Previous reports on heterologously-expressed human P2Y11 receptors have indicated that ATP, but not UTP, is an agonist stimulating both phosphoinositidase C and adenylyl cyclase. Consistent with these findings, we report that in 1321N1 cells expressing human P2Y11 receptors, UTP stimulation did not lead to accumulation of inositol(poly)phosphates under conditions in which ATP gave a robust, concentration-dependent effect. Unexpectedly, however, both UTP and ATP stimulated increases in cytosolic Ca2+ concentration ([Ca2+]c), with both nucleotides achieving similar EC50 and maximal responses. The responses to maximally effective concentrations of ATP and UTP were not additive. The [Ca2+]c increase in response to UTP was less dependent on extracellular Ca2+ than was the response to ATP. AR-C67085 (2-propylthio-β,γ-difluoromethylene-d-ATP, a P2Y11-selective agonist), adenosine 5′-O-(3-thiotriphosphate), and benzoyl ATP were all full agonists with potencies similar to those of ATP and UTP. In desensitization experiments, exposure to ATP resulted in loss of the UTP response; this response was more sensitive to desensitization than that of ATP. Pertussis toxin pretreatment attenuated the response to UTP but left the ATP response unaffected. The presence of 2-aminoethyl diphenylborate differentially affected the responses of ATP and UTP. No mRNA transcripts for P2Y2 or P2Y4 were detectable in the P2Y11-expressing cells. We conclude that UTP is a Ca2+-mobilizing agonist at P2Y11 receptors and that ATP and UTP acting at the same receptor recruit distinct signaling pathways. This example of agonist-specific signaling is discussed in terms of agonist trafficking and differential signal strength.
Currently, seven members of the mammalian P2Y family of G protein-coupled receptors for nucleotides have been cloned. These display differential sensitivity to native nucleotide agonists (Communi et al., 1997, 2001; Fredholme et al., 1997; Boarder and Hourani, 1998; Hollopeter et al., 2001; Zhang et al., 2001). The human P2Y11 receptor (hP2Y11) is coupled to both phosphoinositidase C and adenylyl cyclase and has been described as a receptor for ATP, the most potent native nucleotide (Communi et al., 1997; 1999; van der Weyden et al., 2000; Qi et al., 2001). In these studies, ATP and analogs displayed greater potency and efficacy at the hP2Y11 receptor than the equivalent diphosphates. Interestingly however, for the canine P2Y11 receptor, the reverse is the case: ADP is more effective than ATP (Qi et al., 2001; Zambon et al., 2001; Torres et al., 2002). Pyrimidine nucleotides such as UTP have been reported to be ineffectual at hP2Y11 receptors, based on their inability to stimulate adenylyl cyclase or inositol (poly)phosphate production (Communi et al., 1997; Zambon et al., 2001).
It has been assumed that transfected P2Y11 receptors couple to increased cytosolic Ca2+ concentration ([Ca2+]c) through stimulation of phosphoinositidase C, and subsequent mobilization of Ca2+ from inositol 1,4,5-trisphosphate [Ins(1,4,5)P3]-sensitive stores. We studied the inositol(poly)phosphate and [Ca2+]c responses of 1321N1 cells transfected with hP2Y11 receptors. We have found that UTP is a Ca2+-mobilizing agonist, with a potency and maximal response similar to ATP. Unlike ATP however, UTP did not lead to increased inositol(poly)phosphate production in [3H]inositol-labeled cells. The responses to the two nucleotides show other distinguishing features, indicating that they activate different pathways to Ca2+ mobilization. These results are discussed in terms of agonist-specific signaling, where agonists acting at the same receptor recruit different signaling pathways.
Materials and Methods
hP2Y11-transfected 1321N1 and untransfected control cells were a kind gift from Dr. T. K. Harden, (University of North Carolina, Chapel Hill, NC). AR-C67085 was donated by AstraZenica (Loughborough, UK). Fura 2 acetoxymethyl ester was purchased from Calbiochem (Nottingham, UK). Nucleotides, pertussis toxin, and 2-aminoethyl diphenylborate (2-APB) were from Sigma (, UK). Biotaq was purchased from Bioline (London, UK), Superscript II reverse transcriptase, RNase-free DNase I, oligonucleotides, and TRIzol were all obtained from Invitrogen (Glasgow, UK).
1321N1 neuroblastoma cells were routinely cultured in DMEM supplemented with fetal calf serum (10%), penicillin (50 U/ml), and streptomycin (50 μg/ml) unless otherwise indicated in the text. For the total [3H]inositol (poly)phosphates (InsPx) assay, cells at 80 to 90% confluence in 24-well plates were labeled for 24 h at 37°C in 5% CO2 with [myo-2-3H]inositol (0.037 MBq/ml; 1 μCi/ml; 0.5 ml/well) in serum-free low-inositol medium M199. After 10 min of preincubation with LiCl (10 mM) at 37°C, agonists were added for 20 min. Inositol(poly)phosphates were extracted into 0.5 M trichloroacetic acid; after extraction with tri-n-octylamine/1,1,2-trichlorofluoroethane, inositol(poly)phosphates were purified on small Dowex-AG1 × 8 400 mesh (Cl–) columns. Data are expressed as disintegrations per minute of [3H]InsPx per well.
For measurements of [Ca2+]c, cells were grown on glass coverslips to ∼40% confluence and then maintained for 24 h in serum-free DMEM. For some experiments, pertussis toxin (100 ng/ml) was included for 24 h. Cells were loaded for 60 min with 3 μM fura 2-acetoxymethyl ester in Krebs-HEPES buffer (KHB; 10 mM HEPES, pH 7.4) containing 1% bovine serum albumin and 0.025% pluronic acid F127 at room temperature in the dark. Coverslips were washed twice with KHB, placed in a 100-μl closed chamber on the microscope stage, and continuously perfused at 1 ml/min. Perfusion with agonists typically lasted 30 s, and responses were monitored with a VisiTech imaging system. (VisiTech International, Sunderland, UK). For calcium-free experiments, coverslips were perfused for 2 min with KHB-Ca2+ (nominally Ca2+-free buffer) before the addition of agonists. Data (maximal response above basal) from 8 to 10 cells were pooled for each experiment; the results presented are pooled across three separate experiments, except where otherwise indicated. [3H]InsPx and fura-2 experiments were undertaken on the same days and on the same batches of cells.
For the cyclic AMP assay, cells in 24-well plates were maintained serum-free in DMEM (24 h). A 10-min preincubation at 37° with isobutylmethylxanthine (300 μM) was followed by 5-min stimulation at 37°C with forskolin (10 μM), ATP (300 μM), and UTP (300 μM) as indicated. The incubation was stopped with 0.5 M trichloroacetic acid, followed by extraction with tri-n-octylamine/1,1,2-trichlorofluoroethane. After neutralization with NaHCO3, cyclic AMP levels were measured using the protein binding assay (Brown et al., 1971). Data, pooled across three experiments each performed in triplicate, are presented as picomoles of cyclic AMP per well (mean ± S.E.M.).
For the reverse transcriptase polymerase chain reaction (RT-PCR), 1321N1 wild-type and P2Y11-transfected 1321N1 cell lines were cultured as described above. RNA was prepared (n = 3) from 5 × 106 cells using TRIzol according to the manufacturer's instructions and was treated with RNase-free DNase I (200 U) for 30 min at 37°C and then re-extracted with TRIzol. First-strand cDNA was prepared from 5 μg of RNA using random hexamers and Superscript II reverse transcriptase in a 40-μl reaction volume according to the manufacturer's instructions. PCR reactions for the human P2Y2, P2Y4, and P2Y11 employed the following primer pairs (forward, reverse) designed to amplify partial cDNAs from each sequence: P2Y2, 5′-GACTGCTAAAGCCAGCCTACGGGAC-3′, 5′-CCTGAAGTCCTCACTGCTGCCCAAC-3′, 409 bp; P2Y4, 5′-ACCCTATGGCTCTTCATCTTCCGCC-3′, 5′-AACAAGAGTGACCAGGCAGGGCACG-3′, 488 bp; and P2Y11, 5′-GGCTGGCGGCCTACAGAGCGTATAG-3′, 5′-CTCTGGGTTCCAGCTGTCCCTGTAG-3′, 442 bp. PCR reactions were performed (30 cycles of 94°C for 30 s; 65°C for 30 s, and 72°C for 30 s) using 1.25% (v/v) of each first-strand cDNA or 20 ng of human genomic DNA as a positive control with 2.5 units of Biotaq in a 25-μl amplification containing 1.5 mM MgCl2 according to the manufacturer's instructions. Reverse transcriptions were also performed in the absence of the enzyme as a control for contaminating DNA and resulted in no amplification products when used in PCR, as did PCR reactions performed in the absence of added template. Amplicons (10 μl) were subjected to gel electrophoresis on a 2% (w/v) agarose gel.
Agonist purity was monitored by high-performance liquid chromatography analysis. A creatine phosphokinase regenerating system (10 mM creatine phosphate and 20 U/ml creatine phosphokinase), which maintains UTP and ATP concentrations, was included in some experiments. Data were analyzed using Prism v3.1 (GraphPad Software, San Diego, CA).
[3H]Inositol-labeled 1321N1 cells expressing hP2Y11 receptors responded to increasing concentrations of ATP with a concentration-dependent increase in [3H]InsPx accumulation (log EC50 = –4.9 ± 0.13) as shown in Fig. 1A. UTP, over the same concentration range, gave no response, consistent with expectations from the literature (Fig. 1A).
Unexpectedly, however, fura-2-loaded 1321N1-hP2Y11 cells stimulated with either ATP or UTP responded with similar concentration-dependent increases in [Ca2+]c (Fig. 1B). Although within the same range, the concentration-response curves derived from these responses were significantly different (P < 0.05 by two-way ANOVA). The log EC50 values for ATP and UTP were –5.57 ± 0.11 and –5.21 ± 0.10, respectively. The maximal response (340/380 ratio) to ATP (0.85 ± 0.03) was slightly greater than that to UTP (0.71 ± 0.03). We then conducted an equivalent series of experiments with ATP and UTP in the presence of a nucleotide triphosphate regenerating system. This served to purify the agonists by eliminating contaminating diphosphates and also to prevent interconversion during the stimulation period. The concentration-response curves were essentially unchanged (data not shown).
Figure 2 shows typical experimental traces from individual fura-2-loaded cells stimulated with ATP and UTP in the presence of a nucleotide triphosphate-regenerating system. The two nucleotides elicited similar increases in [Ca2+]c, with an immediate rapidly rising phase, followed by a plateau sustained for the duration of the agonist application. All cells were found to respond. The increase in pooled peak height with increasing agonist concentration depicted in Fig. 1B was caused by an increase in peak response of individual cells, not by recruitment of more cells, as seen in Fig. 2. Considerable overlap in the relative size of responses of individual cells to ATP and UTP was observed (Fig. 2). In control experiments, 1321N1 cells not expressing P2Y11 receptors did not respond to nucleotides either with increases in [Ca2+]c or inositol(poly)phosphates (data not shown).
In a further series of experiments, concentration-response curves were constructed to ATP and three derivatives: benzoyl ATP, ATPγS, and AR-C67085. As shown in Fig. 3, all were potent and full agonists. In [3H]InsPx experiments, AR-C67085 was found to be the most potent agonist (data not shown). These results, together with the data presented in Fig. 1A, establish that the receptors expressed in these transfected 1321N1 cells have the characteristics of hP2Y11 receptors.
To investigate whether the [Ca2+]c increases elicited by ATP and UTP reflected mobilization from intracellular stores, cells were stimulated in the absence of added extracellular Ca2+. Figure 4, A and B, show that in nominally Ca2+-free medium, there was a substantial Ca2+ response to both ATP and UTP. The response to ATP was significantly reduced (P < 0.05) when no extracellular Ca2+ was added, whereas that to UTP was not significantly affected (by two-tailed unpaired t test of responses from three separate experiments of 8–10 cells each). These results are consistent with mobilization of intracellular stores. In addition, they indicate that there is a substantial contribution by extracellular Ca2+ to the response to ATP, but not to UTP.
To determine whether the maximal responses to ATP and UTP were additive, cells were stimulated with maximally effective concentrations of ATP and UTP (100 μM) either separately or simultaneously. As shown in Fig. 4C, the response to combined addition of ATP and UTP was indistinguishable from that to ATP alone. In a variation of this experimental design, cells were perfused with 100 μM ATP until the response had subsided (∼4 min), and then UTP was added in the continued presence of ATP. Under these conditions, no response to 100 μM UTP was recorded (data not shown).
We then investigated whether these responses desensitize with repeated agonist administration. In these experiments, the stimulation period was extended to 4 min with intervals of 1 min between agonist applications. A typical result is shown in Fig. 5. The response to ATP was desensitized after one or more 4-min perfusions with this nucleotide. The response to UTP was also lost (Fig. 5A). However, after application of UTP, the response to ATP recovered. Figure 5B shows that this recovery was caused by time elapsed since the end of the last ATP perfusion (6 min) and was not dependent on the presence of UTP. Responses to UTP were also lost after a 4-min perfusion; however, this did not attenuate a subsequent response to ATP (Fig. 5C). When three applications of UTP were followed by a 6-min perfusion with buffer only, the UTP response did not recover (Fig. 5D). This contrasts with the result with ATP (Fig. 5B), which did recover under the same experimental conditions.
Pertussis toxin pretreatment was found to have differential effects on the [Ca2+]c response of 1321N1-hP2Y11 cells to ATP and UTP. As illustrated in Fig. 6, the concentration-response curve to ATP was unaffected by pertussis toxin pretreatment (100 ng.ml for 24 h). The concentration-response curves were not significantly different (two-way ANOVA), and the maximal response of pretreated cells was 97.8% of untreated cells. By contrast, the response to UTP was reduced by pertussis toxin. The maximal response was down to 43.2% of control value and the depressant effect on the concentration-response curve was highly significant (P < 0.0001 by two-way ANOVA, from three separate experiments with 8–10 cells in each).
2-APB has been shown to affect Ca2+ responses involving Ins(1,4,5)P3 receptors and store-operated Ca2+ channels (Ma et al., 2000; van Rossum et al., 2000; Gregory et al., 2001). We investigated whether this compound had a differential effect on the Ca2+ response to ATP and UTP. Figure 7, A and B, shows responses to ATP in the absence and presence of 2-APB (40 μM), respectively. The responses are unaffected by 2-APB. In contrast, the response to UTP (Fig. 7C) is substantially diminished by 2-APB (Fig. 7D). Pooled across three experiments measuring peak Ca2+ responses (340/380 ratios), results were: ATP alone, 1.043 ± 0.036; ATP plus 2-APB, 0.985 ± 0.102; UTP alone, 0.914 ± 0.042; UTP plus 2-APB, 0.330 ± 0.062 (significantly different from UTP alone, P < 0.001, unpaired t test).
To determine whether UTP stimulation of hP2Y11 receptors regulates adenylyl cyclase, we measured cyclic AMP levels in cells in the presence of the phosphodiesterase inhibitor isobutylmethylxanthine. Stimulations with nucleotides (300 μM) or forskolin (10 μM) were for 5 min at 37°C. Expressed as picomoles per well, results from three separate experiments were: control, 1.58 ± 0.26; ATP, 65.8 ± 19.7; forskolin, 15.9 ± 2.5; UTP, 1.78 ± 0.26; UTP with forskolin, 14.0 ± 0.29. The results confirm that ATP stimulates cyclic AMP accumulation. Contrasting with this UTP had no effect on either basal or forskolin-elevated cyclic AMP levels.
RT-PCR studies were undertaken to investigate the possibility that the P2Y11-transfected 1321N1 cells also express mRNA transcripts for P2Y2 or P2Y4 receptors. RNA from the wild-type and P2Y11-transfected cells was extracted and analyzed using primers specific for hP2Y11, hP2Y2, and hP2Y4 encoding sequences. Genomic DNA was used to provide a positive control in these experiments. The results in Fig. 8 show that no P2Y-encoding transcripts were detected in the extracts from untransfected cells. In extracts from P2Y11-expressing cells, there were no transcripts for either P2Y2 or P2Y4 receptors under conditions in which there was a very strong signal for the presence of P2Y11 transcripts.
In this report, we show that both ATP and UTP are agonists at transfected hP2Y11 receptors when the response measured is elevated [Ca2+]c. This is surprising given the considerable interest shown in P2Y11 as an adenosine nucleotide-selective purinoceptor. Furthermore, we provide evidence that the two agonists, acting on one receptor population, increase [Ca2+]c by different pathways. This may be caused by agonist-directed trafficking, in which the induction of agonist-specific receptor conformations activates two pathways from the same receptor, each leading to Ca2+ mobilization and entry.
With respect to the first of these 2 issues, we are not aware of any published reports that describe the effect of UTP on [Ca2+]c in cells heterologously expressing cloned P2Y11 receptors. Here, we report a robust and reliable [Ca2+]c response to UTP. To confirm that the response recorded was caused by UTP, some experiments were carried out in the presence of a nucleotide triphosphate-regenerating system. In addition, the Ca2+ measurements were carried out using rapid perfusion of a low density of cells. Together these protocols effectively eliminate the involvement of agonist interconversion in the response. High-performance liquid chromatography analysis of agonist stocks also ruled out the involvement of any impurity. We therefore believe that the response measured is to UTP acting directly at the cell surface.
The data we report here are consistent with responses at transfected P2Y11 receptors but not at other known P2Y receptors. For example, we found that UTP stimulation did not generate [3H]InsPx in [3H]inositol-loaded cells, and AR-C67085 was more potent than ATP in leading to accumulation of [3H]InsPx, consistent with results reported by Communi et al. (1997, 1999). When considering increases in [Ca2+]c, ATP, ATPγS, benzoyl ATP, and AR-C67085 were all found to be full and potent agonists. Although AR-C67085 is an antagonist at the P2Y12 receptor, it displays agonist activity only at the P2Y11 receptor. These data are inconsistent with the involvement of P2Y2 receptors, at which UTP and ATP are equally effective at generating a phosphoinositidase C response. In addition, benzoyl ATP and AR-C67085 are not effective agonists at the P2Y2 receptor. The data presented are also not consistent with P2Y4 receptors, at which ATPγS is not an effective agonist (our unpublished data).
Our results cannot be explained by coexpression of other P2Y receptors induced in our cell cultures by the presence of P2Y11 receptors. Not only is the pharmacology inconsistent with any known P2Y receptors apart from P2Y11, as indicated above, but the increases in [Ca2+]c in 1321N1 cells expressing P2Y receptors depend upon the activation of the phosphoinositidase C/Ins(1,4,5)P3 pathway. If, for example, P2Y2 were expressed alongside the P2Y11 receptors in our cells, the [Ca2+]c profiles might display some of the characteristics reported here, but there would also be a [3H]InsPx response, as seen in parallel experiments with P2Y2-transfected 1321N1 cells. To directly explore the possibility of P2Y2 or P2Y4 expression in these cells, we undertook RT-PCR studies, using primers for P2Y2, P2Y4, and P2Y11, and found no transcripts for P2Y2 and P2Y4 under conditions that detect a very strong signal for P2Y11 transcripts. We conclude, therefore, that these results cannot be explained by UTP acting at coexpressed P2Y2 or P2Y4 receptors.
UTP and ATP both generate Ca2+ transients typical of a phosphoinositidase C-coupled receptor, with a rapid peak and slower decline toward baseline. However, the response to UTP was not significantly changed in nominally Ca2+-free medium, compared with a substantial attenuation of the response to ATP. This indicates that the two agonists activate different signal transduction pathways. When ATP and UTP were applied together at a maximally effective concentration, the response was no greater than when one agonist was applied alone, consistent with the notion that the receptor has a single site at which the two agonists bind (Erb et al., 1995; Qi et al., 2001).
The possibility that ATP and UTP, acting on the P2Y11 receptor, raise cytosolic Ca2+ by different mechanisms is also suggested by the observation that ATP, but not UTP, stimulates detectable Ins(1,4,5)P3 formation (current data; Communi et al., 1997). We cannot exclude the possibility that UTP activates a small and transient increase in Ins(1,4,5)P3 synthesis that went undetected in these experiments. Alternatively, UTP may use an alternative pathway, such as cyclic adenosine 5′-diphosphoribose or the recently proposed NADPH pathway (Genazzani and Gallione, 1997; Churchill and Gallione, 2001). There have been previous reports of apparently inositol phosphate-independent increases in [Ca2+]c in response to activation of native P2Y receptors (e.g., Frelin et al., 1993; Albert et al., 1997; White et al., 2000).
The desensitization experiments are also consistent with two agonists acting at one receptor to generate different outcomes. The results illustrate that the UTP response is more sensitive to desensitization than is the response to ATP. One possible explanation is that ATP has a higher efficacy than UTP. The residual signal to the Ca2+ response after desensitization may then be greater for ATP than for UTP and sufficient to generate a substantial response. However, it may be that the two agonists activate distinct pathways that vary in their ability to be down-regulated by desensitization protocols.
The experiments described here with pertussis toxin and 2-APB confirm that the two nucleotide agonists activate different signaling pathways to Ca2+ mobilization. 2-APB was originally reported to be an effective antagonist at Ins(1,4,5)P3 receptor channels (Ma et al., 2000; van Rossum et al., 2000). We therefore used this compound as a tool for investigating the involvement of Ins(1,4,5)P3 receptors in the responses to ATP and UTP. Our hypothesis was that the [Ca2+]c increase evoked by ATP would be blocked by 2-APB, whereas the UTP response, occurring in the absence of detectable inositol phosphate formation, would not. However, the reverse was true; the UTP response was blocked by 2-APB, and the ATP response was unaffected. The literature reveals complexities in the action of 2-APB. In cardiac myocytes, agonist-stimulated Ins(1,4,5)P3-dependent Ca2+ mobilization was unaffected by acute exposure to 2-APB (Gysembergh et al., 1999); in hepatocytes, 2-APB has been shown to leave Ins(1,4,5)P3 mediated Ca2+ mobilization intact but blocks vasopressin and thapsigargin-stimulated Ca2+ influx (Gregory et al., 2001). A clear proposal for the mechanism of action of 2-APB in our experiments is difficult to formulate. Despite this, these results confirm that ATP and UTP elicit a Ca2+ response by different signaling pathways.
The pertussis toxin experiments also provide a clear distinction between the 2 agonists: UTP is sensitive, but ATP is not, reflecting earlier reports in different cells of pertussis toxin sensitivity of UTP responses (e.g., Purkiss et al., 1994). This indicates that the G protein coupling of the P2Y11 receptor is agonist-specific; when activated by UTP; the receptor is coupled to the Gi/o family of G proteins. The observations on cyclic AMP accumulation are of interest here, because failure to stimulate basal levels or inhibit forskolin-stimulation suggests the UTP-activated receptor is not effectively coupled to either Gs or GI, so the pertussis toxin-sensitive Ca2+ response is likely to be through Go. When activated by ATP, the Ca2+ response is pertussis toxin-insensitive, probably through Gq. Earlier work has shown that P2Y11 receptors (when stimulated with ATP but not UTP) also couple to Gs (Communi et al., 1997, 1999; Torres et al., 2002). Taken together, this leads to the interesting hypothesis that the P2Y11 receptor may show agonist-directed promiscuity in coupling to three G proteins: Go coupling may occur with UTP stimulation, whereas ATP elicits receptor conformations that preferentially couple to Gq and Gs.
Kenakin (1995) has pointed out that the two explanations for differential activation of signaling pathways by different agonists are different strengths of signal or agonist trafficking. In the former explanation, a high-efficacy agonist may have the potential to activate both tightly and loosely coupled G proteins, whereas a low-efficacy agonist will activate only the more efficiently coupled G proteins. The result will be differential recruitment of signaling pathways, not because of distinct receptor conformations, but because the lower efficacy agonist produces insufficient activated receptor to produce a significant response from the less efficiently coupled G protein. This promiscuity is fundamentally different from agonist trafficking with agonist-specific receptor conformations, each of which may show fidelity in coupling to G proteins. In this article, we have presented clear evidence for agonist-specific signaling at the P2Y11 receptor, but our results do not permit a clear distinction between these two explanations for the differential activation of P2Y11 receptors by ATP and UTP.
We thank Abigail Knight for technical assistance with some of this work and Dr. Jane Dixon for help in manuscript preparation.
This work was funded by the Wellcome Trust.
ABBREVIATIONS: [Ca2+]c, cytosolic Ca2+; AR-C67085, 2-propylthio-β,γ-dichloromethylene-d-ATP; 2-APB, 2-aminoethyl diphenylborate; DMEM, Dulbecco's modified Eagle's medium; InsPx, inositol polyphosphates; KHB, Krebs-HEPES buffer; RT-PCR, reverse transcriptase-polymerase chain reaction; Bp, base pairs; ANOVA, analysis of variance; ATPγS, adenosine 5′-O-(3-thiotriphosphate); Ins(1,4,5)P3, inositol 1,4,5-trisphosphate; AR-C67085, 2-propylthio-β,γ-difluoromethylene-d-ATP.
- Received October 28, 2002.
- Accepted March 6, 2002.
- The American Society for Pharmacology and Experimental Therapeutics