 4°C followed by centrifugation at 16,000g for 10 min to pellet debris. Total protein samples were removed and assayed for protein content using the Bio-Rad Protein Assay kit (Bio-Rad Laboratories, Hemel Hempstead, UK). SDS-polyacrylamide gel electrophoresis sample buffer was added, and the samples were boiled for 2 min at 100°C to denature the protein. Where appropriate, deglycosylation was performed by incubating protein samples (100 μg) for 1 h at 37°C with 500 units of PNGase F (New England Biolabs, Herts, UK) according to the manufacturer's instructions. Samples were separated on 8% polyacrylamide gels according to standard methods and transferred to polyvinylidene difluoride membranes. Western blotting was performed according to standard protocols, and proteins were visualized using anti glutamic acid-glutamic acid primary antibody (Bethyl Laboratories, Cambridge, UK) and horseradish peroxidase-conjugated secondary antibody (Dako UK Ltd., Ely, UK), both at 1:2000 dilution, followed by detection using the ECL-plus kit (Amersham, Bucks, UK) and Kodak Bio-Max MS film (Sigma, Poole, Dorset, UK).
Electrophysiological Recordings. Whole-cell recordings were made 24 to 48 h after transfection using an EPC9 patchclamp amplifier (HEKA Electronik, Lambrecht, Germany). Membrane potential was held at -60 mV. Recording pipettes (5-7 MΩ) were pulled from borosilicate glass (World Precision Instruments, Sarasota, FL) and filled with an intracellular solution that consisted of 145 mM NaCl, 10 mM EGTA, and 10 mM HEPES. The external solution contained 147 mM NaCl, 10 mM HEPES, 13 mM glucose, 2 mM KCl, 2 mM CaCl2, and 1 mM MgCl2. Agonists were applied in divalent-free solution; cells were otherwise super fused with normal external solution. Osmolarity and pH values of all solutions were 295 to 310 mOsM and 7.3, respectively. All experiments were performed at room temperature. Agonists were applied using a RSC 200 fast-flow delivery system (BioLogic Science Instruments, Grenoble, France). Agonists were applied for 5- or 10-s duration to obtain steady-state responses. Concentration-response curves to ATP and BzATP were obtained by first obtaining a maximum response to agonist, because marked run-up of response was observed at both rat and mouse P2X7 receptors, (Surprenant et al., 1996; Chessell et al., 1998b; Young et al., 2006), and then either applying decreasing or increasing concentrations. In either case, similar curves were obtained, provided that a maximum response had been obtained beforehand. Concentration-response curves were plotted using KaleidaGraph (Abelbeck/Synergy Software, Reading, PA) and Prism version 3.0a software (GraphPad Software Inc., San Diego, CA) using the Hill equation provided in Prism.
Results
Comparison of Mouse and Rat P2X7 Receptors. Although all P2X7Rs are potently inhibited by extracellular divalent cations, there is a significant species difference with Mg2+ and Ca2+ being approximately 10-fold more potent to inhibit human than rat P2X7Rs (Surprenant et al., 1996; Rassendren et al., 1997). In preliminary experiments, we found that mouse P2X7R was also more sensitive (by approximately 5-fold) to inhibition by Mg2+ and Ca2+ than was rat P2X7R (data not shown). Therefore, all agonist responses were recorded in the divalent-free cation solution to rule out possible contributions of differential divalent cation sensitivity to agonist concentration responses. We first compared ATP and BzATP concentration-response curves from wild-type rat and mouse P2X7Rs using equal amounts of cDNA for transfection and using a 10-fold lower concentration of rP2X7R cDNA, which we have found previously results in similar protein expression of rat and mouse P2X7Rs (Young et al., 2006). Typical currents recorded from cells transfected with equal cDNA concentrations are shown in Fig. 2A, and results from all experiments are shown in Fig. 2, B and C. Reducing the rP2X7R cDNA concentration by 10-fold resulted in approximately 50% reduction in maximum currents to ATP or BzATP (Fig. 2B) without a significant change in agonist EC50 values or BzATP/ATP maximum current ratio or EC50 ratio (Fig. 2, D and E). The maximum agonist-evoked currents recorded from cells transfected with mP2X7R were approximately the same as those recorded from cells transfected with the 10-fold lower rP2X7R cDNA concentration (Fig. 2, B and C). In agreement with earlier studies on human and rat P2X7Rs (Surprenant et al., 1996; Wiley et al., 1998; Hibell et al., 2000), we found ATP to be a partial agonist relative to BzATP, with maximum BzATP-evoked currents that were 30 to 45% greater than maximum ATP-evoked currents at both rat and mouse P2X7R (Fig. 2D). EC50 values for BzATP and ATP at the rat P2X7R (3.6 and 123 μM, respectively) were several-fold lower than at the mouse P2X7R (285 and 936 μM) (Table 1). These values yield a striking difference in the BzATP/ATP EC50 ratio, which was 34 at the rP2X7R but only 3.3 at the mouse P2X7R (Fig. 2E).
Introducing Segments of Rat P2X7R into Mouse P2X7R. Rat and mouse P2X7R sequences are 84% identical with 88 specific amino acid differences; most of the nonconservative differences are found in two distinct regions of the ectodomain or in the intracellular C-terminal domain (Fig. 1A). Of the 22 nonconserved amino acids in the ectodomain, 8 are found in the region encompassing residues 115 to 136, 4 are found between residues 282 and 288, and the remaining half are scattered throughout the extracellular loop (Fig. 1A). We therefore made a series of chimeric constructs depicted in Fig. 2B to examine the effects of transposing rat ectodomain, C terminus, and residues 115 to 136 and 282 to 288 (alone and in combination) into the mouse P2X7R on agonist-evoked responses.
Neither agonist concentration-response curves nor agonist-evoked kinetics was altered by transposing rat intracellular C terminus onto mouse P2X7R or vice versa (Fig. 3, A and C). Transposition of rat ectodomain onto mouse P2X7R resulted in ATP and BzATP EC50 values that were the same as for wild-type rat P2X7R (Fig. 3C), but deactivation kinetics were several-fold slower (Fig. 3B). No responses were recorded from cells in which the mouse ectodomain was transposed into the rat P2X7R. The chimeric mouse P2X7R containing residues 115 to 136 of rat P2X7R resulted in a pronounced leftward shift of the BzATP concentration-response curve without any alteration in the ATP-evoked responses (Fig. 4A and Table 1). When residues 282 to 288 of rat P2X7R were inserted into mouse P2X7R, there was a small leftward shift in the BzATP concentration-response curve but a large leftward shift in the ATP concentration response (Fig. 4B and Table 1). Substitution of both regions of rat P2X7R into mouse P2X7R resulted in agonist-evoked concentration-response curves and EC50 values that were not significantly different from wild-type rat P2X7R (Fig. 4C and Table 1).
Introducing Single Amino Acids of Rat P2X7R into Mouse P2X7R. We next substituted individually each rat P2X7R residue in these regions into the mouse receptor and examined agonist responses in these mutant receptors. ATP-evoked responses were not different from wild-type mouse P2X7R for any mutation except at residue 284, at which substitution of asparagine (rat P2X7R) for aspartate (mouse P2X7R) resulted in ATP concentration response and EC50 value that was not significantly different from wild-type rat P2X7R (Fig. 5A and Table 1). In contrast, the sensitivity of the mouse receptor to BzATP was clearly increased over wild-type by substitutions of the equivalent rat residue at several positions. These were D284N (p < 0.01), A127K (p < 0.01), S130H (p < 0.1), R134G (p < 0.05), and K136I (p < 0.05). The largest difference was seen for mouse P2X7[A127K], in which case the EC50 value for BzATP (80 μM) was intermediate between that for the rat (3 μM) and mouse (285 μM) receptor (Fig. 5B and Table 1).
The double substitution fully converted the mouse receptor (P2X7[A128K,D284N]) to the sensitivity of the rat receptor with respect to both ATP and BzATP EC50 (Fig. 5A and Table 1) and was the only construct BzATP/ATP potency ratio was the same as that for wild-type rat P2X7R (Fig. 5C). However, we noticed that at this receptor BzATP was not able to produce as great a maximal current as ATP, which is the opposite for either of the wild-type receptors. Maximum BzATP-evoked currents were approximately 30% greater than maximum ATP-evoked currents at both mouse and rat P2X7R (Fig. 2D), but in the mouse P2X7[A128K,D284N] receptor, the ATP-evoked currents were 36 ± 5% (n = 5) greater than the maximal currents evoked by BzATP.
Involvement of N-Glycosylation. The substitution of Asp284 with asparagine at the mouse P2X7R generates a potential N-glycosylation acceptor sequence. The wild-type rat P2X7R also contains a similar potential acceptor sequence at this position (NESL). Because it has been well-demonstrated at other P2XRs that adding N-glycosylation sites significantly increases, whereas removing N-glycosylation sites decreases, protein expression (Newbolt et al., 1998; Torres et al., 1998; Rettinger et al., 2000; Chaumont et al., 2004), we assayed protein expression in wild-type, chimeric, and mP2X7D284N receptor and asked whether this site was, indeed, glycosylated. As found previously (Young et al., 2006), transfection with equal concentrations of rat and mouse P2X7R cDNA yielded protein level ratios of approximately 3:1 (Fig. 6A). The presence of rat P2X7R ectodomain, but not N terminus, C terminus, or transmembrane domains, yielded protein levels not significantly different from wild-type rat P2X7R, whereas the presence of mouse ectodomain yielded low protein expression equivalent to wild-type mouse P2X7R (Fig. 6A). We then removed N-glycan chains from wild-type mouse and rat P2X7R and mouse P2X7-D284N receptor using PNGase F and examined molecular mass by SDS-polyacrylamide gel electrophoresis. Wild-type rat P2X7R, mouse P2X7-D284N receptor, and mouse P2X7R bands were detected at 78, 78, and 75 kDa, respectively (Fig. 6B). After PNGase F treatment, all receptors were detected at approximately 65 kDa (Fig. 6B). This result is consistent with Asn284 being glycosylated in the wild-type rat receptor and the mouse P2X7[D284N] receptor.
Discussion
The sequence relatedness between the mouse and rat P2X7 receptors (84.9%) is substantially less than for the other P2X receptors (for 1 through 6, 97.5, 97.2, 99.2, 94.7, 94.5, and 92.6% identity, respectively). This may underlie the large species difference in agonist sensitivity, which have generally not been described for other (homomeric) members of the P2X receptor family. The differences are somewhat clustered, and in two parts of the ectodomain, the identity is only 68% (115-136) and 43% (282-288) (Fig. 1A, boxes). Unlike other P2X receptors, in which mutations in transmembrane domains have been demonstrated to significantly alter ATP concentration-response curves (Haines et al., 2001), we found agonist potency to be determined solely by the ectodomain of the P2X7 receptor. We then identified two amino acids that are not conserved in any of the other P2X receptors, one in each of these ectodomain segments, that could account for the differential ATP/BzATP agonist sensitivity at the P2X7 receptor.
The first main finding of the present study is that, of all the differences between rat and mouse P2X7 receptor ectodomain residues, a single amino acid is largely responsible for the difference in sensitivity to ATP. Thus, introducing asparagine in place of aspartate at position 284 in the mouse P2X7R changed the EC50 value from 936 to 146 μM, which is close to the value for the wild-type rat receptor (123 μM, Table 1). The change in sensitivity associated with the aspartate-to-asparagine substitution (from -O- to -NH2) is striking. The asparagine is situated within a sequence commonly found at sites of N-linked glycosylation (N-X-S; NESL in rat P2X7; N-glycosylation acceptor sequence in mouse P2X7[D284N]). We found evidence that the mutated mouse P2X7[D284N] receptor was glycosylated at this position, because the molecular mass was approximately 3 kDa higher than the wild-type mouse receptor. The mass corresponded to that of the wild-type rat receptor (78 kDa). We note that the wild-type rat receptor and mouse receptor both carry the same five potential N-linked glycosylation sites (Asn74, Asn100, Asn106, Asn187, Asn241), whereas only rat P2X7R carries a sixth (Asn284) glycosylation site. Thus, the wild-type rat P2X7R and the mouse P2X7[D284N] mutated receptor each carry the same N-linked glycosylation sites. In both cases, treatment with PNGase F reduced the molecular mass to approximately 65 kDa, which is similar to the calculated molecular mass (68.5 kDa) of the receptor. The large effect that this aspartate-to-asparagine substitution has on the potency of ATP could plausibly indicate that the attached sugar moiety participates directly in ATP binding. An alternative explanation is that the glycan impedes the conformational change leading from binding to gating; this seems less likely because it is the form of the P2X7 receptor with the attached sugar at this position (mouse P2X7[D284N] or rat P2X7) that is more sensitive to ATP than the form without it (mouse P2X7). Rettinger et al. (2000) have found previously that, in the case of the P2X1 receptor, glycosylation affects the potency of ATP. Although the difference in potency was small at the P2X1 receptor (approximately 3-fold), it was also the form of the receptor without attached sugar (P2X1[N210Q]) that was less sensitive. In the P2X1 receptor, the asparagine was situated close to the center of the ectodomain (Asn210 of rat P2X1; Roberts and Evans, 2004). The subject asparagine of the present study, at position 284, is situated in region that is poorly conserved among P2X receptors. Thus, the present results indicate that it plays a key role in determining the unique response to ATP observed at the P2X7 receptor.
The second main finding of the present work was that the sensitivity to BzATP was almost 100-fold different between rat and mouse receptors, 10 times greater than the difference for ATP, and this BzATP differential was largely influenced by residue 127 (lysine in rat and alanine in mouse P2X7R). The residue Asp284 also had an effect on the BzATP sensitivity because substitution of aspartic acid for asparagine at this position did increase the potency of BzATP, but by only approximately 3-fold (Table 1). A much greater further increase of approximately 30-fold was observed when the additional A127K mutation was introduced into the mouse receptor. The observation that the effects of ATP and BzATP are differentially affected by two point mutations might imply that the residues are involved directly in agonist binding rather than in the subsequent gating conformational changes (which might be expected to be more in common for distinct agonists). If this is the case, then one might ask why an alanine-to-lysine substitution might affect the action of BzATP but not ATP. The obvious difference between the two agonists is the presence of the two (O-linked) aromatic moieties at the 2′(3′) position. One might speculate that the cationic lysine residue in the binding site might interact with the π electron cloud on one or another of the aromatic rings, providing a contribution to binding energy that would be unique to BzATP.
Irrespective of the detailed mechanism of the differences between ATP and BzATP, the present results provide a stark qualitative reminder of the criticality of species differences in ATP receptor pharmacology. Important differences in antagonist effectiveness have been reported for P2X1 receptors (chick versus human: Soto et al., 2003), P2X4 receptors (human versus rat: Buell et al., 1996; Soto et al., 1996), and for P2X7 receptors (human versus rat: Humphreys et al., 1998; Jiang et al., 2000; Baraldi et al., 2004; and mammalian versus nonmammalian: Lopez-Castejon et al., 2007). However, differences among species in agonist sensitivity have not been widely described. The findings therefore suggest that caution should be exercised in cross-species extrapolation on studies on P2X7 receptors, in which BzATP is in very widespread use as an experimental agonist.
Acknowledgments
We thank E. Martin and L. Collinson for cell and molecular biology technical support.
Footnotes
-
This work was supported by The Wellcome Trust and the Biotechnology and Biological Sciences Research Council.
-
M.T.Y. and P.P. contributed equally to this study.
-
ABBREVIATIONS: BzATP, 2′-3′-O-(4-benzoylbenzoyl) adenosine 5′-triphosphate; PCR, polymerase chain reaction; KN-62, 1-[N,O-bis(5-isoquinolinesulfonyl)-N-methyl-l-tyrosyl]-4-phenylpiperazine; KN-04, N-(1-(p-(5-isoquinolinesulfonyl)benzyl)-2-(4-phenylpiperazinyl)ethyl)-5-isoquinolinesulfonamide.
-
↵1 Current affiliation: Faculty of Life Sciences, University of Manchester, Manchester, United Kingdom.
- Received August 26, 2006.
- Accepted October 10, 2006.
- The American Society for Pharmacology and Experimental Therapeutics